article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
the wetting flows ( where the triple solid - liquid - fluid contact line is present ) are important for many practical applications ranging from the metal coating to the medical treatment of the lung airways .
the contact line statics and dynamics attracted a lot of attention from the scientific community during the last decades .
it became clear that the hydrodynamics in the region of the liquid wedge close to the contact line ( which we will call clr - contact line region ) differs from the hydrodynamics in the bulk of the liquid .
because of the contact line singularity @xcite , the fluid motion in presence of the contact line appears to be much slower than without it .
this difference can be accounted for by the introduction of the anomalously large energy dissipation inside the clr @xcite . for practical purposes
, one needs to know the dynamics of the liquid surface influenced by this dissipation .
this influence is especially strong in the very common case of ( i ) low viscosity fluids that ( ii ) wet partially the solid with ( iii ) no precursor film on it . in this case
the bulk dissipation is particularly small with respect to the large clr dissipation @xcite .
the contact line motion is very slow and the liquid surface can be described in the quasi - static approximation @xcite .
the dissipation in the liquid ( energy per unit time ) can then be approximated by the dissipation in the clr as @xcite @xmath0 where @xmath1 is the normal component of the contact line velocity and the integration is performed along the contact line .
the generalized dissipation coefficient @xmath2 is a constant that is assumed to be much larger that the liquid shear viscosity .
the expression ( [ diss ] ) assumes that the contribution of a piece of the clr to the dissipation is proportional to the contact line length .
then the @xmath3 term is leading for small @xmath1 .
the dissipation coefficient @xmath2 can be obtained e.g. from measurements of the kinetics of relaxation of an oval sessile drop toward its equilibrium shape ( see @xcite , where it was found @xmath4 times larger than the shear viscosity ) .
it can also be obtained from the measurements of @xmath1 and the dynamic contact angle @xmath5 by using the expression @xmath6 where @xmath7 is the equilibrium value of the contact angle and @xmath8 is the surface tension .
[ cos ] is common for many contact line motion models .
it can be shown @xcite that eq .
[ cos ] follows from the expression ( [ diss ] ) for the cases where @xmath1 does not vary along the contact line ( i.e. for the contact line of constant curvature ) . usually
, one is interested to know the displacement of the contact line ( or its statistical properties in the case of the irregular solid ) because its position serves as a boundary condition for the determination of the shape of the liquid surface .
the recent articles on this subject show that no general approach to this problem is accepted .
it is recognized generally @xcite that the static contact line equation should be non - local because the contact line displacement at one point influences its position at other points through the surface tension . in dynamics , the local relations similar to eq .
[ cos ] were considered universal for a long time . in our previous work
@xcite we developed a non - local dynamic approach .
we showed that if non - locality is taken into account , eq . [ cos ] is not valid in the general case where @xmath1 varies along the contact line .
this variation can be due either to the initially inhomogeneous contact line curvature ( like in @xcite ) or to a inhomogeneous substrate . in this letter
we analyze this latter case .
we use this non - local approach to derive an equation of motion for the very common case of the capillary rise of the liquid along a vertical solid wall with the account of the surface defects .
a cartesian reference system @xmath9 is chosen in such a way that the liquid surface far from the wall ( defined by the @xmath10 plane ) coincides with the @xmath11 plane , see fig .
[ wilh ] . the liquid surface described by the equation @xmath12
is assumed to be weakly deformed so that @xmath13 the contact line can be described by the equation @xmath14 . a piece of the liquid surface of length @xmath15 in the @xmath16-direction ( see fig . [ wilh ] )
is considered so that the final form for the equation of motion is found in the limit @xmath17 .
first , following the approach @xcite , we find the energy @xmath18 of the liquid ( per contact line length ) as a functional of @xmath19 .
it is convenient to break @xmath18 into two terms , @xmath20 . in the approximation ( [ weak ] ) , the first term reads @xmath21,\ ] ] where @xmath22^{1/2}$ ] is the capillary length , @xmath23 is the liquid density and @xmath24 is the gravity acceleration .
the defects are accounted for in the second term @xcite @xmath25 by the fluctuations of the function @xmath26 which is the difference of the surface energies ( in @xmath8 units ) of the gas - solid and liquid - solid interfaces at the point @xmath27 of the solid surface .
according to the young expression , @xmath28 $ ] , where @xmath7 is the local value of the equilibrium contact angle .
when the surface defects exist , @xmath7 is an arbitrary function of @xmath29 . in the quasi - equilibrium approximation
, the surface shape @xmath30 is found by minimization of the functional @xmath18 resulting in the equation @xmath31 which can be solved by separating the variables by using the boundary condition @xmath32 and assuming that the function @xmath33 is bounded at @xmath34 .
the solution reads @xmath35\exp\left(-x\sqrt{l_c^{-2}+\pi^2 n^2/l^2}\right),\ ] ] where the coefficients @xmath36 , @xmath37 , @xmath38 are the coefficients for the fourier series @xmath39.\ ] ] they can thus be related to @xmath40 by the expressions @xmath41 one can notice that @xmath36 is simply the horizontally averaged displacement of the contact line .
the back substitution of eq .
( [ f ] ) with @xmath36 , @xmath37 , @xmath38 replaced by their values ( [ abh ] ) into eq .
( [ u1 ] ) results in the functional @xmath42.\ ] ] eq .
( [ u2 ] ) together with eq .
( [ funcu1 ] ) define the functional @xmath43 $ ] .
now , one needs to calculate the dissipation function @xmath44 ( per unit length ) . in the approximation ( [ weak ] ) , @xmath45 in eq .
( [ diss ] ) and @xmath46 where the dot means a time derivative .
the equation of motion can be written in the quasi - static approximation as @xmath47\over \delta h}=-{\delta t[\dot h]\over \delta \dot h},\ ] ] where @xmath48 means variational derivative . by substituting eqs .
( [ u2],[funcu1],[t ] ) into eq .
( [ var ] ) and taking the variational derivatives , one obtains the dynamic equation of the contact line motion : @xmath49,\ ] ] where @xmath19 and @xmath36 are assumed to be time dependent .
the final form for the integral equation for the contact line motion can be obtained by taking the limit @xmath17 : @xmath50\sqrt{l_c^{-2}+p^2}\biggr\}.\ ] ] to treat the contact line equation , it is convenient to introduce the spatial fluctuation @xmath51 and solve the equations for @xmath52 and @xmath36 separately .
first , we derive the dynamic equation for @xmath36 by integrating eq .
( [ hl ] ) over @xmath16 from @xmath53 to @xmath54 , changing the integration order and dividing by @xmath15 .
the integrals of all terms in the sum are zero and one obtains the following equation for @xmath36 @xmath55 where @xmath56 is the ( time dependent ) horizontal average of @xmath57 . by subtracting eq .
( [ h0 ] ) from eq .
( [ hh ] ) one obtains the equation for @xmath52 which has the same form as eq .
( [ hh ] ) where @xmath57 should be replaced by its fluctuation @xmath58 .
equation for @xmath52 simplifies in the fourier space . by making use of the convolution theorem @xcite ,
one obtains @xmath59 where the tilde over a variable denotes its fourier transform with the parameter @xmath60 , e.g. @xmath61 this concludes the derivation of the equation of contact line motion . the stationary version of eq .
( [ h1f ] ) was derived first by pomeau and vannimenus @xcite .
the static contact line position is defined by the minimum of the energy @xmath43 $ ] .
however , it can exhibit multiple minima that correspond to the metastable states . in this case
one needs to choose between them @xcite to determine correctly the contact line position and corresponding local contact angle ( which are both different in advancing and receding cases ) . in dynamics , eq .
( [ hh ] ) provides us with an unique solution because the system knows " the direction of the contact line motion and its history .
to show the relevance of this approach , we solve rigorously the contact line dynamics for a simple case of a single stripe - shaped defect at the wall , @xmath62 where @xmath63 and @xmath64 are constants .
consider first the contact line displacement at equilibrium described by the equations @xmath65 notice that @xmath56 should be small enough so that @xmath30 defined by eq . ( [ f ] ) satisfies conditions ( [ weak ] ) . obviously , eq . ( [ h0 ] ) results in @xmath66 and @xmath67 where @xmath68 .
to show how this result relates to earlier approaches to the single - defect problem , we analyze first the limit @xmath69 .
the inverse transform can easily be taken using the tables @xcite and results in @xmath70 of the contact line from its horizontally averaged value .
the defect area ( half - width of which is @xmath71 ) is shadowed .
the time values in the units @xmath72 are shown above the corresponding curve.,height=264 ] where @xmath73 is the modified bessel function of zeroth order .
one recognizes the result @xcite .
it leads to an unphysical divergence when @xmath74 : @xmath75 and a cut - off at small @xmath16 is needed @xcite . using the present approach
, one can understand the origin of this divergence .
the limit @xmath69 in the rigorous result ( [ h1eq ] ) is equivalent to the limit @xmath76 since their product enters eq .
( [ h1eq ] ) .
the asymptotics at @xmath76 in the fourier transform corresponds @xcite to the limit @xmath77 in the original function , which means that eq .
( [ hd ] ) is correct only at @xmath77 .
however , it is not guaranteed that its asymptotics at @xmath74 is correct .
let us return now to the rigorous expression ( [ h1eq ] ) .
the fourier transform ( [ h1eq ] ) can be inverted analytically , @xmath78/\pi,\ ] ] where @xmath79 .
it is clear now that @xmath80 remains finite , see the dotted curve in fig .
[ stripe ] .
the dynamic solution can be obtained the same way by using the initial condition @xmath81 : @xmath82\ } , \\
h_1(y , t)=\delta c\left[f(\delta+y)+f(\delta - y)\right.\left .- g(\delta+y , t)-g(\delta - y , t)\right]/\pi , \end{array}\ ] ] where @xmath83 the time evolution of @xmath84 is shown in fig .
[ stripe ] .
these curves can be compared to those obtained experimentally in @xcite where the motion of the contact line over a single defect was studied .
the comparison shows a good qualitative agreement .
the results can not be compared quantitatively since in each of these articles some parameters that enter eq .
[ h1d ] ( @xmath85 in particular ) are missing .
( [ hh ] ) describes the spontaneous contact line motion for _ arbitrary _ distribution of surface energy given by the function @xmath26 ( provided @xmath86 is small enough so that the conditions ( [ weak ] ) are satisfied ) .
this function can be considered random and the equation becomes stochastic .
it can be used to establish any statistical parameter of the contact line in dynamics . in particular ,
the collective effect of defects on the contact line motion can be studied . in this work
we solve a simple example where the defect properties do not vary along the average contact line velocity vector . in the general case , where the local value of the equilibrium contact angle ( or of the function @xmath57 ) varies along this direction ,
the equation of the contact line motion becomes non - linear .
more sophisticated methods are then needed to solve it .
g. d. nadkarni & s. garoff , europhys . lett . *
20 * , 523 ( 1992 ) .
j. m. marsh & a .- m .
cazabat , phys . rev
. lett . * 71 * , 2433 ( 1993 ) ; europhys . lett .
* 23 * , 45 ( 1993 ) .
a. paterson , m. fermigier , p. jenffer , & l. limat , phys .
e * 51 * , 1291 ( 1995 ) . | the wetting flows are controlled by the contact line motion .
we derive an equation that describes the slow time evolution of the triple solid - liquid - fluid contact line for an arbitrary distribution of defects on a solid surface .
the capillary rise along a partially wetted infinite vertical wall is considered .
the contact line is assumed to be only slightly deformed by the defects .
the derived equation is solved exactly for a simple example of a single defect . |
thanks to advances in nanofabrication , experimentalist can routinely manufacture nanostructures that resemble artificial atoms quantum dots that can be manipulated with an extreme degree of control@xcite .
the degree of tunability is such , that one can see a state with a single electron sitting at the dot , and even control and detect its spin . by tuning a gate voltage
one can also control the transport through the dot in what is regarded as a single - electron transistor .
the techniques used to construct these devices can also be applied to build structures with multiple dots , and suggest that scalability to circuits of many quantum dots should be technologically achievable .
another avenue that has been pursued is using molecules to build transistors , instead of quantum dots@xcite .
some examples of working devices have been realized with nanotubes@xcite , benzene molecules@xcite , fullerene@xcite , graphene @xcite , and proteins@xcite , to mention a few .
adding contacts and gating these devices is very challenging , and scalability may be more difficult to attain . a third possibility consists of manipulating single atoms to build complex structures @xcite that could be controlled using an stm tip@xcite .
whether a complex molecule , or a quantum - dot array , the most general transport features of these systems in the absence of interactions can infered from some basic fundamental physics : aharonov - bohm - type interference@xcite , and fano resonances@xcite .
there is an important effort toward engineering devices at the atomic / molecular scale taking advantage of these different phenomena , and that can have useful functionality as classical switches diod , transistor , or novel behavior that could open the doors to quantum circuits with no classical analogy .
these systems could lead to a new generation of electronic devices , and revolutionary applications such as quantum computers .
the chemistry of these molecules , or the geometry of the circuits , can be simple , or not . in this work
we show that regardless of the chemistry , the transport properties of a molecular device can be inferred from symmetry considerations@xcite .
moreover , canonical transformations allow us to establish dualities between quantum circuits , providing us with a tool to design devices that are protected by symmetry .
consider a one dimensional chain of atoms , described by a tight - binding chain with hopping @xmath0 , with a weak link in the center , with hopping @xmath1 .
let us split the chain into left and right halves , and fold it in two , as depicted in fig .
[ map1](a ) .
the corresponding hamiltonian is : @xmath2 where @xmath3 is the electron annihilation operator acting on site @xmath4 of lead @xmath5 ( where the values 0,1 correspond to left and right leads ) , and sites labeled `` 0 '' connect the two leads .
the addition of the term proportional to @xmath6 will become apparent below . for the moment , we take @xmath7 , and consider spinless fermions , but these considerations can be easily generalized to the spin - full case .
we now introduce a symmetric ( + ) and antisymmetric ( - ) combination of operators acting on the left and right leads .
this is nothing else but an application of the reflection symmetry , yielding new even(+ ) and odd(- ) operators : @xmath8 this is a simple change of basis , with the new operators obeying the same fermionic anti - commutation rules .
this is a well known canonical transformation , originally introduced in the context of quantum impurity problems @xcite , and referred to as a `` folding '' transformation @xcite .
it is used extensively in numerical renormalization group ( nrg ) calculations @xcite , and it has been recently used to study impurity problems on rings using the density matrix renormalization group ( dmrg ) @xcite . it has also been used in rather creative ways to simplify complex problems in quantum chemistry , such as conjugated molecules and polymers @xcite . in this work , we shall take an unconventional approach : instead of applying the transformation to the entire system , we shall apply it to only a fraction of it , say @xmath9 sites to the right of the weak link . as a result
, we find a section of the chain transformed to the new basis , connected to the rest of the left and right leads that remain in the original basis , as shown in fig .
1(b ) . at the boundary connecting the transformed and untransformed sections
, we find new hopping terms mixing the two bases as : @xmath10 + \mathrm{h.c . }
\label{boundary}\end{aligned}\ ] ] after some simple algebra , it is easy to see that the hopping at the weak link , maps onto effective chemical potentials in the new basis ( see appendices a and b for details ) : @xmath11 where @xmath12 is the density operator .
this term is just a boundary chemical potential with opposite signs for the `` + '' and `` - '' sections of the leads , at site `` 0 '' .
all the other hopping terms along the `` + '' and `` - '' sections of the lead remain unaltered in the new basis . taking all this into account , and after some visual inspection , it is easy to see that the original chain , after the basis transformation , maps onto the equivalent circuit shown in fig .
[ map1](c ) : a diamond connected to the left and right leads ( that remain in the original basis ) , and the `` + '' and `` - '' leads that have a finite length , and a boundary chemical potential at the ends of magnitude @xmath1 . let us call these leads `` gates '' from now on .
all hoppings inside the diamond are equal to @xmath13 , with the exception of one link that has the opposite sign .
this can be interpreted as a @xmath14 magnetic flux threading the diamond .
the physics of the original system , and the new equivalent one should be completely equivalent , since this mapping is exact .
therefore , let us consider setting the boundary chemical potential of the two `` + '' and `` - '' leads on the equivalent circuit to zero , @xmath15 .
if we look at the original chain , this corresponds to effectively cutting the connection between left and right leads , completely opening the circuit .
therefore , under these conditions , no current can circulate from left to right leads , no matter what the temperature is .
the circuit is completely open , by simple symmetry considerations !
if one wanted to describe the transport properties of the equivalent circuit , one could use ideas of quantum interference@xcite , and would consider the anti - resonances coming from the gates@xcite or combinations of the two @xcite .
however , this simple mapping shows a different aspect of the problem , and allows one to understand all the transport properties by simply studying at the original unfolded tight - binding chain .
so far , we have shown that the equivalent circuit acts like a multi - gate transistor : by simply tuning the potential at the gates , we can open the circuit , and completely suppress transport between the leads . alternatively , at finite tunneling amplitude @xmath16 , we can obtain perfect transport .
another situation that we can consider is connecting one of the gate leads to ground , and the other one to a gate voltage @xmath6 .
this corresponds to choosing @xmath15 , @xmath17 in eq.([link ] ) : @xmath18 the results for the conductance are shown in fig.[cuad - oct ] : at zero voltage @xmath7 , the circuit is open , while at large voltages we recover the perfect conductance .
this behavior is analyzed below , and in appendix c.
one of the undesired aspects of the previous `` device '' is the emergence of a @xmath14 flux threading the loop .
one could easily devise a setup to get rid of the phase .
consider the circuit shown in fig.[map1](d ) , corresponding to a t - shape junction .
after applying the folding transformation to all the sites situated to the left of the connecting link , we recover the same molecular circuit with the diamond in the center , but with no phase shift .
the price we have to pay is that we can no longer interpret the problem in terms of a simple tight - binding chain .
however , this still is a non - interacting circuit , and we can easily understand the transport properties . in this case , the conductance will depend on the parity of the dangling lead in the t - junction .
let us first consider the @xmath7 case . if @xmath19 this segment will have an even number of sites , with a vanishing density of states at @xmath20 .
this means that it can not participate in the transport between the left and right leads , yielding a perfect conductance .
if @xmath21 , the segment is effectively cut in half , and now two situations can arise : ( i ) for @xmath22 , we obtain the same result as before for @xmath19 , or ( ii ) for @xmath23 , a resonance will appear at the fermi level , and the conductance will cancel identically . notice that these are more complex generalizations of the problem studied in ref.@xcite .
the general situation for finite @xmath6 will be studied below .
we can generalize the previous protocol to other geometries , to obtain the equivalent of molecular transistors with a ring structure of @xmath24 sites , similar to the proposed devices based on annulenes@xcite .
consider the geometry illustrated in fig.[map2](a ) .
it depicts a one dimensional chain , that has been folded in half at the weak link , but with the addition of a second `` twist '' plus as dangling link , marked with an arrow in fig.[map2](a ) . by rotating the @xmath25 operators to the left of the @xmath9th site
, we obtain the hexagonal molecule shown in fig.[map2](b ) , connected to leads , and gates , as in the original case described before , with the emerging @xmath14 flux . the main difference with the tight - binding chain
is the dangling sites , which introduce an additional degree of complexity .
if the size of the annulus is @xmath26 , the dangling link will have an even number of sites , and will not participate in the transport properties , yielding the same results as the diamond . for @xmath27 , the dangling link will introduce an anti - resonance , and the conductance will be @xmath28 .
again , the undesired @xmath14 phase can be removed by re - shaping the circuit as a @xmath29-junction ( see fig.[map2](d ) ) . in this case
, we expect a parity effect in the transport behavior , depending on the size of the ring @xmath24 , and the size of the dangling `` gate leads '' , @xmath30 .
we now analyze the case with @xmath15 , @xmath17 in more detail . in general
, the transport properties are related to the green s functions that propagate one electron from the left to the right leads @xmath31 .
the conductance @xmath32 at zero bias can be written as @xcite @xmath33 where @xmath34 is the local density of states ( ldos ) at the leads ( assumed to be semi - infinite tight - binding chains ) and @xmath35 is the fermi level .
since the system is non - interacting , we can calculate the exact green s functions using either the equations of motion formalism@xcite or by applying wick s theorem . in figure [ scheme ]
we have rearranged the geometry of the leads to show more clearly the structure of the system .
independently of the phase , the system always presents a dangling segment .
the conductance is controlled by fano resonances .
two different paths , as represented in the inset of fig .
[ scheme](a ) , can interfere in a constructive or destructive way@xcite .
the parity of the number of sites in the ring will determine whether the phases cancel or add up .
it is convenient to start the discussion by considering the t - junction problem . in this case , the transport properties depend on both , the parity of @xmath24 , and the size of the gate leads @xmath36 .
we first look at a ring with @xmath26 sites . in figure
[ cuad - oct ] we present the results for the conductance for rings with 4 and 8 sites as a function of @xmath6 and gate leads with different lengths @xmath36 , odd and even .
we find two clearly distinct situations depending on the parity of @xmath36 . as shown in this figure , when @xmath36 is even a perfect conductance of @xmath37 independent of @xmath6 is realized , while for odd @xmath36 the conductance depends strongly of the gate potential @xmath6
. to understand this behavior we look at figure [ scheme](a ) .
as explained before , in the @xmath7 limit the upper part of the system is disconnected and the conductance is controlled by the number of sites in the gate leads @xmath36 . if @xmath36 is odd , this segment will have a resonance at the fermi level and again , this will cancel the conductance identically . if @xmath36 is even , there will not be any resonance in the ldos and the conductance in this case is perfect , g=@xmath38 . when @xmath17 the upper part of the system of figure [ scheme](a ) will be connected and the chemical potential at the ends will also start playing a role . the total number of sites of the hanging part will be @xmath39 always odd , producing , in principle , a resonance at the fermi level , and canceling the conductance
. however , at large @xmath6 the chemical potential removes _
one _ site from the hanging segment , with the conductance growing toward @xmath38 with increasing @xmath6 . in order to understand the effects of a finite gate voltage @xmath6 ,
let us consider the system before the transformation , as shown in figures [ map1](a ) and ( d ) , and take @xmath40 , such that we apply the folding transformation only on the link connecting left and right leads @xmath41 ( [ link2 ] ) .
this will yield two orbitals `` + '' and `` - '' that will be connected to the rest of the system through @xmath42 ( [ boundary ] ) . in the large @xmath6 limit , the orbital `` + '' will be empty , and effectively removed from the system , which will have one site / orbital less .
this changes the parity of the problem , converting a resonance into an anti - resonance and viceversa .
we now analyze rings with @xmath27 sites .
results for the conductance as function of @xmath6 for rings with @xmath43 and @xmath44 sites are shown in fig .
[ hex - dec ] .
same as before , we can observe two different regimes .
for @xmath36 odd the conductance is always zero while for @xmath36 even the conductance depends strongly on @xmath6 .
the physics can be explained following similar arguments as the previous case , but with odd and even @xmath36 cases inverted .
we now go back to the case with a flux , corresponding to a single chain .
as shown in figures [ cuad - oct ] and [ hex - dec ] , the conductance is independent of @xmath36 , as we should expect for a tight - binding chain .
however , there is a parity effect corresponding to the size of the ring being @xmath26 or @xmath27 , which dominates the transport properties . as mentioned before ,
this is explained by the effect of the dangling links added to the chain to construct the rings .
depending on the number of sites in the links , we will observe a resonance or anti - resonance at the fermi level ( see fig.[scheme](b ) ) .
the effect of @xmath6 can again be explained as gradually removing a site from the chain ( appendix b ) .
however , since adding or removing a site from a chain does not have any effect on its fundamental transport properties , the conductance interpolates smoothly between the @xmath45 and @xmath46 cases . for both @xmath26 , and @xmath27 cases , with @xmath36 odd , the behavior is exactly the same for the chain and the t - junction .
this is a remarkable result , and leads one to think of a hidden symmetry between the two setups . as a matter of fact
, it is not a symmetry relating the two hamiltonians , but a symmetry in the green s functions that determine the equilibrium transport properties , that is `` magically '' realized when these quantities are evaluated at the fermi level , @xmath20 .
this is analyzed in great detail in the appendix c.
let us look again at fig.[map1](d ) .
we can make the transformation inside the hanging segment of the t - junction , away from the contact to the leads , as shown in fig.[map3](a ) .
it is easy to see that the equivalent circuit results as depicted in fig.[map3](b ) .
the physics of the problem , of course , is just as described in the previous sections .
it is the same system after all !
this shows that connecting the molecule in parallel , or in series , will have exactly the same effect and functionality .
moreover , it may be easier to realize experimentally , and more convenient for scalability purposes .
we have illustrated how three complex quantum `` devices '' , shown in figures [ map1](c ) , [ map2](c ) , and [ map3](b ) can be understood in terms of simple equivalent circuits that have the general topology of either a single chain , or a t - junction .
these exact mappings have been obtained as a straightforward application of the folding transformation .
we can easily understand the transport properties of these simple circuits , and explain the general behavior of the molecular device .
two remarkable aspects are that : ( i ) in the single chain case , with a phase @xmath14 threading the molecule , the gate voltage can be used to effectively `` cut '' the equivalent circuit in two , yielding a complete cancellation of the conductance .
this effect is protected by the symmetry of the problem , and indicates that this behavior as a quantum switch survives at arbitrary temperatures .
( ii ) the zero - temperature transport properties of the chain and the t - junction are the same for @xmath36 odd , and can be understood in terms of a symmetry in the equilibrium green s functions at @xmath20 . even though we have limited our analysis to a single orbital situation , these considerations can be generalized to a multi - orbital case .
the non - interacting limit assumed here applies to a problem that can be described in terms of a hckel - like theory .
the addition of interactions would introduce long range terms in the transformed basis , making the folding transformation impractical .
however , one could imagine studying the interactions within a hartree - fock framework , in which the resulting effective hamiltonian is quadratic . in this picture
, the interactions introduce a shift in the diagonal site energies , and the main effect is to shift and to split the resonance positions @xcite . in such scenario ,
as long as the reflection symmetry is preserved , our treatment is still valid .
inteference - based molecular transistors have already been proposed @xcite , and quantum interference has recently been observed in molecules@xcite .
this study indicates a path toward understanding and designing molecular devices with customized functionality , and we illustrate it with setups that operate as multi - terminal transistors that can work connected in series , and parallel .
these symmetries make us see these systems in a new light , and can also aid _ ab - initio _ calculations .
it is clear that the mapping is not limited to the particular examples presented here , but it is more general and may lead to complex and interesting possibilities that could pave the way toward new advances in nano - electronics .
aef is grateful to nsf for funding under grant dmr-0955707 .
the entire treatment utilized in this work relies on a simple and basic symmetry : reflection , equivalent to a bonding / anti - bonding transformation .
we start by considering a non - interacting one - dimensional chain , folded in two .
it is easy to understand the effects of this symmetry by visualizing them in diagrams , as depicted in figure [ appendix2 ] .
when this transformation is applied to two parallel links , as in diagram [ appendix2](a ) , the transformation leaves the hamiltonian invariant and reads exactly the same in the transformed basis : @xmath48 when applied to a link connecting a site on chain `` 0 '' and its partner on chain `` 1 '' , it leads to a vanishing hopping , and the introduction of two chemical potentials ( see diagram [ appendix2](b ) ) : @xmath49 the interesting situation arises when the transformation is applied to the left of certain boundary , and not to the sites to the right .
then , at the link connecting the transformed sites , and those remaining in the original basis , new hopping terms emerge .
the four possibilities shown in figure [ appendix2](c)-(f ) correspond to : @xmath50 by simply combining these diagrams , and introducing `` twists '' in the chains , we can construct the devices presented in the body of this work . of course , this is just a limited number of the countless possibilities .
in this appendix we give a closer look at the link term connecting the two chains : @xmath51 after the considerations introduced in the previous appendix , it is easy to see that this hamiltonian can be transformed into : @xmath52 when @xmath15 , this reduces to : @xmath53 for large @xmath54 , this term will force the `` + '' orbital to be either fully occupied , or empty , with the consequent suppression of charge fluctuations . therefore , all the transport will occur through the `` - '' orbital , meaning that the total system will behave as though it had one less site .
in order to understand why the conductance is the same for the t - junction and the single chain , _
i. e. _ does not depend on the flux for @xmath26 , and @xmath27 , @xmath36 odd , we need to calculate the propagator @xmath55 appearing in the expression for the conductance . we first realize that , independently of the connection , both systems have a common part shown in fig .
[ appendix1](a ) .
therefore , we calculate the partial propagators for this common segment , @xmath56 and @xmath57 , and we establish the connection to the second lead using a dyson equation . for the t - junction , the left lead must be connected to the site labeled as @xmath58 , while for the the single chain , the connection must be established through site @xmath59 .
this is shown in figures [ appendix1](b ) and ( c ) .
the equations of motion for the propagator @xmath55 for the t - junction case can be written as , @xmath60 while for the chain we have , @xmath61 where @xmath62 , and we had preserved the subindex to distinguish the different cases . the bare green functions @xmath63 are the propagators before reestablishing the connection between sites @xmath58 and @xmath64 .
the green s functions @xmath65 denote the propagators after reconnecting sites @xmath58 and @xmath64 as shown in fig .
[ appendix1](a ) .
comparing eqs .
( [ sinf ] ) and ( [ conf ] ) , we can observe that the term between brackets is common to both cases .
the prefactors will determine when these two green s functions are equal .
note that by symmetry @xmath69 .
the prefactors in eqs .
( [ sinf ] ) and ( [ conf ] ) as well as @xmath70 and @xmath71 from eqs .
( [ partialsin ] ) and ( [ partialcon ] ) will be equal if , @xmath72 notice that , since we are calculating the zero bias conductance , this condition must be proved just for @xmath73 . in order to demonstrate that the condition ( [ condition ] ) is satisfied
, we need to calculate the bare green s functions for a single tight - binding segment of length @xmath74 .
for simplicity we will restrict this treatment to rings with @xmath75 sites , @xmath76 , but the demonstration can be extended to arbitrary ring sizes .
the green s function is the inverse of the matrix @xmath77 where @xmath78 is the hamiltonian of the segment . for the non - interacting case studied here
, @xmath78 is a finite @xmath79 symmetric tridiagonal matrix : @xmath80 for @xmath81 , and @xmath82 for @xmath83 .
there is a simple and elegant formula to calculate the inverse of such matrix @xcite : @xmath84 where the @xmath85 s verify the recurrence relation @xmath86 with @xmath87 and @xmath88 , and the @xmath89 s obey the recurrence @xmath90 with @xmath91 and @xmath92 .
notice that @xmath93 .
we are interested in the green s function at the fermi level , @xmath94 ( we have ignored any imaginary contributions for simplicity ) .
moreover , all the diagonal elements are @xmath95 except @xmath96 .
similarly , all the off - diagonal elements are @xmath97 , with the exception of @xmath98 . using these expressions
, it is easy to verify that @xmath99 for @xmath100 , and @xmath101 for @xmath102 .
moreover , for @xmath102 , @xmath103 and @xmath104 therefore , we can see that the recursions lead to the same numerical result for @xmath105 and @xmath106 as long as the length of the chain obeys @xmath107 , or @xmath36 odd . the important condition to get
this equivalence is the presence of the symmetric chemical potentials @xmath6 in @xmath41 .
notice that these observations do not imply a symmetry relating the hamiltonians of the chain and the t - junction , but a symmetry in the green s functions at the fermi level and , consequently , their equilibrium transport properties . | realizing a quantum transistor built of molecules or quantum dots has been one of the most ambitious challenges in nanotechnology . even though remarkable progress has been made , being able to gate and control nanometer scale objects , as well to interconnect them to achieve scalability
remains extremely difficult .
most experiments concern a single quantum dot or molecule , and they are made at ultra low temperature to avoid decoherence and tunneling .
we propose to use canonical transformations to design quantum devices that are protected by symmetry , and therefore , may be operational at high temperatures .
we illustrate the idea with examples of quantum transistor architectures that can be connected both in series and parallel . |
the pioneer 10/11 missions were the first spacecraft to explore the outer solar system @xcite .
their objectives were to conduct , during the 1972 - 73 jovian opportunities , exploratory investigation beyond the orbit of mars of the interplanetary medium , the nature of the asteroid belt , and the environmental and atmospheric characteristics of jupiter and saturn ( for pioneer 11 ) .
pioneer 10 was launched on 2 march 1972 on top of an atlas / centaur / te364 - 4 launch vehicle .
the launch marked the first use of the atlas - centaur as a three - stage launch vehicle .
the third stage was required to rocket pioneer 10 to the speed of 14.39 km / s needed for the flight to jupiter . its sister craft , pioneer 11 , was launched on 5 april 1973 , like pioneer 10 , also on top of an atlas / centaur / te364 - 4 launch vehicle . after safe passage through the asteroid belt on 19 april 1974 , pioneer 11 s thrusters were fired to add another 63.7 m / s to the spacecraft s velocity .
this adjusted the aiming point at jupiter to 43,000 km above the cloudtops .
the close approach also allowed the spacecraft to be accelerated by jupiter to a velocity of 48.06 km / s so that it would be carried across the solar system some 2.4 billion km to saturn . after the jupiter and saturn ( for pioneer 11 ) encounters ( see figure [ fig : pioneer_inner_path ] ) , the craft followed escape hyperbolic orbits near the plane of the ecliptic on opposite sides of the solar system , continuing their extended missions @xcite .
( see figure [ fig : pioneer_path ] . )
pioneer 10 explored the outer regions of the solar system , studying energetic particles from the sun ( solar wind ) , and cosmic rays entering our portion of the milky way .
pioneer 10 continued to make valuable scientific investigations until its science mission ended on march 31 , 1997 . since that time , pioneer 10 s weak signal has been tracked by the nasa s deep space network ( dsn ) as part of an advanced concept study of communication technology in support of nasa s future interstellar probe mission .
pioneer 10 eventually became the first man - made object to leave the solar system .
the power source on pioneer 10 finally degraded to the point where the signal to earth dropped below the threshold for detection in its latest contact attempt on 7 february , 2003 .
the previous three contacts had very faint signals with no telemetry received .
the last telemetry data point was obtained from pioneer 10 on 27 april 2002 when the craft was 80 au from the sun . following its encounter with saturn , pioneer 11 explored the outer regions of our solar system , studying the solar wind and cosmic rays .
pioneer 11 sent its last coherent doppler data on 1 october 1990 while at @xmath7 au from the sun . in september 1995 ,
pioneer 11 was at a distance of 6.5 billion km from earth . at that distance
, it takes over 6 hours for the radio signal to reach earth .
however , by september 1995 , pioneer 11 could no longer make any scientific observations . on 30 september 1995 , routine daily mission operations were stopped .
intermittent contact continued until november 1995 , at which time the last communication with pioneer 11 took place .
there has been no communication with pioneer 11 since .
the earth s motion has carried our planet out of the view of the spacecraft antenna .
the spacecraft can not be maneuvered to point back at the earth .
it is not known whether the spacecraft is still transmitting a signal .
no further tracks of pioneer 11 are scheduled .
-10pt the pioneers were excellent craft for the purposes of precision celestial mechanics experiments @xcite .
this was due to a combination of many factors , including their attitude control ( spin - stabilized , with a minimum number of attitude correction maneuvers using thrusters ) , power design ( the plutonium-238 powered heat - source rtgs
radioisotope thermoelectric generators
being on extended booms aided the stability of craft and also reduced thermal effects on the craft ) , and precise doppler tracking ( with the accuracy of post - fit doppler residuals at the level of mhz ) .
the result was the most precise navigation in deep space to date @xcite .
( see figure [ fig : pio - craft ] for a design drawing of the spacecraft . ) because of the excellent health and navigational capabilities of pioneer 10 , nasa supported a proposal to initiate search for unmodeled accelerations in 1979 ( when the spacecraft was at a distance of some 20 au from the sun ) .
the main motivation was the search for planet x. eventually both pioneers were used in the search for trans - neptunian objects ; their doppler tracking capabilities also yielded the first ever limits on low frequency gravitational radiation @xcite . the pioneer 10/11 spacecraft had exceptional `` built - in '' acceleration sensitivity naturally allowing them to reach the level of @xmath8 m / s@xmath2 .
as indicated by their radiometric data received from heliocentric distances of 20 - 70 au , the accuracies of their orbit reconstructions were limited by a small , anomalous , doppler frequency drift @xcite . by 1980 , when pioneer 10 already had passed a distance of @xmath9 20 au from the sun and the acceleration contribution from solar radiation pressure on the craft ( away from the sun ) had decreased to less than @xmath10 m / s@xmath2 , the radiometric data started to show presence of the anomalous acceleration ( towards the sun ) .
this blue - shifted drift is uniformly changing with a rate of @xmath11 hz / s , which can be interpreted as a sunward constant acceleration of @xmath12 m / s@xmath2 @xcite .
the detailed study of this anomaly @xcite led to a better understanding of its properties ( summarized in section [ sec : anomaly ] ) ; however , the nature of this anomalous signal is still unknown . in this paper
we are going to discuss recently recovered set of pioneer data and objectives for upcoming new investigation of the anomalous residuals detected in the radiometric doppler data of the pioneers 10 and 11 .
we continue the paper in section [ sec : anomaly ] by discussing the pioneer anomaly and its known properties .
we briefly review the original efforts to understand the signal as well as some recent proposals to explain the pioneer anomaly .
section [ sec : pio - doppler ] presents the current status of the effort to retrieve the entire set of the pioneer radiometric doppler data .
section [ sec : mdr ] reports on the newly recovered pioneer telemetry data in the form of master data records ( mdrs ) .
because of their potentially high value in the investigation of the pioneer anomaly , we discuss the purpose , means of delivery , storage , processing and possible use of the mdrs . in section [ sec : objectives ] we present objectives for the upcoming study of the pioneer anomaly and particularly the use of both the extended set of doppler data and the mdrs to study the on - board systematics , including the development of an accurate thermal - electric - dynamical model . in section [ sec : conclude ] we conclude by presenting the next steps for the analysis of the pioneer anomaly .
the analysis of the pioneer 10 and 11 data ( with some support from the ulysses and galileo spacecraft ) @xcite demonstrated the presence of an anomalous , doppler frequency blue shift drift , uniformly changing with a rate of @xmath13 hz / s @xcite . to understand the phenomenology of the effect ,
consider @xmath14 , the frequency of the retransmitted signal observed by a dsn antenna , and @xmath15 , the predicted frequency of that signal modeling conventional forces influencing the spacecraft s motion including both gravitational and non - gravitational in their origin .
the observed , two - way ( round trip ) anomalous effect can be expressed to first order in @xmath16 as @xmath17 = -2\dot{f}_p\,t.$ ] after accounting for the gravitational and other large forces included in standard orbit determination programs this translates to @xmath18 _ { \tt dsn } = -f_{0}\frac{2a_p~t}{c}. \label{eq : delta_nu_syst}\end{aligned}\ ] ] here @xmath19 is the dsn reference frequency @xcite ( see discussion of the dsn sign convention in ref . [ 38 ] of @xcite ) .
after accounting for all _ known _ ( not modeled ) sources of systematic error ( discussed in @xcite ) , the conclusion was reached that there exists an anomalous sunward constant acceleration signal of @xmath20 the nature of this anomaly remains unexplained .
this signal has become known as the pioneer anomaly .
note that there exist a dualism in interpreting the radiometric doppler data .
the anomaly can be due to a force acting on the craft that produces acceleration eq .
( [ eq : ap ] ) , or , alternatively , this signal can also be interpreted as a time deceleration uniformly changing with rate @xmath3 s / s@xmath2 ( see discussion of this possibility in @xcite ) . by now several studies of the pioneer doppler navigational data have demonstrated that the anomaly is unambiguously present for both pioneer 10 and 11 .
these studies were performed with four independent ( and different ! ) navigational computer programs @xcite , namely : * the jpl s orbit determination program ( odp ) developed in 1980 - 2005 , * the aerospace corporation s chasmp code extended for deep space navigation @xcite , * code written at the goddard space flight center @xcite that was used to analyze pioneer 10 data for the period 1987 - 1994 obtained from the national space science data center ( http://nssdc.gsfc.nasa.gov/ ) , and finally * code developed at the institute of theoretical astrophysics , university of oslo , norway , that was recently successfully used to analyze the set of the pioneer 10 data above @xcite . for the most detailed analysis of the pioneer anomaly to date , @xcite used the following pioneer 10/11 doppler data : * pioneer 10 : the data set had 20,055 data points obtained between 3 january 1987 and 22 july 1998 and covering heliocentric distances @xmath21 au . * pioneer 11 : the data set had 19,616 data points obtained between 5 january 1987 to 1 october 1990 and covering heliocentric distances @xmath22 au .
the recent analyses of the pioneer 10 and 11 radiometric data @xcite have established the following basic properties of the pioneer anomaly : * _ distance : _ it is unclear how far out the anomaly goes , but the pioneer 10 data supports its presence at distances up to @xmath970 au from the sun .
the pioneer 11 data shows the presence of the anomaly as close in as @xmath920 au . * _ direction : _ from typical angular uncertainty of doppler navigation at s - band and spacecraft high gain antennae pointing accuracy set within 3 db gain bandwidth , @xmath23 behaves as a line - of - sight constant acceleration of the craft generally pointing in the innermost region of the solar system . * _ constancy : _
both temporal and spatial variations of the anomaly s magnitude are of order 10% for each craft , while formal errors are significantly smaller .
there are other pieces of information obtained from spot analyses @xcite ; they indicate that : * the anomalous acceleration was present in the pioneer 11 data at shorter distances , as close in as @xmath24 au ( see figures 6 and 7 in @xcite ) . *
the pioneer 11 data also indicated that the anomaly may be much smaller at distances @xmath25 au .
it appears to be amplified ( or turned on ) at a distance of @xmath24 au from the sun .
this is approximately when the craft flew by saturn and entered an hyperbolic , escape trajectory .
this information was used as guidance in investigating the applicability of proposals to explain the pioneer anomaly using both conventional and `` new '' physical mechanisms . in the next section
we briefly review these proposals .
initial announcement of the anomalous acceleration ( e.g. @xcite ) triggered many proposals that invoked various conventional physics mechanisms , all aimed at explaining the origin of the anomaly . finding a systematic origin of the proper magnitude and behavior was the main focus of these proposals , which is yet to be found .
although the most obvious explanation would be that there is a systematic origin to the effect , perhaps generated by the spacecraft themselves from anisotropic heat rejection or propulsive gas leaks , the analysis performed did not find evidence for either of them ; that is , no unambiguous , on - board systematic has been discovered .
this initial search was summarized in @xcite , where possible contributions of various mechanisms to the final solution for @xmath23 were given .
the entire error budget was subdivided in three main types of effects , namely effects due to sources external to the spacecraft ; the contribution of on - board systematics , and computational systematic errors . in this section
, we present a summary of this earlier work .
the first group of effects includes those external to the spacecraft , such as the solar radiation pressure , effects of the solar wind , and the effect of the solar corona on the propagation of radio wave signals .
@xcite also discuss the influences of the kuiper belt s gravity , galactic gravity , and electromagnetic lorentz forces .
errors in the accepted values of the earth s orientation parameters , precession , and nutation are included in the discussion .
the analysis evaluated the contributions of the mechanical instabilities and the location errors of the dsn antenna structures , the phase stabilities of the dsn antennae and clocks , and effects due to the troposphere and ionosphere .
even though some of these mechanisms are near the limit for contributing to the final error budget , it was found that none of them could explain the behavior of detected signal . moreover , some were three orders of magnitude or more too small . in totality , they were insignificant .
the second group of effects includes those that originated on - board and are tied to well - known sources ; this group , as expected , had the largest impact on the final error . among these effects , the radio beam reaction force produced the largest bias to the result , @xmath26 m / s@xmath2 .
as the force exerted by the radio beam necessarily points away from the earth ( and thus from the sun ) , the correction of the measured data increases the amount of the observed anomalous attractive force and makes the pioneer effect larger .
large uncertainties also came from a conjectured differential emissivity of the radioisotope thermoelectric generators ( rtgs ) , radiative cooling of the spacecraft , and propulsive gas leaks from thrusters of the attitude control system : @xmath27
m / s@xmath2 , @xmath28 m / s@xmath2 , and @xmath29 m / s@xmath2 , respectively .
the effect due to expelled helium produced within the rtgs was also considered , as well as the small difference in anomaly determinations between the two pioneers .
the effect of rejected thermal radiation was the second largest bias / uncertainty that has been the most critical systematic bias to quantify .
if heat generated by the on - board power sources was asymmetrically reflected by the body of the craft , a fore / aft acceleration could be produced causing the measured anomaly .
the pioneer spacecraft were powered by snap-19 rtgs ( space nuclear ancillary power ) mounted on long extended booms ( designed to protect the on - board electronics from heat and radiation impact ) @xcite . in principle , there was more than enough heat available on the craft to cause the anomaly . however ,
the spacecraft s spin - stabilized attitude control , special design of the rtgs and the length of the rtg booms that resulted in a relatively small spacecraft surface available for the preferential heat rejection significantly minimized the amount of heat for the mechanism to work .
the analysis of the 11.5 years of pioneer doppler data @xcite can only support an effect as large as @xmath30 m / s@xmath2 . in summary ,
although this group represents conceptually the most likely sources for the anomaly , these mechanisms did not gain enough experimental support . at most one
can obtain @xmath912% of the discovered effect by employing all of these mechanisms .
the possibility exists that a combination of the factors above could amount to the measured effect @xcite , but again , the shortness of the data interval used in the analysis , actual spacecraft design and performance data at hand , and also the complexities of modeling thermal radiative processes on the craft made it difficult to unambiguously support this claim @xcite .
the third group of effects was composed of contributions from computational errors .
the effects in this group dealt with the numerical stability of least - squares estimations , accuracy of consistency / model tests , mismodeling of maneuvers , and the solar corona model used to describe the propagation of radio waves .
it has also been demonstrated that the influence of annual / diurnal terms seen in the data on the accuracy of the estimates was all small @xcite . the total uncorrelated error associated with computational systematics
is estimated to be less than @xmath31 m / s@xmath2 .
these three groups of effects exhausted all available conventional explanations for the anomaly .
the inability to explain the pioneer anomaly with conventional physics has led to a significant number of theoretical proposals that use more unusual mechanisms ( more details are in @xcite ) .
the pioneer anomaly is an effect at the limit of what is detectable with radiometric tracking of a deep space probe , but it is huge in physical terms : the anomaly exceeds the corrections to newtonian motion predicted by general relativity by five orders of magnitude ( at 50 au ) . hence
, if the effect is not a result of conventional systematics it would have a considerable impact on our models of fundamental forces , regardless of the anomaly being due to a deceleration of the spacecraft or a blue shift of the radio signal .
attempts to verify the anomaly using other spacecraft proved disappointing .
this is because the voyager , galileo , ulysses , and cassini spacecraft navigation data all have their own individual difficulties for use in an independent test of the anomaly @xcite .
in addition , many of the deep space missions that are currently being considered either may not provide the needed navigational accuracy and trajectory stability sensitive to accelerations of under @xmath32 m / s@xmath2 or else they have significant on - board systematics that mask the anomaly .
a requirement to have an escape hyperbolic trajectory makes a number of other missions @xcite less able to directly test @xmath23 .
although these missions all have excellent scientific goals and technologies , nevertheless , their orbits lend them a less advantageous position to conduct a precise test of the detected anomaly .
a number of alternative ground - based verifications of the anomaly have also been considered ; for example , using very long baseline interferometry ( vlbi ) astrometric observations . however , the trajectories of spacecraft like the pioneers , with small proper motions in the sky , make it presently impossible to use vlbi in accurately isolating an anomalous sunward acceleration of the size of @xmath23 . efforts to explain the anomaly were originally focused on conventional physics mechanisms generated on - board , such as gas leaks from the propulsion system or a recoil force due to the on - board thermal power inventory .
so far , these mechanisms have been found to be either not strong enough to explain the magnitude of the anomaly or else to exhibit significant temporal or spatial variations contradicting the known properties of the anomaly presented above @xcite .
a number of other conventional physics possibilities have also been addressed . in particular
, it has been proposed that kuiper belt objects or dust could explain the anomaly by a gravitational acceleration , an additional drag force ( resistance ) and a frequency shift of the radio signals proportional to the distance .
of course , one of the most natural mechanisms to generate a putative physical force is the gravitational attraction due to a known mass distribution in the outer solar system ; for instance , due to kuiper belt objects or dust .
however , possible density distributions for the kuiper belt were studied in @xcite and found to be incompatible with the discovered properties of the anomaly .
even worse , these distributions can not circumvent the constraint from the undisturbed orbits of mars and jupiter @xcite .
the density of dust is not large enough to produce a gravitational acceleration on the order of @xmath23 @xcite and also it varies greatly within the kuiper belt , precluding any constant acceleration .
hence , a gravitational attraction by the kuiper belt can , to a large extent , be ruled out . also , the data from the inner parts of the solar system taken by the pioneer 10/11 dust detectors favors a spherical distribution of dust over a disk in this inner region .
ulysses and galileo measurements in the inner solar system find very few dust grains in the @xmath33 kg range @xcite .
ir observations rule out more than 0.3 earth mass from kuiper belt dust in the trans - neptunian region .
the resistance caused by the interplanetary dust is too small to provide support for the anomaly @xcite .
any dust - induced frequency shift of the carrier signal is also ruled out .
finally we note that , motivated by the numerical coincidence @xmath34 , where @xmath35 the speed of light and @xmath36 is the hubble constant at the present time , there have been many attempts to explain the anomaly in terms of the expansion of the universe .
@xcite showed that such a mechanism would produce an opposite sign for the effect .
a study of the effect of cosmic acceleration on the radio signals rather than on the spacecraft themselves was also undertaken .
this mechanism might be able to overcome the apparent conflict that @xmath23 presents to modern solar system planetary ephemerides @xcite .
the apparent difficulty to explain the anomaly within standard physics became a motivation to look for `` new physics . ''
so far these attempts have not produced a clearly viable mechanism for the anomaly .
in particular , the physics of mond ( modified newtonian dynamics ) represents an interesting possibility with its phenomenological long - range modification of gravity invoked to explain the rotation curves of galaxies @xcite .
however , the numerical value for the mondian acceleration @xmath37 is almost an order of magnitude smaller then @xmath23 and is not likely to be observed on the scales of the solar system .
there is also an attempt to explain the anomaly in the framework of a nonsymmetric gravitational theory @xcite .
it has been argued that a skew - symmetric field with a suitable potential could account for @xmath23 @xcite as well as galaxy and cluster rotation curves .
a modification of the gravitational field equations for a metric gravity field , by introducing a general linear relation between the einstein tensor and the energy - momentum tensor has also been claimed to account for @xmath23 @xcite .
both these approaches are currently being further investigated .
various distributions of dark matter in the solar system have been proposed to explain the anomaly , e.g. , dark matter distributed in the form of a disk in the outer solar system with of a density of @xmath38 kg / m@xmath39 , yielding the wanted effect .
dark matter in the form of `` mirror matter '' @xcite is one example .
however , it would have to be a special smooth distribution of dark matter that is not gravitationally modulated as normal matter so obviously is .
@xcite have shown that a generic scalar field can not explain @xmath23 ; on the other hand they proposed that a non - uniformly coupled scalar might produce the wanted effect .
although braneworld models with large extra dimensions offer a richer phenomenology than standard scalar - tensor theories , it is difficult to find a convincing explanation for the pioneer anomaly @xcite .
other ideas include yukawa - like or higher order corrections to the newtonian potential and a theory of conformal gravity with dynamical mass generation , including the higgs scalar ( see discussion in @xcite ) .
we must conclude that there are many interesting ideas proposed to explain the physics of the pioneer anomaly . however , most of them need more work before they can be considered to be viable .
to summarize , the origin of the pioneer anomaly remains unclear .
@xcite have advocated for an analysis of the entire set of existing pioneer doppler data , obtained from launch to the last useful data received from the vehicles .
this data could yield critical new information about the anomaly , especially during the earlier mission phases @xcite .
this led to the initiation of an effort to retrieve the early pioneer data that until recently existed in several places on various obsolete format and media ( e.g. , 9-track and even 7-track magnetic tapes ) .
recovery of radiometric data for a mission operating for more then 30 years is an effort that was never attempted before .
indeed , 30 years is a long time , presenting many unique challenges , including changes in the data formats , navigational software , as well as supporting hardware .
even the dsn configuration had changes new stations were built , and some stations moved , upgraded and reassigned .
the main asset of the entire mission support
its people changed the most . by 2005 all the dsn data formats , navigational software used to support pioneers , all the hardware used to read , write and maintain the data have become obsolete and are no longer operationally supported by existing nasa protocols . despite the anticipated complexities , the transfer of the available pioneer doppler data to modern media formats had been initiated at jpl in early june 2005 and , as of december 2005 , it is nearing completion . in this section
we discuss the current status of this effort to recover the pioneer 10/11 radiometric doppler data .
doppler data is the measure of the cumulative number of cycles of a spacecraft s carrier frequency received during a user - specified count interval .
the exact precision to which these measurements can be carried out is a function of the received signal strength and station electronics , but it is a small fraction of a cycle .
raw doppler data is generated at the tracking station and delivered via a dsn interface to customers . in order to acquire doppler data ,
the user must provide a reference trajectory and information concerning the spacecraft s rf system to jpl s deep space mission system ( dsms ) , to allow for the generation of pointing and frequency predictions .
the user specified count interval can vary from 0.1 sec to 10 minutes , with count times of 10 to 60 seconds being typical @xcite .
the average rate of change of the cycle count over the count interval expresses a measurement of the average velocity of the spacecraft in the line between the antenna and the spacecraft .
the accuracy of doppler data is quoted in terms of how accurate this velocity measurement is over a 60 second count .
the accuracy of data improves as the square root of the count interval .
-12pt 1 .
_ non - coherent doppler data _ ( also known as one - way or f1 data ) is data received from a spacecraft where the downlink carrier frequency is not based on an uplink signal .
the ability of the tracking station to measure the phase of the received signal is the same for non - coherent versus coherent data types , however the uncertainty in the value of the reference frequency used to generate the carrier is generally the dominant error source .
pioneers provided a significant amount of f1 data that is unfortunately not useful for precision orbit determination .
coherent doppler data _ is received from a spacecraft when the reference frequency of the received carrier signal was based on a transmitted uplink signal from the earth .
this is commonly known as two - way ( or f2 ) data , when the receiving and transmitting ground stations are the same , and three - way ( or f3 ) data , when the transmitting and receiving stations are different .
since the frequency of the original source signal is known , this error source does not affect data accuracy .
the accuracy of this data is a function primarily of the carrier frequency , but is affected by transmission media effects .
the f2 and f3 data are the primary source for the pioneer 10 and 11 orbit determination and were the primary focus of the data recovery efforts .
the pioneers used s - band ( @xmath92.2 ghz ) radio signals to communicate with the nasa deep space network ( dsn ) .
the s - band data is available from 26 m , 70 m , and some 34 m antennas of the dsn complex ( see baseline dsn configuration in the figure [ fig : pio - radio - science ] ) .
the 1-@xmath40 accuracy of s - band data is approximately 1 mm / s for a 60 second count interval after being calibrated for transmission media effects .
the dominant systematic error that can affect s - band tracking data is ionospheric transmission delays .
when the spacecraft is located angularly close to the sun , with sun - earth - spacecraft angles of less than 10 degrees , degradation of the data accuracy will occur .
s - band data is generally unusable for sun - earth - spacecraft angles of less than 5 degrees .
( see more details on the modern capabilities of the dsn , especially its radio science performance for extracting precision doppler signal demonstrated in @xcite . )
-10pt the pioneer radiometric data was received by the dsn in `` closed - loop '' mode , i.e. , it was tracked with phase lock loop hardware .
( `` open loop '' data is recorded to tape but not tracked by phase lock loop hardware . )
there are basically two types of data : doppler ( frequency ) and range ( time of flight ) , recorded at the tracking sites of the dsn as a function of ut ground received time @xcite . during their missions , the raw radiometric tracking data from pioneers were received originally in the form of intermediate data record ( idr ) tapes , which were then processed into special binary files called atdfs ( archival tracking data files , see description in @xcite ) , containing doppler data from the standard dsn tracking receivers ( see figure [ fig : pio - data - format - flow ] ) .
note that the `` closed - loop '' data constitutes the atdfs that were used in @xcite .
after a standard processing at the rmdc ( radio metric data conditioning group ) of jpl s navigation and mission design section , atdfs are transformed into odfs ( orbit data files , see description in @xcite ) .
a program called stripper is used to produce the odfs that are , at this point , the main product that is distributed to the end users for their orbit determination needs ( see more discussion on the conversion process in @xcite ) . at jpl , after yet additional processing , these odfs are used to produce sequentially formatted input / output files in navio format that is used by navigators while working with the jpl orbit determination program .
( note that the navio input / output file format is used only at jpl ; other orbit determination programs convert odfs to their particular formats . )
atdfs are files of radiometric data produced by the network operations control center ( nocc ) navigation subsystem ( nav ) ( see figure [ fig : pio - radio - science ] ) .
they are derived from intermediate data records by nav and contain all radiometric measurements received from the dsn station including signal levels ( agc = automatic gain control in dbm ) , antenna pointing angles , frequency ( often referred to simply as `` doppler '' ) , range , and residuals .
doppler data is often used to infer spacecraft radial motion relative to the tracking antenna .
data values in atdfs are reported at rates no higher than 10 per second . during their missions
, the pioneers received frequencies at s - band were recorded at a nominal sample time of one per either 10 , 60 , 300 , 600 or 1980 seconds .
in addition , for several years from 1980 to 1994 there were tracking passes scheduled for the jpl - led gravity wave experiments that delivered data at a smaller count time of 1 per second .
each atdf consists of all tracking data types used to navigate a particular spacecraft ( pioneers had only doppler data type ) and typically include doppler , range and angular types ( in s- , x- , and l - frequency bands ) , differenced range versus integrated doppler , programmed frequency data , pseudo - residuals , and validation data .
( unfortunately there was no range capability implemented on the pioneers @xcite .
early in the mission , jpl successfully simulated range data using modulation of sub - carrier signal and some range data is available , especially for planetary flybys .
later on , this technique of simulating range did was limited by the bandwidth of the sub - carrier signal and did not produce any range data . ) also , atdfs contain data for a single spacecraft , for one or more ground receiving stations , and for one or more tracking passes or days .
the atdf is described in section trk-2 - 25 of @xcite .
each atdf data record contains 117 parameters , stored in records of 288 bytes .
each atdf physical record is 2016 32-bit words in length ( 8064 bytes ) and consists of 28 72-word ( 288-byte ) logical records .
each atdf contains an integer number of these blocks .
the tracking data table format and content used until early 1997 are described by @xcite .
the atdf records are arranged in a sequence that consists of one file identification record , one transponder logical record , tracking data logical records in time order , and software / hardware end - of - file markers .
bit lengths of data fields are variable and cross word boundaries .
( further details of the atdf content and format are given in @xcite . )
each file contains 15 columns per data point specifying a time - tag , s - band , and x - band receiver parameters ( the latter not applicable for the pioneers . ) after standard processing at jpl s rmdc with the use of the stripper program ( see @xcite for more details ) , atdfs are transformed into odfs for use in determining spacecraft trajectories , gravity fields affecting them , and radio propagation conditions .
odfs contain radiometric data that has been converted from the atdf format @xcite .
each odf physical record is 2016 32-bit words in length and consists of 224 9-word logical records per data block .
the odf records are arranged in a sequence that consists of one file label record , one file identifier logical record , orbit data logical records in time order , ramp data logical records in time order , clock offset data logical records in time order , data summary logical records in time order and software / hardware end - of - file markers .
bit lengths of data fields are variable and cross word boundaries .
an odf usually contains most types of records , but may not have them all .
the first record in each of the 7 primary groups is a header record ; depending on the group , there may be from zero to many data records following each header .
the first task was to identify the data storage facilities with potential inventory of the pioneer doppler data . in general
, we expected to find data in four different locations , namely : 1 .
archive of radiometric data at rmdc at jpl : our hope was that radiometric data of pioneers 10 and 11 ( in all formats , namely idr , atdf , and odf ) may still exist at the various data storage archives at jpl and especially at the rmdc archive .
furthermore , we expected to find data in all possible media formats , including 7-track and 9-track magnetic tapes and also digital records on the rdc system .
the next logical place was the national space science data center ( nssdc ) at the nasa goddard space flight center , http://nssdc.gsfc.nasa.gov/planetary/pioneer10-11.html .
we knew that nssdc had copies of atdfs on magnetic tapes submitted there by john d. anderson of jpl , he being the principal investigator on the celestial mechanics experiments , as part of his agreement with nasa .
the data coverage was expected to be from 1978 to 1994 .
in addition to that , we knew that george null of jpl also submitted pioneer doppler data to nssdc prior to 1978 .
it was also know that , in addition to raw doppler data , he sent some orbital solutions and auxiliary information ( maneuvers , spin rates , initial conditions , etc . ) that also would be of interest to us .
we also expected that nasa ames research center ( arc ) , as the center with the project management responsibilities for all the pioneer missions , would have some information useful for our investigation .
we later realized that no raw orbital information was provided to arc ; instead , the trajectory solutions derived by jpl were used for all project needs .
4 . several facilities throughout the country were used to provide overflow storage for magnetic tapes in the 1970s and 1980s .
two of these federal data storage facilities are located in california : in long beach and in palo alto .
we have not yet tried to approach these data archival centers , but plan to do this in the near future .
all the sources of radiometric data listed above were used to recover the pioneer doppler data .
this data was present in different media formats ( magnetic tapes , floptical disks , digital formats on various computer platforms with different hard drive formatting standards ) and were written in different data formats ( i.e. idr , atdf and odf ) .
the goal was to recover as much of this data as possible .
the next section summarizes the results of our retrieval efforts . to recover pioneer radiometric data we first went to the nssdc with a request to provide us with copies of all available navigational data ( idr , atdf , and odf ) for the period 1978 - 1994
that was previously archived at nssdc ( done by j.d .
anderson and e.l .
lau of jpl ) .
the nssdc staff loaded the jpl - supplied idr and atdf tapes on their vms and unix machines .
however , we realized that all the atdf data files that we were receiving from nssdc were corrupted .
the problem was introduced by the process that was used to transfer the data across multiple computer platforms : an additional byte was inserted at record boundaries . in other words , every 8065th byte is a record marker and is not part of the atdf stream .
this problem was first encountered by c. markwardt during his study of the pioneer anomaly @xcite .
fortunately , we were able to recognize the problem early on and were able to fix most of the corrupted files following the procedure suggested by @xcite .
when jpl retrieved back the data using ftp , every 8065th byte was removed from the each file record .
the resulting data is good to use .
another data segment retrieved from the nssdc corresponded to the very beginning of the pioneer missions , namely the period @xmath91972 - 1976 ( files were archived by g. null of jpl ) .
unfortunately , it appears that not only are these data files corrupted , but they were also written in a format called `` type-66 '' ( or t66 ) which has not been used at jpl for more than 25 years , which adds unanticipated delay .
this is why we are still in the process of trying to recover some segments in the hope of adding this data to the new data set .
( note that we were not able to recover any odf data files from nssdc due to unknown / unidentified data formats . )
-2pt -5pt a significant volume of data ( both idr and atdf ) was found at rmdc on 9-track magnetic tapes ( 1978 - 1988 ) that were read on a still available minivax computer .
this segment of the data is in a good condition and is ready to use .
the retrieval of idr data was reasonably fast as each tape contained only one file .
however , the atdf 9-track tape had several atdfs per tape ( usually 10 - 15 files ) .
this increased the time needed to read a single tape on the minivax computer . with regular cleaning of the head of the magnetic tape reader and also cleaning the tapes themselves it usually took almost 90 minutes to read a single tape on this hardware , which was built in the 1980s . with an inventory of more than 200 tapes this was a major effort during the summer of 2005 .
there were a few additional and unexpected finds of pioneer data at jpl .
several historical atdfs for both pioneer 10 and 11 were found and recovered from the rmdc s archival optical disks . also at rmdc
, we found spr ( system performance record ) data ( format trk-2 - 15a ) .
the raw spr data format was used to replace the idr format , and it was utilized after 1985 when jpl / nav upgraded the system from univac to dec / vms hardware .
we were fortunate to recover several additional historical atdfs for both pioneers from the rmdc s system archive tapes ( sat ) ( written on 9-track and 4 mm tapes ) .
these files were also added to the overall file inventory .
the resulting atdf and odf retrieval methods are shown in figure [ fig : recovery - flow - chart ] .
in addition to the data already recovered , we found and recovered some early odf data ( 1973 - 1974 ) that was still available in the personal data archives of our jpl colleagues who worked with pioneer data for other purposes ( e.g. , development of the solar system ephemerides ) .
in fact , we were able to recover several odf files written in the navio input / output form .
these files , obtained from several jpl users , were re - converted back to the odf data format . finally , we have added the data that was already analyzed by @xcite that corresponds to the period of 1987 - 1998 ( for pioneer 10 ) and 1987 - 1990 ( for pioneer 11 ) .
lastly , we added the most recent data for the period 1998 - 2002 .
the resulting 1972 - 2002 data has some redundancy , but mostly it is a very complete data set assembled for the first time . after necessary certification at jpl
this same set will be available for distribution .
the currently available expanded data set includes the following data segments for each spacecraft : * pioneer 10 : the entire available data set covers mission events from mid-1973 ( including jupiter encounter data ) to the last time a pioneer 10 contact returned telemetry data , 27 april 2002 .
this interval spanned heliocentric distances from @xmath41 au to 87 au .
the total @xmath930-year pioneer 10 data set might have @xmath960,000 data points and is about 20 gb in size .
we also have most of the information on maneuvers , spin rate , and initial conditions .
* pioneer 11 : the entire available data set covers the mission from mid-1974 to late 1994 ( including both jupiter and saturn encounters ) .
this interval spanned heliocentric distances from @xmath94 au to 33 au .
the total @xmath920-year pioneer 11 data set might have @xmath950,000 data points and is about 15 gb in size .
we also have most of the information on maneuvers , spin rate , and initial conditions . to summarize , there exists @xmath930 years of pioneer 10 and @xmath920 years of pioneer 11 data ,
most of which had never been well studied for our purposes .
we are still trying to increase the percentage of the f2 and f3 doppler data in the overal retreived doppler dataset , as f1 data is useless for our analysis .
for this we are working together with rmdc in an attempt to improve the output of the atdf - to - odf file conversion effort , which we expect to complete by january 2006 . by the same time we also hope to close the t66 issue and prepare all input information critical for the upcoming re - processing of the doppler data for the entire pioneer 10 and 11 trajectories .
to re - process the complete pioneer trajectories , one would first have to ( re)edit and ( re)process the entire data span ( from 1972 to 2002 ) , using the same editing strategy , initial conditions , and parameter estimation and noise propagation algorithms .
one could also process the high rate doppler data ( i.e. , 1 record per second ) that previously was used very little .
the high rate data can be used to better determine a spacecraft s spin rate and also to improve the maneuver data file information .
also , the spin rate change was found to be highly correlated with a small but significant spacecraft - generated force , probably from gas leaks @xcite .
therefore , one can also use the high - rate data to estimate and/or calibrate valve gas leaks and all the maneuvers . note that the use of the spacecraft telemetry data may be critical in addressing this issue ( see discussion of this possibility in section [ sec : mass - expulsion ] ) .
since the previous analysis @xcite , physical models for the earth s interior and the planetary ephemeris have greatly improved .
this is due to progress in gps- , slr- , llr- and vlbi - enabling technologies , doppler spacecraft tracking , and new radio science data processing algorithms .
one would have to write and/or update existing orbit determination programs using these latest earth models ( adopted by the iers ) and also using the latest planetary ephemeris .
this will improve the solutions for the dsn ground station locations by two orders of magnitude ( 1 cm ) over that of the previous analysis .
additionally , this will allow a better characterization of not only the constant part of any anomalous acceleration , but also of the annual and diurnal terms detected in the pioneer 10 and 11 doppler residuals @xcite .
as pointed out above , we were able to recover almost the entire set of doppler data obtained from both pioneers 10 and 11 .
we were also able to obtain the entire archive of the spacecraft telemetry data in a form called master data records ( mdr ) for both craft . because of their potentially high value in the investigation of the pioneer anomaly , in this section
we discuss the purpose , means of delivery , storage , processing and possible use of the mdrs .
all data received from the pioneer 10 and 11 spacecraft by the nasa deep space network ( dsn ) was initially stored in the form of mdrs .
these records contained information about the dsn station and reception characteristics , in addition to the actual , `` raw '' data records themselves , which contained both scientific measurements and engineering telemetry , as received from the spacecraft .
( as an example , figure [ fig : sensors ] shows locations of several thermal sensors in the spacecraft instrument compartment . )
-10pt -10pt normally , mdrs are seen to be of little use once the relevant information is extracted from them .
scientific measurements are extracted , packaged in the appropriate form , and sent to the corresponding experimenters for further processing and evaluation .
engineering telemetry is used by the spacecraft operations team to control and guide the spacecraft .
mdrs are not necessarily considered worth preserving once the scientific data has been extracted , and the engineering telemetry is no longer needed for spacecraft operations .
indeed , the mdr retention schedule prescribed that the tapes be destroyed after 7 years .
fortunately , most of the mdrs for the pioneer 10 and 11 projects have been preserved nevertheless .
long before we ( lrk and vtt ) heard about the pioneer anomaly investigation , we recognized the potential value of the mdrs for educational purposes ; we envisioned an educational web site where a visitor could observe the pioneers as they traveled through the solar system , while viewing telemetry information that showed the health status and operational characteristics of the two spacecraft .
for this reason , lrk endeavored to preserve the complete set of mdrs for the two projects and copied the data to modern media .
meanwhile , vtt began developing pc - based tools for processing the mdrs and extracting information . far beyond the original expectations ,
this effort is now seen to be of value for the investigation of the pioneer anomaly , as the mdrs , specifically the telemetry data contained therein , may help in the construction of an accurate model of the spacecraft during their decades long journey , including a precise thermal profile , the time history of propulsion system activation and usage , and many other potential sources of on - board disturbances . in the early days of pioneer operations , the original mdrs for the pioneer 10 and 11 missions
were copied to magnetic tape at the dsn receiving stations , and then sent to the appropriate nasa office for further processing . after more than 20 years of storage ,
the medium started to show signs of deterioration and all of the recoverable raw data was copied to 128 megabyte magneto - optical ( `` floptical '' ) disks .
these disks were readable by a microvax using a digital corporation rwz21 magneto optical scsi disk drive .
digital corporation is now part of compaq , the available vax 4000 - 300 computer will soon cease to exist , and the scsi rwz21 drives will not be usable to read these `` floptical '' disks . at present ,
a near complete set of mdr `` floptical '' disks still exists , along with equipment capable of reading this media . during the course of 2004 and 2005
, kellogg has copied the files from `` floptical '' disks to a modern computer , and also created dvd - roms for distribution .
the contents of the available mdr disks can be summarized as follows : pioneer 10 : 155 disks x 128 mb ( 16.33 gb ) + pioneer 11 : 217 disks x 128 mb ( 23.01 gb ) all mdr files are named using a common naming convention .
each file contains the mdrs for one spacecraft for an entire day .
the file name begins with the letter ` m ' , followed by a two - digit spacecraft identifier ( 23 for pioneer 10 , 24 for pioneer 11 ) , and a 5-digit date ( 2 digits for the year , 3-digit doy . )
the filename extension is .mdr .
thus , m2495003.mdr , for instance , would be the name of the pioneer 11 file for january 3 , 1995 . when kellogg copied the files to modern computers , he established a separate directory for each `` floptical '' disk copied .
he also used a consistent naming convention : the directory name began with the two - digit spacecraft identifier , followed by the letter ` p ' , a two - digit year value , and a two - digit sequence number that simply identified the number of the disk for the given year .
so for instance , 23p9602 would be disk 2 for pioneer 10 , for the year 1996 .
the total amount of data stored in these mdr files is approximately 40 gb . according to the original log sheets that record the transcription from tape to `` floptical '' media , only a few days worth of data is missing , some due to magnetic tape damage .
one notable exception is the `` jupiter encounter period '' of pioneer 10 . according to the transcription log sheets ,
doy 332 - 341 from 1973 were not available at the time the `` floptical '' disks were made .
kellogg s investigation revealed that the tapes in question may have been misplaced long before the transcription has taken place .
unfortunately , it is unlikely that those tapes will ever be located , and even if they re found , chances are that they will no longer be readable due to media deterioration .
other significant periods of missing data are listed in table [ tb : missing_mdr ] .
it is not known why these records are not present , except that we know that very few days are missing due to unreadable media ( i.e. , the cause is missing , not damaged , tapes . )
-12pt .pioneer 10/11 missing mdrs ( periods of missing data shorter than 1 - 2 days not shown . ) [ cols="<,<,<",options="header " , ] [ tb : power ] -10pt to utilize this data for our investigation , one would have to reconstruct the direction of heat flow : absorption and re - emission by , and reflection off the craft surfaces .
we see this as the main challenge of the upcoming analysis . how to convincingly use a reading at several points on the craft to reconstruct the heat dynamics ?
( similar questions exist for gas leaks and all other parameters . )
if we do nt see their signature in the doppler signal , we should be able to put an upper limit on the potential contribution from these sources to the pioneer anomaly .
similarly , the temperature data in the mdrs must put an upper limit on the changes of the optical surfaces of the rtgs ( i.e. their emissivity and reflectivity ) .
because of the complexities involved , clearly more resources are needed ; we invite researchers to join us for this part of the upcoming study .
another suggestion related to the rtgs is based on the idea that during the early parts of the missions , there might have been a differential change of the radiant emissivity of the sun - pointing sides of the rtgs as compared to the sides facing deep space .
note that , especially closer to the sun , the inner sides were subjected to the solar wind . on the other hand ,
the outer sides were sweeping through the solar system dust cloud .
these two processes could result in different levels of degradation of the optical surfaces of the rtgs . in turn
this degradation could result in asymmetric patterns of heat radiation away from the rtgs in the fore / aft directions along the spin axis .
therefore , it can be argued that such an anisotropy may have caused the anomaly . using the actual design information on the rtgs and a sophisticated model of the fin structure ,
@xcite estimate an upper limit of 6.12w to be the uncertainty in the contribution from the differential emissivity of the rtgs to the anomalous acceleration of the craft .
this value , in turn , results in an acceleration uncertainty of @xmath42 the authors also point out the significance of radioactive decay for this mechanism and especially its expected contribution to the decrease of the magnitude of @xmath23 , which has not been observed . the effect of the radioactive decay on the thermal output of the rtgs can be directly observed from telemetry .
all rtgs had a temperature sensor located at the root of one of the radiating fins ( `` fin root temperature '' ) .
preliminary analysis suggests that these readings are in agreement with what would be expected from the radioactive decay of the plutonium fuel ( see figure [ fig : root - fin ] ) .
one can try to detect changes in the measurements of fin root temperature during each of the three planetary encounters and see if the rtgs had developed an anisotropic radiating pattern .
( note that the same test can be conducted for the electrical energy dissipation . )
thus , if the temperatures at the fins change during the fly - bys , this may indirectly indicate an alteration of the optical properties of the rtgs during the planetary encounters .
this should allow us to ask the question : `` if the system survived flawlessly than why would it change long after the encounter ? ''
one can also look at the different segments of the thermal history in an attempt to identify different heat decay rates .
however , the planetary flybys might be the strongest test for the system s design and operational longevity .
if the mechanisms above were treated as separate constituents to the anomalous accelerations , one can ask what if a combination of them is responsible for the effect .
it has been recently suggested that most , if not all , of the unmodeled acceleration of pioneer 10 and 11 is due to an essentially constant supply of heat coming from the central compartment , directed out the front of the craft through the closed louvers @xcite .
this is a more subtle version of an earlier proposal @xcite calling on the total electrical power as a mechanism . that proposal was argued against because of the observed lack of decay of the acceleration with time @xcite . for this mechanism to work , one essentially needs to find a constant supply of heat radiated off the back of the spacecraft that would produce a thermal recoil force with the well established properties ( discussed in section [ sec : anomaly ] ) .
however , the assumption of constancy came against the actual design of the pioneers and experimental data on the performance of the vehicles @xcite .
in fact , the lack of constancy of heat dissipated during the longest doppler segment analyses ( i.e. 11.5 years of the pioneer 10 data ) invalidated the hypothesis . however , this claim can and will be further investigated with the newly recovered data , both doppler and telemetry ( see tables [ tab : telemetry],[tb : power ] ) . the newly acquired data ( both doppler and telemetry ) can significantly contribute to addressing this possibility .
for one , we also have a much larger segment of doppler data that will be used to analyze these heat dissipation processes on the vehicles .
in addition , we now have the actual design , fabrication , testing , pre- and in - flight calibration data that characterize the pioneer craft performance for the duration of their missions .
this data can tell precisely at what time the louvers were open and closed , when a certain instrument was powered on and off , what was the performance of the battery , shunt current and all electric parts of the spacecraft .
this is truly a unique data set that will be used to reconstruct the physics of the anomalous acceleration , if it is due to on - board systematics .
in fact we have all the detailed information on properties of the spacecraft and the data needed to reconstruct the behavior of its major components , including electrical power and thermal subsystems .
this information may be used in developing a thermodynamical model of the spacecraft that would help us to establish the true thermal and electrical power dissipation history of the vehicles and also correlate major events on the pioneers ( such as powering `` on '' or `` off '' certain instruments or performing a maneuver ) with the available doppler data ; this possibility is addressed in section [ sec : thermal - model ] .
another possible on - board systematic error may be coming from the expulsion of the he being created in the rtgs from the @xmath43-decay of @xmath44pu ( see discussion in @xcite ) .
the pioneer rtgs were designed so that the he pressure is not totally contained within the pioneer heat source over the life of the rtgs .
instead , the pioneer heat source contains a pressure relief device that allows the generated he to vent out of the heat source and into the thermoelectric converter .
the thermoelectric converter housing - to - power output receptacle interface is sealed with a viton o - ring .
the o - ring allows the helium gas within the converter to be released by permeation into the space environment throughout the mission life of the pioneer rtgs . for this mechanism to work
, what one needs is all he produced to preferentially leave the rtgs in one direction at a rate of 0.77g / yr . assuming a single elastic reflection , @xcite ruled out helium permeating through the o - rings as the cause of @xmath23 .
they also give the estimate for acceleration contribution in the direction away from the sun due to this systematic expulsion @xmath45 it is highly unlikely that this mechanism is responsible for the pioneer anomaly seen on both spacecraft ; the use of mdrs can further support this conclusion .
the possible use of a similar strategy to the above to study the effect of propulsive mass expulsion due to gas leakage has to be assessed .
it is known that , although this effect is largely unpredictable , many spacecraft have experienced gas leaks producing accelerations on the order of @xmath46m / s@xmath2 .
gas leaks generally behave differently after each maneuver .
the leakage often decreases with time and becomes negligible .
( see more discussion of these effects in @xcite . ) -6pt -5pt -6pt -5pt -6pt -5pt gas leaks can originate from the pioneers propulsion system , which is used for mid - course trajectory maneuvers , for spinning up or down the spacecraft , and for orientation of the spinning spacecraft .
( consult @xcite for design of the pioneer propulsion and attitude control systems . ) by studying time histories of the spin rates of the pioneers , @xcite estimated the effect of the gas leakage mechanism on the anomalous acceleration .
this mechanism was found to produce an acceleration uncertainty of @xmath47 furthermore , they concluded that the gas leak mechanism is very unlikely to be the explanation for the anomalous acceleration .
this is because it is difficult to understand why this effect , being a stochastic variable obeying a poisson distribution , would affect pioneers 10 and 11 by the same amount @xcite .
the new analysis can refine the conclusions above .
we have three readings from the propellant tank on board that can help establish the propellant loss history of the two spacecraft , namely propellant tank temperature ( figure [ fig : thrusters - c-130 ] ) , propellant supply pressure ( figure [ fig : thrusters - c-210 ] ) , and propellant supply temperature ( figure [ fig : thrusters - c-327 ] ) .
we also have several temperature readings from the thruster clusters and individual thrusters . not only may this new information be used to precisely establish the time of the maneuvers
, it can also provide some insight as to why the two identically designed systems behaved so differently , especially when changes to their spin rates are concerned ( see figures 11 and 12 of @xcite . )
the corresponding analysis had been initiated and results will be reported . in this section
we will argue that , based on the information provided by the mdrs , one can develop a high accuracy thermal / electrical / dynamical model of the pioneer spacecraft .
such a model can be used to further improve our understanding of the anomalous acceleration and especially to study contribution from the on - board thermal environment to the anomaly . in this section
we will comment on our preliminary evaluation of what is needed to develop such an accurate model .
it is clear that a thermal model for the pioneers spacecraft would have to account for all heat radiation within the spacecraft and out to deep space .
specifically , this model would have start from the rtgs and include description of the heat radiated , absorbed , reradiated , and conducted by each and every instrument , all cables , support structures and etc .
the objective is to develop a broadband ( ir and optical ) thermal map of the spacecraft s exterior , accurate to few degrees kelvin . as input
, this model will use electrical telemetry readings from the on - board sensors that were installed on the spacecraft , while temperature readings ( see figure [ fig : sensors ] ) can be used for its calibration .
we will then attempt to verify this model by using the radiometric doppler data from the attitude maneuvers performed at the earlier mission phases .
it is worth noting that no other mission in the past used telemetry data to improve its navigational capabilities ; what we are going to implement with the pioneers is a novel navigational technique .
pioneer 10/11 may be the first missions to benefit from this technique , designed to reduce the impact of on - board sources to non - gravitational accelerations on the accuracy of trajectory reconstruction .
our proposal uses the fact that the pioneers already have very good `` built - in '' navigational capabilities ( i.e. guidance , navigation and control ) ; our goal is to reach the absolute limit allowed by these capabilities . as a quantitative objective , we hope to improve acceleration contribution due to on - board systematic sources to the level below @xmath48 m / s@xmath2 . to reach the stated level of accuracy , one would have to adjust navigational data by using the developed thermal / electrical / dynamical model of the spacecraft . to achieve this
we will utilize the following available information : 1 .
dynamical properties of the spacecraft , i.e. geometry , dimensions , and weights of all the components of the spacecraft , including cabling , fuel and thermal blankets . in short
, we will need to have precise mass and power budget data for the entire craft .
ideally , we would prefer to use not only design numbers , but rather actual values established during the final pre - launch tests at the kennedy space center , florida ( ksc ) .
this information is still to be identified and retrieved .
2 . thermal , electric and optical ( infrared ) properties of the spacecraft materials used to fabricate all major components of the spacecraft , including thermal blankets and cabling .
again , we would prefer to have these values from the tests at the ksc .
note that we will have access to the initial values of these parameters ; their degradation during the 30 years in space is unknown .
there is very limited information available from long - term exposure tests .
furthermore , for obvious reasons ( tests conducted in a different environment , e.g. , in low earth orbit ( loe ) , for different durations ) , even these results will not be directly applicable to our model .
3 . we will benefit from the design information of the spacecraft , especially information concerning their thermal , power , propulsion , and communication subsystems .
we will use the knowledge on the location of each and every telemetry sensor placed on the spacecraft and use all the recently retrieved pioneer telemetry data .
this information will be used to build and calibrate the thermal model of the spacecraft . in particular
, the following telemetry information will be critical for the upcoming investigation : 1 .
rtg fin root temperature , as an indicator of rtg thermal output 2 .
rtg current and voltage , providing a measure for total electrical power 3 . all thermal data from sensors on the spacecraft and instrument compartment 4 .
temperatures and pulse counts from thruster cluster assemblies 5 .
battery current and temperature 6 .
shunt current 7 .
radio beam gain and transmitted power , twt amplifier temperature 8 .
propellant tank and supply lines temperature 4 .
we would have to analyze the design and thermal properties of the snap-19 rtgs used in the pioneers .
much information is already available , but additional data on their thermal behavior will be very helpful in building a 3-dimensional thermal radiation pattern .
once the thermal model is completed , it must be calibrated for in - flight conditions .
for this , we may need to analyze all the attitude control ( i.e. conscan ) maneuvers in the earlier mission phases @xcite .
the earlier phases had significant variations in the sun - earth line to spin axis angle .
these variations will help us to calibrate the effect of thermal recoil forces on the radiometric data received from the craft and , thus , to precisely evaluate the magnitude of the thermal recoil force .
the closer the craft to the earth is , the better the ultimate accuracy of the model will be .
this is due to the fact that we know the properties of the solar radiation and thermal environment in the inner solar system to a much better accuracy than in the deep space . in fact , it will be hard to do the required calibration after the jupiter flyby , and it will be impossible at distances greater than 10 au . clearly we will benefit from the larger size of the earth - probe - sun angle , which is getting much smaller after 10 au .
after 20 au , both the sun and the earth will be in the same field of view , making it difficult to decouple any thermally induced force from other physical effects .
however , given the significant uncertainties on the degradation of the thermodynamical and optical properties of the materials used to build the pioneers , the outcome of this model is difficult to predict .
nevertheless , we expect that , given the simplicity of the pioneers design , one should be able to build a model with an uncertainty of no more than 15% .
we have initiated this work with anticipated involvement of several navigational and thermal engineers at jpl ; results of this work will be reported .
the plan outlined here is ambitious .
before we embark on this complex exercise , we shall also attempt to build a simpler ( but enthusiastically not `` simplistic '' ) thermal model of the pioneers based on available telemetry data .
such a model would necessarily utilize simplifying assumptions on the spacecraft s geometry , and ignore contributions from small components ( e.g. , structural elements , struts , cables as heat sources ) .
nevertheless , it is possible that even such a model may lead to a much better understanding of the contribution of spacecraft systematics to the anomalous acceleration , and may also guide us to better direct our efforts as we carry out a more elaborate effort .
on march 2 , 2002 nasa s dsn made the last contact with pioneer 10 and confirmed that the spacecraft was still operational thirty years after its launch on march 3 , 1972 ( ut ) .
the uplink signal was transmitted on march 1 from the dsn s goldstone , california facility and a downlink response was received twenty - two hours later by the 70-meter antenna at madrid , spain . at this time the spacecraft was 11.9 billion kilometers from earth at about 79.9 au from the sun and heading outward into interstellar space in the general direction of aldebaran at a distance of about 68 light years from the earth , and a travel time of two million years .
although science operations were officially terminated on march 31 , 1997 , pioneer s onboard rtgs still provided just enough electrical power for operations of the radio transmitter and receiver systems and for the geiger tube telescope experiment of the university of iowa .
after 1997 , nasa arc continued to support occasional tracking of pioneer 10 for development and testing of communications technology that might be used for a future interstellar probe mission .
pioneer 10 s signal was also used as a standard test source for seti radio telescope studies ( i.e. project phoenix , see details at http://www.seti.org/ ) . by 2005 ,
the existence of the pioneer anomaly is no longer in doubt .
further , after much understandable hesitancy , a steadily growing part of the community has concluded that the anomaly should be subject to further investigation and interpretation .
the results of the investigation of the pioneer anomaly would be win - win ; improved navigational protocols for deep space at the least , exciting new physics at the best . a strong international collaboration ( e.g. , http://www.issi.unibe.ch/teams/pioneer/ ) that extends in broader areas of fundamental physics , technology and mission design
is an additional outcome of the discussed program of the study of the pioneer anomaly @xcite .
concluding , we would like to emphasize one immediate impact of the study of the pioneer anomaly , this time on the pioneer 10 mission itself .
recently it became clear that there exists one last opportunity to contact pioneer 10 in the deep space .
specifically , in february - march 2006 , the mutual motions of the earth and the spacecraft will again put the earth in the field of view of pioneer 10 s antenna , thus making it possible to establish a radio contact with the craft . at this time
, pioneer 10 will be at a heliocentric distance of @xmath990.1 au , moving at a speed of nearly 12.1 km / s with round trip light time from the earth of almost 25 hours ( thus , the same dsn antenna , dss 14 at goldstone , is planned for the operations . )
this would be the last time when dsn will initiate contact with pioneer 10 , as the on - board power system on the craft is at its limits .
this intriguing possibility to re - acquire a coherent doppler signal from pioneer 10 is currently being investigated at jpl and results will be reported .
we would like to express our gratitude to our many colleagues who have either collaborated with us on this manuscript or given us their wisdom .
firstly we must acknowledge the many people who have helped us with suggestions , comments , and constructive criticisms .
we specifically thank john d. anderson , sami asmar , and timothy p. mcelrath , who provided us with very valuable comments while this manuscript was in preparation .
invaluable information on the history , spacecraft design and mission operations of the pioneers 10 and 11 , as well as the structure of their telemetry data , came from lawrence lasher and david lozier of the nasa ames research center .
we also thank belinda arroyo , peter j. breckheimer , john e. ekelund , jordan ellis , scott e. fullner , gene l. goltz , olga king , george d. lewis , robert a. jacobson , margaret medina , neil mottinger , teresa thomas of jpl for their help in obtaining , understanding and conditiononing of the pioneer doppler data .
we thank moustafa t. chahine , william folkner , ulf e. israelsson , tomas j. martin - mur , thomas a. prince , and michael m. watkins of jpl for encouragement and stimulating discussions regarding the pioneer doppler data retrieval effort .
john f. cooper and sharlene rhodes of nssdc provided pivotal help in the retrieval and understanding of the historical pioneer doppler data archived with nssdc .
louis k. scheffer of cadence design systems contributed with useful observations on the thermal modeling and analysis .
we especially thank the planetary society for support and , in particular , louis d. freidman , charlene m. anderson , and bruce betts for their interest , stimulating conversations and encouragement .
this work was partially performed at the international space science institute ( issi ) , bern , switzerland , when two of us ( sgt and vtt ) visited issi as part of an international team program . in this respect
we would like to thank roger m. bonnet , vittorio manno , brigitte fasler and saliba f. saliba of issi for their hospitality and support .
the work of sgt was supported in part by the office of the jpl chief scientist under the research and technology development program and carried out at the jet propulsion laboratory , california institute of technology , under a contract with the national aeronautics and space administration .
asmar , s.w . ,
armstrong , j.w .
, iess , l. , and tortora , p. , `` spacecraft doppler tracking : noise budget and accuracy achievable in precision radio science observations , '' _ radio science * 40 * _ , rs2001 ( 2005 ) .
anderson , j.d .
, `` quarterly report to nasa / ames research center , _ celestial mechanics experiment , pioneer 10/11 , _ '' jpl interoffice memorandum , 8 july 1992 ; _ celestial mechanics experiment , pioneer 10/11 , _ 22 july 1992 .
also see the later quarterly report for the period 1 oct .
1992 to 31 dec .
1992 , dated 17 dec .
1992 , letter of agreement arc / pp017 .
the first and last , specifically , contain figure 7 of ref .
@xcite .
anderson , j.d . , laing , p.a . , lau , e.l . , liu , a.s . , nieto , m.m . , and turyshev , s.g . , `` indication , from pioneer 10/11 , galileo , and ulysses data , of an apparent anomalous , weak , long - range acceleration , '' _ phys .
lett . _ * 81 * 2858 ( 1998 ) , gr - qc/9808081 .
anderson , j.d . ,
laing , p.a . , lau , e.l . ,
liu , a.s . ,
nieto , m.m .
, and turyshev , s.g . ,
`` anderson et al .
reply ( to the comment by katz on pioneer 10/11 ) , '' _ phys .
lett . _ * 83 * , 1893 ( 1999 ) , gr - qc/9906112 .
anderson , j.d .
, laing , p.a . , lau , e.l .
, liu , a.s . , nieto , m.m . , and turyshev , s.g .
, `` anderson et al .
reply ( to the comment by murphy on pioneer 10/11 ) , '' _ phys . rev .
lett . _ * 83 * , 1891 ( 1999 ) , gr - qc/9906113 .
anderson , j.d . ,
laing , p.a . , lau , e.l . ,
liu , a.s . ,
nieto , m.m . , and turyshev , s.g .
, `` study of the anomalous acceleration of pioneer 10 and 11 , '' _ phys .
d. _ * 65 * , 082004/1 - 50 ( 2002a ) , gr - qc/0104064 . deep space network , system requirements detailed interface design , jpl document 820 - 13 , rev . a , jet propulsion laboratory , pasadena , ca ( november 1 , 1991 ) , see _ dsn tracking system interfaces , archival tracking data file interface , trk-2 - 25 . _ an example of an atdf for ulysses mission ( part of sfoc - nav-2 - 25 ) at http://www.igpp.ucla.edu/pds3/uly_5101/data/atdf/2040040a.lbl deep space network , system requirements detailed interface design , jpl document 820 - 13 , rev .
a , jet propulsion laboratory , pasadena , ca ( november 1 , 1991 ) , see chapter on _ dsn tracking system interfaces , orbit data file interface , trk-2 - 18 .
_ an example of an odf for near mission ( part of sfoc - nav - trk-2 - 18 ) may be found at + http://pdssbn.astro.umd.edu/data3/near/nreros_2001/erosfb/odf/8357357c.lbl defrere , d. , rathke , a. , `` pioneer anomaly : what can we learn from lisa ? '' in proceedings of `` lasers , clocks , and drag - free , '' zarm , bremen , germany , 30 may 1 june 2005 , ed . c. lmmerzahl , h. dittus , s.g .
turyshev , to be published , ( 2006 ) gr - qc/0509021 dittus , h. , turyshev , s.g . ,
lmmerzahl , c. , theil , s. , foerstner , r. , johann , u. , ertmer , w. , rasel , e. , dachwald , b. , seboldt , w. , hehl , f.w . ,
kiefer , c. , blome , h .- j . ,
kunz , j. , giulini , d. , bingham , r. , kent , b. , sumner , t.j .
, bertolami , o. , pramos , j. , rosales , j.l . ,
christophe , b. , foulon , b. , touboul , p. , bouyer , p. , reynaud , s. , brillet , a. , bondu , f. , samain , e. , de matos , c.j .
, erd , c. , grenouilleau , j.c .
, izzo , d. , rathke , a. , anderson , j.d . ,
asmar , s.w . ,
lau , e.e . ,
nieto , m.m . , and mashhoon , b. , `` a mission to explore the pioneer anomaly , '' in proceedings of 2005 eslab symposium `` trends in space science and cosmic vision 2020 , '' esa / estec , noordwijk , the netherlands , 19 april 2005 , esa publication sp-588 , 3 - 10 ( 2005 ) , gr - qc/0506139 .
fimmel , r. o. , swindell , w. , and burgess , e. , _ pioneer odyssey : first into the outer solar system _ , nasa publication sp-349/396 ( nasa , washington , d.c . , 1977 ) , see electronic version at http://history.nasa.gov/sp-349/sp349.htm turyshev , s.g .
, laing , p.a . , lau , e.l . ,
liu , a.s . , and nieto , m.m .
: 1999 , `` the apparent anomalous , weak , long - range acceleration of pioneer 10 and 11 . '' in : _ gravitational waves and experimental gravity
_ , proceedings of the xviiith workshop of the rencontres de moriond , les arcs , savoi , france ( january 23 - 30 , 1999 ) , ed . by j. dumarchez and j. tran thanh van ( world publishers , hanoi - vietnam , 2000 ) .
481 - 486 ( 1999 ) , gr - qc/9903024 .
turyshev , s.g . , nieto , m.m . , and anderson , j.d , `` lessons learned from the pioneers 10/11 for a mission to test the pioneer anomaly . ''
`` 35th cospar scientific assembly , '' july 18 - 24 , 2004 , paris .
france . to be published in _ advances in space res .
_ , ( 2004 ) , gr - qc/0409117 .
turyshev , s.g . , nieto , m.m . , and anderson , j.d , `` a route to understanding of the pioneer anomaly . ''
`` the xxii texas symposium on relativistic astrophysics , '' stanford university , dec .
13 - 17 , 2004 , edited by p. chen , e. bloom , g. madejski , and v. petrosian .
slac - r-752 , stanford e - conf # c041213 , # 0310 , ( 2005a ) , eprint : http://www.slac.stanford.edu/econf/c041213/ , gr - qc/0503021 .
turyshev , s.g . , nieto , m.m . , and anderson , j.d .
, `` study of the pioneer anomaly : a problem set . '' _ amer .
j. phys . *
73 * _ , 1033 - 1044 ( 2005b ) , physics/0502123 .
turyshev , s.g . ,
nieto , m.m . , and anderson , j.d . , `` the pioneer anomaly and its implications . ''
xxist iap colloquium on `` mass profiles and shapes of cosmological structures '' at the institute dastrophysique de paris , paris , france , 9 july 2005 , to be published by edp sciences , paris , france , ( 2006 ) , gr - qc/0510081 . | the pioneer 10/11 spacecraft yielded the most precise navigation in deep space to date .
however , their radiometric tracking data has consistently indicated the presence of a small , anomalous , doppler frequency drift .
the drift is a blue shift , uniformly changing with a rate of @xmath0 hz / s and can be interpreted as a constant sunward acceleration of each particular spacecraft of @xmath1 m / s@xmath2 ( or , alternatively , a time acceleration of @xmath3 s / s@xmath2 ) .
this signal has become known as the pioneer anomaly ; the nature of this anomaly remains unexplained .
we discuss the current state of the efforts to retrieve the entire data sets of the pioneer 10 and 11 radiometric doppler data .
we also report on the availability of recently recovered telemetry files that may be used to reconstruct the engineering history of both spacecraft using original project documentation and newly developed software tools .
we discuss possible ways to further investigate the discovered effect using these telemetry files in conjunction with the analysis of the much extended pioneer doppler data . in preparation for this new upcoming investigation , we summarize the current knowledge of the pioneer anomaly and review some of the mechanisms proposed for its explanation .
we emphasize the main objectives of this new study , namely analysis of the early data that could yield the true direction of the anomaly and thus , its origin , analysis of planetary encounters , that should tell more about the onset of the anomaly ( e.g. pioneer 11 s saturn flyby ) , analysis of the entire dataset , that should lead to a better determination of the temporal behavior of the anomaly , comparative analysis of individual anomalous accelerations for the two pioneers with the data taken from similar heliocentric distances , the detailed study of on - board systematics , and development of a thermal - electric - dynamical model using on - board telemetry .
the outlined strategy may allow for a higher accuracy solution for the anomalous acceleration of the pioneer spacecraft and , possibly , will lead to an unambiguous determination of the origin of the pioneer anomaly .
slava g. turyshev,@xmath4 viktor t. toth,@xmath5 larry r. kellogg,@xmath6 eunice .
l. lau,@xmath4 and kyong j. lee@xmath4 _ @xmath4jet propulsion laboratory , california institute of technology , + 4800 oak grove drive , pasadena , ca 91109 , usa + @xmath5vttoth.com , 3 - 575 old st patrick st . , ottawa on k1n 9h5 , canada + @xmath6nasa ames research center , moffett field , ca 94035 , usa _ |
the gl 876 planetary system ranks with the most remarkable dicoveries that have emerged from the first decade of extrasolar planet detections . as described by rivera et al .
( 2005 ) , laughlin et al .
( 2005 ) , butler et al .
( 2001 ) marcy et al .
( 2001 ) , delfosse et al .
( 1998 ) and marcy et al .
( 1998 ) , there are a number of reasons why the gl 876 planets are extraordinary .
d " , with its 1.974 day orbit and @xmath4 is the lowest - mass exoplanet known to orbit a nearby main - sequence star ( rivera et al . 2005 ) . planets c " and b " are the only jovian - mass companions known to orbit an m - dwarf , and their clear participation in 2:1 mean - motion resonance gives important dynamical clues to the formation and evolution of the system ( e.g. lee and peale 2002 ) .
furthermore , the gl 876 planets are the closest exoplanets that have been reliably characterized .
their radial velocity amplitudes induce an overall signal - to - noise of nearly 100 , and they have have been observed for well over a decade . the detection of gl 876 d by rivera et al .
( 2005 ) underscored the rapid development of the doppler technique .
the detection of planets having only a few earth masses is the culmination of advances described by butler et al .
( 1996 ) , marcy & butler ( 1998 ) , marcy , cochran , & mayor ( 2000 ) , marcy et al .
( 2005 ) , and lovis et al .
( 2006 ) . nevertheless ,
while radial velocity precision has improved to better than 1 m / s , keplerian orbital fits to doppler velocities give estimates of @xmath5 rather than @xmath6 . in order to ascertain the true mass of a planet
, one must obtain an independent measure of the orbital inclination , @xmath7 , which is most easily measured when the planet is observed to transit across the face of the parent star . the now - celebrated companion to hd 209458 provided the first example of a transiting extrasolar planet , and this body has generated an incredible variety of observational and theoretical investigations .
hd 209458b was initially detected with the radial velocity method , and was then discovered to transit on the basis of follow - up photometric observations at the predicted transit times ( henry et al . 1999 , 2000 ;
charbonneau et al . 2000 ) .
hd 209458b s well - sampled light curve , when combined with high - precision radial velocity data , has permitted precise measurements of the planet s basic attributes ( e.g. mazeh et al . 2000 , brown et al .
2001 , laughlin et al .
2005 , winn & holman 2005 ) . to date , nine additional planets have been found to transit their parent stars , and the varied uses to which these transits have been put are described in the review article by charbonneau et al .
our goal in this paper is to ascertain whether the two outer planets orbiting gl 876 ( il aqr , gj 876 , hip 113020 ) can similarly be observed in transit .
a positive detection would give a further major improvement in our understanding both of this specific system and of planetary properties in general .
rivera et al .
( 2005 ) carried out a photometric search for transits by planet d ( @xmath8 ) , and showed that such transits do not occur .
to date , photometric searches for transits by planet c ( @xmath9 ) and b ( @xmath10 ) have not been reported .
although the _ a priori _ geometric transit probabilities for gl 876 c and b are only @xmath11 and @xmath12 , respectively , the expected photometric transit depths are greater than 10% .
librational motions arising from the resonance between the two outer planets lead to predicted transit epochs for the outer planets that are not spaced evenly in time .
( the resonant structure of the system is discussed further in laughlin & chambers 2001 ; rivera & lissauer 2001 ; kley 2003 ; and beaug , michtchenko , & ferraz - mello 2005 ) .
benedict et al . ( 2002 ) used hst astrometry to derive an orbital inclination for gl 876b of @xmath13 .
their result thus suggests that planet b has a mass equal or close to the rv - determined value of @xmath14 , and that the system is within 1-sigma of being viewed edge - on .
the implication of the benedict et al .
( 2002 ) measurement is that occurence of transits is much more likely than the a - priori geometric probability would suggest .
the benedict et al .
( 2002 ) result is in conflict , however , with the dynamical models of the system presented by rivera et al .
( 2005 ) , who find a best fit to the radial velocity data when the system ( assumed co - planar ) is inclined by @xmath15 with respect to the plane of the sky .
therefore , if transits are detected , then the benedict et al .
( 2002 ) results will be confirmed at the expense of the conclusions of rivera et al .
( 2005 ) .
gl 876b and gl 876c are members of an as - yet unobserved class of giant planets with intermediate surface temperatures .
the expected effective temperatures of @xmath16k and @xmath17k of the planets are much lower than the value @xmath18k measured with spitzer for hd 209458b ( charbonneau et al .
2005 ) , but still considerably higher than the @xmath19k measured for jupiter .
determination of their physical properties would thus provide a useful link between jupiter and saturn on the one hand , and objects such as hd 209458b on the other .
accurate measurement of their radii would also give clues to their interior structure ( see charbonneau et al .
2006 ) .
finally , the gl 876 system with its strongly gravitationally interacting planets and large planet - to - star radius ratios , has the potential to display an extraordinarily informative set of transit light curves .
our model light curves herein show what one can expect when interacting transiting systems are discovered , for example , with ongoing radial velocity surveys or with space - based missions such as kepler ( borucki et al .
2003 ) or corot ( baglin 2003 ) , and allow for a more complete understanding of the multiplanet dynamics .
the plan of this paper is as follows . in 2 , we describe our dynamical model of the system , and illustrate some of the photometric properties that the in - transit light - curves would be expected to display . in 3 , we describe an evaluation of the doppler radial velocity measurements that have been taken during predicted transit intervals . in 4 , we evaluate a range of archival and newly obtained photometric data sets for evidence of transits . in 5 , we discuss our results , as well as the current status and outlook for the ongoing _ transitsearch _ collaboration , and in 6 , conclude .
gl 876 ( m4v ) is the fortieth - nearest star to earth , with a hipparcos - determined distance of 4.69 pc ( perryman et al .
1997 ) ; tycho - ii and ucac positions validate this determination . as discussed in laughlin et al .
( 2005 ) , current estimates of its mass and radius stand at @xmath20 , and @xmath21 .
we adopt these values as fixed in this paper . to date
, 155 doppler velocities measured at the keck telescope and 16 velocities measured at the lick telescope have been published for the gl 876 system ( marcy et al .
2001 ; rivera et al .
the two planets induce a total velocity half - amplitude of the star of nearly @xmath22 . furthermore , as discussed in rivera et al .
( 2005 ) the system displays a very low level of stellar radial velocity `` jitter '' , which further aids in obtaining a detailed characterization of the orbits .
the system has now been observed for more than eighty orbital periods of the middle planet , c. this extensive time baseline allows for a much more detailed study of planet - planet interactions than can be obtained with the other exoplanetary systems known to be in 2:1 resonance , a list which now includes hd 82943 ( mayor et al .
2004 , lee et al .
2006 ) , hd 128311 ( vogt et al .
2005 ) , and hd 73526 ( tinney et al . 2006 ) .
these three systems all have inner planet periods of order ten times longer than gl 876 c. our baseline co - planar @xmath23 orbital model was obtained by rivera et al .
( 2005 ) from a dynamical fit to the keck radial velocity data . for reference ,
the parameters of this model are reproduced in table 1 .
additional fits with @xmath24 were also obtained to the keck data using the algorithm described by rivera et al .
( 2005 ) . as with the @xmath23 baseline model ,
the system is assumed to contain three planets , all orbiting the star in the same plane .
we varied the inclination of the orbital plane to the plane of the sky from @xmath23 to @xmath25 in decrements of @xmath26 , and fitted for each of the 13 + 1 remaining parameters .
n - body integrations of the fit then generate a sequence of transit predictions for the outer two planets .
uncertainties in the predicted transit midpoints were generated with the bootstrap monte - carlo technique described by rivera et al .
( 2005 ) . for both outer planets ,
the uncertainty in the transit midpoint times was found to vary from @xmath271 hr to a few hours .
the variation in uncertainty from transit to transit is plotted in figure 1 .
the predicted transit midpoints are best constrained during the epoch spanned by the radial velocity observations .
errors increase as the transit predictions are generated for earlier or later times .
light curves were produced with a photometric model of the transit which assumes that the planets are opaque disks with radii @xmath28 for c and @xmath29 for b. these equatorial radii are computed using the models of bodenheimer , laughlin , & lin ( 2003 ) .
the specific intensity of the stellar disk is modeled with a linear limb darkening law , given as @xmath30 where @xmath31 is the specific intensity of the stellar disk , @xmath32 is the corresponding limb darkening coefficient appropriate to v - band observations of an m4v primary ( per claret 2000 ) ; and @xmath33 . figures 2 and 3 show a selection of model light curves for planets b and c. for planet b , which has an eccentric ( @xmath34 ) , rapidly precessing orbit ( @xmath35 ) , the shape of the predicted curves depends strongly on the osculating periastron angle . for @xmath24 ,
the transits are both shorter and deeper when the transits occur near periastron .
the orbital eccentricity of planet b is more nearly circular , with @xmath36 . in this case
, there is little variation in the predicted light curves as the orbits precess .
the known transiting planets have transit - to - transit intervals @xmath37 that are evenly spaced to within the current resolution of the observations .
hd 209458b , for example , has had individual transit midpoints measured to an accuracy of several seconds ( brown et al .
2001 , wittenmyer et al .
2005 ) , and an average period measured to an accuracy of 83 ms ( wittenmyer et al . 2005 , winn & holman 2005 ) .
if , however , one is able to measure differences in the time intervals between successive transits , one can derive additional dynamical information .
as discussed by miralda - escud ( 2002 ) , holman & murray ( 2005 ) , and agol et al .
( 2005 ) , measurements of small variations from exact periodicity can be used to infer the presence of additional bodies in the system , potentially with the masses in the range appropriate to terrestrial planets . in the event that either gl 876b or gl 876c were to transit
, one would expect the departures from strict periodicity to be large .
in fact , successive transits of planet c would occur at intervals differing by as much as 4 hours ! by fitting a dynamical model to such transit intervals , one could obtain information regarding all of the orbital elements of the planets ( including inclinations and nodes ) .
during a planetary transit , the occulting disk of the planet passes over portions of the stellar disk containing a varying component of line - of - sight velocity arising from stellar rotation ; this phenomena produces an asymmetry in the star s spectral lines known as the rossiter - mclaughlin effect ( rossiter 1924 , mclaughlin 1924 ) .
the effect has been described in detail in the context of planetary transits by ohta , taruya & suto ( 2005 ) , and has been analyzed for hd 209458 by queloz et al .
( 2000 ) , bundy & marcy ( 2000 ) , and winn et al .
( 2005 ) , and for hd 149026 by wolf et al .
in fact , the transiting planet hd 189733 was initially discovered to transit _ not _ via photometry , but rather via such in - transit radial velocity variations , found by bouchy et al .
( 2005 ) . for our study of gl 876 , we compared the observation epochs of all of the lick ( see marcy et al . 2001 ) and keck ( see rivera et al .
2005 ) radial velocity observations to see if any of these observations occurred during a predicted transit ( assuming @xmath23 ) . for gl 876
, we used a rotational period that was determined to be 96.7 days using rotational modulation ( as reported in rivera et al .
2005 ) ; we used this value to calculate the amplitude of the rossitter - mclaughlin effect .
four such observations were found , three taken during a predicted transit by planet `` c '' , and one taken during a predicted transit by planet `` b '' .
figures 4 and 5 display the relevant parts of the radial velocity model ( both with and without the rossiter - mclaughlin effect ) , with the observations shown .
the radial velocity model curves in both figures were generated by adding template curves which use the stellar and planetary radii and limb darkening law given above ( and which apply the analytic integrals given in ohta et al .
2005 ) , to our fiducial @xmath23 three - planet radial velocity model .
for the transit epoch in the case of planet c ( figure 4 ) , the data is inconsistent with a transit . in figure 5 , for the outer planet b , the situation is unresolved .
we have also compared our model light curves to a variety of photometric datasets for gl 876 taken since 1989 .
the results of these comparisons are as follows .
* hipparcos * : the hipparcos mission ( see perryman et al .
1997 ) produced 67 accepted photometric measurements for gl 876 ( hip 113020 ) .
these measurements have a median magnitude @xmath38 , with @xmath39 .
there is evidence for a weak periodicity when the data is folded at the 96.7 d rotational period of the star ( also noted in rivera et al .
2005 ) . in figure 6
, we plot our model @xmath40 light curves for gl 876 b and c superimposed on the hipparcos epoch photometry .
none of the hipparcos measurements were made near or during a predicted transit window .
* fairborn observatory * : a number of long - term photometric monitoring programs are being carried out with the automated telescopes at fairborn observatory ( e.g. , henry 1999 ; eaton , henry , & fekel 2003 ) and , as described in detail in rivera et al .
( 2005 ) , gl 876 has been on one of the observing programs .
figures 7 and 8 show the fairborn observatory data and our corresponding model light curve for planets b and c. data taken during predicted transit epochs for planet c show no indication of a transit .
the fairborn observatory data indicate that long - term stellar variability from gl 876 is of order 0.05 magnitude , which is somewhat smaller than the variation observed with hipparcos .
* transitsearch and aavso * : seagroves et al .
( 2003 ) describe _ transitsearch _ , a cooperative distributed observing project involving sub - meter class telescopes worldwide .
the transitsearch strategy is to observe known planet - bearing stars at the dates and times when transits are expected to occur .
it is therefore a targeted search , which differentiates it from ongoing wide - field surveys .
we identify observing windows for candidate stars through the use of the bootstrap monte carlo technique described by laughlin et al .
the observational campaigns are prioritized by the a - priori likelihood that a particular candidate planet will display transits .
this likelihood is given by , @xmath41\ , \label{eq2}\ ] ] where _ a _ is the semi - major axis of the orbit , @xmath42 is the radius of the star , @xmath43 is the radius of the planet , @xmath44 is the longitude of periastron , and @xmath45 is the orbital eccentricity .
the first transitsearch dataset for gl 876 was obtained by one of us ( shankland ) from 2003 oct .
27.9583 to 2003 oct .
28.0368 , with a v - filtered , 16-bit speckle ccd at the primary focus of a 40.4 cm f/5.1 reflector ( at air mass from @xmath46 to 1.7 ) .
the photometric time series shows a monotonic flux increase of 0.16 + /-
0.04 mag from 2003 oct .
27.9583 to 2003 , oct 27.9972 .
as shown on the left panel of figure 9 , the depth , duration and timing of this event were consistent with an egress from transit of gl 876 c , in agreement with our predictions .
the possibility that a transit had been observed spurred us to orchestrate a distributed observational campaign during the 2004 season . during the 2004 june 24 opportunity
, observations were attempted by ten observers in australia .
poor weather thwarted all but one observer , and no transit signal was seen in the data , which covered about half of the @xmath47 transit window .
eight australian and japanese observers participated in the following 2004 july 24 planet c campaign .
they were all clouded out .
weather improved for the 2004 august 23 planet c opportunity , permitting observations by five of the nine australian , south african , and german observers .
no transit signal was seen with photometric depth greater than 1% . in 2004 october , a number of aavso and transitsearch observers obtained photometry during a closely spaced pair of transit windows for planets b and c. beginning october 20 11:22 ut and ending october 23 15:27 ut , a total of 2,795 ccd observations were obtained .
in addition , 2,981 photometric observations were obtained in the days immediately before and after the window to set baseline activity and to look for red dwarf flaring . such a large , distributed network of observers allowed coverage of the window with very few observing gaps over the 3-day period .
gaps in the coverage were smaller than the estimated transit period or forecast window , allowing us to conclude that transits for planets b and c did not occur during the october 2004 campaign .
photometric data from the campaign are plotted in figures 9 and 10 .
individual observations and uncertainty estimates from the october 2004 campaign are available from the aavso web site , while the other datasets can be found at the _ transitsearch _ website .
we are continuing to obtain photometry of gl 876 , with the goal of gaining a better understanding of the photometric variability of mature ( age exceeding 1gyr ) red dwarfs .
the _ transitsearch _ photometric network has been in operation for three years , and has participating observers capable of providing fully global longitude and latitude coverage . the network maintains a continuously updated catalog of the known census of extrasolar planets .
the catalog is presently the only available source for ( 1 ) a - priori transit probabilities and predicted transit depths , for the known extrasolar planets , ( 2 ) predicted transit ephemeris and transit window uncertainties based on orbital fits to published radial velocities , and ( 3 ) the results of known photometric searches for transits of planet - bearing stars . to date , the majority of negative transit searches have not been reported in the literature . with this paper
, we hope to spur a reversal of that trend . due to significant uncertainties in the orbital fits ,
it is generally very difficult to definitively rule out transits by planets with @xmath48 .
the gl 876 planets form a notable exception to this rule because of their high signal - to - noise , @xmath49 , and the extensive radial velocity data set allows for a very accurate orbital fit .
campaigns by the transitsearch collaboration have thus far led to transits of a number of other shorter - period planet - bearing stars being ruled out to a high degree of confidence .
the planets for which the possibility of transits have been significantly discounted by our campaigns include hd 217107 `` b '' ( @xmath50 , @xmath51 , see also vogt et al .
2005 ) , hd 168746 `` b '' ( @xmath52 , @xmath53 , see also pepe et al .
2002 ) , and hd 68988 `` b '' ( @xmath54 , @xmath55 , see also vogt et al .
2002 ) .
furthermore , planets for which usable photometry of the parent stars has been obtained by our network during the @xmath56 transit windows , but for which transits can not yet be fully ruled out , include hd 188753ab ( @xmath57 , @xmath58 ) , hd 76700 `` b '' ( @xmath59 , @xmath60 ) , hd 13445 b ( @xmath61 , @xmath62 ) , hd 74156 `` b '' ( @xmath63 , @xmath64 ) , hd 37605 `` b '' ( @xmath65 , @xmath66 ) , and hd 80606 `` b '' ( @xmath67 , @xmath68 ) .
planets for which campaigns are scheduled to begin soon include gl 581 `` b '' ( @xmath69 , @xmath70 ) , and hd 99492 `` b '' ( @xmath71 , @xmath62 ) .
gl 581 is of particular interest because the primary star is an m dwarf , and the planet ( if it transits ) will be a neptune - mass object @xmath72 ( bonfils et al .
the 1.6% transit depth would allow an unprecedented physical characterization of a low - mass extrasolar planet .
to date , our gl 876 campaign has employed the most globally far - reaching distribution , and along with recent leaps in fidelity in similar campaigns , clearly underscores the utility of the network .
what is more , gl 876 b and c constitute a dramatic exception to the emerging aphorism that jupiter - mass planets are rarely associated with m dwarfs .
indeed , a paucity of giant planets orbiting red dwarfs seems to be a natural consequence of the core accretion theory of planet formation ( laughlin , bodenheimer & adams 2004 ) .
the core - accretion theory predicts , however , that neptune - mass and smaller mass planets will be very common around red dwarfs .
this stands in contrast to the predictions of boss ( 2006 ) , who suggests that the gravitational instability mechanism is the dominant mode of giant planet formation .
if this is true , then jupiter - mass planets should be just as common around red dwarfs as they are around solar - type stars .
efforts to detect transits of small planets orbiting small stars are therefore likely to gain increasing importance and focus over the next several years .
there are , in fact , several observational indications that low - mass planets may be common around red dwarfs .
the radial velocity surveys have recently reported the detection of neptune - mass companions in short - period orbits around the red dwarfs gl 581 and gl 436 .
the ogle team of observers using the microlensing method have detected the signature of what appears to be a planet with 5.5 earth masses orbiting a distant red dwarf ( bennet et al .
2006 ) . with these points in mind
, we can put the results of our gl 876 campaign in a broader context . a ground - based ( and `` fiscally viable '' ) photometric network such as transitsearch , especially in collaboration with a dedicated bank of telescopes , can capably monitor individual m dwarfs to search for transits of terrestrial - sized planets . owing to the intrinsic long - term variability of m dwarfs , candidate planets
would have to be identified on the basis of a full transit signature in a single time - series .
a 0.1 @xmath73 m dwarf has @xmath74 , @xmath75 , and @xmath76 ; a habitable planet that receives an earth - equivalent flux from the star thus needs to orbit at a distance of just 0.022 au , which corresponds to an orbital period of 3.85 days .
an earth - like planet with such a period would be rotationally synchronized to the red dwarf , and in the absence of any significant perturbing bodies , its orbit would be almost perfectly circular .
interestingly however , simulations show that habitability is unlikely to be adversely affected by a spin - orbit period synchronization .
joshi , haberle , & reynolds ( 1997 ) used a global climate model to investigate how the earth s climate would respond if the earth were tidally locked to the sun , and found that earth remains habitable in this configuration at 1 au , and perhaps throughout the habitable zone .
the @xmath77 , @xmath78 companion to gl 876 demonstrates that it is not unreasonable to expect terrestrial - mass bodies on short - period orbits of @xmath79 stars .
montgomery & laughlin ( 2006 ) have carried out accretion simulations with the wetherill - chambers method ( wetherill 1996 , chambers 2001 ) that model the accretion of terrestrial - mass planets in short - period orbits about m dwarfs .
these simulations provide further evidence that terrestrial - mass planets will commonly form in the desired short - period orbits .
if habitable planets do commonly form in orbit around low mass m dwarfs , the chances of detecting and characterizing them are surprisingly good .
the transit of an earth - sized planet will block about 1% of the stellar flux of a @xmath80 star . for a planet on a habitable 3.85 day orbit
, the transit will be relatively brief , @xmath81 minutes .
the a - priori geometric probability of observing such a transit is 2% .
a 1% photometric dip is readily detectable .
amateur astronomers who participate in the transitsearch collaboration routinely achieve detection thresholds of considerably better than 1% , as evidenced by confirming detections of hd 149026b which has a transit depth of just 0.3% .
indeed , capable amateur observers have demonstrated the photometric capability to detect the passage of a mars - sized body in front of an 11th magnitude 0.1 solar mass red dwarf .
inevitably one must ask how many suitable red dwarfs are available on the sky as a whole . even though the lowest - mass m stars are the most common type of star , they are also exceedingly dim . a dedicated 3m - class telescope would be required to properly search the magnitude range v=16 - 18 , where observations of the lowest - mass red dwarfs begin to be plentiful ( and indeed , many 0.1 @xmath73 stars within 10 parsecs still likely remain to be discovered ) . in the meantime , however , we recommend that observers obtain high - cadence photometry of nearby single stars listed in the recons catalog of the 100 nearest stellar systems ; these include proxima centauri , barnard s star , wolf 359 , ross 154 , ross 128 , dx cancri , gj 1061 , gj 54.1 , and gj 83.1 .
we estimate that each one has a @xmath82% chance of harboring a detectible transiting , potentially habitable planet - readily detectible by distributed photometric observation .
our study of gl 876 has produced several interesting results . we have demonstrated that transits of planets b and c are not currently occurring , and we have thus modestly constrained the parameter space available to observationally allowed orbital configurations of the gl 876 system .
the absence of transits is consistent with the dynamical study of rivera et al .
( 2005 ) who find that a co - planar system inclination @xmath83 yields the best fit to the radial velocity data .
our dynamical analysis indicates that when planetary transits in a strongly interacting system similar to gl 876 are eventually observed , then a tremendous amount of information will be obtained .
we look forward to the discovery of the first such system . we have also shown that a globally distributed network of small - telescope observers can effectively provide definitive photometric follow - up of known planet - bearing low - mass stars .
our experience with gl 876 indicates that intensive photometric monitoring campaigns of individual , nearby m dwarfs constitute a tractable strategy for planet detection .
the sun s nearest low - mass stellar neighbors should therefore be scrutinized very carefully for transits by short - period terrestrial planets .
a low - budget survey of this type provides a small , but nevertheless non - negligible chance to discover a truly habitable world from the ground , and within a year or two .
such a detection would be a remarkable and exciting discovery .
we thank k. johnston , r. gaume , k. malatesta , j. lazio , d. boboltz , c. shaw , b. casey , e. albert , a. henden , j. larsen , d. katz , c. phillips , w. bollwerk , z. dugan , j. jaglowicz , and t. castellano for useful discussions or other support .
we are grateful to the aavso and its many members who provided campaign support in addition to those authors listed above .
we also gratefully acknowledge the insightful suggestions and constructive feedback offered by our referee .
this material is based upon work supported by the national aeronautics and space administration under grant nng04gn30 g issued through the origins of solar systems program to gl .
gwh acknowledges support from nasa grant ncc5 - 511 and nsf grant hrd-9550561 .
this research has made use of the simbad , usno and _ transitsearch _ databases , operated at cds , strasbourg , france , usno , washington , d.c .
, usa , and ucsc , santa cruz , california , usa , respectively . , taken near jd 2452940.4381 ; relative photometry , which began in astronomical twilight ( _ filled - in circles _ ) _ right : _ dynamical model light curves for gl 876 due to planet `` c '' for @xmath84 through @xmath85 in @xmath26 decrements , taken near a predicted transit centered at jd 2453301.7879 .
the data were compiled , baselined and shifted to employ a mean magnitude of v=10.19 , based on relative photometry obtained by aavso and _ transitsearch _ ( _ filled - in circles _ ) .
the light lines correspond to @xmath23 model light curves arising from bootstrap fits in which the transit is 3@xmath86 early , 1@xmath86 early , and 1@xmath86 late and 3@xmath86 late.,title="fig:",scaledwidth=99.0% ] + llll @xmath87 & 5.89 @xmath88 0.54 @xmath89 & 0.619 @xmath88 0.005 @xmath90 & 1.935 @xmath88 0.007 @xmath90 + @xmath91 ( d ) & 1.93776 @xmath88 0.00007 & 30.340 @xmath88 0.013 & 60.940 @xmath88 0.013 + @xmath92 ( ms@xmath93 ) & 6.46 @xmath88 0.59 & 88.36 @xmath88 0.72 & 212.60 @xmath88 0.76 + @xmath94 ( au ) & 0.0208067 @xmath88 0.0000005 & 0.13030 @xmath88 0.00004 & 0.20783 @xmath88 0.00003 + @xmath45 & 0 ( fixed ) & 0.2243 @xmath88 0.0013 & 0.0249 @xmath88 0.0026 + @xmath95 ( @xmath96 ) & 0 ( fixed ) & 198.3 @xmath88 0.9 & 175.7 @xmath88 6.0 + @xmath97 ( @xmath96 ) & 309.5 @xmath88 5.1 & 308.5 @xmath88 1.4 & 175.5 @xmath88 6.0 + transit epoch & jd 2452490.756 @xmath88 0.027 & jd 2452517.633 @xmath88 0.051 & + [ 3plparam ] | we report the results of a globally coordinated photometric campaign to search for transits by the @xmath0 and @xmath1 outer planets of the 3-planet system orbiting the nearby m - dwarf gl 876 .
these two planets experience strong mutual perturbations , which necessitate use of a dynamical ( four - body ) model to compute transit ephemerides for the system .
our photometric data have been collected from published archival sources , as well as from our photometric campaigns that were targeted to specific transit predictions .
our analysis indicates that transits by planet `` c '' ( @xmath0 ) do not currently occur , in concordance with the best - fit @xmath2 co - planar configuration obtained by dynamical fits to the most recent radial velocity data for the system .
transits by planet `` b '' ( @xmath3 ) are not entirely ruled out by our observations , but our data indicate that it is very unlikely that they occur .
our experience with the gl 876 system suggests that a distributed ground - based network of small telescopes can be used to search for transits of very low mass m - stars by terrestrial - sized planets . |
it is an exciting possibility that dark matter has a supersymmetric origin .
various species can have the properties of dark matter depending on the supersymmetry breaking mechanism and further particulars of the model , with the neutralino and gravitino being the most prominent candidates . in classes of models like anomaly @xcite and mirage @xcite mediation , the gravitino is heavier than the other sparticles and
thus can not constitute dark matter . on the other hand , in gauge @xcite and gaugino @xcite mediation , the gravitino is light and represents a good candidate for dark matter @xcite,@xcite . in gravity mediation
@xcite , the situation is more model - dependent and both relatively heavy and light gravitinos are possible . in this work ,
we study the circumstances under which the gravitino is the lsp in gravity mediation . if we further require r- or matter parity @xcite , which can descend from string theory @xcite , the gravitino is stable and can constitute dark matter .
phenomenology of the gravitino lsp has been an active research subject @xcite@xcite , while in this paper we focus on its supergravity side and identify relevant constraints on fundamental supergravity quantities .
let us review relevant features of the supergravity formalism , following @xcite ( and the original work @xcite ) .
the supergravity scalar potential is given ( in planck units ) by @xmath7 here the subscript @xmath8 ( @xmath9 ) denotes differentiation with respect to the @xmath10-th ( @xmath10-th complex conjugate ) scalar field .
@xmath11 is a function of the khler potential @xmath0 and the superpotential @xmath12 , @xmath13 , and the susy breaking @xmath14-terms are @xmath15 with @xmath16 being the inverse of @xmath17 .
the gravitino mass is given by @xmath18 another quantity we need is the khler metric @xmath19 for the observable fields @xmath20 .
it is found by expanding the khler potential around @xmath21 : @xmath22 then the soft susy breaking terms are given by @xcite @xmath23\;,\nonumber\end{aligned}\ ] ] where @xmath24 are the gauge kinetic functions , @xmath25 @xmath26 , and @xmath27 are the superpotential yukawa couplings .
vanishing of the vacuum energy requires @xmath28 at the minimum of the scalar potential .
thus , the magnitude of the @xmath14-terms depends on the khler metric of the susy breaking fields @xmath17 .
the @xmath14-terms and , consequently , the soft masses can be much larger than the gravitino mass provided the khler metric is sufficiently small .
let us consider the case of a single dominant susy breaking field @xmath29 .
omitting for simplicity complex phases , we have @xmath30 where @xmath31 .
let us further assume universal gauge kinetic functions and khler metrics for the matter fields .
then the soft terms simplify to @xmath32~ m_{3/2}^2 ~ , \nonumber\\ a_0 & = & \sqrt{3\over k''}~ \biggl [ k ' -3 { \tilde k ' \over \tilde k } \biggr]~m_{3/2}~ , \label{softmassessinglefield}\end{aligned}\ ] ] where a prime denotes differentiation with respect to @xmath29 and a double prime differentiation with respect to @xmath29 and @xmath33 . we have assumed that the yukawa couplings are independent of @xmath29 , which allows us to avoid strong constraints from electric dipole moments @xcite .
this case corresponds to the cmssm .
we see that the scalar and gaugino masses can be made arbitrarily large by decreasing @xmath34 .
this requires a non - negligible @xmath35 and @xmath36 .
if this quantity vanishes at the gut scale , large scalar masses are induced by the renormalization group evolution down to the electroweak ( ew ) scale .
since the gravitino mass does not run , we have @xmath37 and the gravitino is the lsp .
the @xmath38-terms usually do not play any significant role unless they are much larger than the other soft parameters .
let us quantify this effect . in the cmssm ,
the lightest superparticle is either a neutralino ( mostly bino ) or a stau .
unless @xmath39 is large , their masses can be approximated by @xcite,@xcite @xmath40 where we have neglected the ew contributions . for moderate @xmath41 ,
the lightest neutralino is lighter than the staus .
then the gravitino is the lsp for @xmath42 for small @xmath41 , the stau is lighter than the neutralino , so that a larger @xmath43 is required for a gravitino lsp .
however , eq . ( [ masses ] ) shows that the change in the lower bound will be rather small , so that eq .
( [ 2.5 ] ) is still a good approximation .
thus , the key parameter is the gaugino mass @xmath43 and as long as the scalar masses squared are non - negative , we obtain the gravitino lsp for @xmath44 where we have used @xmath45 .
this bound does not involve the superpotential nor the khler metric for the observable fields and is therefore largely model - independent . in figs .
[ fig1 ] and [ fig2 ] we illustrate these results with a numerical analysis in the cmssm .
we have used softsusy @xcite to determine the low - energy superpartner spectrum .
the figures display parameter space regions with the gravitino , neutralino and stau lsp .
we have chosen @xmath46 gev to fix the overall mass scale for definiteness , while the qualitative features of the plots are independent of this . at large @xmath39 , eq .
( [ masses ] ) receives corrections mainly from the @xmath47 yukawa coupling so that the @xmath48 becomes the lsp at small @xmath41 .
this is displayed in the right panel of fig .
[ fig1 ] . also , large @xmath38-terms further decrease the stau mass and can lead to tachyons , so that the @xmath48 lsp and the excluded regions are enlarged , cf.fig .
[ fig2 ] .
gev and @xmath49.,title="fig : " ] gev and @xmath49.,title="fig : " ] , but with a large @xmath38-term . ]
let us note that our parameter space is subject to further ( model - dependent ) phenomenological constraints .
these depend on the overall mass scale , assumptions on the cosmological history , whether r - symmetry is exact or approximate and whether the gravitino constitutes all or just part of the observed dark matter . for instance , as the lep bound on the higgs mass requires @xmath50 gev , eq .
( [ softmassessinglefield ] ) implies @xmath51 gev for @xmath52 and @xmath53 . for detailed studies
, we refer an interested reader to refs .
@xcite@xcite . from eq .
( [ bound ] ) , we see that @xmath34 is not required to be very small . in fact , it is of the order of magnitude of typical @xmath34 expected in string theory .
the usual moduli / dilaton khler potential is @xmath54 with @xmath55 so that @xmath56 just to have an idea of the numerics , take @xmath29 to be the dilaton of the heterotic string .
then @xmath57 , @xmath58 and @xmath59 at the minimum of the potential , as required by the observed values of the gauge couplings .
we get @xmath60 , which falls just short of the bound ( [ bound ] ) .
one should keep in mind , however , that the dilaton - dominated susy breaking is not possible with the above khler potential and in realistic cases one has to include either non - perturbative corrections to the khler potential @xcite,@xcite or additional fields . in the former case ,
one typically has @xmath61 at the minimum of the potential @xcite,@xcite , although the zero vacuum energy is not enforced .
when @xmath29 is the modulus associated with the radius of the compact dimensions , one can trust the supergravity approximation for @xmath62 in which case @xmath63 can be arbitrarily small .
in realistic cases , however , one has to include more than one field to have the correct gauge coupling , e.g. @xmath64 .
otherwise , the gauge coupling becomes too small .
one can also entertain the possibility that @xmath29 is a hidden matter - like field with the khler potential @xmath65 , which dominates susy breaking @xcite . in this case
, @xmath66 can be very small due to large moduli , @xmath67 .
however , the gauge kinetic function is then given predominantly by some other field , e.g. the dilaton , so that @xmath68 . consequently , @xmath69 can be sufficiently small for the gravitino to be the lsp , yet @xmath70 would require careful engineering
. the above formulae can be generalized to the case of multiple susy breaking fields in a straightforward manner .
let us illustrate with an example how the gravitino lsp can arise in string - motivated setups .
consider two modulus - type fields @xmath71 and @xmath72 with @xmath73 and @xmath74 . here
@xmath75 is the matter `` modular weight '' , which can be negative , positive or zero @xcite . for @xmath76 , locally stable vacua with zero ( or small ) vacuum energy are possible @xcite
for instance , this is the case when @xmath77 are the overall modulus and the dilaton .
finally , one can choose an appropriate superpotential such that the fields stabilize at @xmath78 in order to have the correct gauge couplings at the gut scale
. `` mixed '' gauge kinetic functions of this type appear in string models with fluxes @xcite,@xcite . requiring zero vacuum energy at the minimum of the scalar potential and neglecting complex phases , we can parametrize supersymmetry breaking by the goldstino angle @xmath79 : @xmath80 then the soft terms read @xmath81 ~m_{3/2 } ~ , \nonumber\\ m_0 ^ 2 & = & \biggl [ 1 + { 3n \over a } ~\cos^2\theta \biggr]~ m_{3/2}^2 ~ , \nonumber\\ a_0 & = & -\sqrt{3}~ \biggl [ \sqrt{a } \left(1 + { 3n \over a } \right ) ~\cos\theta + \sqrt{b } ~ \sin\theta \biggr]~m_{3/2}~.\end{aligned}\ ] ] for a sufficiently large @xmath71 and @xmath82 , the gauginos are heavy so that @xmath83 and the gravitino is the lsp .
note that the @xmath84 dependence cancels out in @xmath85 and @xmath86 .
taking as an example @xmath77 to be the overall modulus and the dilaton , @xmath87 and @xmath88 .
then for @xmath89 , the gravitino lsp imposes the bound @xmath90 .
note that the scalar masses are non - tachyonic at the gut scale for @xmath91 .
let us finally note that the superpotential does not play a role in this discussion as long as it stabilizes the fields at the desired values with vanishing vacuum energy .
this question can be analyzed locally , in terms of @xmath92 and @xmath93 , along the lines of ref .
@xcite . as a result
, the superpotential expansion coefficients @xmath94 , @xmath95 , have to satisfy certain ( model - dependent ) constraints .
we have studied the conditions for the gravitino to be the lsp in models of gravity mediated susy breaking .
this requirement constrains mainly the gaugino mass parameter at the gut scale while the other parameters play a minor role , as long as the scalar masses are non - tachyonic . for cmssm - type models at moderate @xmath39 ,
the resulting constraint on the khler potential and the gauge kinetic function is approximately @xmath2 , where the derivatives are taken with respect to the dominant susy breaking field . for large @xmath39 and @xmath38-terms ,
the above constraint gets modified at small values of the universal scalar mass @xmath41 .
the results are illustrated in figs .
[ fig1 ] and [ fig2 ] .
the condition @xmath2 can easily be satisfied in string - motivated set - ups and thus the gravitino lsp is a reasonable alternative to the neutralino lsp in gravity mediation . as it is hard to obtain an extremely small value of @xmath34 without finetuning , the gravitino mass is still expected to be of the same order of magnitude as the other soft masses .
k. choi , a. falkowski , h. p. nilles and m. olechowski , nucl .
b * 718 * , 113 ( 2005 ) ; m. endo , m. yamaguchi and k. yoshioka , phys .
d * 72 * , 015004 ( 2005 ) ; k. choi , k. s. jeong and k .-
okumura , jhep * 0509 * , 039 ( 2005 ) ; a. falkowski , o. lebedev and y. mambrini , jhep * 0511 * , 034 ( 2005 ) .
m. dine , w. fischler and m. srednicki , nucl .
b * 189 * , 575 ( 1981 ) ; m. dine and w. fischler , phys .
b * 110 * , 227 ( 1982 ) ; m. dine and a. e. nelson , phys .
d * 48 * , 1277 ( 1993 ) ; m. dine , a. e. nelson and y. shirman , phys .
d * 51 * , 1362 ( 1995 ) .
d. e. kaplan , g. d. kribs and m. schmaltz , phys . rev .
d * 62 * , 035010 ( 2000 ) ; z. chacko , m. a. luty , a. e. nelson and e. ponton , jhep * 0001 * , 003 ( 2000 ) ; w. buchmller , k. hamaguchi and j. kersten , phys .
b * 632 * , 366 ( 2006 ) ; w. buchmller , l. covi , j. kersten and k. schmidt - hoberg , jcap * 0611 * , 007 ( 2006 ) ; j. l. evans , d. e. morrissey and j. d. wells , phys .
d * 75 * , 055017 ( 2007 ) . s. k. soni and h. a. weldon , phys .
b * 126 * , 215 ( 1983 ) ; v. s. kaplunovsky and j. louis , phys .
b * 306 * , 269 ( 1993 ) .
s. abel , s. khalil and o. lebedev , phys .
lett . * 89 * , 121601 ( 2002 ) .
d. g. cerdeo , k. y. choi , k. jedamzik , l. roszkowski and r. ruiz de austri , jcap * 0606 * , 005 ( 2006 ) .
s. p. martin and p. ramond , phys .
d * 48 * , 5365 ( 1993 ) .
a. riotto , e. roulet and i. vilja , phys .
b * 390 * , 73 ( 1997 ) .
j. r. ellis , j. giedt , o. lebedev , k. olive and m. srednicki , phys .
d * 78 * , 075006 ( 2008 ) .
b. c. allanach , comput .
commun . * 143 * , 305 ( 2002 ) p. binetruy , m. k. gaillard and y. y. wu , nucl .
b * 481 * , 109 ( 1996 ) . | we study general conditions for the gravitino to be the lightest supersymmetric particle ( lsp ) in models with gravity mediated supersymmetry breaking .
we find that the decisive quantities are the khler potential @xmath0 and the gauge kinetic function @xmath1 . in constrained mssm ( cmssm )
type models , the gravitino lsp occurs if the gaugino mass at the gut scale is greater than approximately 2.5 gravitino masses .
this translates into @xmath2 , where the derivatives are taken with respect to the dominant susy breaking field .
this requirement can easily be satisfied in string - motivated setups .
= 20pt * jrn kersten@xmath3 and oleg lebedev@xmath4 * + @xmath5__university of hamburg , ii .
institute for theoretical physics + luruper chaussee 149 , 22761 hamburg , germany _
_ + @xmath6__desy theory group + notkestrae 85 , 22603 hamburg , germany _ _ |
systems with strongly correlated electrons are a subject of considerable research efforts . among those are the ones with 5@xmath0 electrons , i. e. , involving actinide ions of particular interest .
while 4@xmath0 electrons , because of their closeness to the nuclei , are correlated as strongly as possible on an atom , the 5@xmath0 electrons have a larger spatial extent and tend more towards delocalization .
the degree of delocalization is even more pronounced when 3@xmath1 electrons are considered . by continuing this line of arguments
one finds that @xmath2 electrons are the least correlated ones on an atom and have the largest tendency to delocalize . returning to 5@xmath0 electrons
, the question arises under which circumstances localized features dominate delocalized ones or vice versa .
in particular we want to follow up the suggestion which has been made previously @xcite that 5@xmath0 electrons have a dual character in the sense that some of the 5@xmath0 orbitals lead to delocalization when occupied by electrons while others result in localization .
it was demonstrated that a division of 5@xmath0 orbitals into delocalized and localized ones is able to explain the heavy quasiparticles observed in upt@xmath3 and upd@xmath4al@xmath3 including their mass anisotropies @xcite .
however , the question remained unanswered as regards the microscopic origin of this division of orbitals .
apparently , strong intra - atomic correlations seem to enhance anisotropies in the kinetic energy matrix elements which reflect anisotropies in the hybridization with the chemical environment . the purpose of the present work is to clarify the physical origin of the dual character of 5@xmath0 electrons . for that purpose clusters of two and three sites are considered . because of the large spin - orbit coupling each 5@xmath0 electron is characterized by its total angular momentum @xmath5 and azimuthal quantum number @xmath6 .
we do not include @xmath7 states in our analysis since their energy is higher by approximately 1ev @xcite .
we shall treat clusters where the average number of 5@xmath0 electrons per site is either 2.5 or 2.67 , values which are close to the 5@xmath0 occupancy of u ions in compounds like upt@xmath3 @xcite , upd@xmath4al@xmath3 or uge@xmath4 @xcite . depending on @xmath6
different hopping matrix elements are assumed for the @xmath0 electrons in the cluster .
the key result of the present investigation is to demonstrate that indeed anisotropies in the hopping are strongly enhanced by intra - atomic correlations .
the paper is organized as follows . after defining the model
hamiltonian a two - site and a three - site cluster are investigated .
various properties are calculated , like the ground - state energy and magnetization as well as the density of states and its magnetic field dependence .
this is done for different ratios , of hopping .
partial localization is the subject of the subsequent section [ secpartial ] . a discussion and
the conclusions are found in sect .
the appendices are devoted to perturbation theory , the atomic limit of vanishing inter - atomic electron transfer and to a more detailed analysis of the ground state s composition .
we investigate a microscopic model of 5@xmath0 electrons which includes local coulomb interactions and hopping terms between nearest neighbor sites .
the model hamiltonian is @xmath8 where the first sum is over neighboring sites @xmath9 .
furthermore @xmath10 ( @xmath11 ) , creates ( annihilates ) an electron at site @xmath12 in the @xmath13 state with @xmath14 .
we will consider two and three sites models .
the effective hopping between sites results from the hybridization of the 5@xmath0 states with the orbitals of the ligands and depends generally on the crystal structure . rather than trying to exhaust all possible different lattice symmetries
, we shall concentrate here on the special case that hopping conserves @xmath6 . while this is certainly an idealization
, it allows us to concentrate on our main interest , i. e. , a study of the influence of atomic correlations on the renormalization of hybridization matrix elements .
the parameters @xmath15 are chosen in accordance with density - functional calculations for bulk material which use @xmath16 basis states .
we also include local magnetic fields @xmath17 to account for the influence of ferromagnetic ( fm ) or antiferromagnetic ( afm ) internal fields .
the local coulomb interactions can be written in the form @xmath18 with coulomb matrix elements @xmath19 here @xmath20 denotes the total angular momentum of two electrons and @xmath21 .
the sum is restricted by the antisymmetry of the clebsch - gordan coefficients @xmath22 to even values @xmath23 .
we use in the actual calculations @xmath24 values which are determined from lda wavefunctions for upt@xmath25@xcite , i. e. , @xmath26 , @xmath27 , and @xmath28 .
we expect @xmath29 to always hold for coulomb interactions , independently of the chemical environment .
in contrast , the relative order of the hopping matrix elements will vary strongly from one compound to the next .
the average coulomb repulsion of about 20ev is irrelevant for the low - energy physics of the model .
it simply restricts the relevant configurations to states so that the sites are occupied either by 2 or 3 electrons .
the low - energy sector is solely determined by the differences of the @xmath30 values , which are of the order of 1ev and thus slightly larger than typical bare @xmath0 bandwidths .
the latter are obtained , e.g. , from lda calculations for metallic uranium compounds like upt@xmath3 .
note that restricting the model to @xmath31 and @xmath32 configurations is equivalent to let the different @xmath33 while their differences remain finite .
to mimic the situation in the u - based heavy - fermion compounds we consider the intermediate valence regime .
note that in the absence of a magnetic field all states of the two - site model with 5 electrons will be at least doubly degenerate because of kramers degeneracy .
the hamiltonian eq .
( [ eq : twositehamiltonian ] ) conserves @xmath34 where @xmath35 is the z - component of the total angular momentum of the system and the @xmath36 refer to angular momentum projections on individual sites .
we shall therefore characterize the eigenstates by their @xmath35 value . for isotropic hopping @xmath37
the system is rotationally invariant .
then @xmath38 provides an additional good quantum number .
as mentioned before it should suffice to limit ourselves to the subspace of @xmath31 and @xmath32 configurations , because of the large values of @xmath30 . to assess the validity of this simplification we shall compare the predicted ground states for the @xmath31-@xmath32 restricted model with those obtained by diagonalizing the hamiltonian eq .
( [ eq : twositehamiltonian ] ) in the full space .
the rational for the restriction to @xmath31-@xmath32 is not the reduction of the numerical diagonalization effort by a factor of about two , but the conceptual simplicity gained for the interpretation and presentation of the results for , e.g. , local spin configurations and orbital occupations . in the full model , interactions of the order of @xmath39 are present which arise from virtual @xmath40-@xmath41-@xmath42 transitions ( @xmath43 and @xmath44 are representative values of @xmath45 and @xmath30 , respectively . )
the restricted model ceases to be a good approximation when these interactions become comparable to the bare parameters of the model . for most of the calculations below
we shall assume @xmath46 and @xmath47 as independent parameters , a choice motivated by lda calculations for upt@xmath3 @xcite where @xmath48 , @xmath49 .
however , partial localization , which we will discuss in section [ secpartial ] , is also found for other parameter choices , e. g. , by choosing @xmath50 and @xmath51 as independent parameters .
the origin is qualitatively identical for both choices , as we will demonstrate later .
the results presented below are obtained from numerical studies of the hamiltonian ( [ eq : twositehamiltonian ] ) .
figure [ fig : threeenergies]a shows the variation of the ground - state energy @xmath52 as a function of the hopping parameters @xmath53 and @xmath54 .
this energy is smooth except for a kink along the isotropic line @xmath55 . and
@xmath54 . ( b ) expectation value of the coulomb interaction @xmath56 in the ground state @xmath57 vs. hopping parameters @xmath53 and @xmath54 . , title="fig : " ] and @xmath54 .
( b ) expectation value of the coulomb interaction @xmath56 in the ground state @xmath57 vs. hopping parameters @xmath53 and @xmath54 .
, title="fig : " ] level crossings are apparent from discontinuous jumps of the expectation value @xmath58 of the coulomb interaction with respect to the ground state , see fig .
[ fig : threeenergies]b .
the ground - state energy at vanishing hopping is 69.53 ev which is the sum of the coulomb energy of the @xmath59 @xmath31 configuration and the @xmath60 @xmath32 configuration , @xmath61 in general , the ground - state energy decreases with increasing hopping due to a lowering of the kinetic energy .
this is however partially compensated for by an increase in the interaction energy .
the jumps indicate quantum phase transitions , which result from the competition between coulomb interactions and anisotropic hopping or hybridization .
in order to characterize the ground state of the two - site cluster for different choices of the hopping matrix elements @xmath45 we study the total magnetization characterized by @xmath35 in the presence of a vanishingly small field .
results are shown in fig .
[ f2f3_phasentt2.eps ] .
the different ground states are separated by solid lines when we allow for @xmath31 and @xmath32 configurations only .
dashed lines show the ground - state boundaries when this restriction is lifted . the later
are obtained from diagonalization of @xmath62 in the full space of 5-electron states .
because of the small differences we shall limit ourselves to @xmath31-@xmath32 configurations only .
five different phases are visible in fig .
[ f2f3_phasentt2.eps ] , i.e. , two `` high - spin '' phases ( @xmath63 , @xmath64 ) and three `` low - spin '' phases ( @xmath65 ) . in the `` high - spin '' phases which prevail for @xmath66 the two sites are ferromagnetically correlated , @xmath67 , while in the low - spin phases the correlations are antiferromagnetic ( afm ) , @xmath68 . here and below we denote the two different sites by @xmath69 and @xmath70 .
( = @xmath49 ) vs. @xmath53 plane derived from the total magnetization @xmath35 of the ground state .
solid lines : for the @xmath31-@xmath32 model . dashed lines : with no restriction on the occupancy . ]
a special case is @xmath71 ( isotropic case ) . here
the ground state is always six - fold degenerate with @xmath72 .
this is plausible in the limit of large values of @xmath73 : all electrons will occupy orbitals with different @xmath6 in order to maximize the kinetic energy . with a total of five electrons one of the @xmath6 orbitals
will remain empty , implying a six - fold degenerate ground state . even when @xmath74 , implying the @xmath31-@xmath32 model , but with a difference @xmath75 of any pair of @xmath30 values much larger than @xmath73 , i.e.,@xmath76 , one can see that @xmath72 .
in that case the six - fold degenerate ground states can be explicitely written in the form @xmath77 where @xmath78 is the projector allowing only for atomic @xmath31 and @xmath32 configurations .
one implication is that the kinetic energy is only @xmath79 instead of @xmath80 as for free electrons , because at most three out of five electrons find free places to hop onto .
note that @xmath81 in eq .
( [ corrwavefunction ] ) .
while for finite values of @xmath82 the ground - state wavefunctions become vastly more complex , they remain eigenfunctions of @xmath72 .
simple results are found again in the limit @xmath83 , see appendix [ appperturbation ] . as one moves away from the isotropic line with @xmath72 the degeneracy of different @xmath35 states
is lifted .
depending on @xmath53 and @xmath84 one of them becomes the ground state while @xmath85 is no longer preserved ( see fig .
[ f2f3_phasentt2.eps ] ) .
the phase with @xmath63 appears when @xmath86(@xmath87 ) .
it corresponds to a very simple eigenstate in which all occupied orbitals have @xmath88 .
one finds @xmath89 with kinetic energy @xmath90 .
the @xmath6=5/2 and @xmath6=1/2 orbitals are occupied on both sites while @xmath91 is occupied with probability @xmath92 on each of the two sites . note that @xmath93 is separately an eigenstate of the coulomb as well as the kinetic energy terms . while the coulomb energy is at its absolute minimum
, there are no energy contributions resulting from @xmath94 and @xmath53 due to the limitation to @xmath31 and @xmath32 configurations .
this can be considered as an extreme case of partial localization : the @xmath6=5/2 and @xmath6=1/2 states are fully localized and the @xmath6=3/2 state is maximally delocalized .
when the hopping matrix elements increase , a discontinuous jump takes place into a state with @xmath95 . the corresponding increase in coulomb energy
is seen in fig .
[ fig : threeenergies]b .
the variation of the ground - state energy along the line @xmath96 can be qualitatively understood from the following argument : for sufficiently small hopping matrix elements @xmath53 the high spin @xmath63 state is lowest in energy because it minimizes the coulomb energy . with increasing values of @xmath53
it becomes more favorable to partially break up the local hund correlations in order to gain kinetic energy by partially populating both @xmath91 and the @xmath97 orbitals . in the limit of large
hopping , their angular momenta components cancel and @xmath35 is given by the @xmath6 components of the remaining electrons , i.e. , by @xmath98 resulting in @xmath64 or 1/2 . for intermediate hopping the total angular momentum of states were both @xmath99 and @xmath100 are occupied is given by the remaining @xmath101 orbitals .
for more details on the occupation of the individual @xmath6 orbitals in the ground state see appendix [ appanalysis ] . ) for sites with local @xmath32 ( upper panel ) and @xmath31 configurations ( lower panel ) , respectively . note the scale and the smallness of the deviation from the values @xmath102 and @xmath103 corresponding to fully polarized @xmath32 and @xmath31 configurations . ] while the total angular momentum @xmath85 depends strongly on the anisotropy , the local angular momenta change very little .
[ figq2q3 ] shows for different anisotropies the local spins defined in terms of projectors onto configurations with given local @xmath0-occupation as @xmath104 the projector @xmath105 selects atomic configurations with @xmath106 electrons .
the small range of @xmath107 variation in fig .
[ figq2q3 ] is remarkable .
hopping elements up to 0.5ev reduce the local moments by less than 5% relative to their maximal values 24.75 and 20 for @xmath108 and @xmath109 configurations , respectively .
one observes a reduction of the local spin expectation values at large hopping for both the @xmath108 and the @xmath109 site .
a decrease of @xmath110 implies for @xmath111 , an increase of the coulomb energy .
thus , figure [ figq2q3 ] demonstrates the competition between coulomb interaction and hopping . in the absence of any hopping ,
i.e. , for @xmath53=@xmath94=@xmath49=@xmath112 , the ground state is 180-fold degenerate ( see appendix [ appperturbation ] ) . at finite isotropic hopping this degeneracy
is only partially lifted .
then @xmath113 and @xmath114 are no longer separately conserved , but the total @xmath115 still is .
inset : anisotropic hopping @xmath116 lifts the high degeneracies .
+ ] thus , along the isotropic line , spherical symmetry still leads to rather highly degenerate states . as seen in fig .
[ fig : dossplitting ] , these states are grouped by hund s rule coupling , i.e. , by the energy difference of the values @xmath30 .
the different clusters of energy levels are separated by well - defined gaps of about 1ev . when spherical symmetry is broken
the @xmath35 degeneracy is lifted .
this is seen from the inset of fig .
[ fig : dossplitting ] .
it shows the evolution of the lowest 66 eigenstates as a function of the symmetry breaking parameter @xmath116 . the value of @xmath117 is thereby kept fixed .
five different groups of levels are noticed which for @xmath118 correspond to the values ( ordered by increasing energy ) @xmath95 , 3/2 , 7/2 , 13/2 and a 34 fold degenerate level consisting of @xmath63 and 17/2 states . when @xmath119 the @xmath35 degenerate states split into doublets . for @xmath120 one state from the originally 34-fold degenerate states
becomes lowest in energy .
this marks the transition to the `` high - spin '' state with @xmath63 . in the following we want to discuss the effect of an internal or molecular field on the phase diagram .
[ fig : jzrestrictedmodelmagneticfield ] shows the changes in the phase diagram caused by a ferromagnetic field of three different magnitudes .
one notices the appearance of new `` high - spin '' as well as `` intermediate spin '' phases @xcite .
now the phases with low @xmath121 values require larger hopping matrix elements in order to be realized .
the regions of `` high - spin '' states grow with increasing field and for small values of @xmath53 a new phase with @xmath122 appears .
this state is similar to the @xmath63 state and has the form @xmath123 0.005 , 0.01 , and 0.05ev , respectively . ]
in this section , we extend our analysis to a 3-site cluster . in order to approach the valency @xmath124 present in upt@xmath3
, we choose a cluster with eight electrons ( average @xmath0 valency 2.67 ) .
this choice also avoids the kramers degeneracy which was present in the 2-site cluster and which would again be present in the 3-site model with 7 electrons , i. e. , the other possible electron number with mixed valency .
we put the sites on a ring order to work with a model in which all sites are equivalent .
as a consequence , afm correlations are subject to frustration .
furthermore the sign of the matrix elements @xmath45 s is now relevant : the eigenstates of the kinetic energy term in the hamiltonian consist of a doublet with energy @xmath45 and a singlet state with energy @xmath125 .
thus , for @xmath126 occupancy of @xmath6 orbitals two of the three sites is favored whereas for @xmath127 a single occupancy is .
below , we consider both negative and positive hopping parameters , assuming the same sign for all @xmath45 .
all calculations are done in the subspace of only atomic @xmath31 and @xmath32 configurations . the left side of fig .
[ 3sitepm ] shows the value of the total angular momentum @xmath35 of the three - site cluster as a function of negative @xmath84 and @xmath53 . as for the two - site cluster
, a ferromagnetic high - spin phase appears in the region @xmath129 .
this phase has @xmath130 and can be understood as a straightforward generalization of the @xmath63 phase of the two - site model .
it is an eigenstate of both the coulomb and the kinetic terms . only orbitals with @xmath131 are occupied .
the @xmath132 and @xmath133 orbitals are filled with one electron per site and the remaining two electrons occupy @xmath91 orbitals .
in contrast to this simple ferromagnetic state , all other phases in fig .
[ 3sitepm]a have a more complicated structure . along the isotropic line of the phase diagram , @xmath134
, the ground state has @xmath135 for all finite values of @xmath136 .
the phases differ also in their magnetic correlations .
two of them have well pronounced correlations .
the @xmath137 phase has antiferromagnetic correlations ( @xmath138 ) , whereas the @xmath139 and 12 phases have ferromagnetic correlations .
correlations within the @xmath140 phase are by about a factor 5 smaller than those of the @xmath130 phase .
the other two phases , @xmath141,@xmath142 have generally small magnetic correlations , which can be either ferro- or antiferromagnetic , depending on @xmath94 and @xmath53 .
phase diagram for three sites for negative ( left panel ) and positive hopping ( right panel ) . ]
we now turn to the case with positive hopping matrix elements .
the ground state along the isotropic line has @xmath144 for all finite values of @xmath136 .
away from the isotropic line one finds @xmath145 for @xmath146 and @xmath147 for @xmath148 , see fig .
[ 3sitepm]b . as the kinetic energy penalizes states in which two or more electrons on different sites occupy the same @xmath6 orbital , the ferromagnetic state which was found for @xmath128 is no longer favorable .
most of the phase diagram is dominated by afm correlations . however , for large anisotropies @xmath149 these correlations weaken continously with increasing ratio @xmath150 and vanish near @xmath151 .
in the following we want to demonstrate that strong on - site correlations result in a considerable enhancement of anisotropies in the bare hopping matrix elements . this can lead to a a localization of electrons in orbitals with relatively weak hybridization .
the latter is effectively reduced to zero in those cases . in order to quantify the degree of localization or , alternatively , of the reduction of hopping of a given @xmath6 orbital by local correlations , we calculate the ratio of the @xmath6| projected kinetic energy @xmath152 and the bare matrix element @xmath45 , i. .e .
, @xmath153 the ground - state wavefunction @xmath154 contains the strong on - site correlations
. a small ratio of @xmath155 indicates partial suppression of hopping for electrons in the @xmath156 orbitals .
two kinds of correlations may contribute to that process .
the first one is based on the reduction of charge fluctuations to atomic @xmath31 and @xmath32 configurations .
this is a result for large values of @xmath30 and can be studied by setting all @xmath30 equal to a value much larger that the different @xmath45 .
the second one is due to differences in the @xmath30 values , i. e. , @xmath29 .
the differences in the @xmath30 values are the basis of hund s rules .
hopping counteracts hund s rule correlations and vice versa .
what we want to stress is the fact that those correlations can lead to a complete suppression of hopping channels except for the dominant one which is influenced only little .
results for the ratios @xmath155 are shown in fig .
[ fighop1 ] . for
a better understanding one should consider that figure in parallel with fig .
[ f2f3_phasentt2.eps ] . the parameter range in fig .
[ fighop1 ] corresponds to a straight line in fig .
[ f2f3_phasentt2.eps ] intersecting the abscissa at @xmath157 ev and the ordinate at @xmath158 ev .
this line crosses three different phases with @xmath63 , 5/2 and 3/2 , labeled i , ii and iii respectively .
one notices that in region i only the dominant hybridization of the @xmath159 and @xmath160 orbitals is completely suppressed . in regions
ii and iii the correlation effects on different orbitals are not so extremely different . in a forthcoming paper
we show how the effects of the correlations on the ratios @xmath155 are changed when all @xmath30 values are assumed to be equal . to understand the suppression of the @xmath152 in the limit of very small @xmath45 it is sufficient to treat the hopping perturbatively , this
is discussed in appendix [ appperturbation ] .
for the two - site cluster along the line connecting linearly the points written below the figure . phases with @xmath63 , 5/2 and 3/2 , are labeled with i , ii and iii respectively . ] in the case of a three - sites cluster the sign of the matrix elements @xmath45 matters because of the respective degeneracy of the ground state .
we begin with @xmath126 . shown in fig
. [ fighop3 ] are results for different parameter choices .
again , for a better understanding , one should consider in parallel the phase diagram in the left hand side of fig .
[ 3sitepm ] .
again , three different phases are involved and they are labeled as before by i , ii and iii in fig .
[ fighop3 ] .
altogether , the figure looks similar to fig .
[ fighop1 ] .
there is again a regime of total suppression of the @xmath94 and @xmath49 matrix elements ( region i ) and also in region iii the smaller of the @xmath45 , i. e. , @xmath53 is strongly reduced to an effective @xmath161 .
the case of equal coulomb integrals @xmath30 , i. e. , @xmath162 is discussed in an extended version of this paper .
next we turn to the case of positive @xmath127 .
corresponding results are shown in fig . [ fighop2 ] which should be considered in connection with the right hand side of fig .
[ 3sitepm ] .
there it is seen that two different phases with @xmath147 and 0 are involved with the chosen parameters @xmath45 . in distinction to figs .
[ fighop1 ] and [ fighop3 ] there is no regime of total suppression of hopping matrix elements . but a strong enhancement of anisotropies by the correlations is noticeable also here .
again , the overall features are very similar to those in figs . [ fighop1 ] and [ fighop3 ] except that the intermediate regime is missing here .
this is of little surprise , since we are dealing with local correlations so that the effects should be insensitive to the precise form of the chemical environment . of the three - sites model with @xmath126 along the line connecting linearly the @xmath45 values indicated in the figure .
] of the three - sites model with @xmath127 along the line connecting the @xmath45 values written below the figures . ]
it is instructive to take a closer look at the ground state . for @xmath163
, the ground state is twofold degenerate with @xmath164 ( see fig .
[ 3sitepm ] ) . when @xmath84 are small , @xmath129 , the @xmath31 configurations are strongly dominated by ones in which the orbitals @xmath165 and @xmath166 are occupied with a ferromagnetic alignment of the electrons .
the admixture of other states is less then 1.5 % of the norm for @xmath167 and @xmath168 .
furthermore , the wavefunction is strongly dominated , in fact to 87 % , by configurations in which the @xmath165 , @xmath169 orbitals form an ising - like frustrated afm background for mobile @xmath170 orbitals .
we describe this afm background by states @xmath171 ( @xmath172 and @xmath173 are defined analogously ) .
we further introduce operators @xmath174 for itinerant electrons in the @xmath175 orbitals @xmath176 with a number @xmath177 of order 1 taking into account that due to the background not all three sites are equivalent .
the ground - state wavefunction ( with @xmath147 ) is then of the approximate form @xmath178 where @xmath179 the projector @xmath180 projects onto local @xmath31 and @xmath32 configurations .
furthermore @xmath181 , @xmath182 are obtained from @xmath183 via cyclic permutation of sites .
note that the kinetic energy contributions @xmath184 and @xmath185 calculated with the wavefunction ( [ gutzwiller3 ] ) vanishes exactly . as the ground state can be written approximately as a sum of simple products of itinerant @xmath170 particles and a magnetic background , one expects to find a similar form for the lowest excited state .
this is indeed the case : the first excited state has again @xmath186 but with a fourfold rather than twofold degeneracy .
the excited states can be very well approximated by a difference of any two of the three states @xmath183 , @xmath181 , @xmath182 , e. g. @xmath187 .
for fixed @xmath147 or @xmath188 these states span a subspace of dimension two .
we may view the first excited state as an excitation of the frustrated afm background which is dominated by the @xmath166,@xmath189 orbitals , while the electrons / holes in the @xmath170 orbitals simply adjust to the slow dynamical changes of the magnetic background .
the magnetic correlations of the ground state are also interesting .
the overall magnetic correlations in @xmath190 are weakly ferromagnetic , i. e. , @xmath191 .
but because of @xmath192 one can construct states @xmath193 which have no magnetic correlations .
thus , we have a situation with a localization mechanism akin to a mott transition , but without the magnetic correlations which are usually present in mott systems .
in this investigation we have studied the effects of strong intra - atomic correlations on anisotropies in the hybridization of different 5@xmath0 orbitals with the environment . for that purpose we considered two - site and three - site clusters which allow for a complete numerical diagonalization of the corresponding hamiltonian .
we demonstrated that in particular hund s rule correlations enhanced strongly anisotropies in the hybridizations . for
a certain range of parameters this may result in a complete suppression of the effective hybridization except for the largest one , which remains almost unaffected .
this provides for a microscopic picture of partial localization of 5@xmath0 electrons , a phenomena , observed in a number of experiments on u compounds . in fact , previous work on the de haas - van alphen frequencies and anisotropic effective masses in upt@xmath3 and upt@xmath4al@xmath3 could explain the experimental data very well by assuming , that two of the 5@xmath0 electrons of the u ions remain localized while a third one is itinerant @xcite .
the latter is in an orbital with @xmath91 in both cases , and a @xmath194-@xmath194 coupling scheme is used .
previous work on the molecule uranocene u(c@xmath195h@xmath195)@xmath4 gave a hint on partial localization @xcite , but the present work provides a detailed justification for such a feature . by including an external field we obtained a rich magnetic phase diagram .
we plan to investigate its relevance to compound such as uge@xmath4 .
we note that intra - atomic excitations within the system of two localized 5@xmath0 electrons might provide the required bosonic excitations for itinerant electron attraction and cooper - pair formation @xcite .
one may also speculate about the application of the present findings to multiband hubbard models .
they are used to describe correlations in transition metal compounds @xcite .
this problem is left for a future investigation as well as the one of a u impurity embedded in a weakly correlated chemical environment .
also a molecular - field description of the effects found here , e.g. , by introducing auxiliary fields to model the correlation effects is an interesting subject of future research . in an extended version of this paper we plan to discuss also the phenomenon of charge disproportionation , which may occur in clusters of the type investigated here
in the limit @xmath196 , the 180 states of products of local @xmath31 configurations @xmath197 and local @xmath32 configurations @xmath198 are degenerate .
the corresponding states are direct products of maximal local @xmath115 states , i.e. , three - particle states with @xmath199 and two - particle states with @xmath200 .
finite hopping splits the manifold in first order degenerate perturbation theory .
the transition matrix element for hopping with initial product state @xmath201 @xmath202 and final state @xmath203 @xmath204 is @xmath205 where the @xmath206 are clebsch - gordan coefficients .
in the plane spanned by @xmath53 and @xmath84 , for small @xmath45 transitions occur near @xmath207 and at the isotropic line , see fig .
[ f2f3_phasentt2.eps ] .
these transitions are reproduced with the perturbative treatment .
furthermore , the perturbative analysis also describes the degeneracies of the ground states correctly . of particular interest
is the behavior along the isotropic line @xmath208 where the lifting of degeneracy can be solved in closed form . in this limit @xmath85@xmath209 provides an additional good quantum number .
we therefore consider basis states which are eigenstates of @xmath210@xmath209 , i.e. , @xmath211 and @xmath212 .
they are obtained as linear combinations of the above - mentioned product states @xmath213 and @xmath214 . in this basis , the hopping matrix elements depend only upon @xmath85 and can be expressed in terms of the 6-@xmath194 symbols @xmath215 where @xmath216 4 & { \cal j } & \frac{9}{2 } \end{array } \right)\ ] ] accounts for the variation with @xmath85 of the hopping matrix element .
this result allows for a simple interpretation . in the @xmath217 model ,
the total angular momentum @xmath85 of the 5-electron states @xmath218 and @xmath219 arises from the coupling of three angular momenta .
they characterize @xmath31 configurations at the two sites @xmath69 and @xmath70 with angular momenta 4 and an additional electron which can sit at either site .
the angular momentum @xmath220 of this electron couples with that of the corresponding @xmath31-configurations yielding the angular momentum @xmath221 or @xmath222 of the @xmath32-configurations at site @xmath69 or @xmath70 , respectively . technically speaking , the states @xmath223 and @xmath224 arise from applying different coupling schemes to the three above - mentional angular momenta . hopping between the two sites connects the two types of states . as a consequence , the hopping matrix elements can be expressed in terms of the 6@xmath194-symbols which transform between the two different coupling schemes .
the variation with @xmath85 of the hopping matrix elements is easily evaluated yielding @xmath225 here @xmath85 is restricted to half - integer values @xmath226 .
the variation with @xmath85 of @xmath227 is displayed in fig .
[ fig:6j ] .
the maximum value for @xmath228 is obtained for @xmath229 corresponding to a sixfold degenerate ground state in the isotropic case .
with @xmath85 in the isotropic case.[fig:6j ] ]
at the phase boundaries in fig .
[ f2f3_phasentt2.eps ] the jumps of the @xmath230 component of the total magnetization of a two - site cluster coincide with changes of the orbital - resolved occupation number @xmath231 in the ground - state wavefunction .
there are shown as gray - scale plots in fig .
[ figwwtt ] .
they confirm the scenario developed in the body of the text and allow for some additional observations .
the figure corresponds to the @xmath232 phase diagram shown in fig .
[ f2f3_phasentt2.eps ] in an infinitesimal ferromagnetic field . for weak
hopping with @xmath233 , the @xmath132 and the @xmath159 orbitals are fully occupied on both sites , whereas @xmath234 is half - occupied , allowing for a maximal kinetic energy gain .
the corresponding total angular momentum component @xmath235 .
similarly for large @xmath53 values the @xmath64 phase is formed in which one of the @xmath159 electrons is promoted to @xmath97 . in all other phases ,
the average occupations of all orbitals are close to one half or zero .
occupations near one half could indicate frequent hopping .
such a paramagnetic interpretation is appropriate only if @xmath236 is much larger than the energy differences @xmath237 resulting in hund s rule .
the three phases to be discussed are best classified by looking at the particular @xmath6 orbital which remains unoccupied ( light - colored areas in the figure ) . obviously , for the case @xmath238 , the loss of hopping energy gain is minimal if @xmath97 remains empty , yielding @xmath239 . in the case
@xmath240 , either the orbital @xmath241 or @xmath242 is unoccupied , leading to @xmath243 and @xmath244 , respectively . it should be kept in mind that the coulomb interaction mixes various @xmath6 orbitals .
the concept of occupying @xmath6 states is therefore only approximate .
this is reflected in small changes of the occupation within each phase .
g. zwicknagl , a. n. yaresko , and p. fulde , phys .
b * 65 * , 081103 ( 2002 ) g. zwicknagl and p. fulde , cond - mat/0211706 r.c .
albers , a.m. boring , and n.e .
christensen , phys .
b * 33 * , 8116 ( 1986 ) n. kernavanois , b. genier , a. huxley , e. ressouche , j. p. sanchez , and j. flouquet , phys . rev .
b * 64 * , 174509 ( 2001 ) h. yamagami and a. hasegawa , physica b * 186 - 188 * , 182 ( 1993 ) ; see also a. b. shick and w. e. pickett , phys .
* 86 * , 300 ( 2001 ) k.o .
klepp and w. bronger , z.anorg .
chem * 532 * , 23 ( 1986 ) ; p. mller and w. bronger , z. naturforsch . *
34b * , 1264 ( 1979 ) ; w. bronger , u. ruschewitz , and p. mller , j. alloys compounds * 187 * , 95 ( 1992 ) ; h .- h .
schmidtke , r. packroff , w. bronger , and p. mller , chem .
. lett . * 150 * , 129 ( 1988 ) ; m. koga , w. liu , m. dolg and p. fulde , phys .
b * 57 * , 10648 ( 1998 ) p. fulde , l.l .
hirst and a. luther , z. phys . *
230 * , 155 ( 1970 ) , see also p. fulde , _
handbook on the physics and chemistry of rare earth _ ,
2 , ed . by k. a. gschneidner and l. eyring ( north - holland , amsterdam , 1979 ) n.k . sato , n. aso , k. miyake , r. shiina , p. thalmeier , g. varelogiannis , c. geibel , f. steglich , p. fulde , and t. komatsubara , nature * 410 * , 240 , ( 2001 ) t. hotta and k. ueda , cond - mat/0211706 see , e.g. , chap .
11 of : p. fulde , _ electron correlations in molecules and solids _
( springer , heidelberg 1995 ) | we study the effect of intra - atomic correlations on anisotropies of hopping of 5@xmath0 electrons . it is shown that they may considerably enhance these anisotropies to the extend that electrons in some of the 5@xmath0 orbitals remain localized .
this provide a microscopic basis for a previously made assumption that some of the 5@xmath0 electrons must be partially localized when fermi surfaces and effective masses of the component are calculated .
calculations are performed on two- and three - sites clusters .
they include phase diagrams in the presence and absence of an external magnetic field as function of anisotropic hopping . # 1 # 1 1 |
exact solutions hold a special place in the theoretical condensed matter physics of interacting electron systems .
although they can be obtained only for very specific models , these proved to be very useful in understanding the behaviour of more general models of interacting electrons , or informing us of novel physics that we would otherwise not suspect from approximate treatments . in particular , our present paradigm of two universality classes , fermi liquids versus luttinger liquids , for low - dimensional systems of interacting fermions came out of exact solutions showing separation of the charge and spin degrees of freedom @xcite . in this paper , we report further surprises coming out of the exact solution of models of hardcore bosons and spinless fermions with infinite nearest - neighbor repulsion @xcite .
we consider chain models @xmath0 + u\sum_j n_j ( n_j - { \mathbb{1 } } ) \\ & \quad\ + \expandafter v\sum_j n_j n_{j+1 } , \\
h_{tv}^{(c , f ) } & = -t\sum_j\left[{c_{j}^{\dagger}}{c_{j+1 } } + { c_{j+1}^{\dagger}}{c_{j}}\right ] + v\sum_j n_j n_{j+1 } , \end{aligned}\ ] ] of hardcore bosons ( @xmath1 ) and spinless fermions , as well as a ladder model @xmath2 of spinless fermions . in this ladder model@xcite ,
the correlated hopping @xmath3 is the simplest term we can introduce to blatantly favor the emergence of superconducting order . throughout this paper
, we will specialize to the limit of infinite onsite repulsion @xmath1 and infinite nearest - neighbor repulsion @xmath4 .
more precisely , we admit only configurations in which each site can be occupied by at most one particle , with no simultaneous occupation of nearest - neighbor sites
. we will show how the _ nearest - neighbor excluded chain models _ can be mapped to the _ nearest - neighbor included chain models _ @xmath5 + u\sum_j n_j ( n_j - { \mathbb{1 } } ) , \\
h_t^{(c , f ) } & = -t\sum_j\left[{c_{j}^{\dagger}}{c_{j+1 } } + { c_{j+1}^{\dagger}}{c_{j}}\right ] , \end{aligned}\ ] ] and ultimately solve for the ground states of the former in terms of the one - dimensional fermi sea .
we will also show how the ladder model can be solved exactly in three limiting cases , by mapping their ground states to those of the chain models given in eq . and eq . .
these analytical results were used to guide a density - matrix analysis of correlations for the ladder model , first using the exactly diagonalized ground states @xcite , and later using the density - matrix renormalization group @xcite . for the rest of this paper , we will consistently use uppercase letters @xmath6 and @xmath7 ( @xmath8 and @xmath9 ) to denote hardcore boson ( spinless fermion ) annihilation and creation operators on nearest - neighbor excluded chains , and lowercase letters @xmath10 and @xmath11 ( @xmath12 and @xmath13 ) to denote hardcore boson ( spinless fermion ) annihilation and creation operators .
similarly , @xmath14 ( or @xmath15 ) and @xmath16 ( or @xmath17 ) are the hardcore boson ( spinless fermion ) occupation number operator on the nearest - neighbor excluded and nearest - neighbor included chains respectively .
hereafter , we will also use _ excluded _ to refer to all quantities associated with the nearest - neighbor excluded chain , and _ ordinary _ to refer to all quantities associated with the nearest - neighbor included chain .
our paper will be organized as follows : in sec .
[ sect : maptech ] , we will describe an analytical map that establishes a one - to - one correspondence between the hamiltonian matrices of the excluded and ordinary chains of hardcore bosons and spinless fermions , before developing a systematic expansion that would allow us to calculate ground - state expectations in bosonic and fermionic excluded chains .
we then present and analyze in sec .
[ sect : correlationsnnebosons ] correlation functions calculated using the analytical tools developed in sec .
[ sect : maptech ] for excluded chains of hardcore bosons and spinless fermions . following this , we write down in sec .
[ sect : laddermodels ] the exact ground states of the ladder model given in eq . in three limiting cases , and calculate various ground - state correlation functions , before summarizing our results and discuss the interesting physics they imply in sec .
[ sect : conclusions ] .
in sec . [ sect : nnetonni ] , we establish a one - to - one correspondence between states of the nearest - neighbor excluded and nearest - neighbor included chains .
we explain how the hamiltonian matrices of the two chains , and hence their energy spectra , are identical to one another . in the infinite - chain limit ,
we then show how we can write the ground state of the excluded chain in terms of the ground state of the ordinary chain , and ultimately be written in terms of the one - dimensional fermi sea . in sec .
[ sect : nnegsexp ] , we show how the ground - state expectation between two local operators can be calculated for the excluded chain , by writing it as a systematic sum over conditional expectations , each of which associated with a fixed configuration of intervening particles . for the sake of definiteness , let us consider open chains of a finite length @xmath18 and total particle number @xmath19 .
sites on these chains are indexed by @xmath20 . since the models in eq . and
eq . conserve @xmath19 , the infinite - chain limit is obtained by letting @xmath21 , keeping the density of particles @xmath22 fixed . ultimately , the results we present in this section will not depend on what boundary conditions we impose on the chain ( which is what we would expect in the infinite chain limit ) . for convenience ,
we establish some notations to cover boson and fermion cases together in the same formula .
let us call @xmath23 an excluded configuration , where @xmath24 for hardcore bosons , @xmath25 for spinless fermions , and the sites @xmath26 are such that @xmath27 .
we will also employ the labels @xmath28 and @xmath29 for distinct @xmath19-particle configurations @xmath30 and @xmath31 , i.e. the @xmath19-particle configurations @xmath32 and @xmath33 differ in at least one site .
similarly , let us call @xmath34 an ordinary configuration , where @xmath35 for hardcore bosons , @xmath36 for spinless fermions , and the the sites @xmath26 are such that @xmath37 .
the labels @xmath28 and @xmath29 will again denote distinct @xmath19-particle configurations @xmath38 and @xmath39 .
we will also consistently denote the hamiltonian of an excluded chain by @xmath40 , where @xmath41 for hardcore bosons , @xmath42 for spinless fermions , and the hamiltonian of an ordinary chain by @xmath43 , where @xmath44 for hardcore bosons , @xmath45 for spinless fermions . in this subsection
, our goal is to establish the one - to - one correspondence between states of the excluded and ordinary chains , and to show that as matrices , the hamiltonians and are identical . to do this ,
let us note that an excluded chain with @xmath18 sites has fewer @xmath19-particle states than an ordinary chain of @xmath18 sites , because of the infinite nearest - neighbor repulsion .
therefore , we can form a one - to - one correspondence between excluded and ordinary states only if the length @xmath46 of the ordinary chain is shorter than @xmath18 .
there are several ways to systematically map excluded configurations to ordinary configurations : we can ( i ) delete the site to the right of every particle , if it is not the rightmost particle ; or ( ii ) delete the site to the left of every particle , if it is not the leftmost particle .
we can easily check that these maps produce the same ordinary configurations for finite open chains .
we expect this to hold true even as we go to the infinite chain limit . for the rest of this paper
, we will adopt _ right - exclusion map _ @xmath47 that maps a @xmath19-particle configuration on an open excluded chain of length @xmath18 to a @xmath19-particle configuration on an open ordinary chain of length @xmath46 . the empty site to the right of each occupied site in the open excluded chain
is deleted , to give a corresponding configuration for an open ordinary chain .. ] as illustrated in fig .
[ fig : rightexclusionmap ] , we do not delete any empty site to the right of the @xmath19th particle , and thus the effective length of the open ordinary chain is @xmath48 .
this nearest - neighbor exclusion map was first used by fendley to map a supersymmetric chain of spinless fermions to the @xmath49 chain @xcite .
it tells us that an excluded chain with density @xmath50 gets mapped to an ordinary chain with density @xmath51 thus , in the limit of @xmath52 , @xmath53 -particle configurations on an open excluded chain of length @xmath18 to @xmath19-particle configurations on an open ordinary chain of length @xmath48 , by deleting one empty site to the right of a particle that is not the rightmost particle . ] for @xmath19-particle excluded configurations @xmath30 and @xmath31 , the matrix element @xmath54 is nonzero only when @xmath30 and @xmath31 can be obtained from one another by a single particle hopping to the left or the right . when this is so , the ordinary @xmath19-particle configurations @xmath38 and @xmath39 they map to under the right - exclusion map are also related to each other by a single particle hop .
thus , we have @xmath55 since there is a one - to - one correspondence between @xmath19-particle configurations on an open excluded chain and @xmath19-particle configurations on an open ordinary chain , eq .
tells us that @xmath40 and @xmath43 are identical as matrices in their respective configurational bases .
we therefore conclude that the @xmath19-particle energy spectra of the two chains coincide , and that there is a one - to - one correspondence between the energy eigenstates , @xmath56 for the excluded chain , and @xmath57 for the ordinary chain .
that is , if @xmath58 , and @xmath59 , then @xmath60 and @xmath61 have the same amplitudes in @xmath62 and @xmath63 respectively .
this result has profound implications on the thermodynamics of the two chains , as well as that of the ladder model in the three limiting cases described in sec .
[ sect : briefthreelimits ] , because their partition functions are the same .
however , for the rest of this paper , we limit ourselves to the ground states of the infinite excluded and ordinary chains , as well as those of the infinite ladder . in this subsection , we explain how the expectation @xmath64 of an observable @xmath65 in the ground state of the excluded chain can be computed , by relating it to the expectation @xmath66 in the ground state of the ordinary chain , for an appropriately chosen observable @xmath67 satisfying some basic correspondence requirements that we shall outline . specifically , we are interested in the correlations @xmath68 between two local observables @xmath69 and @xmath70 separated by a distance @xmath71 within the excluded - chain ground state . however ,
the right - exclusion map maps excluded matrix elements @xmath72 to ordinary matrix elements @xmath73 in which @xmath74 and @xmath75 are separated by varying separations . to deal with this problem
, we develop a method of intervening - particle expansion involving a sum over conditional expectations . to begin ,
let us look at the ground states @xmath76 of the excluded and ordinary chains respectively . to take advantage of the equality of amplitudes , @xmath77
if @xmath58 under the right - exclusion map , we want @xmath78 to have a simple relation with @xmath79 . while it is possible to pick @xmath67 such that @xmath80 for all @xmath28 and @xmath29
, we find that it is more convenient to pick @xmath67 such that @xmath81 for example , if @xmath82 , we can pick the corresponding observable to be @xmath83 , in which case we find that @xmath84 which satisfies eq .
we call @xmath65 and @xmath67 a _ corresponding pair of observables _ , if eq .
is satisfied for all @xmath28 and @xmath29 , allowing us to write the very simple relation @xmath85 between their ground - state expectations .
since we are mostly interested in correlation functions within the excluded chain ground state , let us look at expectations of the product form @xmath86 , where @xmath87 acts locally about site @xmath88 , and @xmath89 acts locally about site @xmath90 .
becasue the number of particles @xmath91 between sites @xmath88 and @xmath92 varies from excluded configuration to excluded configuration , these sites get mapped by the right - exclusion map to sites on the ordinary chain with varying separations @xmath93 . therefore , to calculate the excluded chain ground - state expectation @xmath86 in terms of ordinary chain ground - state expectations , we first write down an _ intervening - particle expansion _
@xmath94 where @xmath95 are conditional expectations . here
@xmath96 is a vector of occupation numbers within the intervening sites , and @xmath97 is a string of factors , each of which is either @xmath98 or @xmath99 , @xmath100 .
the sum is over all possible ways to have intervening particles between @xmath87 and @xmath89 . for each excluded term @xmath95 in eq .
, we then write down the corresponding ordinary expectation @xmath101 , and thereafter sum over all corresponding ordinary expectations , @xmath102 making use of eq . .
the vector @xmath103 of occupation numbers is obtained from @xmath96 using the right - exclusion map , and contains the same number @xmath91 of occupied intervening sites . to illustrate how the corresponding expectations @xmath104 can be constructed ,
let us write eq .
out explicitly as @xmath105 each of which contains intervening particles at fixed sites .
we call terms in the expansion with @xmath91 intervening @xmath106 s the _ @xmath91-intervening - particle expectations_. because of nearest - neighbor exclusion , most of the terms in eq . vanish . next , we map each conditional excluded expectation in eq . to a corresponding conditional ordinary expectation following the simple rules given below : 1
. * nearest - neighbor exclusion . * to ensure that we do not violate nearest - neighbor exclusion , we make the assignment @xmath107 note that this is intended not as a statement on the operator algebra , but as a mere bookkeeping device for evaluating expectations .
the assignment @xmath108 follows from eq . .
* right - exclusion map . * the right - exclusion map described in section [ sect : nnetonni ] is then implemented by making the substitution @xmath109 the assignment @xmath110 follows from eq . .
re - indexing . * because the right - exclusion map in eq .
merges the occupied site @xmath111 and the empty site @xmath112 to its right , operators to the right of site @xmath112 must be re - indexed .
the index @xmath111 on the excluded chain becomes @xmath113 on the ordinary chain .
thus , two ending operators @xmath71 sites apart in the @xmath91-intervening - particle excluded expectation becomes @xmath93 sites apart in the corresponding @xmath91-intervening - particle ordinary expectation .
in this section , we make use of the tools developed in sec . [
sect : maptech ] to calculate three simple correlation functions within the ground states of the excluded chains of hardcore bosons and spinless fermions .
in general , the intervening - particle expansion for excluded chain ground - state correlations must be evaluated numerically ( even when each ordinary chain ground - state expectations in the sum can be expressed in closed form ) , keeping in mind a excluded chain with density @xmath114 maps to an ordinary chain with density @xmath115 ( see eq . ) . in sec .
[ sect : flchain ] , sec .
[ sect : cdwchain ] , and sec .
[ sect : scchain ] , we show numerical results for the two - point functions @xmath116 and @xmath117 , and the four - point functions @xmath118 , @xmath119 and @xmath120 respectively . for the sake of easy reference , we will call the two - point functions @xmath116 and @xmath117 fermi - liquid ( fl ) type correlation functions , even though their spatial structures depend on particle statistics .
we will also call the four - point functions @xmath121 cdw type correlation functions , and the the four - point functions @xmath119 and @xmath120 sc type correlation functions .
both cdw and sc type correlations are identical for hardcore bosons and spinless fermions on the excluded chain , but the latter has the ` superconducting ' interpretation only for fermions . in sec .
[ sect : flchain ] , we will also explain in detail how nonlinear curve fits of the numerical correlations to reasonable asymptotic forms as a function of @xmath71 are done .
based on results from the nonlinear curve fits , we show how the luttinger s theorem manifests itself , and how meaningful power - law exponents can be extracted .
similar analyses are done in sec .
[ sect : cdwchain ] and sec .
[ sect : scchain ] , as well as in sec .
[ sect : laddermodels ] for the three limiting ground states of the ladder model . in the intervening - particle expansions of the two - point functions @xmath116 and @xmath117 , the nonvanishing terms map to @xmath91-intervening - particle expectations of the form @xmath122 and @xmath123 . both can be evaluated in terms of two - point functions @xmath124 of the one - dimensional fermi sea , after invoking the jordan - wigner transformation ( see appendix [ sect : jwtrans ] ) for the former . as shown in the inset of fig .
[ figure : flbn20n25n30 ] , the fl correlation @xmath116 was found to consists of a simple power law part , decaying with a smaller exponent @xmath125 , and an oscillatory power law part , decaying with a larger exponent @xmath126 .
multiplying @xmath116 by various simple powers of @xmath71 , we find that @xmath127 asymptotes to a constant with large @xmath71 ( as shown in the main plot of fig .
[ figure : flbn20n25n30 ] ) , which suggests that @xmath128 .
this is the correlation exponent predicted by efetov and larkin , in their study of the ordinary chain of hardcore bosons @xcite . as a function of @xmath71 for the particle densities @xmath129 ( red circles ) , @xmath130 ( green squares ) , @xmath131 ( blue diamonds ) .
the colored curves shown are nonlinear curve fits of @xmath132 to the asymptotic form @xmath133 .
( inset ) plot of @xmath116 as a function of @xmath71 , showing that it consists of a simple power law part and an oscillatory power law part . ]
another important result comes from the unrestricted nonlinear curve fits of @xmath127 to the asymptotic form @xmath134 . in fig .
[ figure : flbkfit ] we show a plot of the fitted wave number @xmath135 as a function of the density @xmath114 of the excluded chain of hardcore bosons .
as we can see , the fitted wave numbers fall neatly onto the straight line @xmath136 , where @xmath137 is the fermi wave number .
the fact that @xmath138 appears naturally in the numerical correlations is expected from luttinger s theorem , which states that the volume of the reciprocal space bounded by the noninteracting fermi surface is invariant quantity not affected by interactions , and applies in both fermi and non - fermi liquids @xcite . from this point
onwards , we restrict the wave number of the oscillatory part of the correlation functions to @xmath138 , @xmath139 , or @xmath140 in the nonlinear curve fits .
( solid circles ) against the density @xmath114 of the excluded chain of hardcore bosons .
the parameter @xmath135 is obtained from the unrestricted nonlinear curve fit of @xmath127 to the asymptotic form @xmath141 .
the straight line is @xmath142 for the fermi wave number . ] from the restricted nonlinear curve fits , we find that @xmath143 is large when @xmath144 is large , and small when @xmath144 is small .
this suggests that neither of these parameters can be accurately determined from our nonlinear curve fits , unless we further constrain what values @xmath144 can take .
we also find that the quality of the nonlinear curve fit is good when @xmath114 is far from @xmath145 , but deteriorates as we approach quarter filling .
this suggests important physics in the fl correlation @xmath116 near quarter filling , which can not be adequately accounted for by an asymptotic form @xmath146 .
this loss of fit also affects the phase shift @xmath147 , presumably to a smaller extent , and the amplitude @xmath148 of the simple power law , to an even smaller extent .
these two parameters are plotted as functions of the density in fig .
[ figure : flbaphifit ] . in the limit @xmath149 , we have essentially a dilute gas of hardcore ( otherwise noninteracting ) bosons , and thus @xmath116 should include an overall factor of @xmath114 .
thus we expect @xmath150 as @xmath149 . in the limit
@xmath151 , the excluded chain of hardcore bosons becomes increasingly jammed , and the relevant degrees of freedom are holes . for a dilute chain of holes , we expect @xmath116 to be proportional to the hole density , and thus @xmath150 as @xmath151 . from our numerical results
alone , it is hard to tell whether @xmath148 reaches a maximum at @xmath152 ( corresponding to a quarter - filled , @xmath153 , ordinary chain of hardcore bosons ) or @xmath154 ( quarter - filled excluded chain of hardcore bosons ) .
it is also hard to say anything definite about the phase shift @xmath147 , which might in fact be constant .
( top ) of the simple power law part , and the fitted phase shift @xmath147 ( bottom ) of the oscillatory power law part , as functions of the density @xmath114 of the excluded chain of hardcore bosons . ]
in contrast , the fl correlation @xmath117 contains no simple power law part .
a preliminary unrestricted nonlinear curve fit of this correlation to the asymptotic form @xmath155 suggests that @xmath156 for all densities , i.e. may be universal @xmath157 just as for noninteracting spinless fermions .
however , a more careful restricted nonlinear curve fit @xmath158 show systematic deviations , as shown in fig .
[ figure : flfn10n25n40 ] , and therefore we perform an unrestricted fit to @xmath159 .
the fitted parameters @xmath143 , @xmath160 , and @xmath147 are shown in fig .
[ figure : flffits ] .
as functions of @xmath71 ( solid circles ) for particle densities @xmath161 ( top ) , @xmath162 ( middle ) , and @xmath163 ( bottom ) .
nonlinear curve fits of @xmath164 to the restricted asymptotic form @xmath165 ( solid curves ) show systematic deviations , which can be accounted for by a unrestricted asymptotic form @xmath166 . ] ( top ) , the fitted exponent @xmath126 ( middle ) , and the fitted phase shift @xmath147 of the leading oscillatory power - law decay in the fl correlation @xmath117 , as functions of the density @xmath114 of the excluded chain of spinless fermions . in these plots ,
the solid curves are for fits to @xmath167 , whereas the dashed curves are for fits to @xmath168 $ ] .
the dotted line in the bottom plot is a straight line from @xmath169 at @xmath170 to @xmath171 at @xmath172 to guide the eye . ] at very low densities @xmath149 , our dilute chain of nearest - neighbor excluding spinless fermions will behave like noninteracting fermions , so we expect @xmath173 from our curve fits , we see that in this limit , @xmath174 , @xmath175 , and @xmath176 , and thus the fl correlation does indeed go to the low density limit in eq . .
also , in the half - filling limit @xmath151 , the chain become more and more congested , making it increasingly difficult to annihilate a spinless fermion at site @xmath92 , find an empty site @xmath88 to create a spinless fermion , without running afoul of the nearest - neighbor exclusion constraint .
this tells us that @xmath143 must vanish as @xmath151 , which is hinted at in fig .
[ figure : flffits ] . however , the vanishing amplitude is only half of the story in this limit , the other half being the rate at which the correlation decay with increasing separation .
in fact , very close to @xmath172 , we expect the ground - state physics of the chain of rung - fermions with infinite nearest - neighbor repulsion to be describable in terms of a low density of holes .
naively , we would expect from such a low - density - of - holes argument that @xmath177 decay as @xmath178 .
instead , the nonlinear curve fits at @xmath179 tells us that @xmath180 .
thinking about this nearly - half - filled limit more carefully , we realized that what we called ` holes ' are really domain walls separating a region in which the spinless fermions sit on odd sites , from a region in which the spinless fermions sit on even sites .
the fl correlation @xmath117 , which can be written as a hole - hole correlation function , then depends on how many holes there are between @xmath88 and @xmath92 .
the idea is that , in order to annihilate a hole ( create a spinless fermion ) at site @xmath92 and create a hole ( annihilate a spinless fermion ) at site @xmath88 , we must first find a configuration with a hole at @xmath92 .
such a configuration will have spinless fermions at sites @xmath181 , @xmath182 ,
, until we encounter another hole at @xmath183 , and then the sequence of spinless fermions will thereafter be at sites @xmath184 , @xmath185 , .
if @xmath71 is even , @xmath117 receives nonzero contributions only from those configurations with an even number of intervening holes , whereas if @xmath71 is odd , @xmath117 receives nonzero contributions only from those configurations with an odd number of intervening holes .
this is very similar in flavor to the intervening - particle expansion of the two - point function @xmath186 of a chain of ordinary hardcore bosons , except that @xmath186 receives positive contributions from configurations with an even number of intervening particles , and negative contributions from an odd number of intervening particles .
therefore , in the limit @xmath151 , we find that the fl correlation @xmath117 maps to a string correlation of holes .
bosonization calculations then show that this string correlation of holes decay as a power law , with correlation exponent @xmath187 @xcite .
important physics can also be learnt from the nonlinear curve fitting of the cdw correlations @xmath188 .
first we tried to fit the subtracted cdw correlation to the asymptotic form @xmath189 , but found the quality of fit deteriorates as @xmath190 , as shown in fig .
[ figure : cdwbn10n25n45 ] .
we understand this as follows : for @xmath149 , the dimensionless quantity @xmath191 is small , and the poor fit indicates that @xmath118 contains contributions from a term that decays more rapidly than @xmath192 . if we assume that this faster decaying term is a simple power law of the form @xmath193 , and fit the cdw correlation to @xmath194 , we found the quality of the nonlinear curve fit is improved , after dropping data points @xmath195 from the fit , as shown in fig .
[ figure : cdwbn10n25n45 ] . we can include the simple power law decay term in the nonlinear curve fit throughout the entire range of @xmath114 , but the parameters @xmath196 and @xmath197 can not be reliably determined beyond @xmath145
. therefore , the parameters @xmath198 , @xmath199 , and @xmath200 are the only parameters that can be reliably determined across the whole range of densities . from fig .
[ figure : cdwbfits ] , we find that @xmath200 changes very little over the whole range of densities , and remains close to @xmath201 .
on the other hand , the leading correlation exponent @xmath199 appears to be density - dependent , and is very close to being @xmath202 ( solid circles ) in the bosonic / fermionic ground states of the excluded chain , at densities @xmath161 ( top ) , @xmath162 ( middle ) , and @xmath203 ( bottom ) . above quarter filling ( middle and bottoms plots ) , the subtracted cdw correlations can be fitted very well to the simple asymptotic form @xmath204 ( solid curves ) , whereas at low densities ( top plot ) , the subtracted cdw correlations deviate significantly from this simple asymptotic form .
the nonlinear curve fit improves only after we add a simple power - law correction term @xmath205 , giving the dashed curves . ]
( top ) , the fitted exponent @xmath199 ( middle ) and the fitted phase shift @xmath200 ( bottom ) of the leading oscillatory power law decay of the subtracted cdw correlation @xmath206 , as functions of the density @xmath114 of the excluded chain of hardcore bosons / spinless fermions . ]
just as for the fl correlation , we need to again think hard about the hole physics of the ground - state cdw correlation very close to half - filling . in this limit
, the cdw correlation @xmath207 can be written in terms of the hole - density - hole - density correlation @xmath208 , where @xmath209 is the hole occupation number operator on site @xmath88 . using an argument similar to the one used for the fl correlation @xmath177 near half - filling
, we realized that if @xmath71 is even , @xmath210 receives nonzero contribution from configurations with an even number of intervening holes , whereas if @xmath71 is odd , @xmath210 receives nonzero contribution from configurations with an odd number of intervening holes .
this means that our hole - density - hole - density correlation @xmath210 must be mapped to a string correlation of a chain of noninteracting spinless fermions .
bosonization calculations tell us that this string correlation decays as a power law with correlation exponent @xmath211 , consistent with the conjectured behaviour , eq . , at the special value @xmath172 @xcite .
in contrast to the fl and cdw correlations , the sc correlation @xmath212 on the excluded chain has a rather more complex structure . the sc correlation is always negative , and oscillations are highly suppressed , suggesting that it is the sum of a simple power law and an oscillatory power law . to improve the reliability of the nonlinear curve fits ,
we prescale the sc correlation by multiplying it by @xmath213 .
this strange exponent is chosen because it is closest to the rate at which the simple power law decay for various densities .
after dropping data points for @xmath214 , good fits to the asymptotic form @xmath215 were obtained .
the nonlinear curve fits were improved marginally by adding a correction term of the form @xmath216 ( see fig .
[ figure : scfn10n25n45 ] ) .
the fitted parameters are shown in fig .
[ figure : scffits ] as functions of the excluded chain density @xmath114 .
( solid circles ) in the bosonic / fermionic ground states of the excluded chain , at densities @xmath161 ( top ) , @xmath130 ( middle ) , and @xmath203 ( bottom ) . after dropping data points from @xmath217 ,
the sc correlations can be fitted well to the simple asymptotic form @xmath218 ( solid curves ) .
the nonlinear curve fits are marginally improved by adding a correction term of the form @xmath219 ( dashed curves ) . ] and @xmath220 ( top ) , the fitted exponents @xmath221 and @xmath222 ( middle ) , and the fitted phase shift @xmath223 of the leading simple power - law decay and the subleading oscillatory power - law decay in the sc correlation @xmath212 , as functions of the density @xmath114 of the excluded chain of hardcore bosons / spinless fermions . ] unlike for the fl and cdw correlations , there are no analytical bosonization calculations to help suggest values for the sc correlation exponents , so we used an ad - hoc process where we imposed trial values of the exponents , and let the nonlinear curve fitting program find the appropriate amplitudes and phase shifts .
we found visually that the best fit of the numerical sc correlations appears to the mixed asymptotic form @xmath224
in this section we show how the analytical machinery developed in sec .
[ sect : maptech ] can be adapted to calculate ground - state correlations in the ladder model of interacting spinless fermions given in eq .
, in three limiting cases where the ground states can be deduced from simple energetic arguments .
an overview is given in sec .
[ sect : briefthreelimits ] , before we move on to detailed analyses and discussions of the three limiting cases in sec .
[ sect : infinitelystrongcorrelatedhops ] , sec .
[ sect : independentlegs ] , and sec .
[ sect : independentrungs ] . as with the chain models , we assume that the ladder is finite , with @xmath225 , and subject each of its legs @xmath226 to open boundary conditions . exact solution for the infinite ladder
is then obtained by taking @xmath21 keeping the particle density @xmath227 fixed . just as for the chain models , we expect in this limit that the ladder exact solutions would not depend on which boundary conditions we used . for the ladder model described by eq . , with @xmath228 , the ground state is determined by the two independent model parameters , @xmath229 and @xmath230 , and the density @xmath227 . for fixed @xmath227 , the two - dimensional region in the ground - state phase diagram
is bounded by three limiting cases , 1 .
the _ paired limit _
@xmath231 , which we will discuss in detail in sec .
[ sect : infinitelystrongcorrelatedhops ] . in this limit ,
we find sc correlations dominating at large distances ( though , as for hardcore bosons , cdw correlations inevitably dominate at short distances ) .
based on our numerical studies in sec .
[ sect : infinitelystrongcorrelatedhops ] , the leading sc correlation exponent appears to be universal , with a value of @xmath232 , while the leading cdw correlation exponent @xmath29 is nonuniversal . in this limit , fl correlations are found to decay exponentially .
a staggered form of long - range cdw order also appears ; 2 .
the _ two - leg limit _ @xmath233 , @xmath234 , which we will discuss in detail in sec . [ sect : independentlegs ] . in this limit ,
the two legs of the ladder are coupled only by infinite nearest - neighbor repulsion .
the dominant correlations at large distances are those of a power - law cdw , for which we find numerically to have what appears to be an universal correlation exponent of @xmath235 . in this limit
, the leading sc correlation exponent was predicted analytically to be @xmath236 , while fl correlations are found to decay exponentially ; 3 .
the _ rung - fermion limit _ @xmath237 , @xmath238 , which we will discuss in detail in sec .
[ sect : independentrungs ] . in this limit
, the particles are effectively localized onto the rungs of the ladder . when the ladder is quarter - filled
, a true long - range cdw emerges in the two - fold degenerate ground state . below quarter - filling ,
we find numerically that the cdw power - law correlation dominate at large distances , with a leading non - universal correlation exponent @xmath239 .
the leading fl correlation exponent was also found numerically to be non - universal , with values going from @xmath240 to @xmath241 .
the sc correlation exponent , on the other hand , was found numerically to be universal , with value @xmath242 . to zeroth order ( i.e. without plunging into first - order perturbation theory calculations )
, the ground - state phase diagram can be obtained by interpolating between these three limiting cases .
there will be three lines of quantum phase transitions or cross - overs , which at quarter - filling , separate the long - range cdw ( lr - cdw ) , power - law cdw ( pl - cdw ) , and sc phases . at quarter - filling ,
we expect these three lines of critical points or cross - overs to meet at a point on the phase diagram
. if we have three lines of true critical points , this point would be a quantum tricritical point .
we therefore end up with a ground - state phase diagram which looks like that shown in figure [ fig : roughphasediagram ] . in this subsection , we solve for the ground - state wave function , and calculate various ground - state correlations in the paired limit @xmath243 . in this limit , the hamiltonian in eq .
simplifies to @xmath244 in sec .
[ sect : boundpairsdof ] , we explain how pairs of spinless fermions are bound by correlated hops in this limit , and the degrees of freedom in the system become mobile bound pairs with infinite nearest - neighbor repulsion .
these bound pairs come in two flavors , determined by the specific arrangement of the two bound - pair particles around a plaquette .
these flavors are conserved by correlated hops if the length of the ladder is even , and hence the ladder ground state is two - fold degenerate .
we then describe how these two degenerate ladder ground states can be mapped to a excluded chain of hardcore bosons , then to an ordinary chain of hardcore bosons , and finally to a chain of noninteracting spinless fermions . in sec .
[ sect : tonkscorr ] , we calculate the sc and cdw correlations , using the intervening - particle expansion described in sec .
[ sect : nnegsexp ] .
we then use a restricted - probability argument in sec .
[ sect : explainflcorr ] to show that fl correlations decay exponentially with distance , governed by a density - dependent correlation length , in this paired limit .
we find , as expected from making the absolute correlated hopping amplitude @xmath245 large , that sc correlations dominate at large distances . in the paired limit @xmath231 ,
we solve for the ground state of the simplified hamiltonian given by eq .
, which admits only correlated hops .
because of this , isolated spinless fermions can not hop at all ; by contrast , a pair occupying diagonal corners on a plaquette can perform correlated hops . therefore , for an even number of spinless fermions , ground - state configurations consist of well - defined bound pairs , which are effectively bosons .
we say that a bound pair at @xmath246 and @xmath247 has even ( resp .
odd ) flavor if its two sites are even ( resp .
odd ) sites . in this limit of @xmath248 , a particle on rung @xmath88 can only hop to rung @xmath249 .
this moves the bound pair s center of mass by one lattice constant , without changing its flavor .
the degrees of freedom in this limiting case thus becomes bound pairs with definite flavors hopping along a one - dimensional chain .
subject to periodic boundary conditions . for ladders of odd length @xmath18 subject to periodic boundary conditions , the flavor of a bound pair changes as it goes around the boundary of the ladder , and
so the conserved quantum numbers are not the even and odd flavors , but are instead the symmetric and antisymmetric combinations of the two flavors .
this mixing between even and odd flavors lifts the ground - state degeneracy , giving a nondegenerate many - bound - pair ground state whose quantum number is the antisymmetric combination of flavors . in this paper
we consider only ladders of even length @xmath18 , because we want to work with ground states containing bound pairs with a definite flavor .
] we write these hardcore boson operators in terms of the spinless fermion operators as @xmath250 and @xmath251 where we order first with respect to the leg index , and then with respect to the rung index of of the ladder .
since bound pairs can not move past each other along the chain , the @xmath19-bound - pair hilbert space breaks up into many independent sectors , each with a fixed sequence of flavors .
the @xmath19-bound - pair problem in one sector is therefore an independent problem from that of another @xmath19-bound - pair sector .
the minimum energy in each sector can be very crudely determined by treating the @xmath19-bound - pair problem as a particle - in - a - box problem , where each bound pair is free to hop within a ` box ' demarcated by its flanking bound pairs .
as shown in figure [ fig : closestapproach ] , two bound pairs with the same flavor can get within a separation @xmath252 of each other , whereas two bound pairs with different flavors can only achieve a closest approach with separation @xmath253 . therefore , for a fixed separation between flanking bound pairs , the kinetic energy of the ` boxed ' bound pair is lowest when all three bound pairs have the same flavor . repeating this argument for all bound pairs
, we realized therefore that the two - fold degenerate ground state lies within the all - even and all - odd sectors . in these sectors ,
bound pairs can not occupy nearest - neighbor plaquettes , i.e. we are dealing with an excluded chain of hardcore bosons . by one , thereby increasing the particle density ( and hence the energy of that sector s ground state ) . ]
the twofold degeneracy between all - even and all - odd sectors represents a symmetry breaking with long - range order of a staggered cdw type ( in terms of fermion densities ) .
it may be viewed as breaking the invariance under reflection about the ladder axis of the original hamiltonian as given in eq . . which permits a tiny tunnel amplitude between the even and odd sectors in finite ladders .
the symmetrized ground states are @xmath254 . ]
thus the quantum - mechanical problem of a ladder with density @xmath227 is mapped to the quantum - mechanical problem of an excluded chain with density @xmath255 .
three simple correlation functions , the fl , cdw , and sc correlations , were computed for the excluded chain of hardcore bosons in sec .
[ sect : correlationsnnebosons ] . on the ladder model in the paired limit , these correlations must be interpreted differently . in mapping the ladder model to the excluded chain ,
we replace a pair of spinless fermion by a hardcore boson operator , i.e. @xmath256 .
thus the fl correlation @xmath116 of hardcore bosons actually corresponds to a sc correlation of the fermion model on the ladder .
depending on which of the two degenerate ladder ground states we are looking at , the sc operators are @xmath257 such that @xmath258 because of eq .
, we shall drop the indices @xmath259 and @xmath260 from here on . from sec . [
sect : correlationsnnebosons ] , we know that @xmath261 decays with separation @xmath71 asymptotically as the sum of a simple ( leading ) power law and an @xmath262-oscillatory ( subleading ) power law . the leading correlation exponent has been determined to be @xmath263 , while the subleading correlation exponent @xmath222 can not be reliably determined .
the simplest cdw correlations are @xmath264 which we call the cdw-@xmath265 correlations .
these are not easy to calculate , because they can not be written simply in terms of the expectations of hardcore boson operators .
in contrast , the cdw-@xmath266 correlations correlations @xmath267 can not be written as simple linear combinations of eight - point functions because a term like @xmath268 will pick up contributions from configurations that @xmath269 will not . this tells us that @xmath269 is some messy linear combination of eight - point , twelve - point , sixteen - point , , @xmath270-point functions .
] @xmath271 can be evaluated with the help of the intervening - particle expansion in eq . .
this was done in sec .
[ sect : correlationsnnebosons ] , where we found the subtracted cdw-@xmath266 correlation @xmath272 decaying asymptotically with separation @xmath71 as a simple power law @xmath273 , with a nonuniversal leading correlation exponent @xmath274 , and a universal phase shift of @xmath275 . unlike the sc and cdw-@xmath266 correlations
, the fl correlations can not be calculated easily in this paired limit , because the operators involved can not be written in terms of hardcore boson operators .
nevertheless , we can still calculate it by making use of the fact that this correlation is very close to being the probability of finding a restricted class of configurations in the ground state .
we then make use of the scaling form reported in ref . to calculate the probability analytically .
this idea is exploited again in sec . [
sect : weakxy ] in this paired limit , the ground state consists exclusively of a superposition of bound pair configurations .
therefore , if we annihilate a spinless fermion on leg @xmath276 , we must create another on the same leg elsewhere , and thus the only nonzero fl correlations are of the form @xmath277 .
in fact , to start with a paired configuration and end up with another paired configuration , after annihilating a spinless fermion at @xmath90 and creating a spinless fermion at @xmath88 , the initial and final configurations must contain a compact cluster of pairs between rung @xmath88 and rung @xmath92 , as shown in fig .
[ fig : compactclusterargument ] .
-bound - pair cluster configurations making nonzero contribution towards the expectation of the fl operator product @xmath278 , which annihilates a spinless fermion on the right end of the compact @xmath91-bound - pair cluster , and creates a spinless fermion on the left end of the compact @xmath91-bound - pair cluster . @xmath279 and @xmath280 are the ground - state amplitudes of the initial and final configurations respectively . ]
based on this compact cluster argument , we know that @xmath281 when @xmath71 is odd .
when @xmath282 is even , @xmath283 receives contributions from all pairs of configurations with a compact @xmath91-bound - pair cluster between the rungs @xmath88 and @xmath284 . clearly , these products of amplitudes will depend on where the other bound pairs are on the ladder . however , if the ladder is not too close to half - filling , we expect @xmath285 , so that on an infinite ladder , @xmath286 is very nearly the probability of finding a compact @xmath91-bound - pair cluster
,- particle sector of the cluster density matrix of a @xmath287-site cluster , and look at the matrix element between a configuration with an empty site at the left end of the cluster and a configuration with an empty site at the right end of the cluster .
however , the relevant cluster density matrix is that of a system of hardcore bosons . while this hardcore - boson cluster density matrix should be simply related to the noninteracting - spinless - fermion cluster density matrix , this relation has not been worked out , for use on this problem of finding fl correlations at large @xmath71 for the bound - pair ground states on a two - legged ladder . ]
@xmath288 after using the relation between excluded and ordinary expectations , where @xmath289 is the density of the ordinary chain .
here @xmath290 is the determinant of the noninteracting - spinless - fermion cluster green - function matrix @xmath291 for a cluster of @xmath91 sites , which we can write as@xcite @xmath292 where @xmath293 are the eigenvalues of the cluster green - function matrix @xmath291 , and @xmath294 are the single - particle pseudo - energies of the cluster density matrix @xmath295 , for the cluster of @xmath91 sites in an infinite chain of noninteracting spinless fermions . for @xmath296
, @xmath291 has approximately @xmath297 eigenvalues which are almost one , and approximately @xmath298 eigenvalues which are almost zero , the determinant of @xmath291 is thus determined predominantly by the approximately @xmath298 eigenvalues which are almost zero . for these @xmath293 , @xmath299 , and thus @xmath300 where @xmath301 is such that @xmath302 . converting the sum into an integral , and using the approximate scaling formula in ref .
, we find that @xmath303 i.e. the probability of finding a compact @xmath91-bound - pair cluster decays exponentially with @xmath91 in the limit of @xmath296 . with this simple compact cluster argument
, we conclude that the ladder fl correlation @xmath304 decays exponentially with separation @xmath71 as @xmath305,\ ] ] with a density - dependent correlation length @xmath306 in the strong correlated hopping limit . from ref .
we know that the scaling function @xmath307 depends only very weakly on @xmath308 , and thus , at very low ladder densities @xmath309 , the correlation length @xmath310 attains its minimum value of @xmath311 and the fl correlation @xmath312 decays most rapidly in this regime of @xmath309 .
this is expected physically , since a long cluster of occupied sites is very unlikely to occur at very low densities , with or without quantum correlations . in the regime of @xmath313
, we find @xmath314 , and thus the correlation length @xmath315 diverges according to eq . .
this diverging correlation length tells us nothing about the amplitude of the fl correlation .
indeed , when the ladder becomes half - filled , the two degenerate ground states are inert bound - pair solids .
each of the half - filled - ladder ground - state wave functions consists of a single configuration whereby all available plaquettes are occupied by a bound pair , and it is not possible to annihilate a spinless fermion at the @xmath316th rung and create another at the @xmath88th rung .
the fl correlation @xmath312 is thus strictly zero in this half - filled - ladder limit .
this subsection concerns the ground state in the two - leg limit @xmath233 , @xmath317 .
based on energetic considerations , we argue in section [ sect : nointerleggs ] that there will be two degenerate ground states , within which successive spinless fermions are on alternate legs of the ladder .
we call these the _ staggered ground states _ , and write their wave functions in terms of the fermi sea ground - state wave function with the help of a _ staggered map _ between ladder configurations and ordinary chain configurations .
we then calculate various ground - state correlations in sec .
[ sect : weakxy ] , sec .
[ sect : weakxycdw ] , and sec .
[ sect : weakxysc ] , where we show that the non - vanishing fl correlations decay exponentially with distance , governed by a density - dependent correlation length , while the cdw and sc correlations decay with distance as power laws .
we find in this two - leg limit that the antisymmetric cdw correlation dominates at large distances . in the limit of @xmath318 , each spinless fermion on the two - legged ladder carries a permanent leg index , and thus the number of spinless fermions @xmath319 on leg @xmath276 are good quantum numbers .
furthermore , successive spinless fermions along the ladder can not move past each other , even if they are on different legs , because of the infinite nearest - neighbor repulsion acting across the rungs .
consequently , the hilbert space of the @xmath19-spinless - fermion problem breaks up into many independent sectors , each with a fixed sequence of leg indices .
the @xmath19-spinless - fermion problem in one such sector is therefore an independent problem from that of another @xmath19-spinless - fermion sector . noting that the closest approach between two particles on the same leg is @xmath252 , whereas that between two particles on different legs is @xmath320
, we invoke the same `` particle - in - a - box '' argument used for the paired limit in sec .
[ sect : boundpairsdof ] to find the ground state for @xmath19 spinless fermion on a ladder of even length @xmath18 to be in a _ staggered _ sector , where successive particles are on different legs .
there are two such sectors in a ladder with open boundary conditions , which we call sector 1 when the first fermion ( from the left ) is on leg 1 , or sector 2 when it is on leg 2 .
evidently this is a twofold symmetry breaking .
( the broken symmetry is that of reflecting the configuration about the ladder axis , which is a valid symmetry within the staggered sector . )
this state has a form of long range order , in that the flavor alternates ; however , that can not be represented by any local order parameter , but only by a `` string '' order parameter @xcite .
let us write @xmath321 and @xmath322 for the ground states in sectors 1 and 2 , respectively .
a staggered configuration of @xmath19 ladder spinless fermions in sector 1 can be mapped to a chain of @xmath19 noninteracting spinless fermions using the _ staggered map _ @xmath323 using the same formula , with an exchange of leg index @xmath324 , configurations in sector 2 are similarly mapped . if @xmath325 , the sectors become connected with a tiny tunnel amplitude , in a finite system ; in this case only the symmetry - restored states @xmath326 are actual eigenstates . ] because the staggered map maps a ladder with density @xmath227 onto a ordinary chain with density @xmath327 , we want a ladder observable @xmath328 and its corresponding chain observable @xmath329 to be such that @xmath330 this is analogous to eq . , which we derived when we map from a excluded chain to an ordinary chain .
we use the subscripts ` ladder ' and ` chain ' just this once to distinguish between ladder and chain expectations .
this notation is cumbersome , so we will not use it again . whether an expectation is a ladder expectation or
a chain expectation should be clear from the context .
having solved the staggered ground states in terms of the one - dimensional fermi sea , we calculate the fl , cdw , and sc correlations .
there are four fl correlations at range @xmath71 , @xmath331 , @xmath332 , @xmath333 and @xmath334 . from the staggered nature of @xmath326 , we know that @xmath335 in both ground states .
the inter - leg fl correlations vanish , i.e. @xmath336 because annihilating a particle on one leg and creating a particle on the other leg disrupts the stagger configuration . ,
followed by creation of a spinless fermion at site @xmath246 , within a staggered ground - state configuration leads to a staggered ground - state configuration , when there are no intervening particles between rungs @xmath88 and @xmath90 . ]
the intra - leg fl correlation @xmath337 , which is nonzero , receives contributions only from initial and final staggered configurations in which there are no intervening particles between rungs @xmath88 and @xmath92 , for example , those shown in fig .
[ fig : staystagger ] .
this tells us that @xmath338 when we map the ladder model to the chain model .
this correlation is evaluated numerically , and shown in fig .
[ fig : ladderlinfxyweakflr ] , where we see the staggered ground - state fl correlations decaying exponentially with separation @xmath71 .
this asymptotic behaviour can again be understood using a constrained probabilities argument similar to that used in sec .
[ sect : tonkscorr ] , except that instead of a compact cluster , the relevant probability @xmath339 is that of finding a gap at least @xmath71 in length within the one - dimensional fermi - sea ground state . , @xmath226 , as a function of the separation @xmath340 for ladder densities @xmath341 , 0.25 and 0.30 , in the two - leg limit @xmath342 , @xmath317 .
] applying a restricted probability argument similar to the one outlined in sec .
[ sect : explainflcorr ] , we know this probability is simply the zero - particle weight @xmath343 of the density matrix of a cluster of @xmath71 contiguous sites in the chain of noninteracting spinless fermions . for @xmath344 , the cluster green - function matrix @xmath345 has approximately @xmath346 eigenvalues which are almost zero , and @xmath347 eigenvalues which are almost one .
the determinant of @xmath348 is thus essentially determined by the approximately @xmath347 eigenvalues which are almost one .
using this fact , we calculate the asymptotic form of @xmath349 to be @xmath350 where @xmath307 is the universaling scaling function identified in ref . .
eq . explains the observed exponential decay of @xmath337 in fig .
[ fig : ladderlinfxyweakflr ] .
we note further that as @xmath314 ( or equivalently , @xmath313 ) , the fl correlations decay fastest exponentially , whereas as @xmath351 ( equivalent to @xmath309 ) , the exponential decay is the slowest .
we expect these behaviours physically , because it is more likely to find a long empty cluster when the density is low , and less likely to find a long empty cluster when the ladder is closed to half - filled .
next , we calculate the cdw correlations , for which the four simplest at separation @xmath71 are , @xmath352 because of the staggered nature of the ground states , configurations making nonzero contributions to @xmath353 are those which map to noninteracting spinless fermion configurations in which the sites @xmath88 and @xmath92 are occupied , with an odd number of intervening particles between them .
similarly , configurations making nonzero contributions to @xmath354 , are those which map to noninteracting spinless fermions in which the sites @xmath88 and @xmath92 are occupied , with an even number of intervening particles between them . defining the density operators @xmath355 which are symmetric and antisymmetric with respect to reflect along the ladder axis , we find that @xmath356 and @xmath357 which we call the _
cdw@xmath358 correlation_. this is identical to the cdw correlation of the one - dimensional fermi sea , which we know decays as an oscillatory power law @xmath359 there is also the _ cdw@xmath360 correlation _ , @xmath361 associated with @xmath362 .
this is identical to the subtracted cdw@xmath360 correlation , since @xmath363 . evaluating this expectation numerically
, we find that at all ladder densities @xmath227 , @xmath364 oscillates about a zero average with wave vector @xmath262 , and a decaying amplitude .
a preliminary unrestricted nonlinear curve fitting to the asymptotic form @xmath365 where @xmath273 is the leading asymptotic behaviour , and @xmath366 is a correction term , suggests that the leading correlation exponent may actually be universal , taking on the value @xmath367 .
further nonlinear curve fitting , restricting @xmath367 , tells us that only the parameters @xmath198 and @xmath200 of the leading asymptotic term can be reliably determined .
these are shown in fig .
[ figure : weakxycdwmfits ] . from this restricted curve fit
, it appears that the phase shift might also be universal , taking on value @xmath368 .
( top ) and fitted phase shift @xmath200 ( bottom ) of the leading oscillatory power - law decay , as functions of the ladder density @xmath227 , for the cdw@xmath360 correlation in the staggered ground state of the ladder model , in the two - leg limit @xmath342 , @xmath317 . ] the simplest sc correlations at separation @xmath71 are @xmath369 correlations of the type @xmath370 receive nonzero contributions from configurations containing an even number of intervening particles between rungs @xmath371 and @xmath92 , whereas correlations of the type @xmath372 receive nonzero contributions from configurations containing an odd number of intervening particles between rungs @xmath371 and @xmath92 . defining the paired operators @xmath373 which are symmetric and antisymmetric with respect to reflection about the ladder axis , we find that @xmath374 and that the sc@xmath358 correlation @xmath375 at large separations
. the sc@xmath360 correlation @xmath376 must be evaluated numerically .
we find that , just like @xmath377 , @xmath378 oscillates about a zero average with wave vector @xmath262 , and a rapidly decaying amplitude . to improve the quality of the nonlinear curve fitting , we fit @xmath379 to the asymptotic form @xmath380 where @xmath381 is the leading asymptotic behavior , while @xmath382 is a correction term .
a preliminary unrestricted fit suggests that the leading correlation exponent is universal , and takes on value @xmath383 .
further restricted nonlinear curve fitting tells us that only the parameters @xmath220 and @xmath223 can be reliably determined .
these are shown in fig .
[ figure : weakxyscmfits ] , where we see that the amplitude @xmath220 exhibits symmetry about quarter filling , which is a kind of particle - hole symmetry , and that the phase shift @xmath384 $ ] is non - universal , but depends linearly on the density @xmath227 .
( top ) and fitted phase shift @xmath223 ( bottom ) of the leading oscillatory power - law decay , as functions of the ladder density @xmath227 , for the sc@xmath360 correlation in the staggered ground state of the ladder model , in the two - leg limit @xmath342 , @xmath317 . ] in this subsection , we look at the rung - fermion limit @xmath237 , @xmath317 .
we argue in section [ sect : stronginterleggs ] that in this limit , each spinless fermion spends most of its time hopping back and forth along the rung it is on , and only very rarely hops along the legs to an adjacent rung .
therefore , each spinless fermion will be in a quantum state very close to the symmetric eigenstate of one rung , and we can think of the ladder of spinless fermions with density @xmath227 in this limit as essentially an excluded chain of fermions with density @xmath385 . for @xmath386 ,
the ground state of this excluded chain of rung - fermions has been solved in sec .
[ sect : nnetonni ] .
the fl , cdw , and sc correlations have also been calculated in sec .
[ sect : correlationsnnebosons ] , so we will not repeat them here . at @xmath387 ,
the ground state is a ` dynamic solid ' phase , in which rung - fermions occupy either all the even rungs , or all the odd rungs , and can not hop along the legs to adjacent rungs because of the infinite nearest - neighbor repulsion between them . for @xmath388 , we describe in sec .
[ sect : phasesep ] how the system will phase separate into a high - density inert solid phase , in which spinless fermions can not hop at all , and the lower - density ` dynamic solid ' phase . in this phase separation regime
, the fl , cdw , and sc correlations can not be calculated . in the limit of @xmath389 ,
a spinless fermion spends most of its time hopping back and forth along a rung , and only very rarely hops along the leg to an adjacent rung , where it will spend a lot of time hopping back and forth , before hopping along the leg again .
because of this long dwell time on a rung , the spinless fermion is very nearly in the rung ground state @xmath390 let us call a spinless fermion in the rung ground state a _ rung fermion _ in short .
rung - fermions inherit the infinite nearest - neighbor repulsion of the bare spinless fermions , and therefore two rung - fermions in adjacent rungs experience infinite nearest - neighbor repulsion as well . with this insight
, we find that the full many - body problem of spinless fermions with infinite nearest - neighbor repulsion on the two - legged ladder with density @xmath227 reduces to the problem of an excluded chain with density @xmath391 of spinless rung - fermions .
the latter problem was solved in sec .
[ sect : maptech ] and sec .
[ sect : correlationsnnebosons ] for excluded chain densities @xmath392 . in the special case of quarter - filling on the ladder , @xmath387 ,
spinless fermions occupy alternate rungs .
these are free to hop along the rungs that they reside on , but can not hop along the legs , for non - vanishing values of @xmath393 .
even virtual processes in which a spinless fermion on rung @xmath88 hops along the leg to an adjacent rung and back are essentially forbidden by the infinite nearest - neighbor repulsion , because such virtual processes , which has a time scale of @xmath394 , would not be complete when the spinless fermion on the next - nearest - neighbor rung hops across the rung , which occurs on a time scale of @xmath395 .
virtual processes such as these only become energetically feasible when the two time scales become comparable , i.e. when @xmath396 . therefore , over a wide range of anisotropies @xmath393 , the spinless fermions in the quarter - filled ladder with @xmath317 can hop back and forth along the rungs they are on , but can not hop to the neighboring rungs .
this gives rise to a symmetry breaking , where the spinless fermions are either all on the even rungs , or they are all on the odd rungs . because translational symmetry along the ladder axis is broken in the quarter - filled ladder ground states , we think of these as ` dynamic solids ' , since the constituent spinless fermions are constantly hopping back and forth along the rungs . in this limit ,
the only non - vanishing correlation is the rung - fermion cdw correlation @xmath397 i.e. there is true long - range order in the quarter - filled ladder ground state in the limit of @xmath237 , @xmath317 . in this rung - fermion limit @xmath237 , @xmath317 , the system phase separates for ladder densities @xmath398 . as shown in fig .
[ fig : stronginterlegphaseseparation ] , when the ladder is above quarter - filling , some of the spinless fermions will go into a high - density inert solid phase with density @xmath399 , where spinless fermions are arranged in a staggered array , and therefore can not hop at all .
these spinless fermions contribute nothing to the ground - state energy .
if @xmath400 is comparable to @xmath401 , the rest of the spinless fermions will go into a fluid phase , whose density is @xmath386 .
these spinless fermions are free to hop back and forth on the rungs they are on , and occasionally to the neighboring rungs , when permitted by nearest - neighbor exclusion .
these contribute a density - dependent total kinetic energy to the ground - state energy .
the ground - state composition depends on whether the kinetic energy gained per particle , by removing a spinless fermion from the solid phase and adding it to the fluid phase , outweighs the decrease in kinetic energy per particle that results from the fluid becoming more congested . , and a low - density fluid phase ( mobile spinless fermions shown with arrows ) with @xmath387 , in the rung - fermion limit @xmath237 , @xmath317 . ]
when @xmath401 becomes large compared to @xmath400 , which is the limit we are interested in , it becomes energetically favorable , always , to remove one spinless fermion from the inert solid phase , and add it to the fluid phase , if its density is @xmath386 .
this is because the kinetic energy penalty to make the fluid becoming more congested , which is of @xmath402 , is more than compensated for by the kinetic energy gain of @xmath401 for an extra spinless fermion freed to hop back and forth along a rung .
iterating this argument , we find then that , for @xmath237 , and the overall density @xmath403 , the system will phase separate into an inert solid phase with density @xmath404 , and a dynamic solid phase with density @xmath405 . for example , if the overall density is @xmath406 , we will find that @xmath407 of the total number of spinless fermions will be in the inert solid phase , while the other @xmath408 of the total number of spinless fermions will be in the dynamic solid phase .
in this paper , we established a one - to - one correspondence between @xmath19-particle configurations on the excluded chain and @xmath19-particle configurations on the ordinary chain using the right - exclusion map .
we then showed that the hamiltonian matrices of the models given in eq . and eq . are identical , therefore solving for the ground states of the former in terms of those of the latter .
these results were obtained for finite chains subject to open boundary conditions , but continues to hold for infinite chains . based on this one - to - one correspondence between ground states
, we showed that the ground - state expectation @xmath64 of an excluded chain observable @xmath65 can be evaluated using eq . in terms of the ground - state expectation @xmath66 of a carefully chosen corresponding observable @xmath67 on the ordinary chain .
we then developed the method of intervening - particle expansion , to write the ground - state expectation @xmath86 of a product of local excluded chain operators @xmath87 and @xmath89 , first as a sum over excluded chain expectations @xmath409 conditioned on the occupations of the sites between @xmath88 and @xmath92 , and then as a sum over the corresponding ordinary chain expectations @xmath410 . using these analytical results from sec .
[ sect : maptech ] , we calculated the fl , cdw , and sc correlations of the excluded chains of hardcore bosons and spinless fermions in sec .
[ sect : correlationsnnebosons ] . based on nonlinear curve fits of the numerically evaluated correlations , to reasonable asymptotic forms , we find all three types of correlations decaying with separation @xmath71 as power laws , for hardcore bosons as well as for spinless fermions .
more interestingly , we find for both hardcore bosons and spinless fermions a universal exponent @xmath411 for the oscillatory power - law decay of the sc correlation , but a non - universal , density - dependent , exponent @xmath412 for the oscillatory power - law decay of the cdw correlation . also , the leading asymptotic behaviour for the hardcore boson fl correlation was found to a non - oscillating power - law decay with universal exponent @xmath128 , while that for the spinless fermion fl correlation was found to be oscillations in a power - law envelope , with a non - universal exponent that approaches @xmath157 as @xmath149 , and @xmath187 as @xmath151 .
we then analyzed our spinless - fermion ladder model , eq .
, in sec .
[ sect : laddermodels ] .
this ladder model can be solved exactly in three limiting cases : ( i ) the paired limit @xmath413 ; ( ii ) the two - leg limit @xmath414 ; and ( iii ) the rung - fermion limit @xmath415 . in the paired limit , which we solved in sec .
[ sect : infinitelystrongcorrelatedhops ] , spinless fermions form correlated - hopping bound pairs , and so the ladder model can be mapped to the excluded chain of hardcore bosons .
the ground state of this latter model was solved exactly in sec .
[ sect : maptech ] and its ground - state correlations calculated in sec .
[ sect : correlationsnnebosons ] . by reinterpreting the excluded chain correlations in ladder terms
, we realized that ladder sc correlations dominates at large distances over ladder cdw correlations , both of which decay as power laws with separation , with leading exponents @xmath416 and @xmath417 respectively , @xmath227 being the ladder density . we also showed , using a restricted probabilities argument , that ladder fl correlations decay exponentially with separation , with a density - dependent correlation length . next , in the two - leg limit , which we solved in sec .
[ sect : independentlegs ] , we argued based on a `` particle - in - a - box '' picture that successive spinless fermions in the two - fold degenerate _
staggered ground states _ occupy different legs of the ladder .
we write these ground states exactly in terms of the one - dimensional fermi sea in sec .
[ sect : nointerleggs ] , before calculating correlations in sec .
[ sect : weakxy ] .
we found , using a different restricted probabilities argument , that fl correlations decay exponentially with separation , with a density - dependent correlation length .
cdw and sc correlations symmetric ( antisymmetric ) with respect to a reflection about the ladder axis decay as power laws , with universal leading exponents @xmath418 and @xmath419 respectively . finally , in the rung - fermion limit , we mapped the ladder model to an excluded chain of spinless fermions in sec .
[ sect : independentrungs ] . since we have already solved this latter model in sec .
[ sect : maptech ] and calculated its ground - state correlations in sec .
[ sect : correlationsnnebosons ] below half - filling ( which corresponds to quarter - filling on the ladder ) , we discussed the phase separations that occurs on ladders with greater than quarter filling in sec . [
sect : stronginterleggs ] .
correlation exponents obtained for the three limiting cases of our ladder model eq . , as well as those for the excluded chains of hardcore bosons and spinless fermions , are summarized in table [ table : summaryofexponents ] .
.a summary of the leading correlation exponents and wave vectors of various correlation functions that decay as power laws in the ( i ) paired limit @xmath231 ; ( ii ) two - leg limit @xmath233 , @xmath317 ; and ( iii ) rung - fermion limit @xmath237 , @xmath317 .
the wave vector @xmath135 of the leading terms in the correlation functions are reported in terms of @xmath420 , where @xmath421 is the excluded chain density .
the suffixes @xmath266 and @xmath422 indicate further symmetries possible in the ladder model . [ cols="^,^,^,^ " , ] in this study , we find the emergence of surprising universal correlation exponents . in the luttinger liquid paradigm ,
all correlation exponents can be written in terms of the exponents @xcite @xmath423 appearing in the quantum - mechanical propagator , also called the ( equal - time ) two - point function @xmath424 } \cos k_f r \\ & = a_1\ , r^{-\frac{1}{4}\left[(k_{\rho } + k_{\rho}^{-1 } ) + ( k_{\sigma } + k_{\sigma}^{-1})\right ] } \cos k_f r. \end{aligned}\ ] ] the parameters @xmath425 and @xmath426 depend generically on the filling fraction and the interaction strength , and thus all correlation exponents are non - universal .
in particular , various theoretical approaches ( see review by slyom @xcite ) tell us that the charge density waves ( cdw ) , spin density wave ( sdw ) , singlet superconductivity ( ssc ) and triplet superconductivity ( tsc ) correlations decay as power laws [ eqn : expkrhoksigma ] @xmath427 in a tomonaga - luttinger liquid .
when the chain of interacting spinfull fermions is spin - rotation invariant ( for example , in the absence of an external magnetic field ) , the spin stiffness constant must be @xmath428 , and the ground - state properties become completely determined by the single nontrivial luttinger parameter @xmath425 .
the spinfull power laws thus become @xmath429 for spinless fermions , there is only one independent stiffness constant @xmath430 @xcite , so that the spinfull power laws which have proper spinless analogs are @xmath431 in the luttinger liquid paradigm , universal correlation exponents only arise in the special case of the fermi liquid , where we have @xmath432 .
consequently , the two - point function decays as @xmath433 , while the cdw and sc correlations both decay as @xmath434 .
however , the universal correlations exponents that we find in our exact solutions are different from these . furthermore , the correlation exponents @xmath28
, @xmath29 , and @xmath435 of the fl , cdw , and sc correlations ought to obey definite relations in a luttinger liquid , because they can all be written in terms of a single luttinger parameter @xmath436 . again , the universal and non - universal correlation exponents we find in our exact solutions do not obey these relations .
these observations bring us to the paper by efetov and larkin , who first calculated the universal fl correlation exponent for an ordinary chain of hardcore bosons to be @xmath437 @xcite .
if we accept for the moment that the luttinger paradigm is correct , and that universal correlation exponents can only be found at the fermi liquid fixed point , then we are led to the conclusion that @xmath211 must be a correlation exponent of the fermi liquid .
clearly , this exponent does not belong to the fermi liquid fl correlation ( which should be @xmath240 , so what correlation does it belong to ? in the seminal paper by jordan and wigner , the ordinary chain of hardcore bosons is mapped to the ordinary chain of spinless fermions using the jordan - wigner transformation ( see appendix [ sect : jwtrans ] ) . in this transformation ,
the hardcore boson point operators @xmath11 and @xmath438 are each mapped to spinless fermion _ string operators _ @xmath439 and @xmath440 respectively .
the fl correlation @xmath186 between two hardcore boson point operators thus become the expectation @xmath441 of the string operator @xmath442 . which efetov and larkin found to decay with separation @xmath71 as @xmath443 . string correlations such as this
have never been systematically studied .
one reason for this lack of interest is that typical string correlations , which receive contributions only from restricted classes of configurations , decay exponentially with @xmath71 , as we have seen for the fl correlation in the paired limit ( sec .
[ sect : tonkscorr ] ) and the two - leg limit ( sec .
[ sect : weakxy ] ) . however , there appear to many string correlations that decay with separation @xmath71 as power laws .
these power law decays are associated with ( quasi-)long - range order that we have not been creative enough to imagine . in sec .
[ sect : nointerleggs ] , we found in the two - leg limit that the staggered ground state has long - range order , in that if we know the @xmath91th particle is on leg @xmath444 , then we know for certain that the @xmath445th particle is on leg @xmath444 , and the @xmath446th particle is on leg @xmath447 , even as @xmath448 , and even though we have no idea where these particles are on the ladder .
this long - range order is not the usual kind of long - range order , which can be written in terms of the correlation between local order parameters , but is a long - range _ string _ order .
the map from the ordinary chain ground state to the staggered ladder ground state , which is the inverse of the one constructed in sec .
[ sect : nointerleggs ] , implicitly involves string operators , in that if we take the @xmath91th particle in the ordinary ground state configuration , we will know whether to map it to a particle on leg @xmath444 or leg @xmath447 , _ after _ we know which legs the preceding particles are on . also , while it is deceptively simple to describe what the string operator in this inverse map does , which is to project out any combination of more than or equal to two consecutive particles on the same leg of the ladder , we know of no compact way to write down the string operator , even in this simple limit , unlike for the case of the jordan - wigner string .
what we do know , drawing parallels from the jordan - wigner map from hardcore bosons on ordinary chains to noninteracting spinless fermions , is that a string map from one model to another will map some products of local operators to string operators , for example , the hardcore boson @xmath449 to the spinless fermion @xmath450 , and other products of local operators to products of local operators , for example , the hardcore boson @xmath451 to the spinless fermion @xmath452 .
having understood this , we realized that the cdw@xmath358 and sc@xmath358 correlations in the staggered ground state get mapped to the the correlation of local operators , because the string operators involved in the map multiply and cancel each other . on the other hand , when we map the cdw@xmath360 and sc@xmath360 staggered ground - state correlations to correlations of a chain of noninteracting spinless fermions , the string operators involved in the map do not cancel each other , and thus the resulting ordinary chain spinless - fermion correlations are string correlations .
we also realized that these string correlations are operationally defined by the intervening - particle expansions we used to compute them .
since all the exact solutions we have obtained in this paper can ultimately be mapped to the one - dimensional fermi sea , we conjecture that all correlation exponents are universal .
we claim that : ( i ) all exponents that are explicitly universal are simple rational polynomials of the single universal spinless fermi liquid parameter @xmath453 ; and ( ii ) non - universal exponents are the result of ( under)fitting linear combinations of universal power laws to a single power law . for example , in the two - leg limit , the leading universal exponent @xmath367 of the cdw@xmath360 correlation in the staggered ground state can be shown using a bosonization calculation of the string correlation it is mapped to , to follow automatically from the universal fermi liquid parameter @xmath454 @xcite .
in this same limiting case , the leading universal correlation exponent @xmath455 of the sc@xmath360 correlation , which gets mapped to a significantly more complicated string correlation , can conceivably be written as the combination @xmath456 of the universal fermi liquid parameter @xmath454 , even though the bosonized form of this string correlation is not known . for the excluded chain of hardcore bosons or spinless fermions , nonlinear curve fitting of the sc correlation to the sum of one leading power - law decay and one
subleading power - law decay leads to weakly non - universal correlation exponents for both power laws , whereas a complicated sum of power - law decays , eq . , produces a better fit visually .
we believe good fits can also be obtained , using similar complicated sum of power - law decays , for those numerical correlations which we found to have strongly non - universal correlation exponents .
finally , we asked ourselves whether all these string correlations that we have predicted will decay with separation @xmath71 slower than the two - point function @xmath457 can be measured in a chain of noninteracting spinless fermions . since these string operators are nonlocal observables , they do not in general couple to local measurements , so direct experimental measurement would be challenging , if not downright impossible
however , we would like to suggest the following possibility : for a given string correlation of the one - dimensional fermi sea , cook up in the laboratory an experimental system in which the corresponding correlation is a point correlation .
if the ground - state of the experimental system can be mapped to the one - dimensional fermi sea , we expect a measurement of the point correlation exponent in the experimental system to be an indirect measurement of the string correlation exponent in the fermi sea .
on a one - dimensional chain , hardcore bosons can not move past each other , as one boson must first hop on top of the other a move explicitly forbidden by the hardcore condition for this to happen . for a different reason ( the pauli exclusion principle ) , but to the same effect , noninteracting spinless fermions on a one - dimensional chain can not exchange positions . therefore , in one dimension , the hardcore - boson and noninteracting - spinless - fermion hamiltonians are also identical in structure , and thus the ground state of a chain of hardcore bosons is related to the fermi - sea ground state of a chain of noninteracting spinless fermions in a simple way .
a translation machinery exists to map back and forth between these two ground states .
this is the jordan - wigner transformation @xcite @xmath458 which maps hardcore bosons to spinless fermions , where the product @xmath459 is called the _ jordan - wigner string_. in section [ sect : infinitelystrongcorrelatedhops ] we saw how pairs of spinless fermions bound by correlated hops in the limit @xmath460 can be mapped to hardcore bosons with infinite nearest - neighbor repulsion , and then to hardcore bosons using the right - exclusion map described in section [ sect : nnetonni ] , and then finally to noninteracting spinless fermions . in section [
sect : nnegsexp ] , we saw how excluded hardcore - boson expectations are related to appropriately chosen ordinary hardcore - boson expectations .
this relation between excluded hardcore - boson expectations and ordinary hardcore - boson expectations will typically involve the intervening - particle expansion eq . .
as such , we will encounter hardcore - boson expectations of the form @xmath461 a lot , where there are @xmath91 hardcore - boson occupation number operators @xmath462 , at sites @xmath463 , and @xmath464 hardcore - boson operators @xmath465 , at sites @xmath466 , between the hardcore boson operators @xmath467 at site @xmath276 and @xmath468 at site @xmath469 . to evaluate these expectations
, we first invoke the jordan - wigner transformation to replace all the hardcore - boson occupation number operators @xmath16 by spinless - fermion occupation number operators @xmath17 in .
then , to account for the two unpaired hardcore - boson operators at the ends of the hardcore - boson operator product , we write as the spinless - fermion expectation @xmath470 noting that all jordan - wigner string operators @xmath471 commutes with @xmath472 and @xmath473 , for @xmath474 and @xmath475 , and that @xmath476 we can bring the jordan - wigner string @xmath477 associated with the annihilation operator @xmath478 through the intervening spinless - fermion operators to obtain @xmath479 then , using the fact that @xmath480 we can finally write the hardcore - boson expectation @xmath481 as a spinless - fermion expectation , where @xmath91 is the number of occupied sites between @xmath276 and @xmath469 . the suffixes ` empty ' or ` filled ' in the products in refer to the sites between @xmath276 and @xmath482 which are empty or filled respectively .
this research is supported by nsf grant dmr-0240953 , and made use of the computing facility of the cornell center for materials research ( ccmr ) with support from the national science foundation materials research science and engineering centers ( mrsec ) program ( dmr-0079992 ) .
sac also acknowledge support from the the nanyang technological university startup grant sug 19/07 . | by removing one empty site between two occupied sites , we map the ground states of chains of hardcore bosons and spinless fermions with infinite nearest - neighbor repulsion to ground states of chains of hardcore bosons and spinless fermions without nearest - neighbor repulsion respectively , and ultimately in terms of the one - dimensional fermi sea .
we then introduce the intervening - particle expansion , where we write correlation functions in such ground states as a systematic sum over conditional expectations , each of which can be ultimately mapped to a one - dimensional fermi - sea expectation .
various ground - state correlation functions are calculated for the bosonic and fermionic chains with infinite nearest - neighbor repulsion , as well as for a ladder model of spinless fermions with infinite nearest - neighbor repulsion and correlated hopping in three limiting cases .
we find that the decay of these correlation functions are governed by surprising power - law exponents . |
gamma - ray bursts ( grbs ) are powerful brief transient phenomena of high - energy radiation that appear randomly in the sky .
they were first detected in 1969 by the _
vela _ satellites ( klebesadel et al . 1973 ) and have remained for many years one of the most elusive mysteries in astrophysics . before the launch of the _ bepposax _ and _ rossixte _ satellites in 1996
, they had not been detected in any other wavelength region , and their distance scale remained unknown .
the discovery of x - ray afterglows by both satellites revolutionized the field , because they are able to provide very accurate grb error boxes within a few hours ( circular error boxes with radius up to @xmath4 ) , which enables very rapid follow up observations at longer wavelengths .
eight grbs detected by _
bepposax _ have been detected at optical and infrared wavelengths ; grb 970228 ( guarnieri et al . 1997a , van paradijs et al .
1997 ) , grb 970508 ( bond 1997 , djorgovski et al .
1997 , castro - tirado et al .
1998a ) , grb 971214 ( halpern et al .
1997 , gorosabel et al .
1998 , ramaprakash et al . 1998 ) , grb 980326 ( groot et al .
1998 ) , grb 980329 ( klose 1998 , palazzi et al .
1998a , 1998b , taylor et al .
1998 ) , grb 980519 ( jaunsen et al .
1998 , maury et al . 1998 ) , grb 980613 ( hjorth et al .
1998 , djorgovski et al .
1998a , 1998b ) , and grb 980703 ( frail et al . 1998 , zapatero - osorio et al . 1998 ) .
spectral information taken for grb 970508 showed that the gamma event occurred at a redshift @xmath50.83 , therefore supporting the cosmological origin of grbs ( metzger et al .
recently , kulkarni et al . ( 1998 ) measured the redshift of the host - galaxy of grb 971214 ( @xmath6=3.42 ) , placing it at a distance of @xmath712 billion light years ( assuming the universe to be about 14 billion years old ) .
this large distance implies an energy release at least of @xmath8 erg , which can not be explained by the neutron star merger model , at least in its simplest form ( narayan et al .
1992 ) .
grb 970111 was the first grb observed by _
bepposax _ that was promptly followed up at other wavelengths . in fact , the results presented in this paper represent the fastest follow - up observation performed for grb 970111 in all wavelengths .
it was detected as a three - peak gamma - ray burst on january 11 , 1997 by the wide field camera of the x - ray _ bepposax _ satellite ( costa et al .
the burst was localized at @xmath9 , @xmath10 ( equinox 2000.0 ; error box radius= @xmath11 ) .
soon after that , butler et al .
( 1997 ) reported the presence of two faint x - ray sources in the field of grb 970111 ; 1sax j1528.8 + 1944 and 1sax j1528.8 + 1937 , labeled as `` a '' and `` b '' respectively .
then , hurley et al . ( 1997a ) reduced the error box of grb 970111 by means of the _ ulysses _ and batse data , and found that only the _ bepposax _ source `` a '' lied within it . unfortunately , due to a misalignment of the wide field cameras of _ bepposax _
( int zand et al . 1997 ) , the error box reported by costa et al . (
1997 ) had to be shifted of @xmath12 from its previous position .
the new error box provided by hurley et al .
( 1997b ) , seven times smaller than the former , placed both x - ray sources far away ( @xmath13 ) from this new error box , thus showing both sources to be unrelated to grb 970111 .
1sax j1528.8 + 1944 has been detected by rosat ( voges et al . 1997 ) and it was found to be related to the variable radiosource vla j1528.7 + 1945 ( frail et al .
an optical spectrum revealed two objects that lied within @xmath14 from vla j1528.7 + 1945 , showing redshifts @xmath150.636 and @xmath150.458 , respectively ( kulkarni et al .
frontera et al .
( 1997 ) provided more precise coordinates for this x - ray source , however no object showed remarkable variations in the optical , neither in the field of grb 970111 nor in the error boxes of the two x - ray sources ( guarnieri et al .
radio observations at 840 mhz , 1.4 ghz and 1.5 ghz carried out between 26.4 hours and 120 days did not reveal any steady source in the intersection between the wfc @xmath16 error box with a new improved ipn annulus ( galama et al .
1997a , galama et al .
one month after the event , the error box was scanned by the bima array at 3.5 mm , without detecting any fading source at millimeter wavelengths ( smith et al .
very recently the wfc team improved again the error box to an irregular circle of @xmath17 radius , still consistent with the previous one ( heise et al .
furthermore a recent reanalysis of the mecs data revealed a previously unknown x - ray source , 1sax j1528.1 + 1937 , which is almost entirely contained within the @xmath18 wfc error box ( feroci et al .
however this source is only marginally consistent with the last ipn annulus reported by galama et al .
( 1997b ) .
the above mentioned new @xmath19-ray and x - ray positions of the grb 970111 encouraged us to reanalyze the results already presented elsewhere ( castro - tirado et al .
furthermore new observations were carried out , which enabled us to observe the grb 970111 up to 6 months after the gamma - ray event .
section @xmath202 will briefly describe the data reduction and calibration techniques , while section @xmath203 will present and discuss the results .
finally , section @xmath204 will draw the conclusions .
lcccccccr date & & telescope + of 1997 & u & b & v & r & i & z & free & + jan .
12&- & 1.5 & - & 0.6&- & - & - & 2.2 caha + jan .
17&- & 2.7 & - & 0.9&- & - & - & 1.5 loiano + feb .
10&- & 1.5 & - & 0.6&- & - & - & 2.2 caha + feb .
11&- & - & - & 0.6&- & - & - & 2.2 caha + mar .
5&- & - & - & 1.8&- & - & - & 1.5 loiano + mar .
10&0.6 & 0.6 & - & 0.6&0.6&0.6&- & 1.54 danish + mar .
12&- & 1.8 & - & - & - & - & - & 1.54 danish + mar .
14&- & - & - & - & - & - & 7.2&1.54 danish + mar .
15&- & 3.6 & - & - & - & - & - & 1.54 danish + jun .
26&- & - & 4.8&- & - & - & - & 1.54 danish + jun .
27&- & - & - & 3.2&- & - & - & 1.54 danish + jul . 1 & - & 2.7 & - & - & - & - & - & 1.54 danish + jul
. 3 & - & - & - & 1.8&1.6 & - & - & 1.54 danish + jul .
4 & - & - & - & - & 2.9 & - & - & 1.54 danish + + [ table2 ] lcccccccr date & & telescope + of 1997 & u & b & v & r & i & z & free & + jan .
12 & - & 1.5 & - & 0.6 & - & - & -&2.2 caha + jan .
14 & - & 1.2 & - & 1.2 & - & - & -&1.5 loiano + jan .
17 & - & 2.7 & - & 0.9 & - & - & -&1.5 loiano + jan .
31 & - & 3.0 & 1.2 & 0.9 & - & - & -&1.5 loiano + feb .
10 & - & 1.5 & - & 0.6 & - & - & -&2.2 caha + feb .
11 & - & - & - & 0.6 & - & - & -&2.2 caha + feb .
14 & - & - & 1.8 & - & - & - & -&1.5 loiano + feb .
17 & - & - & 1.62 & 1.8 & - & - & -&1.5 loiano + feb .
18 & - & 4.5 & 2.7 & 1.8 & - & - & -&1.5 loiano + mar . 5 & - & - & - & 1.8 & - & - & -&1.5 loiano + mar . 13 & - & 3.6 & - & 1.8 & - & - & -&1.5 loiano + + lcccccccr date & & telescope + of 1997 & u & b & v & r & i & z & free & + jan .
12 & - & 1.5 & - & 0.6 & - & - & -&2.2 caha + jan .
14 & - & 1.2 & - & 1.2 & - & - & -&1.5 loiano + feb .
10 & - & 1.5 & - & 0.6 & - & - & -&2.2 caha + feb .
11 & - & - & - & 0.6 & - & - & -&2.2 caha + + .[table4 ] filters used in the observations carried out at la silla .
[ cols="^,^,^,^,^ " , ]
we compared the images taken at caha on january 12 with the ones obtained on february 10 and 11 .
variable sources have neither been found within the x - ray error boxes of 1sax j1528.8 + 1944 and 1sax j1528.8 + 1937 , nor in the whole 16@xmath21 diameter field .
[ dif ] shows the magnitude differences in the b and r filters for objects inside the error boxes of grb 970111 when the frames taken in caha 19 hours after the burst and @xmath7 1 month later are compared .
if a suspected variable object is found , it is compared to the images taken at loiano and la silla .
as it is shown there is no object varying by more than @xmath22 neither in the r nor in the b filter observations , as reported elsewhere ( castro - tirado et al .
1997 ) . furthermore
, none of the objects within the different grb 970111 error box has changed in brightness in the images taken at loiano and la silla .
the fields of both x - ray sources 1sax j1528.8 + 1944 and 1sax j1528.8 + 1937 were observed at loiano and caha using b and r filters .
no variations in brightness were found for any of the objects within the two x - ray sources error boxes .
[ figure2 ] shows the field of both sources in the b - band .
b and r magnitudes for the objects inside the x - ray error boxes of 1sax j1528.8 + 1944 and 1sax j1528.8 + 1937 are shown in tables [ table6 ] and [ table7 ] . for objects with magnitudes fainter than 23.5 in b and 22.5 in r
, the errors introduced by the photometry do not allow to get reliable values of the b - r colour index .
thus , in tables [ table6 ] and [ table7 ] only objects brighter than the above mentioned magnitudes are shown .
object # 11 of table [ table6 ] is coincident with the radio source vla 1528.7 + 1945 and has been proposed as the optical counterpart of 1sax j1528.8 + 1944 ( frontera et al .
1997 , guarnieri et al .
1997b , masetti et al .
this object consists of a blend of at least 2 galaxies with different redshifts ( kulkarni et al .
1997 ) and shows a fuzzy and complicated structure .
the b and r magnitudes reported for object # 11 comprises the flux of all its components .
this fact explains the discrepancy of the b and r magnitudes with those reported by kulkarni et al .
we also note object # 7 inside error box ( table [ table6 ] ) which shows b - r=-0.1 .
however the large error in the colour index makes this candidate less likely to be related to 1sax j1528.8 + 1944 .
( -87,-195) ( -203,-25)mg ii ( -60,-25)h@xmath23 ( -46,-25)[o iii ] we would like to remark that one of the objects in the 1sax j1528.8 + 1937 field ( object # 4 , table [ table7 ] ) is the bluest one with b @xmath0 21 found in the @xmath24 radius caha image .
two optical spectra of this object were taken with cafos at the 2.2-m telescope of caha using the b200 grism ( 3500 - 6500 ) and the r200 grism ( 6500 - 10000 ) .
the resolution is about 12 , given by the @xmath25 slit .
exposure time was 600 sec in b200 and 1200 sec in r200 .
[ spectrum ] shows the spectrum of object # 4 , revealing it as a typical seyfert-1 galaxy with broad emission lines of mg ii ( 2795.5 , 2802.7 ) and h@xmath23 ( 4861 ) , and the narrow emission lines of [ o iii ] ( 4959 , 5007 ) observed at wavelengths of 8216 and 8298 .
the broad mg ii line is observed at a central wavelength of 4636 , which is very consistent with the redshift determined from the narrow lines .
the fwhm of the mg ii and h@xmath23 lines imply internal velocities of @xmath26 km s@xmath27 and @xmath28 km @xmath29 , respectively .
the spectrum was flux calibrated using instrumental responses determined in a photometric night several days later . taking into account an interstellar absorption of @xmath70.2 mag at @xmath7750 nm derived from the reddening ( see @xmath203.3 ) and using the flux level taken from the spectrum , we determine the blue restframe luminosity of this agn to be between @xmath30 -21.6 and @xmath31 -22.5 depending on cosmology ( @xmath32 , @xmath33 km s@xmath27 mpc@xmath27 and @xmath34 km s@xmath27 mpc@xmath27 , respectively ) .
cumulative surface densities of quasars / seyfert-1 have been determined observationally by several authors ( hartwick and schade 1990 , hawkins and veron 1995 ) .
these authors report the number of quasars and seyfert-1 galaxies with b@xmath021 and @xmath35 2.2 to be about 30 per square degree , rather independent on the selection criteria employed .
therefore , in our 1@xmath21 radius x - ray error box the _ a priori _ probability for finding a quasar or seyfert-1 galaxy with b@xmath36 and @xmath37 2.2 is roughly 2.7 % .
since object # 4 has b=20.6 , this probability represents an upper limit . for this reason , we propose object # 4 as the optical counterpart of the x - ray source .
astrometry of object # 4 yields : @xmath38 @xmath39 ( equinox j2000 ) .
in order to detect any object with peculiar colours , colour - colour diagrams were constructed for all objects inside the wfc error box .
we consider the error box reported by int zand et al .
( 1997 ) , in order to assure that the possible error boxes are included . for 157 objects inside the above mentioned error box
we were able to measure the b , v , r and i magnitudes , that enabled us to construct a b - r vs v - i diagram ( see fig . [
figure6]a ) .
many objects ( @xmath7300 ) were only detected in the r and b bands . in order to distinguish objects with colours
not expected in stars and normal galaxies , we constructed a colour - colour diagram based on the unreddened colours of 5138 stars of the solar neighbourhood from the catalogue compiled by lanz ( 1986 ) , which we have represented by a dotted band in the colour - colour diagram .
this band is composed by main sequence , giants and super - giants stars .
all the objects with error bars not overlapping with the band are candidates to high - redshift galaxies , quasars , active galactic nuclei or whatever object with an optical emission different from that of normal stars in our galaxy .
we have also overplotted the colours for grb 970508 derived from the bvr@xmath40i@xmath40 photometry performed from zharikov et al .
the r filter photometric calibration carried out from schaefer et al .
( 1997 ) and the offsets between johnson and kron - cousins filters ( frei and gunn 1994 ) were taken into account .
as it is shown , the colours of grb 970508 are clearly away ( at least 2@xmath41 ) from the band that represents the main sequence stars . in spite of the larger errors derived from the observations in the u filter in comparison to the other filters
, we constructed the u - b vs b - v diagram in order to distinguish objects with ultraviolet excesses , but u , b , v , r and i magnitudes are available only for 31 objects ( see fig .
[ figure6]b ) .
the uncertainties for the colour indices were calculated from : @xmath42 @xmath43 @xmath44 @xmath45 the colour excess e(b - v ) of the objects is related to the hydrogen column density following bohlin et al .
( 1978 ) : e(b - v)@xmath46 n(hi ) @xmath47 + the column density along the line of sight of the field @xmath48 ) is n(hi)@xmath49@xmath47 , which implies a colour excess e(b - v)@xmath7 0.1 .
once e(b - v ) is known the colour excess for the others filters can be obtained following cardelli et al .
( 1989 ) , resulting ; e(u - b)@xmath7 0.08 , e(b - r)@xmath70.15 , e(v - i)@xmath70.13 .
thus the reddening is comparable to the errors on the photometry .
furthermore , taking into account that the distance of the sources are unknown , we will consider the colour - colour diagrams as the intrinsic colour - colour diagrams for the objects in the field .
[ figure6 ] shows both diagrams . as it is shown in the colour - colour diagrams , there are no sources that deviate over more than a 2@xmath41 level from the colours expected for the usual components of the galaxies ( main - sequence , giants or supergiants stars ) .
among the 31 objects inside the wfc error box for which four colours are available , there are no sources bluer than u - b@xmath0 - 0.70 , b - [email protected] , b - v @xmath00.45 , v - r @xmath00.32 and r - i @xmath00.17 .
in order to determine the identity of the objects in the wfc error box , we used a supervised classification method that was originally developed for the calar alto deep imaging survey ( cadis ) and is used for object classification and accurate redshift estimation of galaxies and quasars there ( wolf et al .
this method calculates likelihoods for each individual object to resemble either a star , a galaxy or a quasar , based on a comparison of its observed colours with libraries of template colours .
these colour libraries have been calculated from spectral libraries by synthetic photometry based on the ubvri filterset and the system responses . as an input , we used the gunn & stryker ( 1983 ) catalogue of stellar spectra , the galaxy template spectra of kinney et al .
( 1996 ) and a quasar model library derived from the quasar template spectrum of francis et al .
( 1991 ) , which is detailed in wolf et al .
( 1998b ) . since the galaxy and quasar libraries are parameterized in redshift , it is also possible to estimate the redshift for these two classes of objects .
the performance of this supervised multicolour classification strongly depends on the colour information available and , of course , on the choice of good libraries . according to wolf et al .
( 1998a ) , this method detects broad - line agns very reliably , and estimates redshifts for galaxies and quasars , which are mostly accurate to within 5% , if seven or more filters are available with photometric accuracies better than 3% , and the above mentioned libraries are used .
given only four or five broadband colours , redshift estimates are bound to be less reliable , but classification will still distinguish well between stars and galaxies . as it was explained in @xmath20 [ colour ] , for only 31 objects inside the wfc error box ubvri photometry was possible , while for 157 objects only bvri magnitudes were measured . in this case
, the low number of bands available and the higher photometric errors of the faint objects left many objects unclassifiable .
for the set of 31 sources , the proportion between stars , galaxies and unclassifiable objects were : 74% stars , 18% galaxies and 18% unclassifiable objects .
for the set of 157 objects the number of unclassifiable objects was higher : 27% stars , 16% galaxies and 56% were not classified .
probably most of the unclassifiable objects in this set are galaxies because the second set is composed for fainter objects than the set of 31 objects , and so the fraction between galaxies and stars should be higher . for twelve objects
that were classified as galaxies we were able to roughly estimate their redshift , yielding values in the range of @xmath50 .
none of the classified objects appears unusual in any respect .
wx ser , a mira - type variable is detected @xmath51 outside the grb 970111 error box .
its period of is 425.1 days and a maximum magnitude of v=12.0 and a minimum of v@xmath016 .
however , comparing the b images we taken on 12 - 13 march and 15 - 16 march 1997 to those obtained on july 1997 , we found a variability in the b band of @xmath52b@xmath53 mag , ranging from b=20.4@xmath54 to b @xmath0 15.4 . on the other hand , the variable object found by masetti et al .
( 1997b ) is located outside the field of view of cafos and dfosc , so we can not provide further measurements .
grb 980329 is the tenth grb localized with the wfc on board _ bepposax_. the presence of a radio counterpart inside the nfi error box that peaked @xmath7 3 days after the gamma - ray event ( taylor et al .
1998 ) , enabled the detection of a near infrared and an optical ( i and r band ) counterpart ( djorgovski et al .
1998c , klose 1998 ,
larkin et al .
1998a , 1998b , palazzi et al .
1998a , 1998b ) .
grb 970111 and grb 980329 are the most intense grbs detected by the wfc and the grb monitor ( grbm ) , showing a prominent emission above 40 kev .
in fact , their fluence in the 50 - 300 kev range is about more than four times larger than the largest of the other grbs detected by _
bepposax_. on the other hand , they displayed the hardest spectra of the _ bepposax _ grbs , showing a hardness ratio ( hr ) between 0.6 and 0.7 ( int zand et al .
therefore , at first sight one could speculate that both grbs were originated under similar physical conditions and nearer than the other grbs of the _ bepposax _ sample .
if this was the case , the optical decay curves should be somehow similar .
this fact could give us a clue for explaining the non detectability of grb 970111 optical transient 19 hours after the gamma - ray event .
according to the power law decay of grb 980329 ( @[email protected] ) and the magnitude of the grb 980329 optical transient [email protected] 20 hours after the gamma - ray event ( palazzi et al .
1998b ) , the magnitude 19 hours after the grb would be r=23.5 , barely detectable in our images of grb 970111 .
taking into account the fluence of grb 971214 and assuming that grbs resemble standard candles it could imply that both grbs were originated from a nearby source in comparison to the source that produced grb 971214 ( @xmath6=3.42 ) .
in fact , palazzi et al . (
1998b ) suggest that grb 980329 could be arised from a strongly obscured starburst galaxy at @xmath151 .
the new wfc and nfi positions reported for grb 970111 ( heise et al .
1998 , feroci et al . 1998 ) and the only marginal overlap between the later and the ipn error box , made us to consider the grb error box as several non overlapping regions which had to be analyzed independently .
any possible fading was @xmath57 mag for objects with b@xmath021 , [email protected] and @xmath58 mag for those down to b=23 and r=22.6 .
no fading object was detected , within the 1sax j1528.8 + 1937 and 1sax j1528.8 + 1944 error boxes , being any fading @xmath0 0.1 mag for objects with b@xmath021 and r @xmath0 20.8 .
the colour - colour diagrams constructed for each non overlapping grb error box show neither objects with unusual colours nor objects with ultraviolet excesses or highly reddened galaxies .
the low interstellar extinction in the direction to grb 970111 makes the colour - colour diagrams similar to the unredenned ones .
no objects bluer than u - b@xmath0 - 0.70 , b - r @xmath0 0.60 , b - v @xmath00.45 , v - r @xmath00.32 , r - i @xmath00.17 were found inside the different areas studied .
a photometric classification and redshift estimation of the objects in the grb error box revealed no particularly conspicuous object . on the basis of the the b - r index we have found a possible candidate with b=20.6 for the x - ray source detected by sax , called 1sax j1528.8 + 1937 which is the bluest object found in the @xmath24 radius image .
spectroscopic observations revealed that it is a seyfert-1 galaxy at redshift @xmath1 . according to previous surveys , the a _ priori _ probability that a quasar or seyfert-1 galaxy with @xmath59 2.2 and b@xmath021 is located by chance inside the 1@xmath21 x - ray error box , is at most @xmath7 3% .
therefore , we propose the seyfert-1 galaxy as the source of 1sax j1528.8 + 1937 . on other hand , we report the large amplitude found in the variable star wx ser ( @xmath52b @xmath60 5 mag ) located @xmath51 outside the grb 970111 error box .
grb 970111 shows similar characteristics to grb 980329 , being the most intense and showing the hardest spectra of the grbs detected by _
bepposax_. if they also shared a similar optical decay , the magnitude 19 hours after the grb would be r=23.5 ( barely detectable in our images of grb 970111 ) .
this fact could explain the lack of detection of the grb 970111 optical transient .
we are grateful to r. castillo for the help provided at la silla and to m. de santos - lle for fruitful discussions . this work has been partially supported by spanish cicyt grant esp95 - 0389-c02 - 02 and by the university of bologna ( funds for selected research topics ) .
jochen heidt and thomas seitz acknowledge support by the deutsche forschungsgemeinschaft through sfb 328 . | we present the results of the monitoring of the grb 970111 field that started 19 hours after the event .
this observation represents the fastest ground - based follow - up performed for grb 970111 in all wavelengths .
as soon as the detection of the possible grb 970111 x - ray afterglow was reported by feroci et al .
( 1998 ) we reanalyzed the optical data collected for the grb 970111 field .
although we detect small magnitude variability in some objects , no convincing optical counterpart is found inside the wfc error box . any change in brightness 19 hours after the grb is less than 0.2 mag for objects with b @xmath0 21 and r @xmath0 20.8 .
the bluest object found in the field is coincident with 1sax j1528.8 + 1937 .
spectroscopic observations revealed that this object is a seyfert-1 galaxy with redshift @xmath1 , which we propose as the optical counterpart of the x - ray source .
further observations allowed to perform multicolour photometry for objects in the grb 970111 error box .
the colour - colour diagrams do not show any object with unusual colours .
we applied a photometric classification method to the objects inside the grb error box , that can distinguish stars from galaxies and estimate redshifts .
we were able to estimate photometric redshifts in the range @xmath2 for several galaxies in this field and we did not find any conspicuous unusual object .
we note that grb 970111 and grb 980329 could belong to the same class of grbs , which may be related to nearby sources ( @xmath3 ) in which high intrinsic absorption leads to faint optical afterglows . |
since the top quark was discovered at the fermilab tevatron in 1995@xcite , it has always been one of the hottest topics in particle physics .
so far , the top quark is the heaviest particle discovered , with a mass close to the electroweak symmetry breaking scale .
thus it is a wonderful probe for the electroweak breaking mechanism and new physics ( np ) beyond the standard model ( sm ) . as a genuine top quark factory
, the lhc will copiously produce the top quark events and can provide a good opportunity to scrutinize the top quark properties . in order to solve the hierarchy problem ,
the little higgs model was proposed @xcite , where the higgs boson is constructed as a pseudo - goldstone boson .
the littlest higgs model ( lh)@xcite provides an economical realization for this theory , however , this model suffers strong constraints from electroweak precision tests @xcite .
a feasible way to relax this constraints is to introduce a discrete symmetry called t - parity@xcite in the lh model .
this resulting model is referred to as the littlest higgs model with t - parity ( lht ) .
the lht model predicts many new particles , such as heavy gauge bosons , mirror fermions and heavy top partners , they can interact with the top quark and contribute to the top pair production at the loop level . at the lhc
, top quarks can be mostly produced in @xmath4 production through strong interactions and up to now the @xmath4 production has been measured in different channels with remarkable accuracy@xcite . in general , qcd controls the theoretical predictions for the @xmath4 production in both the sm and np at hadron colliders , and the qcd high order corrections play a key role for the accurate theoretical predictions .
therefore , it is necessary to perform the qcd high order calculations in order to test the sm and search for np . in our previous work
, we have studied the one loop electroweak effects on the @xmath4 production process in the lht at the lhc@xcite . in this paper
, we consider the latest experimental constraints and calculate the one loop qcd corrections to the @xmath4 production process in the lht at the lhc . since the new interactions between top quark and lht particles can not only affect the @xmath4 production rate but also the spin polarization@xcite , we also discuss the lht corrections to the spin polarization in the @xmath4 production process .
this paper is organized as follows . in sec.ii
we give a brief review of the lht model related to our work .
in sec.iii we give a brief description for the one - loop qcd calculations in the lht model .
in sec.iv we show the numerical results and some discussions .
finally , we make a short summary in sec.v .
the lht model is based on a @xmath5 non - linear @xmath6 model . at the scale @xmath7 , the global group @xmath8 is spontaneously broken into @xmath9 by a @xmath10 symmetric tensor and the gauged subgroup @xmath11^{2}$ ] of @xmath8
is broken into the sm gauge group @xmath12 .
after the symmetry breaking , four new heavy gauge bosons @xmath13 appear , whose masses up to @xmath14 are given by @xmath15 with @xmath16 and @xmath17 being the sm @xmath18 and @xmath19 gauge couplings , respectively . in order to preserve the t - parity , a copy of quarks and leptons with t - odd quantum number are added .
we denote the mirror quarks by @xmath20 , where @xmath21 are the generation index . in order to cancel the quadratic divergences to the higgs boson mass arising from the sm top quark ,
an additional t - even top partner @xmath22 that has its associated t - odd mirror quark @xmath23 are introduced . the new fermions which can contribute to the one loop qcd correction of top quark pair production are @xmath24 , whose
masses up to @xmath14 are given by @xmath25\\ & & m_{t^{-}}=\frac{f}{v}\frac{m_{t}}{\sqrt{x_{l}}}[1+\frac{v^{2}}{f^{2}}(\frac{1}{3}-\frac{1}{2}x_{l}(1-x_{l}))]\end{aligned}\ ] ] where @xmath26 are the diagonalized yukawa couplings of the mirror quarks , @xmath27 is the mixing parameter between the sm top - quark @xmath28 and the new top - quark @xmath22 .
at the tree - level , the feynman diagrams of the process @xmath29 are shown in fig.1 .
the complete one - loop qcd corrections to the process @xmath30 can be generally divided into several parts : self - energies , vertex corrections , boxes and their relevant counter terms . if the amplitudes are performed at the order @xmath31 , we find that the new particles in the lht model only can contribute to the self - energies and the relevant counter terms , which means that there are only the lht qcd one - loop virtual corrections to the process @xmath30 .
the relevant feynman diagrams for the subprocesses @xmath32 and @xmath33 in the lht model are depicted in fig.2 .
we can see that the lht qcd one - loop virtual corrections ( @xmath34 ) come from the fermion loops .
the one - loop qcd corrected production cross section of the process @xmath35 can be obtained by @xmath36 in the t hooft - feynman gauge , we use the dimensional reduction method to regulate the ultraviolet ( uv ) divergences in the fermion loops and adopt the on - shell renormalization scheme to remove them .
we list the explicit expressions of these amplitudes in appendix .
we have checked that the uv divergences in the renormalized propagator have been canceled . due to no massless particles in the loop , there are no infrared ( ir ) singularities in the one - loop integrals . in our numerical calculations , we use the parton distribution function cteq10@xcite with renormalization / factorization scale @xmath37 .
the relevant lht parameters are the scale @xmath38 , the mixing parameter @xmath27 and the yukawa couplings @xmath39 . for the mirror fermion masses ,
we get @xmath40 at @xmath41 and assume that the masses of the first two generations are degeneracy : @xmath42 recently , both of the cms and atlas collaborations reported their search results of the fermionic top partner and respectively excluded the masses regions below 557 gev @xciteand 656 gev@xcite at 95% cl . in our numerical calculations , we consider the constraints above and scan the parameter regions : @xmath43gev , @xmath44 , @xmath45 .
and we require our samples to satisfy the constraints from refs.@xcite .
we take the input parameters of the sm as@xcite @xmath46 we will discuss the lht qcd one - loop corrections to the ( un)polarized top pair production by using the following observables : * for the unpolarized @xmath4 production , we calculate the relative corrections for total @xmath4 production cross section(@xmath47 ) , which is defined by : @xmath48 * for the polarized @xmath4 production , we calculate the spin correlation(@xmath49)@xcite , which is defined by : @xmath50 here , the subindices _
l(r ) _ represent left- ( @xmath51 ) and right - handed(@xmath52 ) top(antitop ) quarks , respectively .
in fig.3(a ) , we show the relative correction of the @xmath4 production cross section @xmath53 as the function of the scale @xmath38 at the lhc with @xmath0 tev respectively .
we can see that the maximum value of the relative correction to the @xmath4 production cross section can reach @xmath54% for @xmath55 tev and @xmath56% for @xmath57 tev , respectively .
the results indicate that the size of the corrections and the sensitivity on these effects are larger at 8 tev than at 14 tev .
this is because the correction of the subprocess @xmath58 is positive and the correction of the subprocess @xmath59 is negative so that they cancel each other . due to
the main correction comes from @xmath60 , the relative correction of the @xmath61 total cross section is negative .
when the center - of - mass energy @xmath62 varies from 8 tev to 14 tev , the correction of the subprocess @xmath63 increases quickly so that the cancel between this two subprocesses become stronger .
besides , when the scale @xmath38 increases , the relative corrections @xmath53 become small , which indicates that the lht qcd one - loop effects on @xmath4 production cross section will decouple at the high scale @xmath38 . in fig.3(b ) , we show the relative correction of the @xmath4 production cross section @xmath53 as the function of the yukawa couplings @xmath64 at the lhc with @xmath0 tev respectively .
we can see that the relative corrections @xmath65 decrease with the yukawa couplings @xmath64 increasing , which indicates that the lht qcd one - loop effects on @xmath4 production cross section also decouple at the heavy mirror quark masses . in fig.3(c )
, we show the relative correction of the @xmath4 production cross section @xmath53 as the function of the mixing parameter @xmath27 at the lhc with @xmath0 tev respectively .
the distribution behaviors of the samples can be explained as follows . according to the eqs.(4 , 5 )
, we can see that the heavy top quark @xmath22 and @xmath23 masses have a strong dependence on the mixing parameter @xmath27 .
when @xmath66 , the masses of @xmath22 and @xmath23 will become heavy and their contribution become very small .
when @xmath67 , the masses of @xmath22 will become heavy while the masses of @xmath23 will become light . as a result , the effect of @xmath23 will still reside in the @xmath4 production .
recently , the cms collaboration reported their measurement of the @xmath4 spin correlation coefficient @xmath68 in the helicity basis@xcite , which is agree with the sm predictions . in fig.4
, we show the relative correction of the @xmath4 spin correlation @xmath49 as the function of @xmath69 for the lhc with @xmath0 tev , respectively .
we can see that the behaviors of @xmath49 versus @xmath70 are similar to those of the relative correction @xmath53 .
the maximum value of @xmath49 can reach about @xmath2 for @xmath55 tev and @xmath3 for @xmath57 tev , which may be detected at the lhc@xcite .
in this paper , we studied the one - loop @xmath31 qcd corrections to @xmath4 production in the lht model at the lhc for @xmath0 tev .
we presented the numerical results for the relative correction to @xmath4 production cross section and @xmath4 spin correlation at the lhc .
the relative correction of the @xmath4 production cross section is negative and only can reach @xmath1 .
the relative correction of the @xmath4 spin correlation @xmath49 can reach about @xmath2 for @xmath55 tev and @xmath3 for @xmath57 tev , which may be a potential probe to detect the lht effects at the lhc .
they can be represented in form of 1-point and 2-point standard functions @xmath71 . here
@xmath72 and @xmath73 denote the momenta of the top and antitop respectively , and they are assumed to be outgoing . where @xmath75\\ & & \delta z_{gg}=-\frac{\partial\sigma^{gg}_{t}(k^{2})}{\partial k^{2}}|_{k^{2}=0}\\ & & \hat{\sigma}^{gg}_{\mu\nu}=\hat{\sigma}^{gg}_{\mu\nu}(u_{h})+\hat{\sigma}^{gg}_{\mu\nu}(d_{h})+\hat{\sigma}^{gg}_{\mu\nu}(t^{\pm})\end{aligned}\ ] ] @xmath76\nonumber\\ & & + [ -\frac{2}{3}a_{0}+\frac{2}{3}m^{2}_{t^{+}}b_{0}-\frac{2}{3}k^{2}b_{1}(k , m_{t^{+}},m_{t^{+}})+\frac{2}{3}m^{2}_{t^{+}}-\frac{k^{2}}{9}]\nonumber\\ & & + [ -\frac{2}{3}a_{0}+\frac{2}{3}m^{2}_{u_{h}}b_{0}-\frac{2}{3}k^{2}b_{1}(k , m_{u_{h}},m_{u_{h}})+\frac{2}{3}m^{2}_{u_{h}}-\frac{k^{2}}{9}]\nonumber\\ & & + [ -\frac{2}{3}a_{0}+\frac{2}{3}m^{2}_{d_{h}}b_{0}-\frac{2}{3}k^{2}b_{1}(k , m_{d_{h}},m_{d_{h}})+\frac{2}{3}m^{2}_{d_{h}}-\frac{k^{2}}{9}]\nonumber\\ & & -[\frac{2}{3}m^{2}_{t^{-}}\frac{\partial b_{0}}{\partial k^{2}}-\frac{2}{3}b_{1}(0,m_{t^{-}},m_{t^{-}})-\frac{1}{9}]k^{2}\nonumber\\ & & -[\frac{2}{3}m^{2}_{t^{+}}\frac{\partial b_{0}}{\partial k^{2}}-\frac{2}{3}b_{1}(0,m_{t^{+}},m_{t^{+}})-\frac{1}{9}]k^{2}\nonumber\\ & & -[\frac{2}{3}m^{2}_{u_{h}}\frac{\partial b_{0}}{\partial k^{2}}-\frac{2}{3}b_{1}(0,m_{u_{h}},m_{u_{h}})-\frac{1}{9}]k^{2}\nonumber\\ & & -[\frac{2}{3}m^{2}_{d_{h}}\frac{\partial b_{0}}{\partial k^{2}}-\frac{2}{3}b_{1}(0,m_{d_{h}},m_{d_{h}})-\frac{1}{9}]k^{2}\}\end{aligned}\ ] ] n. arkani - hamed , a. g. cohen , and h. georgi , phys .
b 513 : 232 ( 2001 ) ; n. arkani - hamed , et al . , jhep 0208 : 020 ( 2002 ) ; jhep 0208 : 021 ( 2002 ) ; i. low , w. skiba , and d. smith , phys . rev .
d 66 : 072001 ( 2002 ) ; d. e. kaplan and m. schmaltz , jhep 0310 : 039 ( 2003 ) . n. arkani - hamed , a. g. cohen , e. katz , and a. e. nelson , jhep 0207 : 034 ( 2002 ) ; s. chang , jhep 0312 : 057 ( 2003 ) ; t. han , h. e. logan , b. mcelrath , and l. t. wang , phys
d 67 : 095004 ( 2003 ) ; m. schmaltz and d. tucker - smith , ann . rev .
nucl . part .
55 : 229 ( 2005 ) . c. csaki , j. hubisz , g.d .
kribs , p. meade , and j. terning , phys .
d 67 : 115002 ( 2003 ) ; phys . rev .
d 68 : 035009 ( 2003 ) ; j. l. hewett , f. j. petriello , and t. g. rizzo , jhep 0310 : 062 ( 2003 ) ; m. c. chen and s. dawson , phys . rev . d 70 : 015003 ( 2004 ) ; m. c. chen , et al . , mod .
a 21 : 621 ( 2006 ) ; w. kilian , j. reuter , phys . rev .
d 70 : 015004 ( 2004 ) .
h. c. cheng and i. low , jhep 0309 : 051 ( 2003 ) ; jhep 0408 : 061 ( 2004 ) ; i. low , jhep 0410 : 067 ( 2004 ) ; j. hubisz and p. meade , phys . rev .
d 71 : 035016 ( 2005 )
. g. aad et al .
( atlas collaboration ) , eur .
j. c 71 : 1577 ( 2011 ) ; s. chatrchyan et al .
( cms collaboration ) , eur .
j. c 71 : 1721 ( 2011 ) ; s. chatrchyan et al .
( cms collaboration ) , jhep 1107 : 049 ( 2011 ) ; g. aad et al .
( atlas collaboration ) , jhep 1205 : 059 ( 2012 ) ; s. chatrchyan et al .
( cms collaboration ) , phys .
d 85 : 112007 ( 2012 ) ; g. aad et al .
( atlas collaboration ) ( 2012 ) , [ arxiv:1205.2067 ] .
j. cao , l. wang , l. wu , j. m. yang , phys . rev .
d 84 : 074001 ( 2011 ) ; j. cao , l. wu , j. m. yang , phys .
d 83 : 034024 ( 2011 ) ; j. cao , z. heng , l. wu , j. m. yang , phys .
d 81:014016 ( 2010 ) .
h. l. lai , m. guzzi , j. huston , z. li , p. m. nadolsky , j. pumplin , c. p. yuan , phys .
d 82 : 074024 ( 2010 ) .
cms collaboration , phys .
b 716 : 103 ( 2012 ) .
atlas collaboration , phys .
b 718 : 1284 - 1302 ( 2013 ) .
j. hubisz , p. meade , a. noble , and m. perelstein , j. high energy phys .
0601 : 135 ( 2006 ) ; a. belyaev , c .- r .
chen , k. tobe , and c. p. yuan , phys .
d 74 : 115020 ( 2006 ) ; q .- h . cao and c
chen , phys .
d 76 : 075007 ( 2007 ) ; j. reuter , m. tonini , jhep 0213 : 077 ( 2013 ) ; x. f. han , l. wang , j. m. yang , j. y. zhu , phys .
d 87 : 055004 ( 2013 ) ; j. reuter , m. tonini , m. de vries , arxiv:1307.5010 [ hep - ph ] , arxiv:1310.2918 [ hep - ph ] ; c. c. han , a. kobakhidze , n. liu , l. wu , b. f. yang , nucl .
b 890 : 388 - 399 ( 2014 ) . j. beringer et al . , ( particle data group ) , phys . rev .
d 86 : 010001 ( 2012 ) ; g. aad et al .
[ atlas collaboration ] , phys .
b 716:1 - 29 ( 2012 ) ; s. chatrchyan et al .
[ cms collaboration ] , cms - hig-12 - 028 , arxiv:1207.7235 .
j. h. kuhn , nucl .
b 237 : 77 ( 1984 ) ; v. d. barger , j. ohnemus and r. j. n. phillips , int .
a 4 : 617 ( 1989 ) ; g. mahlon and s. j. parke , phys .
b 411 : 173 ( 1997 ) ; phys . rev .
d 53 : 4886 ( 1996 ) ; phys . rev .
d 81 : 074024 ( 2010 ) ; t. stelzer and s. willenbrock , phys .
b 374 : 169 ( 1996 ) ; w. bernreuther , a. brandenburg , z. g. si , p. uwer , phys . rev . lett .
87 : 242002 ( 2001 ) ; w. bernreuther and z. g. si , nucl . phys .
b 837 : 90 ( 2010 ) ; g. mahlon , arxiv:1007.1716 [ hep - ph ] . | in this work , we investigate the one - loop qcd correction to top pair production in the littlest higgs model with t - parity at the lhc .
we calculate the relative correction of the top pair production cross section and top - antitop spin correlation at the lhc for @xmath0 tev .
we find that the relative corrections of top pair production cross section can reach about @xmath1 , and the top - antitop spin correlation can reach @xmath2(@xmath3 ) at the 8(14 ) tev lhc in the favorable parameter space . |
scattered starlight is a valuable tool for measuring the properties of interstellar dust . in optical reflection
nebulae @xcite , dark clouds @xcite and the diffuse galactic light @xcite the scattering grains typically lie many arcseconds from the illuminating sources . the dust can be nearer to or further from us than the source and the detected photons have been deflected through a wide range of angles .
further information about interstellar grains comes from stellar x - rays scattered by foreground dust within a few arcseconds of the star @xcite .
the measurements sample the whole column of material between us and the star because the grains are strongly forward scattering at x - ray energies . here
we show that high - contrast imaging with a space - based telescope such as the proposed terrestrial planet finder coronagraph enables a new kind of measurement for probing dust at projected separations similar to the x - ray measurements .
the technique samples a range of scattering angles because at visible wavelengths the scattering is only mildly biased in the forward direction . as a consequence ,
most of the light comes from foreground grains lying close to the star .
the scattered light halo carries information on the interstellar dust passing through the cavity opened in the interstellar gas by the stellar wind .
evidence for the existence of interstellar grains within our own heliosphere was obtained by the ulysses spacecraft when 55 dust impacts as measured by the cosmic dust experiment were identified as interstellar by their speed , mass and arrival direction @xcite .
the overall contribution of interstellar dust to the zodiacal cloud the debris disk of the solar system is unknown , although the fraction is presumably greater in the outer solar system .
collisional debris from the asteroid belt dominates inside 3 au @xcite .
however the interstellar grains , charged by photoionization and strongly influenced by the solar gravity , radiation pressure and magnetic field , are able to penetrate deep into the inner solar system .
a uniform incoming spatial distribution of interstellar grains becomes strongly clumped as a function of particle size and phase of the solar cycle , although the contribution of this component to the all - sky thermal emission as viewed from 1 au is negligible due to the local dominance of asteroidal and cometary material @xcite .
starlight scattered by interstellar dust also is a source of noise in coronagraphic planet searches .
we calculate the expected brightness and distribution of the scattered light and show that the halo is unlikely to affect the direct detection of earth - like planets around the nearest stars .
we compute the trajectories of interstellar grains under the stellar gravity , radiation and magnetic forces and show that the stellar wind produces observable signatures in the scattered light . for a large number of stars at distances greater than 100 pc , high - contrast imaging can potentially yield detailed information about the stellar wind and the adjacent interstellar medium .
the remaining five sections of the paper cover the radiative transfer method ( 2 ) , the results for uniformly - distributed dust , obtained analytically assuming isotropic scattering ( 3 ) and numerically including the anisotropy ( 4 ) , the results for dust passing through a model stellar wind ( 5 ) , and a summary and conclusions ( 6 ) .
the transfer of visible light through the optically - thin interstellar medium is dominated by scattering and the radiative transfer equation at wavelength @xmath2 is @xmath3 where @xmath4 is the monochromatic specific intensity , @xmath5 the optical depth , @xmath6 the albedo of the scattering grains and @xmath7 the phase function describing scattering from all directions @xmath8 into the beam direction @xmath9 .
the star is approximated by a disk of uniform specific intensity @xmath10 , subtending a small angle @xmath11 as seen from a dust grain on our line of sight ( figure [ fig : sketch ] ) .
the scattering phase function is taken to be constant across the face of the star , and the solid angle filled by the star is @xmath12 .
the transfer equation reduces to @xmath13 where @xmath14 is the scattering angle between the rays joining star to grain and grain to observer .
the intensity of the scattered light is calculated by integrating eq .
[ eqn : halort ] along the line of sight . the angles @xmath14 and @xmath15 vary with the distance from the observer , @xmath16 . writing the distance in place of the optical depth using @xmath17 , neglecting emission from any distant background , and taking the simple case of an interstellar medium with uniform opacity , the intensity is @xmath18
we first solve equation [ eqn : si ] assuming isotropic scattering , with @xmath19 independent of @xmath14 and @xmath16 . along a line of sight offset by a small angle @xmath20 from the direction to the star ,
the stellar angular radius @xmath15 varies as @xmath21 ^ 2\right)^{1/2}},\ ] ] where the star has radius @xmath22 and lies at distance @xmath23 . with a change of variable to @xmath24 ,
the solution to the transfer equation is @xmath25 the integral is very nearly @xmath26 because @xmath27 is almost @xmath28 and the integrand falls off rapidly with @xmath29 . as a specific case
we take the sun observed from a distance of 100 pc at wavelength 547 nm through a uniform interstellar medium with 0.1 h atoms @xmath0 , similar to the densities found in the solar neighborhood @xcite . for milky way grains ,
the total opacity @xmath30 @xmath31 h@xmath32 h @xmath33 @xmath34 and albedo @xmath35 @xcite .
the optical depth to the star @xmath36 .
the predicted surface brightness ratio at an angular separation @xmath20 is @xmath37 and falls off inversely with the separation .
the integrated flux within an angle @xmath38 is proportional to the optical depth , @xmath39 and the flux within one arcsecond is eighty times greater than the nominal threshold @xmath40 for detection by the proposed terrestrial planet finder coronagraph .
interstellar grains scatter visible light anisotropically .
we treat this effect using a standard model for grains in the diffuse interstellar medium of the milky way , with the phase function @xmath41 computed from mie theory by @xcite .
forward scattering is moderately preferred and at the chosen wavelength of 547 nm , each event has a 50% chance of deflecting the photon through an angle less than 43 .
the interstellar medium is uniform with the same density as in section [ sec : isotropic ] .
the transfer equation [ eqn : halort ] is numerically integrated along the line of sight to the observer using a fourth - order runge - kutta method .
we have checked that the method reproduces the analytic solution given by equation [ eqn : isotropic ] in the case of isotropic scattering . including the preferential forward scattering yields a halo 48% brighter .
the results for a solar analog star at a distance of 100 pc are shown in figure [ fig : anisotropic ] .
also shown for comparison is a solar analog exozodiacal cloud from appendix 1.b of @xcite .
the light scattered from the interstellar grains is brighter than that from the exozodiacal cloud outside 50 milliarcsec .
the exozodiacal cloud is inclined 60 from face - on and its brightness was calculated including anisotropic scattering .
much of the interstellar scattered light comes from grains passing close by the star , due to the inverse square falloff of the stellar illumination .
most comes from grains whose distance from the star is less than ten times the projected separation in the plane of the sky .
the preferential forward scattering has a smaller effect , making the dust in front of the star contribute more than the dust at the same distance behind .
the contributions from dust at different positions along a line of sight passing 0.1 arcsecond from a star at 100 pc are shown in figure [ fig : where ] .
overall , the scattered light halo is most sensitive to the distribution of interstellar dust in the vicinity of the illuminating star .
measurements and models of the optical properties of the dust in the diffuse interstellar medium allow a range of values for the albedo and the degree of forward scattering @xcite .
the @xcite model used above lies near the middle of the range in both the albedo and the scattering asymmetry as measured by the mean cosine scattering angle @xmath42 at 547 nm .
the appearance of the halo is qualitatively independent of the dust properties over the allowed range .
the halo brightness is simply proportional to the albedo ( equation [ eqn : isotropic ] ) .
the scattering asymmetry affects both the halo brightness and the location of the grains giving most of the light .
we examine the dependence by re - computing the halo using a standard henyey - greenstein phase function with asymmetry parameters up to the maximum @xmath43 consistent with the range of the measurements .
half the photons are deflected through angles less than 19 .
the resulting halo is 2.19 times brighter than with the @xcite phase function .
the fraction of the light 0.1 arcsecond from the star that is scattered by grains lying more than 200 au from the star rises from 7% to 21% .
a star surrounded by the more forward - throwing grains has a brighter halo with a slightly lower contrast for any features arising from dust passing close by the star .
sub - micron interstellar grains entering the solar system are deflected by the solar gravity , radiation pressure and solar wind electromagnetic forces @xcite .
the highest densities of grains 0.1 micron in radius occur where the particles pile up against the solar wind and where the magnetic forces concentrate particles over the solar magnetic poles @xcite . in this section
we explore whether similar effects can be detected in the scattered light halos around other stars .
the star is taken to be a solar analog and its wind is modeled following @xcite , @xcite and @xcite by applying the expanding corona model of the solar wind magnetic field @xcite .
the gas streaming outward from the star draws out the stellar magnetic field lines so that near the star the field is approximately radial .
the more distant field lines are dragged into the shape of an archimedean spiral as the star rotates .
the radial and azimuthal components of the field are @xmath44 and @xmath45 where @xmath46 is the stellar angular velocity , @xmath47 is the astrocentric colatitude angle , @xmath48 is the stellar wind speed ( chosen as 400 km s@xmath49 ) and @xmath50 is the distance at which the reference field strength @xmath51 is taken .
the lorentz force is then @xmath52\ ] ] where @xmath53 is the grain charge and @xmath54 the grain speed . the grain charge @xmath53 is given by @xmath55 where @xmath56 is the grain radius , @xmath57 cv@xmath49m@xmath49 is the permittivity of free space and @xmath58 is the grain surface potential .
dust in the stellar wind cavity is expected to be positively charged due to the dominance of stellar photoionization over the competing mechanism of the ` sticking ' of stellar wind electrons @xcite . since the stellar wind density roughly follows an inverse square law decrease with astrocentric distance , particles
are charged to + 5 - 10 volts with little distance variation .
we adopt a surface potential @xmath58 of + 5 volts .
notice that the dominant electric part in equation [ eqn : lorentz ] is proportional to @xmath59 and therefore to @xmath60 .
this can make the instantaneous lorentz force dominate on small grains at large astrocentric distances .
the ballerina model proposed by @xcite gives an explanation for solar system spacecraft measurements which indicate that near the ecliptic plane the interplanetary magnetic field is directed inward in certain regions and outward in others . in this picture
a current sheet separates plasma from either hemisphere carrying fields of opposite polarity .
as the sheet rotates with the sun the small up and down displacements in the sheet , similar to the wave motion of the skirt of a spinning ballerina , explain the observed effect .
we use a simplified model of this phenomenon to describe the magnetic field strength averaged over a stellar rotation by the inclusion of a current sheet which changes its tilt angle @xmath61 to the stellar equatorial plane at a constant angular velocity of @xmath62 radians per year over the stellar magnetic activity cycle @xcite , so that the sheet completes one 360 rotation in 22 years .
the averaged field at latitude @xmath63 is then a fraction @xmath64 of the unipolar field .
the line of nodes defined by the intersection of the current sheet and stellar equatorial plane will rotate through 360 in one 27 day stellar rotation .
the effect of radiation pressure is to reduce the stellar gravitational force , the magnitude of the radiation pressure being given by @xmath65 where @xmath66 is the flux , @xmath67 is the speed of light and @xmath68 is the radiation pressure efficiency of the particle .
grains larger than one micron are rare in the diffuse interstellar medium according to extinction curve fitting results @xcite .
grains smaller than 0.01 micron have little optical cross - section and furthermore are largely excluded from the heliosphere by magnetic forces due to their high charge - to - mass ratios .
grains in the 0.1 micron size range experience a radiation pressure force larger than the gravitational force but nevertheless are able to penetrate deep into the solar system @xcite .
most of the scattered light from interstellar grains passing through the astrosphere of a solar analog star is likely due to grains about 0.1 micron in size .
we therefore concentrate on grains with radii between 0.02 and 0.3 microns .
the ratio @xmath69 of the radiation pressure force to the gravitational force may be expressed in terms of @xmath68 as @xmath70 where @xmath71 is the density of the particle @xcite .
the optical efficiencies of the grains are found from mie theory .
we assume astronomical silicate composition and optical constants @xcite .
clearly , the particle spatial distributions will vary with the particle size ; for example , the gravitational force will dominate for larger particles , whereas the interaction with the magnetic field will become increasingly important at the small particle end of the size range .
the distribution of particles will also depend on the stellar magnetic cycle , so that the particle positions will represent a ` snapshot ' in time rather than an equilibrium configuration .
taking into account gravity , radiation pressure and the stellar magnetic field model outlined above , we follow the trajectories of a uniform distribution of dust particles at infinity as they approach the star from the upstream direction at a speed of 26 km s@xmath49 .
the equations of motion are integrated at small intervals of time and the position , speed and acceleration of the particles are calculated .
the local properties of the stellar wind and magnetic field are updated as the integration progresses .
we continue the integration until the heliosphere analog is filled with dust particles and stop the run at a given phase of the stellar magnetic field cycle with the magnetic field orientation matching that of the sun in the year 1990 .
the dust density is then calculated by binning the particles in cubes 2 au on a side and normalizing to the upstream interstellar density .
the resulting distributions are broadly similar for particles 0.02 to 0.3 @xmath72 m in radius .
the distribution of 0.1 @xmath72 m grains is shown in figure [ fig : density ] .
the particles are pushed by the magnetic forces into the equatorial current sheet , and pile up 10 au upstream from the star ( left panel ) , where the density reaches six times that in the interstellar medium .
the greatest density enhancement of all is a factor fifteen and occurs in two fins over the magnetic poles and 60 au downstream from the star ( right panel ) .
these result from the focusing of the grains by the inclined magnetic fields off the equatorial plane .
synthetic images of the scattered starlight are constructed by solving the transfer equation [ eqn : halort ] to find the intensity on a grid of sky points , treating the anisotropic scattering as described in section [ sec : anisotropic ] with the @xcite mie theory phase function .
no exozodiacal cloud is included . selected results are shown in figures [ fig : halo ] and [ fig : cut ] . in addition to an overall deficit of scattered light relative to the case of uniformly - distributed dust , several features produced by the magnetic field are clearly visible , including the density enhancements in the equatorial disk and the polar fins .
we have shown that stars appear surrounded by halos of light scattered by interstellar dust .
the halos are bright enough for detection with the proposed terrestrial planet finder coronagraph around stars at distances of 100 pc and greater .
the halo reveals the distribution of dust in the vicinity of the star , providing opportunities to probe the interstellar medium and measure the properties of the stellar wind .
particles of different sizes have different distributions , with larger particles approaching the star most closely @xcite .
this size - sorting can potentially be used to measure the distribution of grain sizes at selected locations in the interstellar medium and to resolve whether particles larger than one micron are present in significant numbers , as suggested by dust impact detector data from the ulysses and galileo spacecraft @xcite . since the grains are deflected by stellar wind magnetic forces , the features in the scattered light halo
can be used to estimate the wind speed and the relative motion between the star and the surrounding interstellar medium .
the results will complement the wind mass loss rates estimated using the lyman-@xmath73 absorption by the gas collected near stellar wind bow shocks @xcite .
the scattering halos will make it possible in addition to determine the wind magnetic field strength and orientation ( figure [ fig : halo ] ) and , by observations spanning a decade or so , the variation over the stellar activity cycle @xcite . imaging
the halos will yield fresh information about the winds of stars with a range of ages and could lead to a better understanding of stellar spin histories @xcite and the long - term evolution of stellar activity .
this work was carried out at the jet propulsion laboratory , california institute of technology with support from the internal research and technology development fund innovative spontaneous concepts program .
levine , m. , shaklan , s. , & kasting , j. 2006 , terrestrial planet finder coronagraph science and technology definition team report , publication d-34923 of the jet propulsion laboratory , california institute of technology | we show that space - based telescopes such as the proposed terrestrial planet finder coronagraph will be able to detect the light scattered by the interstellar grains along lines of sight passing near stars in our galaxy .
the relative flux of the scattered light within one arcsecond of a star at 100 pc in a uniform interstellar medium of 0.1 h atoms @xmath0 is about @xmath1 .
the halo increases in strength with the distance to the star and is unlikely to limit the coronagraphic detection of planets around the nearest stars .
grains passing within 100 au of sun - like stars are deflected by radiation , gravity and magnetic forces , leading to features in the scattered light that can potentially reveal the strength of the stellar wind , the orientation of the stellar magnetic field and the relative motion between the star and the surrounding interstellar medium . |
hadron spectroscopy plays a valuable role in particle physics .
it was crucial in validating quantum chromodynamics ( qcd ) and the quark substructure of matter .
it provides a stage for understanding nonperturbative techniques , not only in qcd but potentially elsewhere in physics .
hadron spectra are crucial in separating electroweak physics from strong - interaction effects , as in charm and beauty decays .
quarks and leptons have an intricate level and weak coupling structure for which we have no fundamental understanding .
sharpening spectroscopic techniques may help solve this problem .
i review hadron spectroscopy with emphasis on new results from heavy - quark studies , discussing no - quark states ( glueballs ) , light - quark states , charmed hadrons , charmonium , beauty hadrons , and the @xmath0 family , concluding with a homework assignment .
in qcd , quarkless states ( `` glueballs '' ) may be constructed from pure - glue configurations .
if @xmath1 is the gluon field - strength tensor , one can form @xmath2 states as @xmath3 , @xmath4 as @xmath5 , and higher - spin configurations using derivatives or more than two gluon fields .
all such states should be flavor - singlet with isospin @xmath6 .
although equal coupling to light nonstrange and strange quarks has usually been assumed , @xmath7 couplings could be favored @xcite .
lattice qcd calculations predict the lowest glueball to be @xmath8 with @xmath9 gev .
many other @xmath6 levels , e.g. , @xmath10 , @xmath11 ( @xmath12 = gluon ) , @xmath13 , can mix with such a state . one must study @xmath6 levels and their mesonic couplings to separate out glueball , @xmath14 , and @xmath7 components .
understanding the rest of the _ flavored _ @xmath10 spectrum for the same @xmath15 thus is crucial .
two mechanisms for producing glueballs are shown in fig .
[ fig : prod ] . in double pomeron exchange ( as utilized , for example , in cern experiment wa-102 @xcite ) , the centrally produced state is expected to have @xmath16 .
radiative @xmath17 decay is about 10% of three - gluon decay and also should populate @xmath6 states if the photon couples to the charmed quark .
the bes collaboration has obtained a sample of 58 million @xmath17 decays , and the cleo - c detector hopes to record an even larger number .
other processes for glueball production include @xmath18 and @xmath19 annihilation @xcite .
one can diagnose the flavor of neutral states @xmath20 by noting that @xmath21 favors ( nonstrange , strange ) @xmath20 , with @xmath22 testing the flavor content of @xmath20 .
a glueball is expected to have a small decay width to @xmath23 .
one pattern of glueball masses is predicted in ref .
the lowest state has @xmath2 and mass around 1.7 gev .
the next - lowest are @xmath24 gev ) and @xmath25 gev ) .
these predictions are difficult both to make and to test , since mixing with decay channels and @xmath26 configurations can occur .
the best prospects are to sort out the @xmath8 @xmath6 mesons @xcite , for which several candidates exist .
the @xmath27 ( with @xmath28 mev ) , a possible qcd analogue of the higgs boson , appears to be a dynamically induced @xmath29 resonance .
it can be generated from @xmath29 scattering using just current algebra , crossing symmetry , and unitarity @xcite .
it also can be interpreted as a candidate for a light @xmath30 state @xcite .
the @xmath31 is a narrow resonance with @xmath32100 mev ) which appears to be strongly correlated with @xmath33 threshold .
it has been interpreted variously as @xmath7 , @xmath34 , an @xmath35 molecule , or a mixture of these states .
for a recent discussion , see @xcite .
three states have been proposed as mixtures of a glueball and @xmath36 @xmath10 states @xcite .
( 1 ) the @xmath37 , with @xmath38500 mev , viewed mainly as @xmath39 , is seen in @xmath40 .
( 2 ) the @xmath41 , with @xmath42 mev , is not seen in @xmath40 but appears in @xmath19 annihilations and central @xmath33 production .
@xcite consider it to be mainly a glueball @xmath43 .
( 3 ) the @xmath44 , with @xmath45 mev , is considered in ref.@xcite to be mainly @xmath7 since it is prominent in @xmath46 .
it could be a glueball if @xmath47 couplings are enhanced @xcite .
the scheme of ref .
@xcite is consistent with production in @xmath48 , @xmath49 , and @xmath50 [ @xmath51 processes but needs to be tested via @xmath52 .
the @xmath53 candidates listed in @xcite are @xmath54 , @xmath55 , @xmath56 , @xmath57 , @xmath58 , and @xmath59 .
the first two are well - established @xmath60 states .
higher states with masses within 100 mev of one another may not be distinct if their observed properties depend on production .
the @xmath61 are in the range expected for radial @xmath10 excitations .
the highest - lying ones are in the expected glueball range , however .
their flavor diagnosis via decay modes or suppressed @xmath23 production is crucial .
@xmath4 candidates @xcite are @xmath62 , @xmath63 , @xmath64 , @xmath65 , and @xmath66 , while a state @xmath67 , decaying to @xmath68 , has recently been reported in @xmath69 @xcite . there is room for glueball content of these states but two of the three between @xmath70 and 1.5 gev are likely radial @xmath71 excitations , while the glue content of @xmath72 is limited by the observed @xmath73 rate @xcite .
the state @xmath67 is consistent with the enhancement at @xmath19 threshold reported earlier by bes in @xmath74 @xcite if @xmath75 .
( the @xmath19 enhancement was also consistent with @xmath8 but such a state can not decay to @xmath68 . )
it could be a glueball with a large @xmath76 content in its wave function ( but is below the lowest expected @xmath4 mass @xcite of 2.6 gev ) , or a @xmath19 `` baryonium '' resonance @xcite .
baryon - antibaryon enhancements near threshold appear not only in radiative @xmath17 decays but also in @xmath77 , @xmath78 , and @xmath79 decays @xcite , and in @xmath80 @xcite .
mechanisms for production of these enhancements are shown in fig .
[ fig : bbbar ] .
the low - mass @xmath81 enhancement seen in @xmath77 appears to be associated with fragmentation , since antiprotons are seen to follow the @xmath82 .
the other enhancements seen in @xmath83 decays also appear to favor a fragmentation mechanism .
the @xmath84 enhancement seen in @xmath17 decays has @xmath85 mev , @xmath86 mev .
the decay @xmath87 , with a contribution from the @xmath88 intermediate state , measures the @xmath29 scattering length difference @xmath89 , predicted to be @xmath90 @xcite . by studying @xmath91 in a @xmath92 atom , the dirac collaboration measured @xmath93 @xcite .
on the basis of @xmath94 @xmath95 decays and a cleanly observed cusp in the @xmath96 spectrum near @xmath92 threshold , the cern na48 collaboration finds @xmath97 @xcite . as mentioned earlier , a dynamical @xmath98 resonance in @xmath99
is predicted using only current algebra , crossing symmetry , and unitarity @xcite .
a pole with large imaginary part occurs at or below @xmath100 .
the effects of this pole are very different in @xmath99 [ where an adler zero suppresses the amplitude at low @xmath101 and ( e.g. ) @xmath102 , where no such suppression occurs @xcite .
evidence for such an effect has been cited in the decays @xmath103 @xcite .
a broad scalar resonance @xmath104 near 800 mev , likely also of dynamical origin , is seen in the @xmath105 system in the decay @xmath106 , and a model - independent phase shift analysis shows resonant @xmath107 behavior in the @xmath105 system @xcite .
elastic @xmath108 scattering does not show similar behavior , casting doubt on the nave application of watson s theorem to the inelastic process .
the @xmath104 is also seen by the bes collaboration in @xmath109 decays @xcite .
an effective supersymmetry between scalar @xmath110 color triplet states and spin-1/2 quarks was suggested some time ago @xcite .
jaffe and wilczek @xcite have proposed that @xmath111 , @xmath112 , @xmath31 , and @xmath113 could be viewed as a nonet of diquark antidiquark pairs . in that case
@xmath114=0 diquarks would have effective masses @xmath115 ) \simeq 300 $ ] mev , @xmath116 ) = m([ds ] ) \simeq 490 $ ] mev .
however , one can also view @xmath117 , and @xmath118 as meson - meson ( `` molecular '' ) effects .
selem and wilczek @xcite have performed successful fits to baryon regge trajectories using a quark - diquark picture .
interesting features of their treatment are the derivation of a universal string tension giving a universal slope for trajectories , and a prediction of the slope for trajectories of baryons with one heavy quark .
if diquarks are really pointlike , the pauli principle does nt mind putting @xmath119,~[us],~[ds]$ ] together in a single hadron , making an `` @xmath120 '' dibaryon whose mass would be below @xmath121 gev .
the absence of such a state means that one must be careful about correlations between quarks in different diquarks .
shuryak @xcite has noted that instanton interactions allow one to account for such pauli effects , and lead to a short - distance repulsive core ( 0.35 fm ) between diquarks .
qcd predicts that in addition to @xmath10 states there should be @xmath11 ( `` hybrid '' ) states containing a constituent gluon @xmath12 .
one signature of them would be states with quantum numbers forbidden for @xmath10 but allowed for @xmath11 .
for @xmath10 , @xmath122 , so @xmath123 . the forbidden @xmath10 states are then those with @xmath124 and @xmath125 .
a consensus in quenched lattice qcd is that the lightest exotic hybrids have @xmath126 and @xmath127 gev , @xmath128 gev , with errors 0.10.2 gev @xcite .
( unquenched qcd opens a pandora s box of mixing with @xmath30 and meson pairs . )
candidates for hybrids include @xmath129 ( seen in some @xmath130 final states , e.g. , in @xmath81 annihilations ) and @xmath131 ( seen in @xmath132 , @xmath133 , @xmath134 ) .
brookhaven experiment e-852 published evidence for a @xmath135 state called @xmath131 @xcite . a recent analysis by a subset of e-852 s participants @xcite , however , does not require this particle if a @xmath136 contribution is assumed .
such a state would be the regge recurrence of the pion , expected in a picture @xcite of a rotating qcd string .
cleo sees @xmath137 events above background of @xmath138 in @xmath139 pb@xmath140 of @xmath141 collisions at @xmath142 gev implying @xmath143 and @xmath144 mev @xcite .
this is consistent with lattice predictions , including one @xcite of @xmath145 mev . in the past couple of years
the lowest @xmath146 and @xmath147 @xmath148 states turned out to have masses well below most expectations . if they had been as heavy as the already - seen @xmath148 states with @xmath149 , the @xmath150 [ @xmath151 and @xmath152 [ @xmath153 ) , they would have been able to decay to @xmath154 ( the @xmath155 state ) and @xmath156 ( the @xmath147 state ) .
instead several groups @xcite observed a narrow @xmath157 decaying to @xmath158 and a narrow @xmath159 decaying to @xmath160 .
these decays violate isospin .
should we have been surprised ?
the selection rules in decays of these states show their @xmath15 values are consistent with @xmath155 and @xmath147 .
low masses are predicted @xcite if these states are viewed as parity - doublets of the @xmath161 and @xmath162 @xmath148 ground states in the framework of chiral symmetry .
the splitting from the ground states is 350 mev in each case .
the splitting of the lightest charmed - nonstrange @xmath155 and @xmath147 mesons from their @xmath163 and @xmath164 partners ( @xmath165 and @xmath166 ) is a bit larger , as anticipated in @xcite : @xmath167 mev , @xmath168 mev @xcite or @xmath169 mev , @xmath170 mev @xcite ; @xmath171 mev , @xmath172 mev @xcite or @xmath173 mev , @xmath174 mev @xcite .
the cleo collaboration is concerned with finding the best strategy for scanning @xmath175 in @xmath141 annihilations to optimize @xmath176 production .
one can invoke a simple regularity of resonance formation to suggest optimal energies . if a meson @xmath177 and a meson @xmath178 have quarks which can mutually annihilate , they form at least one resonance when @xmath179 mev/@xmath180 @xcite .
denote this resonance mass by @xmath181 .
@xmath182 states resonate at @xmath183 , below the predicted value @xmath184 gev .
@xmath185 + c.c . and
@xmath186 + c.c.resonances have been seen at 3872 and 3940 mev .
the @xmath187 is a satisfactory candidate for @xmath188 .
this regularity predicts at least one @xmath189 resonance below 4.0 gev and one @xmath190 + c.c .
resonance below 4.08 gev .
the @xmath191 ( see below ) is a candidate for a @xmath192 resonance from this standpoint .
detailed estimates of higher charmonium production have been performed in @xcite .
an old estimate of the cross section shape based on coupled - channel considerations @xcite would be very useful with updated @xmath193 and @xmath194
. an excited @xmath195 candidate is seen by the selex collaboration in @xmath196 ( dominant ) and @xmath197 modes @xcite . in the first mode ,
its yield is 40% that of the @xmath176 , which is surprising .
it is not seen by babar , belle , cleo , or focus @xcite .
doubly - charmed baryon candidates are also seen by selex .
they are produced by a ( 67% pure ) beam @xmath198 beam . a @xmath199 candidate is seen @xcite with @xmath200 mev , width resolution - limited , lifetime @xmath201 fs , decaying to @xmath202 .
the signal is 22 events above a background of 6.1 .
surprisingly , the @xmath203 yield is 40% that of @xmath204 . a weak selex signal ( 5.4/1.6 events ) in @xmath205
also is seen @xcite .
there is the claim of an excited doubly charged state @xmath206 decaying to @xmath207 , but in contrast to expectations for a strong decay the final state is not 100% @xmath208 .
no other experiments have reported any doubly - charmed baryon signals .
data on singly - charmed baryons continue to accumulate , following contributions by cleo and others .
the reported levels are summarized in fig .
[ fig : lexb ] .
the belle collaboration reports an excited @xmath209 candidate decaying to @xmath210 , with mass about 510 mev above @xmath211 @xcite
. it could be an @xmath149 excitation of the spin-1 isotriplet s - wave @xmath212 quark pair which , in the ground state with respect to the @xmath180 quark , forms the @xmath209 and @xmath213 .
it could have @xmath214 .
the value shown in fig .
[ fig : lexb ] is a guess , using the diquark ideas of @xcite .
belle also confirms a @xmath215 seen earlier by cleo . ,
scaledwidth=98.0% ] in fig .
[ fig : lexb ] the first excitations of the @xmath204 and @xmath216 are similar , scaling well from the first @xmath217 excitations @xmath218 and @xmath219 .
they have the same cost in @xmath220 , and their @xmath221 splittings scale as @xmath222 or @xmath223 .
higher @xmath204 states may correspond to excitation of a spin - zero @xmath119 $ ] pair to @xmath224 , leading to many allowed @xmath15 values up to @xmath225 . in @xmath209
the light - quark pair has @xmath226 ; adding @xmath149 allows @xmath227 .
cleo has recently remeasured mass differences and widths of the singly - charmed baryon @xmath228 @xcite .
mass splittings between the doubly - charged and neutral states are quite small , in accord with theoretical expectations .
the prediction of heavy quark symmetry that @xmath229 is borne out by the data , in which @xmath230 , @xmath231 .
the elusive @xmath232 state of charmonium has been observed by cleo @xcite via @xmath233 with @xmath234 . whereas s - wave hyperfine charmonium splittings are @xmath235 mev for 1s and @xmath236 - m(\eta'_c )
\simeq $ ] 48 mev for 2s levels , p - wave splittings should be less than a few mev since the potential is proportional to @xmath237 for a coulomb - like @xmath238 interaction .
lattice qcd @xcite and relativistic potential @xcite calculations confirm this expectation of a small p - wave hyperfine splitting .
one expects @xmath239 mev .
earlier @xmath240 sightings ( see @xcite for references ) , based on @xmath19 production in the direct channel , include a few events at @xmath241 mev seen in cern isr experiment r704 ; a state at @xmath242 mev , decaying to @xmath243 , reported by fermilab e760 but not confirmed by fermilab e835 ; and a state at @xmath244 mev , decaying to @xmath245 with @xmath246 , reported by e835 with about a dozen candidate events @xcite . in the cleo data ,
both inclusive and exclusive analyses see a signal near @xmath247 .
the exclusive analysis reconstructs @xmath248 in 7 decay modes , while no @xmath248 reconstruction is performed in the inclusive analysis .
the exclusive signal is shown on the left in fig .
[ fig : hc ] .
a total of 19 candidates were identified , with a signal of @xmath249 events above background .
the mass and product branching ratio for the two transitions are @xmath250 mev ; @xmath251 . the result of one of two inclusive analyses
is shown on the right in fig .
[ fig : hc ] .
these yield @xmath252 mev , @xmath253 . combining exclusive and inclusive results yields @xmath254 mev , @xmath255 .
the @xmath240 mass is @xmath256 mev below @xmath247 , barely consistent with the ( nonrelativistic ) bound @xcite @xmath257 and indicating little p - wave hyperfine splitting in charmonium .
the value of @xmath258 agrees with theoretical estimates of @xmath259 .
cleo has reported a new measurement of @xmath260 ev based on 14.4 fb@xmath140 of @xmath141 data at @xmath26111.30 gev @xcite .
the result is compatible with other measurements when they are corrected for cleo s new @xmath262 and @xmath263 .
the errors given are statistical , systematic , and @xmath264 ) .
one can average the cleo measurement with a corrected belle result @xcite to obtain @xmath265 ev . using the fermilab e835 value of @xmath266 mev @xcite ,
@xmath267 one finds @xmath268 hadrons ) = @xmath269 mev .
this can be compared to @xmath270 , taking account of qcd radiative corrections @xcite , to obtain @xmath271 .
cleo has reported a number of new results based on about 3 million @xmath272 decays .
the branching fractions for @xmath273 have been remeasured , with the results shown in table [ tab : incpsip ] @xcite .
the inclusive @xmath274 rates are above the world average @xcite values .
the decay @xmath275 serves as a key calibration for the @xmath276 measurement mentioned above .
.branching ratios for @xmath273 measured by cleo @xcite .
[ tab : incpsip ] [ cols="^,^,^",options="header " , ] cleo has recently reported results on @xmath277 partial widths , based on the exclusive process @xmath278 @xcite . the results are shown in table [ tab : psipprad ] . the exclusive analysis has no sensitivity to @xmath279 since @xmath280 is small .
although the @xmath281 partial width is expected to be high , it must be studied in the inclusive channel , which has high background , or by reconstructing exclusive hadronic @xmath279 decays .
even with the maximum likely @xmath282 , one thus expects @xmath283(2% ) . several searches for @xmath284 , including vp , @xmath285 , and multi - body final states , are under way .
the value of @xmath286 also is being re - checked .
two preliminary analyses @xcite find no evidence for any light - hadron @xmath287 mode above expectations from continuum production except @xmath288 .
upper limits on the sum of 26 modes imply @xmath289 ( light hadrons ) ] @xmath290 .
cross sections at 3.77 gev are consistent with continuum extrapolated from 3.67 gev , indicating no obvious signature of non-@xmath182 @xmath287 decays .
known modes account for at most a few percent of @xmath287 decays , to be compared with a possible discrepancy of 12 nb in @xmath291 or @xmath29220% . a value based on bes data @xcite , @xmath293 = ( 0.72\pm0.46\pm 0.62)$ ] nb
does not answer the question .
could a more careful treatment of radiative corrections solve the problem ?
a remeasurement of @xmath286 by cleo , preferably through an energy scan , is crucial .
many charmonium states above @xmath182 threshold have been seen recently .
the @xmath294 , discovered initially by belle in @xmath83 decays @xcite but confirmed by babar @xcite and in hadronic production @xcite , decays predominantly into @xmath295 .
evidence for it is shown in fig . [
fig : x3872 ] @xcite .
since it lies well above @xmath182 threshold but is narrower than experimental resolution ( a few mev ) , unnatural @xmath296 is favored .
it has many features in common with an s - wave bound state of @xmath297 with @xmath298 @xcite .
its simultaneous decay of @xmath294 to @xmath299 and @xmath300 with roughly equal branching ratios is a consequence of this `` molecular '' assignment .
analysis of angular distributions
@xcite in @xmath301 favors the @xmath302 assignment @xcite .
the detection of a small @xmath303 mode ( @xmath304 of @xmath305 ) @xcite confirms the assignment of positive @xmath306 and suggests some admixture of @xmath238 in the wave function . for the @xmath294 region @xcite .
[ fig : x3872],scaledwidth=94.0% ] the @xmath300 final state in @xmath307 shows a peak above threshold at @xmath308 mev @xcite .
this could be a candidate for an excited p - wave charmonium state , perhaps the @xmath309 .
the corresponding @xmath310 states @xmath311 have been seen to decay to @xmath312 @xcite .
a charmonium state distinct from this one but also around 3940 mev is produced recoiling against @xmath17 in @xmath313 and is seen to decay to @xmath314 + c.c .
but not @xmath300 .
since all lower - mass states observed in this recoil process have @xmath107 ( the @xmath315 and @xmath316 ; see fig .
[ fig : pakh ] @xcite ) , it is tempting to identify this state with @xmath317 ( not @xmath318 , which would decay to @xmath182 ) . in @xmath319 @xcite .
[ fig : pakh],scaledwidth=98.0% ] belle has recently reported a candidate for @xmath320 in @xmath321 collisions @xcite , decaying to @xmath182 .
the spectrum is shown on the left in fig .
[ fig : spect ] .
the angular distribution of @xmath322 pairs is consistent with @xmath323 as expected for a state with @xmath324 .
it has @xmath325 mev , @xmath326 mev , and @xmath327 , all of which are reasonable for a @xmath328 state . finally , babar reports a state @xmath191 produced in the radiative return reaction @xmath329 and seen in the @xmath330 spectrum @xcite .
its mass is consistent with being a @xmath331 level ( see @xcite for this interpretation ) since it lies about 230 mev above the @xmath332 candidate ( to be compared with a similar @xmath331-@xmath332 spacing in the @xmath0 system ) .
however , it could also be a hybrid state @xcite , as it lies roughly in the expected mass range , a @xmath333 state @xcite , or an effect associated with @xmath334 threshold .
the cdf collaboration has identified events of the form @xmath335 , allowing for the first time a precise determination of the mass : @[email protected]@xmath3371.1 mev @xcite .
this is in reasonable accord with the latest lattice prediction of 6304@xmath33712@xmath338 mev @xcite .
we are still waiting for the observation of @xmath339@xmath340 mixing at the expected level of @xmath341 ps@xmath140 : the current lower limit is @xmath342ps@xmath140 @xcite .
cleo data continue to yield new results on @xmath310 spectroscopy .
new values of @xmath343 = ( 2.39 \pm 0.02 \pm 0.07 , 2.03\pm0.03\pm0.08,2.39\pm0.07\pm0.10)\%$ ] @xcite imply lower values of @xmath344 , which will be important in updating comparisons with perturbative qcd .
the study of @xmath345 decays @xcite has provided new measurements of e1 transition rates to @xmath346 states .
searches in these data for the forbidden m1 transitions to spin - singlet states of the form @xmath347 have excluded many theoretical models .
the strongest upper limit , for @xmath348 , @xmath349 , is @xmath350 ( 90% c.l . ) .
searches for the lowest @xmath310 spin - singlet , the @xmath351 , using the sequential processes @xmath352 and @xmath353 @xcite are being conducted .
the direct photon spectrum in @xmath354 decays has been measured using cleo iii data @xcite .
the ratios @xmath355 are found to be @xmath356 , @xmath357 , @xmath358 .
@xmath359 is consistent with an earlier cleo value of @xmath360 .
the transitions @xmath361 have been observed for the first time @xcite .
one looks for @xmath362 in cleo data consisting of 5.8 million 3s events .
( see fig .
[ fig : chipipi ] . ) events with at least one detected soft pion are used . in the @xmath363 sample
7 events are seen above [email protected] background in the @xmath364 sample 17 events are seen above [email protected] background . assuming that @xmath365 ,
both are found equal to @xmath366 kev , which is in satisfactory agreement with theoretical expectations .
analysis of @xmath367 is in progress .
states illustrating the transitions @xmath368 .
[ fig : chipipi ] ]
we have made remarkable progress in sorting out level structures and transition rates for hadrons by appealing to their quark structure and using the underlying theory based on qcd .
the quarks and leptons themselves present a much greater challenge .
the underlying source of their masses and transitions , though these are being mapped out with impressive accuracy , is unknown . will the successes of qcd and the quark model repeat themselves at another level ? some suggest that quark and lepton masses and couplings are as random as the orbits of the planets in the solar system .
i hope they are wrong .
mendeleev spotted gaps in his periodic table by dealing out elements on cards .
are we playing with a full deck of quarks and leptons ?
does our `` periodic table '' have gaps ?
can we understand it as fundamentally as we understand mendeleev s ?
hadron spectroscopy continues to provide both long - awaited states such as @xmath240 and surprises such as low - lying p - wave @xmath176 mesons and @xmath294 , @xmath369 , @xmath370 , @xmath371 ( the @xmath23 state ) and @xmath191 .
we continue to learn how qcd works at strong coupling .
we may have evidence for molecules , @xmath332 and @xmath372 candidates , and @xmath331 or hybrid charmonium .
qcd may not be the last strongly coupled theory with which we have to deal .
electroweak symmetry breaking or the very structure of quarks and leptons may require us to apply similar or related techniques .
these insights largely come from experiments at the frontier of intensity and detector capabilities rather than energy , illustrating the importance of a diverse approach to the fundamental structure of matter .
i am grateful to j. appel , t. browder , f. close , i. danko , r. s. galik , b. golob , c. hearty , h. mahlke - krger , s. olsen , y. sakai , k. seth , j. shan , s. f. tuan , s. uehara , b. yabsley , s. ye , and j. yelton for sharing data and for helpful discussions .
this work was supported in part by the united states department of energy under grant no .
de fg02 90er40560 , and performed in part at the aspen center for physics .
i thank m. tigner for hospitality of the laboratory for elementary - particle physics at cornell and the john simon guggenheim foundation for partial support .
f. wilczek , hep - ph/0409168 , published in _ from fields to strings : circumnavigating theoretical physics : ian kogan memorial collection _
, edited by m. shifman _
et al . _ ,
1 ( world scientific , singapore , 2005 ) , pp .
77 - 93 ; a. selem , senior thesis , m. i. t. , 2005 ( unpublished ) .
b. aubert _ et al . _ [ babar collaboration ] , phys .
lett . * 90 * , 242001 ( 2003 ) ; d. besson _ et al . _
[ cleo collaboration ] , phys .
d * 68 * , 032002 ( 2003 ) ; k. abe _ et al .
_ [ belle collaboration ] , phys .
lett . * 92 * , 012002 ( 2004 ) . | some aspects of hadron spectroscopy are reviewed as of summer 2005 .
address = enrico fermi institute and department of physics , university of chicago , 5640 s. ellis avenue , chicago il 60637 usa |
sio maser sources are usually identified as oxygen - rich ( o - rich ) evolved stars , and it has been long thought that o - rich evolved stars do not significantly contribute to the enrichment of organic matter in the universe unlike carbon stars do .
however , recent observations of iras 19312 + 1950 , which is an sio maser source ( nakashima & deguchi 2000 ) , suggest that o - rich stars are non - negligible in the chemical evolution of the universe , because a rich set of molecular lines are detectable toward iras 19312 + 1950 .
if iras 19312 + 1950 is really an o - rich evolved star , its chemical properties are quite suggestive in terms of the interstellar chemistry , but we must securely identify its evolved - star status before going to further astrophysical / astrochemical interpretation , because some observational properties of this object are not consistent with the evolved - star status .
the extended infrared nebulosity of iras 19312 + 1950 was first realized in the 2mass images by nakashima & deguchi ( 2000 ) , and soon after that a more fine near - infrared image was obtained by uh 2.2 m / sirius .
the sirius imaging revealed that the envelope of iras 19312 + 1950 exhibits a point symmetric structure .
deguchi et al . ( 2004 ) and nakashima et al . (
2004 ) searched molecular millimeter lines toward iras 19312 + 1950 using the nro 45 m telescope . even though the molecular lines search was not unbiased survey ( i.e.
, selected lines were observed ) , they detected 22 different molecular species .
nakashima & deguchi ( 2005 ) made interferometric observations with bima in the bright molecular lines .
they concluded that the properties of the broad component seen in the line profiles are explained by an expanding sphere model .
maser emission detected from iras 19312 + 1950 enables us to investigate the kinematics in the vicinity to the central star if we use the vlbi technique .
we have recently made vlba observations in the sio ( @xmath00 , @xmath1 and 2 ) and h@xmath2o ( 6@xmath35@xmath4 ) maser lines .
in addition , we have made merlin observations in the h@xmath2o ( 22 ghz ) and oh ( 1612 mhz ) maser lines . in the h@xmath2o observations
, we clearly detected a double peak in the total flux line profile , and confirmed that the red- and blue - shifted components are seated in spatially separated regions ( see , figure 1 ) .
our current interpretation of iras 19312 + 1950 is a rare agb or post - agb star embedded in a small dark cloud , even though we have no clear idea about why an agb / post - agb star is embedded in a small dark cloud . a recent model calculation by murakawa et al . ( 2007 )
have also suggested that the mass loss rate of the central star is consistent with that of an agb star , but they have also withheld the discussion about the origin of the ambient material . to securely confirm the evolutional status of this object , presumably , observations listed below will be a key : ( 1 ) monitoring observations in maser lines with vlbi technique , ( 2 ) measuring a @xmath5c/@xmath6c ratio through , for example , radio observations in the h@xmath5co@xmath7 and h@xmath6co@xmath7 lines . | iras 19312 + 1950 is a unique sio maser source , exhibiting a rich set of molecular radio lines , although sio maser sources are usually identified as oxygen - rich evolved stars , in which chemistry is relatively simple comparing with carbon - rich environments .
the rich chemistry of iras 19312 + 1950 has raised a problem in circumstellar chemistry if this object is really an oxygen - rich evolved star , but its evolutional status is still controversial . in this paper , we briefly review the previous observations of iras 19312 + 1950 , as well as presenting preliminary results of recent vlbi observations in maser lines .
pdf file of the poster is available from http://www.geocities.jp/nakashima_junichi/ |
nonequilibrium processes , their stationary states and their phase transitions have been of considerable interest in natural science as well as in medicine and sociology for many years .
here we are interested in processes that can be modelled by growth and decay of populations with spatially local interaction rules .
the transition between survival and extinction of a population is a nonequilibrium continuous phase transition phenomenon and is characterized by universal scaling laws .
it is well known that also for systems far from equilibrium the concept of universality classes with respect to their critical properties in the vicinity of a continuous phase transition is applicable . for the description of transitions in systems that show active and absorbing inactive states , percolation models play an outstanding role .
some years ago it was conjectured @xcite that markovian growth models with one - component order parameters displaying a transition into an absorbing state in the absence of any special conservation law generically belong to the universality class of directed percolation ( dp ) .
besides dp @xcite this universality class includes e.g. reggeon field theory ( rft ) @xcite , the contact process @xcite , certain cellular automata @xcite and some catalysis models @xcite
( for a recent review of dp processes see @xcite ) . despite the fact that a large variety of different models belong to the dp universality class ,
_ there is still no experiment where the critical behavior of dp was seen _ @xcite . in a recent paper @xcite ,
hinrichsen compares suggested experiments and discusses possible reasons why the observation of dp critical exponents is obscured or even impossible .
one of these reasons might be that the basic feature of the dp class , the existence of an absorbing state , is quite difficult to realize in nature .
small fluctuations will always affect this state and may be strong enough to soften the transition like a small particle source , which works as an external field @xcite .
another reason might be the influence of spatial quenched disorder which is abundant in reality .
we have shown @xcite that in contrast to equilibrium systems , the critical scaling properties of dp processes are not only altered by frozen randomness , but fully destroyed .
for the analytic description of universal behavior near a critical nonequilibrium transition , it is often useful to model the universality class by mesoscopic stochastic processes involving the order parameter and other relevant fields . in case of the dp class a representation by the langevin equation for the time - development of the particle density ,
the gribov process ( the stochastic version of the so - called schlgl model @xcite ) , is appropriate .
the name gribov process was coined by grassberger who showed that rft is a markov process in disguise rather than a quantum theory @xcite . on the level of a formulation of stochastic processes by means of path integrals
, there is superficially no difference between rft and the gribov process .
however rft uses creation and annihilation operators for particles as the principal fields in contrast to the particle density and its conjugate response field in the gribov process .
microscopically the rft - description of dp starts with special reactions between diffusing individuals on a given @xmath1-dimensional lattice such as birth : @xmath2 , competition : @xmath3 , and death : @xmath4 .
these reactions are represented by a master equation that is mapped onto a second - quantized bosonic operator representation , which is in turn mapped onto a bosonic field theory using the continuum limit @xcite . at the critical point
the rates of birth and death have to balance to yield a vanishing overall production of individuals . but this condition leads to strong local correlations and the microstates of the system consist typically of clusters of individuals embedded in the vacuum @xcite .
thus , a fluctuating density description for the gribov process is appropriate on a mesoscopic level .
note that _ the replacement of the spontaneous single particle death reaction by a two particle reaction @xmath5 leads to strong anticorrelations because now the birth rate itself has to vanish at the critical point_. in this case , the microstates consist of separated lonely wanderers that only sporadically interact and no clustering of individuals sets in .
_ for such problems of branching and annihilating random walks one is forced to use the creation - annihilation operator formulation _ @xcite and a naive stochastic density description for the fundamental field would lead to wrong results . as a general rule : one can show that branching processes lead to positive correlations and annihilation processes to anticorrelations with the exception of spontaneous decay , which does not generate any correlations .
nonvanishing branching and spontaneous decay are needed to yield positive correlations together with the possibility of a vanishing overall production of particles at the critical point . in such cases
a mesoscopic density description is correct .
_ instead of considering only one species of particles as is usually done for processes belonging to the dp class , it is of interest to introduce processes with several interacting species as for instance in mathematical biology @xcite , which are also of relevance for a special model of surface growth @xcite .
it is the purpose of this paper to describe the detailed field - theoretic investigation based on renormalization group methods of such _ colored and flavored directed percolation processes _ ( mdp ) realized by the gribov process for several species .
we group different colored species in the same flavor class , if they have equal transport properties . in the next chapter
we introduce the model and define its renormalization and the one - loop calculation in the third chapter . in chapter iv , we present the general renormalization group analysis and find the asymptotic scaling behavior to one - loop order . in chapter
v , we include the two - loop results of the appendices and show the crossover to unidirectionality of the couplings between the different species
. in chapter vi , we present our considerations on symmetries and the general fixed point properties of the model .
we show that the permutation symmetry of a multicolored process is spontaneously broken .
a brief account of this work has been presented in @xcite .
in addition , in chapter vii , we will show that the universal multicritical features of a recently introduced model of unidirectional coupled directed percolation processes ( ucdp ) by tuber et al .
@xcite , which contains an additional linear coupling between the species , is completely described by the mdp class . in the last chapter
we discuss the results and give an application to biomathematics .
three appendices present technical details , e.g. the @xmath6-expansion of the dp - exponents , known for a long time , but as yet unpublished .
the mesoscopic description of the dynamics of physical systems is based on a correct choice of the complete set of fundamental slowly - developing fields . in general
these are the order parameter densities and the densities of conserved quantities .
the multispecies processes under consideration are completely described by the particle densities @xmath7 of the percolating colored and flavored individuals .
we assume that there does not exist any conservation law .
next one has to find out the general form of the stochastic equations of motions of the fundamental fields as timelocal ( markovian ) langevin equations .
these langevin equations have to respect symmetries and general principles characterizing the universality class under consideration .
the mdp - class is characterized by the following four principles : 1 .
errorfree self - reproduction ( `` birth '' ) and spontaneous annihilation ( `` death '' ) of individuals .
the rates for birth and death may be different for each color .
interaction between the individuals ( `` competition '' , `` saturation '' ) with color - dependent couplings . 3
. diffusion ( `` motion '' , `` spreading '' ) of the individuals in a @xmath1-dimensional space with flavor - depending transport coefficients .
4 . the states with at least one extinct color are absorbing . in the language of chemistry , the mdp
may be realized microscopically by an autocatalytic reaction scheme of the form @xmath8 , @xmath9 , @xmath10 , where the last reaction subsumes the interactions of individuals with colors @xmath11 , @xmath12 and @xmath13 may be the integers @xmath14 .
a description in terms of particle densities typically arises from a coarse - graining procedure where a large number of microscopic degrees of freedom are averaged out .
the influence of these is simply modelled by gaussian noise - terms in the langevin equation which however have to respect the absorbing state condition .
the stochastic reaction - diffusion equations for the particle densities in accordance with the four principles given above are of the form @xmath15 where the first term on the right hand side models the ( diffusive ) motion , and the @xmath16 are the overall reproduction rates of the particles with color @xmath11 .
these deterministic terms are constructed proportional to @xmath17 in order to ensure the existence of an absorbing state for each species . near the absorbing transition
the particle densities @xmath18 are small quantities . expanding the rates @xmath16 in powers of @xmath18 results in @xmath19 the gaussian noises @xmath20
must also respect the absorbing state condition , whence @xmath21 subleading terms in the expansions ( [ 2],[3 ] ) as well as additional terms with derivatives of the spatial @xmath22-function in the first line of eq .
( [ 3 ] ) are not displayed .
it can be shown that they are irrelevant in the renormalization group sense as long as the stability condition @xmath23 for all @xmath24 is fulfilled .
the breakdown of this condition signals the occurrence of a discontinuous transition to compact growth and the appearance of tricritical phenomena at the border between first and second order transitions .
such a behavior is expected in microscopic models based on more complicated particle reactions @xcite .
the `` temperature '' variables @xmath25 measure the difference of the rates of death and birth of the color @xmath11 .
thus the temperatures may be positive ore negative .
we are interested in the case where all @xmath26 ( up to fluctuation corrections ) which defines the multicritical region . under these conditions
all the species live on the border of extinction .
the next step is a mean field investigation of homogeneous steady state solutions of the equations of motion . neglecting all fluctuations in eq.([1 ] ) and using the expansion ( [ 2 ] ) up to first order in @xmath27
we find easily that all @xmath28 as long as all @xmath29 , meaning the vacuum is absorbing for each color .
as soon as some of the temperatures become negative , stationary solutions with @xmath30 emerge , satisfying the corresponding equations @xmath31 .
thus , in general a multitude of hypersurfaces of first and second order transitions exist in the phase space spanned by the relevant temperature variables @xmath32 which separate the phases where a specific color becomes extinct .
whenever @xmath33 changes from a negative to a positive value , the order parameter @xmath34 undergoes a continuous or a discontinuous phase transition from an inactive absorbed state with @xmath35 to an active state with @xmath30 .
all hypersurfaces of phase transitions meet in the multicritical point where all temperatures are zero .
all homogeneous states are globally stable because the evolution of the total particle density of homogeneous states in time is given by @xmath36 .
thus all solutions of the mean field equations of motion are bounded to a finite region in the space of positive @xmath34 as long as the stability condition mentioned in the foregoing paragraph holds . in the following we focus on the effect of fluctuations on the scaling behavior of correlation and response functions in the vicinity of the multicritical point where the strongly relevant parameters \{@xmath37 } are small . in order to apply field - theoretic methods and the renormalization group equation in conjunction with an @xmath38-expansion about the upper critical dimension @xcite , it is convenient to use the path - integral representation of the underlying stochastic processes @xmath39 @xcite . with the imaginary - valued response fields denoted by @xmath40 , the generating functional of the connected response and correlation functions , the greens functions , takes the form @xmath41=\ln \int { \cal d}\bigl[\widetilde{n},n \bigr]\exp \biggl(-{\cal j}\bigl[\widetilde{n},n\bigr]+\int d^{d}x\int dt \bigl(hn+\widetilde{h}\widetilde{n}\bigr)\biggr)\ .
\label{pathint}\ ] ] the response fields @xmath42 correspond to the conjugated auxiliary variables of the operator formulation of statistical dynamics by kawasaki @xcite and martin , siggia , rose @xcite ) .
the dynamic functional @xmath43 $ ] and the functional measure @xmath44 $ ] , which in symbolic notation is proportional to @xmath45 , is understood to be defined using a prepoint ( ito ) discretization with respect to time .
the prepoint discretization leads to the causality rule @xmath46 in response functions which forbids response propagator loops in the diagrammatical perturbation expansion @xcite . the langevin equations ( [ 1]-[3 ] ) are recast as a dynamic functional @xmath47 in the second line we have neglected subleading terms .
correlation and response functions can now be expressed as functional averages of monomials of the @xmath17 and @xmath48 with weight @xmath49 . a glance at eq .
( [ pathint ] ) shows that the responses are defined with respect to additional local particle sources @xmath50 in the langevin equations ( [ 1 ] ) .
a rescaling of the fields @xmath51 , @xmath52 leaves the functional @xmath53 forminvariant but transforms the coupling constants @xmath54 and @xmath55 .
thus invariant coupling constants are given by @xmath56 .
a suitable rescaling is defined by @xmath57 .
if we choose this normalization , we denote the rescaled fields by @xmath58 and @xmath59 . the scaling by a suitable mesoscopic length and time scale , @xmath60 and @xmath61 ( with @xmath62 )
respectively , leads to @xmath63 , @xmath64 where @xmath65 .
hence @xmath66 is the upper critical dimension .
it is now easy to show that all neglected possible subleading terms in the expansion of @xmath67 and the noise correlation , as well as higher gradient terms in the langevin equation ( [ 1 ] ) and non - gaussian and non - markovian noise correlations , have coupling constants with negative @xmath68-dimensions near the upper critical dimension .
therefore , under the renormalization group flow , they are renormalized to zero and we can safely neglect them because we are interested in the leading universal critical behavior .
these couplings can be reintroduced if one is interested in corrections to scaling . in a renormalized field theory context @xcite , the @xmath68-dimensions of the coupling constants are equal to the so called naive or engineering dimensions .
all coupling constants with positive naive dimensions ( the relevant couplings with respect to the gaussian fixed point as the starting point of the perturbation expansion ) need a renormalization because the corresponding vertex functions are the only ones which develop primitive divergencies in perturbation theory .
thus the field theory based on the functional @xmath69 ( eq . ( [ 4 ] ) ) is renormalizable and the calculated scaling properties are universal for the full class of mdp - processes .
now we are in a position to develop the perturbation theory in the inactive phase with all the @xmath29 . to begin with we separate the anharmonic `` interaction '' terms from the dynamic functional @xmath69 , eq .
( [ 4 ] ) , and retain only the harmonic ones : @xmath70 here we have included the external sources @xmath71 and @xmath72 . the gaussian path integral @xmath73\exp ( -{\cal j}_{0})=\exp ( ( h , g\widetilde{h}))$ ] involves only non - negative fields @xmath74 as long as @xmath75 .
it yields the propagators for the fourier transformed fields @xmath76 etc .
, as @xmath77 where the heaviside theta - function is defined with @xmath78 following from the ito - discretization of the path - integral and ensuring causality .
besides the propagators , the anharmonic coupling terms in @xmath53 , eq.([4 ] ) , define the elements of the graphical perturbation expansion .
they are depicted in fig . 1 where an arrow marks a @xmath79-leg and we draw diagrams with the arrows always directed to the left ( ascending time ordering from right to left ) .
1 . elements of the graphical perturbation expansion fig .
1 shows that the color of a @xmath79-leg on the left side of a vertex is not annihilated : at least one @xmath80-leg on the right side displays the same color .
thus , going from left to right , i.e. backward in time , through a diagram , colors have only sources and no sinks .
this property is a consequence of the existence of absorbing states in the model .
the perturbation expansion of a translationally invariant field theory can be analyzed by the calculation of the vertex functions @xmath81 corresponding to the one - particle irreducible amputated diagrams . here
the sets @xmath82 and @xmath83 denote the colors of the amputated outer @xmath79-legs and @xmath80-legs respectively .
going backward in time we conclude from the conservation property of the @xmath79-colors that the colors of the set @xmath84 must appear as a subset of @xmath83 .
therefore the only nonzero two and three point vertex functions are @xmath85 , @xmath86 , and @xmath87 . moreover
, another property follows directly from color conservation : the vertex function @xmath88 does not depend on parameters @xmath89 , @xmath90 , and @xmath91 with colors other than the ones of the vertex functions itself .
thus , @xmath85 , @xmath92 , and @xmath93 are only functions of the parameters @xmath94 , @xmath25 , and @xmath57 and are in particular independent from the interspecies couplings @xmath95 with @xmath96 .
therefore these vertex functions are the same as the corresponding functions of the well analyzed one - species gribov process .
to calculate them one can set the interspecies couplings @xmath97 .
in this case the model shows rapidity reversal invariance @xmath98 from which follows the equality @xmath99 .
now we are ready to consider the renormalization of the model .
it is known that the perturbation expansion of a field theory with a momentum cutoff @xmath100 develops divergencies if @xmath101 @xcite .
if the model is renormalizable one can absorb all these `` primitive divergencies '' order by order in a loop expansion in a suitable reparametrization of the parameters .
absorbing the primitive divergencies regularizes the full model .
the primitively divergent vertex functions have nonnegative naive @xmath68 -dimensions . here
, they are @xmath102 and @xmath103 @xmath104 , logarithmic at the upper critical dimension , ( fig . 2 ) .
fig . 2 .
primitively divergent vertex functions taking into account the general properties of the vertex functions found in the foregoing paragraph , we see that the following renormalization scheme renders the theory finite @xmath105 \tau _
{ \alpha } & \rightarrow & \mathaccent"7017{\tau } _ { \alpha } = \frac { z_{\tau } ^{(\alpha ) } } { z_{\lambda } ^{(\alpha ) } } { \tau } _ { \alpha } \ , \quad f_{\alpha \beta } \rightarrow \mathaccent"7017{f}_{\alpha \beta } = g_{\varepsilon } ^{\,-1}\mu ^{\varepsilon } \,\frac{z_{u}^{(\alpha \beta ) } } { z_{s}^{(\beta ) } z_{\lambda } ^{(\alpha ) } z_{\lambda } ^{(\beta ) } } \,u_{\alpha \beta } \ .
\label{7}\end{aligned}\ ] ] here @xmath106 is a convenient constant . instead of a momentum - cutoff regularization ,
we use dimensional regularization and minimal renormalization in the following . from the discussion above we learn that the renormalization factors @xmath107 with @xmath108 and @xmath109 , which are determined from @xmath85 and @xmath110 , are already known from the one - species gribov process and depend only on @xmath111 . thus @xmath112 .
the new renormalization factors @xmath113 with @xmath114 stem from the interspecies couplings .
they depend only on the couplings @xmath115 , @xmath116 , @xmath117 , @xmath118 , and the ratio @xmath119 .
we will now explicitly calculate the renormalizations to one - loop order .
the primitively divergent one - loop diagrams are shown in ( fig . 3 ) .
primitively divergent one - loop diagrams using dimensional regularization , we express the contribution of the self - energy diagram , fig .
3(a ) , as a function of external momentum and frequency , @xmath120 and @xmath121 : @xmath122 here , we have retained only terms linear in @xmath121 and @xmath123 .
these are the terms that display poles in @xmath124 . to determine the primitive divergencies of the vertex functions we can set external momenta and frequencies to zero and use equal temperatures @xmath125 as infrared regularisators .
the contributions of the three diagrams fig .
3(c , d , e ) add to @xmath126 from the zero - loop contributions and the results of our short calculation , eqs .
( [ 8],[9 ] ) , we obtain the ( as yet unrenormalized ) one - loop vertex functions to the desired order in @xmath121 and @xmath123 : @xmath127 and @xmath128 an explicit calculation of diagram ( b ) of fig .
3 demonstrates that indeed @xmath129 if @xmath130 . to absorb the @xmath6-poles in the renormalization @xmath131-factors we note that the vertex functions are renormalized by the scheme eq .
( [ 7 ] ) as @xmath132 using again the renormalization scheme eq .
( [ 7 ] ) , we find the renormalized vertex functions from eqs .
( [ 10],[11 ] ) as @xmath133 and @xmath134 therefore the vertex functions become finite by choosing @xmath135 up to higher orders in the coupling constants @xmath136 . as anticipated , for @xmath137 we have found the well known renormalization factors of the reggeon field theory .
next we will explore the scaling properties of multicolored directed percolation .
scaling properties describe how physical quantities will transform under a change of length scales . at the end of chapter ii
we have introduced the arbitrary mesoscopic length scale @xmath68 .
the freedom in the choice of @xmath68 , keeping the unrenormalized bare parameters @xmath138 , and in cutoff regularization the momentum cutoff @xmath139 fixed , can be used to derive in a routine fashion the renormalization group ( rg ) equation for the connected correlation and response functions , the green functions @xmath140 we denote @xmath68-derivatives at fixed bare parameters by @xmath141 . from @xmath142 and the renormalization scheme eq .
( [ 7 ] ) , which leads to @xmath143 , we then find the rg equations @xmath144g^{\{n,\widetilde{n}\}}=0
\label{17}\ ] ] with the renormalization group differential operator @xmath145 here we have introduced the gell - mann low functions @xmath146 with @xmath147 , and the wilson functions @xmath148 the rg equations ( [ 17 ] ) can be solved in terms of a single flow parameter @xmath149 using characteristics . following this method ,
flowing parameters are defined by the characteristic equations @xmath150 and the rg equations ( [ 17 ] ) of the green functions become @xmath151g^{\{n,\widetilde{n}\}}(\{{\bf x } , t\},\{\bar{\tau}(l)\},\{\bar{u}(l)\},\{\bar{\lambda}(l)\},l\mu ) = 0\ . \label{21b}\ ] ] here , the functions @xmath152 , @xmath153 , and @xmath154 are independent of the color @xmath155 and the interspecies couplings .
the flow equations ( [ 21a ] ) describe how the parameters transform if we change the momentum scale @xmath68 according to @xmath156 .
being interested in the infrared ( ir ) behavior of the theory , we must study the limit @xmath157 .
in general we expect that in this ir limit the coupling constants @xmath158 flow to a stable fixed point @xmath159 according to the first set of eq .
( [ 21a ] ) .
in particular , the intraspecies couplings @xmath160 then flow to a color independent fixed point @xmath161 because @xmath162 , and this gell - mann low function @xmath163 is equal to the corresponding function known from the one - species gribov process .
thus , the fixed point value @xmath164 is independent from any coupling to other species .
the solutions of the second and third set of the flow equations ( [ 21b ] ) are readily found in terms of the functions @xmath165 . in the ir limit @xmath166
they have the scaling form @xmath167 where the @xmath168 are nonuniversal amplitude factors .
the scaling exponents @xmath169 are already known from directed percolation . from eq .
( [ 21b ] ) we find the solution in the ir limit as @xmath170 with @xmath171 , @xmath172 , and the dp anomalous field scaling exponent @xmath173 omitting the nonuniversal amplitude factors @xmath168 and taking into account dimensional analysis in space and time @xmath174 where @xmath175 , we get by combination of eqs .
( [ 24],[26 ] ) the asymptotic scaling form of the green functions @xmath176 this has the important consequence that all scaling properties of the dp processes remain unaffected by the introduction of many colors .
moreover , all intraspecies green functions are completely independent of the coupling between the species , which follows from the absorbing state conditions for each color . however , the interspecies coupling constants @xmath177 with @xmath178 determine the properties of the interspecies scaling functions
. therefore we will now consider the consequences of the flow equations for these parameters .
for this purpose we need the gell - mann low functions @xmath179 explicitly . from the last of eqs .
( [ 19 ] ) we know that each of these functions begins with the zero - loop term @xmath180 , and the higher order terms are determined by the wilson functions .
these functions , the logarithmic derivatives of the @xmath131 -factors , are given by @xmath181 .
in minimal renormalization the @xmath131-factors have a pure laurent expansion with respect to @xmath6 : @xmath182 .
thus recursively in the loop expansion the wilson functions also have a pure laurent expansion and , because they are finite for @xmath183 , this expansion reduces to the constant term , i.e. all @xmath6-poles have to cancel in the logarithmic derivation .
hence , we obtain the wilson functions simply from the formula @xmath184 .
now it is easy to get these functions from the one loop results of the @xmath131-factors eqs .
( [ 15 ] ) .
we find to this order @xmath185 from which one finds the gell - mann low functions eq .
( [ 19 ] ) as @xmath186 the gell - mann low functions of the intraspecies couplings , the `` diagonal '' @xmath163-functions , follow as @xmath187 giving the stable fixed point values @xmath188 . with the help of eqs.([19],[23],[25 ] ) , this stable fixed point leads to the well known one loop order dp exponents @xmath189 , @xmath190 , and @xmath191 . using the fixed point values
@xmath192 to obtain the gell - mann low functions of the interspecies couplings of a pair of colors @xmath114 , we get @xmath193 in addition to the unstable decoupled fixed point values @xmath194 , the equations @xmath195 are solved by a fixed point line @xmath196 this equation is the key result of this section . clearly , however , the one - loop calculation can not give us any information whether the degeneracy of all the points on this line is fundamental to our model or is lifted by higher loop corrections .
thus , we must proceed to two - loop order , presented in the following chapter .
we note , however , that there are two special points of unidirectional coupling on the fixed line : @xmath197 in the next chapter we will show that for colors with the same flavor , meaning @xmath198 , these two points are indeed the only stable points on the line . because @xmath199 or @xmath200 are fixed point values at any loop order , we conjecture that unidirectionality is generic for the asymptotic behavior of coupled dp processes , irrespective of whether colors belong to different flavors or not .
from the analysis in the foregoing chapters we know that the interaction of two colors does not depend on the existence of other ones .
thus in this chapter we consider a coupled model of two colors @xmath201 and @xmath202 with the same flavor , i.e. equal kinetic coefficients @xmath203 and intraspecies couplings @xmath204 .
in contrast , the temperatures @xmath25 may be different .
thus , the ( unrenormalized ) dynamic functional is now given by @xmath205 note that in the case @xmath206 , the dynamics of species @xmath207 completely decouples from species @xmath202 and vice versa .
it follows that @xmath208 , if @xmath209 and @xmath114 .
the detailed calculation of the two - loop contributions is presented in appendices a to c. adding eq .
( [ a9a ] ) to the zero- and one - loop parts of the selfenergy eq .
( [ 10 ] ) , we get after renormalization , using eq .
( [ 12 ] ) and the scheme eq .
( [ 7 ] ) , @xmath210 where the constants @xmath211 are given in eq .
( [ a9b ] ) , and @xmath212 .
note that in comparison with eq .
( [ 13 ] ) now the one - loop terms acquire renormalizations to @xmath213 in order to be consistent in @xmath214 of the perturbation expansion .
these renormalized one - loop terms are needed to compensate non primitive singular terms proportional to @xmath215 arising now by the @xmath38-expansion of @xmath216 .
of course only primitive uv - divergencies , which means here @xmath217-independent ones , have to be regularized by renormalization to make the theory finite . eliminating the @xmath218-poles from eq .
( [ 35 ] ) by the @xmath131-factors , we find @xmath219 in the same way as for the selfenergy we find the other vertex functions .
renormalizing eq .
( [ 11 ] ) and adding the two - loop contribution eq .
( [ a22 ] ) we get @xmath220 where @xmath221 . here
the @xmath6-expansion also shows that the non primitive divergencies @xmath222 cancel .
we eventually find @xmath223 & & + \frac{1}{16\varepsilon } \biggl(\biggl(\frac{36}{\varepsilon } -\frac{19}{2 } \biggr)u^{2}+\biggl(\frac{8}{\varepsilon } -\frac{5}{4}\biggr)uu_{\beta \alpha } + \biggl(\frac{2}{\varepsilon } -1\biggr)u_{\alpha \beta } u_{\beta \alpha } \nonumber \\ & & + \biggl(\frac{8}{\varepsilon } -\frac{9}{4}+3\ln \frac{4}{3}\biggr ) uu_{\alpha \beta } + \biggl(\frac{1}{\varepsilon } -\frac{3}{2}\ln \frac{4}{3 } \biggr)\bigl(u_{\alpha \beta } ^{2}+u_{\beta \alpha } ^{2}\bigr)\biggr ) + o(u^{3})\ , \label{38}\end{aligned}\ ] ] and in particular @xmath224 we are now in a position to calculate the renormalization group functions eq .
( [ 19 ] ) from eqs .
( [ 36],[38],[39 ] ) with the help of eq .
( [ 20 ] ) .
they are given by @xmath225 and @xmath226 setting @xmath227 , we find the nontrivial stable fixed point value @xmath228 from which the scaling exponents eqs .
( [ 23],[25 ] ) of the green functions eq .
( [ 27 ] ) follow to second order of the @xmath6 -expansion as @xmath229 the first two expansions have been known for a long time from reggeon field theory @xcite where a different definition of the exponents is used .
the expansion of the exponent @xmath230 was presented by the author in @xcite .
the order parameter exponent @xmath163 , which enters the scaling law for the mean particle number in the active state @xmath231 we now turn to the interspecies coupling constants @xmath177 with @xmath232 .
the fixed point values as the solutions of the equation @xmath233 , where @xmath179 is the gell - mann low function eq .
( [ 40a ] ) , are of three different types : 1 .
the decoupled fixed point @xmath234 , totally instable for @xmath235 ; 2 .
the two unidirectional coupled fixed points @xmath236 , @xmath237 and @xmath238 , @xmath239 ; 3 .
the symmetric fixed point @xmath240 . to discuss the stability and the crossover between the last two types we try an ansatz of the form @xmath241 with @xmath242 .
we already know from the one - loop result that @xmath177 is driven by the renormalization flow to the fixed line @xmath243 with a crossover exponent @xmath244 . setting therefore @xmath245 in @xmath179 we get @xmath246 for @xmath247 this equation shows that the symmetric fixed point @xmath248 is instable in contrast to the stable unidirectional coupled fixed points @xmath249 .
the solution of the flow equation @xmath250 leads to the crossover @xmath251 with the very small crossover exponent @xmath252 a qualitative picture of the flow of the interspecies couplings in the plane @xmath247 under renormalization is shown in fig .
4 . fig .
flow of the interspecies couplings under renormalization in addition to the four fixed points d ( decoupled ) , s ( symmetric ) , u ( unidirectional ) , the topology of the flow is determined by the symmetry line @xmath253 which acts in the first quadrant as a separatrix between the regions of attraction of the two unidirectional fixed points .
there exists another separatrix at the border of these regions , given to first order in @xmath6 by @xmath254 , where the flow is driven to the hatched line @xmath255 .
on the left of this line we have the region of instability i , in which the condition @xmath256 for all positive @xmath257 is violated .
therefore we conjecture that interspecies couplings with @xmath258 ultimately lead to first order transitions . in fig .
4 the fixed point line of the one - loop calculation is also shown .
this line is the support of the slow crossover to the unidirectional fixed points which therefore describe the ultimate critical behavior of the mdp universality class .
we conjecture that in the case of colors with different flavors , @xmath259 , all properties are smooth functions of the ratio @xmath119 as long as this ratio is sufficiently close to one . thus generalizing to different flavors , the topology of the renormalization flow displayed in fig .
4 should only be smoothly deformed but not destroyed . in particular , the unidirectional coupled fixed points u should be stable also for @xmath260 .
having found the detailed fixed point structure of the model eq.([33 ] ) of two colors with the same flavor in the two - loop approximation , we will now investigate which results are valid to all orders of perturbation theory . according to the considerations of the third chapter , one demonstrates easily that for @xmath114 the vertex function @xmath261 if @xmath262 .
thus the lines of unidirectionally coupled models in fig . 4 , namely @xmath263 or @xmath264 , respectively , are invariant under the renormalization flow .
trivially , the decoupled fixed point d is the intersection point of these lines .
for @xmath265 the dynamic functional @xmath53 eq .
( [ 33 ] ) possesses the symmetry @xmath266 , @xmath267 , and @xmath268 from which we find that @xmath269 .
thus , these two vertex functions need the same @xmath131-factor : @xmath270 .
it follows that the symmetry line in fig . 4 , @xmath253 , is invariant under renormalization .
is there a condition that determines the crossover line ?
the answer is yes . we change to variables corresponding to the total and relative particle numbers , respectively , @xmath271 such linear transformations do not alter the measure of the functional integrals , and the dynamic functional changes , in the special case @xmath272 , to @xmath273 we see that in the case @xmath274 the dynamics of the total particle number decouples from the dynamics of the relative one in the sense that all vertex functions containing @xmath79- , but no @xmath275 -legs are zero if @xmath276-legs are attached .
in particular @xmath277 and @xmath278 , which leads to the renormalized interspecies couplings with @xmath279 and determines the crossover line to all loop - orders .
note that on the symmetry line @xmath280 holds , and the functional @xmath53 is then invariant against @xmath281 , @xmath282 .
the different fixed point values of the interspecies couplings are now fully determined as the intersection points of the several invariant lines ( see fig .
4 ) : the unidirectional lines @xmath263 , @xmath264 , the symmetry line @xmath253 , and the crossover line @xmath279 .
thus , we find to all orders for the symmetric fixed point s : @xmath283 and for the stable unidirectional fixed points u : @xmath284 @xmath285 and @xmath286 @xmath287 , respectively
. these are of course the results that we have found explicitly in the two - loop approximation . the unidirectionally coupled model , which describes the generic scaling properties of the mdp processes , exhibits another symmetry . using @xmath206 and @xmath288
, we write the dynamic functional of this model in the form @xmath289 here we have introduced a further harmonic unidirectional coupling @xmath290 , which corresponds to an additional linear term : @xmath291 in the langevin equation for species @xmath202 . this term was first considered by tuber et al .
@xcite in their study of the nonequilibrium critical behavior in unidirectionally coupled dp processes . such a term
does not alter the general renormalizations as long as the corresponding composed field @xmath292 is treated as a soft insertion .
of course , @xmath293 is a relevant parameter like the temperatures @xmath25 and needs its own renormalization factor @xmath294 determined by @xmath295 in such a way that the renormalized vertex function with an insertion @xmath296 is finite .
the settings @xmath297 define the multicritical point . from reggeon field theory
one knows a transformation called rapidity reversal : @xmath298 , which is broken by the dp transition to an active state .
here we generalize it to read @xmath299 under this transformation the functional @xmath300 changes to @xmath301 we learn from this relation that in the case @xmath302 , @xmath303 gains a higher symmetry at the multicritical point , which is not destroyed by renormalization .
it follows in this case that @xmath293 renormalizes in the same way as the @xmath25 , which implies @xmath304 , and that the renormalized couplings are related by @xmath305 at the fixed point , as we already know from above . from these considerations
follows that the crossover exponent , which is defined by the scaling invariants @xmath306 , is given by @xmath307 .
recently tuber et al .
@xcite ( in the following abbreviated by thhg ) introduced a general unidirectionally coupled dp process with species - independent diffusion coefficients , that was motivated by a study of alon et al .
@xcite on a nonequilibrium growth model for adsorption and desorption of particles which displays a roughening transition .
the thhg - model reads in our dynamic functional language as @xmath308 in contrast to thhg , we have introduced an additional cross diffusion term @xmath309 here that is indispensable for a complete renormalization of the general model . in physical terms
, coarse graining will always produce cross diffusion in this coupled model .
but coarse graining does more : it also produces a term proportional to the time derivative of the density of the first species in the langevin equation of the second one besides further irrelevant couplings .
accordingly , in the functional @xmath310 a term proportional to @xmath311 arises .
however , such a term can be eliminated by a suitable redefinition of the fields @xmath312 so that the harmonic parts with the time derivatives in the dynamic functional remain diagonal .
note that in the special case @xmath313 , @xmath314 the functional @xmath310 eq .
( [ 53 ] ) is identical to @xmath303 , eq . ( [ 49 ] ) , from which we know that it is fully renormalizable , and , in particular , in the case @xmath315 by the @xmath131-factors eqs .
( [ 36],[39 ] ) . in the following
we will demonstrate that the asymptotic properties of the thhg - model indeed belong to the universality class described by @xmath300 .
it is well known @xcite that the infinitesimal generators of a continuous transformation of the fundamental fields , which leads to a forminvariance of the describing statistical functional , define redundant operators ( composite fields ) . in the case that these redundant operators are relevant or marginal , they unnecessarily contaminate the renormalization group
. therefore they should be avoided from the outset . here
we introduce a linear , homogeneous transformation between the fields , which does not change the form of the functional @xmath310 eq.([53 ] ) .
let us call it the @xmath11-transformation : @xmath316 note the difference to the @xmath163-transformation eq .
( [ 54 ] ) , which is exploited to eliminate a @xmath317-term from @xmath318 .
this @xmath11-transformation leads to new coupling constants that we denote with a bar : @xmath319 now we renormalize the thhg - model by the scheme @xmath320 where @xmath321 , @xmath322 , and the @xmath323 with @xmath324 are given by eqs .
( [ 36],[39 ] ) . besides @xmath136 ,
dimensionless coupling constants are defined as @xmath325 . the scheme eq.([57 ] )
is chosen in such a way that the renormalized dynamic functional reads @xmath326 note that the counter term @xmath327 serves to cancel primitive divergencies arising in the vertex function @xmath328 .
in dimensional regularization and minimal renormalization the counterterms @xmath329 and @xmath330 are given by series in @xmath331 beginning with simple poles @xmath332 and @xmath333 , respectively .
we have formerly shown that only the residua of these poles determine the renormalization group functions .
it is now appropriate to define @xmath11-transformation invariant dimensionless coupling constants as @xmath334 the somewhat lengthy but simple calculation of all the one - loop renormalizations leads to the gell - mann low functions of the renormalization group equation . in part , in the case @xmath335 , they can be derived from the results of thhg @xcite .
if we set @xmath136 to its fixed point value @xmath336 and define as usual @xmath337 , where @xmath338 is any of the coupling constants , we obtain for the @xmath163 -functions of the invariant couplings @xmath339 the function @xmath340 results from the additive renormalization @xmath341 and mixes the renormalized fields in a correlation or response function under application of the renormalization group as @xmath342\bigl\{\tilde{s}_{1},s_{1 } , \tilde{s}_{2},s_{2}\bigr\}=\bigl\{-a\tilde{s}_{2},0,0,-as_{1}\bigr\}\;. \label{64}\ ] ] here @xmath343 is the renormalization group differential operator now given by @xmath344 with @xmath345 @xmath346 , and eq .
( [ 64 ] ) acts on green functions .
the @xmath347 result from the additive renormalizations @xmath330 .
the function @xmath348 is found to be @xmath349 the new renormalizations yield @xmath350 in order to determine the fixed - point solutions of eqs .
( [ 63 ] ) , @xmath351 , we impose the condition @xmath352 .
this yields @xmath353 , with @xmath354 ( unstable ) or @xmath355 ( stable ) .
it can easily be checked that these solutions are consistent with the full set of eqs .
( [ 63 ] ) and ( [ 66 ] ) .
these are the solutions found by thhg @xcite .
note that on the fixed point lines generated from the stable fixed point by the @xmath356-transformation a minimally coupled fixed point with @xmath357 is found .
now one has to prove stability of the fixed points of the full equations ( [ 63 ] ) without using the constraints @xmath352 .
a linearization about @xmath358 and either @xmath359 or @xmath360 shows that the flow of @xmath361 is unstable for @xmath359 , whereas it shows full stability of the fixed point for @xmath362 .
this vindicates the neglect of @xmath363 and the corresponding counterterms @xmath341 and @xmath364 in @xcite but only at the stable fixed point line , which is generated from the fixed point by the @xmath11-transformation .
however without further knowledge this statement is only correct in the one - loop calculation and could be violated in higher loop orders .
we will show that the stable fixed point is given by @xmath365 , @xmath355 to all orders of the loop expansion . as a consequence , the fixed point values @xmath366 are zero , and @xmath367 .
this leads to a crossover exponent @xmath368 where @xmath369 determines the scaling of @xmath370 .
indeed , we see from eq .
( [ 59 ] ) that the stable one - loop order fixed point belongs , up to an @xmath11-transformation , to the dynamic functional @xmath318 , eq .
( [ 58 ] ) , with coupling constants @xmath371 and @xmath372 , i.e. model @xmath373 , eq .
( [ 53 ] ) , with the additional constraint @xmath315 . above it was shown that this equality leads to rapidity reversal eq.([51 ] ) as a higher symmetry .
this higher symmetry is preserved under renormalization and , because @xmath300 is fully renormalizable , we have the result @xmath374 , @xmath375 and @xmath307 to all loop orders .
computations based on the dynamic functional @xmath300 , eq .
( [ 53 ] ) , are much easier to perform than calculations using the complete model @xmath376 , eq .
( [ 58 ] ) .
thus it may be possible to find the equation of state for @xmath377 to second order and check the assumptions made in @xcite on the reexponentiation of logarithms to yield the new order parameter exponent @xmath378 of that paper ( for a calculation of the equation of state for @xmath379 to two - loop order see @xcite ) .
the model @xmath300 describes the coupled dp processes near the multicritical point @xmath380 . what is needed for a thorough calculation of @xmath378
is a theory that comprises the limit @xmath381 .
therefore our considerations here do not solve the problem addressed in @xcite , namely the determination of the scaling exponent @xmath378 that controls the scaling @xmath382 where species @xmath207 is in its active phase .
thhg calculate @xmath383 by reexponentiation of logarithms .
( we have done a recalculation and find a slightly different value @xmath384 .
the difference arises from a subleading term resulting from the ominous peculiar diagram fig .
9(c ) in @xcite .
) however the approach of thhg relies on the assumption that simple reexponentiation is possible . to derive such scaling properties faithfully ,
one indeed has to solve the crossover problem @xmath385 which ( possibly ! ) induces a new scaling at infinity for the correlations of species @xmath202 .
some features of this crossover remind us of the crossover from special to ordinary behavior in the theory of surface transitions @xcite , with @xmath293 corresponding to the surface enhancement @xmath276 and the species @xmath207 and @xmath202 corresponding to the bulk and surface respectively .
the crossover problem of interest here is thus as yet unsolved .
in this paper , we have studied multicolored directed percolation processes ( mdp ) . in particular , we have shown that the scaling behavior of these coupled dp processes near their absorbing state transition is determined by the same critical exponents as known from simple one - species dp and therefore is independent from the number of colors .
the characteristic asymptotic feature of mdp shows up in the asymptotic unidirectional coupling of each pair of colors .
a special result of our analysis is the very slow crossover to this asymptotic unidirectionality which may be seen in computer simulations .
the unidirectional behavior of the couplings of an interacting population is summarized in the following graphical picture .
consider a graph where each node represents one color .
the stable fixed points are then represented by the so - called tournaments , that are the complete graphs with directed edges @xcite .
the directed edge from color @xmath11 to color @xmath163 stands for the influence of @xmath386 on @xmath163 in the respective equations of motion . in particular , for a population of three species there exist two different tournaments ( up to permutation of the colors ) which we call `` cyclic '' and `` hierarchic '' , fig . 5 . fig
cyclic and hierarchic tournament of three species one gets a graphical picture for the not fully stable fixed points either by deleting the directionality of the edges between the nodes for symmetric couplings of the corresponding colors , or by completely deleting the edges for the uncoupled pairs ( uncomplete tournaments ) . to get a simple qualitative impression of the behavior of the dynamic system of a population corresponding to a tournament of colors with the same flavor , we consider , in the active region with all@xmath387 , a renormalized mean - field theory of the equations ( [ 1],[2 ] ) for spatially homogeneous densities @xmath388 and set @xmath389 with @xmath390 .
we redefine the time scale @xmath391 and get @xmath392 then the decision , which of all the @xmath393 equal @xmath394 , defines the tournament .
the edge between a pair of species is directed from @xmath163 to @xmath11 if @xmath395 and vice versa .
the directions of the edges therefore represent the unidirectional `` pressure '' on the reproduction rate resulting from one color to another . despite the simplicity of the eqs .
( [ 68 ] ) they can generate a complex dynamic behavior ( see e.g. @xcite ) . first , let us consider the stationary states of eqs .
( [ 68 ] ) for a population consisting of three species . in the ternary phase diagram spanned by the positive rates
@xmath396 in the subspace @xmath397 , one finds different regions with one , two ore all three species alive , fig . 6 .
fig . 6 .
phase diagrams of the cyclic and hierarchic tournament these regions are bounded by critical lines with absorbing state transitions where some colors become extinct .
the dynamic behavior of the hierarchic tournament fig .
5(b ) is relatively simple : from each nonequilibrium initial state @xmath398 , the system relaxes to the stationary state which is a stable node . in the case of the cyclic tournament fig .
5(a ) , for rates @xmath399
such that we have a three species stationary state , one also finds regions for which the ultimate relaxation behavior is characterized by attracting nodes .
but for rates which lead to stationary states belonging to the crosshatched area in the ternary diagram fig . 7 spanned by the @xmath400 , we find stable spirals ( damped cyclic relaxation ) as the ultimate relaxation behavior ( qualitatively pictured by the trajectory in fig
. 7 ) .
fig . 7 .
dynamic behavior of the cyclic tournament for stationary points near the middle of fig . 7 the damping is very small and cyclic behavior dominates the dynamics . especially in the middle of the diagram , i.e. for equal rates @xmath401
, the motion is not damped anymore . in this case , the dynamic system eq .
( [ 68 ] ) is known as a special form of the may - leonard model @xcite that has been extensively studied in mathematical biology @xcite .
finally , after a relaxation in the plane @xmath402 , there exists another constant of motion @xmath403 and the dynamic behavior is characterized by limit cycles around the neutral stationary point @xmath404 .
summarily we have shown that also in stochastic multispecies models of populations that evolve near the extinction threshold of all colors and therefore have many absorbing states , the critical properties at the multicritical point and at all continuous transitions are governed by the well known gribov process ( reggeon field theory ) exponents ( for a previous simulational result on a two - species system which seems to agree with our findings see @xcite ) . in other regions of the phase diagram of course more complicated critical behavior may arise such as multicritical points with different scaling exponents @xcite .
the models considered here have many absorbing states ( each combination of colors may go extinct irrespective of the other ones ) , and therefore are different from models which were considered by grinstein et al .
@xcite where it was shown that multispecies systems with one absorbing state belong to the gribov universality class .
in addition we have shown that the universal properties of interspecies correlations and the phase diagram are determined by totally asymmetric fixed point values of the renormalized interspecies coupling constants .
this eventually leads to a system working cooperatively .
it is interesting that the asymmetry between the species seems to be the condition for this cooperation near extinction .
the model considered here is a simple but universal model of such a cooperative society and should therefore have many applications in all fields of natural and even social science .
competition and extinction is of course a subject much considered in theoretical biology .
the main differences between the present work and the topics covered e.g. in the monograph of hofbauer and sigmund @xcite are the more realistic local description of the interactions between the species , their diffusional motion , and the inclusion of local fluctuations .
thus here the equations of motion are local stochastic partial differential equations .
coarse graining and renormalization lead to an universal macroscopic picture of cooperativity near the critical states of extinction .
we thank uwe tuber for many fruitful discussions , stephan theiss for a critical reading of the paper and beate schmittmann for numerous valuable remarks leading to the final version of the manuscript .
this work has been supported in part by the sfb 237 ( `` unordnung und groe fluktuationen '' ) of the deutsche forschungsgemeinschaft .
in the calculation we will encounter momentum integrals of the type @xmath405 where @xmath406 , @xmath407 , and @xmath408 .
they can be derived from two `` mother '' integrals @xmath409 by taking derivatives with respect to the parameters @xmath410 .
discarding nonsingular terms , we find in dimensional regularization @xmath411 these formulas yield the singular parts of the integrals ( [ a1 ] ) as @xmath412
fig . 8 .
two - loop selfenergy diagrams in fig .
8 the two - loop selfenergy diagrams are drawn .
diagram fig .
( 8a ) leads to @xmath413 where @xmath414 is the propagator ( eq .
( [ 6 ] ) ) and we always set @xmath415 as an ir regulator . noting that e@xmath416e@xmath417 e @xmath418e@xmath419 factorizes in the integral , the time integrations over the intervals @xmath420 , @xmath421 , @xmath422 are easily performed . the expansion in @xmath423 and @xmath424 to linear order eventually yields @xmath425 where the @xmath426 denote the integrals defined in eq .
( [ a1 ] ) . extracting the singular parts using eq .
( [ a4 ] ) , we obtain @xmath427 in the same way we calculate the second selfenergy diagram fig
. 8(b ) : @xmath428 up to higher orders in @xmath423 and @xmath424 . extracting the singular parts again , we find @xmath429 summing up
, we finally get from @xmath430 and @xmath431 the two - loop contribution of the selfenergy as @xmath432 where @xmath433
fig . 9 . two - loop vertex diagrams ; the numbers show the other time orderings fig .
9 presents the eleven two - loop vertex diagrams .
they can be calculated by the same method as the selfenergy diagrams . here
the external frequencies and momenta can be set to zero because the expansion in these variables does not lead to primitive divergencies . as an example
we show the calculation of the diagram fig .
9(i ) explicitly .
the symmetry factor of the diagrams is one , but one also has to add the two diagrams arising from the interchange of the color indices @xmath163 and @xmath434 .
after the factorization of the propagators , the integration over the time intervals , indicated by the broken lines , is trivial und leads to the expression ( with the abbreviations @xmath435 , @xmath436 ) @xmath437 thus we obtain the contribution of the diagram fig .
9(i ) to the vertex function @xmath87 : @xmath438 m. doi , j. phys .
a : math .
* 9 * , 1479 ( 1976 ) ; p. grassberger and p. scheunert , fortschr .
* 28 * , 547 ( 1980 ) ; l. peliti , j. phys .
( france ) * 46 * , 1469 ( 1984 ) ; b. p. lee , j. phys . a : math . gen .
* 27 * , 2633 ( 1994 ) .
u. c. tuber , m. j. howard , and h. hinrichsen , phys .
lett . * 80 * , 2165 ( 1998 ) ; y. y. goldschmidt , phys .
* 81 * , 2178 ( 1998 ) ; y. y. goldschmidt , m. j. howard , h. hinrichsen , and u. c. tuber , phys .
e * 59 * , 6381 ( 1999 ) .
k. kawasaki , in _ proc .
varenna summer school on critical phenomena _ , edited by m. s. green , ( academic , new york , 1971 ) ; in _
phase transitions and critical phenomena _ , vol .
* 5a * , edited by c. domb and m. s. green , ( academic , london , 1976 ) . | a model of directed percolation processes with colors and flavors that is equivalent to a population model with many species near their extinction thresholds is presented .
we use renormalized field theory and demonstrate that all renormalizations needed for the calculation of the universal scaling behavior near the multicritical point can be gained from the one - species gribov process ( reggeon field theory ) . in addition
this universal model shows an instability that generically leads to a total asymmetry between each pair of species of a cooperative society , and finally to unidirectionality of the interspecies couplings .
it is shown that in general the universal multicritical properties of unidirectionally coupled directed percolation processes with linear coupling can also be described by the model .
consequently the crossover exponent describing the scaling of the linear coupling parameters is given by @xmath0 to all orders of the perturbation expansion .
as an example of unidirectionally coupled directed percolation , we discuss the population dynamics of the tournaments of three colors .
key words : multicolored directed percolation , field - theoretic renormalization group , stochastic population dynamics , |
most of the pulsar spin - down power is carried away by a relativistic , magnetized wind .
the pulsar wind injects this energy into the surrounding nebula in the form of relativistic electron - positron pairs and magnetic fields therefore such nebulae , or plerions , emit synchrotron radiation from the radio to the gamma - ray band .
the most famous and well studied example of the plerion is the crab nebula ; the spectrum of this source is measured from about 10 mhz to dozens tev . according to mhd models ( rees & gunn 1974 ; kennel & coroniti 1984 ; emmering & chevalier 1987 ; begelman & li 1992 ) , the pulsar wind terminates in a standing shock at a radius defined by the condition that the confining pressure balances the momentum flux of the wind . in the crab case ,
the shock radius was estimated to be @xmath1 cm in excellent agreement with observations ; according to chandra results ( weisskopf et al .
2000 ) , the radius of the shock in the equatorial plane is @xmath2 cm . the observed brightness and the spectral index distributions are generally consistent with the assumption that the relativistic particles are accelerated at the termination shock and then fill in the nebula , spending the acquired energy on synchrotron emission and @xmath3 work .
the generic observational feature of plerions is a flat radio spectrum ; the spectral flux may be presented as a power law function of the frequency , @xmath4 , with the spectral index @xmath5 . at high frequencies the spectrum steepens and
the typical spectral slope in the x - ray band is @xmath6 .
the overall spectrum of the crab may be described as a broken power law with the spectral breaks around @xmath7 hz , few@xmath8 hz and around 100 kev .
the synchrotron lifetime of the radio emitting electrons ( and positrons , below by electrons i mean both electrons and positrons ) significantly exceeds the plerion age therefore one can not exclude a priori that they were injected at the very early stage of the plerion evolution ( kennel & coroniti 1984 ; atoyan 1999 ) . in this case
the overall spectrum depends on history of the nebula .
however the spectral break at @xmath7 hz may be simply accounted for the synchrotron burn off effect assuming that particles emitting from the radio to the optical bands are injected more or less homogeneously in time with the single power law energy distribution .
this view is strongly supported by gallant & tuffs ( 1999 , 2002 ) who found that the infra - red spectral index in the central parts of the crab is close to that in the radio , and gradually steepens outward .
recent observations of wisps in the radio band ( bietenholz & kronberg 1992 ; bietenholtz , frail & hester 2001 ) suggest unambiguously that the radio emitting electrons are accelerated now in the same region as the ones responsible for the optical to x - ray emission .
if the radio emitting electrons have been injecting into the crab nebula till the present time , the injection rate of electrons should be about @xmath9 s@xmath10 .
it is interesting that the observed pulsed optical emission from the crab pulsar suggests that about the same amount of electrons is ejected from the pulsar magnetosphere ( shklovsky 1970 ) so the pulsar does supply the necessary amount of particles .
the observed spectral slope in the radio band , @xmath11 , implies the energy distribution of the injected electrons of the form @xmath12 . in this case , most of the particles find themselves at the low energy end of the distribution whereas particles at the upper end of the distribution dominate the energy density of the plasma .
taking into account that no sign of a low frequency cut - off is observed in the crab spectrum down to about 10 mhz whereas the high frequency break lies in the ultra - violet band ( recall that the break at about @xmath7 hz is attributed to synchrotron cooling but not to the injected energy distribution ) , one concludes that the above distribution extends from @xmath13 mev to @xmath14 mev . at @xmath15
the distribution becomes steeper ; the spectral slope in the x - ray band , @xmath16 , corresponds , with account for the synchrotron burn off effect , to @xmath17 . the distribution further steepens at @xmath18 mev as it follows from the gamma - ray spectrum of the nebula .
thus the injected electrons have a very wide energy distribution , their number density being dominated by low energy electrons whereas the plasma energy density being dominated by tev electrons .
the above considerations place severe limits on the pulsar wind parameters and possible mechanisms of the particle acceleration at the termination shock . according to the widespread view
, the pulsars emit poynting - dominated winds however the electro - magnetic energy is efficiently transferred to the plasma flow such that the magnetization parameter @xmath0 , defined as the ratio of the poynting flux to the kinetic energy flux , is already very small when the flow enters the termination shock .
the mechanisms of such an energy transfer still remain unclear ( the so called @xmath0-problem ) however dynamics of the flow in the nebula suggests that the magnetic pressure is small just beyond the termination shock , which implies that the poynting flux just upstream the shock is very small ( rees & gunn 1974 ; kundt & krotscheck 1980 ; kennel & coroniti 1984 ; emmering & chevalier 1987 ; begelman & li 1992 ) . in this case
it is kinetic energy of the upstream flow that converts into the energy of accelerated particles when the plasma flow is randomized at the shock .
then the characteristic downstream `` temperature '' is about the upstream particle kinetic energy , @xmath19 , so the average particle energy does not vary considerably across the shock . a high energy tail may be formed in the particle energy distribution ( the particle acceleration at the relativistic shocks is considered by hoshino et al .
( 1992 ) ; gallant & arons ( 1994 ) ; bednarz & ostrowski ( 1998 ) ; gallant & achterberg ( 1999 ) ; kirk et al .
( 2000 ) ; achterberg et al .
( 2001 ) ) however this tail merges , at its low energy end , with the quasi - thermal distribution at @xmath20 .
therefore if the pulsar spin - down power , @xmath21 , is converted into the kinetic energy of the wind , the available upper limit on the low - frequency break in the crab spectrum suggests that the wind lorentz - factor , @xmath22 , does not exceed few hundred . on the other hand , the observed flat spectrum may be formed only if the energy per electron in the wind is much larger than @xmath23 .
gallant et al.(2002 ) , modifying the original idea of hoshino et al .
( 1992 ) and gallant & arons ( 1994 ) , suggested that the wind is loaded by ions ; in this case the radio emitting electrons are accelerated by resonant absorption of ion cyclotron waves collectively emitted at the shock front . the necessary ion injection rate , @xmath24 s@xmath10 , vastly exceeds the fiducial goldreich - julian elementary charge loss rate , @xmath25 s@xmath10 .
although one can not exclude by observations that pulsars emit the required amount of ions , the available pulsar models do not assume an ion outflow with the rate exceeding the goldreich - julian charge loss rate ( cheng & ruderman 1980 ; arons 1983 ) .
here i consider an alternative possibility , which does not imply a radical modification of the basic pulsar model . in the equatorial belt of the wind from an obliquely rotating pulsar magnetosphere , the sign of the magnetic field alternates with the pulsar period forming stripes of opposite magnetic polarity ( michel 1971 , 1982 ; coroniti 1990 ; bogovalov 1999 ) .
observations of x - ray tori around pulsars ( weisskopf et al .
2000 ; helfand , gotthelf & halpern 2001 ; gaensler , pivovaroff & garmire 2001 ; pavlov et al .
2001 ; gaensler et al . 2002 ; lu et al .
2002 ) suggest that it is in the equatorial belt where most of the wind energy is transported ; theoretical models ( e.g. , bogovalov s ( 1999 ) solution for the oblique split monopole magnetosphere ) support this conclusion .
therefore the fate of the striped wind is of special interest . in the striped wind ,
the poynting flux converts into the particle energy flux when the oppositely directed magnetic fields annihilate ( coroniti 1990 ; lyubarsky & kirk 2001 , henceforth lk ; lyutikov 2002 ; kirk & skjraasen 2003 ) . until now
this dissipation mechanism was considered only in the unshocked wind .
it was found that the flow acceleration in the course of the energy release dilates the dissipation timescale so that the wind may enter the termination shock before the alternating field annihilates completely . in this case driven annihilation of the magnetic field at the shock
may provide the energy necessary to form the flat particle distribution . on the other hand ,
the formed distribution may extend down to low enough energy because the kinetic energy of the flow in the striped wind is lower than the total energy .
the aim of this research is to consider properties of the termination shock in the striped wind .
it will be shown that the alternating field completely annihilates at the shock so that the downstream parameters of the flow are the same as if the field has already annihilated upstream the shock .
therefore the available limits on the upstream magnetization parameter , @xmath0 , should be attributed not to the total poynting flux but to the poynting flux associated with the averaged magnetic field ( see also rees & gunn 1974 ; kundt & krotscheck 1980 ) .
the upstream flow may be poynting dominated provided most of the poynting flux is transferred by alternating magnetic field .
on the other hand , driven reconnection of the magnetic field within the shock radically alter the particle acceleration process .
it follows from both analytical and numerical studies that particles are readily accelerated in the course of the magnetic field reconnection to form the energy distribution with the power - law index @xmath26 ( romanova & lovelace 1992 ; zenitani & hoshino 2001 ; larrabee , lovelace & romanova 2002 ) .
electrons with such a distribution emit synchrotron radiation with a flat spectrum therefore romanova & lovelace ( 1992 ) and birk , crusius - wtzel & lesch ( 2001 ) suggested that reconnection plays a crucial role in flat spectrum extragalactic radio sources .
one can naturally assume that radio emission of plerions is generated by electrons accelerated in the course of reconnection of the alternating magnetic field at the pulsar wind termination shock .
then the steeper high - energy spectrum may be attributed to the fermi acceleration of particles preaccelerated in the reconnection process .
a simple model for the particle acceleration in the shocked striped wind is presented here .
it will be shown that the minimal energy of the power law energy distribution may be low in this case , even less then the upstream kinetic energy , whereas the energy density of the plasma will be dominated by high - energy electrons in agreement with the observed spectra of plerions .
the article is organized as follows .
jump conditions for a shock in the striped wind are obtained in sect .
2 . making use of these jump conditions i demonstrate in sect .
3 that the alternating magnetic field dissipates completely at the pulsar wind termination shock . in sect .
i discuss particle acceleration by driven reconnection .
particle acceleration at the shock in the striped wind is analyzed in sect .
the obtained results are summarized in sect .
let us find jump conditions for a shock in a flow with an alternating magnetic field .
the plasma upstream the shock is assumed to be cold everywhere with the exception of narrow current sheets separating stripes with opposite magnetic polarity .
the pressure balance implies that the magnetic field strength in adjacent stripes differs only by sign but not by the absolute value ; on the other hand the width of stripes with opposite polarity may not be the same so that the average magnetic field may be nonzero ( fig .
1 ) . one generally finds jump conditions from the conservation laws and some prescription for the magnetic flux passing the shock . in the standard mhd shock
, the magnetic flux is conserved ; in the case under consideration one should choose more general prescription to take into account possible annihilation of the magnetic field at the shock .
outside the shock , the magnetic field is frozen into the plasma , @xmath27 . in the case of interest
the magnetic field lies in the plane of the shock . in the frame
where the flow velocity is perpendicular to the shock plane @xmath28 .
then the faraday law may be written as @xmath29 together with the continuity equation @xmath30 this implies @xmath31 where @xmath32 and @xmath33 are the lorentz factor and the proper plasma density , correspondingly , and @xmath34 is a constant for each fluid element .
the energy and momentum fluxes are @xmath35 @xmath36 where @xmath37 and @xmath38 are the gas pressure and specific enthalpy , correspondingly .
taking into account the above considerations , one can write @xmath39 where the effective pressure and enthalpy are @xmath40 so the flow may be described by the hydrodynamical equations . in the standard mhd theory , the magnetic flux conserves and the flow
is considered as homogeneous both upstream and downstream the shock so that @xmath34 is a global constant . in our case
@xmath34 is an alternating function of the lagrangian coordinate however only @xmath41 enters the conservation laws therefore the only essential difference from the standard mhd theory is that the modulus @xmath34 may decrease in the shock because of the field annihilation . let us introduce a parameter @xmath42 to measure decreasing in the magnetic flux across the shock .
the indexes 1 and 2 are referred to quantities upstream and downstream the shock , correspondingly .
one can express the downstream parameters via the upstream ones making use of the conservation of the particle , energy and momentum fluxes across the shock .
radiation losses may be safely neglected at the shock width scale . in the striped wind
, one should take the fluxes averaged over the wave period .
let us assume , for the sake of simplicity , that the current sheets between the stripes are so narrow that one can neglect the contribution of the plasma within the sheets into the conserving fluxes .
outside the sheets , the magnetic field and plasma density are constant across a stripe and the adjacent stripes are differ only by the sign of the magnetic field and by the width . taking into account that the energy and momentum fluxes are proportional to the magnetic field squared ,
one can write the conservation laws like in the case of homogeneous magnetic field : @xmath43 @xmath44 @xmath45 the only formal difference from the case of the standard mhd shock is in the additional parameter @xmath46 ; in the standard shock @xmath47 . one can resolve the above equations as follows .
the plasma downstream the shock is relativistically hot , @xmath48 , therefore one can easily eliminate @xmath38 and @xmath37 from eq.(3 ) and express @xmath49 via @xmath50 and @xmath51 . substituting the obtained relation into eq.(7 ) , one then eliminates @xmath50 making use of eq.(6 ) and eliminates @xmath51 making use of eq.(5 ) .
taking into account that the plasma upstream the shock is cold , @xmath52 and @xmath53 , one gets the resulting equation for the downstream velocity in the form @xmath54@xmath55 where @xmath56 is the magnetization parameter .
this equation is exact . in case @xmath57
one can take @xmath58 .
then eq.(8 ) reduces to @xmath59 now the downstream parameters may be found explicitly : @xmath60 @xmath61 } } { 1+\chi+\sqrt{1 + 14\chi+\chi^2 } } ; \eqno(11)\ ] ] @xmath62@xmath63 where @xmath64 at @xmath47 one recovers the downstream parameters for the relativistic mhd shock in the homogeneous medium ( kennel & coroniti 1984 ; appl & camenzind 1988 ) . in this case
the quantity @xmath65 is the ratio of the upstream poynting flux to the total energy flux . in a striped wind ,
@xmath65 is determined by the poynting flux corresponding to the magnetic flux passed the shock .
it will be shown in the next section that the alternating magnetic field annihilates completely at the pulsar wind termination shock ; then @xmath65 is simply the ratio of the poynting flux associated with the average magnetic field , @xmath66 , to the total energy flux . at the equator of the flow
the average field is zero , @xmath67 , therefore one gets @xmath68 like at the nonmagnetized relativistic shock .
when the average magnetic flux is large , @xmath69 , @xmath70 , the downstream flow is relativistic with the lorentz factor @xmath71
in a striped wind , the magnetic field forms toroidal stripes of opposite polarity , separated by current sheets ( michel 1971 , 1982 ; bogovalov 1999 ) .
such a structure arises in the equatorial belt of the pulsar wind and may be considered as an entropy wave propagating through the wind .
usov ( 1975 ) and michel ( 1982 ) noticed that the waves must decay at large distances , since the current required to sustain them falls off as @xmath72 whereas the density of available charge carriers in the wind decreases as @xmath73 . hence the alternating magnetic fields should eventually annihilate .
it was shown in lk that the distance beyond which the available charge carriers are unable to maintain the necessary current exceeds the radius of the standing shock where the wind terminates so that only some fraction of the magnetic energy is converted into the particle energy before the plasma reaches this shock front .
this fraction depends on the magnetic field reconnection rate .
a lower limit may be obtained assuming that the dissipation keeps the width of the current sheet to be equal to the particle larmor radius , which is roughly the same condition as that the current velocity is equal to the speed of light ( corontiti 1990 ; michel 1994 ; lk ) . then about 10% of the poynting flux dissipates before the wind reaches the termination shock . at higher reconnection rate
the fraction of the dissipated energy is larger ; assuming that the reconnection velocity is as high as the sound velocity one can find that the alternating field annihilates completely before the flow enters the termination shock ( kirk & skjraasen 2003 ) . here
i adopt the slow reconnection model by lk to estimate the flow parameters just upstream the termination shock . neglecting the dependence of the wind parameters on latitude ,
one can express them only via the pulsar spin - down power , @xmath21 , and the total amount of the ejected electrons , @xmath74 .
lk used instead the ratio of the gyro - frequency at the light cylinder to the angular velocity of the neutron star and the so called multiplicity coefficient arising in the models of the pair production in pulsar magnetospheres .
these parameters may be expressed via @xmath21 and @xmath74 taking into account that the magnetic field at the light cylinder is estimated as @xmath75 , where @xmath76 is the light cylinder radius , @xmath77 the pulsar period .
the characteristic distance beyond which the available charge carriers are unable to maintain the necessary current ( eq.(14 ) in lk ) may be written as @xmath78 where @xmath79 is the electron charge .
for the crab @xmath80 cm . note that the dimensionless parameter @xmath81 used by kirk & skjraasen ( 2003 ) is proportional to @xmath82 . in the poynting
dominated flows , dissipation of even a small fraction of the poynting flux implies significant acceleration of the flow . while the dissipated energy is small , the lorentz factor of the wind ( eq.(30 ) in lk ) may be estimated as @xmath83 where @xmath84 is the lorentz factor attained by the wind if all the spin - down power is converted into the kinetic energy of the plasma .
note that @xmath85 is the fraction of the spin - down power transferred to the plasma therefore the magnetization parameter of the wind may be written as @xmath86 the current sheet width is conveniently measured by a fraction @xmath87 of a wavelength @xmath88 occupied by the two current sheets ; eq.(31 ) in lk may be written as @xmath89 in the crab , the termination shock is observed at the distance @xmath2 cm ( weisskopf et al . 2000 ) therefore just upstream the shock @xmath90 , @xmath91 ; @xmath92 , where @xmath93 . at the termination shock ,
the flow sharply decelerates so the plasma is compressed .
the proper density grows @xmath94 times . according to eqs.(10 ) and
( 13 ) , the less @xmath95 ( i.e. the larger fraction of the magnetic flux dissipates at the shock ) the less the downstream velocity so one can place a lower limit on the compression factor assuming @xmath47 . in this case @xmath96 ; substituting the estimated above parameters of the crab wind , one gets @xmath97 .
such a huge compression factor implies significant heating of the plasma , especially within the current sheets .
let us show that after such a compression the particle larmor radius should exceed the wavelength so that the alternating magnetic field dissipates completely within the shock . the downstream temperature , which determines the larmor radius , depends on the fraction of the alternating magnetic field annihilated at the shock .
let us first assume that this fraction remains small so that the structure sketched in fig .
1 preserves downstream the shock .
the width of the current sheet in the wind frame can not be less than the particle larmor radius , @xmath98 where @xmath99 is the plasma temperature within the sheet and @xmath100 the magnetic field in the wind frame of reference .
the wavelength in the wind frame is @xmath101 therefore the minimum fraction of the wavelength occupied by the two current sheets is @xmath102 the temperature obeys the condition of the hydrostatic equilibrium of the plasma in the sheet , @xmath103 , where @xmath104 is the number density of the plasma in the sheet .
the magnetic field is frozen into the plasma outside the sheet therefore @xmath105 , where @xmath106 is the plasma number density outside the sheet
. now one can write ( see eq.(17 ) in lk ) @xmath107 the left - hand side of this relation is the fraction of particles carried in the sheet ; of course this fraction should be less than unity . in the unshocked wind @xmath108 ( lk ) ; with the above estimate for the upstream @xmath87 , one obtains that upstream the shock this fraction is about 0.1 .
it was demonstrated above that @xmath32 decreases about 1000 times when the flow passes the shock , therefore downstream the shock @xmath109 , which is impossible .
this means that contrary to the initial assumption , the fraction of the magnetic field dissipated within the shock is not small and @xmath95 defined by eq.(4 ) is not close to unity . in this case
@xmath65 ( eq.(13 ) ) is also not close to unity and then , according to eqs.(10 ) and ( 12 ) , the downstream flow is non- or mildly relativistic , @xmath110 , whereas the downstream temperature is about the particle kinetic energy upstream the shock , @xmath111 . with the above estimates for the upstream lorentz factor and magnetization parameter
, one can easily see that the larmor radius downstream the shock , @xmath112 cm , vastly exceeds the wavelength , @xmath113 cm , therefore the alternating magnetic field should completely annihilate at the shock .
this means that the parameter @xmath65 should be determined as the ratio of the poynting flux associated with the average magnetic field , @xmath114 , to the total energy flux
. therefore the downstream parameters are independent of whether the alternating magnetic field dissipated upstream the shock or just at the shock .
the available upper limits on the magnetization parameter in the pulsar wind ( kennel & coroniti 1984 ; emmering & chevalier 1987 ; begelman & li 1992 ) was found from the standard mhd shock conditions applied to the downstream parameters estimated from the analysis of the plasma dynamics in the nebula downstream the shock . in the case of the striped wind these upper limits should be attributed not to the total poynting flux but to the poynting flux associated with the average magnetic field ( see also rees & gunn 1974 ; kundt & krotscheck 1980 ) .
the upstream flow may be poynting dominated provided most of the electromagnetic energy is carried by the alternating magnetic field , which annihilates at the termination shock .
the particle acceleration by driven reconnection may be qualitatively described as follows .
let us consider a box with two stripes of the oppositely directed magnetic field and a current sheet between them ( fig .
when the box is compressed , the electric field @xmath115 arises , which has the same sign in the domains of opposite magnetic polarity .
close enough to the zero line of the magnetic field , particles may move freely along this line and gain energy from the electric field .
of course the real picture should be much more complicated than the presented one - dimensional sketch .
the magnetic reconnection might proceed in separate x - points but in average the process remains one - dimensional because the released energy is confined to a layer around the field reversal .
acceleration of relativistic electrons close to an x - point was considered by romanova & lovelace ( 1992 ) ; zenitani & hoshino ( 2001 ) , larrabee et al.(2002 ) who found a power - law particle distribution with the slope @xmath26 .
reconnection in a long current sheet at a time - scale large enough that particles pass many x - points has not been considered yet .
let us assume that a power - law distribution is formed in this case also and estimate the maximal energy particles attain when the magnetic field annihilates completely .
let the particle energy distribution within the current sheet be @xmath116 at @xmath117 with @xmath118 .
at such a distribution function , the particle density is dominated by low energy electrons , @xmath119 whereas the energy density is dominated by high - energy electrons , @xmath120 let us assume that the power law index @xmath121 remains fixed in the course of compression and only @xmath122 and @xmath123 vary .
one can find these variations considering particles and energy balance within the box .
the particle balance is written as @xmath124 where @xmath125 is the sheet width , @xmath126 the box width , @xmath104 and @xmath106 the particle number densities in the sheet and outside it , correspondingly , and the index 0 is referred to the initial state .
the energy balance implies that the variation of the total energy within the box is equal to the work done on the box by the outer pressure ( @xmath127 ) : @xmath128=-\frac{b^2}{8\pi}dl,\ ] ] where @xmath129 is the plasma energy density in the sheet .
taking into account that the magnetic field is frozen into the cold plasma , @xmath130 and that the plasma pressure in the sheet , @xmath131 , is counterbalanced by the magnetic pressure , @xmath132 one can write the energy balance equation in the form @xmath133 let us assume that the sheet width is equal to the maximal larmor radius @xmath134 the maximal lorentz factor , @xmath123 , is found from eqs .
( 21 ) , ( 22 ) , ( 24 ) and ( 25 ) as @xmath135 now one can write @xmath136 in order to find the plasma parameters in the course of compression one should solve eqs.(23 ) , ( 26 ) and ( 28 ) for @xmath137 , @xmath138 , @xmath139 from @xmath140 to smaller @xmath126 .
the magnetic field dissipates completely when the sheet width becomes comparable with the box width , @xmath141 . at the last stage ,
one is unable to separate unambiguously the total plasma volume onto a hot current sheet and a cold magnetized plasma therefore the presented equations may be only used to estimate roughly the final plasma parameters . taking this into account
, one can simplify the problem even more and solve these equations in the limit @xmath142 , when they are formally applicable , and take the limit @xmath143 in the obtained solutions .
let us introduce the dimensionless variable @xmath144 . in the zeroth order in small @xmath87 , both eq.(23 ) and eq.(26 ) reduce to the same equation
@xmath145 so the system is nearly degenerate . in order to find the second equation
, one should expand eqs.(23 ) and ( 26 ) to the first order in @xmath87 and eliminate the zeroth order term . introducing one more dimensionless variable @xmath146 , one gets @xmath147 transforming eq.(28 ) to the dimensionless variables @xmath87 and @xmath148 and substituting @xmath106 from eq.(29 ) , one obtains @xmath149 now one can eliminate @xmath126 from eqs.(30 ) and ( 31 ) to get finally @xmath150 the solution to this equation is @xmath151 at @xmath152 ( but still @xmath153 ) @xmath148 goes to a constant , @xmath154 , independently of the initial conditions . taking into account that in the pulsar wind @xmath155 ( lk ) , one can take for estimates @xmath156 .
now one can estimate the compression factor , @xmath157 , necessary for the magnetic field to dissipate completely .
substituting @xmath158 into eq.(31 ) , one gets @xmath159 substituting @xmath160 into eq.(27 ) yields an estimate for the maximal lorentz factor attained when the magnetic field dissipates completely @xmath161 where @xmath162 is the initial magnetization parameter .
it follows from eqs.(16 ) , ( 18 ) and ( 19 ) that in the wind upstream the shock @xmath163 therefore the particles may be accelerated up to large energies , @xmath164 , even at a moderately large @xmath0 .
at the termination shock , the flow is compressed and the energy of the alternating magnetic field is released . close to the equator of the flow , the average magnetic field is small and nearly all the poynting flux is transferred to the particles .
it was shown in sect . 2 that the downstream velocity is non - relativistic in this case , therefore the proper density of the plasma increases @xmath165 times where @xmath22 is the wind lorentz factor upstream the shock . for typical parameters ( see sect .
3 ) this compression factor significantly exceeds that given by eq . ( 33 ) ; this confirms the conclusion that the alternating magnetic field annihilates completely at the termination shock .
the flow is decelerated in the shock by the pressure of the hot downstream plasma and magnetic field . in the collisionless shock
the deceleration scale is about the larmor radius of those particles , which make a major contribution to the downstream pressure .
these particles penetrate upstream by their larmour radius and exert , via the magnetic field , the decelerating force on the upstream flow . in the standard mhd shock ,
the downstream temperature is about the particle kinetic energy in the upstream flow therefore the upstream particles penetrate about all the shock width immediately after they enter the shock .
so there is only one characteristic spatial scale in this case , namely that of the larmor radius corresponding to the upstream kinetic energy .
in the shock in the striped wind , the particles gain energy from the annihilating magnetic field so that the downstream pressure is determined by particles with the energy significantly exceeding the upstream kinetic energy .
the larmor radius of these particles significantly exceeds not only the larmor radius corresponding to the kinetic energy in the upstream flow but even the strip width .
therefore the width of such a shock , or the deceleration scale , exceeds all `` internal '' scales in the upstream flow and hence the the field annihilation proceeds locally in the proper frame like in the plasma smoothly compressed by an external force .
only when the alternating field dissipates completely , the particle larmor radius , calculated with account of both thermal and kinetic energy , becomes comparable with the shock width and the flow decelerates further on like in the standard shock .
so one can roughly separate the shock into two zones . in the first zone ,
the flow is decelerated and compressed by the pressure of high energy particles entering from the second zone .
the magnetic field dissipates in the first zone roughly according to the simple picture outlined in sect .
the plasma heated by the field annihilation in the first zone enters the second one therefore the second zone resembles the standard shock with a hot upstream plasma .
the compression factor , @xmath166 , necessary for the magnetic field to dissipate completely was estimated in the previous section .
the continuity equation in the relativistic flow , @xmath167 , implies that the flow lorentz factor at the end of the field dissipation stage is @xmath168 .
adopting the outlined in the previous section picture of the particle acceleration by driven reconnection , one concludes that a power - law particle distribution is formed at this stage . in the proper plasma frame
moving with the lorentz factor @xmath169 , this distribution extends from @xmath170 up to @xmath171 .
now the alternating field is already dissipated and the flow decelerates further on , from @xmath172 down to @xmath173 , like in the standard mhd shock . applying the particle and energy flux conservation , @xmath174 and @xmath175 , one finds the maximal energy in the particle distribution downstream the shock , @xmath176 .
the upper limit on the minimal energy may be estimated from the condition that the energy of the most of particles is simply randomized but does not changes considerably ; then @xmath177 .
making use of eqs.(33 ) and ( 34 ) , one gets @xmath178 @xmath179 substituting @xmath180 and parameters of the crab pulsar wind upstream the termination shock ( see sect.3 ) , one gets @xmath181 ; @xmath182 , which is roughly compatible with the parameters inferred from the observed spectrum .
these simple estimates show that particle acceleration at the shock in a striped wind may form such a particle distribution that the energy of most of the particles is significantly less than the upstream particle kinetic energy whereas the plasma energy density is dominated by a relatively small amount of high energy particles .
these high energy particles may be accelerated further on by the 1-st order fermi mechanism thus forming a high energy tail at @xmath183 .
recent investigations ( bednarz & ostrowski ( 1998 ) ; gallant & achterberg ( 1999 ) ; kirk et al .
( 2000 ) ; achterberg et al .
( 2001 ) ) have shown that in ultra - relativistic shocks , the tail is formed with the power - law index @xmath184 compatible with the observed x - ray spectra of the crab and other plerions .
so the observed broken power - law spectrum with a flat low frequency part may be attributed to the particle acceleration at the termination shock in a striped pulsar wind .
the observed spectra of plerions from the radio to the gamma - ray band imply a very wide electron energy distribution , from less than few hundreds mev to @xmath185 ev .
most of electrons are accumulated at the low - energy end of this distribution .
although the synchrotron life time of these electrons exceeds the plerion age , there is strong evidence to suggest that they are accelerated now together with high energy electrons responsible for the hard radiation from the nebula ( bietenholz & kronberg 1992 ; gallant & tuffs 1999 , 2002 ; bietenholtz et al .
the observed flat radio spectrum implies that the acceleration mechanism transfers most of the available energy to a small fraction of particles and retains most of particles at relatively low energy .
it is proposed in this article that the flat energy distribution is formed in the course of the particle acceleration by driven reconnection of the alternating magnetic field at the pulsar wind termination shock .
it is widely believed that just upstream the termination shock the magnetic energy is negligible as compared with the plasma kinetic energy because plasma dynamics in the nebula implies small magnetization downstream the shock ( rees & gunn 1974 ; kennel & coroniti 1984 ; emmering & chevalier 1987 ; begelman & li 1992 ) .
however the plasma magnetization and dynamics depend only on the average upstream magnetic field so the upstream flow may be poynting dominated provided most of the magnetic energy is associated with the alternating magnetic field . in this case
the average particle energy grows significantly at the shock where the magnetic field annihilates .
therefore the particle distribution with the power - law index @xmath186 is formed readily .
the proposed mechanism may also explain why no low - frequency cutoff is observed in the crab radio spectrum .
at the kinetic energy dominated shocks , the power - law tail is formed only at the energies exceeding the downstream temperature , which is about the particle kinetic energy upstream the shock .
therefore if the crab pulsar wind were kinetic energy dominated , only rather low lorentz factor of the wind would be compatible with the radio data , which would imply extremely highly mass loaded wind . in the striped wind , most of the energy
is contained in the magnetic field therefore the wind lorentz factor is lower than in the kinetic energy dominated wind .
moreover the presented here simple model of the shock in a striped wind predicts that in this case the downstream temperature may be considerably less than the upstream particle kinetic energy .
therefore the lack of the low - frequency turnover in the observed plerion spectra may be naturally explained within the scope of the proposed model .
of course the outlined here qualitative picture of the particle acceleration in plerions may be considered only as preliminary ; it may be justified only by numerical simulations of the shock in a striped wind .
i am grateful to david eichler for stimulating discussions .
achterberg a. , gallant y.a . ,
kirk j.g . ,
guthmann a.w . , 2001 , mnras , 328 , 393 appl s. , camenzind m. , 1988 , a & a , 206 , 258 arons j. , 1983 , in positron - electron pairs in astrophysics , eds .
m.l.burns , a.k.harding , r.ramaty ( ny , aip ) , p. 163 atoyan a.m. , 1999 , a & a , 346 , l49 bednarz j. , ostrowski m. , 1998 , phys.rev.lett . , 80 , 3911 begelman m.c . , li z .- y . , 1992 , apj , 397 , 187 bietenholtz m.f .
, frail d.a .
, hester j.j . , 2001 , apj , 560 , 254 bietenholtz m.f .
, kronberg p.p . , 1992 ,
apj , 393 , 206 birk g.t .
, crusius - wtzel a.r . , lesch h. , 2001 , apj , 559 , 96 bogovalov s.v . , 1999 ,
a&a , 349 , 101 cheng a.f . ,
ruderman m.a . , 1980 ,
apj , 235 , 576 coroniti f.v . , 1990 , apj , 349 , 538 emmering r.t . ,
chevalier r.a . , 1987 , apj , 321 , 334 gaensler b.m . , arons j. , kaspi v.m . , pivovaroff m.j . , kawai n. , tamura k. , 2002 , apj , 569 , 878 gaensler b. m. , pivovaroff m. j. , garmire g. p. , 2001 ,
apj , 556 , l107 gallant y.a .
, achterberg a. , 1999 , mnras , 305 , l6 gallant y.a . , arons j. , 1994 , apj , 435 , 230 gallant y. a. , tuffs r. j. , 1999 , pulsar astronomy - 2000 and beyond , asp conference series , vol . 202 ; proc .
iau coll .
177 ( san francisco : asp ) .
eds . : m. kramer , n. wex , and n. wielebinski , p. 503
gallant y. a. , tuffs r. j. , 2002 , in neutron stars in supernova remnants , asp conference series , vol .
271 , eds .
slane and b.m.gaensler .
asp conf .
series vol.271 , p.161 gallant y.a . , van der swaluw e. , kirk j.g .
, achterberg a. , 2002 , in neutron stars in supernova remnants , eds . : p.o.slane and b.m .
asp conf .
series vol .
271 , , p. 99
helfand d. j. , gotthelf e. v. , halpern j. p. , 2001
, apj , 556 , 380 hoshino m. , arons j , gallant y.a .
, langdon a.b . , 1992 ,
apj , 390 , 454 kennel c. f. , coroniti f. v. , 1984 , apj , 283 , 694 kirk j.g . ,
guthman a.w .
, gallant y.a . ,
achterberg a. , 2000 , apj , 542 , 235 kirk j.g . , skjraasen o. , 2003 , apj , in press kundt w. , krotscheck e. , 1980 , a&a , 83 , 1 larrabee d.a . , lovelace r.v.e . ,
romanova m.m . , 2002 , astroph/0210045 lu f.j . , wang q.d . , aschenbach b. , durouchoux p. , song l.m . , 2002 ,
apj , 568 , l49 lyubarsky y.e . ,
kirk j.g . , 2001 ,
apj , 547 , 437 ( lk ) lyutikov m. , 2002 , astro - ph/0210353 michel f.c . , 1971 , comments astrophys.space phys . , 3 , 80 michel f.c . , 1982 , rev.mod.phys . , 54 , 1 michel f.c . , 1994 ,
apj , 431 , 397 pavlov g.g . ,
kargaltsev o.y . , sanwal d. , garmire g.p . , 2001 , apj , 554 , l189 rees m. j. , gunn , j. e. , 1974 , mnras , 167 , 1 romanova m. m. , lovelace r. v. e. , 1992 , a&a , 262 , 26 shklovsky i. s. , 1970 , apj , 159 , l77 usov v.v .
, 1975 , apss , 32 , 375 weisskopf c. et al . ,
2000 , apj , 536 , l81 zenitani s. , hoshino m. , 2001 , apj , 562 , l63 | the origin of radio emission from plerions is considered .
recent observations suggest that radio emitting electrons are presently accelerated rather than having been injected at early stages of the plerion evolution .
the observed flat spectra without a low frequency cutoff imply an acceleration mechanism that raises the average particle energy by orders of magnitude but leaves most of the particles at the energy less than about few hundred mev .
it is suggested that annihilation of the alternating magnetic field at the pulsar wind termination shock provides the necessary mechanism .
toroidal stripes of opposite magnetic polarity are formed in the wind emanated from an obliquely rotating pulsar magnetosphere ( the striped wind ) . at the termination shock , the flow compresses and the magnetic field annihilates by driven reconnection .
jump conditions are obtained for the shock in a striped wind .
it is shown that postshock mhd parameters of the flow are the same as if the energy of alternating field has already been converted into the plasma energy upstream the shock .
therefore the available estimates of the ratio of the poynting flux to the matter energy flux , @xmath0 , should be attributed not to the total upstream poynting flux but only to that associated with the average magnetic field .
a simple model for the particle acceleration in the shocked striped wind is presented .
acceleration of particles magnetic fields mhd shock waves pulsars : general supernova remnants |
among the iii - v semiconductors the nitride - based semiconductors have received both scientific and technological attention over the past years owing to their excellent physical properties.@xcite the fascinating mechanical properties , such as hardness , high melting temperature , high thermal conductivity , and large bulk moduli make them useful for protective coatings . in particular , the electronic properties , characterized by large band gaps and relatively low dielectric constants , indicate tremendous potential applications in optical devices working in the blue and ultraviolet ( uv ) wavelengths .
recently , high - power blue and green light emitting diodes ( leds ) have been fabricated by using ingan - based multiquantum - well structures and are commercially available .
@xcite blue and uv nitride semiconductor laser diodes ( lds ) have also successfully been demonstrated .
@xcite consequently there has been great interest in the fundamental properties , the understanding of which can render the improvement of the materials and device quality possible .
heterostructures based on gan and ga@xmath0al@xmath1n have become important candidates for optoelectronic materials .
great attention has been paid to the nitride heterostructures and layered structures .
short - period superlattices ( sl s ) have been successfully fabricated by advanced growth techniques .
@xcite phonon spectra of these sl s were investigated by infrared ( ir ) spectroscopy . @xcite
it has been known that gan and aln are crystallized in the stable wurtzite structure and metastably in the cubic structure .
phonons in cubic gan / aln systems have drawn interest in recent years .
theoretical studies on phonons in cubic gan / aln sl s have been reported by several groups .
@xcite there have been , however , few studies of phonons in hexagonal gan / aln sls,@xcite where the constituents have wurtzite structures . in this paper
we develop a planar force - constant model ( pfcm ) to describe the longitudinal phonons in wurtzite gan and aln propagating along the [ 0001 ] direction .
we then apply the pfcm to study the phonons in hexagonal ( 0001)-gan / aln sl s .
the paper is organized as follows .
the pfcm is described in section ii .
the calculated results for hexagonal ( 0001)-gan / aln sl s are presented in section iii and conclusions are given in section iv .
in general , a three - dimension calculation is necessary in the discussion of phonons because of the coupling between longitudinal and transverse vibrations . for the vibrations of wurtzite compounds along the [ 0001 ] direction , however , the longitudinal and transverse vibrations can be decoupled . as a result , the longitudinal and transverse vibrations can be treated separately .
the pftm for wurtzite gan or aln is schematically displayed in fig .
the dynamical matrix elements are given by @xmath2 \right\ } , \label{dy}\end{aligned}\ ] ] where @xmath3 is the wave - vector ; @xmath4 is the index of a unit cell and @xmath5 stands for the index of atoms in the unit cell ; @xmath6 is the mass of the atom @xmath5 ; @xmath7 is the planar force - constant between the atoms @xmath8 and @xmath9 ; @xmath10 is the position of the atom @xmath11 .
the dynamical matrix subjects to the following secular equation @xmath12 the eigenvalues and eigenvectors of the vibrational modes can be obtained by solving eq .
( [ secular ] ) .
four planar force - constants shown in fig .
[ fig1 ] are used for either gan or aln .
they are two cation - anion planar force - constants , one cation - cation force - constant and one anion - anion force - constant .
the on - site force - constants @xmath13 can be obtained from the sum rules @xmath14 the values of the four planar force - constants are obtained by a fitting procedure .
the existing experimental data that are useful in fitting are the a@xmath15(lo ) phonon frequency at the zone - center determined by raman spectroscopy .
therefore , the four planar force - constants are obtained by fitting them to the three - dimensional calculations.@xcite the fitted a@xmath15(lo ) phonon frequency , two b@xmath15 phonon frequencies and the elastic constant @xmath16 are given in table [ tab1 ] together with the corresponding experimental data and other theoretical results .
it is obvious that the fitted a@xmath15(lo ) phonon frequencies of both gan and aln are in good agreement with the raman experimental data .
the two b@xmath15 phonon frequencies agree fairly well with other theoretical calculation .
the fitted elastic constants @xmath16 are also in good agreement with the experiment .
the resulting four planar force - constants are given in table [ tab2 ] .
the calculated phonon dispersion based on the obtained planar force - constants for wurtzite gan and aln for the longitudinal polarization along the [ 0001 ] direction are shown in fig .
it is found that the four planar - force - constants can give very satisfactory phonon dispersion of both gan and aln compared with other three - dimensional calculations.@xcite
from fig . [ fig2 ] some general features of phonon modes in hexagonal gan / aln sl s can be postulated .
if a sl mode is located inside gan or aln continuum while gan and aln phonon continua do not overlap , vibrations of this mode will sharply be confined to one of the constituent layers since the other constituent can not sustain this kind of vibrations . if a mode is located in a region where gan and aln phonon continua overlap the mode will be a resonant , quasi - confined one because the vibrations in another constituent layers will be excited .
if there is a sl mode with its frequency lies neither in gan continuum nor aln continuum the vibrations of this mode will be constrained in the interface region .
it is a microscopic interface mode . with the pfcm introduced in the above section phonons in any hexagonal ( 0001)-(gan)@xmath17/(aln)@xmath18
sl s can be studied , where @xmath19 and @xmath20 are the number of bilayers@xcite of gan and aln , respectively . for gan
/ aln sl s , the planar force - constants are assumed to have the same values as corresponding bulk except that across an interface the planar force - constants are taken to be the average of those for gan and aln .
the calculated phonon dispersion and vibrational patterns is given in fig .
[ fig3 ] for a ( 0001)-(gan)@xmath21/(aln)@xmath21 sl . it is obvious that the vibrations of modes 1 - 3 are sharply confined to aln layers .
they are aln - like modes originated from the a@xmath15 branch of aln .
there is nearly no dispersion for these modes due to the strong confinement of vibrations .
mode 4 is also an aln - like mode , but it is from the optical b@xmath15 branch of aln .
the frequency of mode 5 is slightly above the a@xmath15 branch of gan .
it is an aln - like mode , from the b@xmath15 branch of aln .
finite excitations in gan layers exist near the interfaces .
modes 6 - 9 are resonant , quasi - confined gan - like modes due to the fact that their frequencies are inside the overlap region of the b@xmath15 branch of aln and a@xmath15 branch of gan .
vibrations in aln layers could be excited .
consequently , there exists finite dispersion for these modes .
mode 10 is a gan - like mode with vibrations confined to gan layers .
modes 11 - 13 are aln - like modes and they are from the acoustic b@xmath15 branch of aln .
the remaining modes are folded acoustic modes .
the minigaps in the zone - center and zone - boundary are due to the difference of the elastic properties of gan and aln .
no interface modes are found in hexagonal gan / aln sl s .
this may stem from the fact that there are no _
new bonds _ across the interface . for a confined mode the vibrations
are sharply confined to one of constituent layers . in consequence
, the confined mode can be viewed as the standing wave , described by an effective bulk wave - vector @xmath22 where @xmath20 is the nominal number of bilayers , @xmath23 is the order of a confined mode , @xmath24 is the nearest - neighbor cation - cation or anion - anion distance and @xmath25 is a parameter describing the degree of penetration of the vibrations of a confined mode into the adjacent constituent . the frequency of the confined mode can be derived simply from the bulk dispersion @xmath26 instead solving the secular equation about the dynamical matrix if the parameter @xmath25 is properly chosen .
the above discussions are only valid for the confined modes not for the resonant , quasi - confined modes or the extended modes . to test the above idea we give in fig .
[ fig4 ] the comparison of the bulk dispersion and the frequencies of the confined modes of a ( gan)@xmath27/(aln)@xmath28 sl mapped according to eq .
( [ conf ] ) .
a good correspondence is established between the confined modes and the bulk dispersions for both gan and aln when @xmath25 is chosen to be 1 for all branches .
this means that the vibrations of gan or aln confined modes extend about one bilayer into the adjacent constituent . for those resonant , quasi - confined modes
significant vibrations exist in the other constituent layers .
their frequencies are somewhat dependent on the thickness of the constituents . therefore standing wave picture
is no longer valid .
this can be clearly seen from fig .
in the present work a pfcm is developed for wurtzite gan and aln for the longitudinal phonons propagating along the [ 0001 ] direction .
we apply the pfcm to the study of the phonons in hexagonal ( 0001)-gan / aln sl s .
there are confined modes and resonant , quasi - confined modes depending on their frequencies .
no interface modes are found in this system .
the vibrational confinement is investigated . for confined modes either
gan - like or aln - like their vibrations can be viewed as standing waves described by an effective wave - vector .
their frequencies can be obtained directly from the bulk dispersion .
s. nakamura , m. senoh , s. nagahama , n. iwasa , t. yamada , t. matsushita , h. kiyoku , and y. sugimoto , jpn .
part 2 * 35 * , l74 ( 1996 ) ; _
ibid_. , l217 ( 1996 ) ; k. itaya , m. onomura , j. nishio , l. sugiura , s. saito , m. suzuki , j. rennie , s. nunoue , m. yamamoto , h. fujimoto , y. kokubun , y. ohba , g. hatakoshi , and m. ishikawa , _ ibid_. , l1315 ( 1996 ) ; i. akasaki , s. sota , h. sakai , t. tanaka , m. koike , and h. amano , electron . lett . * 32 * , 1105 ( 1996 ) .
.comparison of the fitted phonon frequencies at the zone - center and elastic constants of wurtzite gan and aln with other experimental and theoretical results .
the phonon frequencies are in units of @xmath29 and elastic constants are in units of gpa .
[ cols= " < , < , < , < , < , < " , ] | a planar force - constant model is developed for longitudinal phonons of wurtzite gan and aln propagating along the [ 0001 ] direction .
the proposed model is then applied to the study of the phonon modes in hexagonal ( 0001)-gan / aln supperlattices in longitudinal polarization .
the confinement of the superlattice phonon modes is discussed . |
weak decays of @xmath3 mesons encode valuable information about the quark - flavour sector of the standard model ( sm ) of particle physics . the conceptually simplest observables are branching ratios , which describe the probability for the considered decay to occur .
measurements of @xmath0-decay branching ratios at hadron colliders would require precise knowledge of the @xmath0 production cross section , which is not available , and rely therefore on experimental control channels and the ratio of the @xmath4 fragmentation functions ( for a detailed discussion , see ref .
@xcite ) . at the @xmath5 @xmath6 factories operating at the @xmath7 resonance , @xmath0-decay
branching ratios can be extracted since the total number of produced @xmath0 mesons can be determined separately @xcite . the neutral @xmath0 mesons exhibit @xmath3@xmath8 mixing . measuring the time - dependent angular distribution of the @xmath9 decay products @xcite , lhcb has recently established a non - vanishing difference @xmath10 between the decay widths of the @xmath0 mass eigenstates @xcite : @xmath11 where @xmath12 is the inverse of the average @xmath0 lifetime @xmath13 . since a discrete ambiguity
could also be resolved @xcite , we are left with the sign in ( [ ys ] ) , which agrees with the sm expectation . a sizable value of @xmath10 was theoretically expected since decades @xcite . in view of the sizable @xmath10
, special care has to be taken when dealing with the concept of a branching ratio , and the question of how to convert measured experimental " @xmath0-decay branching ratios into theoretical " @xmath0 branching ratios arises .
this issue is the central topic of this writeup , summarizing the results of refs .
a special emphasis will be put on the rare decay @xmath1 .
the untagged rate of a @xmath0 decay , were no distinction between initially present @xmath3 or @xmath8 mesons is made , is a sum of two exponentials : @xmath14 which can be rewritten as @xmath15.\label{untagged-2}\end{aligned}\ ] ] here the parameter @xmath16 was introduced in ( [ ys ] ) , and @xmath17 is an observable depending on the final state @xmath18 .
the branching ratios given by experiments are extracted from total event yields , without taking time information into account , and can be defined as follows @xcite : @xmath19
@xmath20 = \frac{\tau_{b_s}}{2}\left(r^f_{\rm h } + r^f_{\rm l}\right ) \left[\frac{1 + { \cal a}^f_{\delta\gamma}\ , y_s}{1-y_s^2 } \right].\ ] ] on the other hand , theorists usually consider and calculate the following cp - averaged branching ratios : @xmath21 where the @xmath3@xmath8 oscillations are switched off " by choosing @xmath22 .
it should be noted that this @xmath0 branching ratio concept allows a straightforward comparison with branching ratios of decays of @xmath23 or @xmath24 mesons through the @xmath25 flavour symmetry of strong interactions .
the experimental branching ratios defined in ( [ br - exp ] ) can be converted into the theoretical branching ratios defined in ( [ br - theo ] ) by means of @xmath26 { \rm br}\left(b_s \to f\right)_{\rm exp},\label{brratio-1}\end{aligned}\ ] ] where the term in square brackets would equal one for a vanishing @xmath0 decay width difference . however , for the experimental value of @xmath16 in ( [ ys ] ) , the theoretical @xmath27 branching ratios can differ from their experimental counterparts by up to @xmath28 , depending on the the final state @xmath18 .
using theoretical input , in particular the @xmath25 flavour symmetry , the @xmath29 observables can be estimated for specific decays ( for examples of recent analyses , see refs .
@xcite ) , as compiled in ref . @xcite .
once time information for the untagged @xmath0 decay data sample becomes available , the theoretical input for determining @xmath30 can be avoided in the extraction of the theoretical branching ratio ( [ br - theo ] ) @xcite .
using the effective @xmath0 decay lifetime @xmath31 , \label{taueff}\end{aligned}\ ] ] we obtain @xmath32{\rm br}\left(b_s \to f\right)_{\rm exp } , \label{brratiot}\end{aligned}\ ] ] where only measurable quantities appear on the right - hand side .
the measurement of effective @xmath0 decay lifetimes is hence an integral part of the extraction of the theoretical branching ratios ( [ br - theo ] ) from the data and not only an interesting topic to constrain the @xmath3@xmath8 mixing parameters @xcite .
the use of the theoretically clean relation in ( [ brratiot ] ) is advocated for the compilation of @xmath0 decay properties in particle listings . for a discussion of experimental subtleties related to the measurement of @xmath0 decay branching ratios and effective lifetimes ,
the reader is referred to ref .
@xcite .
the branching ratio measurements of @xmath33 decays into two vector mesons , such as @xmath34 , @xmath35 and @xmath36 , are also affected by the sizable width difference @xmath10 @xcite . here
an angular analysis of the decay products of the vector mesons has to be performed in order to disentangle the cp - even ( @xmath37 ) and cp - odd ( @xmath38 ) final states , with @xmath39 the experimental branching ratios can then be converted into the theoretical branching ratios through @xmath40 \mbox{br}_{\rm exp}^{vv}\ ] ] with the help of theoretical information on the @xmath41 observables , or by means of the relation @xmath42
f_{vv , k}^{\rm exp}\ ] ] utilizing effective lifetime measurements @xcite .
a key probe of new physics ( np ) is the rare decay @xmath43 , which receives only loop contributions from box and penguin topologies in the sm . here
the ( theoretical ) branching ratio is predicted as follows @xcite : @xmath44 in the presence of np , the branching ratio may be affected by new particles in the loops or new contributions at the tree level @xcite .
the error of ( [ br - sm ] ) is dominated by lattice qcd input for non - perturbative physics @xcite .
the limiting factor for the @xmath1 branching ratio measurement at hadron colliders is the ratio @xmath45 , where the fragmentation functions @xmath46 describe the probability that a @xmath47 quark fragments in a @xmath48 meson . a new method for determining @xmath45 using nonleptonic @xmath49 , @xmath50
, @xmath51 decays @xcite was recently implemented at lhcb @xcite , with a result in good agreement with measurements using semileptonic decays @xcite .
the @xmath52-breaking form - factor ratio entering the non - leptonic method has recently been calculated with lattice qcd @xcite .
searches for the @xmath1 decay were performed by the cdf @xcite , d0 @xcite , atlas @xcite , cms @xcite and lhcb collaborations .
the latter experiment has recently reported the currently most stringent upper bound on the branching ratio , which corresponds to @xmath53 at the 95% confidence level @xcite , and is approaching the sm prediction ( [ br - sm ] ) .
for a recent review of the experimental studies , see ref .
@xcite . in the analyses of the @xmath1 decay ,
the impact of @xmath10 was not taken into account . in view of the discussion in the previous section ,
the question arises how the sizable value of @xmath10 affects the theoretical interpretation of the @xmath1 data and whether we can actually take advantage of this decay width difference @xcite . in order to address this issue
, we use the general low - energy effective hamiltonian describing the decay @xmath54 as the starting point . using a notation similar to ref .
@xcite , where a model - independent np analysis was performed , it can be written as @xmath55,\label{heff}\end{aligned}\ ] ] where @xmath56 and @xmath57 are the fermi and qed fine - structure constants , respectively , and the @xmath58 are elements of the cabibbo kobayashi
maskawa ( ckm ) matrix .
the short - distance physics is encoded in the wilson coefficients @xmath59 , @xmath60 of the four - fermion operators @xmath61 where @xmath62 denotes the @xmath47-quark mass , @xmath63 , and the @xmath64 are obtained from the @xmath65 through the replacements @xmath66 .
it should be noted that only operators with non - vanishing contributions to @xmath54 are included in ( [ heff ] ) ; in particular the matrix elements of operators involving the @xmath67 vector current vanish .
the hadronic structure of the leptonic @xmath54 decay is very simple and can be expressed in terms of a single , non - perturbative parameter , which is the @xmath0-meson decay constant @xmath68 @xcite . in the sm
, only the @xmath69 operator contributes with a real wilson coefficient @xmath70 , leading to the prediction in ( [ br - sm ] ) .
the sensitivity to ( pseudo-)scalar lepton densities entering the @xmath71 and @xmath72 operators is an outstanding feature of the @xmath54 channel .
the corresponding wilson coefficients are still largely unconstrained , thereby leaving ample space for np @xcite . for the calculation of the @xmath73 observables ,
it is convenient to go to the rest frame of the decaying @xmath8 meson and to use the notation @xmath74 to distinguish between the left - handed ( @xmath75 ) and right - handed ( @xmath76 ) muon helicity configurations . in this
setting , the @xmath77 and @xmath78 states are simply related to each other through cp transformations .
thanks to @xmath3@xmath8 mixing , we get interference effects between the @xmath79 and @xmath80 decay processes that are described by the observable @xmath81.\ ] ] here @xmath82 is the @xmath3@xmath8 mixing phase , whereas @xmath83 denotes a convention - dependent phase which is associated with cp transformations @xcite . expressing the @xmath84 decay amplitude as @xmath85 and performing appropriate cp transformation results eventually in the following expression @xcite : @xmath86,\ ] ] where all convention - dependent quantities ( such as the @xmath83 phases ) cancel , @xmath87 , and @xmath88 @xmath89 these combinations of wilson coefficient functions
have been introduced in such a way that we simply have @xmath90 and @xmath91 in the sm , whereas @xmath92 and @xmath93 carry , in general , non - trivial cp - violating phases @xmath94 and @xmath95 ( see also ref .
it should be noted that the non - perturbative @xmath0-meson decay constant @xmath96 , which arises in the parametrization of ( [ ampl ] ) and affects the sm prediction ( [ br - sm ] ) , cancels in the observable ( [ xi - obs ] ) . before having a closer look at the branching ratio , it is interesting to consider the following time - dependent rate asymmetries , which require tagging information and knowledge of the muon helicity @xmath97 : @xmath98 @xmath99 here @xmath100 denotes the mass difference between the heavy and light @xmath0 mass eigenstates while @xmath16 is given in ( [ ys ] ) . neglecting the impact of @xmath10 , such cp asymmetries were considered for @xmath101 decays within various np scenarios in the previous literature @xcite .
the observables entering ( [ asym-1 ] ) are governed by @xmath102 in ( [ xi - obs ] ) and take the following expressions @xcite : @xmath103\ ] ] @xmath104 @xmath105 which are theoretically clean .
note that @xmath106 and @xmath107 do not depend on the muon helicity @xmath97 . in the discussion given above
, it was assumed that np enters only through the wilson coefficients governing the @xmath54 decay and that the @xmath3@xmath8 mixing phase @xmath108 takes its sm value @xmath109 , which is cancelled in ( [ xi - def ] ) through the ckm factors of the ratio of decay amplitudes . in ( [ s - lam ] ) and ( [ adg - lam ] ) , np in @xmath3@xmath8 mixing can straightforwardly be included through the replacements @xmath110 .
the lhcb data for cp violation in @xmath111 decays already constrain @xmath112 to the few - degree level @xcite . consequently , this effect is negligible from the practical point of view for the following considerations .
it is difficult to measure the muon helicity . in order to circumvent this problem
, we consider the @xmath113 rate and its counterpart for initially present @xmath8 mesons , which can be combined into the cp asymmetry @xmath114 since a non - zero value would immediately signal new cp - violating phases , it would be most interesting to measure this asymmetry .
this feature was recently highlighted in ref .
@xcite for minimal @xmath115 models @xcite .
unfortunately , despite the independence on the muon helicity , this is still challenging from the practical point of view as tagging and time information are required .
an analogous expression holds for the rare @xmath116 decays , where @xmath117 is negligibly small . from the practical point of view , the branching ratio extracted from untagged data samples , ignoring the decay - time information , is the first measurement : @xmath118 here the untagged @xmath119 rate is given in general terms in ( [ untagged-1 ] ) and ( [ untagged-2 ] ) . since @xmath120 in ( [ adg - lam ] )
does actually not depend on the muon helicity , i.e. @xmath107 , we can apply ( [ brratio-1 ] ) to extract the theoretical branching ratio from the experimental branching ratio ( [ defbrexp ] ) : @xmath121 { \rm br}(b_s \to \mu^+\mu^-)_{\rm exp}.\label{brratio}\end{aligned}\ ] ] the former is considered and calculated by the theoretical community ( see , e.g. , refs .
@xcite ) , and satisfies @xmath122 ) , @xmath30 depends sensitively on np entering through the wilson coefficients which govern the @xmath1 channel .
consequently , this observable is currently unknown .
varying @xmath123 $ ] yields @xmath124 which has to be added to the experimental error of ( [ defbrexp ] ) .
on the other hand , within the sm , we have the theoretically clean prediction @xmath125 .
if we rescale the theoretical sm branching ratio in ( [ br - sm ] ) correspondingly by a factor of @xmath126 and use ( [ ys ] ) , we obtain @xmath127 this is the sm branching ratio reference value for the comparison with the experimental branching ratio ( [ defbrexp ] ) . once the @xmath43 decay has been observed and more experimental data become available , it is possible to include also the decay time information in the analysis so that the effective @xmath43 lifetime @xmath128 , which is defined as in ( [ taueff ] ) , can be measured . since @xmath30 in ( [ adg - lam ] )
does not depend on the muon helicity , this observable can be extracted from the effective lifetime with the help of the relation @xmath129,\ ] ] and results in @xmath130 { \rm br}\left(b_s \to \mu^+\mu^-\right)_{\rm exp}.\label{brmumu - correct}\end{aligned}\ ] ] this expression takes the same form as ( [ brratiot ] ) and allows the conversion of the experimental @xmath131 branching ratio into its theoretical counterpart , irrespective of whether there are np contributions present or not .
consequently , the error in ( [ br - error ] ) can then be eliminated .
the effective @xmath43 lifetime and the extraction of @xmath30 from untagged data samples is an important new measurement for the high - luminosity upgrade of the lhc .
extrapolating from the currently available analyses of the effective @xmath132 and @xmath133 lifetimes performed by the cdf and lhcb collaborations , a precision of @xmath134 or better appears feasible @xcite .
detailed experimental studies of this exciting new feature of the @xmath43 channel are strongly encouraged .
looking at the expression for @xmath135 in ( [ adg - lam ] ) , it is obvious that this observable and the effective lifetime @xmath128 may well be affected by np .
the @xmath10 effects propagate also into the constraints on np parameters that can be obtained from the comparison of the experimental information on the @xmath131 branching ratio with the sm branching ratio , where it is useful to introduce @xcite @xmath136[r - def ] using ( [ adg - lam ] ) and ( [ brratio ] ) , the ratio @xmath137 takes the form @xmath138 \left ( |p|^2 + |s|^2\right)}\nonumber\\ & & \hspace*{-0.9truecm}=\left[\frac{1+y_s\cos2\varphi_p}{1-y_s^2 } \right ] |p|^2 + \left[\frac{1-y_s\cos2\varphi_s}{1-y_s^2 } \right ] |s|^2.\label{r - expr}\end{aligned}\ ] ] combining ( [ br - sm ] ) and ( [ lhcb - bound ] ) yields the bound @xmath139 , where the theoretical uncertainty of the sm prediction of the @xmath1 branching ratio was neglected . the ratio @xmath137 would fix a circle in the @xmath140@xmath141 plane for @xmath142 , i.e. for a vanishing @xmath0 decay width difference . on the other hand , for non - zero values of @xmath16 , @xmath137 can be converted into ellipses which depend on the cp - violating phases @xmath143 . as the latter quantities are in general unknown , @xmath137 results in a circular band , with the upper bounds @xmath144 .
since the @xmath145 and @xmath146 contributions can not be separated through experimental information on @xmath137 , as can be seen in ( [ r - expr ] ) , there may still significant np contributions be present in @xmath43 , even if the branching ratio should eventually be measured close to the sm expectation . as was pointed out in ref .
@xcite , the measurement of the effective lifetime @xmath128 and the associated untagged @xmath30 observable allows a resolution of this situation .
the point is that @xmath147 fixes a straight line through the origin in the @xmath140@xmath141 plane . for illustrations ,
the reader is referred to the figures shown in ref .
@xcite . in the most recent analyses of the constraints on np parameter space that are implied by the experimental upper bound on the @xmath131 branching ratio for various extensions of the sm ,
authors have now started to take the effect of @xmath10 into account ( see , for instance , the papers listed in refs .
the non - vanishing width difference of the @xmath0-meson system , which has recently been established by lhcb , leads to subtleties in the extraction of @xmath0 branching ratio information from the data but offers also new observables .
the differences between the experimental and theoretical branching ratios can be as large as @xmath28 , depending on the final state .
both branching ratios can be converted into each other either through theoretical considerations or through the measurement of the effective @xmath27 decay lifetimes . as the latter involves only experimental data , it is generally the preferred avenue .
the rare decay @xmath43 is also affected by @xmath10 , where the theoretical branching ratio in ( [ br - sm ] ) has to be rescaled by @xmath126 for the comparison with the experimental branching ratio , resulting in the sm reference value of @xmath148 .
thanks to @xmath10 , the @xmath1 decay offers a new observable , which is the effective lifetime @xmath128 .
it allows the inclusion of the @xmath10 effects in the conversion of the experimental into the theoretical branching ratio .
moreover , it offers also a new , theoretically clean np probe that may still show large np effects , in particular those originating from the ( pseudo-)scalar @xmath149 densities entering the four - fermion operators .
this observable may even show np should the @xmath1 branching ratio be found close to the sm prediction .
the measurement of @xmath128 and the associated @xmath30 observable is an exciting new topic for the high - luminosity upgrade of the lhc .
detailed feasibility studies are strongly encouraged .
i am grateful to the organizers , in particular giulia ricciardi , for hosting and inviting me to another excellent meeting of the capri flavour physics workshop series , and would like to thank my phd students and colleagues for the enjoyable collaboration on the topics discussed above .
r. fleischer , n. serra and n. tuning , phys .
d * 82 * ( 2010 ) 034038 [ arxiv:1004.3982 [ hep - ph ] ] .
a. drutskoy _ et al .
_ ( belle collaboration ) , phys . rev .
d * 76 * ( 2007 ) 012002 [ hep - ex/0610003 ] .
m. pepe altarelli , these proceedings .
r. aaij _ et al . _
[ lhcb collaboration ] , phys .
* 108 * ( 2012 ) 241801 [ arxiv:1202.4717 [ hep - ex ] ] . for a recent review ,
see a. lenz , arxiv:1205.1444 [ hep - ph ] . k. de bruyn , r. fleischer , r. knegjens , p. koppenburg , m. merk and n. tuning , phys .
d * 86 * ( 2012 ) 014027 [ arxiv:1204.1735 [ hep - ph ] ] .
k. de bruyn , r. fleischer , r. knegjens , p. koppenburg , m. merk , a. pellegrino and n. tuning , phys .
* 109 * ( 2012 ) 041801 [ arxiv:1204.1737 [ hep - ph ] ] .
i. dunietz , r. fleischer and u. nierste , phys .
d * 63 * ( 2001 ) 114015 [ hep - ph/0012219 ] .
r. fleischer and r. knegjens , eur .
j. c * 71 * ( 2011 ) 1789 [ arxiv:1109.5115 [ hep - ph ] ] ; r. knegjens , these proceedings .
r. fleischer , r. knegjens and g. ricciardi , eur .
j. c * 71 * ( 2011 ) 1832 [ arxiv:1109.1112 [ hep - ph ] ] ; eur .
j. c * 71 * ( 2011 ) 1798 [ arxiv:1110.5490 [ hep - ph ] ] .
k. de bruyn , r. fleischer and p. koppenburg , eur .
j. c * 70 * ( 2010 ) 1025 [ arxiv:1010.0089 [ hep - ph ] ] .
r. aaij _ et al . _
[ lhcb collaboration ] , phys .
b * 709 * ( 2012 ) 50 [ arxiv:1111.4183 [ hep - ex ] ] .
s. descotes - genon , j. matias and j. virto , phys .
d * 85 * ( 2012 ) 034010 [ arxiv:1111.4882 [ hep - ph ] ] .
a. j. buras , pos beauty * 2011 * , 008 ( 2011 ) [ arxiv:1106.0998 [ hep - ph ] ] .
for a recent review , see a. j. buras and j. girrbach , arxiv:1204.5064 [ hep - ph ] .
e. gamiz , these proceedings .
r. fleischer , n. serra and n. tuning , phys .
d * 83 * ( 2011 ) 014017 [ arxiv:1012.2784 [ hep - ph ] ] .
r. aaij _ et al . _
[ lhcb collaboration ] , phys .
lett . * 107 * ( 2011 ) 211801 [ arxiv:1106.4435 [ hep - ex ] ] .
r. aaij _ et al . _
[ lhcb collaboration ] , phys .
d * 85 * ( 2012 ) 032008 [ arxiv:1111.2357 [ hep - ex ] ] .
j. a. bailey , a. bazavov , c. bernard , c. m. bouchard , c. detar , d. du , a. x. el - khadra and j. foley _ et al .
_ , phys .
d * 85 * ( 2012 ) 114502 [ arxiv:1202.6346 [ hep - lat ] ] .
t. aaltonen _ et al . _ [ cdf collaboration ] , phys . rev .
* 107 * ( 2011 )
191801 [ arxiv:1107.2304 [ hep - ex ] ] . v. m. abazov _ et al . _ [ d0 collaboration ] , phys .
b * 693 * ( 2010 ) 539 [ arxiv:1006.3469 [ hep - ex ] ] .
g. aad _ et al . _ [ atlas collaboration ] , phys .
b * 713 * ( 2012 ) 387 [ arxiv:1204.0735 [ hep - ex ] ] ; m. bona , these proceedings .
s. chatrchyan _ et al .
_ [ cms collaboration ] , jhep * 1204 * ( 2012 ) 033 [ arxiv:1203.3976 [ hep - ex ] ] ; g. tonelli , these proceedings .
r. aaij _ et al . _ [ lhcb collaboration ] , phys .
* 108 * ( 2012 ) 231801 [ arxiv:1203.4493 [ hep - ex ] ] ; j. albrecht , these proceedings .
j. albrecht , arxiv:1207.4287 [ hep - ex ] .
w. altmannshofer , p. paradisi and d. m. straub , jhep * 1204 * ( 2012 ) 008 [ arxiv:1111.1257 [ hep - ph ] ] .
r. fleischer , phys .
* 370 * ( 2002 ) 537 [ hep - ph/0207108 ] .
huang and w. liao , phys .
b * 525 * ( 2002 ) 107 [ hep - ph/0011089 ] ; phys . lett .
b * 538 * ( 2002 ) 301 [ hep - ph/0201121 ] .
a. dedes and a. pilaftsis , phys . rev .
d * 67 * ( 2003 ) 015012 [ hep - ph/0209306 ] .
h. chankowski , j. kalinowski , z. was and m. worek , nucl .
b * 713 * ( 2005 ) 555 [ hep - ph/0412253 ] .
a. j. buras and j. girrbach , arxiv:1206.3878 [ hep - ph ] .
r. barbieri , these proceedings .
o. buchmueller , r. cavanaugh , m. citron , a. de roeck , m. j. dolan , j. r. ellis , h. flcher and s. heinemeyer _ et al .
_ , arxiv:1207.7315 [ hep - ph ] ; t. hurth and f. mahmoudi , arxiv:1207.0688 [ hep - ph ] ; w. altmannshofer and d. m. straub , arxiv:1206.0273 [ hep - ph ] ; d. becirevic , n. kosnik , f. mescia and e. schneider , arxiv:1205.5811 [ hep - ph ] ; f. mahmoudi , s. neshatpour and j. orloff , arxiv:1205.1845 [ hep - ph ] ; t. li , d. v. nanopoulos , w. wang , x. -c .
wang and z. -h .
xiong , jhep * 1207 * ( 2012 ) 190 [ arxiv:1204.5326 [ hep - ph ] ] . | the lhcb experiment has recently established a sizable width difference between the mass eigenstates of the @xmath0-meson system .
this phenomenon leads to a subtle difference at the 10% level between the experimental branching ratios of @xmath0 decays extracted from time - integrated , untagged data samples and their theoretical counterparts .
measuring the corresponding effective @xmath0-decay lifetimes , both branching ratio concepts can be converted into each other .
the rare decay @xmath1 and the search for new physics through this channel is also affected by this effect , which enhances the standard - model reference value of the branching ratio by @xmath2 , while the effective lifetime offers a new observable to search for physics beyond the standard model that is complementary to the branching ratio .
@xmath0 decays , branching ratios , effective lifetimes , new physics |
spectrum sensing plays an important role in cognitive radio ( cr ) systems , and cyclostationary feature detection is one of the main technologies for spectrum sensing in low snr cases [ 1 - 2 ] .
the idea of cyclic detection is that one cr node samples the rf signals , transform the time domain signals into cyclic frequency domain , then decide primary user s occupancy of a target band regarding whether the cyclic spectrum on a significant cyclic frequency is above a certain threshold [ 2 ] .
+ although the simulation results of cyclostationary feature detection has been studied [ 3 - 5 ] , less attention has been paid to analyze the noise distribution on cyclic frequency domain . consequently , theoretical function between detection threshold @xmath1 and miss detection probability @xmath0 is not available , which makes it difficult for practical system design . for comparison , the theoretical function between detection threshold @xmath1 and miss detection probability @xmath0 of energy detection is available by using central and non - central chi - square distribution to model the distribution of the observation variable affected by time domain gaussian noise [ 6 ] , which makes energy detection to be a practical method for spectrum sensing .
however , its performance in low snr cases is much poorer than cyclostationary feature detection [ 7 ] . in this paper ,
noise distribution on cyclic frequency domain has been analyzed and generalized extreme value distribution is found to be a precise match of the observation variable affected by time domain gaussian noise . a fast cyclic frequency domain feature detection algorithm [ 7 ] has been introduced to evaluate the coincidence between theoretical roc curve and the simulated roc curve , which proves the reliability of theoretical distribution model and feasibility of practical system design .
+ the rest part of the paper is organized as follows : section ii describes the system model of cyclostationary feature detection .
noise distribution on cyclic frequency domain is analyzed in section iii .
a fast cyclic frequency domain feature detection algorithm has been introduced in section iv .
simulation results are given in section v. finally , conclusions are drawn in section vi .
the spectrum sensing problem can be modeled as hypothesis testing . it is equivalent to distinguishing between the following two hypotheses : @xmath2 @xmath3 , @xmath4 and @xmath5 denote the received signal , the primary user s transmit signal and the gaussian noise , respectively . @xmath6 and @xmath7 represent the hypothesis that the primary user is active or inactive . due to the existence of noise , a certain threshold @xmath1 should be set to decide whether a primary user is active or not .
probability of detection ( @xmath8 ) and false alarm ( @xmath0 ) are defined to evaluate the detection performance : @xmath9 the goal of detection is to maximize @xmath8 while maintain a given @xmath0 .
+ when feature detection is applied , the detection model ( 1 ) changes into : @xmath10 @xmath11 is the spectrum correlation density ( scd ) of the received signal @xmath3 , @xmath12 and @xmath13 is the scd of @xmath4 and @xmath5 , respectively [ 8 ] . theoretically , gaussian noise @xmath5 is not a cyclostationary statistic process , then @xmath14 when @xmath15 [ 8 ] . as for cyclostationary signal @xmath4
, there is a significant frequency set \{@xmath16 } , on which @xmath17 . due to the ideal non - cyclostationary characteristic of gaussian noise ,
any pre - set threshold @xmath1 on a significant frequency will lead to @xmath18 and @xmath19 .
in practice , scd is calculated for limited length signals , therefore , @xmath20 when @xmath15 [ 9 ] . as shown in fig.1 ,
the background noise is obvious on @xmath21 square when calculating scd of a noise interfered am modulated signal . in order to analyze the background noise distribution , a limited - length gaussian noise sequence @xmath22
is considered : @xmath23 where @xmath24 is the length of the analysis window , @xmath25 is the index of analysis window .
noise data in each analysis window are transformed into cyclic frequency domain , and then mapped from @xmath21 square to @xmath26 axis through the following expression : @xmath27 for each cycle frequency @xmath16 , the cyclic spectrum value is aligned to a set \{@xmath28 } , @xmath29 . according to the extreme definition of @xmath30 in ( 5 ) , generalized extreme value ( gev ) distribution is adopted to model the cyclic frequency domain noise [ 10 ] . the density function is : + when @xmath31 @xmath32 when @xmath33 @xmath34 where @xmath35 is the shape parameter , @xmath36 is the position parameter , @xmath37 is the scale parameter . and
the parameters @xmath35 , @xmath36 , @xmath38 can be estimated by maximum likelihood estimation based on the noise sequence \{@xmath39}. for most cases , @xmath40 , then the likelihood function is defined as follows : @xmath41 let @xmath42,@xmath43 then : @xmath44 where @xmath45 and @xmath46 are the estimated value of @xmath36 and @xmath38 . by solving ( 9 ) , @xmath45 and @xmath46
are obtained .
+ after that , the likelihood function for @xmath35 is defined as : @xmath47\nonumber\\ & & -\sum_{i=1}^n[1+\kappa(\frac{n_i-\hat{u}}{\hat{\sigma}})]^{-\frac{1}{\kappa}}\end{aligned}\ ] ] let @xmath48 , the estimation of @xmath49 can be obtained by solving ( 10 ) .
fig.2 shows the block diagram of a cyclic frequency domain feature detector [ 7 ] : in this detector , time domain signals are transformed to cyclic frequency domain , and then mapped to @xmath26 axis and get extreme value for each @xmath26 , finally , compared with a pre - set threshold @xmath1 to determine the occupancy of primary user . to decrease the computational complexity of feature detection , only one cycle frequency @xmath16 for a modulated signal
is calculated for spectrum sensing [ 7 ] .
the probability of detection ( @xmath8 ) and false alarm ( @xmath0 ) is defined by : @xmath50 where @xmath51 is cdf of generalized extreme value distribution .
substitution of ( 6 ) and ( 7 ) into ( 11 ) yields : @xmath52 for a pre - set @xmath0 , the threshold can be estimated as : @xmath53 where @xmath54 .
in this section , monte - carlo simulation results are presented to prove the reliability of the upper analytical results between @xmath0 and the threshold @xmath1 .
+ frequency smoothing method in [ 11 - 12 ] is applied to estimate scd of a time domain noise signal .
simulation parameters are listed in table 1 .
length of the analysis window is set to be 4096 and totally 10000 cyclic spectrum values on cyclic frequency @xmath55 , aligned by analysis window index , are considered to evaluate the theoretical curve .
.simulation parameters list [ cols="^,^",options="header " , ] the histogram of the aligned cyclic spectrum values is shown in fig 3 , with compared to gev distribution .
it is proved that the gev distribution precisely match the cyclic frequency domain noise data .
+ for further proof of the proposed model , the curves of _ receiver operating characteristics ( roc ) _ , which are theoretically derived from gev distribution , are plotted to compare with those derived from monte - carlo simulation . as to the theoretical curve ,
@xmath0 is pre - set according to system requirement . by using ( 13 ) , a theoretical threshold @xmath1 is obtained . finally , received signals under hypothesis @xmath6 are compared with the threshold to obtain the statistics result of @xmath8 .
as to the simulated curve , threshold are chosen to be the same as the theoretical threshold \{@xmath1 } , after that , received signals under hypothesis @xmath6 and @xmath7 are compared with each @xmath1 to obtain the statistics results of @xmath8 and @xmath0 .
finally , plot these ( @xmath8,@xmath0 ) points to form a continuous curve .
+ experiments results are shown in fig.4 , we can see that the theoretical roc curve ( red highlighted ) precisely match the simulated roc curve ( green highlighted ) for different received signal power levels .
it is proved that the generalized extreme value distribution is efficient to model the noise distribution on cyclic frequency domain .
in this paper , noise distribution on cyclic frequency domain is studied and generalized extreme value ( gev ) distribution is found to be an efficient method to model the cyclic frequency domain noise . maximum likelihood estimation is applied to estimate the parameters of gev .
sensing threshold is consequently derived from system requirements ( a pre - set @xmath0 ) and theoretical cdf of gev distribution .
monte carlo simulation has been carried out to prove that the simulated roc curve is precisely coincided with the theoretical roc curve .
the project is founded by the corporation research department of huawei technology , the national 863 project of china , no .
2007aa01z237 , and the fund of ministry of science and technology of china , no .
2008dfa11950 .
william a. gardner , and chad m. spooner , `` detection and source location of weak cyclostationary signals : simplifications of the maximum - likelihood receiver '' , ieee trans .
vol-41 , no . 6 , june 1993 .
paul d. sutton , keith e. nolan , and linda e. doyle , `` cyclostationary signatures in practical cognitive radio applications '' , selected areas in commun , ieee journal vol.26 , issue 1 , pp .
13 - 24 , jan . 2008 .
jun ma , guodong zhao , and ye li , `` soft combination and detection for cooperative spectrum sensing in cognitive radio networks , '' ieee transactions on wireless communications , vol . 7 , no . 11 , pp . 4502 - 4507 , november 2008 .
gan xiaoying , xu hao , xu youyun , qian liang , liu jing , `` noise analysis for limited length cyclostationary detection in cognitive radio systems '' , journal of pla university of science and technology,2008 , vol . 9 , no.6 , pp:633 - 636 . | in cognitive radio systems , cyclostationary feature detection plays an important role in spectrum sensing , especially in low snr cases . to configure the detection threshold under a certain noise level and a pre - set miss detection probability @xmath0 , it s important to derive the theoretical distribution of the observation variable . in this paper , noise distribution in cyclic frequency domain has been studied and generalized extreme value ( gev ) distribution is found to be a precise match .
maximum likelihood estimation is applied to estimate the parameters of gev .
monte carlo simulation has been carried out to show that the simulated roc curve is coincided with the theoretical roc curve , which proves the efficiency of the theoretical distribution model . cognitive radio , spectrum sensing , cyclic feature detection , noise distribution on cyclic frequency domain |
stars form in dense cores of dust and gas within molecular clouds , but remain cocooned in the material of their natal envelopes .
though young stars and protostars can be extremely luminous , the enveloping dust absorbs nearly all of the ultraviolet and optical starlight and reradiates it at longer wavelengths . in the submillimetre regime , the dust emission is generally optically thin .
submillimetre observations thus sample all the dust emission within the cloud along the line of sight , and can be used to determine the column density and mass of material present @xcite .
interpretation of the observations , however , depends on the assumed dust grain temperature , composition and size distribution .
the effects of composition and grain size on the dust emission are usually combined into a dust opacity , @xmath4 , which can be reasonably fit empirically with a power law dependence on frequency in the submillimetre . [ cols="^,^,^,^,^,^ " , ] source brightnesses required to determine the dust spectral index @xmath5 in 12 hours of observing using an ifts on scuba-2 .
scuba-2 is the next generation submillimetre instrument currently in development to replace scuba in 2006 @xcite .
an imaging fts ( ifts ) is being developed as an ancillary instrument to scuba-2 @xcite .
based on detector noise calculations alone , the ifts will be at least 10 times more sensitive per pixel than the single bolometer fts used in these observations @xcite .
the extremely accurate sky subtraction possible with the ifts will provide an even more substantial increase in the observational precision .
when also considering the variable atmosphere noise , the per - pixel sensitivity of the ifts increases to at least 100 times greater than in the observations presented in this paper . using the detector sensitivities in @xcite ,
we have calculated that it will be possible to determine in 12 hours of observing with the ifts the spectral index @xmath6 of the submillimetre dust emission at 850 to @xmath2 for sources of only 400mjy / beam in brightness , and to @xmath3 for sources only @xmath7mjy / beam in brightness .
similar calculations for observations at 450 are listed with those at 850 in table [ tab : ifts ] .
we have determined the spectral index of the dust emission of three hot molecular cores solely within the 350ghz ( 850 ) passband using an fts on the jcmt .
we find an average dust spectral index @xmath1 , in agreement with the spectral indices calculated using archived scuba data at 450 and 850 .
these results for @xmath0 are consistent with measurements of the spectral index in other hmc .
the uncertainties in @xmath0 from these observations are dominated by the difficulties in subtracting the very bright and variable submillimetre atmosphere from the data . with better sky subtraction techniques , as will be possible with an imaging fts on scuba-2
, these uncertainties will be greatly reduced .
the per - pixel sensitivity of the imaging fts planned for scuba-2 will be greater by a factor of @xmath8 than that of the single pixel detector used in this study , and will allow the determination of the spectral index of dust emission of sources with brightnesses of only a few hundred mjy / jcmt beam .
we thank the referee , jenny hatchell , for her suggestions which improved this paper .
we also thank all tss involved in fts runs for their invaluable assistance , as well as g. j. tompkins , b. gom and i. chapman for much fts - related support .
additional thanks to g. fuller for providing the source list .
rkf also thanks b. hesman for helpful advice and discussion .
this work was funded in part by the university of victoria ( rkf ) and by grants from nserc canada ( dj , dan and grd ) .
the jcmt is operated by the joint astronomy centre in hilo , hawaii on behalf of the parent organizations particle physics and astronomy research council in the united kingdom , the national research council of canada and the netherlands organization for scientific research .
scuba-2 is a jointly funded project through the jcmt development fund with substantial additional contributions from the uk office of science and technology and the canada foundation for innovation .
99 aanestad , p.a .
1975 , apj , 200 , 30 archibald , e.n . , jenness , t. , holland , w.s .
2002 , mnras , 336 , 1 beuther , h. , schilke , p. & wyrowski , f. 2004 , apj , in press chapman , i. msc thesis , lethbridge , 2001 chapman , i.m . ,
naylor , d.a .
& phillips , r.r . 2004 , mnras , 354 , 621 churchwell , e. , wolfire , m.g . & wood , d.o.s .
1990 , apj , 354 , 247 davis , g.r . , naylor , d.a . ,
griffin , m.j . ,
clark , t.a . & holland , w.s .
1997 , _ icarus _ , 130 , 387 draine , b.t . &
lee , h.m .
1984 , apj , 285 , 89 goldsmith , p.f . ,
bergin , e.a .
, & lis , d.c .
1997 , apj , 491 , 615 gom , b.g . & naylor , d.a . 2004 . in proc .
spie , vol .
5498 , millimeter and submillimeter detectors for astronomy iia , ed .
zmuidzinas , j. , holland , w.s . , withington , s. ( spie ) 695 groesbeck , t.d . , phillips , t.g . & blake , g.a .
1994 , apjs , 94 , 147 hatchell , j. , thompson , m.a . ,
millar , t.j .
& macdonald , g.h .
1998 , a&as , 133 , 29 hatchell , j. , fuller , g.a . ,
millar , t.j . ,
thompson , m.a . , macdonald , g.h .
2000 , a&a .
357 , 637 hatchell , j. & van der tak , f.f.s .
2003 , a&a , 409 , 589 hildebrand , r.h .
1983 , qjras , 24 , 267 hoare , m.g . ,
roche , p.f .
& glencross , w.m . 1991 , mnras , 251 , 584 hogerheijde , m.r . &
sandell , g. 2000 , apj , 534 , 880 holland , w.s . , duncan , w.d . ,
kelly , b.d . , irwin , k.d . , walton , a.j . , ade , p.a.r . , robson , e.i . 2003 , in proc .
spie , vol .
4855 , millimeter and submillimeter detectors for astronomy , ed .
phillips , t.g . , zmuidzina , j. ( spie ) 1 hunter , t.r .
1998 , pasp , 110 , 634 johnstone , d. , boonman , a.m.s .
& van dishoeck , e.f .
2003 , a&a , 412 , 157 kurtz , s. , cesaroni , r. , churchwell , e. , hofner , p. , & walmsley , c.m .
2000 , in protostars & planets iv , ed . v. mannings , a.p .
boss , & s.s .
russell ( tucson : univ .
arizona press ) lis , d.c .
& menten , k.m .
1998 , apj , 507 , 794 naylor , d.a . ,
davis , g.r . ,
griffin , m.j .
, clark , t.a . , gautier , d. & marten , a. 1994 , a&a , 291 , l51 naylor , d.a . ,
davis , g.r .
, gom , b.g . ,
clark , t.a . ,
griffin , m.j .
2000 , mnras , 315 , 622 naylor , d.a . ,
gom , b.g . ,
schofield , i.s . , tompkins , g.j . ,
davis , g.r .
2003 , proc .
spie , in proc .
spie , vol .
4855 , millimeter and submillimeter detectors for astronomy , ed .
phillips , t.g .
, zmuidzina , j. ( spie ) 540 naylor , d.a . &
gom , b.g . 2004 , in proc .
spie , vol .
5159 , imaging spectrometry ix , ed .
shen , s.s . ,
lewis , p.e .
( spie ) 91 naylor , d.a .
, gom , b. g. , ade , p. a. r. , davis , j. e. 1999 , rev .
instruments , 70 , 4097 ossenkopf , v. & henning , th .
1994 , a&a , 291 , 943 osorio , m. , lizano , s. & dalessio , p. 1999
, apj , 525 , 808 pollack , j.b . ,
hollenbach , d. , beckwith , s. , simonelli , d.p . , roush , t. & fong , w. 1994 , apj , 421 , 615 privett g. , jenness t. , matthews h. , 1998
. starlink user note 213.2 serabyn , e. & weisstein , e.w .
1995 , apj , 451 , 238 thompson , m.a .
& macdonald , g.h .
1999 , a&as , 135 , 531 visser , a.e .
, richer , j.s . ,
chandler , c.j .
& padman , r. 1998 , mnras , 301 , 585 walsh , a.j . ,
macdonald , g.h .
, alvey , n.d.s . ,
burton , m.g . , lee , j.k .
2003 , a&a , 410 , 597 williams , s.j .
, fuller , g.a .
& sridharan , t.k .
2004 , a&a , 417 , 115 wood , d.o.s . &
churchwell , e. 1989 , apjs , 69 , 831 | we have used a fourier transform spectrometer on the james clerk maxwell telescope to study the submillimetre continuum emission from dust in three hot molecular cores ( hmc ) .
the spectral index @xmath0 of the dust emission for these sources has been determined solely within the 30ghz wide 350ghz ( 850 ) passband to an accuracy comparable to those determined through multi - wavelength observations .
we find an average @xmath1 , in agreement with spectral indices determined from previous submillimetre observations of these sources and with those determined for hmc in general .
the largest single source of uncertainty in these results is the variability of the atmosphere at 350ghz , and with better sky subtraction techniques we show that the dust spectral index can clearly be determined within one passband to high accuracy with a submillimetre fts . using an imaging fts on scuba-2 , the next generation wide - field submillimetre camera currently under development to replace scuba at the jcmt in 2006 , we calculate that at 350ghz it will be possible to determine @xmath0 to @xmath2 for sources as faint as 400mjy / beam and to @xmath3 for sources as faint as 140mjy / beam .
[ firstpage ] dust instrumentation : spectrographs stars : formation submillimetre |
bose - einstein condensation ( bec ) is a very robust phenomenon . because of bose statistics , non - interacting bosons seek out the lowest single particle energy state and ( below a critical temperature @xmath0 ) condense into it , even though many almost degenerate states may be nearby . at temperature @xmath1
, the condensate contains all the particles in the system @xcite .
although this remarkable phenomenon was originally predicted by einstein for non - interacting systems @xcite , it became understood , starting with the work of f. london @xcite , that it occurs in strongly interacting systems such as @xmath2he . while interactions remove particles from the condensate into other states , the energy gain by macroscopically occupying the lowest energy state ( or another state , such as a vortex ) is sufficiently great that the interactions in a bose condensed system manage only to deplete a fraction of the condensate , but not destroy it @xcite .
this is the case for liquid @xmath2he @xcite as well as dilute gases of bosonic alkali atoms @xcite . in certain situations ,
however , a system does not condense into a _
condensate @xcite . in this paper
we explore the physics of condensation when the ground state can contain several condensates
situations of _ fragmented _ condensation .
this possibility arises naturally from the very concept of bec
i.e. , that the _ non - degenerate _ ground state of a single particle hamiltonian @xmath3 is macroscopically occupied .
how , instead , does condensation take place when the ground state of @xmath3 is degenerate , with two or more states competing simultaneously for condensation ?
what happens if the ground state degeneracy @xmath4 is not just of order unity , but of order @xmath5 , the number of particles ? and what happens if @xmath4 becomes much greater than @xmath5 , or even approaches infinity ?
how do the bosons distribute themselves in these competing levels ?
in all these cases , interaction effects play an important role in determining the structure of the many - body ground state . different types of interactions produce different fluctuations ( such as those of phase , number , or spin ) and lead to different classes of ground states .
exact degeneracy is difficult to achieve , since small external fields or weak tunneling effects are generically present .
such energy - splitting effects lead to a non - degenerate ground state and therefore favor the formation of a single condensate . only when interaction effects dominate over energy splittings can a single condensate break up
. considerations of the effect of ground state degeneracies are not merely theoretical exercises .
rather , such degenracies occur in a wide range of current experiments in cold atoms .
the cases where the degeneracy @xmath4 is of order unity are related to bosons with internal degrees of freedom .
examples include a pseudospin-1/2 bose gas made up , e.g. , of two spin states @xmath6 and @xmath7 ) of @xmath8rb @xcite , and a spin-1 or spin-2 bose gas such as @xmath9na or @xmath8rb in an optical trap @xcite . in the former case ,
even though the two spin states of @xmath8rb are separated by a hyperfine splitting of order ghz , they can be brought to near degeneracy , with @xmath10 , by applying an external rf field . in the spin-1 bose gas ,
the three spin states are degenerate at zero external magnetic field , and @xmath11 .
the case of @xmath12 is encountered in one dimensional geometries @xcite , where the density of states has a power law singularity at low energies .
the case of @xmath13 is realized for bosons in optical lattices @xcite with @xmath4 sites each having a few bosons ; in the limit of zero tunneling , @xmath4 equivalent sites compete for bosons .
the case @xmath14 is realized in rotating bose gases with very large angular momentum , @xmath15 , in a transverse harmonic trap @xcite . as @xmath15 increases , the rotation frequency @xmath16 of the atom cloud approaches the frequency of the transverse harmonic trap , @xmath17 , causing the single particle states to organize into landau orbitals , which become infinitely degenerate as @xmath18 .
the great diversity of phenomena in these experiments is a manifestation of the physics of bose condensation for varying degrees of degeneracy .
as we shall see , when the degeneracy is low ( @xmath19 ) , a single condensate can break up into @xmath4 condensates . following nozires @xcite
, we refer to such a state as a _ fragmented condensate _
( define more precisely below ) .
we caution , however , that there are many sorts of fragmented states .
even degeneracy as small as @xmath10 can give rise to fragmented condensates with distinct properties which depend on whether interaction effects favor phase or number fluctuations .
system with larger degeneracies usually have a larger number of relevant interaction parameters and are more easily influenced by external fields . as a result , they have a greater variety of fragmented states . for degeneracies
@xmath4 comparable to or much greater than the particle number , the system does not have enough particles to establish separate condensates .
instead , interactions effects tend to distribute bosons among different degenerate single particle states in a coherent way , establishing correlations between them . in such _ strongly correlated _
systems , interaction effects obliterate all traces of a conventional singly condensed state .
we stress at the outset that while interaction effects can cause fragmentation in the presence of degenerate single particle states , the presence of near degeneracies does not force the condensate to fragment . in many cases the ground state for a _ macroscopic _
bose system is a conventional single condensate .
creating a fragmented state typically requires carefully tuning the parameters of the system , and whether such a state can occur in practice is very much dependent on the system . in optical lattices and double well systems , where the tunneling between wells can be tuned arbitrarily finely , a fragmented state can easily be achieved .
yet in other systems such as a spin-1 bose gas in a single trap , or a rotating bose gas , the parameter range allowing the existence of fragmented states scales like @xmath20 , making it difficult to realize these ground states unless the number of particles is reduced to @xmath21 or fewer . in this paper , we focus on fragmentation in the ground state of bose systems .
we consider a number of canonical examples ( double - well systems , spin-1 bose gases , and rotating bose gases ) , which have increasing degeneracy in the single particle hamiltonian .
these examples illustrate the origin of fragmentation , the variety of fragmented states , and their key properties .
we will not discuss here fragmentation in optical lattices or in dynamical processes , for they are a sufficiently large subject to require separate discussions .
to begin , we first define condensate fragmentation more precisely , and the discuss general properties of certain classes of fragmented states .
the concept of bose - einstein condensation was generalized to interacting systems by penrose and onsager@xcite in the 1950s , by defining condensation in terms of the single particle density matrix , @xmath22 where @xmath23 creates a scalar boson at position @xmath24 , and @xmath25 is the thermal average at temperature @xmath26 . since @xmath27 is a hermitian matrix with indices @xmath28 and
@xmath29 , it can be diagonalized as , @xmath30 where the @xmath31 are the eigenvalues and @xmath32 the orthonormal eigenfunctions of @xmath27 ; @xmath33 . setting @xmath34 and
integrating over @xmath28 we have @xmath35 , where @xmath5 is the number of particles .
we label the eigenvalues in descending order , i.e. , @xmath36 .
equation ( [ rho1 ] ) implies that if one measures the number of bosons in the single particle state @xmath37 , one finds @xmath38 .
this does not mean that the wavefunction of the many particle interacting system is a product of such single particle eigenstates . unless there are special reasons ( such as strict symmetry constraints , e.g. , translational invariance ) , the eigenfunctions @xmath39 need not be the same as the single particle eigenstates @xmath40 of the single particle hamiltonian of the non - interacting system .
two relevant examples are a homogeneous system of particles in free space , and a system in a harmonic trap . in the former case , where the momentum @xmath41 is a good quantum number , we have @xmath42 , and @xmath43 and @xmath44 are the same plane - wave momentum eigenstates . in an inhomogeneous trapped system
, there is no simple relation between the @xmath39 and the @xmath45 @xcite .
the usual situation of bose condensation corresponds to the one eigenvalue @xmath46 being of order @xmath5 , while other eigenvalues are of order unity , i.e. , @xmath47 or simply @xmath48 where @xmath49 , and @xmath50 denotes terms with eigenvalues @xmath51 . since the macroscopic term in eq .
( [ bec ] ) is identical to the density matrix of the pure single particle quantum state , @xmath52 , the function @xmath53 is often referred to as the macroscopic wavefunction " of the system .
systems in which @xmath27 has only one macroscopic eigenvalue , eq .
( [ bec ] ) , have _
single condensates_. the advantage of the penrose - onsager characterization of bec , eq . (
[ bec ] ) , is that it applies to both interacting and non - interacting systems , since it makes no reference to dynamics .
penrose and onsager also demonstrated the remarkable fact that eq .
( [ bec ] ) holds for a jastrow function , which is a reasonable approximation to the ground state of a system of hard core bosons , therefore substantiating eq .
( [ bec ] ) as a general property of interacting bose systems . the penrose - onsager characterization can be easily generalized to bosons with internal degrees of freedom , labeled by an index @xmath54 . with field operator @xmath55 ,
the single particle density matrix is @xmath56 in a conventional noncondensed system , such as a zero temperature gas of non - interacting fermions , or a high temperature gas of bosons , all the occupation numbers are small : @xmath57 .
a conventional singly condensed system has one large eigenvalue @xmath58 , with all other eigenvalues small , @xmath59 .
a fragmented system has @xmath60 large eigenvalues , @xmath61 .
there is clearly a range of other possiblilites such as having an extremely large number @xmath62 eigenvalues , each of which are of size @xmath63 .
this latter case occurs in an interacting system of one dimensional bosons , and is associated with a phase incoherent _ quasicondensate _ @xcite . before examining the origin of fragmentation in detail ,
let us consider a basic example of fragmentation the nozires model @xcite .
consider a system of @xmath5 bosons each of which have available two internal states ; 1 and 2 .
as we consider in detail later , this model can also be used to describe atoms in a double well potential .
the hamiltonian of nozires model consists solely of an interaction between atoms in the two spin states , @xmath64 where the @xmath65 create bosons in state @xmath66 , and @xmath67 is the number of particles in @xmath68 .
the interaction between the particles can be either repulsive ( @xmath69 ) or attractive ( @xmath70 ) .
the eigenstates have a definite number of particles in each well @xmath71 , with @xmath72 , and energy @xmath73 clearly , for @xmath69 , the ground state is two - fold degenerate , with @xmath74 , or @xmath75 ; these states have single condensates , whose density matrices have eigenvalues 0 and @xmath5 . on the other hand , for @xmath70 , the state with @xmath76
has lowest energy .
fock_-state , @xmath77 has a fragmented condensate ; the corresponding single particle density matrix , @xmath78 has two macroscopic eigenvalues .
one can contrast this fragmented state to the coherent state @xmath79 in which @xmath5 bosons are condensed into the single particle state @xmath80 .
the coherent state is an example of a single condensate , where the single particle density matrix , @xmath81 , is @xmath82 the difference between eqs .
( [ rhoc ] ) and ( [ rhof ] ) is the absence of the off - diagonal term @xmath83 in the latter , which represents the loss of phase coherence in state ( pseudospin ) space .
the fock state ( [ f ] ) is an average over all coherent phase states @xmath84 , as we see from the relation , @xmath85 the latter relation holds for @xmath86 . as we discuss in the next section
, this connection is very useful for understanding the origin of various ground states .
an important implication of this relation is that _ for a macroscopic system , the expectation value of any @xmath87-body operator _ , @xmath88 , _ in the fock state is indistinguishable from that in an ensemble of coherent phase states @xmath89 , as long as @xmath90 _
, @xmath91 this equation , which we shall prove momentarily , shows that by measuring quantities associated with few body operators , one can not distinguish a fock state from an ensemble of coherent states with random phases @xcite .
an illustration of this effect is the interference of two condensates initially well separated from each other .
prior to any measurement process , the system is in a fock state , eq.([f ] ) , since there is no phase relation between the two condensates .
experimentally , in any single shot measurement ( a photo of the interfering region ) , one finds an interference pattern consisting of parallel fringes whose location is specified by a phase @xmath92 , as if the two far away condensates actually had a well defined relative phase with this value @xcite .
the value of @xmath92 , however , varies randomly from shot to shot , so that if one averages over all the measurements , the interference fringes average out , as described simply by eq .
( [ f = c ] ) @xcite . to prove eq .
( [ f = c ] ) , we use eq .
( [ f - c ] ) to write , @xmath93 it is then sufficient to show that @xmath94 vanishes unless the phases @xmath92 and @xmath95 are very close to each other .
we note that @xmath96 for @xmath86 . for @xmath87-body operators of the form @xmath97 , @xmath98 } \langle \phi'_{n - p}| \phi_{n - p}\rangle.\end{aligned}\ ] ] for @xmath99 , we see from eq .
( [ phi - phi ] ) that @xmath100 , from which eq .
( [ f = c ] ) follows .
we now develop three different examples that illustrate the origin of fragmentation .
these examples are chosen to illustrate the increasingly complex behavior of fragmentation when the number of degenerate single particle states increases .
after the model of sec .
[ noz ] , the simplest model with a fragmented ground state is that of bosons in a double well potential with tunneling between the wells . unlike in the previous example
, this model produces two distinct types of fragmented states .
we label the wells by @xmath66 ; we assume that there is only one relevant state in each well , and that particles within a given well have an interaction , @xmath101 , which can be either repulsive ( @xmath102 ) or attractive ( @xmath103 ) .
we take the hamiltonian to be @xmath104 , \label{2site}\ ] ] where the @xmath65 creates a boson in well @xmath68 , and @xmath67 is the number of particles in well @xmath68 .
the first term describes tunneling between the wells via a tunneling matrix element @xmath105 ( which we assume to be real and positive ) .
the form @xmath106 is the usual contact interaction @xmath107 reduced to the single mode in each well . for fixed number of particles @xmath108 , the interaction term
can be written simply as @xmath109 \label{u}\ ] ] this model , simple as it is , has wide applicability to many physical situations : atoms in a double well potential @xcite , internal hyperfine states coupled by electromagnetic fields @xcite , atoms in a rotating toroidal trap @xcite , or wave packets in an optical lattice @xcite . in solving this model
it is useful to write eq .
( [ 2site ] ) in the wigner - schwinger pseudospin representation @xcite .
we introduce the operators @xmath110 which obey the angular momentum commutation relation @xmath111= i\epsilon_{ijk}j_{k}$ ] , @xmath112 , and satisfy @xmath113 the hamiltonian ( [ 2site ] ) can be written in terms of @xmath114 as @xmath115 as we shall see , the hamiltonian ( [ 2site ] ) can be solved exactly .
nonetheless , the mean - field solutions illustrate much of the physics of the true ground state . they also allow one to see the kind of fluctuations about the mean - field state that lead to condensate fragmentation .
the mean - field states are of the form of ( pseudo)spinor condensates @xmath116 where @xmath117 and @xmath118 .
the matrix elements of the density matrix in this state are @xmath119 , @xmath120 , and @xmath121 .
the angles @xmath122 and @xmath92 therefore characterize the density and phase difference between the bosons in the two wells . in the pseudospin language
, the state ( [ n ] ) describes a ferromagnet with total spin , @xmath123 , where @xmath124 is the unit vector with polar angles @xmath125 .
according to eq .
( [ spinh ] ) , its energy is @xmath126 for a repulsive interaction , @xmath102 , @xmath127 is minimum at @xmath128 , @xmath129 ; or @xmath130 .
the mean - field approach therefore selects the non - interacting ground state @xmath131 as optimal .
condensates in the neighborhood of @xmath132 have energy @xmath133 from this result one can begin to see problems with the mean - field solution : as @xmath134 with fixed @xmath101 , the energy of phase fluctuations ( @xmath92 ) vanishes ; therefore quantum fluctuations begin to mix in many nearly degenerate phase states , @xmath135 .
the mean - field solution for attractive @xmath101 is very different from that of repulsive @xmath101 .
the solution depends on whether @xmath136 or @xmath137 . in the former case
, @xmath132 is locally stable since the energy of fluctuations , described by eq .
( [ d2mean ] ) , remain positive definite . however , the stiffness constant for density fluctuations ( @xmath138 ) is lower than that for phase fluctuations ( @xmath92 ) .
thus , as @xmath139 increases , density fluctuations become dominant .
note , however , that the condition @xmath136 only occurs for a bounded number of particles , and under ordinary circumstances is not expected to be achievable in macroscopic systems . on the other hand , when @xmath137 , we see from eq .
( [ euv ] ) that the optimal states satisfy @xmath140 .
there are two degenerate solutions : @xmath141 and @xmath142 . for @xmath143 ,
these two states approach @xmath144 and @xmath145 , corresponding to all particles being in well 1 or 2 respectively .
we now construct the exact ground state of the hamiltonian , ( [ 2site ] ) , and see how interactions can cause condensate fragmentation . at the same time
, we can see how different types of interaction cause different types of fragmented states .
_ non - interacting case : _ let us first consider the simplest case of non - interacting bosons , with hamiltonian @xmath146 .
the single particle eigenstates are the symmetric state @xmath147 and antisymmetric state @xmath148 with energy @xmath149 and @xmath105 respectively . for a system of @xmath5 bosons ,
the ground state is @xmath150 with energy @xmath151 .
the single particle density matrix of this state is @xmath152 which has a single macroscopic eigenvalue @xmath153 .
the ground state is therefore a single condensate with condensate wavefunction @xmath154 ( the superscript
@xmath26 " stands for transpose ) .
since the ground state @xmath132 is a linear combination of number states @xmath155 , the number of particles in each well fluctuates .
we calculate the number fluctuations of the coherent state @xmath132 by writing it in the number basis . for even @xmath5
, we have @xmath156 where @xmath157 , and @xmath158 the number fluctuations are then @xmath159 which , despite the approximations made in this derivation , coincides with the exact result . _ interacting case : _ the many - body physics of this double well system is completely tractable .
while one can calculate the properties the ground state numerically to arbitrary accuracy , we derive below all the essential features of the ground state by studying the effect of interactions on the non - interacting ground state , i.e. , the coherent state @xmath132 .
we shall see that depending on whether the interactions are repulsive or attractive , the coherent state can be turned into one of two distinct fragmented states : a ` fock - like ' state or a ` schrdinger cat - like ' state .
we first look at the schrdinger equation of this system . writing the ground state in the number basis ,
@xmath160 we can write the schrdinger equation , @xmath161 , where @xmath162 is given by eq .
( [ 2site ] ) , as @xmath163 @xmath164 the many - body problem then reduces to a one - dimensional tight binding model in a harmonic potential .
the special feature of this model is that the tunneling matrix element @xmath165 is highly non - uniform @xcite : @xmath166 for @xmath167 , and @xmath168 for @xmath169 .
this non - uniformity is a consequence of bosonic enhancement , @xmath170 and @xmath171 , which increases the matrix element by a factor of @xmath172 when removing a particle from a system with @xmath5 bosons . as a result
, @xmath173 is maximum when both wells have equal number of bosons ( i.e. , @xmath174 ) , and drops rapidly when the difference in boson numbers between the wells begins to increase , ( i.e. , @xmath175 ) .
a consequence is that hopping favors wavefunctions @xmath176 having large amplitudes near @xmath177 .
for example , in the non - interacting case the wavefunction , eq .
( [ ao ] ) , is a sharply peaked gaussian at @xmath178 .
the interaction term , @xmath179 , leads to a harmonic potential in eq .
( [ sch ] ) .
repulsive interactions @xmath180 suppress number fluctuations , meaning that the gaussian distribution [ eq . ( [ ao ] ) ] of the coherent state will be squeezed into an even narrower distribution . in the limit of zero number fluctuation ,
@xmath181 the system becomes the fock state @xmath182 which is clearly fragmented , since it is made up of two independent condensates .
this fragmentation shows up clearly in the single particle density matrix , @xmath183 which has two macroscopic eigenvalues , corresponding to independent condensation in each well .
the evolution from the coherent state @xmath132 to the fock state @xmath184 can be captured by the family of states @xmath185 as @xmath186 varies from @xmath20 to much greater than unity , the initial coherent state @xmath132 becomes more and more fock - like .
indeed , exact numerical solution of eq .
( [ sch ] ) shows that eq .
( [ fockfamily ] ) is an accurate description of the evolution from a coherent state towards a fock state . during the collapsing process when the wavefunction @xmath176 still extends over many number states ( but fewer than @xmath187 )
, one can take the continuum limit of eq .
( [ sch ] ) , which reduces to the equation for a particle in a harmonic oscillator potential . from its gaussian ground state wavefunction
, we extract @xmath188 . using this continuum approximation
, we calculate the off - diagonal matrix element , @xmath189 which leads to a single particle density matrix @xmath190 the eigenvalues are @xmath191 , and @xmath192 .
the relative number fluctuations are @xmath193 as @xmath186 varies from @xmath20 to a number much larger than unity , the eigenvalues @xmath194 varies from @xmath195 to @xmath196 , and @xmath197 varies from @xmath172 to 0 .
this transition of the coherent state into a fock state , described by the family eq .
( [ fockfamily ] ) , is due to increasing phase fluctuations , as discussed in sec .
the phase fluctuation effect can be seen by writing eq .
( [ fockfamily ] ) in terms of phase states , @xmath198 where @xmath199 are the coefficients given by eq.([ao ] ) .
the family of q.([fockfamily ] ) then becomes @xmath200 where @xmath201 .
as @xmath202 varies from @xmath203 ( coherent state ) to @xmath204 , the gaussian in eq .
( [ phaseaverage ] ) changes from a @xmath205-function , @xmath206 to a uniform distribution , thereby driving the coherent state towards a fock state .
_ attractive interaction : _ when @xmath103 , the potential energy @xmath179 in eq .
( [ u ] ) favors a large number difference between the two wells , in particular , the states @xmath207 and @xmath208 .
it therefore acts in the opposite direction as hopping , which favors a gaussian distribution of number states around @xmath209 .
the effect of interaction is then to split the gaussian peak of the coherence state eq .
( [ ao ] ) into two peaks , a process which can be described by the family of states @xcite , @xmath210 where @xmath211 is the separation between the peaks , @xmath212 is the width of the peaks , and @xmath213 is the normalization constant .
as @xmath214 varies from 0 toward @xmath215 , and @xmath212 shrinks at the same time from @xmath216 to 0 , the state @xmath217 evolves from the coherent state @xmath132 to a schrdinger - cat state @xmath218 the schrdinger - cat state is fragmented in the sense that its single particle density matrix has two large eigenvalues , @xmath219 identical to that of fock state . on the other hand ,
contrary to the fock state , it has huge number fluctuations , @xmath220 details of how @xmath214 and @xmath212 depend on the ratio @xmath221 are found in @xcite , where it is also shown that the family eq .
( [ catfamily ] ) accurately represents the numerical solution of eq .
( [ sch ] ) .
this double - well example brings out the important point that fragmented condensates can not be characterized by the single particle density matrix alone .
higher order correlation functions such as number fluctuations are needed .
this example also shows how a coherent state can be brought into a fock state ( or schrdinger - cat state ) through the phase ( or number ) fluctuations caused by repulsive ( or attractive ) interaction .
the spin-1 bose gas , which is only marginally more complicated , provides an excellent illustration of the role of symmetry in condensate fragmentation . here
the degeneracies that give rise to fragmentation are due to a symmetry : rotational invariance in spin space .
we will see that in the presence of local antiferromagnetic interactions all low energy singly condensed states break this symmetry .
the true ground state , which is a quantum superposition of all members of this degenerate manifold , is a fragmented condensate .
we consider a spin-1 bose gas with @xmath222 particles in the single mode approximation , where each spin state has the same spatial wavefunction .
several recent experiments have been carried out in this limit @xcite ( note that several of these experimental also explored regimes where the single mode approximation was not valid @xcite ) . for conceptual clarity
we consider the extreme case where the trap is so deep that all particles reside in the lowest harmonic state , trivially leading to the single mode limit . in this extreme situation
there are no particle fluctuations : the only relevant interaction is scattering between different spin states .
we denote the creation operator of a boson with spin projection @xmath54 = ( 1,0,-1 ) in the lowest harmonic state by @xmath223 ; the number of bosons with spin @xmath54 is @xmath224 .
the most general form of the scattering hamiltonian which is rotational invariant in spin space is @xcite @xmath225 where @xmath226 , with @xmath227 the spin-1 matrices , ( @xmath228 ) , and @xmath229 the interaction constant .
we consider the case @xmath230 .
more complicated hamiltonians , with more interaction terms , are allowed for atoms with higher spins @xcite .
the variety of fragmented states proliferate rapidly as the atomic spin increases . to illustrate how condensate fragmentation is affected by external perturbations , we include the linear zeeman effect , with hamiltonian , @xmath231 where @xmath87 is the zeeman energy proportional to the external magnetic field @xmath232 .
to explain current experiments one also needs to include the quadratic zeeman effect : the atomic energy levels of atoms are not linear in @xmath232 due to hyperfine interaction between electron spins and nuclear spins .
we ignore these nonlinearities as they are irrelevant for describing fragmentation . for later discussion
, we also include a term of the form @xmath233 which mixes spin states @xmath234 and @xmath235 , where @xmath138 is a constant . terms of this form can be generated by magnetic field gradients @xcite , in which case @xmath138 is proportional to the square of the field gradient .
since both @xmath162 and @xmath236 conserve @xmath237 , the density matrix @xmath238 for the ground state is diagonal in the presence of these terms alone , and the system is generally fragmented unless the density matrix happens to have only one macroscopic eigenvalue .
the effect of @xmath239 is to mix the 1 and -1 states , bringing the system into coherence , similar to the role of the tunneling term in the double well system . like the double well problem in the previous section , the hamiltonian in eq .
( [ spin1h ] ) is exactly soluble .
as previously , we first consider the mean - field solution so that we can relate the fragmented condensate to a linear combination of singly condensed states . the general form of a singly condensed spinor bose condensate is @xmath240 using the fact that @xmath241
, it is easy to see that the number fluctuations are @xmath242 . writing the hamiltonian as , @xmath243 we have , up to terms @xmath244 , @xmath245 where @xmath246 .
it is easy to see that eq.([mean1 ] ) is minimized by @xmath247 where , as before , @xmath26 stands for transpose , @xmath92 is the relative phase between the 1 and @xmath235 component , @xmath248 , and the difference @xmath249 is given by the nearest integer to @xmath250 .
note that @xmath251 , and the entire family @xmath252 is degenerate , reflecting the invariance of eq .
( [ mean1 ] ) under spin rotation about the @xmath253-axis . in the absence of a magnetic field , @xmath254
, we have @xmath255 .
however , since eq .
( [ mean1 ] ) becomes fully rotationally invariant at zero field , any arbitrary rotation of the state @xmath256 is also an optimal spinor condensate .
the entire degenerate family ( referred to as the polar " family in literature ) is given by @xmath257 with energy @xmath258 before discussing the exact ground states , we introduce the cartesian " operators @xmath259 the important properties of these operators is that that under a spin rotation @xmath260 , where @xmath261 , @xmath262 rotates like a cartesian vector , i.e. @xmath263 , where @xmath264 is a rotational matrix in cartesian space ( @xmath265 ) .
it is easily verified that @xmath266 = \delta_{ij}$ ] , and @xmath267 @xmath268 , \,\,\,\,\,\,\,\,\,\,\ ,
\theta = { \bf a}^{2 } .\ ] ] and the operator that creates a singlet pair is , @xmath269 the mean - field state eq .
( [ mf ] ) can now be written as @xmath270 which has average spin @xmath271 the optimal mean - field state is determined by minimizing @xmath272 - i
p n \hat{\bf z } \cdot \vec{\alpha}^{\ast}\times \vec{\alpha}.\ ] ] in zero field ( @xmath273 ) , the optimal mean - field state is @xmath274 , where @xmath275 is a _ real _ unit vector .
the polar family mentioned above can now be conveniently represented by all possible directions of @xmath275 , and the polar state in eq .
( [ polarfamily ] ) corresponds to @xmath276 in the absence of a magnetic field , @xmath277 . since @xmath162 is proportional to the angular momentum operator , the ground state for @xmath278 ( with an even number of bosons @xmath5 ) is a singlet .
the singlet state of a single mode spin-1 bose gas is unique @xcite , and hence any manifestly rotationally invariant state must be the ground state .
one can therefore obtain the ground state by forming singlet pairs , which for an even number of particles yields @xcite @xmath279 since this state is rotationally invariant , its single particle density matrix @xmath238 is proportional to the identity matrix ( shur s theorem ) . with the constraint @xmath280
, we then have @xmath281 the condensate is therefore fragmented into three large pieces .
the energy of the ground state is exactly zero , whereas that of the mean - field state is @xmath282 ( as shown in eq .
( [ polarenergy ] ) ) . to relate the exact ground to the optimal mean - field state ( the polar family @xmath283 )
, we note that @xmath284 i.e. , the exact ground state is an average over the family of optimal mean field states .
this is the analog of the symmetry averaging relation in eq .
( [ f = c ] ) .
this averaging process represents the quantum fluctuations within the family of degenerate mean - field states ; the effect of these fluctuations is to reduce the mean - field energy from @xmath282 to zero .
next we consider the case of non - zero magnetic field @xmath285 .
the hamiltonian is @xmath286 .
the ground state is @xmath287 or simply denoted as @xmath288 , where @xmath289 is an integer closest to @xmath250 .
the state @xmath288 can be easily obtained by changing a singlet pair into a triplet pair , and we have @xmath290 where @xmath291 is a normalization constant , and bose statistics require @xmath289 to be even if @xmath5 is even .
since the hamiltonian @xmath292 conserves @xmath237 , the single particle density matrix remains diagonal , @xmath293 .
as shown in @xcite , the eigenvalues @xmath294 have a very interesting behavior as a function of @xmath289 , namely @xmath295 @xmath296 if @xmath289 is of order 1 , all three eigenvalues , @xmath297 are macroscopic . on the other hand , if @xmath289 becomes macroscopic , ( i.e. , @xmath298 is less than but of order 1 ) , both @xmath299 and @xmath300 remain macroscopic ( with @xmath301 and @xmath302 ) while @xmath303 becomes of order unity .
this means as @xmath289 increases , the @xmath304 component is completely depleted .
although the system is still fragmented , the number of fragmented pieces is reduced from three to two , even for a tiny spin polarization .
further analysis @xcite shows that as one increases @xmath289 , the fluctuations , @xmath305 , drop rapidly .
the reduction in the number of fragmented pieces mirrors the previously discussed reduction in the size of the space of degenerate mean - field states . in the absence of a magnetic field , polar states with arbitrary @xmath306
are degenerate , while a magnetic field in the @xmath307 direction favors those with @xmath306 pointing in the x - y plane . to connect the exact polarized states to the mean - field states
, we use the relation @xmath308 , eq .
( [ theta ] ) , to write @xmath309 where the exact form of @xmath310 is not important for our discussion .
if we now act on @xmath311 by @xmath312 , the coefficients of the _ large @xmath87 _ states in eq .
( [ sum ] ) will be bosonically enhanced via @xmath313 .
further acting on the same state by @xmath314 eventually picks out the term with largest @xmath87 , which is @xmath315 , in the sum eq .
( [ sum ] ) .
this process is equivalent to replacing @xmath316 by @xmath317 .
we then have @xmath318 from our previous analysis of the two - well system it is clear that this state is formed from an angular average of eq .
( [ meanzeta ] ) . the field gradient term ( [ hg ] )
plays the role of tunneling in the two - well system and can drive the system from a fragmented to coherent state . to complete the connection between these examples , we note that when the occupation of the @xmath319 state is negligible , we can set @xmath320 , and only two spin states are required to describe the system . under these conditions ,
the hamiltonian for the spin 1 gas [ eqs .
( [ spin1h ] ) , ( [ magt]),([hg ] ) ] reduces to the two - well hamiltonian ( [ 2site ] ) with @xmath321 , and @xmath322 , and an asymmetry between the wells given by @xmath87 .
this analogy breaks down when @xmath323 and the @xmath319 state becomes occupied . in the previous cases of pseudospin-1/2 and spin-1 bose gases ,
the number of degenerate single particle states is of order unity , far fewer than the number of bosons .
bose gases with large amounts of angular momentum have just the opposite behavior ; the number of nearly degenerate states can be much larger than the number of particles . as we shall see , for relatively small angular momentum , fragmentation similar to that in pseudospin-1/2 and spin-1 bose gases
can occur .
however , for very large angular momentum , it is energetically more favorable for a repulsive bose gas to organize itself into a quantum hall state , which is an extreme form of fragmentation in which all traces of conventional condensation disappear . before discussing the fragmentations of rotating bose gas ,
we first discuss how the single particle energy levels of a rotating bose gas turn into landau levels and achieve high degeneracy as the angular momentum of the system increases . for simplicity ,
we limit the discussion to two dimensional systems .
the single particle hamiltonian for a particle in the rotating harmonic trap is @xmath324 where @xmath325 is the trap frequency , and @xmath16 is the rotational frequency of the trap .
equation ( [ 2drot ] ) can be written as @xmath326 if one rewrites @xmath327 as @xmath328 , with @xmath329 , one sees that eq .
( [ new2drot ] ) is identical to the hamiltonian of an electron in a magnetic field @xmath330 in a harmonic potential with reduced frequency @xmath331 . to diagonalize the single - particle hamiltonian , eq .
( [ 2drot ] ) , we begin by noting that the 2d simple harmonic oscillator @xmath332 is diagonalized as @xmath333 , where @xmath334 , @xmath335 , and @xmath336 . defining @xmath337 , we have @xmath338 and @xmath339 the eigenstates are therefore @xmath340 with eigenvalues @xmath341 where @xmath342 . the energy levels , shown in fig .
[ landau ] , are organized into landau levels , labeled by @xmath343 , separated by @xmath344 .
states within a given landau level are labeled by @xmath345 with spacing @xmath346 . at criticality ,
@xmath347 , each landau level becomes infinitely degenerate . .
for clarity only seven values of @xmath345 are shown , so that the infinite degeneracies at @xmath348 appear to be only seven - fold . the inset shows the energy states at @xmath349 as a function of their angular momentum @xmath15 .
( colors : each landau level is colored with a distinct hue . )
[ landau ] ] the many - body hamiltonian , including a contact interaction @xmath350 , is @xmath351 as @xmath352 , mixing is strongest among the states in the same landau level . to simplify matters
, we consider only extremely weak interactions , @xmath353 .
this limit is naturally reached when @xmath16 is close to @xmath354 , where the centrifugal potential largely cancels the trapping potential and the cloud becomes large and dilute . for such weak interactions , only the lowest landau level is populated .
the eigenfunctions @xmath355 in the lowest landau level , @xmath356 are angular momentum eigenstates with @xmath357 , and a spatial peak at @xmath358 . the many - body wavefunction for a systems of @xmath5 bosons is then @xmath359 } c_{[m ] } u_{m_{1}}({\bf r}_{1 } ) ...
u_{m_{n}}({\bf r}_{n}),\ ] ] where @xmath360 denotes the set of non - negative integers @xmath361 . since , apart from the gaussian factor , the single particle wavefunctions are of the form @xmath362 , the many - body wave function is @xmath363 where @xmath364 is an analytic function which is symmetric in @xmath365 . in particular
, a bose - condensed state corresponds to @xmath366 if @xmath367 is also an eigenstate of the total angular momentum @xmath368 , then @xmath364 in eq .
( [ many ] ) must be a homogeneous symmetric polynominal of @xmath369 of degree @xmath15 .
moreover , the single particle density matrix in the angular momentum basis must be diagonal , i.e. , @xmath370 thus a bose condensed state ( eq .
( [ w ] ) ) that is also an angular momentum eigenstate must be fully condensed into the single particles state @xmath371 , i.e. , @xmath372 .
the total angular momentum must then be @xmath373 .
such a coherent state , however , offers no flexibility to redistribute particles to lower the energy , and is therefore unlikely to be the ground state at fixed angular momentum or fixed rotational frequency @xmath16 , except for certain special cases in which parameters of the system are carefully tuned .
a general bose condensed state , eq .
( [ w ] ) , would be one where @xmath374 contains more than a single angular momentum state , @xmath371 .
such states break rotational symmetry and are not angular momentum eigenstates .
furthermore , due to rotational symmetry , any rotation of @xmath375 ( @xmath376 ) will have the same energy and angular momentum , @xmath377 .
one can therefore form an angular momentum eigenstate with angular momentum @xmath15 by forming the average @xmath378 such averaging always lowers the energy of the system , and generically results in a fragmented state .
specific examples follow .
note that with small modification these arguments also apply in the general case when interactions are not weak .
for the rotating bose gas with attractive interaction , wilkin , gunn , and smith @xcite pointed out the ground state with non - zero angular momentum is one with all angular momentum carried by the center of mass , and has the form @xmath379 where @xmath380 is the center of mass and @xmath381 is a normalization constant , @xmath382 it is straightforward to show that the single particle density matrix of this state is @xmath383 the ground state is fragmented because eq .
( [ fragatt ] ) has a distribution of large eigenvalues .
later , pethick and pitaevskii @xcite pointed out that this state is of the form @xmath384 where @xmath385 is the wavefunction of the center of mass @xmath386 , and @xmath387 is a product of single particle states , where the particle coordinates are @xmath388 . given this structure , it is natural to refer to this state as being singly condensed in the center of mass frame .
here , we show that this exact ground state can also be written as a symmetry average of broken symmetry states of the form eq .
( [ angleave ] ) , as in many of the previous examples .
attractive interactions favor particles clumping together . in homogeneous systems ,
such clumping leads to collapse . in the fast rotating limit
, however , the analyticity of the wavefunction in the lowest landau level and the gaussian factor impose strong constraints on the degree of localization possible .
the most localized state is @xmath389 , a gaussian of a width given by the trap length @xmath390 .
a similar localized wave - packet at location @xmath391 is @xmath392 which gives @xmath393 a many - body coherent state formed from these single particle states is @xmath394 this state carries angular momentum @xmath395
@xmath396 @xmath397 , with @xmath398 ^ 2 - 1\right ) = n |a / d|^2.\ ] ] the single particle density matrix is @xmath399 in which @xmath400 from this degenerate set of coherent states we construct an eigenstate of angular momentum by taking the superposition @xmath401 where @xmath402 .
similar fragmentation is found in the repulsive case , which for @xmath403 is described by a vortex lattice .
a general mean field state is of the form eq .
( [ w ] ) , with @xmath404 , @xmath405 where the zeroes @xmath214 are the locations of the vortices @xcite .
the angular momentum carried by this state can be obtained by noting that within the lowest landau level , @xmath406 which increases as the density of vortices increases . following the arguments which we used for
the attractive case , we are once again lead to a fragmented condensate ( see @xcite for numerical studies ) .
the rotating bose gas with repulsive interactions is , however , much richer than that with attractive interactions , since the mean - field picture breaks down in a fundamental way at large values of angular momentum .
one sees this breakdown by first noting that with increase of the angular momentum , the vortex density increases and the particle density decreases @xcite .
eventually , the density of vortices is comparable to the density of particles and one can significantly improve the energy of the system by correlating the positions of the vortices with the positions of the particles .
these correlations can not be captured by any simple manipulation of the mean - field states ; the general description of the system with large angular momentum is quite complicated @xcite . despite this complexity
, there is a limit in which we can find the exact ground state .
imagine that the system is rotating with @xmath16 sufficiently close to @xmath354 that the energy spacing within the lowest landau level may be treated as a perturbation .
one would find the ground state by first minimizing the interaction energy in the lowest landau level , and then perturbatively including the level spacing . for short ranged interactions
the energy is minimized by any state for which the wavefunction vanishes whenever two particles come together .
degenerate perturbation theory then says that the lowest energy state of all these wavefunctions is the one with lowest angular momentum . in the lowest landau level , the lowest angular momentum bosonic wavefunction which vanishes when two particles touch is the @xmath407 laughlin state , @xmath408 ) = \prod_{i > j}(z_{i}-z_{j})^2 e^{-|z_{i}|^2/2d^2}.\ ] ] for an infinite system , one readily sees that @xmath409 for all @xmath345 .
not only are there no eigenvalues of order @xmath5 , but they are all less than unity .
all traces of the conventional bose condensation are obliterated .
the examples shown in sec .
iii share a number of common features .
many of the fragmentation processes we have discussed , such as those in eqs .
( [ phaseaverage ] ) , ( [ spin1average ] ) , and ( [ rotatingaverage ] ) , can be described by a family of quantum states that are weighted averages of broken symmetry states over the space of broken symmetry , typically of the form @xmath410 where @xmath411 is a coherent state with broken symmetry parameter @xmath412 ( e.g. , a spin direction ) , @xmath413 is a distribution function in the space of broken symmetry , and @xmath414 is the parameter that controls the fragmentation of the system .
if , as @xmath414 changes , say , from @xmath415 to 1 , @xmath413 changes from a distribution sharply peaked at @xmath416 to a completely uniform distribution , eq .
( [ evolution ] ) will evolve from the coherent state @xmath416 to a fragmented state .
such changes in the distribution function reflect a growing fluctuation about the initial coherent state @xmath416 singled out by a tiny symmetry - breaking field in the coherent regime .
for example , in the case of repulsive bose gas in a double well , @xmath412 is the relative phase between the condensates in the two wells and @xmath417 . in the spin-1 bose gas , @xmath412 is the vector @xmath275 and @xmath416 is a direction normal to an infinitesimal external magnetic field . in the rotating attractive bose gas , @xmath412 is the location in space of the coherent state , and @xmath416 is the equilibrium location determined by , e.g. , a weak potential that breaks rotational symmetry .
exactly how the fluctuations about @xmath416 grow depends on the specific dynamics of the system .
large fluctuations about @xmath416 directly reflect the competition different degenerate states for bose condensation , the ultimate cause of fragmented structure , as discussed in the introduction .
another important feature of states of the type ( [ evolution ] ) is that , as a consequence of bosonic enhancement , the range of the control parameter @xmath414 over which the system switches from a coherent to a fragmented state shrinks with particle number @xmath5 . for a repulsive bose gas in a double well , as discussed in sec .
iii.a.2 , the transition from a fragmented to a coherent takes place around @xmath418 @xcite .
a similar situation occurs in the spin-1 bose gas , where the singlet ( fragmented ) state gives way to a coherent state for field gradient @xmath419 @xcite . since the window for fragmentation vanishes as @xmath5 increases , fragmented condensates will not be realized in condensed matter systems , with @xmath420 particles . on the other hand , in mesocopic systems like quantum gases with typically @xmath421 particles
, fragmented states can exist in parameter ranges accessible to experiment .
although our discussions have mainly focussed on the class of states eq .
( [ evolution ] ) , other fragmentation processes do not lead to states that are most naturally expressed in this manner , e.g. , the schrdinger - cat family in eq .
( [ catfamily ] ) .
one can have distinct fragmented states , such as the fock state and the schrdinger - cat state , with entirely different properties but identical single particle density matrices .
such different states are therefore indistinguishable within the penrose - onsager scheme . to tell them apart , it is necessary to examine second order correlation functions such as number fluctuations .
even higher order correlation functions are needed to distinguish the fragmented states of more complicated systems , such as those of @xmath422 bose gases of @xmath8rb , and the recently realized spin-3 bose gas of @xmath423cr @xcite . in general
, one can expect a large variety of fragmented states , differing from each other by high order correlation functions .
the relative number fluctuations of a fragmented state ( such as those in the pseudo - spin 1/2 and spin-1 bose gases discussed in sec .
iii ) are a measure of the stability of the state .
huge fluctuations , such as in the schrdinger - cat state and the singlet state of the spin-1 bose gas indicate that the system is easily damaged by external perturbations .
consider , for example , a perturbation of the form @xmath424 , where @xmath425 is a boson in one of the wells in the double well example , or a boson in the @xmath426 spin state of a spin-1 bose gas , @xmath427 adds a particle in a different atomic state of the same boson ( e.g. , a plane wave state in the background gas ) , and @xmath428 is a very small parameter .
the perturbation @xmath429 acting on the fock state @xmath430 simply changes it to another fock state , @xmath431 , which also has zero number fluctuations , @xmath432 .
in contrast , @xmath429 acting on the schrdinger - cat state @xmath433 collapses it into the state @xmath434 , immediately reducing the enormous number fluctuations of the schrdinger - cat state to zero . more generally , if the cat state @xmath435 is a sum of two gaussians in number space , as shown in eq .
( [ catfamily ] ) , with one gaussian ( @xmath436 ) peaked at @xmath437 and the other ( @xmath438 ) peaked at @xmath439 , where @xmath440 , and @xmath441 , the action of @xmath429 on @xmath442 considerably enhances @xmath436 and suppresses @xmath438 .
the large effect of the small perturbation @xmath429 is due to bosonic enhancement , which gives @xmath443 and @xmath444 .
since @xmath441 , the norm of @xmath445 is considerably smaller than that of @xmath446 , which is no longer a schrdinger - cat state . in a spin-1 bose gas , a similar action changes the singlet state @xmath447 to one with much smaller relative number fluctuations . since schrdinger
cat - like states can collapse into fock states with the slightest perturbation , and fock states can easily be reassembled into a single condensate by any small amount of tunneling between different fragmented pieces , why should one bother with fragmented states ?
is fragmentation relevant ?
the point , as mentioned before , is that even though fragmented states can not be realized in macroscopic systems , the situation is different for mesoscopic systems like trapped quantum gases .
the huge reduction in particle number considerably relaxes the constraint of formation of fock states and schrdinger cat - like states , and fragmented ground states become realizable .
the phenomena of fragmentation becomes even richer if it takes place in both real and spin space , such as with high spin bosons in an optical lattices close to the mott limit .
the combined effect of spin degeneracy and spatial degeneracy ( due to different isolated wells ) produces a great variety of quantum phases as the spin of boson increases .
in addition , the singlet ground state of a spin-1 bose gas can also be viewed as a resource " for singlets and may therefore be useful in developing quantum teleportation protocols in optical lattices with spin-1 bosons .
our understanding of the properties of fragmented states , and in particular their dynamics , is at such an early stage that it leaves open considerable room for inventive ideas , which is where the excitement lies .
we stress that we have described only the simplest types of fragmented states .
many other kinds which do not fall into the category we considered . the @xmath407 laughlin state discussed in sec .
iii.c.3 is an example at the other extreme .
there are no large eigenvalues singled out in the density matrix . instead
, all the eigenvalues are identical and of order unity .
in addition to level degeneracy , temperature effects can cause condensates to fragment , a subject to which we turn in the next section . up to this point
, the examples of fragmentation given in this paper have focussed on the case where interaction - driven quantum fluctuations break up the condensate .
it is intuitively clear that thermal fluctuations can play a similar role .
a trivial example is given by a non - interacting josephson junction , governed by the hamiltonian @xmath448 this hamiltonian is diagonal in the basis of symmetric and asymmetric states , for which the creation operators are @xmath449 , and @xmath450 , so the single particle density matrix is diagonal in this basis . at zero temperature
only the symmetric state is occupied a single condensate . at very large temperature
each of these states is equally occupied so the system is fragmented . for intermediate temperatures ,
the occupation of each of these states is @xmath451 the upper signs denote the symmetric state ( s ) and the lower the antisymmetric state ( a ) .
the crossover between the singly condensed state at @xmath1 and the fragmented state at @xmath452 is smooth .
extensions of this argument are relevant for spinor condensates in which one can , in principle , have a hierarchy of transition temperatures , where the @xmath453 mode becomes macroscopically occupied below some temperature @xmath454 , but order in the spin - channel does not occur until a lower temperature @xcite .
a vortex lattice provides a qualitatively different example of how finite temperature fragments a condensate .
imagine a bucket of @xmath2he rotating at frequency @xmath16 .
the ground state of the system contains a triangular array of vortices with @xmath455 vortices per unit area . at finite temperature
the vortex lattice is thermally excited , giving rise to a decay in the phase correlations across the sample . if we let @xmath456 be the superfluid order parameter coarse - grained on a scale large compared to the vortex spacing , then according to @xcite , the correlation function @xmath457 decays as @xmath458 , for large separations .
the exponent @xmath459 is proportional to the ratio of the distance between particles to the distance between vortices . here
the particle number density is @xmath460 , the thermal wavelength is @xmath461 and the distance between vortices is @xmath462 .
this algebraic decay of correlations in real space corresponds to an algebraic decay in momentum space . for @xmath463 , the occupation of the @xmath453 mode scale as @xmath464 , and as as @xmath465 the occupations of the @xmath466 modes scale as @xmath467 .
thus the number of macroscopically occupied modes scales as @xmath468 , and for carefully chosen @xmath15 and @xmath428 a fragmented state can result . for a typical helium experiment @xmath469 , and
the depletion caused by this effect is negligible .
experiments on alkali gases , e.g. , @xcite , create vortex lattices with @xmath470 m and a particle density @xmath471 @xmath472 , for which @xmath473 , also a minor correction .
experiments with smaller vortex lattices in paris @xcite have a similar vortex spacing @xmath474 m , and density @xmath471 @xmath472 , yielding a comparable value for @xmath428 .
although vortex lattices in current experiments are not thermally fragmented , there does not appear to be any fundamental impediment to making @xmath428 larger . not surprisingly , given the fascinating nature of fragmented states ,
several recent papers ( all cited in the relevant sections of this paper ) have been investigating models similar to those discussed here . as a guide to the reader
, we briefly summarize these works , presenting them in roughly the same order as they appear in the main text .
the two - state system was explored by nozires @xcite , who pointed out the existence of fragmentation in that system .
the generic stability of a two - state fragmented condensate was considered by rokhsar @xcite .
one realization of this model would be to place atoms in bonafide double - well potentials .
due to mode mixing such a system is not identical to a simple two - mode model , and several authors have explored more realistic models . spekkens and
sipe @xcite used a variational approach to compare fragmented and singly condensed states in such potentials .
more recently , streltsov and cederbaum @xcite , along with moiseyev @xcite produced similar results through a multi - mode mean - field theory .
the evolution of the ground state from a fock state to a coherent state and from a coherent state to a schrdinger - cat state was studied by ho and ciobanu @xcite .
experiments have recently produced a condensate confined in a double - well trap @xcite .
the fragmented nature of the ground state of the spin-1 bose gas was first noted by nozires and saint james @xcite .
later studies by law , pu , and bigelow @xcite investigated the ground state properties and the spin dynamics in terms of the basis used in eq .
( [ sum ] ) . ho and
yip @xcite showed that the fragmented singlet state has huge fluctuations ; they also showed the relation between this fragmentation and spontaneous symmetry breaking .
they explicitly showed how magnetic fields and field gradients drive the system into a singly condensed state .
similar considerations were addressed by koashi and ueda @xcite .
javanainen discussed issues involving the measurement of the fragmented spin-1 ground state @xcite .
the rotating attractive gas was first studied by wilkin et al .
@xcite , who noted that the ground state was fragmented .
the connection between this fragmentation and symmetry breaking was first discussed by pethick and pitaevskii @xcite .
there have been extensive studies of the properties of the rotating repulsive gas in the lowest landau level many of which have focussed on the structure of the order in the single particle density matrix .
these works have mainly used a combination of exact diagonalization and variational techniques .
a particularly relevant paper is the exact diagonalization study of liu et al .
@xcite , which focusses on the symmetry - breaking nature of the vortex states . a related paper by jackson et al .
@xcite compares mean - field and exact wavefunctions at various values of the angular momentum .
the goldstone mode associated with vortex nucleation was discussed by ueda and nakajima @xcite .
fragmentation occurs in clumped " bosonic systems with attractive interactions .
ueda and leggett used a two - mode approximation @xcite to study fragmentation and soliton formation in a one dimensional attractive bose gas .
a quite thorough comparison of mean - field theory and exact diagonalzation is found in the articles by kanamoto et al .
@xcite , as well as the closely related work of kavoulakis @xcite .
montina and arecchi @xcite use a monte - carlo scheme to investigate the degree of fragmentation of this system . in three dimensions , elgaroy and pethick
@xcite showed that a harmonically trapped gas of atoms with attractive interactions does not form a stable fragmented state .
other systems with symmetry breaking and fragmentation include phase separated two - component gases @xcite , and rotating gases during a phase - slip event @xcite .
boson ground states where the condensate is broken into a macroscopic number of pieces include the mott insulator @xcite , fractional quantum hall states @xcite , and the low dimensional bose gas @xcite .
aspects of the dynamics of regaining phase coherence among condensates are discussed by yi and duan @xcite , a paper closely related to discussion of how the measurement process is influenced by fragmentation @xcite .
finally we mention the lecture notes produced by castin and herzog for the 2000 cargse summer school @xcite , which lucidly introduces fragmentation and analyzes the case of spin-1 bosons and of the one - dimensional attractive gas .
they included an extended discussion of the role of symmetry breaking .
bose condensation is remarkably robust . it is therefore exciting to search for zero - temperature bosonic states which are not condensed and to understand the process that breaks up the condensate .
fragmentation , where the condensate breaks up into a few pieces , is the first step in this journey , which eventually ends at strongly - correlated states possessing no trace of condensation .
we have described some of the canonical models which have fragmented ground states , and extracted their properties .
we see that there is a rich variety of fragmented ground states .
often fragmentation is associated with restoring a broken symmetry .
sometimes it is accompanied by order in higher - order correlation functions .
the one unifying feature appears to be a combination of near - degeneracies and interactions .
the higher the degree of degeneracy , the more fragmented the condensate may become .
parts of this research were performed at the aspen center for physics and the kavli institute for theoretical physics in santa - barbara .
the authors acknowledge the following support : ejm nsf grant phy-0456261 and the sloan foundation ; tlh nsf grant dmr-0426149 and nasa grant - nag8 - 1765 ; mu grants - in - aids for scientfic research ( grant no . 17071005 ) and a crest program of the jst ; and gb nsf grant phy03 - 55014 and phy05 - 00914 .
y. castin and c. herzog , `` bose - einstein condensates in symmetry breaking states '' , comptes rendus de lacademie des sciences de paris , tome 2 , serie iv , 419 ( 2001 ) . also available as cond - mat/0012040 .
myatt , e.a .
burt , r.w .
ghrist , e.a .
cornell and c.e .
wieman , phys .
lett . * 78 * , 586 ( 1997 ) ; m.r .
matthews , d.s . hall , d.s .
jin , j.r .
ensher , c. wieman , e.a .
cornell , f. dalfovo , c. minniti , and s. stringari , phys .
lett . * 81 * , 243 ( 1998 ) ; d. s. hall , m.r .
matthews , j.r .
ensher , c.e .
wieman , and e.a .
cornell , phys .
81 , 1539 ( 1998 ) ; d.s .
hall , m.r .
matthews , c.e .
wieman , and e.a .
cornell , phys .
lett . * 81 * , 1543 ( 1998 ) ; m.r .
matthews , b.p .
anderson , p.c .
haljan , d.s .
hall , m.j .
holland , j.e williams , c.e .
wieman , e.a .
cornell , phys .
lett . * 83 * , 3358 ( 1999 ) .
a. grlitz , t.l .
gustavson , a.e .
leanhardt , r. lw , a.p .
chikkatur , s. gupta , s. inouye , d.e .
pritchard , and w. ketterle , phys .
lett . * 90 * , 090401 ( 2003 ) ; d.m .
stamper - kurn , h .- j .
miesner , a.p .
chikkatur , s. inouye , j. stenger , and w. ketterle : phys .
lett . * 83 * , 661 ( 1999 ) .
miesner , d.m .
stamper - kurn , j. stenger , s. inouye , a.p .
chikkatur , and w. ketterle : phys .
lett . * 82 * , 2228 ( 1999 ) .
j. stenger , s. inouye , d.m .
stamper - kurn , h .- j .
miesner , a.p .
chikkatur , and w. ketterle : nature * 396 * , 345 ( 1998 ) .
stamper - kurn , m.r .
andrews , a.p .
chikkatur , s. inouye , h .- j .
miesner , j. stenger , and w. ketterle : phys .
lett . * 80 * , 2027 ( 1998 ) ; j. m. higbie , l. e. sadler , s. inouye , a. p. chikkatur , s.r .
leslie , k.l .
moore , v. savalli , and d.m .
stamper - kurn .
lett . * 95 * , 050401 ( 2005 ) ; m .- s . chang , c.d .
hamley , m.d .
barrett , j.a .
sauer , k.m .
fortier , w. zhang , l. you , and m.s .
chapman , phys .
* 92 * , 140403 ( 2004 ) ; h. schmaljohann , m. erhard , j. kronjger , m. kottke , s. van staa , l. cacciapuoti , j.j .
arlt , k. bongs , and k. sengstock , phys .
lett . * 92 * , 040402 ( 2004 ) ; and t. kuwamoto , k. araki , t. eno , and t. hirano , phys .
a * 69 * , 063604 ( 2004 ) .
cooper and n.k .
wilkin , phys .
b * 60 * , r16279 ( 1999 ) ; s. viefers , t.h .
hansson , and s.m .
reimann , phys .
a * 62 * , 053604 ( 2000 ) ; n.r .
cooper , n.k .
wilkin , and j.m.f .
gunn , rev .
lett . * 87 * , 120405 ( 2001 ) ; b. paredes , p. fedichev , j.i .
cirac , and p. zoller , rev . lett . * 87 * , 010402 ( 2001 ) ; j.w .
reijnders , f.j.m .
van lankvelt , k. schoutens , and n. read , phys .
* 89 * , 120401 ( 2002 ) ; b. paredes , p. zoller , and j.i .
cirac , phys .
a * 66 * , 033609 ( 2002 ) ; n. regnault , and th .
jolicoeur , phys .
lett . * 91 * , 030402 ( 2003 ) ; and t. nakajima and m. ueda , phys .
lett . * 91 * , 140401 ( 2003 ) .
t. papenbrock , phys . rev . a * 67 * , 041601(r ) ( 2003 ) ; d.m .
gangardt , and g.v .
shlyapnikov , phys .
lett . * 90 * , 010401 ( 2003 ) ; m.d .
girardeau , e.m .
wright , and j.m .
triscari , phys .
a * 63 * , 033601 ( 2001 ) ; d.s .
petrov , g.v .
shlyapnikov , and j.t.m .
walraven , phys .
lett . * 85 * , 3745 ( 2000 ) ; and a. lenard , j. math . phys .
* 7 * , 1268 ( 1966 ) .
m. r. andrews , c. g. townsend , h .- j .
miesner , d. s. durfee , d. m. kurn , and w. ketterle , science * 275 * , 637 ( 1997 ) ; m. r. andrews , d. m. kurn , h .- j .
miesner , d. s. durfee , c. g. townsend , s. inouye , and w. ketterle , phys .
lett . * 79 * , 553 ( 1997 ) ; _ ibid . _
* 80 * , 2967 ( 1998 ) . the real puzzle is why each single shot measurement yields an interference pattern of two phase - coherent condensates with an apparent relative phase @xmath92 .
the explanation , which is not contained in eq.([f = c ] ) , lies in quantum measurement theory ; the very act of measurement actually establishes a phase between the two initially incoherent condensates , a phenomenon that is a consequence of bosonic enhancement @xcite . by a suitable change of variables
one can make the hopping uniform , but at the cost of making the potential more complicated .
see g. scharf , w. f. wreszinski , and j. l. van hemmen , j. phys .
20 * , 4309 ( 1987 ) .
v. schweikhard , i. coddington , p. engels , v. p. mogendorff , and e. a. cornell , phys .
* 92 * , 040404 ( 2004 ) ; i. coddington , p. c. haljan , p. engels , v. schweikhard , s. tung , and e. a. cornell , phys .
a * 70 * , 063607 ( 2004 ) ; and references therein .
y. kagan , v.a .
kashurnikov , a.v .
krasavin , n.v .
prokofev , and b.v .
svistunov , phys .
a * 61 * , 043608 ( 2000 ) ; j.m .
kosterlitz and d.j .
thouless , j. phys .
c * 6 * 1181 ( 1973 ) ; d.s .
petrov , m. holzmann , and g.v .
shlyapnikov , phys .
* 84 * , 2551 ( 2000 ) ; d.s . fisher and p. c. hohenberg , phys .
b , * 37 * , 4936 ( 1988 ) ; v. n. popov , _ functional integrals in quantum field theory and statistical physics _ , ( reidel , dordrecht , 1983 ) . | we present the theory of bosonic systems with multiple condensates , unifying disparate models which are found in the literature , and discuss how degeneracies , interactions , and symmetries conspire to give rise to this unusual behavior .
we show that as degeneracies multiply , so do the types of fragmentation , eventually leading to strongly correlated states with no trace of condensation . |
agent - based models have attracted significant interest across a broad range of disciplines @xcite .
an increasingly popular application of these models has been the study of financial markets @xcite .
the motivations for this focus on financial markets are two - fold : the promise that data - rich financial markets are good candidates for the empirical study of complex systems , and the inadequacy of standard economic models based on the notions of equilibria and rational expectations .
currently many different agent - based models exist in the econophysics literature , each with its own set of implicit assumptions and interesting properties @xcite . in general
these models manage to exhibit some of the statistical properties that are reminiscent of those observed in real - world financial markets ; for example , fat - tailed distributions of returns and long - timescale volatility correlations . despite their differences , these models draw on several of the same key ideas : feedback , frustration , adaptability and evolution .
the underlying goal of all this research effort is to generate a microscopic agent - based model which ( i ) reproduces all the stylized facts of a financial market , _ and _ ( ii ) makes sense on the microscopic level in terms of financial market microstructure . while each of these goals is separately achievable , the combination of ( i ) _ and _ ( ii ) within a single model represents a fascinating challenge , and underpins our motivation for writing the present paper . challet and zhang s minority game @xcite offers possibly the simplest paradigm for a system containing the key features of feedback , frustration and adaptability .
the mg has remarkably rich dynamics , given the simplicity of its underlying binary structure , and has potential applications in a wide range of fields .
consequently challet and zhang s mg model , together with the original bar model of arthur @xcite , represent a major milestone in the study of complex systems .
given the mg s richness and yet underlying simplicity , the mg has also received much attention as a financial market model @xcite .
the mg comprises an odd number of agents _ n _ choosing repeatedly between the options of buying ( 1 ) and selling ( 0 ) a quantity of a risky asset .
the agents continually strive to make the minority decision i.e. buy assets when everyone else is selling and sell when everyone else is buying .
this strategy seems to make sense on first inspection . for example
, a majority of agents selling will force the price of the asset down and thus ensure a low buy price ( which is favorable ) for the minority of the agents who decided to buy ; hence the minority group has ` won ' .
however , now consider the scenario where this pattern repeats over and over .
the minority of agents , who have been buying more and more assets as the price falls , now find themselves holding a huge inventory of assets which are worth very little .
when these agents try to cash in their assets , they will find themselves very much poorer .
_ surely then the minority group has actually lost ? _
this paradox has sparked much discussion as to the suitability of the mg as the basis of a microscopic model of financial markets .
consequently a diverse range of modified ( yet mg - related ) models have emerged , all seeking to address this point in different ways @xcite .
some models include a combination of different types of agents - in particular , agents who are rewarded for trading in the minority , and agents who are rewarded oppositely for trading in the majority ( see ref .
@xcite for an early example ) . in such models ,
the overall class of the game ( minority or majority ) is ambiguous , and is likely to change as differing numbers of each group flood in and out of the market .
other models try to combine the two different reward structures , thus requiring only one class of agents who strive to be neither exclusively in the minority group nor exclusively in the majority group @xcite .
we refer the reader to the econophysics website @xcite for further examples ( see also refs .
@xcite ) . the purpose of this paper is to consolidate the search for a suitable microscopic reward scheme for use in agent - based financial markets .
the paper comprises a synthesis of several earlier reports , each aimed at addressing the topic of self - consistent agent - based market modelling within a binary game structure .
these earlier reports are re - organized in light of several recent papers on specific reward structures , and supplemented by additional simulation results , observations and thoughts @xcite .
we start by discussing various strategy reward schemes in which the agents seek to maximize their wealth in a direct way .
we then illustrate a disturbing feature whereby such reward structures , while appearing microscopically acceptable , can lead to unrealistic market dynamics ( e.g. instabilities ) .
finally , we discuss various mechanisms which are responsible for re - stabilizing the market in reality .
this discussion leads to a ` toolbox ' of processes from which , we believe , successful market models can be constructed in the future .
consider the situation of a population of @xmath0 agents with some limited global resource .
a specific example could be arthur s famous bar problem @xcite where there are @xmath0 regular bar - goers but only @xmath1 seats . the reward scheme in arthur
s bar problem is simple : bar - goers are successful if they attend _ and _ they manage to obtain a seat .
this reward scheme implicitly assumes that the bar - goers value sitting down above other criteria . while this makes sense for many situations , it is not universal . for example
, it is our experience that customers of college - based bars do not view seating as a necessary requirement for a successful evening !
the motto ` the more the merrier ' often seems more appropriate .
the mg represents a very special case of the bar problem , in which the seating capacity @xmath2 _ and _ all the attendees wish to sit down . in more general and realistic situations ,
the correct reward scheme is likely to be less simple than either arthur s bar model or challet and zhang s mg .
until a specific reward scheme is defined , the bar model remains ill - specified .
putting this another way , the precise reward scheme chosen is a fundamental property of the resulting system and directly determines the resulting dynamical evolution .
hence it is crucial to understand the effect that different microscopic reward schemes have on the macroscopic market dynamics . as first discussed in ref .
@xcite , the bar problem is somewhat analogous to a financial market where the bar - goers are replaced by traders . in the same way that a general bar problem requires a non - trivial reward scheme , any ` correct ' agent - based market model will need a non - trivial reward scheme in order to avoid inconsistencies with financial market microstructure .
this provides us with the motivation to look beyond the bar model and mg , in order to design an agent - based model with financial relevance .
having said this , the binary structure of the mg provides an appealing framework for formulating a market model without introducing too many obvious pathologies .
given these considerations , our discussion from here on will not assume any specific reward scheme , but will employ a binary game structure based on ref .
@xcite .
we start by reviewing the binary game structure .
there are three basic parameters ; the number of agents @xmath0 , the ` memory ' of the agents @xmath3 , the number of strategies held by each agent @xmath4 .
the agents all observe a common binary source of information , but only remember the previous @xmath3 bits .
hence the global information available to each agent at time @xmath5 is given by @xmath6 $ ] where in decimal notation @xmath6 \in\left\ { 0\ldots p-1\right\ } $ ] with @xmath7 .
each strategy @xmath8 contains as its elements @xmath9 , and represents a response @xmath10 to each of the possible @xmath11 values of the global information @xmath12 .
there are hence @xmath13 possible strategies .
the agents are randomly allocated a subset of @xmath4 of these strategies at the outset of the game , and are not allowed to replace these .
the agents also keep a score @xmath14 reflecting their strategies previous successes , whether the strategy is played or not .
the net action of the agents @xmath15 $ ] is defined as : @xmath16 = \sum_{r=1}^{2^{p}}n_{r}\left [ t\right ] a_{r}^{\mu\left [ t\right ] } % \ ] ] where @xmath17 is the number of agents playing strategy @xmath18 .
the agents always play their highest performing strategy , i.e. the strategy with the highest @xmath14 . the strategy score is updated with payoff @xmath19 such that @xmath20 = s_{r}\left [ t\right ] + g_{r}\left [ t+1\right ] $ ] .
the central question is the following : what should we take as ` success ' in the context of a market and hence what are the payoffs ? in short , what is the microscopic reward scheme ?
this is equivalent to asking what the game actually is , since a game requires a definition of the players goals and hence their rewards . in order to address this question ,
we start by considering a binary version of arthur s bar problem with general @xmath21 , i.e. an mg - like model which is generalized to @xmath22 @xcite .
the two responses would be ` attend ' or ` do nt attend ' , and there are two possible scenarios : ( i ) the attendance is less than the seating capacity . in this case , the attendees are successful while those who did nt attend are unsuccessful .
( ii ) the attendance exceeds the seating capacity . in this case , the attendees are unsuccessful while those who did nt attend are successful . setting @xmath2 yields the mg , and the two scenarios collapse into one : the successful agents are those who predict the minority group , i.e. attend when the minority attend , or do nt attend when the majority attend .
[ for a study of the binary game with @xmath22 , see ref . @xcite .
] in the specific case of the mg , success is classified as having correctly predicted the minority outcome .
the payoff function for the mg is thus of the form : @xmath23 = \chi\left [ -a_{r}^{\mu\left [ t\right ] } a\left [ t\right ] \right ] \label{eq strat payoff mg}%\ ] ] where @xmath24 $ ] is an odd , increasing function of @xmath25 usually chosen to be either @xmath24 = x$ ] or @xmath24 = \operatorname*{sgn}\left [ x\right ] $ ] .
the feedback in the mg arises through the global information @xmath6 $ ] where the most recent bit of the binary information is defined by the sign of the net action of the agents @xmath15 $ ] : @xmath26 = 2\mu\left [ t\right ] -p\operatorname{h}\left [ \mu\left [ t\right ] -\frac{p}{2}\right ] + \operatorname{h}\left [ a\left [ t\right ] \right ] \label{eq history update mg}%\ ] ] where @xmath27 $ ] is the heaviside function @xcite . in the more general form of the game where the resource level @xmath22 , the argument in the heaviside function @xmath28 $ ] would depend on the value of @xmath21 @xcite .
more generally it could become a complicated function of all the game parameters , or include feedback from the macroscopic dynamics in the past , or even the effects of exogenous news arrival .
it is now accepted that a generalization of the mg in which agents can sit out of the game ( i.e. @xmath29 } = 0 $ ] ) when they are not confident of the success of any of their @xmath4 strategies , provides a better model of financial trading since it allows the number of agents active in the market to vary .
this more general model is the grand canonical minority game ( gcmg ) which was first studied in ref .
the gcmg typically has two more parameters than the mg : namely @xmath30 which is a time horizon over which strategy scores are forgotten ( i.e. @xmath20 = \left ( 1 - 1/t\right ) s_{r}\left [ t\right ] + g_{r}\left [ t+1\right ] $ ] ) and @xmath31 which is a threshold value of @xmath14 below which the agents will opt not to participate in the game ( @xmath29 } = 0 $ ] ) .
the gcmg seems to reproduce the stylized statistical features of a financial time - series over a wide parameter range ( see the discussion in ref .
it is commonly believed that in a financial market , excess demand ( i.e. the difference between the number of assets offered and the number sought by the agents ) exerts a force on the price of the asset .
furthermore it is believed that a positive excess demand will force the price up and a negative demand will force the price down . a reasonable suggestion for the price formation process
could then be @xcite : @xmath32 \right ] -\ln\left [ p\left [ t-1\right ] \right ] = \frac{d\left [ t^{-}\right ] } { \lambda}% \label{eq price formation mult}%\ ] ] or @xmath33 -p\left [ t-1\right ] = \frac{d\left [ t^{-}\right ] } { \lambda}\label{eq price formation add}%\ ] ] where @xmath34 $ ] represents the excess demand in the market prior to time @xmath5 , when the new price @xmath35 $ ] is set and the buy / sell orders are executed .
the scale parameter @xmath36 represents the ` market depth ' ( or liquidity ) i.e. how sensitive a market is to an order imbalance . in general
we would expect @xmath36 to be some increasing function of the number of traders @xmath0 trading in that asset .
also , the functional forms may not be linear as in equations [ eq price formation mult ] and [ eq price formation add ] .
however , studies carried out on several markets have found the forms of equations [ eq price formation mult ] and [ eq price formation add ] to be reasonable @xcite . in ref .
@xcite , we suggested that in markets involving a market - maker who takes up the imbalance of orders between buyers and sellers , the market - maker will wish to manipulate the price of the asset in order to minimize her inventory of stock @xmath37 $ ] . whilst the market - maker s job is to execute as many of the orders as is possible
, she does not wish to accumulate a high position in either direction . to this end ,
the market - maker lowers the price to attract buyers when she holds a long position @xmath38 and raises it to attract profit - taking sellers when she has a short position @xmath39 .
this modifies the price formation of equation [ eq price formation mult ] to : @xmath40 \right ] -\ln\left [ p\left [ t-1\right ] \right ] = \frac{d\left [ t^{-}\right ] } { \lambda}-\frac{\phi_{m}\left [ t^{-}\right ] } { \lambda_{m}}\label{eq price formation mult mm}%\ ] ] where @xmath41 is the market - maker s sensitivity to her inventory .
our puzzle is to form a sensible model of a collection of traders buying and selling a financial asset .
it is clear from sections [ sec mg basics ] and [ sec price process ] that we have two pieces of the puzzle .
the mg gave us a simple paradigm for a group of agents interacting in response to global information and adapting their behavior based on past experience .
the output of this model was a ` net action ' @xmath15 $ ] .
in addition , we now have some simple systems for price - formation based on an excess demand @xmath42 $ ] .
but what relationship should we assume between @xmath42 $ ] and @xmath15 $ ] ? to answer this question , we should clarify what we want the individual actions @xmath43 } $ ] of the strategies to resemble .
in short , the question is ` what does a strategy map from and to ? ' in most studies of the mg as a market model , researchers have taken the actions @xmath44 } \in\left\ { -1,1\right\ } $ ] to be the actions of selling and buying a quanta of the asset respectively .
thus the net action of all agents @xmath15 $ ] naturally becomes the excess demand , i.e. we have : @xmath45 = a\left [ t-1\right ] \label{eq demand mg}%\ ] ] given this and the fact that the price change is an increasing function of excess demand ( for equations [ eq price formation mult ] and [ eq price formation add ] ) , the global information becomes the directions ( up or down ) of the previous price changes .
so the answer here is that a strategy maps from the history of past asset price movements to a buy / sell signal .
a strategy mapping _ from _ the history of past asset price movements seems a natural and sensible choice , since financial chartists do look at precisely this information to decide upon a course of action .
we therefore propose that in general , for this class of market model , we should always use the following equation ( eq . [ eq history update gen ] ) to update the global information , irrespective of whether equation [ eq demand mg ] holds .
@xmath26 = 2\mu\left [ t\right ] -p\operatorname{h}\left [ \mu\left [ t\right ] -\frac{p}{2}\right ] + \operatorname{h}\left [ \delta p\left [ t+1,t\right ] \right ] \label{eq history update gen}%\ ] ] it is worth now spending a few moments considering the timings in this model . equation [ eq demand mg ] is for the excess demand at time @xmath46 , i.e. prior to time @xmath5 when all the orders are processed and a new price @xmath47 $ ] is formed .
thus @xmath34 $ ] can only result from all the information that is available at time @xmath48 , i.e. @xmath49 $ ] and @xmath50 \right\ } $ ] . from this information ,
the agents take actions @xmath51 } $ ] producing a net action of @xmath52 $ ] .
the agents actions ( orders ) however do not get realized ( executed ) until time @xmath5 when the new price @xmath35 $ ] is known .
the mg payoff function of equation [ eq strat payoff mg ] , hence rewards agents positively for deciding to sell ( buy ) assets ( i.e. @xmath51 } = -1 $ ] @xmath53 ) when the order execution price @xmath35 $ ] is above ( below ) the price level at the time they made the decision ( i.e. @xmath54 $ ] ) .
why is this mechanism of moving in the minority a physically reasonable ambition for the agents ?
let us consider the ` notional ' wealth of an agent @xmath55 to be given by : @xmath56 = \phi_{i}\left [ t\right ] p\left [ t\right ] + c_{i}\label{eq agent wealth}%\ ] ] where @xmath57 is the number of assets held and @xmath58 is the amount of cash held .
it is clear from equation [ eq agent wealth ] that an exchange of cash for asset at any price does not in any way affect the agents notional wealth .
however , the point is in the terminology .
the wealth @xmath59 $ ] is only _
notional _ and not real in any sense - the only real measure of wealth is @xmath58 which is the amount of capital the agent has available to spend .
thus it is evident that an agent has to do a ` round trip ' ( i.e. buy ( sell ) an asset then sell ( buy ) it back ) to discover whether a _ real _ profit has been made .
let us consider two examples of such a round trip .
in the first case the agent trades in the minority while in the second he trades in the majority : * moving in minority : + [ c]|l|l|l|l|l|l|@xmath5 & action @xmath60 $ ] & @xmath61 $ ] & @xmath62 $ ] & @xmath35 $ ] & @xmath63 $ ] + 1 & submit buy order & 100 & 0 & 10 & 100 + 2 & buy ... , submit sell order & 91 & 1 & 9 & 100 + 3 & sell & 101 & 0 & 10 & 101 + * moving in majority : + [ c]|l|l|l|l|l|l|@xmath5 & action @xmath60 $ ] & @xmath61 $ ] & @xmath62 $ ] & @xmath35 $ ] & @xmath63 $ ] + 1 & submit buy order & 100 & 0 & 10 & 100 + 2 & buy ... , submit sell order & 89 & 1 & 11 & 100 + 3 & sell & 99 & 0 & 10 & 99 + as can be seen , moving in the minority creates wealth for the agent upon completion of the necessary round - trip , whereas moving in the majority loses wealth .
however if the agent had held the asset for any length of time between buying it and selling it back , his wealth would also depend on the rise and fall of the asset price over the holding period .
thus the mg mechanism seems perfectly reasonable for a collection of traders who simply buy / sell on one timestep and sell / buy back on the next , but this is not of course what real financial traders do in general .
this pinpoints the main criticism of the mg as a market model .
the minority mechanism at play in the mg , arises solely from equation [ eq strat payoff mg ] for the payoff ( reward ) @xmath64 $ ] given to each strategy @xmath18 based on its action @xmath44 } $ ] .
thus it can be trivially conjectured that changing the structure of the payoff function will change the class of game being played . in section [
sec mg to mm ] we pointed out that although _ trading _ in the minority was beneficial , the minority payoff structure itself made no allowance for the rise or fall in the value of the agent s portfolio @xmath57 of assets .
let us try to rectify this by examining the form of the agent s notional wealth , equation [ eq agent wealth ] .
if we differentiate the notional wealth , we get an expression for @xmath65 = w_{i}\left [ t+1\right ] -w_{i}\left [ t\right ] $ ] given by : @xmath66 = \delta c_{i}\left [ t+1,t\right ] + p\left [ t+1\right ] \delta\phi_{i}\left [ t+1,t\right ] + \phi_{i}\left [ t\right ] \delta p\left [ t+1,t\right]\ ] ] the first two terms cancel because the amount of cash lost @xmath67 $ ] is used to buy the extra @xmath68 $ ] assets at price @xmath69 $ ] .
this leaves us with : @xmath66 = \phi_{i}\left [ t\right ] \delta p\left [ t+1,t\right ] \label{eq agent wealth chng}%\ ] ] we can then use equation [ eq agent wealth chng ] to work out an appropriate reward @xmath64 $ ] for each strategy based on whether its action @xmath44 } $ ] induced a positive or negative increase in notional wealth .
let us first use the fact that the price change @xmath70 $ ] is roughly proportional to the excess demand @xmath71 $ ] : this can be seen explicitly from our earlier equation for the price formation , equation [ eq price formation add ] .
then let s use our mg interpretation of the net action @xmath72 $ ] , equation [ eq demand mg ] .
we thus have from equation [ eq agent wealth chng ] : @xmath73 & \varpropto\phi_{r}\left [ t\right ]
d\left [ \left ( t+1\right ) ^{-}\right ] \\ & \varpropto\phi_{r}\left [ t\right ] a\left [ t\right]\end{aligned}\ ] ] we then identify the accumulated position of a strategy at time @xmath5 , @xmath74 $ ] , to be the sum of all the actions ( orders ) made by that strategy which have been executed within times @xmath75 . remembering that at time @xmath5 , the action ( order ) @xmath44 } $ ] has _ not _ yet been executed ( it gets executed at @xmath76 ) , this gives @xmath74 = \sum_{i=0}^{t-1}a_{r}^{\mu\left [ i\right ] }
let us then set the payoff given to a strategy @xmath64 $ ] , to be an increasing ( odd ) function @xmath77 of the notional wealth increase @xmath78 $ ] for that strategy .
we thus arrive at : @xmath23 = \chi\left [ \sum_{i=0}^{t-1}a_{r}^{\mu\left [ i\right ] } \times a\left [ t\right ] \right ] \label{eq strat payoff nwealth}%\ ] ] we could also propose a locally - weighted equivalent of equation [ eq strat payoff nwealth ] , in which the reward given to a strategy is more heavily weighted on the result of its recent actions than the actions it made further in the past .
this gives @xmath23 = \chi\left [ \sum_{i=0}^{t-1}\left ( 1 - 1/t\right ) ^{t-1-i}a_{r}^{\mu\left [ i\right ] } \times a\left [ t\right ] \right ] \label{eq strat payoff nwealth t}%\ ] ] where @xmath30 enumerates a characteristic time - scale over which the position accumulated by the strategy is ` forgotten ' .
the @xmath79 limit of equation [ eq strat payoff nwealth t ] yields @xmath64 = \chi\left [ a_{r}^{\mu\left [ t-1\right ] } a\left [ t\right ] \right ] $ ] , i.e. only the position resulting from the most recently executed trade is taken into account . with @xmath79 ,
this payoff structure essentially rewards a strategy at time @xmath76 based on whether the notional wealth change @xmath80 $ ] was _ more positive _ than it would have been if action @xmath81 } $ ] had _ not _ been taken .
if equation [ eq strat payoff nwealth t ] is used in an agent - based market model , the agents play the strategy they hold which has accumulated the highest ` virtual ' notional wealth .
we mean ` virtual ' in the sense that the strategy itself will not have _ actually _ accumulated this notional wealth unless it has been played without fail since time @xmath82 .
the agents in this model are thus all striving to increase their notional wealth and are allowed to do so by taking arbitrarily large positions @xmath57 .
* we will call this incarnation of the market model $ g[eq strat payoff nwealth t ] * : the term ` $ g ' is named after the ` the $ -game ' of ref .
@xcite where agents also strive to maximize notional wealth , and ` [ eq strat payoff nwealth t ] ' is named after equation [ eq strat payoff nwealth t ] for the strategy payoff .
we stress that this incarnation of the market model is by no means unique . for example
, it may be more appropriate to consider a model wherein each agent is only allowed one position in the asset at any time , i.e. @xmath83 = -1,0,1 $ ] .
this amounts to us redefining the strategies @xmath18 as mapping to the actions ` go long / short ' instead of ` buy / sell ' , i.e. @xmath44 } = \phi_{r}\left [ \left ( t+1\right ) ^{-}\right ] $ ] . note the reference to @xmath84 here because the action @xmath44 } $ ] is not executed until time @xmath76 : after this time @xmath44 } = \phi_{r}\left [ \left ( t+1\right ) \right ] .$ ] in this new model , the net action of the agents @xmath15 $ ] therefore represents the overall desired position of the population at time @xmath84 , i.e. @xmath15 = \sum_{i=1}^{n}\phi_{i}\left [ \left ( t+1\right ) ^{-}\right ] $ ] .
it therefore follows that the excess demand is given by : @xmath85 & = \sum_{i=1}^{n}\left ( \phi _ { i}\left [ \left ( t+1\right ) ^{-}\right ] -\phi_{i}\left [ t\right ] \right ) \nonumber\\ & = a\left [ t\right ] -a\left [ t-1\right ] \label{eq demand dolg}%\end{aligned}\ ] ] similarly , for this system , the payoff to strategies based on the notional wealth increase is given by equation [ eq agent wealth chng ] as : @xmath86 & \varpropto\chi\left [ \phi_{r}\left [ t\right ] \delta p\left [ t+1,t\right ] \right ] \nonumber\\ & = \chi\left [ a_{r}^{\mu\left [ t-1\right ] } \left ( a\left [ t\right ] -a\left [ t-1\right ] \right ) \right ] \label{eq strat payoff posnwealth}%\end{aligned}\ ] ] * we will call this incarnation of the market model $ g[eq strat payoff posnwealth]*. the payoff structure for this one - position - per - agent system is essentially the same as that reported in equation [ 3 ] of ` the $ -game ' @xcite .
in the previous section , we introduced different microscopic reward schemes for use in binary multi - agent market models .
the aim of these new strategy payoff structures , was to get the agents to maximize their notional wealth as given by equation [ eq agent wealth ] .
this seems to be a more physically realistic goal than to have the agents competing to always be in the minority group , as in the mg and its variants .
the question now must be : how do these new payoff structures affect the resultant dynamics of the market model ? to investigate this , we identify the form of the excess demand in the model and then employ a price formation structure - from section [ sec price process ] for example .
the table below illustrates three possible market models : [ c]|l|l|l|l|l|name & @xmath87 $ ] & @xmath88 $ ] & @xmath64 = \operatorname*{sgn}\left [ \ldots\right ] $ ] & @xmath89 $ ] + gcmg & @xmath90 + a_{r}^{\mu\left [ t-1\right ] } $ ] & @xmath15 $ ] & @xmath91 } d\left [ \left ( t+1\right ) ^{-}\right ] $ ] & @xmath71 /\lambda$ ] + $ g[eq strat payoff nwealth t ] & @xmath90 + a_{r}% ^{\mu\left [ t-1\right ] } $ ] & @xmath15 $ ] & @xmath92 } d\left [ \left ( t+1\right ) ^{-}\right ] $ ] & @xmath71 /\lambda$ ] + $ g[eq strat payoff posnwealth ] & @xmath81 } $ ] & @xmath15 -a\left [ t-1\right ] $ ] & @xmath81 } d\left [ \left ( t+1\right ) ^{-}\right ] $ ] & @xmath88 /\lambda$ ] + we have simulated the three market models in the above table with the same parameters for each model : @xmath93 agents , @xmath94 strategies per agent , @xmath95 timesteps for the strategy score time - horizon , and an @xmath96 payoff point confidence - to - play threshold .
each simulation was run for agents with low memory @xmath97 and for agents with high memory @xmath98 .
the resultant price @xmath35 $ ] and volume @xmath99 = \sum_{r=1}^{2^{p}}n_{r}$ ] are displayed in figure [ fig mm price vol ] .
[ ptb ] figure1.eps the two different values of the agent s memory length ( or rather the ratio @xmath100 ) describe markets in two distinctly different ` phases ' .
arguably the most important feature of these types of model is that there is a common perception of a strategy s ( virtual ) success , i.e. @xmath101 $ ] is a global variable .
this implies that the number of agents adopting the same action ( i.e. the effect of crowding or herding ) is dependent on the number which hold each strategy . at low memory ( low @xmath100 )
there are few strategies available and many agents , consequently large numbers of agents hold the same strategy .
this implies that in this ` crowded phase ' there are large groups of agents adopting the same action @xmath44 } $ ] ( i.e. crowds ) and only small groups adopting the opposite action @xmath102 } $ ] ( i.e. anti - crowds ) @xcite .
the result of this high agent coordination is a volatile market with large asset price movements .
conversely , for high @xmath3 ( high @xmath100 ) hardly any agents hold the same strategy .
hence the crowd size is roughly equal to the anti - crowd size , resulting in a market with low agent coordination and consequently lower volatility and fewer large changes . with this distinction between low and high @xmath3 regimes in mind ,
let us briefly describe the dynamics in each market model .
* * gcmg*. first consider low @xmath3 .
the confidence - to - play threshold is set high enough in order that occasional stochasticity is injected via timesteps when @xmath15 = 0 $ ] ( n.b .
the next state of @xmath103 $ ] is then decided with a coin toss ) .
the gcmg can be seen to reproduce many of the statistical features observed in real markets .
figure [ fig mm price vol ] shows that the activity ( volume ) is generally low and bursty .
the asset price series is thus characterized by frequent large movements ( giving fat - tailed distributions of returns ) and clustered volatility .
no long periods of correlation exist in the gcmg asset price movement : as soon as lots of agents start taking the same action , then the strategies which produce that action are penalized , as can be seen from equation [ eq strat payoff mg ] . at high @xmath3 ,
the absence of agent co - ordination leads to an absence of clustering of activity .
hence the series of asset prices appears more random at high @xmath3 . *
* $ g[eq strat payoff nwealth t]*. at low @xmath3 , the asset price very soon assumes an unstoppable trend : agents with at least one trend - following strategy ( @xmath104 > 0,a\left [ t-1\right ] > 0,a\left [ t\right ] > 0\rightarrow a_{r}=1 $ ] ) join the trend and benefit ( notionally ) from the consequent asset price movement .
because the strategies ( and hence agents ) are allowed to accumulate limitless positions , the trend is self - reinforcing since @xmath105 just keeps getting bigger for the trend - following strategies . at high @xmath3 , the lack of agent coordination means that it is harder for the model to find this attractor .
however sooner or later , a majority of trend - following strategies will have accumulated sufficient score @xmath14 and position @xmath106 to be played successfully . from then on , the pattern of success is self - reinforcing and again the unstoppable trend is created .
this result has to be seen as the natural consequence of wanting the agents to maximize _
notional _ wealth at the same time as being allowed arbitrarily large positions . * * $ g[eq strat payoff posnwealth]*. at low @xmath3 , the model has interesting and irregular dynamical properties .
for example , the series of asset price movements contains periods of very high correlation ( and high volume ) and periods of very high anti - correlation ( and lower volume ) .
these two behaviors arise from the payoff structure of equation [ eq strat payoff posnwealth ] : just as in $ g[eq strat payoff nwealth t ] , strategies are rewarded for having a positive ( negative ) position when the asset price is trending up ( down ) , however they now get penalized for _ joining _ that trend .
this implies that persistent behavior will tend to follow persistent behavior , while anti - persistent behavior will follow antipersistent behavior .
this finding is very much in line with the school of thought that says that markets have distinct ranging and breakout phases . unlike the $ g[eq strat payoff nwealth t ] model ,
the trends are always capped by the limitation that agents can only hold one position .
once as many agents as possible are long ( short ) the asset price can not rise ( fall ) any further : consequently the resulting dynamics are strongly mean - reverting . at high @xmath3 , due to the lack of agent coordination
, it is hard for large enough crowds to arise to form trending periods . thus anti - persistence dominates this regime .
in the mg , there is no unstable attractor which could give rise to a continuously diverging price .
the reason is that each agent is trying to do the opposite of the others .
this results in no net long - term market force .
the mg agents can see no advantage in collectively forcing a price up ( down ) to increase the value of their long ( short ) position , since they are totally unaware of their accumulated position .
section [ sec mm behavior ] showed however that , when we give the agents the goal of maximizing their notional wealth ( cash plus value of position ) , the model can be dominated by the unstable dynamics of price trending . although the structure of the $ g models seems more realistic than the mg , the price dynamics do not reproduce many of the stylized features of financial asset price series .
it seems that an improvement in the microscopic model structure ` spoils ' the macroscopic results . to help explain this , we now look at some of the model deficiencies which still remain
. we hinted at this effect in section [ sec mg to mm ] .
the wealth @xmath63 $ ] that we have forced the agents of the $ g models to maximize , is purely _
notional _ - it is the wealth they _ would _ have if they could unwind their market positions at today s price @xmath47 $ ] .
the agents are clearly not able to do this .
there are two reasons for this .
even if an agent put in an order to unwind his position at time @xmath5 , it would not be executed until time @xmath76 when the price is @xmath107 $ ] .
the second reason is that in this simplistic structure , the agents can only trade one quanta of asset at any time .
hence a position of @xmath57 assets will take @xmath57 timesteps to fully unwind .
all other things being equal , an agent should expect on average to get less cash back than @xmath83 p\left [ t\right ] $ ] when unwinding his position .
this is ` market impact ' .
we can see this effect directly with the following argument : the change in cash of agent @xmath108 from buying @xmath109 $ ] assets at time @xmath5 is given by : @xmath110 = -\delta\phi_{i}\left [ t , t-1\right ] p\left [ t\right ] $ ] .
let us then use equation [ eq price formation add ] to give @xmath35 = p\left [ t-1\right ] + d\left [ t^{-}\right ] /\lambda$ ] .
then we break down the demand as follows : @xmath45 = d_{i}+\sum_{j\neq i}^{n}d_{j}\left [ t^{-}\right]\ ] ] where @xmath111 is the demand of agent @xmath112 .
however we know that the demand of agent @xmath108 is @xmath113 $ ] assets , hence we get @xmath114 & = -\delta\phi_{i}\left [ t , t-1\right ] p\left [ t\right ]
\\ & = -\delta\phi_{i}\left [ t , t-1\right ] p\left [ t-1\right ] -\frac { 1}{\lambda}\left ( \delta\phi_{i}\left [ t , t-1\right ] \right ) ^{2}-\frac { 1}{\lambda}\delta\phi_{i}\left [ t , t-1\right ] \sum_{j\neq i}^{n}d_{j}\left [ t^{-}\right]\end{aligned}\ ] ] if we now assume that the remaining agents @xmath115 form an effective ` background ' of random demands @xmath111 of zero mean , then averaging over this ` mean - field ' of agents gives : @xmath116 \right\rangle _ { j}=-\delta \phi_{i}\left [ t , t-1\right ] p\left [ t-1\right ] -\frac{1}{\lambda}\left ( \delta\phi_{i}\left [ t , t-1\right ] \right ) ^{2}\label{eq mkt impact}%\ ] ] equation [ eq mkt impact ] tells us that an agent will on average receive an amount of cash @xmath117 \right\rangle $ ] , which is equal to the amount he may have naively thought he would get ( if he could have traded at price @xmath118 $ ] ) minus a positive amount proportional to the square of his trade size and inversely proportional to the market liquidity .
consequently , an agent who considered his market impact in calculating his wealth would on average expect the cash - equivalent value of his wealth to be given by @xmath63 -\phi_{i}\left [ t\right ] ^{2}/\lambda$ ] , i.e. his notional wealth minus the average impact from unwinding his position at time @xmath76 .
this is the wealth that the agents _ should _ be attempting to maximize .
we could therefore propose , in the spirit of ref .
@xcite , a modification to the strategy reward function @xmath119 $ ] which takes account of this market impact : @xmath23 \varpropto\chi\left [ \phi_{r}\left [ t\right ] \delta p\left [ t+1,t\right ] -\phi_{r}\left [ t\right ] ^{2}/\lambda\right ] \label{eq strat payoff impact}%\ ] ] let us now re - visit the issue of trend following , this time with reference to a model whose strategies are updated using equation [ eq strat payoff impact ] .
* we will refer to this model as $ g[eq strat payoff impact]*. when a trend forms such that @xmath87 $ ] and @xmath120 $ ] have the same sign , strategy @xmath18 will be rewarded : this continues until the moment when the extra profit gained from increasing the position to benefit from the favorable price movement , is smaller than the added market impact which would be faced from unwinding the larger position . at this point
the strategy will start to be penalized instead of rewarded for supporting the trend .
this is a mechanism which could halt the price - divergent behavior of $ g[eq strat payoff nwealth t ] type models .
however equation [ eq strat payoff impact ] gives the turnover point in strategy reward due to market impact , to be given by @xmath121 = \lambda\delta p\left [ t+1,t\right ] $ ] .
if we then consider that during an ` unstable ' market trend the average price change is of order @xmath122 ( see , for example , the gradient of the $ g[eq strat payoff nwealth t ] price in figure [ fig mm price vol ] ) - and that @xmath123 increments by only @xmath124 every timestep - then trends which are formed before the market impact mechanism starts having a strong effect , will be of order @xmath125 timesteps in length .
the market impact mechanism on its own , is thus clearly too weak to stabilize the $ g models in order to form a realistic price series .
financial agents participating in a real market have finite resources and so can not keep buying and/or selling assets indefinitely .
this hard cut - off of agents resources in turn imposes a hard limit on the magnitude of price trends .
this effect was seen clearly in the $ g[eq strat payoff posnwealth ] model where the agents were only permitted ( or , equivalently because of finite resources , were only able to ) hold one position at any one time .
the $ g[eq strat payoff posnwealth ] and $ g[eq strat payoff nwealth t ] models can thus be thought of as opposite extremes in this respect , i.e. the first mimics extremely tightly - limited resources while the second mimics infinite resources .
it seems natural to expect that a real financial market lies somewhere between these two extremes ; resources are large for the market - moving agents but still finite .
we can include the effect of limited agent resources in our market model by allocating agent @xmath108 an initial capital @xmath126 $ ] ( and position @xmath127 $ ] ) and then updating this capital using @xmath61 = c_{i}\left [ t-1\right ] -\delta\phi_{i}\left [ t , t-1\right ] p\left [ t\right ] $ ] @xcite .
the agents can not then trade at time @xmath5 if the following conditions are met : @xmath128 & > 0\cap c_{i}\left [ t-1\right ] < \delta\phi_{i}\left [ t , t-1\right ] p\left [ t-1\right ] \\ \delta\phi_{i}\left [ t , t-1\right ] & < 0\cap\phi_{i}\left [ t-1\right ] < 0\end{aligned}\ ] ] in other words , an agent will not submit a buy order unless he at least has the capital to buy the asset at the quoted price @xmath118 $ ] , and also will not submit an order to short sell if he already holds a short position .
if we imposed no limit on short selling , an unstable state of the system would exist wherein all agents short - sell indefinitely .
* we will refer to this model as $ g[eq strat payoff nwealth t]w * where w represents ` wealth ' .
let us initialize this market model with agent wealths such that the agents initial buying power is equal to their initial selling power , i.e. @xmath129 , \phi_{i}\left [ 0\right ] \right\ } = \left\ { n\times p\left [ 0\right ] , n-1\right\ } .$ ] hence initially , the agents have the power to buy or sell @xmath130 assets ( n.b . recall that they are allowed up to one short sell ) .
if we then raise @xmath130 from @xmath131 upwards , we can see a qualitative change in the dynamical behavior of the model between the two extremes of $ g[eq strat payoff posnwealth ] and $ g[eq strat payoff nwealth t ] , with the periods of trending growing longer as @xmath130 increases . to investigate this behavior numerically , we fix @xmath130 and run the model for 1500 timesteps , recording the value of the global information @xmath6 $ ] in the last 500 timesteps @xmath5 .
we then count the number of times within this last 500 timesteps that either @xmath132 ( implying negative price movements over the last @xmath3 timesteps ) or @xmath133 ( positive price movements ) occur .
we denote the frequency with which these states occur as @xmath134 .
if the model visited all states of the global information @xmath6 $ ] equally , we would expect @xmath135 .
figure [ fig ftrend ] , showing the variation of @xmath134 with @xmath130 , has several interesting features .
first it can be seen that , as @xmath130 is increased and the consequent wealth available to the agents grows , the tendency of the model to be dominated by price trending increases dramatically .
also we see that at low @xmath130 , @xmath134 is below that of the ( random ) equally visited @xmath12-states case .
this is due to the high degree of anti - persistence in the system as was seen in the $ g[eq strat payoff nwealth t ] model .
the large spread in the results arises from the tendency of the model to exhibit clustering of activity states : persistence follows persistence and anti - persistence follows antipersistence as described in section [ sec mm behavior ] .
[ ptb ] figure2.eps we note that in the work of giardina and bouchaud ( see refs .
@xcite ) the strategy payoff function has the same basic form as eq .
[ eq strat payoff posnwealth ] , i.e. it is a $ g[eq strat payoff posnwealth]w - style model .
these authors consider proportional scoring ( i.e. @xmath136 ) and include an interest rate which resembles a resource level @xmath21 ; these features will not change the qualitative results presented here . however , in the models of refs .
@xcite , it seems that only active strategies are rewarded thereby breaking the common perception of a given strategy s success among agents which is arguably a central feature of the binary model structure .
we expect this to dramatically affect the dynamical properties of the system as the mechanism for agent coordination and thus crowd formation has been altered .
the interesting dynamical behavior of market models based on the binary strategy approach of the mg , arises largely as a consequence of the heterogeneity in the strategies held by the agents and the common perception of a strategy s success .
the extent to which agents will agree on the best course of action , and the consequent crowding , is solely a function of how the @xmath13 available strategies are distributed amongst the population of agents . if the way in which the strategies are initially allocated is biased , then the resulting dynamics will reflect this @xcite .
the introduction of an agent wealth , as discussed in the previous subsection , also brings about diversity in the market model .
even if we initiate the model with all agents having an equal allocation of wealth in the form of cash plus assets , the wealths of the agents @xmath137 will soon become heterogenous as a direct result of their heterogenous strategies .
figure [ fig wealth pdfs ] shows the heterogeneity of agents wealth growing with time during a $ g[eq strat payoff nwealth t]w simulation .
after many timesteps have elapsed , the distribution of agents wealth seems to reach a steady - state : many agents have lost the majority of their wealth to a minority of agents who themselves have accumulated much more . in everyday terminology ,
` the rich get richer while the poor get poorer ' .
[ ptb ] figure3.eps the heterogeneity of agents wealth in the $ g[eq strat payoff nwealth t]w model is fed back into the system through the buying power of the agents alone . however , although wealthier agents have the potential to buy and sell more assets , they still only trade in single quanta of the asset at any given timestep , the same as poorer agents .
it may be more reasonable to propose that the agents trade in sizes proportional to their resources @xcite , i.e. @xmath128 & = \gamma\frac{c_{i}\left [ t-1\right ] } { p\left [ t-1\right ] } \text { \ \ for \ \ } a_{r_{i}^{\ast}}^{\mu\left [ t-1\right ] } = 1\\ \delta\phi_{i}\left [ t , t-1\right ] & = \gamma\phi_{i}\left [ t-1\right ] \text { \ \ for \ \ } a_{r_{i}^{\ast}}^{\mu\left [ t-1\right ] } = -1\end{aligned}\ ] ] where @xmath138 is the highest scoring strategy @xmath18 of agent @xmath108 .
the factor @xmath139 then enumerates what fraction of the agents resources ( cash for buying , assets for selling ) they will transact at any given time . in this system
, assets will generally need to be divisible , i.e. the sense of agents buying and selling a quanta of the asset as in $ g[eq strat payoff nwealth t]w is lost .
this in turn means that instead of the degree of trending being controlled by the level of initial resource allocation @xmath130 , it is instead determined by @xmath140 , since this effectively determines the number of trades agents can make in any trending period before hitting the boundary of their capital resources .
also with this system of trading in proportion to wealth , trends will start steep and end shallow as agents run out of resources and thus make smaller and smaller trades .
apart from these stylized differences , the system has a very similar dynamical behavior to the more straight - forward $ g[eq strat payoff nwealth t]w model .
diversity in strategies and wealth are the two big sources of agent heterogeneity that we have covered so far .
this agent heterogeneity has led to a market model with dynamical behavior which is interesting and diverse over a large parameter range .
however , the typical price / volume output only starts to become representative of a real financial market at higher @xmath3 , as can be seen in figure [ fig dolgw price vol ] . as discussed in section [ sec mm behavior ] , the high @xmath3 regime represents a ` dilute ' market where very few traders act in a coordinated fashion
. however one would expect that a real financial market is _ not _ in a dilute phase at all , since it _ does _ have large groups of agents forming crowds which rush to the market together , thereby creating the bursty activity pattern observed empirically .
why then does this model , when pushed into the low @xmath3 regime , produce endless bubbles of positive and negative speculation as shown in the left - hand panel of fig .
[ fig dolgw price vol ] .
although a real financial market does show _ some _ suggestion of oscillatory bubble formation , it is nowhere near so pronounced and is certainly not on such a short timescale .
[ ptb ] figure4.eps to answer this question , we push further the subject of agent diversity . although our agents may have differing sets of strategies and consequently different wealths , they all act on the same timescale . when we look at charts such as figure [ fig dolgw price vol ] , we see patterns not only on a small tick - by - tick scale but also on a much larger scale
. in fact we can identify patterns all the way up to the ` macro ' scale of the boom - bust speculatory bubbles . from a knowledge of these patterns
, we would form opinions about what will happen next and would trade accordingly to maximize our wealth .
the agents we have modelled , however , can not view the past price - series in this way ; instead they are forced to only consider patterns of length @xmath3 timesteps .
patterns of any length greater than this , go un - noticed by the agents and hence are not traded upon .
this explains why these patterns can arise and survive .
by contrast , when a pattern is traded upon ( even in the non - mg framework ) it is slowly removed from the market .
consider the following pattern : [ c]llllllllllllllltime @xmath5 & 1 & 2 & 3 & 4 & 5 & 6 & 7 & 8 & 9 & 10 & 11 & 12 & 13 & 14 + price @xmath35 $ ] & 14 & 10 & 11 & 12 & 13 & 15 & 14 & 10 & 11 & 12 & 13 & 15 & 14 & 10 if we were able to identify this pattern repeating , our best course of action would be to submit a buy order between times @xmath141 and @xmath142 , i.e. @xmath143 = 1 $ ] .
we could then buy the asset at @xmath144 = 10 $ ] .
then between times @xmath145 and @xmath146 , we would submit an order to sell : @xmath147 = -1 $ ] and hence would have sold the asset at @xmath148 = 15 $ ] .
we then continue : @xmath149 = 1 $ ] , @xmath150 = -1 $ ] ...
this ensures we always buy at the bottom price and sell at the top
. however , trading in this way is _ against _ the trend as @xmath60 a\left [ t\right ] < 0 $ ] .
it is in effect minority trading .
hence trading to maximize our wealth with respect to this pattern , leads to the weakening of the pattern itself just as in the mg .
we conclude therefore that the presence of strong patterns in the $ g[eq strat payoff nwealth t]w and other similar market models at low @xmath3 , is simply due to the absence of agents within the model who can identify these patterns and arbitrage them out . from the proceeding discussion ,
it is clear that a realistic market model should include agents who can analyze the past series of asset price movements over different time - scales .
a first thought on how to do this , is to include a heterogeneity in the memory length @xmath3 of the agents . in this framework
, agents look at patterns not only of differing length but also of differing complexity .
it is therefore appealing to propose a generalization to the way in which the agents interpret the global information of past price movements , in such a way as to allow observation of patterns over different times - scales but having the _ same _ complexity ( @xmath3 ) .
this can be simply achieved by allowing the agents to have a natural information bit - length @xmath151 such that the global information available to them @xmath152 $ ] is updated according to the following generalization of equation [ eq history update gen ] : @xmath153 = 2\mu_{\tau}\left [ t\right ] -p\operatorname{h}% \left [ \mu_{\tau}\left [ t\right ] -\frac{p}{2}\right ] + \operatorname{h}% \left [ \delta p\left [ t+1,t+1-\tau\right ] \right]\ ] ] the table below then shows how the previous pattern would be encoded by agents having @xmath154 : [ c]|l|l|l|l|l|l|l|l|l|l|l|l|l|l|l|time @xmath5 & @xmath155 & @xmath156 & @xmath157 & @xmath158 & @xmath159 & @xmath146 & @xmath160 & @xmath161 & @xmath162 & @xmath163 & @xmath164 & @xmath165 & @xmath166 & @xmath167 + price @xmath35 $ ] & @xmath167 & @xmath163 & @xmath164 & @xmath165 & @xmath166 & @xmath168 & @xmath167 & @xmath163 & @xmath164 & @xmath165 & @xmath166 & @xmath168 & @xmath167 & @xmath163 + ` best ' action @xmath169 $ ] & @xmath155 & @xmath155 & @xmath155 & @xmath155 & @xmath170 & @xmath171 & @xmath155 & @xmath155 & @xmath155 & @xmath155 & @xmath170 & @xmath170 & @xmath155 & @xmath155 + @xmath172 \right ] $ ] & & @xmath173 & @xmath174 & @xmath174 & @xmath174 & @xmath174 & @xmath173 & @xmath173 & @xmath174 & @xmath174 & @xmath174 & @xmath174 & @xmath173 & @xmath173 + @xmath175 \right ] $ ] & & & @xmath173 & @xmath174 & @xmath174 & @xmath174 & @xmath174 & @xmath173 & @xmath173 & @xmath174 & @xmath174 & @xmath174 & @xmath174 & @xmath173 + @xmath176 \right ] $ ] & & & & @xmath173 & @xmath174 & @xmath174 & @xmath174 & @xmath173 & @xmath173 & @xmath173 & @xmath174 & @xmath174 & @xmath174 & @xmath173 + @xmath177 \right ] $ ] & & & & & @xmath173 & @xmath174 & @xmath174 & @xmath173 & @xmath173 & @xmath173 & @xmath173 & @xmath174 & @xmath174 & @xmath173 + if each agent only considered the past two bits of information , i.e. @xmath178 , then the ` best ' strategies for different values of the information bit - length @xmath151 would be as shown in the table below .
these ` best ' strategies have been obtained from inspection of the table above ; in particular , by looking at when the different @xmath178 bitstrings @xmath179 occur and seeing the respective ` best ' action given this bit string @xmath180 $ ] .
[ c]cc & @xmath151 + @xmath6 $ ] & [ c]|l|l|l|l|l| & @xmath155 & @xmath156 & @xmath157 & @xmath158 + @xmath181 & @xmath155 & @xmath155 & @xmath155 & @xmath182 + @xmath183 & @xmath155 & @xmath155 & @xmath170 & @xmath170 + @xmath184 & @xmath155 & @xmath155 & @xmath155 & @xmath155 + @xmath185 & @xmath182 & @xmath182 & @xmath182 & @xmath155 + the ` best ' strategies ( @xmath186 ) in the above table , show a question mark ( ? ) next to a particular value of the global information @xmath6 $ ] .
this denotes that for this state , the best action is sometimes @xmath187 } = 1 $ ] and sometimes @xmath188 } = -1 $ ] .
it can thus be seen that only when we include longer time - scale patterns @xmath189 , do we get a clear signal of when is the optimal time to sell ( @xmath187 } = -1 $ ] ) .
shorter timeframe patterns provide no clear indication when to do so strategy could encode enough information to optimally arbitrage the pattern .
however in general , for longer patterns , this is less likely to be true . ] .
this then demonstrates that an agent holding strategies of different bit - length @xmath151 could identify optimal times to buy and sell , and hence arbitrage patterns of length very much greater than the memory length @xmath3 .
so far this section has discussed how instabilities and inefficiencies in the model market ( as indicated by repeating , un - arbitraged patterns ) could be remedied by inclusion of more realistic , sophisticated and diverse strategic agents .
however the agents we have considered have all acted in the same fashion , i.e. with the goal of maximizing their wealth by employing a strategic methodology based on the observable endogenous market indicators . in terms of a real financial market
, we have considered modelling the subset of agents who consider the past history of asset price movements as the only relevant information .
although in reality we expect this subset to be large , we certainly acknowledge the presence of agents who strategically use other endogenous market indicators such as volume as a source of global information .
the strategic use of other endogenous market indicators simply gives a more sophisticated model within the same basic framework .
what we have so far neglected to include is the presence of agents using exogenous information to influence their trading decisions .
possible sources of exogenous information are as vast as human imagination itself : they could for example range from economic indicators through to company reports and rumor , or even ` gut - feeling ' and astrology ! within the context of our model framework , strategies employed based on these external sources of information are un - strategic with respect to the endogenous information , and as such represent a stochastic influence .
such stochasticity , representing the response of agents to exogenous information , can be incorporated into the models in a variety of ways .
for example a background level of stochastic action could be assumed at each timestep , or be made to appear as a poisson process with a given mean frequency .
however , it seems most physically satisfying to incorporate stochastic action in the form of an extra pair of strategies for each ( or at least some ) agent : one buy strategy , @xmath190 , and one sell strategy , @xmath191 .
agents could then be given a probability of using each of these strategies in place of their ` usual ' chartist strategies .
furthermore , the probabilities of using these strategies could be allowed to evolve .
this system of modelling response to an exogenous signal is essentially the ` genetic game ' of ref .
alternatively , the probability of using one of these strategies representing the response to an exogenous signal , could be based on an endogenous market indicator .
an example of this is the mechanism of ref .
@xcite of switching to ` fundamentalist behavior ' . in the model of ref .
@xcite , agents use strategy @xmath192 \right ] } $ ] with a probability @xmath193 which increases with @xmath194 -p_{0}\right| $ ] , where @xmath195 represents a perceived ` fundamental ' value for the asset . this mechanism , which can be thought of as encoding the behavior of irrational fear , clearly acts to break the formation of long trends since the probability of trading against the trend increases with the duration of the trend itself .
more generally , we suspect that among the agent - based market models in the literature which manage to reproduce stylized facts , the inclusion of some level of stochasticity within the model is crucial for helping to break up unphysical dynamics and hence achieve market - like behavior .
in the framework of the mg , each strategy @xmath8 comprises the elements @xmath9 which represent a response @xmath10 to each of the possible @xmath11 values of the global information @xmath12 ( see section [ sec mg basics ] ) . in section [ sec mg to mm ] , we identified the different states of the global information @xmath12 as corresponding to different patterns in the past history of asset price movements ( binary : up / down ) .
we can therefore think of a strategy as a book of chartist principles recommending an action for each and every possible pattern of length @xmath3 .
each of these strategy ` books ' thus has @xmath196 pages , one for each pattern .
there are @xmath197 possible books an agent can buy and use as guidelines on how to trade .
an agent in possession of one of these strategy ` books ' , whichever he chooses to buy , can find a page giving guidelines on how to trade in _ every _ possible market state .
the books are thus ` complete ' .
it seems unlikely however , that in reality such a ` complete ' book would exist for any arbitrary value of @xmath3 .
indeed , even if such a book did exist , it is unlikely that a given market participant would consider all the pages as being true .
for example , page one of a strategy book may say that if the asset price has fallen three days running , one should sell assets on the fourth day .
page three may say that if the asset price has fallen , then recovered and then fallen again , it is advisable to buy .
the agent holding this book may well believe in page one but think that the guidelines of page three are rubbish : in particular , he considers the pattern down - up - down to correspond to no trading signal at all . this agent would therefore continue to hold any previous position he had if he saw the pattern down - up - down .
we therefore propose a generalization to the strategy structure of the mg to account for the fact that strategy ` books ' may be incomplete , or equivalently that agents do not trust some of the trading signals .
before trading commences , each agent goes through each of her ( complete ) strategy ` books ' page by page , and decides whether she believes in the trading guidelines given for each of the possible @xmath7 market ` signals ' . the agents choose to accept each page of each of their books with a probability @xmath198 .
if the agent rejects a page of the strategy ` book ' , they replace the suggested action for time @xmath5 , i.e. @xmath44 } \in\left [ -1,1\right ] $ ] , by a null action @xmath44 } = \ast$ ] .
when trading commences , the agents are then faced with deciding what course of action to take when their strategy ` books ' register the null action at time @xmath5 .
one type of possible behavior would be to maintain their prior course of action , i.e. @xmath44 } = a_{r}^{\mu\left [ t-1\right ] } $ ] .
another alternative behavior would be to maintain their prior position , i.e. @xmath199 = \phi_{i}\left [ t\right ] $ ] .
( this results in @xmath44 } = 0 $ ] in $ g[eq strat payoff nwealth t ] type models and @xmath44 } = a_{r}^{\mu\left [ t-1\right ] } $ ] in $ g[eq strat payoff posnwealth ] models ) .
we now examine the qualitative differences between a model with this incomplete strategy structure , and a model with complete strategies . for this purpose
we will assume a gcmg type model , i.e. a model with strategy payoff given by equation [ eq strat payoff mg ] .
figure [ fig terngw price vol ] shows the resulting price and volume from two market simulations with incomplete strategies , the probability of non - null action being @xmath200 .
these are compared with the gcmg which has complete strategies .
[ ptb ] figure5.eps the top - left chart of figure [ fig terngw price vol ] shows price and volume for a market where agents maintain their previous course of action when they encounter a null strategy action .
this market shows high volatility in price , a high number of very large movements , and a background of high volume trading with occasional sharp spikes .
we suggest that this behavior arises from the tendency of agents to keep repeating their previous action in the absence of a new market pattern .
clearly then the ` large - movement ' states of up - up - up or down - down - down will be occupied for longer than the ` ranging ' states ( up - down - up etc . ) because these are the only states of @xmath12 which can map onto themselves ( @xmath6 = \mu\left [ t+1\right ] = \mu\left [ t+2\right ] \ldots$ ] ) thus providing no new market pattern .
the top right chart of figure [ fig terngw price vol ] shows price and volume for a market where agents maintain their previous position ( i.e. @xmath44 } = 0 $ ] in this model ) when they encounter a null strategy action .
the volume in this market simulation is very low , giving a consequent low volatility .
this is due to the fact that the agents will not trade unless they receive a market signal for which they have a non - null strategy action .
this model still shows a high probability for price movements which are large compared to the volatility .
these large movements occur in moments of high agent coordination , i.e. when many agents find they have a strategy with the same non - null action .
however in contrast to the previous model , these large movements do nt seem to persist over many timesteps .
it is interesting to compare the statistical features of these models featuring incomplete strategies , with the features of the gcmg .
figure [ fig terngw momes ] shows the volatility and kurtosis of the price movements @xmath89 $ ] for a model with incomplete strategies , as a function of the probability @xmath198 that a strategy has a non - null action for a given global information state .
the market model used has the structure of the gcmg ( n.b . for @xmath201 ,
it _ is _ the gcmg ) and features the behavior wherein agents maintain their previous course of action when they encounter a null strategy action . as a consequence , the volatility of the price movements rises as @xmath198 is reduced since it leads to more persistence of agent action and consequently bigger asset price movements .
this affect can also be seen in the kurtosis which grows rapidly away from the gaussian case , since the probability of large moves increases with falling @xmath198 . [ for more details on the nature of large movements in the gcmg , we refer to ref .
] [ ptb ] figure6.eps
in the preceding sections we have discussed mechanisms through which financial market agents come to make trading decisions about whether to buy or sell an asset . in section [ sec price process ]
we made some suggestions about how the combined action of all these agents decisions produced a demand , and how that demand then affected the movement of the asset price .
so far we have not explicitly mentioned a mechanism by which the market clears , i.e. by which sellers of the asset meet buyers .
the discussion of specific clearing mechanisms amounts to a discussion of specific market microstructure .
however our goal here is to seek to be as general as possible , and hence not limit the scope of our market models to any particular sector or style of financial market .
we thus limit ourselves here to a simplistic discussion of market - making .
the first important aspect of the models we have considered so far is that the agents make a decision about whether to buy or sell assets at time @xmath5 based only on the information up to ( but not including ) time @xmath5 .
the orders as such are considered as ` market orders ' because they are unconditional on the actual traded price @xmath35 $ ] . orders
whose execution depends on the traded price @xmath35 $ ] are known as ` limit orders ' and are beyond the scope of the model types presented here .
the presence of only market orders in our model brings us to the second important consideration : the number of buy orders in general does not balance the number of sell orders , i.e. in general @xmath42 \neq0 $ ] since all orders seek execution at time @xmath5 irrespective of the price .
there are two options then of how the market could function .
the first option is to assume that the number of assets sold by the agents must equal the number bought by the agents . in this case
the total number of assets traded is given by , @xmath202 = \min\left [ n_{\text{buy}}\left [ t^{-}\right ] , n_{\text{sell}}\left [ t^{-}\right ] \right]\ ] ] where @xmath203 is the number of agents who seek to buy , and @xmath204 the number who seek to sell .
with this system , the agents in the majority group ( buyers or sellers ) only get their order partially executed : for example if @xmath205 and @xmath206 , the sellers can each only sell @xmath207 /n_{\text{sell}% } = 40/50=80\%$ ] of a quanta of assets .
this is the system proposed in ref .
an alternative system assumes that there is a third party , a market - maker , who operates between buyers and sellers and makes sure that each and every order is fulfilled by taking a position in the asset herself .
we then have that : @xmath202 = \max\left [ n_{\text{buy}}\left [ t^{-}\right ] , n_{\text{sell}}\left [ t^{-}\right ] \right]\ ] ] and that the market - maker s inventory is : @xmath208 = \phi_{m}\left [ t-1\right ] -d\left [ t^{-}\right ] \label{eq mm inventory}%\ ] ] the market - maker , unlike the agents who just seek to maximize their wealth , has the joint goals of wealth maximization and staying market - neutral , i.e. as close to @xmath37 = 0 $ ] as possible .
if the nature of the market is highly mean - reverting , the second of these conditions will be automatically true since @xmath209 \sim0 $ ] . in these circumstances ,
the market - maker should employ an arbitrage - free price formation process such as equation [ eq price formation mult ] or equation [ eq price formation add ] .
if on the other hand the market is not mean - reverting , then the market - maker s position is unbounded . in these circumstances
a market - maker should employ a price - formation process such as equation [ eq price formation mult mm ] which encourages mean reversion through manipulation of the asset price based on @xmath210 $ ] , as described in section [ sec price process ] .
a market - maker using a price formation such as equation [ eq price formation mult mm ] can however be arbitraged by the agents . whether the market - maker is actually arbitraged is then purely dependent on the underlying dynamics of the market .
let us demonstrate this by following ref .
@xcite using a $ g[eq strat payoff posnwealth ] model . for this model
we have : @xmath45 = a\left [ t-1\right ] -a\left [ t-2\right ] \label{eq demand dolga}%\ ] ] as shown in section [ sec payoff mods ] . from equation [ eq mm inventory ]
we then get that @xmath37 = \sum_{j=0}^{t-1}d\left [ t\right ] = \phi_{m}\left [ 0\right ] -a\left [ t-1\right ] $ ] this immediately tells us that the market - maker s inventory is bounded
. the natural choice of price formation process to use would then be the arbitrage - free case of , for example , equation [ eq price formation mult ] .
if we combine equation [ eq price formation mult ] with equation [ eq demand dolga ] we arrive at : @xmath33 = p\left [ 0\right ] e^{\frac{a\left [ t-1\right ] } % { \lambda}}\label{eq price dolg nomm}%\ ] ] we can thus see how the $ g[eq strat payoff posnwealth ] model produced the tightly bounded , jumpy price series seen in figure [ fig mm price vol ] : the price at time @xmath5 is no longer directly dependent on the price at time @xmath48 and is bounded by @xmath211 e^{-\frac{n}{\lambda}}\leq p\left [ t\right ] \leq p\left [ 0\right ] e^{\frac{n}{\lambda}}$ ] . given the behavior of the $ g[eq strat payoff posnwealth ] model , the use of the price formation process of equation [ eq price formation mult mm ] seems rather unnecessary : the market maker does not need to manipulate the price in order to stay market - neutral on average . however , let us demonstrate the effect of using equation [ eq price formation mult mm ] as in ref .
@xcite . combining equations [ eq price formation mult mm ] and [ eq demand dolga ] with the market maker s inventory of @xmath212 = -a\left [ t-1\right ] $ ] we arrive at : @xmath33 = p\left [ t-1\right ] \exp\left [ \frac{1}{\lambda}a\left [ t-1\right ] + \left ( \frac{1}{\lambda_{m}}-\frac{1}{\lambda}\right ) a\left [ t-2\right ] \right ] \label{eq price dolg mm}%\ ] ] unlike equation [ eq price dolg nomm ] , equation [ eq price dolg mm ] gives us again a multiplicative price process for finite values of @xmath41 .
although it seems an inappropriate step , we would thus expect that introducing equation [ eq price formation mult mm ] as a price formation process would help regain some of the ` stylized features ' of a financial market inherent from the multiplicative price process .
however , closer consideration of equation [ eq price dolg mm ] reveals something different .
let us consider the case of @xmath213 = a\left [ t-1\right ] $ ] . in this case
the price change is simply an increasing function of @xmath52 $ ] .
if more agents are long rather than short at time @xmath5 ( @xmath52 > 0 $ ] ) , then the price movement @xmath214 $ ] is positive .
the agents will then have benefitted from a positive increment in their notional wealth from simply holding their position .
this manipulation of the ( naive ) market - maker leads to an unstable state whereby the agents arbitrage the market - maker simply by holding their long / short positions and watching the price rise / fall .
this unstable state of the market will occur with increasing probability as the agents become more coordinated in their actions , i.e. as @xmath3 is decreased .
for high @xmath3 , where there is no longer significant agent coordination and no crowding affects , the $ g[eq strat payoff posnwealth ] model with equation [ eq price dolg mm ] for the price formation is able to avoid the unstable attractor and thus produce more ` market - like ' time - series as demonstrated in ref .
we can also examine the changing wealth of the market - maker by simply employing equation [ eq agent wealth chng ] .
we know that @xmath212 = -a\left [ t-1\right ] $ ] , so equation [ eq agent wealth chng ] combined with equation [ eq price dolg mm ] gives us : @xmath215 & = \phi_{m}\left [ t\right ] \delta p\left [ t+1,1\right ] \nonumber\\ & = -\frac{1}{\lambda}a\left [ t\right ] a\left [ t-1\right ] -\left ( \frac{1}{\lambda_{m}}-\frac{1}{\lambda}\right ) a\left [ t-1\right ] ^{2}\label{eq mm wealth chng}%\end{aligned}\ ] ] in the case of @xmath216 as in ref .
@xcite , we see can see - by taking a time average of equation [ eq mm wealth chng ] - that whether the market - maker s wealth increases or decreases with time will depend on the dynamics of the market . if the market is persistent as in the unstable case at low @xmath3 , the market maker loses wealth since @xmath217 a\left [ t-1\right ] \right\rangle > 0 $ ] .
conversely if the market is anti - persistent , the market - maker will gain wealth since @xmath218 a\left [ t-1\right ] \right\rangle < 0 $ ] . at first sight
, it therefore seems that the $ g[eq strat payoff posnwealth ] model can generate a ` reasonably market - like ' state wherein the price is not unstable , it exhibits market - like dynamics and has a market - maker who is able to make a profit from her role .
however , in order to generate this state it was necessary to use what seems to be an inappropriate market - making mechanism and also push the model into the uncrowded phase .
neither of these choices seem desirable in the quest for producing a realistic market model .
following the introduction to the community of the minority game , there have been many proposed agent based market models , many of which exhibit the ` stylized facts ' of financial market timeseries in certain parameter ranges .
it seems however that many of these market models include one or more basic assumptions which seem implausible when compared to the actions of real financial market agents . with the first ` wave ' of market models it seemed remarkable that so many of the stylized features of real financial timeseries were reproduced so well . with more and more market model contenders entering the community , each with a quite diverse set of assumptions , it started to look as though it was easy to find models which would reproduce realistic market - like dynamics .
it has been proposed @xcite that general systems which have activity based on a waiting time problem , will all enjoy the scaling properties seen in financial market data .
notwithstanding this observation , we believe that it is not at all obvious how one can produce a self - consistent market model which embodies , in a _ non - stochastic _
way , the behavior of realistic financial market agents , and yet can reproduce the ` stylized facts ' of real financial markets . as we have shown in this work , adaptation of the mg payoff structure to incorporate a more realistic objective function results in a model increasingly dominated by trending as agents resources are increased . as a result of this
we are forced to look to a wider class of agent behavior and heterogeneity in order to recover the dynamical behavior we are looking for .
this paper has provided a discussion of different aspects of market models and how they interact with each other . instead of proposing a unique new market model of our own ( thereby adding to the increasing family in the literature ) , we have examined the different features one may want to include in a market model , and then discussed what their effect would be .
we hope that the resulting discussion has managed to provide a basic of toolbox of model features , brought together by a common formalism .
perhaps some optimal combination of a subset of these model features will provide a new self - consistent , realistic yet minimal market model .
however it is likely that one needs a diversity not only of agent parameter values but also of agent behavior , if one is intent on forming _ the _ market model .
finally we emphasize that this paper falls into the category of an _ interim progress report _ on our research in this area to date - a synthesis of our work over the past few years and thoughts , particularly in light of recent proposals for modified agent - based models .
we would therefore welcome any comments and , in particular , criticisms of our results and opinions .
we also thank the many people with whom we have enjoyed discussions about financial modelling , both within and outside the econophysics community .
see econophysics website www.unifr.ch/econophysics .
see also j. bouchaud and m. potters , _ theory of financial risks _
( cambridge university press , cambridge , 2000 ) ; r. mantegna and h. stanley , _ econophysics _
( cambridge university press , cambridge , 2000 ) ; b. huberman , p. pirolli , j. pitkow , and r. lukose , science * 280 * , 95 ( 1998 ) ; t. lux and m. marchesi , nature * 397 * , 498 ( 1999 ) ; r.g .
palmer , w.b .
arthur , j.h .
holland , b. lebaron and p. tayler , physica d * 75 * , 264 ( 1994 ) .
t. chordia , r. roll and a. subrahmanyam , in press j. fin . econ .
( 2001 ) ; j. lakonishok , a. shleifer , r. thaler and r. vishny , j. fin .
econ . * 32 * , 23 ( 1991 ) ; s. maslov and m. mills , physica a * 299 * , 234 ( 2001 ) ; d. challet and r. stinchcombe , physica a * 300 * , 285 ( 2001 ) .
m. hart , p. jefferies , n.f .
johnson and p. m. hui , physica a * 298 * , 537 ( 2001 ) ; n.f .
johnson , m. hart , and p.m. hui , physica a * 269 * , 1 ( 1999 ) ; m. hart , p. jefferies , n.f .
johnson , and p. m. hui , phys .
e * 63 * , 017102 ( 2000 ) ; n.f .
johnson , p.m. hui , d. zheng , and m. hart , j. phys .
a : math . gen . * 32 * , l427 ( 1999 ) ; t.s .
lim , p.m. hui and n.f .
johnson , physica a * 287 * , 313 ( 2000 ) ; p. jefferies , m. hart , n.f .
johnson , and p.m. hui , j. phys .
a * 33 * , l409 ( 2000 ) ; m. hart , p. jefferies , p.m. hui and n.f .
johnson , eur .
j. b * 20 * , 547 ( 2000 ) .
johnson , p.m. hui , r. jonson and t.s .
lo , phys .
lett . * 82 * , 3360 ( 1999 ) ; t.s .
lo , p.m. hui , and n.f .
johnson , phys .
e * 62 * , 4393 ( 2000 ) ; p.m. hui , t.s .
lo , and n.f .
johnson , physica a * 288 * , 451 ( 2000 ) ; n.f .
johnson , d.j.t .
leonard , p.m. hui , and t.s .
lo , physica a * 283 * , 568 ( 2000 ) . | in light of the growing interest in agent - based market models , we bring together several earlier works in which we considered the topic of self - consistent market modelling . building upon the binary game structure of challet and zhang , we discuss generalizations of the strategy reward scheme such that the agents seek to maximize their wealth in a more direct way .
we then examine a disturbing feature whereby such reward schemes , while appearing microscopically acceptable , lead to unrealistic market dynamics ( e.g. instabilities ) . finally , we discuss various mechanisms which are responsible for re - stabilizing the market in reality .
this discussion leads to a ` toolbox ' of processes from which , we believe , successful market models can be constructed in the future . 0.5 in oxford center for computational finance working paper : _ occf/010702 _ |
a convenient way to characterize disorder - dominated phases is through the statistics of some appropriate `` weights '' . in mean - field models , these weights represent either weights of pure states , as in the replica analysis of the sherrington - kirkpatrick model @xcite , or weights of microscopic configurations , as in the random energy model @xcite or in the directed polymer model on the cayley tree @xcite .
it turns out that in these three cases , the weights statistics is the same as in lvy sums with some index @xmath29 @xcite , where the index @xmath30 depends on the temperature : for instance in the random energy model @xcite or in the directed polymer model on the cayley tree @xcite , it is simply @xmath31 . in @xcite , the corresponding probability distributions of the weights were found to exhibit characteristic singularities at some integer inverses .
similar derrida - flyvbjerg singularities also occur in many other contexts , such as randomly broken objects @xcite , in population genetics @xcite , in random walk excursions or loops @xcite . for disordered systems in finite dimensions
, it seems appropriate to consider the weights associated to a local degree of freedom , to characterize to what extent it is frozen . to the best of our knowledge
, this idea has first been introduced to characterize the freezing of a folded polymer with random self - interactions @xcite .
it was then used in the context of secondary structures of random rna to analyse the fraction of frozen pairs between degenerate ground states @xcite , and to characterize the freezing transition @xcite . in this paper , we study the statistics of the weights @xmath6 of the end - point of a directed polymer in a random potentiel @xcite .
we focus here on the low - temperature @xmath0 disorder - dominated phase both in dimension @xmath1 ( where @xmath2 ) and @xmath3 ( where @xmath4 ) , since we have studied elsewhere @xcite the weights statistics at criticality in @xmath32 , where multifractal behavior occurs .
we are not aware of previous studies on these weights in the physics literature . on the contrary , in the mathematical litterature
, the weight of the favourite site has been considered as a localization criterion @xcite , and a more detailed description of end - point weights was then given via the notion of @xmath33-atoms @xcite .
the paper is organized as follows .
the model and observables are introduced in section [ model ] .
we then present a detailed study of the weights of the end - point of the directed polymer both in dimensions @xmath1 and @xmath3 . for clarity ,
the statistical properties of the weights alone , independently of the distances involved are described in section [ poids ] , whereas the study of spatial properties is postponed to section [ spatial ] .
we summarize our results in section [ conclusion ] .
in this paper , we present numerical results for the random bond version of the model defined by the recursion relation on a cubic lattice in @xmath34 and @xmath32 @xmath35 the bond energies @xmath36 are random independent variables drawn from the gaussian distribution @xmath37 in this paper , we consider the following boundary conditions .
the first monomer is fixed at @xmath38 , i.e. the initial condition of the recurrence of eq .
( [ dp1 ] ) reads @xmath39 the last monomer is free , i.e. the full partition function of the polymer of length @xmath5 is then obtained by summing over all possible positions @xmath40 at @xmath41 @xmath42 this model has attracted a lot of attention because it is directly related to non - equilibrium properties of growth models @xcite . within the field of disordered systems , it is also very interesting on its own because it represents a ` baby - spin - glass ' model @xcite . at low temperature
, there exists a disorder dominated phase , where the order parameter is an ` overlap ' . in finite dimensions ,
a scaling droplet theory was proposed @xcite , in direct correspondence with the droplet theory of spin - glasses @xcite , whereas in the mean - field version of the model on the cayley , a freezing transition very similar to the one occurring in the random energy model was found @xcite .
the phase diagram as a function of space dimension @xmath43 is the following @xcite . in dimension
@xmath44 , there is no free phase , i.e. any initial disorder drives the polymer into the strong disorder phase , whereas for @xmath45 , there exists a phase transition between the low temperature disorder dominated phase and a free phase at high temperature @xcite . in the following ,
we will focus on the statistical properties of the weights @xmath46 normalized to ( eq . [ ztot ] ) @xmath47 the numerical results given below have been obtained using polymers of various lengths @xmath5 , with corresponding numbers @xmath48 of disordered samples with the values @xmath49 in @xmath34 , and the values @xmath50 in @xmath32 . in the following , @xmath51 denotes the average of @xmath52 over the disorder samples . in analogy with the weight statistics in lvy sums and in the random energy model @xcite , we have numerically computed the probability distribution @xmath7 of the maximal weight ( eq . [ defw ] ) @xmath53 as well as the probability distribution @xmath54 of the second maximal weight .
another useful way to characterize the statistical properties of the weights @xcite is to consider the moments of arbitrary order @xmath27 @xmath55 which represents the probability that the last monomer of the polymer of length @xmath5 is at the same point in @xmath27 different thermal configurations of the same disordered sample .
we have measured the probability @xmath9 of the parameter @xmath56 as well as the moments @xmath57 for @xmath58 .
finally , we have also computed the weights density @xmath59 giving rise to the moments @xmath60 the normalization condition for the density @xmath61 is @xmath62 in the following , we will also present histograms of the associated entropy @xmath63
: probability distributions @xmath7 and @xmath54 of the largest and second largest weight seen by the last monomer ( see eq [ wmax ] ) ( a ) at @xmath64 ( @xmath13 ) for @xmath65 : the characteristic derrida - flyvbjerg singularities at @xmath66 and @xmath67 are clearly visible .
( b ) at @xmath68 ( @xmath69 ) for @xmath70 : the distribution @xmath7 does not reach @xmath66 anymore , and the distribution @xmath54 does not reach @xmath67 anymore , title="fig:",height=226 ] : probability distributions @xmath7 and @xmath54 of the largest and second largest weight seen by the last monomer ( see eq [ wmax ] ) ( a ) at @xmath64 ( @xmath13 ) for @xmath65 : the characteristic derrida - flyvbjerg singularities at @xmath66 and @xmath67 are clearly visible .
( b ) at @xmath68 ( @xmath69 ) for @xmath70 : the distribution @xmath7 does not reach @xmath66 anymore , and the distribution @xmath54 does not reach @xmath67 anymore , title="fig:",height=226 ] : probability distribution @xmath7 of the largest weight seen by the last monomer ( see eq [ wmax ] ) ( a ) at @xmath71 ( @xmath72 ) , @xmath73 ( @xmath74 ) , @xmath75 ( @xmath76 ) for @xmath77 . (
b ) at @xmath78 for @xmath77 , title="fig:",height=226 ] : probability distribution @xmath7 of the largest weight seen by the last monomer ( see eq [ wmax ] ) ( a ) at @xmath71 ( @xmath72 ) , @xmath73 ( @xmath74 ) , @xmath75 ( @xmath76 ) for @xmath77 .
( b ) at @xmath78 for @xmath77 , title="fig:",height=226 ] the probability distributions @xmath7 and @xmath54 of the largest ( eq . [ wmax ] ) and second largest weights of the last monomer in dimension @xmath1 are shown on fig .
[ fig1dw ] for two temperatures .
these curves show that there exists a temperature @xmath79 such that \(i ) for @xmath13 ( see fig .
[ fig1dw ] a ) the distribution @xmath7 reaches the point @xmath80 with a singularity parametrized by a temperature - dependent exponent @xmath81 beyond this main singularity , @xmath7 also present characteristic derrida - flyvbjerg singularities at @xmath82 @xcite : in particular , the singularity of @xmath7 at @xmath67 is clealy visible on fig .
[ fig1dw ] a. similarly , the distribution @xmath54 reaches the point @xmath83 ( see fig .
[ fig1dw ] a ) \(ii ) for @xmath84 ( see fig .
[ fig1dw ] b ) the distribution @xmath7 does not reach @xmath66 anymore , but vanishes at some maximal value @xmath85 @xmath86 with some exponent @xmath87 . similarly , the distribution @xmath54 does not reach the point @xmath88 ( see fig .
[ fig1dw ] b ) this temperature @xmath28 in @xmath1 where @xmath2 also exists in @xmath3 where @xmath89 is finite , as shown on fig .
[ fig3dw ] . on fig .
[ fig3dw ] a , the distribution @xmath7 is shown for three temperatures below @xmath28 with exponents @xmath90 , @xmath91 and @xmath92 . on fig .
[ fig3dw ] b , the distribution @xmath7 is shown for temperature @xmath93 in the region @xmath94 .
note that here we only describe the low - temperature phase @xmath0 and we refer to @xcite for a detailed study of the weights statistics at criticality @xmath95 where multifractal behavior occurs . : probability distribution @xmath96 of the last monomer entropy @xmath97 ( a ) at three temperatures below @xmath28 namely @xmath71 ( @xmath72 ) , @xmath75 ( @xmath74 ) , @xmath98 ( @xmath76 ) for @xmath99 : the derrida - flyvbjerg singularities at @xmath100 , @xmath101 and @xmath102 are clearly visible .
( b ) at @xmath103 for @xmath104 : the histogram does not reach @xmath100 .
, title="fig:",height=226 ] : probability distribution @xmath96 of the last monomer entropy @xmath97 ( a ) at three temperatures below @xmath28 namely @xmath71 ( @xmath72 ) , @xmath75 ( @xmath74 ) , @xmath98 ( @xmath76 ) for @xmath99 : the derrida - flyvbjerg singularities at @xmath100 , @xmath101 and @xmath102 are clearly visible .
( b ) at @xmath103 for @xmath104 : the histogram does not reach @xmath100 .
, title="fig:",height=226 ] probability distribution @xmath96 of the last monomer entropy @xmath97 ( a ) at @xmath105 for @xmath77 : the derrida - flyvbjerg singularities at @xmath100 , @xmath101 and @xmath102 are clearly visible .
( b ) at @xmath78 for @xmath106 : the histogram does not reach @xmath100 .
, title="fig:",height=226 ] probability distribution @xmath96 of the last monomer entropy @xmath97 ( a ) at @xmath105 for @xmath77 : the derrida - flyvbjerg singularities at @xmath100 , @xmath101 and @xmath102 are clearly visible .
( b ) at @xmath78 for @xmath106 : the histogram does not reach @xmath100 .
, title="fig:",height=226 ] the derrida - flyvbjerg singularities found for @xmath13 at @xmath14 for the weights translate into singularities in the histograms of the last - monomer entropy ( eq . [ entropy ] ) at @xmath107 .
in particular , the main singularity of @xmath7 for @xmath80 ( eq . [ p1singw1 ] ) yields a corresponding singularity at @xmath108 @xmath109 for @xmath110 , whereas a gap @xmath111 appears for @xmath69 .
this is shown on fig .
[ fig1ds ] for @xmath34 and on fig .
[ fig3ds ] for @xmath32 . : probability distribution @xmath9 of the parameter @xmath113 ( a ) at @xmath114 for @xmath65 : the derrida - flyvbjerg singularities at @xmath115 , @xmath116 and @xmath117 are clearly visible .
( b ) at @xmath118 for @xmath70 : the histogram does not reach @xmath115 .
, title="fig:",height=226 ] : probability distribution @xmath9 of the parameter @xmath113 ( a ) at @xmath114 for @xmath65 : the derrida - flyvbjerg singularities at @xmath115 , @xmath116 and @xmath117 are clearly visible .
( b ) at @xmath118 for @xmath70 : the histogram does not reach @xmath115 .
, title="fig:",height=226 ] : probability distribution @xmath9 of the parameter @xmath113 ( a ) for @xmath64 ( @xmath119 ) and @xmath73 ( @xmath77 ) : the derrida - flyvbjerg singularities at @xmath115 , @xmath116 and @xmath117 are clearly visible .
( b ) for @xmath120 with @xmath121 , title="fig:",height=226 ] : probability distribution @xmath9 of the parameter @xmath113 ( a ) for @xmath64 ( @xmath119 ) and @xmath73 ( @xmath77 ) : the derrida - flyvbjerg singularities at @xmath115 , @xmath116 and @xmath117 are clearly visible .
( b ) for @xmath120 with @xmath121 , title="fig:",height=226 ] the parameter @xmath112 defined in eq .
[ y2def ] can reach the value @xmath122 only if the maximal weight @xmath123 also reaches @xmath124 . as a consequence , the probability distribution @xmath9 has the same singularity near @xmath125 as in eq .
( [ p1singw1 ] ) @xmath126 for @xmath110 . beyond this main singularity ,
the distribution @xmath9 presents the characteristic derrida - flyvbjerg singularities at @xmath15 as shown on fig .
[ fig1dy2 ] a for @xmath34 and on fig .
[ fig3dy2 ] a for @xmath32 .
again for @xmath69 , a gap appears as shown on fig .
[ fig1dy2 ] b for @xmath34 and on fig .
[ fig3dy2 ] b for @xmath32 . : weight density @xmath127 ( a ) at @xmath128 where @xmath129 for @xmath99 : the weight density @xmath127 diverges at @xmath130 and @xmath80 .
( b ) at @xmath131 for @xmath99 : these curves show the temperature dependent singularity at @xmath66 .
, title="fig:",height=226 ] : weight density @xmath127 ( a ) at @xmath128 where @xmath129 for @xmath99 : the weight density @xmath127 diverges at @xmath130 and @xmath80 .
( b ) at @xmath131 for @xmath99 : these curves show the temperature dependent singularity at @xmath66 .
, title="fig:",height=226 ] : weight density @xmath127 ( a ) for @xmath64 ( @xmath119 ) and @xmath132 ( @xmath77 ) where @xmath129 : the weight density @xmath127 diverges at @xmath130 and @xmath80 .
( b ) for @xmath133 ( @xmath134 ) : these curves show that the gap appears around @xmath135 , title="fig:",height=226 ] : weight density @xmath127 ( a ) for @xmath64 ( @xmath119 ) and @xmath132 ( @xmath77 ) where @xmath129 : the weight density @xmath127 diverges at @xmath130 and @xmath80 .
( b ) for @xmath133 ( @xmath134 ) : these curves show that the gap appears around @xmath135 , title="fig:",height=226 ] the density @xmath127 introduced in eq .
( [ densityfdef ] ) is shown on fig .
[ fig1df ] and fig .
[ fig3df ] for @xmath34 and @xmath32 respectively . by construction , this density coincides with the maximal weight distribution @xmath7 for @xmath136 , with the sum ( @xmath137 ) of the two largest weight distributions for @xmath138 , and so on @xcite . as a consequence , @xmath127 has the same singularity near @xmath80 as @xmath7 ( eq . [ p1singw1 ] ) , and the same gap ( eq . [ p1gapw0 ] ) as long as @xmath139 .
the only other singularity is near @xmath130 where @xmath127 diverges in a non - integrable manner , because in the @xmath140 , there is an infinite number of vanishing weights ( only the product @xmath141 has to be integrable at @xmath142 as a consequence of the normalization condition of eq .
[ normay1 ] ) . : ( a ) decay of the moments @xmath143 of eq .
[ ykdecay ] as a function of @xmath144 for @xmath145 and @xmath146 ( b ) exponent @xmath17 as measured from the slope of the log - log decay in the asymptotic region.,title="fig:",height=226 ] : ( a ) decay of the moments @xmath143 of eq .
[ ykdecay ] as a function of @xmath144 for @xmath145 and @xmath146 ( b ) exponent @xmath17 as measured from the slope of the log - log decay in the asymptotic region.,title="fig:",height=226 ] : ( a ) decay of the moments @xmath143 of eq .
[ ykdecay ] as a function of @xmath144 for @xmath147 and @xmath148 ( b ) exponent @xmath17 as measured from the slope of the log - log decay in the asymptotic region .
, title="fig:",height=226 ] : ( a ) decay of the moments @xmath143 of eq .
[ ykdecay ] as a function of @xmath144 for @xmath147 and @xmath148 ( b ) exponent @xmath17 as measured from the slope of the log - log decay in the asymptotic region .
, title="fig:",height=226 ] for @xmath149 , where @xmath7 and @xmath127 behaves near @xmath80 as in eq .
[ p1singw1 ] , the decay in @xmath27 of the averaged moments @xmath25 ( eq . [ ykdef ] ) follow a power - law of exponent @xmath17 @xmath150 the behavior of @xmath151 for @xmath144 are shown on fig .
[ fig1dmu ] a for @xmath34 and fig .
[ fig3dmu ] a for @xmath32 .
the corresponding exponent @xmath17 are shown on fig .
[ fig1dmu ] b for @xmath34 and fig .
[ fig3dmu ] b for @xmath32 . for @xmath69 where there exists a gap @xmath152 for @xmath7 , the behavior of eq .
[ p1gapw0 ] also applies to @xmath127 as long as @xmath139 ( since @xmath153 for @xmath136 as mentioned above ) and thus the decay is then exponential @xmath154
in the previous section , we have studied in details the statistics of the weights independently of their spatial organization . in this section , we study the statistics of the transverse spatial correlation @xmath155 centered on the preferred position @xmath156 of maximal weight ( eq . [ wmax ] ) .
we first recall the predictions of the droplet scaling analysis @xcite that will be useful to analyse our numerical results . the droplet theory for directed polymers @xcite , is very similar to the droplet theory of spin glasses @xcite .
it is a scaling theory that can be summarized as follows . at very low temperature @xmath157 ,
all observables are governed by the statistics of low energy excitations above the ground state .
an excitation of large length @xmath158 costs a random energy @xmath159 where @xmath160 is a positive random variable distributed with some law @xmath161 having some finite density at the origin @xmath162 .
the exponent @xmath163 is the exponent governing the fluctuation of the energy of the ground state is exactly known in one - dimension @xmath164 @xcite and for the mean - field version on the cayley tree @xmath165 @xcite . in finite dimensions @xmath166 , the exponent @xmath167 has been numerically measured , and we only quote here the results of the most precise study we are aware of @xcite for dimensions @xmath168 : @xmath169 and @xmath170 . from ( [ ground ] ) ,
the probability distribution of large excitations @xmath171 reads within the droplet theory @xmath172 where the factor @xmath173 comes from the notion of independent excitations @xcite .
in particular , its average over the disorder follows the power - law @xmath174 this prediction describes very well the numerical data in the regime @xmath175 in dimensions @xmath176 @xcite .
since correlation functions at large distance are directly related to the probability of large excitations , we already see that the low temperature phase is very non - trivial from the point of view of correlations lengths : the typical exponential decay ( [ rhodroplet ] ) indicates a finite typical correlation length @xmath177 , whereas the averaged power - law behavior ( [ rhoavt0 ] ) means that the averaged correlation length @xmath178 is actually infinite in the whole low temperature phase @xmath179 note that within the droplet theory of spin - glasses @xcite , the correlation length @xmath178 is also infinite in the whole low temperature phase for the same reasons .
according to the droplet scaling theory @xcite the whole low temperature phase @xmath180 is governed by a zero - temperature fixed point .
however , many subtleties arise because the temperature is actually ` dangerously irrelevant ' .
the main conclusions of the droplet analysis @xcite can be summarized as follows .
the scaling ( [ ground ] ) governs the free energy cost of an excitation of length @xmath158 , provided one introduces a longitudinal correlation length @xmath181 to rescale the length @xmath158 @xmath182 here as before , @xmath160 denotes a positive random variable distributed with some law @xmath183 having some finite density at the origin @xmath184 . as a consequence ,
the probability of a droplet of size @xmath185 follows the scaling form @xmath186 in particular , the typical behavior follows an exponential decay with exponent @xmath163 @xmath187 whereas the average over the disorder follows the power - law @xmath188 this average is governed by the rare events having @xmath189 .
a droplet of longitudinal size @xmath158 corresponds a transverse distance @xmath190 , where @xmath191 is the roughness exponent of the low temperature phase @xcite . via the change of variable @xmath190 , the droplet distribution of eq .
[ droplet ] translates into the following scaling form for the correlation at large distance @xmath192 @xcite @xmath193 where the transverse correlation length reads @xmath194^{\zeta}$ ] , i.e. finally @xmath195 as a consequence , the typical behavior follows an exponential decay with exponent @xmath196 @xmath197 whereas the average over the disorder is governed by the rare events and follows the power - law @xmath198 we now describe our numerical data and compare with these predictions . in @xmath34 :
finite - size scaling of ( eq .
[ cavfss ] ) : @xmath200 as a function of @xmath201 for @xmath202 ( a ) for @xmath203 ( b ) for @xmath204 , title="fig:",height=226 ] in @xmath34 : finite - size scaling of ( eq .
[ cavfss ] ) : @xmath200 as a function of @xmath201 for @xmath202 ( a ) for @xmath203 ( b ) for @xmath204 , title="fig:",height=226 ] : disorder averaged correlation @xmath199 in @xmath32 finite - size scaling of ( eq .
[ cavfss ] ) : ( eq . [ cavfss ] ) : @xmath205 as a function of @xmath206 for @xmath64 @xmath207 .,height=226 ] the previous predictions concern the limit @xmath140 . to compare with our numerical data , we now recall the corresponding finite - size behaviors within the droplet theory @xcite . the power - law behavior ( eq .
[ cavlinfty ] ) for the averaged correlation translates into the following scaling form for finite @xmath5 @xmath208 where the scaling function @xmath209 behaves as the following power - law @xmath210 to recover eq .
[ cavlinfty ] as @xmath140 .
we show on fig .
[ fig1dcav ] the finite - size scaling analysis of eq .
[ cavfss ] in @xmath34 for two temperatures , one below and one above @xmath28 . in both cases ,
the agreement with the droplet scaling ansatz is very good , confirming the zero - temperature fixed point picture .
the corresponding finite - size scaling analysis for the disorder averaged correlation in @xmath32 is shown on fig .
[ fig3dcavt0.1 ] .
in @xmath34 for @xmath212 and sizes @xmath213 : ( a ) @xmath214 as a function of @xmath215 ( see eq .
[ ctyplinfty ] ) ( b ) finite - size scaling of the same data : @xmath216 as a function of @xmath217 ( see eq .
[ ctypfss ] ) .
, title="fig:",height=226 ] in @xmath34 for @xmath212 and sizes @xmath213 : ( a ) @xmath214 as a function of @xmath215 ( see eq .
[ ctyplinfty ] ) ( b ) finite - size scaling of the same data : @xmath216 as a function of @xmath217 ( see eq .
[ ctypfss ] ) . , title="fig:",height=226 ] in @xmath32 for @xmath64 and sizes @xmath207 ( a ) @xmath218 as a function of @xmath219 ( see eq .
[ ctyplinfty ] ) ( b ) finite - size scaling of the same data : @xmath220 as a function of @xmath221 ( see eq .
[ ctypfss ] ) .,title="fig:",height=226 ] in @xmath32 for @xmath64 and sizes @xmath207 ( a ) @xmath218 as a function of @xmath219 ( see eq .
[ ctyplinfty ] ) ( b ) finite - size scaling of the same data : @xmath220 as a function of @xmath221 ( see eq .
[ ctypfss ] ) .,title="fig:",height=226 ] the typical correlation of eq .
[ ctyplinfty ] are shown on fig . [ fig1dclogav ] a for @xmath34 and on fig .
[ fig3dclogavt0.1 ] a for @xmath32 respectively : the collapse for small @xmath192 is satisfactory . to take into account the @xmath5 dependent cut - off for large @xmath192
, we have tried the following finite - size scaling form @xmath222 where the scaling function @xmath223 behaves as the following power - law @xmath224 to recover eq .
[ ctyplinfty ] as @xmath140 .
the results for @xmath34 and @xmath32 are shown on fig .
[ fig1dclogav ] b and [ fig3dclogavt0.1 ] b respectively .
: probability distribution @xmath225 of the correlation @xmath226 ( eq . [ corre ] ) for @xmath99 ( a ) for @xmath203 the distribution @xmath225 reaches the maximal value @xmath227 for any distance @xmath192 : here @xmath228 ( b ) probability distribution @xmath229 of the correlation of two neighboring sites for @xmath132 where @xmath129 , @xmath98 where @xmath230 and @xmath231 where the histogram does not reach the maximal possible value @xmath227 .
.,title="fig:",height=226 ] : probability distribution @xmath225 of the correlation @xmath226 ( eq . [ corre ] ) for @xmath99 ( a ) for @xmath203 the distribution @xmath225 reaches the maximal value @xmath227 for any distance @xmath192 : here @xmath228 ( b ) probability distribution @xmath229 of the correlation of two neighboring sites for @xmath132 where @xmath129 , @xmath98 where @xmath230 and @xmath231 where the histogram does not reach the maximal possible value @xmath227 .
.,title="fig:",height=226 ] since typical and disorder averaged correlations are very different , we have also studied in @xmath34 the probability distribution @xmath225 of the correlation @xmath226 ( eq . [ corre ] ) between the preferred position and a site at transverse distance @xmath192 . for @xmath13 ,
the distribution @xmath225 reaches the maximal possible value @xmath227 for any distance @xmath192 as shown on fig .
[ fig1dhistocorre ] a. this maximal value @xmath232 corresponds to the case where the maximal weight and second maximal weight are both of order @xmath233 and are at distance @xmath192 . for large @xmath234 , this corresponds to the droplet excitations .
for @xmath69 , the distribution @xmath225 does not reach the maximal possible value @xmath227 anymore , as shown on fig .
[ fig1dhistocorre ] b.
we have studied the weight statistics in the low - temperature @xmath0 disorder - dominated phase of the directed polymer in a random potentiel in dimension @xmath1 ( where @xmath2 ) and @xmath3 ( where @xmath4 ) . in particular , we have found a temperature @xmath12 with the following properties . for @xmath13 , the histograms of weight observables and of spatial correlation display characteristic derrida - flyvbjerg singularities . in particular , there exists a temperature - dependent exponent @xmath17 that governs the main singularities of @xmath18 , @xmath19 and @xmath235 . as well as the power - law decay of the moments @xmath20 .
the exponent @xmath17 grows from the value @xmath21 up to @xmath22 . for @xmath23 ,
the distribution @xmath7 vanishes at some value @xmath24 , and accordingly the moments @xmath25 decay exponentially as @xmath26 in @xmath27 .
finally , our numerical results concerning typical and averaged correlations are in full agreement with the droplet scaling theory both below and above @xmath28 . together with our previous study on the freezing transition of random rna secondary structures
@xcite , this shows that the weight statistics is an efficient tool to characterize to which extent local degrees of freedom are frozen . moreover , the position of @xmath28 with respect to @xmath236 gives a better understanding of the transition . in the rna case
where @xmath237 , the interpretation is that there exists frozen pairs in the high - temperature phase , but only of finite size @xcite . here ,
in the directed polymer case where @xmath12 , there exists frozen monomers only below @xmath28 , whereas for @xmath23 , the localization occurs on a tube of finite extent @xmath238 .
b. derrida , spin glasses , random boolean networks and simple models of evolution , proceedings of the trieste conference on nonlinear cooperative phenomena in biological systems , august 1997 , p 216 - 226 in non linear cooperative phenomena in biological systems ed l. matsson , world scientific ( 1998 ) .
m. prhofer and h. spohn , physica * a279 * , 342 ( 2000 ) ; m. prhofer and h. spohn , phys .
lett . * 84 * , 4882 ( 2000 ) ; m. prhofer and h. spohn , j. stat . phys .
* 108 * , 1071 ( 2002 ) ; m. prhofer and h. spohn , j. stat . phys . * 115 * , 255 ( 2002 ) . | we consider the low - temperature @xmath0 disorder - dominated phase of the directed polymer in a random potentiel in dimension @xmath1 ( where @xmath2 ) and @xmath3 ( where @xmath4 ) . to characterize the localization properties of the polymer of length @xmath5 , we analyse the statistics of the weights @xmath6 of the last monomer as follows .
we numerically compute the probability distributions @xmath7 of the maximal weight @xmath8 $ ] , the probability distribution @xmath9 of the parameter @xmath10 as well as the average values of the higher order moments @xmath11 .
we find that there exists a temperature @xmath12 such that ( i ) for @xmath13 , the distributions @xmath7 and @xmath9 present the characteristic derrida - flyvbjerg singularities at @xmath14 and @xmath15 for @xmath16 . in particular , there exists a temperature - dependent exponent @xmath17 that governs the main singularities @xmath18 and @xmath19 as well as the power - law decay of the moments @xmath20 .
the exponent @xmath17 grows from the value @xmath21 up to @xmath22 .
( ii ) for @xmath23 , the distribution @xmath7 vanishes at some value @xmath24 , and accordingly the moments @xmath25 decay exponentially as @xmath26 in @xmath27 .
the histograms of spatial correlations also display derrida - flyvbjerg singularities for @xmath13 . both below and above @xmath28 ,
the study of typical and averaged correlations is in full agreement with the droplet scaling theory .
# 1#2 # 1#2 |
in the past few years , substantial progress has been made in optical identification of _ hubble space telescope _
counterparts to _ chandra _ x - ray sources in several globular clusters ( gcs ) .
valuable information was obtained for the population of cataclysmic variables ( cvs ) , chromospherically active binaries and quiescent low - mass x - ray binaries ( qlmxbs ) ( edmonds et al .
2004 , heinke et al .
2003b , haggard et al . 2004 ) .
for the first time we can compare populations of such binaries in gcs and in the galactic field , as well as calculate the rates of formation of these binaries and their population characteristics . in particular , as many as 22 cvs have been already identified in 47 tuc , posing a number of interesting questions : e.g. , their ratio of x - ray flux to optical flux is higher than in field cvs ( edmonds et al .
2003 ) and they do not show dwarf nova outbursts , which are common for field cvs ( shara et al . 1996 ) . also , it has been found that the specific incidence of harder x - ray sources ( primarily cvs ) depends more weakly on density than that of qlmxbs ( heinke et al .
2003a ) .
the present population of close binaries with a white dwarf ( wd ) accretor in gcs has not necessarily evolved from primordial binaries , since it can be influenced by dynamical encounters . in this contribution we study these processes by combining an advanced binary population code and careful treatment of all dynamical interactions with a simplified dynamical cluster background model .
we first consider what mechanisms are likely to produce a cv or an am cvn and then compare our predictions with the results of numerical simulations .
we also discuss how the formation of mass - transfer systems with a wd accretor is different from the formation of these with systems with a neutron star ( ns ) accretor .
there are several ways to destroy a primordial binary in a globular cluster .
if the binary is in a dense region , dynamical encounters play a significant role in its evolution . for instance
, there is a high probability to `` ionize '' a soft binary as a result of a dynamical encounter . in the case of a hard binary
, destruction can happen through a physical collision during the encounter .
additionally , a primordial binary can be destroyed through an evolutionary merger ( most important for hard binaries ) or following to a sn explosion .
overall , we find that , if a typical cluster initially had as many as 100% of its stars in primordial binaries , the binary fraction at an age of 10 - 14 gyr will be only @xmath0 ( ivanova et al .
2005 ) .
the typical formation scenario for cvs in the field ( low density environment ) involves the common envelope ( ce ) evolution .
the minimum initial orbital period for the progenitor binary is about 100 days and the primary mass has to be smaller than 8 @xmath1 . in the core of a gc with core density @xmath2 pc@xmath3 , a binary with such a period will experience a dynamical encounter before its primary leaves the main sequence . the primordial channel for the cv formation could succeed therefore only if this binary entered the dense cluster core after the ce event .
the contribution of the primordial channel depends mainly on the cluster half - mass relaxation time , which regulates how fast binaries will segregate into the central dense region .
the situation is even more striking for the formation of am cvn . in this case
, the standard formation channel requires the occurrence of two ce events , and the primordial binary is expected to be even wider .
there are two main types of dynamical encounters that lead to the dynamical formation of a binary consisting of a main sequence star ( ms ) and a wd : exchange interactions , and physical collisions between a red giant ( rg ) and a ms star .
however , only a fraction of the dynamically formed ms - wd binary systems is capable of becoming a cv
. we will consider first the post - encounter binary evolution of a dynamically formed ms - wd binary ( to find out which post - encounter parameters favor cv formation ) , and then determine which encounters give these specific parameters .
a close ms - wd binary looses its angular momentum through magnetic braking ( mb ) and gravitational wave ( gw ) emission
. the orbital shrinkage due to the synchronization of the ms star with the orbital motion can be neglected : less than a few percent of the total orbital angular momentum is required to spin - up a ms .
a post - exchange hard binary has an average eccentricity @xmath4 , following a thermal distribution , and a post - collision binary has eccentricity @xmath5 ( lombardi et al .
2005 ) . .
@xmath6 is the post - encounter orbital period and @xmath7 is the mass of a ms secondary .
the short - dashed lines show the binary periods for collision times of 1 gyr and 10 gyr .
the dotted lines delineate the binaries that will shrink within 1 gyr due to gravitational wave emission , for different post - exchange eccentricities .
the long - dashed lines delineate the binaries that will shrink within 1 gyr and 10 gyr for the standard magnetic braking prescription , and dash - dotted line for reduced magnetic braking . below the solid line
the binary is in contact .
* right panel : * wd - wd binaries where the primary is a wd of 0.6 @xmath1 .
@xmath6 is the post - encounter orbital period and @xmath8 is the mass of a wd secondary .
the dashed and dotted lines are the same as in the left panel .
the solid line shows binaries that can be formed through a collision between a wd of 0.6 @xmath1 and a rg of 0.8 @xmath1 ( @xmath9 ) .
, title="fig : " ] .
@xmath6 is the post - encounter orbital period and @xmath7 is the mass of a ms secondary .
the short - dashed lines show the binary periods for collision times of 1 gyr and 10 gyr .
the dotted lines delineate the binaries that will shrink within 1 gyr due to gravitational wave emission , for different post - exchange eccentricities .
the long - dashed lines delineate the binaries that will shrink within 1 gyr and 10 gyr for the standard magnetic braking prescription , and dash - dotted line for reduced magnetic braking . below the solid line
the binary is in contact .
* right panel : * wd - wd binaries where the primary is a wd of 0.6 @xmath1 .
@xmath6 is the post - encounter orbital period and @xmath8 is the mass of a wd secondary .
the dashed and dotted lines are the same as in the left panel .
the solid line shows binaries that can be formed through a collision between a wd of 0.6 @xmath1 and a rg of 0.8 @xmath1 ( @xmath9 ) .
, title="fig : " ] depending on the mb prescription the standard mb ( skumanich 1972 , pylyser & savonije 1988 ) or the reduced mb ( ivanova & taam 2003 ) in a dynamically formed eccentric binary with @xmath10 , mb is at most comparable to gw emission ( see fig . 1 ) . in order to become a cv within 1 gyr
, a binary should have a post - encounter period @xmath11 day ( the separation is then @xmath12 ) , or @xmath13 day and very high eccentricity , @xmath14 .
it is not clear which post - exchange binaries will dominate in producing cvs . at first glance , in the encounter with a hard binary ,
the post - exchange separation is comparable to the separation in the pre - encounter binary .
therefore , considering the collision time , the formation of a wider binary through exchange is more likely . on the other hand ,
it is rather unlikely to produce such a very high eccentricity through an exchange . in the case of a ms - wd binary with @xmath15 days and a moderate eccentricity @xmath16 , dynamical interactions that occur after binary formation
can not harden such a binary significantly .
let us consider a binary consisting of a 1 @xmath1 ms and a 0.6 @xmath1 wd .
even if each hardening encounter could reduce the orbital separation by @xmath17 , the hardening of this binary from 10 days to 1 day ( at this period mb starts to be efficient ) will take about 20 gyr .
in contrast to hardening , another mechanism is important : _
eccentricity pumping_. the mean time between successive collisions @xmath18 gyr and therefore a binary can experience many encounters . if the acquired eccentricity @xmath19 , the binary can shrink through gw emission even if its initial period is bigger than 10 days .
as the post - exchange separation in a binary that becomes a cv is comparable to the separation in the pre - encounter binary , the pre - encounter binary has to be very tight .
it is not likely that , in the case of a tight binary ( with moderate eccentricity ) , the pre - encounter binary was a ms - ms binary or a ms - wd binary : an encounter with so close a binary will likely lead to a physical collision rather than to an exchange ( fregeau et al .
this restriction , together with @xmath20 few gyr for pre - collision binaries of @xmath21 day , predicts that most of the post - exchange cvs can be formed in a three - body encounter where the pre - encounter binary had @xmath22 days and the post - encounter binary has high eccentricity @xmath23 ( or this eccentricity was increased in subsequent encounters ) .
no hardening is expected . for binaries formed through ms - rg collisions
, the binary separation can be estimated using the standard ce prescription with @xmath9 ( iben & livio 1993 ) .
the consideration of a rg inner structure through its evolution predicts that only collision of a ms star with a rg with core mass @xmath24 or with a giant at the core helium burning stage with @xmath25 will lead to the formation of a tight enough binary ( see also fig . 1
, right panel ) .
a wd of 0.3 @xmath1 can not be formed unless it evolved through a ce event or a physical collision .
a post - ce binary for this wd is very hard and has @xmath26 gyr , so the number of single @xmath27 wds is negligible .
this restricts the exchange formation channel for cvs with a low - mass wd : all cvs with a low mass wd companion must be formed either through a ce event ( in a primordial binary or in a dynamically formed binary with @xmath28 days ) , or as a result of a physical collision .
let us consider first a hard post - exchange wd - wd binary . as we described previously , the typical eccentricity is @xmath29 , the separation is comparable to the pre - exchange separation ; there is no post - exchange hardening .
the main difference with ms - wd binaries is that post - exchange wd - wd binary periods that will allow a binary to evolve to mass transfer ( mt ) are several times smaller ( see fig . 1 , right panel ) .
the collision time for pre - encounter binaries is about 10 gyr , therefore making the formation an am cvn binary through three - body exchange ( with subsequent gw decay ) rather unlikely .
a variant for this channel is the dynamical formation of a relatively wide ms - wd binary that will subsequently evolve through ce .
the collisional time for these relatively wide ms - wd binaries is less than 1 gyr and therefore a significant fraction of such binaries may participate in some other encounter ( destruction , hardening or eccentricity pumping ) before the ce occurs .
the second important channel is again a physical collision , in this case of a single wd with a rg ( see fig . 1 , where we show possible outcomes of a such a collision ) .
we therefore expect that only a post - ce system can become an am cvn , where the post - ce system could be from a primordial binary , a post - collision binary , or a dynamically formed binary .
to study stellar evolution in a dense system , we use a monte carlo approach which couples one of most advanced binary population synthesis codes currently available ( belczynski et al .
2002 ) , a simple model for the cluster , and a small @xmath30-body integrator for accurate treatment of all relevant dynamical interaction processes ( fregeau et al . 2004 ) .
the complete description of the method is provided in ivanova et al .
here we treated physical collisions with a simple ce prescription with post - collision @xmath31 , though in a more realistic treatment the post - collision binary parameters should depend on the impact parameter ( c.f .
lombardi et al . 2005 ) .
we studied a `` typical '' cluster , starting with @xmath32 stars and initial binary fraction 100% .
this high primordial binary fraction is needed in order to match the observed binary fraction in gc cores today ( ivanova 2005 ) .
the core density @xmath33 , one - dimensional velocity dispersion @xmath34 km / s , escape speed from the cluster @xmath35 km / s , and half - mass relaxation time @xmath36 gyr .
we also considered a 47 tuc - type cluster , characterized by @xmath37 , @xmath38 km / s , @xmath39 and @xmath40 gyr .
we adopt the broken power law imf of kroupa ( 2002 ) for single primaries , a flat mass - ratio distribution for secondaries and the distribution of initial binary periods constant in logarithm between contact and @xmath41 d. the mass of the cluster at 11 gyr is @xmath42 . in fig .
2 we show the main formation channels for cvs that operate in a typical cluster at 11 gyr .
the most important channel for cv formation is through an exchange encounter it provides 40% of cvs , and also 7% of cvs are in binaries that experienced a merger during the last three - body encounter ; 15% of cvs were formed as a result of a physical collision between a rg and a ms star ; in 10% the ce occured in a dynamically formed binary ; and 28% of cvs are provided by the primordial channel . in total , the number of post - ce systems and post - exchange systems is about the same .
typical participants of a successful physical collision ( leading to cv formation ) are a ms star of @xmath43 and a rg of about @xmath44 with a core around @xmath27 or he core burning giant with core mass around @xmath45 .
cvs formed this way are similar to post - ce cvs from primordial binaries , the typical post - exchange period is about 0.2 day . for cvs formed through an exchange , there are two types of successful three - body encounters : ( a ) a single , relatively heavy wd of about @xmath46 and a ms - ms binary with a total mass usually @xmath47 ; ( b ) a single relatively massive ms star of around the turn - off mass and a ms - wd or wd - wd binary . in most cases the wd - ms or wd - wd binary was not a primordial binary , but a dynamically formed binary .
the number of successful encounters between ms star and wd - wd binary is relatively small , and no successful four - body encounter occurred .
a post - exchange binary typically has a wd heavier than a post - ce binary . as a result ,
the distribution of wd masses in cvs is more populated at the high mass end ( see fig . [ hist ] ) compared to the field population , which has two very well distinguished peaks at @xmath48 and @xmath49 @xmath1 and exponentially decreases with wd mass ( willems , priv . comm . ) .
as expected , most cvs with a heavy wd were formed dynamically . in fig .
4 we show the main formation channels for am cvns that occur in a typical cluster . as predicted , the main difference with cv formation is that there are only post - ce channels .
the role of physical collisions for am cvns is slightly more important than for cvs , although this is only for the relative fraction .
the number of cvs formed by physical collisions is about 3 times higher than the number of am cvns formed by physical collisions . at the moment of a physical collision , in 2/3 of cases , the participants are a rg of @xmath50 and a relatively massive ms star with well developed he core .
the am cvns we see are the result of mt that started when the donor was at the late ms stage , stripped down to the he core ( the mass of the he donor in this system is typically about @xmath51 ) , or the system has passed through another ce event . about a third of physical collisions occur between a rg and heavy wd .
overall , this channel provides am cvns with accretor masses at present ( 11 gyr ) from @xmath52 to @xmath53 , compared to the primordial channel where masses of accretors are mainly distributed between @xmath52 and @xmath54 ( fig . 3 ; see also nelemans et al .
2001 ) . the exchange formation channel for am cvns involves an exchange encounter and a ce event .
a single wd acquires a relatively massive companion that becomes later a rg or , with a 0.6 - 0.8 @xmath1 ms star , a second encounter occurs , with either a physical collision between ms stars or exchange with a more massive ms . the total numbers of cvs and am cvns present in a typical cluster are not very different from these in field population they are comparable ( per whole cluster population ) and only 2 - 3 times larger in the core than in the field ( per unit mass ) .
dynamical formation is responsible only for 60%-70% of cvs in the core , but there is also a fraction of cvs that never entered the core . the gc core density variation therefore does not play a significant role , in contrast to the case of ns binaries , where almost systems were formed dynamically ( ivanova et al .
2004 ) and the numbers have a direct strong dependence on the cluster collision rate ( pooley et al . 2003 ) . we expect to have about 1 cv per @xmath55 in the core of a typical cluster and about 1 cv per @xmath56 in a 47 tuc type cluster ( about 100 cvs in the core of 47 tuc , in quite reasonable agreement with observations , edmonds et al .
the number of am cvns is typically 3 - 4 times smaller than the number of cvs .
one of the remaining questions is why cvs in gcs do not show nova outbursts .
there are many differences between the binary properties of gc and field cv populations . from our simulations
, we see that an accreting wd in a gc cv is typically more massive than in field , and the ms donor is less massive .
a cv therefore is characterized by a smaller mt rate , @xmath57 , which is much smaller than the typical value for dwarf novae ( @xmath58 ) .
nevertheless , we should mention this this value is large enough to provide @xmath59 , which is about the maximum observed luminosity for cvs in , e.g. , 47 tuc .
small mt rates also possibly explain why the optical fluxes from these objects are unexpectedly low in uv ( edmonds et al .
according to the disk instability model which is currently used to describe dwarf novae cycles ( frank , king & raine 2002 ) , the condition for instability to occur is described in terms of a critical mt rate .
for instance , there is a critical accretion rate below which a cv disk is cold and stable .
the accreting wd did not necessarily evolve through ce .
this possibly results in a higher magnetic field wd . a higher than usual magnetic field could also help suppress disk instabilities even further . | the population of compact binaries in dense stellar systems is affected strongly by frequent dynamical interactions between stars and their interplay with the stellar evolution . in this contribution , we consider these effects on binaries with a white dwarf accretor , in particular cataclysmic variables and am cvns . we examine which processes
can successfully lead to the creation of such x - ray binaries . using numerical simulations ,
we identify predominant formation channels and predict the expected numbers of detectable systems .
we discuss also why the distribution of cataclysmic variables has a weaker dependence upon the cluster density than the distribution of quiescent low - mass x - ray binaries and why dwarf nova outbursts may not occur among globular cluster cataclysmic variables .
address = department of physics and astronomy , 2145 sheridan rd , evanston , il 60208 , usa address = department of physics and astronomy , 2145 sheridan rd , evanston , il 60208 , usa |
quantum cryptography key distribution ( qckd ) is at the moment the most advanced and challenging application of quantum information .
qckd offers the possibility that two remote parties , sender and receiver ( conventionally called alice and bob ) , can exchange a secret random key , called sifted key ( string of qubits ) , to implement a secure encryption / decryption algorithm based on a shared secret key , without the need that the two parties meet . in practical qckd ,
alice and bob use a quantum channel , along which sequences of signals are either sent or measured at random between different bases of orthogonal quantum states .
alice can play the role of either setting randomly the polarization basis of photons and sending them to bob ( faint laser pulses as photon source ) , or measuring photons randomly in any one of the selected bases ( entangled photon source ) .
bob , randomly and independently from alice , measures in one of the bases .
the sifted key consists of the subset of measurements performed when alice s and bob s bases are in an agreed configuration according to the protocol used , obtaining at this point a deterministic outcome whose security relies on the laws of quantum physics , for they previously agreed upon the correspondence between counting a photon in a specific state and the bit values 0 or 1 .
in contrast , the security of conventional cryptography relies upon the unproven difficulty in factorizing large numbers into prime numbers by a conventional algorithm .
we note that there is no guarantee that such an algorithm does not exist . the underlying feature of qckd
, namely the reliance of the security of the distributed secret key on the laws of quantum physics , gives it an advantage over the public key cryptography .
in other words , the uncertainty principle prohibits one from gaining information from a quantum channel without disturbing it .
basically , qckd is founded on the principle that when a third party ( eve ) performs a measurement on a qubit exchanged , she induces a perturbation , yielding errors in the bit sequence transmitted , revealing her presence .
any attempt by eve to obtain information about the key leads to a nonzero error rate in the generated sifted key .
nevertheless this last claim must be somewhat softened because of practical realization of quantum channels @xcite .
unfortunately , in practical systems , errors also happen because of experimental leakage , like losses in optics , detection , electronics and noise .
also , even when no eavesdropper is disturbing the bit exchange , there will be errors in the transmission and alice s and bob s strings will not coincide perfectly .
thus in practice there is no way to distinguish an eavesdropper attack from experimental imperfections , making it necessary to establish an upper bound on tolerable experimental imperfections in the realization of the quantum channels to implement an error correction procedure .
following the first proposal by bennett and brassard and later the ekert protocol invoking entangled states @xcite , various systems of qckd have been implemented and tested by groups around the world .
recently some research groups @xcite performed the first qckd experiments based on polarization - entangled photon pairs , and brassard _
et al_. @xcite proved theoretically that qckd schemes based on spontaneous parametric down - conversion ( spdc ) offer enhanced performance , mostly in terms of security , compared to qckd based on weak coherent pulses .
entangled photons , generated by spdc in nonlinear crystals , have proved largely successful for quantum optical communication @xcite and quantum radiometry @xcite .
furthermore , basic experimental tests of the foundation of quantum mechanics had been performed by exploiting the entanglement of this source @xcite . however , quantum noise in a spdc quantum state may significantly limit performance of the proposed quantum optical communication and information technologies . in this paper
we provide a general model for an _ a priori _ evaluation of some crucial parameters of a general qckd scheme based on polarization entangled photons .
we basically adopt the formalism of quantum operations @xcite to describe the dynamics of an open quantum system subject either to the interactions with the environment or to a quantum device performing a measurement on it .
these unwanted or induced interactions show up as noise in quantum - information processing systems , degrading their ideal performance . exploiting the quantum - operation formalism we present a model to quantify precisely both quantum and ultimately statistical noise in quantum - information experiments performed using an entangled photon source . in section
ii we consider quantum noise in a lossless measurement system where noise is due to the coupling between the polarization mode of the source with the polarizing beam splitter ports . in section iii
we discuss the case of a lossy system , where noise is induced by detection deficiencies , such as losses of correlated photons from the presence of optical elements , non - ideal detectors , and electronic devices in either channel , and detector dark - counts .
the model concludes with the calculation of an overall probability of total coincidence counts , including an imperfect time - correlation measurement , ultimately yielding an estimate of accidental coincidences ( section iv ) . this result is used in the calculation of the quantum bit error rate ( qber ) for qckd protocols , i.e. , bb84 and the two variants of the ekert s protocol , based on chsh and wigner s inequalities respectively ( section v ) .
previsions are also presented about the sifted and the corrected key , the quantum - bit error rate ( qber ) before and after a standard error correction procedure .
finally we evaluate the performance of security criteria for ekert s protocols based on both the chsh inequality and wigner s inequality ( section vi ) .
in this section we consider a real either non - maximally entangled or partially mixed state resulting from both imperfect entangled state generation by spdc and imperfect polarization state selection by real polarizing beam splitters ( pbss ) . in fig .
1 we depict the typical scheme for quantum cryptography key distribution as implemented using entangled photons generated by spdc .
because of a nonlinear interaction in a @xmath0 crystal , some pump photons ( angular frequency @xmath1 ) spontaneously split into a lower - frequency pair of photons , historically called signal and idler , perfectly correlated in all aspects of their state ( direction , energy , polarization , under the constraints of conservation of energy and wavevector momentum , otherwise known as phase - matching)@xmath2 these entangled states show perfect correlation for polarization measured along orthogonal but arbitrary axes .
the qckd performed by pure entangled states relies on the realization of two quantum correlated optical channels .
these channels yield single - photon polarization states , such that whenever alice performs a polarization measurement on a photon of the pair , automatically the other photon is projected in a defined polarization state , i.e. , alice plays the role of triggering bob s measurement .
in actuality , the light field emerging from the output of the nonlinear crystal is a polarization - entangled multimode state .
however , it can be described as a polarization entangled two - photon state in only two effective modes ( one for channel @xmath3 and the other for channel @xmath4 ) , as signal and idler pairs can be easily emitted non - collinearly with the pump by proper phase - matching rules @xcite .
this scheme eventually is exploited in quantum information applications by using one channel as the trigger or reference ( @xmath3 ) and the other channel as the probe ( @xmath4 ) . according to fig . 1
, we denote alice s detector apparatus to be the trigger and bob s to be the probe .
alice and bob detection apparatus consist of polarization - analyzer systems ( pas ) for proper single - beam polarization rotation , polarizing beam splitters ( pbs ) , photon detectors ( @xmath5 ) , data storage systems ( computers ) and synchronization systems .
let us consider in the following type ii spdc entangled states @xcite , where the output two - photon states are a quantum superposition of orthogonally polarized photons , i.e. the singlet state @xcite : @xmath6 practically , pure entanglement may not be achieved because of imperfect source generation and because of incomplete entangled photon collection . according to refs .
@xcite an uncompensated for coherence loss , induced in the state by the coupling between polarization and frequency modes because of a birefringent environment , may produce a partially mixed or non - maximally entangled state .
also the collection of the same number of entangled photons on both channels is unlikely , mostly due to the imperfect positioning of the detection systems along the true directions of entangled photons on the spdc cones .
two complex variables , @xmath7 ( @xmath8 ) and @xmath9 , characterize the imperfect compensation of dephasing and decoherence in the crystal and the misalignment in collecting entangled photons in the optical paths , respectevely .
the net result is a non - maximally entangled , or partially mixed state , written as @xmath10 .\end{aligned}\ ] ] we analyze now the quantum noise introduced via an imperfect polarization state selection by the pbss depicted in fig . 1 . here
, the entangled photons are detected accordingly to their polarizations by using imperfect polarizing beam splitters ( pbss ) on both arms and perfect single - photon detectors .
the pbss project photons onto a polarization basis @xmath11 while the polarization analyzer systems ( pass ) induce the transformations represented by the unitary operators @xmath12 in each arm ( @xmath13 ) , according to @xmath14 in the case of ideal pbss , channel @xmath15 transmits state @xmath16 ( @xmath17 ) and reflects state @xmath18 ( @xmath19 ) , i.e. , there is a perfect coupling between the output ports of the pbss and the projections of the photon polarization state . in the approach so far adopted in the literature a perfect coupling is always assumed because the measurement process is considered as a projection on polarization states @xmath20 @xmath18,@xmath21 @xmath22 thus assuming detectors @xmath23 are sensitive to polarization@xmath2 here we consider a further noise effect induced by the presence of real pbss , where a small part of the photons projected onto @xmath24 @xmath25 are erroneously transmitted , and some photons projected in the state @xmath26 @xmath27 are erroneously reflected ( generally less than 2 % and 5 % , respectively ) , as shown in fig .
2 . for this purpose , we extend the hilbert space to describe these photon states as @xmath28 where @xmath29 represents the photon crossing the output port of the pbs towards detector @xmath30 and @xmath31 represents the photon crossing the input port of the pbs , with @xmath32 .
@xmath33 is the transmittance of photons in the @xmath34 polarization state , and @xmath35 @xmath36 is the corresponding reflectance , whose phase relation is @xmath37 .
analogously , @xmath38 and @xmath39 are the transmittance and reflectance of photons in the @xmath40 state .
so far , we have considered only lossless pbss .
the effect of photon losses due to all optical devices used in the channels are treated in sections iii and iv . polarized state should be reflected , but some of these are wrongly transmitted
. moreover all photons projected in @xmath41 polarized state should be transmitted , but some of these are erroneously reflected .
, width=8,height=5 ] we define the input density matrix for the pbs ports as @xmath42 and the total input density matrix @xmath43 of the photon system as @xmath44 the formalism of quantum operation is the most suitable to describe the evolution of a quantum system coupled with another quantum system or with the environment , as well as the evolution of a quantum system subject to measurement @xcite . in this context , we consider the set of non - trace - preserving quantum operations @xmath45 defined as @xmath46 describing the process of detection of the photon pair by the detectors @xmath47 and @xmath48 ( @xmath49 ) . in this expression ,
the unitary operator @xmath50 describes the action of the pass and the unitary transformation @xmath51 describes the coupling between the photon pair polarization state and the pbs ports .
the explicit form of @xmath51 is deduced by eq.s ( [ eq5 ] ) , and calculations are reported in appendix a. because the operators @xmath52 and @xmath53 independently act on the corresponding subspaces @xmath3 and @xmath4 of the hilbert space of polarization and induce linear transformation , they are ineffective on the hilbert space of the pbs ports .
thus , @xmath54 is subject to a global transformation written as an unitary operator @xmath55 where @xmath56 @xmath57 is the projector representing the detection process by the two detectors @xmath58 and @xmath59 the probability of detection of the photon pair by the detectors @xmath47 and @xmath48 is @xmath60 = \notag \\ & = & \text{tr}\left ( \widehat{p}_{x_{a}y_{b}}\widehat{u}_{\psi \text{-pbs}}% \widehat{s}_{\psi } \widehat{\rho } ^{\text{in}}\widehat{s}_{\psi } ^{\dagger } % \widehat{u}_{\psi \text{-pbs}}^{\dagger } \widehat{p}_{x_{a}y_{b}}^{\dagger } \right ) .
\label{eqste}\end{aligned}\ ] ] @xmath61 is trace - preserving because the probabilities of the distinct outcomes sum to one , i.e. , tr@xmath62 = \sum_{x_{a}y_{b}}p(x_{a},y_{b})=1 $ ] for all possible input @xmath63
in the following we consider the noise contribution to the photon counts because of an imperfect collection of photons and a noisy and lossy detection system . for the experimental setup in fig . 1 , we calculate the total probability @xmath64 of @xmath65 counts by any detector @xmath30 by separately calculating the probabilities of counts associated with correlated photons ( @xmath66 ) , with uncorrelated photons ( @xmath67 ) , and with detector dark counts ( @xmath68 ) . to describe the counting process we adopt the formalism of quantum operations , where we consider a general density matrix representing photons on a channel @xmath69 in term of number of photons , i.e. @xmath70 as @xmath71 the evolution of the system @xmath72
is evaluated according to the formalism of quantum operations . in this way
, we define the set of non - trace - preserving quantum operations as @xmath73 as @xmath74 which describes the detection of @xmath75 photons by the system @xmath76 . in this expression the unitary operator @xmath77 represents the interaction between the quantum system q@xmath78 of photons in the channel @xmath79 in the initial state @xmath80 and the lossy and noisy environment e@xmath81 in the initial state @xmath82 .
the action of @xmath83 on the state `` number of photons '' is @xmath84 where @xmath85 is the probability of measuring @xmath75 photons out of @xmath65 present in the channel @xmath79 because of losses .
@xmath86 are the measurement operators .
thus , the probability of measuring @xmath75 counts by the detection system @xmath76 is @xmath87 = \sum_{n = m}^{\infty } \rho _ { \,nn}^{\nu } \left| d^{\mu } ( m , n)\right| ^{2 } \label{pisp}\ ] ] where @xmath88 is the probability of having @xmath65 photons on the channel @xmath69 . in the following we concentrate primarily on the measurement of correlated photons by a lossy detector , @xmath30 .
we define the density matrix of the number of photon pairs , @xmath89 , as @xmath90 analogous to eq .
( [ rhopi ] ) , we write a density matrix of single photons of the pairs ( sp ) along channel @xmath15 : @xmath91 the counting of correlated photons on channel @xmath15 by the detector @xmath30 is described by mapping @xmath92 on the set of non - trace - preserving quantum operation @xmath93 .
the explicit form of @xmath94 is deduced by analogy to eq .
( [ exz ] ) , by replacing the interaction unitary operator with @xmath95 and the measurement operator with @xmath96 , given that q@xmath97 is the quantum system of single photons of the pair on the channel @xmath15 and e@xmath98 is the lossy and noisy environment in the initial state @xmath99 .
the @xmath100 action on the state @xmath101 is completely described by means of coefficients @xmath102 in complete analogy with eq .
( [ uxz ] ) , while we have @xmath103 .
the probability of @xmath75 counts by the detector @xmath30 corresponding to correlated photons becomes @xmath104 the probability of @xmath65 pairs is given by @xmath105 @xmath106 being the time of measurement and @xmath107 the mean rate of photon pairs in the alice and bob channels @xcite .
the terms @xmath108 ^{m } \\ & & \left [ 1-\xi _ { x_{a}}\sum_{y_{b}}p(x_{a},y_{b})\right ] ^{n - m } , \\
\left| d^{sp , y_{b}}(m , n)\right| ^{2 } & = & \left ( \begin{array}{c } n \\ m \end{array } \right ) \left [ \xi _ { y_{b}}\sum_{x_{a}}p(x_{a},y_{b})\right ] ^{m } \\ & & \left [ 1-\xi _ { y_{b}}\sum_{x_{a}}p(x_{a},y_{b})\right ] ^{n - m}.\end{aligned}\ ] ] are the probabilities that only @xmath75 out of @xmath65 photons in the channel @xmath3 ( @xmath4 ) are counted by the detector @xmath47 ( @xmath48 ) . losses due to electronics ( @xmath109 ) , detection efficiencies ( @xmath110 ) , as well as optical losses ( @xmath111 ) are summed up in the term @xmath112 @xcite@xmath113 while we refer to the appendix b for the analysis of dead time in this context .
the term @xmath114 incorporates all losses in the alice and bob optical path , such as from crystals , filters , lenses , pbss , pass and fibers .
the terms @xmath115 are the probability that each photon of the pair may be counted randomly by any arbitrary detector ( eq . [ eqste ] ) .
probability @xmath116 is derived according to eq .
( [ pisp ] ) , giving @xmath117 with mean count rates given by @xmath118 here we consider counts from any detector @xmath30 from stray light , uncorrelated photons , and dark counts eventually contributing to noise in the distributed key . the density matrix associated with stray light and uncorrelated photons is @xmath119 by pursuing the same formalism as before the detection of uncorrelated photons by the detector @xmath30 is described by means of the set of non - trace - preserving quantum operation @xmath120 the map @xmath121 follows in analogy with eq .
( [ exz ] ) .
the unitary operator @xmath122 describes the interaction between the quantum system , q@xmath123 , of uncorrelated photons on the channel @xmath30 and the lossy environment , e@xmath123 , in the initial state @xmath124 .
the measurement operator is @xmath125 the action of @xmath126 on the state @xmath127 follows from eq .
( [ uxz ] ) with the decomposition coefficients , @xmath128 .
thus , the probability of measuring @xmath75 counts of uncorrelated photons by the detector @xmath30 is @xmath129 where @xmath130 is the probability of @xmath65 uncorrelated photons in the channel @xmath30 . according to refs .
@xcite , we assume that we have @xmath131 where @xmath132 is the mean rate of uncorrelated photons . the term @xmath133 is the probability of @xmath75 out of @xmath65 uncorrelated photons counted by the detector @xmath30 .
the @xmath67 derived accordingly from eq .
( [ piu ] ) is @xmath134 the main source of noise in detectors is due to dark counts , whose distribution is regarded merely from a statistical point of view as the probability of @xmath65 dark counts @xmath135 with the mean dark - count rate being @xmath136 .
as real counters can not distinguish among counts due to correlated photons , counts due to uncorrelated photons , and dark counts , the total probability of measuring @xmath137 counts by detector @xmath30 is calculated according to @xcite , @xmath138 giving @xmath139 where the mean rate of total counts measured by the detector @xmath30 is @xmath140
we build up a model for the probability @xmath141 of measuring @xmath65 coincidences by a pair of detectors @xmath47 and @xmath142 in order to estimate crucial quantities of a typical qckd experiment , such as the sifted key and the qber before and after the error correction procedure , whenever different protocols are applied .
we distinguish between the probability distribution of true coincidences ( @xmath143 due to correlated photons ) and the probability distribution of accidental coincidences ( @xmath144 , because of imperfections in the detection electronics ) .
we consider the density matrix in terms of counted pair states ( eq . ( [ rhopi ] ) ) , and we describe its evolution exploiting the formalism of quantum operations as described in section iv by defining another set of non - trace - preserving quantum operations @xmath145 @xmath146 describes the measurement of @xmath75 coincidences originated by the detection of the two photons of a pair by the detectors @xmath47 , @xmath142 .
its explicit expression is found from eq .
( [ exz ] ) , except for the interaction between the quantum system q@xmath147 of photon pairs in the initial state @xmath89 and the lossy and noisy environment e@xmath148 in the initial state @xmath149 represented by the unitary operator @xmath150 and the measurement operator @xmath151 the action of @xmath152 on the state `` number of photon pairs '' is @xmath153 thus , the probability of measuring @xmath75 true coincidences corresponding to photon pairs by the pair of detectors @xmath154 @xmath48 is @xmath155 realizing that a true coincidence may occur only if both photons of the pair are not lost , we emphasize that the terms @xmath156 are the probabilities that only @xmath75 pairs are detected as coincidences by the pair of detectors @xmath47 , @xmath48 when @xmath157 photons are present in the alice s and bob s channels .
it is straightforward to deduce the explicit form of @xmath158 as @xmath159 ^{m } \\ & & \left [ 1-\xi _ { x_{a}}\xi _ { y_{b}}p(x_{a},y_{b})\right ] ^{n - m}.\end{aligned}\ ] ] probability @xmath160 is derived according to eq .
( [ pipi ] ) , obtaining @xmath161 where @xmath162 is the mean rate of true coincidences seen by the pair of detectors @xmath47 and @xmath142 .
the presence of the temporal coincidence window@xmath163 during which coincidences are measured , modifies the mean total coincidence counts , thus forcing one to distinguish between true and accidental coincidence statistics .
we assume that true coincidences occur in the middle of the coincidence temporal window .
then we deduce the probability distribution of accidental coincidences and finally the probability distribution of total coincidences@xmath164 accounting for true and accidental coincidences , assuming @xmath165 , where @xmath166is the dead time in the @xmath15 channel according to appendix b. we regard @xmath167 as the probability distribution of photons counted by the detector @xmath30 that may contribute to accidental coincidences in the time interval @xmath168 . by observing that the probability distributions @xmath64 and @xmath169 are poisson
, it is simple to demonstrate that we have @xmath170 where @xmath171 , and @xmath172 are the mean count rates possibly contributing to accidental coincidences from the detectors @xmath47 and @xmath48 , respectively .
let us denote by @xmath173 the probability that at least one photon in @xmath48 is counted in the coincidence window @xmath174 @xmath175 because detectors @xmath47 are here considered as triggers .
the term in eq .
( [ eqq ] ) is intended to account for the contribution of single detectors @xmath48 . because detectors @xmath48 are statistically independent , the probability that both detectors count a photon producing an accidental coincidence is @xmath176 .
the final probability of accidental counts from detector @xmath48 is obtained by subtracting half the probability that both detectors in bob s channel count an accidental photon , in formula @xmath177 according to @xcite , we calculate the probability distribution of accidental coincidences in the time measurement @xmath178 by applying the discrete convolution between the poisson distribution of `` triggering '' counts and the binomial distribution with parameter @xmath179 @xmath180 with @xmath181@xmath182 giving @xmath183 with @xmath184 lastly the probability distribution of total coincidence counts @xmath185 is obtained by @xmath186 where the mean rate of total coincidence measured by an arbitrary pair of @xmath47 and @xmath142 detectors is @xmath187 to characterize a particular qckd procedure we embody the effect of the transformations @xmath52 and @xmath53 on photon polarization by the rotation matrices @xmath188 we rewrite eq .
( [ lambdac ] ) in terms of the rotation angles @xmath189 induced by transformations @xmath190 and @xmath191 on the polarization state of photons , by replacing @xmath192 with @xmath193 whose complete expression is in appendix a@xmath2
more specifically , the calculated mean coincident counts are made explicit in terms of angular settings @xmath194 and @xmath195 as @xmath196
to characterize a particular qckd procedure and to assess its advantages , we evaluate particular quantities such as the qber and the sifted key for different types of qckd protocols so far experimentally implemented , i.e. bb84 protocol and ekert s protocols based on chsh and wigner s inequalities , respectively .
the qber is a parameter for describing the signal quality in the transmission of the sifted key , defined as the relative frequency of errors induced by accidental coincidences , i.e. the number of errors divided by the total size of the cryptographic sifted key ( @xmath197 ) @xcite . in other words ,
the qber is given by total coincidence provided by those detectors `` wrongly '' firing in coincidence according to the chosen protocol .
in fact the protocol establishes which pair of detectors should fire to contribute to the key .
here we examine the bb84 protocol variant proposed for entangled states in ref .
recall that alice and bob measure photons randomly and independently between two bases of orthogonal quantum states .
one basis corresponds to horizontal and vertical linear polarization ( @xmath198 ) , while the other to linear polarizations rotated by 45@xmath199 ( @xmath200 ) .
only half of the photon pairs can contribute to the sifted key , as only the subset of measurements performed with the two analyzers in the same basis contributes .
the sifted key is given by @xmath201 t , \notag\end{aligned}\ ] ] where @xmath202 is the probability to measure in the right basis ( @xmath203 and @xmath204 ) , while @xmath205 is the probability to measure in a particular analyzer setting ( @xmath206 or @xmath207 ) .
all detectors contribute to the sifted key @xmath197 , but only coincidences between @xmath208 and @xmath209 correspond to the expected anticorrelation when measurements are performed in the same basis ( @xmath210 ) , while qber contributions come from the coincidences between detectors @xmath211 and @xmath212 ( as it is clear from eq.s ( [ a1 ] ) in appendix a ) .
therefore the qber@xmath213 explicit formula is @xmath214 t}{4 k_{\text{bb84}}(\theta _ { a})}. \label{qber1}\ ] ] to test the behavior of qber , we simulate a realistic experiment with parameters @xmath215 ( quantum efficiency of the four detectors)@xmath216 @xmath217 ( transmittance of the four channels)@xmath216 @xmath218 s@xmath219 ( dark count rate of the four detectors ) , @xmath220 ns ( total dead time of the alice s and bob s detection systems ) , @xmath221 and @xmath222 ns . the entanglement parameters are @xmath223 and @xmath224 the correlation level in the alice channel is @xmath225 ( @xmath226 ) , and the correlated photon rate is @xmath227khz . and @xmath228,width=8,height=5 ] in fig .
3 we show the dependence of qber versus the optical properties of real pbs , i.e. the transmittance @xmath229 and @xmath230 ( @xmath13 ) for the states @xmath231 and @xmath232 respectively .
results show how strongly the qber can be affected by the optical properties of pbss , whose influence has been neglected so far . and
the correlated photon rate @xmath233 for two noise levels in the channels ( @xmath234 and @xmath235 , where @xmath236 and @xmath237).,width=8,height=5 ] in fig .
4 the behavior of the qber is presented versus the level of correlation , @xmath238 , in the alice channel and the rate , @xmath107 , of the correlated photon pairs for two different noise levels , i.e. , the ratio between the mean rate of photons in the two channels .
far from ideal conditions , the presence of uncorrelated events in the two channels induces a non - linear increase of the qber .
the other parameters are set as in fig . 3 for realistic pbss parameters , @xmath239 and @xmath240 .
ekert s protocol has the peculiarity of relying on the completeness of quantum mechanics for security .
therefore , the possible combined choices between alice and bob for analyzer settings split into three groups : the first for key distribution , the second containing the security proof , and the third garnering the discarded measurements . here
we consider two possible variants of ekert s protocol : the variant based on the clauser - horne - shimony - holt inequality ( chsh ) , similar to the one proposed in ref .
@xcite , and the variant based on wigner s inequality @xcite . to increase the number of measurements devoted to the key distribution we consider the case where alice and bob measure randomly among four analyzer settings and use the chsh inequality to test eavesdropping . in this scheme ,
alice s choices for the analyzer settings are @xmath241 and bob s are @xmath242 .
the key distribution is performed when alice and bob s settings are the same or orthogonal , so that the sifted key is @xmath243 t,\end{aligned}\ ] ] where ( @xmath244 are angular settings generating the key and @xmath245 in a maximally entangled state configuration , the detectors contributing to the key should be @xmath211 and @xmath246 for the orthogonal analyzer settings and @xmath208 and @xmath247 for the parallel settings . thus the qber is calculated according to @xmath248 t\right\ } , \ ] ] where the intuitive notation @xmath249 indicates the sum over @xmath208 and @xmath247 detectors and @xmath250 indicates the sum over @xmath211 and @xmath251 as in ref .
@xcite , we consider the case of the ekert s variant where the security of the quantum channels follows from wigner s inequality . in this case ,
alice and bob measure randomly among four analyzers settings whose choices are @xmath252 for alice and @xmath253 for bob .
the key distribution is performed when alice s and bob s settings are the same so that the sifted key is @xmath254 with @xmath255 the qber is calculated according to @xmath256 t}{k_{\text{wi}}(\theta _ { a})},\ ] ] by taking the detectors contributing to the wrong bits as @xmath211 and @xmath257 and the entanglement parameter @xmath258 .
the parameters settings are @xmath259 @xmath260 s@xmath261@xmath262 @xmath263 @xmath264 @xmath265 @xmath266 s@xmath219 , @xmath220 ns , @xmath222 ns and @xmath267 , taken from typical and experimental realistic values so far implemented .
pbss are considered real with @xmath268 and @xmath240 ( @xmath269).,width=8,height=5 ] in fig .
5 we present a comparison of qber levels for the bb84 protocol and the ekert protocols considering both chsh and wigner s inequality versus the analyzers angular setting @xmath270 and the entanglement parameter @xmath9 .
experimental conditions are the same ( low noise ) as for figs . 3 and 4 .
results highlight that the qber is sensitive to the angle @xmath271 when the ideal entanglement is not achieved for both variants of ekert s protocols . in the wigner s case ,
the sensitivity is so remarkable that this protocol has to be considered less robust than bb84 and the chsh based ekert s protocol .
the security of the bb84 variant protocol is based on a public comparison between alice and bob s measurements on a sufficiently large random subset of the sifted key , e.g. more than half is recommended in @xcite . the security proof for the chsh - inequality - based ekert s protocol is evaluated with the specific choices of settings by the chsh inequality , @xmath272 where we have @xmath273 here @xmath274 is the normalized coincidence rate as a function of the analyzer settings and detector choices .
the terms @xmath275 are commonly stored when experiments are performed . for maximally entangled states we have @xmath276 , while for any realistic local theory we have @xmath277 it
is expected that the presence of an eavesdropper will reduce the observed value of @xmath278 giving @xmath279 when the eavesdropper measures photons over either one or both ( total eavesdropping ) of alice s and bob s channels @xcite . in the case of wigners - inequality - based ekert s protocol ,
the wigner s inequality result has @xmath280 giving for the maximally entangled states , @xmath281 and @xmath282 for any local realistic theory .
as for the chsh inequality , it can be proved that the limit becomes @xmath283 for eve detecting only one photon of the pair , while in the case of total eavesdropping there is no boundary condition @xcite . and w@xmath284 versus the coincidence window @xmath285 and the correlation level in the alice channel @xmath286 .
in the case of maximally entangled states we have w@xmath287s@xmath2871 , while the lower cut is defined by the eavesdropping limit w@xmath288 and s@xmath288 .
the choice of parameter values are the same of fig . 4 except for @xmath289 and @xmath290 .
the lower two surfaces represented correspond to w@xmath284 and s@xmath284 in the case of real pbss , with parameters @xmath291 and @xmath292 the higher surface represents s@xmath293=w@xmath284 in the case of ideal pbss.,width=8,height=5 ] in fig .
6 we compare the behaviors of the chsh and wigner s inequality parameters , @xmath294 = @xmath295 and @xmath296 versus the coincidence window @xmath285 and the correlation level in alice s channel , @xmath297 the lower surfaces represent the case of real pbss , where @xmath298 @xmath299 , while the upper surface corresponds to ideal pbss , where @xmath300 @xmath301 .
we observe that , given the same noise level in the system and real pbss , wigner s parameter reaches the eavesdropping limit faster than the chsh one does , revealing the intrinsic weakness of wigner s test against experimental parameters .
furthermore , the wigner s security test guarantees against eavesdropping strategies only for the detection of one photon of the pair , while the chsh security is independent on the adopted strategy ( see refs .
@xcite ) .
a satisfactory protocol must be able to recover from noise as well as from partial leakage , allowing alice and bob to reconcile the two strings of bits measured and distill from the sifted key a corrected key .
a strong need for the application of any error - correction method is an _ a priori _ knowledge of the qber , which provides information regarding how many times the error - correction procedure must be applied to reduce the qber to a certain agreed level , commonly 1 % . here , we show an example of error correction on an _ a priori _ evaluated qber according to a common approach reported in ref .
@xcite , to show that our model allows for prediction of the corrected key length . in general
, alice and bob can not distinguish between errors caused either by an eavesdropper or by the environment .
thus , they must assume that all errors are due to an eavesdropper and evaluate the leaked information from the qber . also ,
even though by the error - correction procedure one can disregard incorrect bits by simply dropping them off in building the distilled key , the residual knowledge of an eavesdropper may still not be faithfully quantified by the reduced qber obtained after the correction .
the effects of eve s strategy is in fact equivalent to quantum noise yielding eventually accidental coincidences , these last contributing to both incorrect and correct bits transmitted , as it is clear from eqs .
( [ k1 ] , [ qber1 ] ) .
hence the error - correction procedure is not sufficient to cancel a potential eve s knowledge of part of the key due to accidental coincidence . to prove this last assertion we introduce the quantum accidental bit rate ( qabr ) , a quantity related to accidental coincidences and , in this sense , analogous to the qber
however , the error - correction procedure can not reduce the qabr at the qber level . as an example
, we give the qabr in the case of ekert s protocol based on wigner s inequality , @xmath302 and the correlation parameter in the alice channel @xmath303 : the sawtooth shape is due to the application of the error correction procedure . the parameters are set as in figure 6 except that , in this case , ideal pbss are considered .
, width=8,height=5 ] figure 7 shows the qber with and without the application of the error - correction procedure together with the 1/2 qabr vs the coincidence window and the correlation parameter in alice s channel .
the error - correction procedure is very ineffective at reducing the qabr and consequently the possible effect of eve s knowledge on the corrected key .
this paper is concerned with an _ a priori _ evaluation of qckd crucial parameters when entangled photons produced by spdc are exploited .
the basic experimental feature consists in the detection of coincident photons . toward this aim
, we developed a statistical model to calculate the probability of accidental coincidences contributing to errors in the sifted key , not completely accounted for by simple experimental means .
we investigated the noise contribution due to imperfect source generation and selection , imperfect polarizing beam splitters performing polarization states analysis , and noisy and lossy measurement system for photon - number detection .
we emphasized some basic system imperfections such as uncorrelated photons collection , detection system deficiencies , detection system noise due to detector dark counts , and electronic system imperfections associated with non - ideal - time - correlation measurements .
we discussed how this model can be adopted for the evaluation of the qber and the sifted key for different well - known protocols , i.e. , bb84 and ekert s protocols based on both chsh and wigner s inequality , and compared them to expected results . given that this model predicts precisely the qber and the sifted key
, it ultimately guarantees a method to compare different security criteria of the hitherto proposed qckd protocols and provides an objective assessment of performances and advantages of different systems .
thus , it yields a method for an _ a priori _ evaluation of the tolerable experimental imperfections in a practically implemented quantum system to establish the degree of security and competitiveness of qckd systems . finally , we used the model in a standard error - correction procedure , observing that this does not completely cancel the possible residual eavesdropping knowledge on the corrected key .
we emphasize that this model yields also the degree of security of the corrected key , if a precise modelling of system imperfections is provided . acknowledgment .
this work was developed in collaboration with elsag s.p.a . ,
genova ( italy ) , within a project entitled quantum cryptographic key distribution " co - funded by the italian ministry of education , university and research ( miur ) - grant n. 67679/ l. 488 .
in addition stefania castelletto acknowledges the partial support of the darpa quist program .
we explicitly calculate the unitary transformation @xmath305 according to eq.s ( [ eq5 ] ) obtaining a 16@xmath30616 matrix @xmath307 where we indicate @xmath308 ; \;\;\;\;\;u_{2}=\left [ \begin{array}{cccc } { \cal t}_{a}{\cal r}_{b } & 0 & 0 & 0 \\ 0 & { \cal t}_{a}{\cal r}_{b}^{\perp } & 0 & 0 \\ 0 & 0 & { \cal t}_{a}^{\perp } { \cal r}_{b } & 0 \\ 0 & 0 & 0 & { \cal t}_{a}^{\perp } { \cal r}_{b}^{\perp } \end{array } \right ] \\
u_{3 } & = & \left [ \begin{array}{cccc } { \cal r}_{a}{\cal t}_{b } & 0 & 0 & 0 \\ 0 & { \cal r}_{a}{\cal t}_{b}^{\perp } & 0 & 0 \\ 0 & 0 & { \cal r}_{a}^{\perp } { \cal t}_{b } & 0 \\ 0 & 0 & 0 & { \cal r}_{a}^{\perp } { \cal t}_{b}^{\perp } \end{array } \right ] ; \;\;\;\;\;\;\;\;u_{4}=\left [ \begin{array}{cccc } { \cal r}_{a}{\cal r}_{b } & 0 & 0 & 0 \\ 0 & { \cal r}_{a}{\cal r}_{b}^{\perp } & 0 & 0 \\ 0 & 0 & { \cal r}_{a}^{\perp } { \cal r}_{b } & 0 \\ 0 & 0 & 0 & { \cal r}_{a}^{\perp } { \cal r}_{b}^{\perp } \end{array } \right]\end{aligned}\ ] ] and @xmath309 , @xmath310 @xmath311 analogously for @xmath312 and @xmath313 according to eq.s ( [ eqste ] ) and ( [ ti2 ] ) we obtain for @xmath314 the following : @xmath315 + \\
\sin ^{2}(\theta _ { b})\left [ \begin{array}{c } \cos ^{2}(\theta _ { a})\left ( \left| \mathcal{t}_{a}\mathcal{t}_{b}\right| ^{2}+\left| \mathcal{t}_{a}^{\perp } \mathcal{t}_{b}^{\perp } \epsilon \right| ^{2}\right ) + \sin ^{2}(\theta _ { a})\left ( \left| \mathcal{t}_{a}^{\perp } \mathcal{t}% _ { b}\right| ^{2}+\left| \mathcal{t}_{a}\mathcal{t}_{b}^{\perp } \epsilon \right| ^{2}\right ) \end{array } \right ] + \\ \sin ( 2\theta _ { a})\cos ( 2\theta _ { b})\mathrm{re}(\epsilon \zeta /2)\left
( \begin{array}{c } -\left| \mathcal{t}_{a}\mathcal{t}_{b}\right| ^{2}+\left| \mathcal{t}% _ { a}^{\perp } \mathcal{t}_{b}\right| ^{2}+ \left| \mathcal{t}_{a}\mathcal{t}_{b}^{\perp } \right| ^{2}-\left| \mathcal{t}% _ { a}^{\perp } \mathcal{t}_{b}^{\perp } \right| ^{2 } \end{array } \right ) \end{array } \right\ } , \ ] ] @xmath316 + \\ \sin ^{2}(\theta _ { b})\left [ \begin{array}{c } \cos ^{2}(\theta _ { a})\left ( \left| \mathcal{t}_{a}\mathcal{r}_{b}\right| ^{2}+\left| \mathcal{t}_{a}^{\perp } \mathcal{r}_{b}^{\perp } \epsilon \right| ^{2}\right ) + \sin ^{2}(\theta _ { a})\left ( \left| \mathcal{t}_{a}^{\perp } \mathcal{r}% _ { b}\right| ^{2}+\left| \mathcal{t}_{a}\mathcal{r}_{b}^{\perp } \epsilon \right| ^{2}\right ) \end{array } \right ] + \\ \sin ( 2\theta _ { a})\cos ( 2\theta _ { b})\mathrm{re}(\epsilon \zeta /2)\left ( \begin{array}{c }
-\left| \mathcal{t}_{a}\mathcal{r}_{b}\right| ^{2}+\left| \mathcal{t}% _ { a}^{\perp } \mathcal{r}_{b}\right| ^{2}+ \left| \mathcal{t}_{a}\mathcal{r}_{b}^{\perp } \right| ^{2}-\left| \mathcal{t}% _ { a}^{\perp } \mathcal{r}_{b}^{\perp } \right| ^{2 } \end{array } \right ) \end{array } \right\ } , \ ] ] @xmath317 + \\
\sin ^{2}(\theta _ { b})\left [ \begin{array}{c } \cos ^{2}(\theta _ { a})\left ( \left| \mathcal{r}_{a}\mathcal{t}_{b}\right| ^{2}+\left| \mathcal{r}_{a}^{\perp } \mathcal{t}_{b}^{\perp } \epsilon \right| ^{2}\right ) + \sin ^{2}(\theta _ { a})\left ( \left| \mathcal{r}_{a}^{\perp } \mathcal{t}% _ { b}\right| ^{2}+\left| \mathcal{r}_{a}\mathcal{t}_{b}^{\perp } \epsilon \right| ^{2}\right ) \end{array } \right ] + \\ \sin ( 2\theta _ { a})\cos ( 2\theta _ { b})\mathrm{re}(\epsilon \zeta
/2)\left ( \begin{array}{c } -\left| \mathcal{r}_{a}\mathcal{t}_{b}\right| ^{2}+\left| \mathcal{r}% _ { a}^{\perp } \mathcal{t}_{b}\right| ^{2}+ \left| \mathcal{r}_{a}\mathcal{t}_{b}^{\perp } \right| ^{2}-\left| \mathcal{r}% _ { a}^{\perp } \mathcal{t}_{b}^{\perp } \right| ^{2 } \end{array } \right ) \end{array } \right\ } , \ ] ] @xmath318 + \\
\sin ^{2}(\theta _ { b})\left [ \begin{array}{c } \cos ^{2}(\theta _ { a})\left ( \left| \mathcal{r}_{a}\mathcal{r}_{b}\right| ^{2}+\left| \mathcal{r}_{a}^{\perp } \mathcal{r}_{b}^{\perp } \epsilon \right| ^{2}\right ) + \sin ^{2}(\theta _ { a})\left ( \left| \mathcal{r}_{a}^{\perp } \mathcal{r}% _ { b}\right| ^{2}+\left| \mathcal{r}_{a}\mathcal{r}_{b}^{\perp } \epsilon \right| ^{2}\right ) \end{array } \right ] + \\ \sin ( 2\theta _ { a})\cos ( 2\theta _ { b})\mathrm{re}(\epsilon \zeta /2)\left ( \begin{array}{c } -\left| \mathcal{r}_{a}\mathcal{r}_{b}\right| ^{2}+\left| \mathcal{r}% _ { a}^{\perp } \mathcal{r}_{b}\right| ^{2}+ \left| \mathcal{r}_{a}\mathcal{r}_{b}^{\perp } \right| ^{2}-\left| \mathcal{r}% _ { a}^{\perp } \mathcal{r}_{b}^{\perp } \right| ^{2 } \end{array } \right ) \end{array } \right\ } .\ ] ] in the case of maximally entangled states , i.e. @xmath319 and @xmath320 , and ideal pbss , i.e. , @xmath321 and@xmath322 the @xmath51 is simply given by @xmath323 ; \;\;\;\;\;\ ; u_{2}=\left [ \begin{array}{cccc } 0 & 0 & 0 & 0 \\ 0 & i & 0 & 0 \\ 0 & 0 & 0 & 0 \\ 0 & 0 & 0 & 0 \end{array } \right ] ; \;\;\;\;\;\ ; u_{3}=\left [ \begin{array}{cccc } 0 & 0 & 0 & 0 \\ 0 & 0 & 0 & 0 \\ 0 & 0 & i & 0 \\ 0 & 0 & 0 & 0 \end{array } \right ] ; \;\;\;\;\;\ ; u_{4}=\left [ \begin{array}{cccc } 0 & 0 & 0 & 0 \\ 0 & 0 & 0 & 0 \\ 0 & 0 & 0 & 0 \\ 0 & 0 & 0 & 1 \end{array } \right]\end{aligned}\ ] ] and @xmath324
according to refs .
@xcite , in the case of non - extending dead time , the correction is @xmath325 where @xmath326 are the mean number of photons counted in the @xmath15 channel , i.e. @xmath327 with @xmath328 calculated in the absence of dead time @xmath329 , and @xmath330 . we can therefore write down the dead time correction in this case as @xmath331 \,\,d_{a}\right\ } ^{-1}\ ] ] and @xmath332 \,\,d_{b}\right\ } ^{-1}\ ] ] by noting that , when several devices are used in series , a good approximation considers the whole apparatus to be a black box , with a non - extending dead time equal to the largest of dead times of the single component @xcite .
we showed in ref .
@xcite that @xmath333provides a satisfactory approximation for @xmath334 .
99 c. bennett and g. brassard , in _ proceedings of the ieee international conference on computers , systems and signal processing , bangalore _ , ( ieee , new york 1984 ) , p. 175 .
a. k. ekert , phys .
rev . lett . * 67 * , 661 ( 1991 ) . c. h. bennett , g. brassard , and n. d. mermin , phys .
lett . * 68 * , 557 ( 1992 ) .
g. brassard , n. lutkenhaus , t. mor , and b. c. sanders , phys .
* 85 * , 1330 ( 2000 ) .
a. k. ekert , j. g. rarity , p. r. tapster , and g. m. palma , phys .
. lett . * 69 * , 1293 ( 1992 ) .
a. v. sergienko , m. atature , z. walton , g. jaeger , b. e. a. saleh , and m. c. teich , phys .
a * 60 * , 2622 ( 1999 ) .
t. jennewein , c. simon , g. weihs , h. weinfurter , and a. zeilinger , phys .
* 84 * , 4729 ( 2000 ) .
d. s. naik , c. g. peterson , a. g. white , a. j. berglund , and p. g. kwiat , phys .
lett . * 84 * , 4733 ( 2000 ) .
w. tittel , j. brendel , h. zbinden , and n. gisin , phys .
lett . * 84 * , 4737 ( 2000 ) .
l. mandel , j. opt .
b * 1 * , 108 ( 1984 ) .
e. jakeman and j. g. rarity , opt .
commun . * 59 * , 219 ( 1986 ) .
j. g. rarity , p. r. tapster , and e. jakeman , opt . commun . * 62 * , 201 ( 1987 ) .
j. g. rarity and p. r. tapster , appl .
b * 55 * , 298 ( 1992 ) .
d. c. burnham and d. l. weinberg , phys .
lett . * 25 * , 84 ( 1970 ) .
d. n. klyshko , _ photons and nonlinear optics _ , ( gordon and breach science publishers , new york , amsterdam , 1988 ) .
j. g. rarity , k. d. ridley , and p. r. tapster , appl . opt . * 26 , * 4616 ( 1987 )
. a. n. penin and a. v. sergienko , appl . opt .
* 30 * , 3582 ( 1991 ) .
p. g. kwiat , a. m. steinberg , r. y. chiao , p. h. eberhard , and m. d. petroff , appl . opt . * 33 * , 1844 ( 1994 ) . a. l. migdall , r. u. datla , a. v. sergienko , j. s. orszak , and y. h. shih , metrologia * 32 * , 479 ( 1996 ) . s. castelletto , a. godone , c. novero , and m. l. rastello , metrologia * 32 * , 501 ( 1996 ) . m. m. hayat , a. joobeur , and b. e. a. saleh , j. opt .
a * 16 * , 348 ( 1999 ) .
g. brida , s. castelletto , c. novero , and m. l. rastello , j. opt .
b * 16 * , 1623 ( 1999 ) .
g. brida , s. castelletto , i. p. degiovanni , c. novero , and m. l. rastello , metrologia * 37 * , 625 ( 2000 ) .
p. r. tapster , j. g. rarity , and p. c. m. owens , phys .
lett . * 73 * , 1923 ( 1994 ) .
w. tittel , j. brendel , h. zbinden , and n. gisin , phys .
* 81 * , 3563 ( 1998 ) .
g. weihs , t. jennewein , c. simon , h. weinfurter , and a. zeilinger , phys .
lett . * 81 * , 5039 ( 1998 ) .
m. nielsen and i. chuang , _ quantum computation and quantum information _ , ( cambridge university press , new york , 2000 ) .
n. boeuf , d. branning , i. chaperot , e. dauler , s. guerin , g. jaeger , a. muller , and a. migdall , opt . eng . * 39 * , 1016 ( 2000 ) .
p. g. kwiat , k. mattle , h. weinfurter , a. zeilinger , a. v. sergienko , and y. shih , phys .
lett . * 75 * , 4337 ( 1995 ) .
entangled states may be successfully generated for this purpose by two type i nonlinear crystals , according to p. g. kwiat , e. waks , a. g. white , i. appelbaum , and p.h .
eberhard , phys .
a * 60 * , r773 ( 1999 ) , so that the general state is @xmath335\,\left| v_{a}\right\rangle \left| v_{b}\right\rangle .$ ] p. g. kwiat , a. j. berglund , j. b. altepeter , and a. g. white , science * 290 * , 498 ( 2000 ) .
a. j. berglund , arxiv : quantph/0010001 v2 ( 2000 ) .
s. castelletto , i. p. degiovanni , and m. l. rastello , in _ quantum communication , computing , and measurement 3 , _ proceedings of the fifth international conference on quantum communication , measurement and computing , capri , italy , edited by p. tombesi and o. hirota , ( kluwer academic , new york , 2001 ) , p. 131 .
j. perina jr . , o. haderka , and j. soubusta , phys .
a * 64 * , 052305 ( 2001 ) .
s. castelletto , i. p. degiovanni , and m. l. rastello , j. opt .
b * 19 * , 1247 ( 2002 ) .
s. castelletto , i. p. degiovanni , and m. l. rastello , metrologia * 37 * , 613 ( 2000 ) .
s. castelletto and i. p. degiovanni , _ private communication . _ s. castelletto , i. p. degiovanni , and m. l. rastello , in _ advanced mathematical and computational tools in metrology v _ , edited by p. ciarlini , m. cox , e. filipe , f. pavese , and d. richter , ( world scientific company , singapore , 2001 ) , p. 41 . | quantum - cryptography key distribution ( qckd ) experiments have been recently reported using polarization - entangled photons . however , in any practical realization , quantum systems suffer from either unwanted or induced interactions with the environment and the quantum measurement system , showing up as quantum and , ultimately , statistical noise . in this paper , we investigate how ideal polarization entanglement in spontaneous parametric downconversion ( spdc ) suffers quantum noise in its practical implementation as a secure quantum system , yielding errors in the transmitted bit sequence . because all spdc - based qckd schemes rely on the measurement of coincidence to assert the bit transmission between the two parties , we bundle up the overall quantum and statistical noise in an exhaustive model to calculate the accidental coincidences .
this model predicts the quantum - bit error rate and the sifted key and allows comparisons between different security criteria of the hitherto proposed qckd protocols , resulting in an objective assessment of performances and advantages of different systems . |
for the case of several known gravitational lens systems , the deflector is either not yet detected or it is so faint that its redshift remains unknown ( see the gravitational lensing web page at ` http://vela.astro.ulg.ac.be/grav_lens/ ` ) .
measurement of the deflector redshift is mandatory in order to properly model the corresponding gravitational lens systems .
usually , the lens redshift is determined from spectroscopic observations of either emission lines due to the lensing galaxy or of galaxy absorption lines detected in the spectra of the qso images . in this work ,
we present an alternate method to derive the redshift of a too faint and/or an invisible lens .
first , we describe the method which is based upon the fitting of photometric observations with a redshifted extinction model and we then show results from monte carlo simulations .
the latter turn out to be very useful in order to assess the errors affecting the estimate of the lens redshift .
next , we attempt to apply this method to several known gravitational lens systems and outline all observational requirements needed in order to derive successful results .
finally , we propose a generalization of this method .
conclusions form the last section .
figure [ schema ] depicts a gravitational lens system .
the light rays emitted from the quasar are deflected by the lensing galaxy and the resulting macro images are reddened because of extinction effects in the deflector . due to the different path lengths across the galaxy , differential reddening results between the macro lensed components .
the flux received by the observer from component @xmath0 at a wavelength @xmath1 is : @xmath2 where @xmath3 is the flux that would be observed from the quasar in the absence of lensing , @xmath4 the macro lens amplification of component @xmath0 , ext(@xmath5 ) the galaxy dust opacity at the wavelength @xmath6 ( @xmath7 being the lens redshift ) and @xmath8 ( see cardelli _ et al . _
@xcite ) , and @xmath9 the path length of component @xmath0 through the deflector . when deriving eq .
( [ flux ] ) , we have of course implicitly assumed that the macro lensed images are not contaminated by light from the lensing galaxy , that they are not affected by microlensing effects , that no intrinsic colour variations exist between the lensed components , that the law of extinction ( first assumed to be alike that of the galaxy ) is similar all through the deflector and that the observed reddening takes place in a single lensing plane .
a check on the applicability of most of these conditions can be made as suggested in section [ colour - colour_diagrams ] .
the colour of component @xmath0 , between the wavelengths @xmath10 and @xmath11 , is then merely given by : @xmath12 having chosen one of the macro - lensed images as a reference component ( i.e. @xmath13 ) , the relative colour of component @xmath14 ( i.e. @xmath15 ) may be expressed as : @xmath16 \label{rel_colour}\end{gathered}\ ] ] where @xmath17 represents the relative path length of component @xmath14 .
if we assume that the gravitational lens system consists of @xmath18 macro lensed components and that observations are being carried out through @xmath19 filters , we easily find that there are @xmath20 independent colours for each of the @xmath21 components ( the reference component is of course not included ) .
thus , there results a set of @xmath22 equations alike eq .
( [ rel_colour ] ) .
besides , there are 2 unknowns which do not depend on the numbers of filters and components : @xmath7 and @xmath23 .
moreover , for each of the @xmath21 components , there is one additional unknown : @xmath24 .
consequently , we have a total of @xmath25 unknowns .
since we must have more equations than unknowns , we easily establish the necessary condition : @xmath26 for @xmath18 = 2 , 3 , or 4 , we easily find out that @xmath27 4 , 3 or @xmath28 , i.e. a minimum of 4 , 3 or 3 filters is requested , respectively . therefore , observational data of multiply imaged quasars acquired through several photometric filters are mandatory . note that if data obtained in @xmath19 filters come from several different observing runs , it is then safer to only use those colours formed between filters from the same run and hence we finally get less than @xmath20 useful colours .
the above necessary condition may consequently not be any longer fulfilled . before applying the proposed method , we must make sure that in a colour - colour diagram , and in accordance with eq .
( [ rel_colour ] ) , the measurements of the various lensed components are approximately located along a straight line . otherwise , it may mean that one or more of the conditions listed in section [ generalities ] are not fulfilled .
secondly , the differential reddening for most of the components has to be significantly larger than the observational errors .
otherwise , the slope of the straight line in the colour - colour diagram remains too uncertain .
for example , the components of the gravitational lens systems ( kristian _ et al .
_ @xcite ) and ( yee & ellingson @xcite ; yee & bechtold @xcite ) have very similar colours and hence these cases are not appropriate for application of this method .
promising candidates are ( see an example of colour - colour diagram in fig .
[ diagram ] ) and especially which is known to be very stable ( moore & hewitt @xcite ) and probably not affected by significant microlensing effects ( mcleod _ et al . _
@xcite ) .
the left side of equation ( [ rel_colour ] ) , which corresponds to an experimental relative colour , is directly derived from photometric observations of the macro lensed quasar images . the right side of equation ( [ rel_colour ] ) , and more explicitly the quantity @xmath29 " , is systematically computed for different contiguous values of @xmath23 and @xmath7 using the galaxy extinction law from cardelli _ et al . _
( @xcite ) .
the relative path lengths @xmath24 are fitted using a routine that minimizes the following @xmath30 : @xmath31 figure [ carte_simul ] presents such a @xmath30 map obtained from a set of simulated data ( see below ) .
in order to study the errors affecting the estimate of the lens redshift using the fitting method described above , we have carried out a wide range of monte carlo simulations .
when doing these , we first fix the values of the following parameters : the lens redshift @xmath7 within a defined interval , the parameter @xmath23 usually set to 3.1 - corresponding to the average galactic value - , the relative macro - amplification @xmath32 and the relative path length @xmath24 through the deflector for each lensed component so that we can calculate precise relative magnitudes for as many different filters as we wish .
then we add to these magnitudes a gaussian noise of a certain level and apply the above fitting method .
we repeat these two last steps 100 times and calculate the dispersion @xmath33 in the estimate of the lens redshift and the @xmath23 parameter . carrying out such simulations for different contiguous values of the theoretical lens redshift and for levels of noise comprised between 0.01 mag . and 0.05 mag .
, we obtain the curves illustrated in fig .
[ noise ] for which we used the set of the five u , b , v , r and i johnson filters .
we see that the dispersion @xmath33 affecting the estimate of @xmath7 is about 0.1 and even less at higher redshifts . the visible oscillations " in fig .
[ noise ] correspond to the positions of the wavelengths of the different selected filters with respect to the redshifted 2175 absorption bump ( see cardelli _
@xcite ) . as an example ,
[ plot_abs ] represents the extinction law of cardelli _ et al . _
( @xcite ) as a function of the reciprocal wavelength @xmath34 for @xmath35 . the redshifted ( @xmath36 )
positions of the u , b , v , r , i , j , h , k and l filters are indicated .
furthermore , we have illustrated in fig .
[ histos_rv ] dispersions affecting the derived values of @xmath23 for the cases of 5 filters and two different levels of noise : 0.01 mag . and 0.05 mag .
we see that , in this latter case , the dispersion gets rather large . next , fig .
[ filters ] shows on a logarithmic scale differences in the lens redshift dispersion when using the u , b , v , r and i filters or the infrared i , j , h , k and l filters .
since the method is sensitive to the extinction gradient between two filter wavelengths , the much higher dispersion derived for the infrared filters is mainly accounted for by the shallower shape of the extinction law in the infrared range .
moreover , the lower extinction in the infrared range also affects the dispersion in the estimate of the lens redshift because the experimental relative colours are smaller and hence the effects of noise get more pronounced .
finally , we have also used other extinction laws such those proposed by pei ( pei @xcite ) for the milky way ( `` mw '' ) , the large magellanic cloud ( `` lmc '' ) and the small magellanic cloud ( `` smc '' ) . from the same set of simulated data as in fig .
[ carte_simul ] , fig .
[ courbes_pei ] shows the resulting @xmath30 curves .
comparable minima and lens redshift solutions are usually found for the mw and lmc laws .
however , high quality photometric observations are mandatory in order to efficiently select the most likely lens redshift among acceptable solutions for the different extinction laws .
note that because of the noise in the simulated photometric data , the minimum for the mw curve in fig .
[ courbes_pei ] is found at @xmath37 , instead of the theoretical value @xmath38 .
[ [ section ] ] figure [ carte_2237 ] represents the @xmath30 map for the derived from the published sets of photometric observations listed in table [ phot_2237 ] .
the minimum is found for @xmath39 . however , the redshift of this gravitational lens is known to be 0.04 ( huchra _ et al . _
the small discrepancy between these two redshift values is very likely due to the fact that the data were obtained during two different runs when photometric quality was not good enough ( @xmath40 mag . for approximately half the reported values in table [ phot_2237 ] ) .
moreover , this gravitational lens system is very likely affected by microlensing effects on a time scale of several years ( irwin _ et al .
_ @xcite ; corrigan _ et al .
_ @xcite ; stensen _ et al . _
the contours in fig .
[ carte_2237 ] correspond to @xmath41 and delineate , at a confidence level of @xmath42 , acceptable ranges for the values of @xmath7 and @xmath23 .
very high signal - to - noise photometric observations of the einstein cross , using several filters during a single run and reliable tools to derive the photometry of the multiple qso images , are badly needed to confirm that the different experimental relative colours reported between the lensed components are effectively , and solely , due to extinction effects in the deflector . [
cols="^,^,^,^,^,^,^ " , ] + @xmath43(1 )
mcleod et al .
@xcite ; ( 2 ) falco et al
[ phot_mg0414 ]
from eq . ( [ flux ] ) and considering the @xmath14 component relative to the _ ref_(-erence ) component , we may also write : @xmath44 } = \mathcal{a}_j\,e^{-\text{ext}(\lambda',r_v)\mathcal{l}_j}\ ] ] where @xmath45 .
from this equation , we straightforwardly derive : @xmath46 \ , -2.5 \log(\mathcal{a}_j ) .
\label{rel_mag}\ ] ] we see that in fact , one can also directly work with the relative magnitudes @xmath47 but that the relative macro lens amplifications @xmath48 then appear as additional unknowns which also need to be fitted .
following a reasoning similar to that adopted in section [ constraints ] , the necessary condition is also found to be : @xmath49 , where @xmath19 and @xmath18 have the same meaning . before applying this generalized method to real photometric observations
, one should make sure that intrinsic or extrinsic ( cf .
induced by micro - lensing ) light variations are not affecting the individual lensed components .
these conditions are of course more difficult to be checked than those imposed on the observed colours .
finally , it is also straightforward to establish from eq .
( [ rel_mag ] ) that the redshifted extinction law @xmath50 may be directly retrieved as a function of @xmath51 from the observed spectra of at least three multiple lensed qso images .
very high signal - to - noise , low spectral resolution , vlt observations of gravitational lens systems like and should enable one to derive such a redshifted extinction law and compare it with the known mw , lmc and smc ones .
monte carlo simulations show that we need very accurate photometric data ( typically @xmath52 mag . ) , in several ( @xmath53 ) filters , in order to estimate the lens redshift from the differential reddening observed between the lensed images of a multiply imaged quasar .
furthermore , the macro - lensed images must not be affected by light contamination from the lensing galaxy , by microlensing effects or by intrinsic colour variations reflected with a time dependence between the lensed components .
unfortunately , all these conditions are rarely fulfilled at the same time and existing photometric data are often not good enough or not sufficiently reliable .
consequently , accurate photometric observations of gravitational lenses obtained in several filters during the same run are badly needed .
finally , we have seen that the redshifted extinction law of the lensing galaxy could be directly retrieved by means of a similar method provided that we may record the individual spectra of at least 3 lensed components , over a wide wavelength range , assuming that no light contamination , no microlensing effects , affect the individual lensed components .
very good such candidates to be observed with the vlt are and .
cj is supported by contract arc94/99 - 178 action de recherche concerte de la communaut franaise ( belgium ) " and ple dattraction interuniversitaire , p4/05 ( sstc , belgium ) " .
js would like to acknowledge support from the sstc / prodex project `` observations of gravitational lenses '' .
_ note added in proof : _ on the basis of spectroscopic observations obtained for with the keck telescope ( see preprint astro - ph/9809063 submitted on 5 september 1998 ) , j.l .
tonry and c.s .
kochanek report the redshift determination ( @xmath54 ) of the very faint gravitational lens . based upon preliminary photometric observations of this system , our own redshift estimate is found to be in reasonable agreement with this value ( see fig . [ carte_mg0414 ] ) .
afanasev , v.l . ,
vlasyuk , v.v . ,
dodonov , s.n . ,
zhdanov , v.i . ,
mandzhos , a.v . ,
spiridonova , o.i . & khmil , s.v . ,
1996 , kinematics and physics of celestial bodies , vol .
5 , pp . 711 burud , i. , stabell , r. , magain , p. , et al . , 1998 , available at ` http://vela.astro.ulg.ac.be/preprint/index.html ` cardelli , j.a . ,
clayton , g.c . &
mathis , j.s . , 1989 ,
apj , 345 , 245 corrigan , r.t . , irwin , m.j . , arnaud , j. , et al . , 1991 ,
aj , 102 , 34 drew , j.e .
& boksenberg , a. , mnras , 1984 , 211 , 813 falco , e.e . ,
lehr , j. & shapiro , i.i . , 1997 , aj , 113 , 540 hazard , c. , morton , d.c . ,
terlevich , r. & mcmahon , r. , 1984 , apj , 282 , 33 huchra , j , gorenstein , m. , kent , s. , shapiro , i. , smith , g. , horine , e. & perley , r. , 1985 , aj , 90 , 691 irwin , m.j . ,
webster , r.l . , hewett , p.c . ,
corrigan , r.t .
& jedrzejewski , r.i . , 1989 ,
aj , 98 , 1989 kristian , j. , et al . , 1993 , aj , 106 , 1330 magain , p. , surdej , j. , swings , j .- p . , et al . , 1988 , nature , 334 , 325 mcleod , b.a . , bernstein , g.m . ,
rieke , m.j . & weedman , d.w . , 1998 , aj , 115 , 1377 moore , c.b .
& hewitt , j.n . , 1997 , apj , 491 , 451 stensen , r. , et al . , 1996 ,
a&a , 309 , 59 stensen , r. , et al . , 1997 , a&as , 126 , 393 pei , y.c .
, 1992 , apj , 395 , 130 turnshek , d.a .
, foltz , c.b . ,
grillmair , c.j . & weymann , r.j . , 1988 ,
apj , 325 , 651 turnshek , d.a . ,
lupie , o.l . , rao , s.m .
, et al . , 1997 , apj , 485 , 100 yee , h.k.c . , 1988 ,
aj , 95 , 1331 yee , h.k.c . & bechtold , j. , 1996 , aj , 111 , 1007 yee , h.k.c .
& ellingson , e. , 1994 , aj , 107 , 28 | light rays from a multiply imaged quasar usually sample different path lengths across the deflector .
extinction in the lensing galaxy may thus lead to a differential obscuration and reddening between the observed macro - lensed qso images .
these effects naturally depend on the precise shape of the extinction law and on the redshift of the lens . by means of numerical monte - carlo simulations , using a least - squares fitting method and assuming an extinction law similar to that observed in the galaxy , we show how accurate photometric observations of multiply imaged quasars obtained in several spectral bands could lead to the estimate of the lens redshift , irrespective of the visibility of the deflector .
observational requirements necessary to apply this method to real cases are thoroughly discussed .
if extinction laws turn out to be too different from galaxy to galaxy , we find out that more promising observations should consist in getting high signal - to - noise low resolution spectra of at least three distinct images of a lensed quasar , over a spectral range as wide as possible , from which it should be straightforward to extract the precise shape of the redshifted extinction law . very high
signal - to - noise , low spectral resolution , vlt observations of and should enable one to derive such a redshifted extinction law . |
the top quark , the heaviest fundamental particle with a mass about @xmath2 @xcite , plays an important role in the study of the electroweak symmetry breaking ( higgs boson ) as well as in the search of physics beyond the standard model ( bsm ) .
moreover , the production of top quark anti - quark pairs is one of the main backgrounds in many of the processes related with the standard model ( sm ) and bsm , and it is crucial to measure its production cross section with very high precision . the large hadron collider ( lhc ) has been in operation since 2009 , producing proton - proton collisions with a center of mass energy of @xmath3 until 2011 and @xmath1 in 2012 .
this document presents the most recent measurements of the inclusive @xmath4 cross section with data collected using the atlas @xcite and the cms @xcite detectors .
the most recent theoretical predictions for the top quark pair production cross section ( @xmath5 ) are @xmath6 and @xmath7 @xcite for a top - quark mass of @xmath8 . according to the sm , top quarks decay into a w boson and a b quark almost 100% of the times .
this leads to final states with two w bosons and two jets coming from the b quark fragmentation . when both w bosons decay leptonically , the event contains two high momentum leptons with opposite charge , two undetected neutrinos which are measured as missing energy in the plane transverse to the beam axis ( @xmath9 ) , and at least two jets , where two of them originate from b quarks .
the dilepton channel is optimal to measure the inclusive @xmath4 cross section due to the small background contribution from other sm processes . as will be presented in the following sections , the uncertainty on @xmath5 measurement in the dilepton channel
is dominated by systematic errors .
this analysis performed by the atlas collaboration @xcite exploits the fact that the @xmath4 , @xmath12 and @xmath11 processes dominate the muon - electron final state .
the simultaneous measurement of all three cross sections is performed at @xmath3 using the full 2011 dataset .
the cross section of these processes is extracted through a fit in a two - dimensional parameter space defined by the missing transverse energy ( @xmath9 ) and the jet multiplicity ( @xmath13 ) .
events from @xmath4 production have large @xmath9 and large @xmath13 , whereas @xmath12 events have large @xmath9 but small @xmath13 , and finally , drell - yan ( @xmath11 ) events are characterized by small @xmath9 and small @xmath13 .
the results obtained for the @xmath4 process are : @xmath14 the fiducial phase space is defined by the lepton cuts applied in the analysis : @xmath15 and @xmath16 , and it includes the @xmath17 decay mode .
figure [ atlas_simul ] shows the best fit for the cross section values with likelihood contours obtained from the simultaneous fit , overlayed with theoretical cross section predictions in the case of the total ( nnlo ) and fiducial ( nlo ) phase space .
the theoretical predictions are calculated for different pdf sets .
a better data - mc agreement is obtained when nnlo calculations are used . and drell - yan / ww processes @xcite .
in the case of the total cross section , the result is compared to nnlo available prediction , while the fiducial cross section is compared to the nlo calculation .
, title="fig:",scaledwidth=40.0% ] and drell - yan / ww processes @xcite . in the case of the total cross section , the result
is compared to nnlo available prediction , while the fiducial cross section is compared to the nlo calculation .
, title="fig:",scaledwidth=40.0% ] + ( a)(b )
the following analysis is a dedicated measurement of the @xmath4 cross section in the muon - electron decay mode using the full 2011 and 2012 datasets @xcite .
the @xmath4 production cross section ( @xmath5 ) is determined by counting the numbers of @xmath18 events with exactly one and exactly two b - tagged jets . in order to reduce the systematic uncertainties ,
the @xmath5 is extracted simultaneously with the combined probability to reconstruct and tag a b - jet from the top - quark decay .
the obtained cross section results in the @xmath18 channel are shown in equation [ 7 - 8tevatlas ] with the corresponding ratio ( @xmath20 ) .
figure [ atlas_btag ] shows the b - tagged jet distribution for the @xmath0 and @xmath1 datasets .
@xmath21 events in ( a ) @xmath22 and ( b ) @xmath23 data @xcite .
the data is compared to the expectation from signal and background simulations .
, title="fig:",scaledwidth=40.0% ] events in ( a ) @xmath22 and ( b ) @xmath23 data @xcite .
the data is compared to the expectation from signal and background simulations .
, title="fig:",scaledwidth=40.0% ] + ( a)(b )
the amount of data studied in this analysis corresponds to @xmath24 of proton - proton collisions taken by the cms detector at @xmath1 @xcite .
the extraction of @xmath5 is performed in all three dileptonic final states ( @xmath25 , @xmath26 and @xmath18 , including @xmath27 as signal ) .
the final @xmath5 result is obtained with a simple and robust cut - based analysis method .
equation [ 8tev_cms ] presents the results for all three channels .
figure [ cms_dilepton_8tev ] shows the b - tagged jet multiplicity and the @xmath9 distributions for the muon - electron channel , the decay mode with the smallest background contamination .
@xmath28 distributions for the muon - electron decay mode @xcite .
the expected distributions for @xmath4 signal and background sources are shown by stacked histograms ; data are shown by black dots .
, title="fig:",scaledwidth=40.0% ] distributions for the muon - electron decay mode @xcite .
the expected distributions for @xmath4 signal and background sources are shown by stacked histograms ; data are shown by black dots .
, title="fig:",scaledwidth=40.0% ] + ( a)(b )
the cms collaboration has measured the production cross section of @xmath4 pairs by considering dilepton decays where one of the w bosons decays into taus ( @xmath30 ) @xcite .
the identification of the @xmath29 lepton is performed by means of its hadronic decay products .
this decay channel has particular interest because it contributes as a background process in the search for a charged higgs boson @xcite ( @xmath31 with @xmath32 ) .
the @xmath4 cross section is extracted with a cut and count method using the @xmath33 of @xmath1 data collected by the cms detector in 2012 .
equation [ taucms ] presents the results for the muon - tau ( @xmath34 ) and electron - tau ( @xmath35 ) channels .
figure [ cms_dilepton_8tev ] shows the reconstructed @xmath36 and the b - tagged jet multiplicity distributions for the two channels combined .
@xmath37 ) and ( b ) the b - tagged jet multiplicity for the @xmath38 and @xmath34 channels combined @xcite . data ( points ) are compared with the sum of signal and background yields .
, title="fig:",scaledwidth=40.0% ] ) and ( b ) the b - tagged jet multiplicity for the @xmath38 and @xmath34 channels combined @xcite . data ( points ) are compared with the sum of signal and background yields . ,
title="fig:",scaledwidth=40.0% ] + ( a)(b )
the lhc has a very strong program in the measurement of the inclusive top quark pair production cross section with its two main detectors : atlas and cms .
the results presented in this document have reached an unprecedented precision of @xmath39 challenging the current theoretical predictions at nnlo+nnll .
figure [ cms_atlas_summary ] shows the most precise measurements performed by cms and atlas detectors at @xmath0 and @xmath1 compared with the last theoretical calculations .
all results are found to be in agreement with the standard model prediction . )
compared to the nnlo qcd calculation complemented with nnll resummation .
the measurements and the theory calculation is quoted at @xmath40 .
, scaledwidth=76.0% ]
9 atlas , cdf , cms , dcollaboration , `` first combination of tevatron and lhc measurements of the top - quark mass '' , http://www.arxiv.org/abs/1403.4427[arxiv:1403.4427 ] .
atlas collaboration , `` the atlas experiment at the cern large hadron collider'',http://iopscience.iop.org/1748 - 0221/3/08/s08003[jinst 3 ( 2008 ) s08003 ] .
cms collaboration , `` the cms experiment at the cern lhc '' , http://iopscience.iop.org/1748-0221/3/08/s08004[jinst 3 ( 2008 ) s08004 ] .
m. czakon , p. fiedler , and a. mitov , `` total top - quark pair - production cross section at hadron colliders through @xmath41 '' , ( 2013 ) . http://www.arxiv.org/abs/1303.6254[phys .
( 2013 ) 252004 ] .
atlas collaboration , `` simultaneous measurements of the top quark pair , @xmath10 , and @xmath42 @xmath43 @xmath44 production cross sections in pp collisions with the atlas detector at @xmath22 '' , _
http://arxiv.org/abs/1407.0573[arxiv:1407.0573 ] _ atlas collaboration , `` measurement of the @xmath4 production cross - section using @xmath45 events with b - tagged jets in pp collisions at @xmath46 and @xmath47 with the atlas detector '' , _
c74 ( 2014 ) 3109 ] _ cms collaboration , `` measurement of the @xmath4 production cross section in the dilepton channel in pp collisions at @xmath48 '' , _
http://link.springer.com/article/10.1007%2fjhep02%282014%29024[jhep 02 ( 2014 ) 024 ] _ cms collaboration , `` measurement of the ttbar production cross section in pp collisions at @xmath1 in the @xmath35 and @xmath34 dilepton final states '' , _
b 739 ( 2014 ) 23 ] _ j. f. gunion , h. e. haber , g. l. kane , and s. dawson , `` the higgs hunter s guide '' .
frontiers in physics .
addison - wesley , 1990 .
cms collaboration , `` measurement of the ratio b(@xmath49)/b(@xmath50 ) in pp collisions at @xmath51 '' , _
http://dx.doi.org/10.1016/j.physletb.2014.06.076[plb 736 ( 2014 ) 33 ] _ atlas collaboration , `` measurement of the @xmath4 production cross section in the @xmath29+jets channel using the atlas detector '' , _
j. c , 73 3 ( 2013 ) ] _ cms collaboration , `` measurement of the @xmath4 production cross section in the @xmath52 channel in pp collisions at @xmath22 '' , _
http://dx.doi.org/10.1140/epjc/s10052-013-2386-x[epj c73 ( 2013 ) 2386 ] _ | this document presents recent results of inclusive top - quark pair production cross section measurements at @xmath0 and @xmath1 .
the results are obtained analyzing the data collected by the cms and atlas detectors at the lhc accelerator .
studies are performed in the dilepton channel , where the smallest uncertainty is reached , with different approaches .
the most precise results of both experiments are combined and confronted with the most precise theoretical calculation ( nnlo - nnll ) . |
for the investigation of qcd in the strongly coupled regime , non - perturbative numerical methods such as lattice qcd and truncated dyson - schwinger equations can be employed .
similar to lattice qcd , numerical dyson - schwinger studies of qcd n - point functions are usually performed in euclidean space ( for recent reviews see @xcite ) .
although dyson - schwinger studies rely on truncation schemes , they have the great advantage , that they can in principle be solved in the continuum limit and with much higher numerical accuracy .
an extraction of dynamical properties , in particular the spectral functions of propagators , is however highly desirable . in quantum monte carlo simulations , the maximum entropy method ( mem )
turns out to be an especially suited tool and has been successfully applied in condensed matter physics ( see @xcite for a review ) , lattice qcd in the vacuum ( see @xcite for a review ) , as well as at finite temperatures @xcite . in this work
we show , that it can also successfully be employed for the extraction of spectral functions in dyson - schwinger studies .
the starting point for the mem is the linear relation between the spectral function and numerically determined euclidean correlation functions via generalized kllen - lehmann representations .
since the inversion of the latter is in general ill - posed due to the spectral properties of the linear operator , further knowledge has to be implemented non - linearly .
this can be done from a regularization point of view leading to the `` historical maximum entropy '' @xcite or in a bayesian approach leading to the `` classic maximum entropy '' @xcite and to `` bryan s method '' @xcite .
the key idea in the latter case is the interpretation of the spectral function as a probability distribution due to its special properties and a proper consideration of the numerical error . as a result
, the mem determines the most plausible or expected spectral function for a given euclidean correlator with known errors and some prior knowledge .
it does not rely on a special form of the function and should , with decreasing errors , converge towards the exact solution . for numerical dyson - schwinger studies
, the method seems especially reliable , since the calculations can usually be performed with much higher numerical accuracy and for much more momentum points of the correlators than in lattice qcd .
this paper is organized as follows : in section [ specfunc ] we collect those properties of spectral functions for fermions and bosons , which make them suitable for the application of the mem . in section [ mem ]
we discuss the mem procedure itself , in particular its adaptation to dyson - schwinger studies and to fermions . in section [ numimpl ]
we sketch the numerical implementation . after this
we show in section [ colddense ] some results for massless color - superconducting fermions in dense quark - gluon matter .
finally we summarize and conclude in section [ sumconc ] .
performing calculations for propagators in euclidean space or rather within the imaginary time formalism means the determination of matsubara propagators .
these are related to spectral functions in minkowski space via the generalized kllen - lehmann representation ( see e.g. @xcite ) . using the euclidean conventions of @xcite we have @xmath0 for a dirac fermion propagator , where @xmath1 denotes a given matsubara frequency and @xmath2
is the chemical potential .
the spectral function is given by @xmath3 therefore @xmath4 is hermitian and has only positive eigenvalues in a given hilbert space .
furthermore it has to fulfill the sum rule @xmath5 as a consequence of the ( anti-)commutation relations .
thus @xmath4/4 can be identified with a probability distribution , which is the key property for motivating the use of the mem . for a massive relativistic fermion in an isotropic ,
even parity and @xmath6-symmetric phase , we can parameterize @xmath7 and knowing that all eigenvalues have to be positive , we get @xmath8 and furthermore the sum rules @xmath9 for the application discussed in section [ colddense ] , we will for simplicity restrict ourself to the chiral limit , i.e. massless fermions ( the case of massive fermions is discussed at the end of section 3 ) . in this case
, we can rewrite the propagators by using the energy projectors @xmath10 and obtain @xmath11 the spectral functions @xmath12 then fulfill @xmath13 with @xmath14 for completeness we also wish to show , how the solutions of the inhomogeneous bethe - salpeter equation ( bse ) can be used to determine the spectral functions of mesons or diquarks . for a given current @xmath15 ,
the solution for the corresponding bethe - salpeter amplitude @xmath16 in momentum space ( see e.g. @xcite ) determines the time - ordered product @xmath17 such that the current - current correlator is given by @xmath18 with @xmath19 again @xmath20 possesses a generalized kllen - lehmann representation and the spectral function has to fulfill positivity conditions .
the maximum entropy method is a numerical tool for the inversion of potentially ill - posed linear equations by the implementation of additional information , i.e. constraints .
it can be easily viewed from a standpoint of regularization as adding some non - linear auxiliary conditions , leading to the so - called `` historical maximum entropy '' @xcite .
this is known to underfit the data by overestimating the effective number of degrees of freedom , thus leading to solutions that are closer to the prior estimate .
usually , the somewhat converse bayesian viewpoint is considered , since it essentially adjusts the number of effective degrees of freedom and also allows for an error estimation @xcite .
we briefly review this here , emphasizing the adaption to our problem . given a ( numerically evaluated ) euclidean correlator , which is treated as data @xmath21 , the objective is to determine the most plausible ( related to `` classic maximum entropy '' @xcite ) or the most expected ( related to `` bryan s method '' @xcite ) spectral function @xmath22 by taking into account prior knowledge @xmath23 of the solution , regulated by the prior estimate @xmath24 to be defined below .
the key entity is the plausibility functional @xmath25 $ ] for the spectral function @xmath26 under given @xmath21 and @xmath23 . with help of the so called `` bayesian theorem '' for conditional plausibilities @xmath27 = p[x|y]p[y ] = p[y|x]p[x],\end{aligned}\ ] ] this can be brought into the form @xmath28 & = & \frac{p[d|\rho h(m)]p[\rho|h(m)]}{p[d|h(m ) ] } \nonumber\\&\propto & p[d|\rho h(m)]p[\rho|h(m)].\end{aligned}\ ] ] here , we have introduced the `` likelihood function '' @xmath29 $ ] for the plausibility of the data @xmath21 under given @xmath26 and @xmath23 and the `` prior probability '' @xmath30 $ ] for the plausibility of @xmath26 under the prior knowledge @xmath23 . the constant plausibility @xmath31 $ ] of the data @xmath21 under the prior knowledge @xmath23 can be dropped , since we normalize the plausibility functional at the end . considering the function ( or sequence at finite temperature ) @xmath21 in an interval @xmath32 $ ] as uncorrelated data points obeying a gaussian distribution functional , we have @xmath33 & = & \exp\left(-l\left[\rho\right]\right),\end{aligned}\ ] ] with the likelihood @xmath34 & = & \frac{1}{b - a}\,\int_{a}^{b}\!dp_{4 } \frac{|d(p_{4})-d[\rho](p_{4})|^{2}}{2\sigma(p_{4})^{2 } } \nonumber\\&\simeq & \frac{1}{b - a } \sum_{i}\delta p_{4,i } \frac{|d_{i}-d[\rho]_{i}|^{2}}{2\sigma_{i}^{2}},\end{aligned}\ ] ] where @xmath35 $ ] is given by the generalized kllen - lehmann representation and a measure @xmath36 for the discretized integral @xmath37 the prior probability @xmath30 $ ] is usually somewhat arbitrary and essentially implements the positivity conditions non - linearly , at least from the regularization point of view .
it can be motivated by the law of large numbers or axiomatically constructed ( see e.g.@xcite ) .
the key idea in the latter case is to consider the spectral function as a probability distribution and derive the most general functional , fulfilling the requirements of subset independence , coordinate invariance , system independence and scaling .
it is then of the form @xmath38 & = & \int_{0}^{\infty}\!\!\ !
d\alpha\,\,\ , p[\rho|h(\alpha m)]p[\alpha|h(m)],\end{aligned}\ ] ] with the scaling factor @xmath39 for the prior estimate @xmath24 and the plausibility for the scaled prior estimate @xmath40 & = & \exp\left(\alpha s\left[\rho\right]\right),\end{aligned}\ ] ] where @xmath41 $ ] is the negative semi - definite entropy @xmath42 & = & \int_{-\infty}^{\infty}\!\!\!d\omega\,\,\ , \left ( \rho(\omega)-m(\omega ) -\rho(\omega)\ln\left(\frac{\rho(\omega)}{m(\omega)}\right ) \right ) \nonumber\\&\simeq & \sum_{i } \delta \omega_{i } \left ( \rho_{i}-m_{i } -\rho_{i}\ln\left(\frac{\rho_{i}}{m_{i}}\right ) \right ) \nonumber\\&= & -\sum_{i } 2\delta \omega_{i } \left(\sqrt{\rho_{i}}-\sqrt{m_{i}}\right)^{2 } + o\left((\sqrt{\rho_{i}}-\sqrt{m_{i}})^{3}\right)\end{aligned}\ ] ] with the measure @xmath43 in the saddle point approximation around the prior estimate .
the scaling factor @xmath39 basically scales the maximum of the entropy and will be balanced by the likelihood .
furthermore we will assume , that the plausibility @xmath44 $ ] can be dropped , which is called laplace s rule and can be justified a posteriori , as discussed in our example in section [ colddense ] . from eq .
( [ prho ] ) we therefore finally get @xmath45 & = & \frac{1}{z}\exp\left(q\left[\rho\right]\right),\end{aligned}\ ] ] with the negative semi - definite functional @xmath46=\alpha s\left[\rho\right]-l\left[\rho\right]$ ] , @xmath47 determined by normalization and the measure @xmath48 it is worth noting , that the `` historic maximum entropy '' @xcite simply determines the maximum , which is unique if it exists @xcite , of the functional @xmath46 $ ] with @xmath39 chosen , such that @xmath49 .
we will however follow `` bryan s method '' @xcite , aiming at the most expected spectral function , by computing @xmath50 \nonumber\\&\simeq & \int_{0}^{\infty}\!\!\ ! d\alpha\,\rho_{\alpha }
\,\,\,\frac{1}{z}\int\!\!\mathcal{d}\rho\ , \ , \exp\left(q\left[\rho\right]\right),\end{aligned}\ ] ] where it is assumed that @xmath25 $ ] is sharply peaked around its maximum @xmath51 .
we therefore define @xmath52 & = & \frac{1}{z } \int\!\!\mathcal{d}\rho\,\,\exp\left(q\left[\rho\right]\right ) \nonumber\\&\simeq & \frac{1}{z } \exp\left ( q\left[\rho_{\alpha}\right]+ \frac{1}{2}\sum_{k}\ln \left(\frac{\alpha\delta\omega_{k}}{\lambda_{k}}\right ) \right),\end{aligned}\ ] ] with @xmath53 being the eigenvalues of @xmath54 and finally get the most expected spectral function via @xmath55.\end{aligned}\ ] ] it should be noted , that @xmath56 $ ] is formally not integrable due to the saddle point approximation .
however , this becomes only relevant for very large values of @xmath39 .
in any numerically considered interval the function decreases exponentially for precise enough and many data points . for consistency
, the choice for the upper cutoff for @xmath39 should always be quoted . in the `` classical maximum entropy '' ,
the most plausible spectral function @xmath57 is determined by maximizing @xmath46 $ ] and @xmath56 $ ] simultaneously . in our case ,
due to ( in principle ) arbitrarily many data points , this turns out and is known @xcite to agree with the most expected spectral function @xmath22 . in comparison to previous applications , we have formulated the mem for arbitrarily discretized functions @xmath21 and @xmath51 , since we want to deal with ( in principle ) continuous functions from truncated dyson - schwinger calculations .
therefore , the single - value decomposition as proposed in the bryan algorithm @xcite for the numerical determination of @xmath51 does not work .
however , our new treatment of the spectral function opens the possibility of a better suited discretization , which can be adopted to the specific form of @xmath51 . in this way , we are also able to significantly reduce the number of points , which are needed for the numerically discretized spectral function .
as already mentioned , the bayesian approach offers the possibility of an error estimation .
if we consider an interval @xmath58 $ ] of the spectral function , the expectation value of @xmath59 for fixed @xmath39 in this interval for constant weighting is given by @xmath60 and , hence , for the most plausible function by @xmath61.\end{aligned}\ ] ] for the variance around this value , we first consider @xcite @xmath62 with fixed @xmath39 and @xmath63 . therefore in the given interval @xmath64 , we get @xmath65 and @xmath66.\end{aligned}\ ] ] at the end of this section , we want to indicate , how the mem can be applied for massive fermions . for @xmath67 ,
the upper procedure can be performed analogously due to the properties given in eqs .
( [ posrhoe ] ) and eqs.([sumrules ] ) . on the other hand
, we would need to extend the procedure for the whole propagator by utilizing that @xmath68 can be considered as a probability distribution . with
the propagator again denoted as data @xmath21 , we generalize @xmath69 & \rightarrow & \frac{1}{b - a}\,\int_{a}^{b}\!dp_{4 } \frac { \mathrm{tr}\left((d(p_{4})-d[\rho](p_{4}))^{\dagger } ( d(p_{4})-d[\rho](p_{4}))\right ) } { 2\sigma(p_{4})^{2 } } , \nonumber\\ s\left[\rho\right ] & \rightarrow & \int_{-\infty}^{\infty}\!\!\!d\omega\,\,\ , \mathrm{tr}\left(\left ( \rho(\omega)-m(\omega ) -\ln\left(\rho(\omega)m(\omega)^{-1}\right)\rho(\omega ) \right)\gamma_{4}\right),\end{aligned}\ ] ] where the prior estimate @xmath24 is now matrix - valued .
since @xmath68 is hermitian and positive , it can be written as @xmath70 , with @xmath71 and @xmath72 a diagonal matrix with positive eigenvalues .
the unconstrained measure @xmath73 then becomes @xmath74 where @xmath75 is the @xmath76-th eigenvalue and @xmath77 the haar measure for the group @xmath78 at an energy @xmath79 .
however , the path integral is usually constrained by symmetries , i.e. @xmath80 for @xmath81 .
it can be easily seen that the group needs to be only integrated over the factor group of @xmath78 and the conjugate closure of @xmath82 and that only the independent eigenvalues need to be considered .
for the case given by eq .
( [ massiverho ] ) , we obtain with @xmath83 where @xmath84 , simply @xmath85 in the approximation of setting the integrand of the @xmath86-integration to be constant and equal to the value @xmath87 given by the maximization of the functional @xmath88 $ ] , the practical mem procedure again becomes similar to the upper case .
for the numerical determination of the most plausible spectral function , we first need the data @xmath21 with errors @xmath89 on an interval @xmath32 $ ] .
furthermore we have to choose a suitable interval @xmath90 $ ] for the considered part of the spectral function and a prior estimate @xmath24 .
the interval @xmath90 $ ] is usually suggested by the involved scales and can be chosen to be rather large .
the prior estimate @xmath24 is in our case taken as a constant and can be estimated from the sum rules ( eqs .
( [ sumrules ] ) ) , if the spectral function varies only in a certain interval @xmath91 .
it can also be adopted to the knowledge from other methods .
since the main purpose is the implementation of positivity , the results turn out to be comparatively insensitive to its choice .
the advantage of our method is that , after an eventual test calculation , we can choose the discretization , i.e. the abscissas and weights , of the spectral function , such that it is interpolated by a small number of points .
we summarize the procedure as follows : * take the data @xmath21 with errors @xmath89 on an interval @xmath32 $ ] and choose a prior estimate @xmath24 on an interval @xmath90 $ ] . *
determine the maximum of the functional @xmath88=l[\rho]-\alpha s[\rho]$ ] for a fixed value of @xmath39 .
due to its simple form , the marquardt - levenberg method @xcite is very well suited .
* choose a discretized interval for @xmath39 , such that @xmath56 $ ] is strongly peaked .
@xmath47 is determined by normalization . for consistency , the choice for the upper cutoff for @xmath39 should always be quoted . *
calculate @xmath22 and eventually @xmath92 and @xmath93 for a chosen interval @xmath64 .
as an example , we now want to present results for spectral functions of massless quarks in dense matter at vanishing temperature , as they have been determined in @xcite .
we consider the gapped channel in the color - superconducting @xmath94 phase at a quark chemical potential of @xmath95 .
the propagator is then of the form @xmath96 and @xmath97 is related to @xmath98 by eq .
( [ specreppm ] ) . for the bare normal quark propagator with a constant gap @xmath99 , the spectral function
is then given by @xmath100 with @xmath101 and @xmath102 .
we will see below , how a non - trivial @xmath103-dependence generates a finite width . as described in section [ mem ] , the main input for
the mem is the data with a proper error estimate . in dyson - schwinger studies ,
these are obtained by self - consistent solutions of truncated integral equations .
to lowest order , the error of @xmath104 therefore scales with the error of the numerical integrals , which determine the normal and anomal self energies @xmath105 and @xmath106 ( see @xcite for details ) . in our case , we have chosen a simple riemann quadrature for the multidimensional integrals , due to the principle - value - type behavior around the fermi surface for ungapped channels .
the error is therefore of order @xmath107 , where @xmath108 is a scaling factor of the integration mesh . for the error estimation
, we therefore calculate the propagator for two different @xmath108 and extrapolate linearly to @xmath109 .
the data are then taken as the result for the smaller scaling factor @xmath108 and the error as the difference between these data and the extrapolated result .
in addition , the errors around nearest neighbors are averaged in order to avoid ( artifically appearing ) vanishing errors
. we also have to justify , that correlations between the data points are negligible for the likelihood in the form given in eq .
( [ likel ] ) .
since our numerical integrals for different values of @xmath103 are in principle independent , this is assumed to be true , at least when the discretized data is coarser than the numerical integral of the self energies .
the input data for @xmath97 for the following example is chosen on an interval @xmath110 $ ] for the gapped channel in the @xmath94 phase at a quark chemical potential of @xmath95 . as an illustration , the input for momentum @xmath111 is shown in fig .
[ datasplus ] .
we consider the input as continuous due to our many data points and it has small errors of less than @xmath112 in absolute value above @xmath113 . for the data input with given errors , we now need to choose an interval for the spectral function and a non - vanishing prior estimate .
we take the comparatively large interval @xmath114 $ ] .
furthermore we choose an interval for @xmath56 $ ] as discussed in the following subsection . for different prior estimates @xmath115 and @xmath116 at momentum @xmath111 , we then obtain the most expected spectral functions as shown in fig .
[ varym ] .
it turns out , that the extracted spectral function is remarkably insensitive on the variation of the prior estimate , even when varying it by more than three orders of magnitude .
this is mainly related to the small errors of our data .
we therefore fix @xmath117 in the following .
it is also interesting to investigate the @xmath39-dependence of the maximum of the functional @xmath88 $ ] as well as of @xmath56 $ ] .
since @xmath56 $ ] shows a pronounced maximum at @xmath118 , we choose the interval @xmath119 $ ] for @xmath56 $ ] to be non - vanishing and normalized , such that @xmath56>10^{-1}\times p[\alpha_{max}|dh(m)]$ ] for @xmath120 .
again , for momentum @xmath111 , the results are shown in fig .
[ palpha ] . on the left - hand side , we show @xmath56 $ ] , normalized on @xmath64 . on the right - hand side ,
we present the maximum of @xmath88 $ ] for @xmath121 and @xmath122 .
they are only weakly varying , even when comparing the border of the interval @xmath64 to the maximum .
this also substantiates laplace s rule for @xmath123 $ ] a posteriori , assuming that its @xmath39-dependence is weaker . apart from this , in fig .
[ palpha ] , the most expected spectral function is equal to the most probable spectral function given at @xmath118 and therefore not shown .
thus the `` classic maximum entropy '' gives very similar results as `` bryan s method '' in our case of `` many '' data points and small errors .
finally we are able to perform an error estimation around the expected value of a given interval .
since we have two pronounced peaks for quasiparticles and quasiparticle - holes in the spectral function , we choose the intervals associated to their full width at half maximum ( fwhm ) .
the result for momentum @xmath111 is shown on the right in fig .
[ rho2sc ] . again , the errors turn out to be very small . on the left - hand side
we show a contour plot of the spectral density as a function of the energy @xmath124 and momentum @xmath125 .
the light ( online yellow ) line shows the maxima of the quasiparticle and quasiparticle - hole branches . for fixed momentum @xmath125 ,
the difference between the dark ( online blue ) lines below and above the light ( online yellow ) line gives the fwhm of the corresponding peak .
the latter is neglected in bcs - type spectral functions as given in eq .
( [ rhobcs ] ) .
( 10,150 ) ( 0,-2 ) phase at @xmath95 as described in the text ( left ) and the spectral function for momentum @xmath111 with the expectation value within the fwhm and its error estimate as shaded background ( right).,title="fig:",width=245 ] ( 205,150 )
in this paper we have outlined , how the mem can be adapted to numerical dyson - schwinger studies , in particular to fermions . it turns out , that the extracted spectral functions are much more reliable and stable against variation , than in applications of the mem in lattice qcd .
reasons for this are the comparatively small errors on the input functions and the almost arbitrary large number of data points .
comparing however to the systematic error , coming from necessary truncations in dyson - schwinger studies , this error is negligible .
therefore this method can be useful for further applications in mesons or diquarks investigations , since currently all calculations have to be extended to complex momenta ( see @xcite ) . avoiding this and reducing the numerical effort drastically
, the mem might help to improve or extent known truncation schemes .
i thank reinhard alkofer and jochen wambach for helpful discussions as well as a critical reading of the manuscript .
99 r. alkofer and l. von smekal , phys .
rept . * 353 * , 281 ( 2001 ) [ arxiv : hep - ph/0007355 ] . c. d. roberts and s. m. schmidt , prog .
* 45 * , s1 ( 2000 ) [ arxiv : nucl - th/0005064 ] . c. s. fischer , j. phys .
g * 32 * , r253 ( 2006 ) [ arxiv : hep - ph/0605173 ] .
m. jarrell , j. e. gubernatis , phys .
rep . * 269 * , 133 ( 1996 ) .
m. asakawa , t. hatsuda and y. nakahara , prog .
phys . * 46 * , 459 ( 2001 ) [ arxiv : hep - lat/0011040 ] .
f. karsch _ et .
b * 530 * , 147 ( 2002 ) [ arxiv : hep - lat/0110208 ] ; s. datta _ et . al . _ ,
d * 69 * , 094507 ( 2004 ) [ arxiv : hep - lat/0312037 ] .
w. h. press _ et .
_ , _ numerical recipes in c _ , ( cambridge univ . press , cambridge , 1992 ) .
j. skilling , in _ maximum entropy and bayesian methods _ edited by j. skilling ( kluwer academic publishers , dordrecht , 1989 ) , p. 45 .
s. f. gull , in _ maximum entropy and bayesian methods _ edited by j. skilling ( kluwer academic publishers , dordrecht , 1989 ) , p. 53 .
r. k. bryan , eur .
biophys .
j. * 18 * , 165 ( 1990 ) . m. le bellac , _ thermal field theory _ , ( cambridge univ .
press , cambridge , 1996 ) .
d. nickel , j. wambach and r. alkofer , phys .
d * 73 * , 114028 ( 2006 ) [ arxiv : hep - ph/0603163 ] .
r. alkofer , p. watson and h. weigel , phys .
d * 65 * , 094026 ( 2002 ) [ arxiv : hep - ph/0202053 ] .
r. alkofer , w. detmold , c. s. fischer and p. maris , phys .
d * 70 * , 014014 ( 2004 ) [ arxiv : hep - ph/0309077 ] . | it is shown how to apply the maximum entropy method ( mem ) to numerical dyson - schwinger studies for the extraction of spectral functions of correlators from their corresponding euclidean propagators .
differences to the application in lattice qcd are emphasized and , as an example , the spectral functions of massless quarks in cold and dense matter are presented .
maximum entropy method , dyson - schwinger equations |
in recent years , accumulating evidence supports the hypothesis that agn reflects an important stage of galaxy formation .
observations reveals the fact that agn activity and star formation frequently occur together ( e.g. , cid fernandes et al .
2001 ; kauffmann et al . 2003
; see reviews in gonzalez delgado 2002 ) .
an important clue of co - evolution of agn and starburst is the fundamental , tight relationship between mass of supermassive black hole ( smbh ) and the velocity dispersion of bulge where the smbh resides in ( e.g. , magorrian et al .
1998 ; gebhardt et al .
2000 ; tremaine et al .
2002 , greene & ho 2006 ) . by analyzing the spectra of narrow - line agns from sloan digital sky survey ( sdss , york et al .
2000 ) , heckman et al .
( 2004 ) found that most of accretion - driven growth of smbh takes place in the galaxy with relatively young stellar population .
recently , as essential relationships relating to agn s activity , the eigenvector 1 space ( boroson & green 1992 , hereafter bg92 ) was found to be related with ages of the stellar populations as assessed by infrared color @xmath7(60,25 ) ( wang et al . 2006 ) .
it is then natural to ask a question `` what is a young agn ? '' .
mathur ( 2000 ) argued that narrow - line seyfert 1 galaxy ( nls1 ) with high eddington ratio and small black hole mass might be a young agn .
if so , the co - evolution of agn and starburst implies that nls1s are expected to be associated with relatively young stellar populations . to understand the elusive agn - starburst connection
, we need to carefully examine the properties of agn s host galaxy
. however , optical spectroscopic study is possible for only a few type i agns because the strong nuclear emission usually masks the relatively faint stellar components . by fitting deep off - nuclear optical spectra canalizo & stockton ( 2001 ) and nolan et al .
( 2001 ) determined the ages of circumnuclear stellar populations for a few ir and optical - selected qsos respectively . in the happy case ,
brotherton et al .
( 1999 ) discovered a spectacular `` post - starburst quasar '' unj1025 - 0040 with a @xmath3400 myr old stellar population .
wang et al . ( 2004 ) identified sdssj022119.84 + 005628.4 as a nls1 ( [email protected] ) whose spectrum displays a clear balmer jump .
however , the authors did not perform further stellar population analysis . in this paper
, we perform a detailed examination of the physical properties of post - starburst agn sdssj085338.27 + 033246.1 and discuss its implications on study of agn s evolution .
the optical spectrum is extracted from sdss data release 4 ( adelman - mccarthy et al . 2006 ) .
its appropriate spectrum allows us to separate it into a young stellar population and a typical nls1 spectrum .
in addition , the emission line profiles allows us to determine the location on the line ratio diagnostic diagrams .
we begin with a brief description of the data reduction in 2 .
the results and implications are presented in 3 .
the cosmological parameters : @xmath9 , @xmath10 , and @xmath11 are adopted for calculations .
we extract the optical spectrum of sdssj085338.27 + 033246.1 when we carry out a systematic search for broad line agns from the objects listed in the catalogues released by kauffmann et al .
( 2003 ) and heckman et al .
totally hundreds of such objects are extracted from the parent sample .
the details of the search and properties of the sub - sample will be presented in subsequent papers . among these objects sdssj085338.27 + 033246.1 has unique properties with strong enough typical agn spectrum superposed on a spectrum of a - type star .
it is coincident that we find the object has been identified as a member of a sample of 74 post - starburst broad line agns by zhou et al .
( 2005 ) when our paper is being prepared .
the observed spectrum is smoothed with a boxcar of 3 pixels ( @xmath35 ) to enhance s / n ratio and precision of the spectra measurements .
standard iraf procedures are adopted to reduce the raw data , including galactic extinction correction ( with @xmath12 from nasa / ipac extragalactic database ( ned ) and assuming @xmath13 ) and de - redshift to rest frame with @xmath14 .
the total light spectrum at rest frame is displayed in figure 1 ( the second line from top to bottom ) .
the spectrum clearly shows a prominent balmer jump at blue even diluted by an agn continuum and high order balmer series . in the next step
, the principal component analysis ( pca ) method is developed to remove the stellar light from the observed spectrum ( e.g. , li et al . 2005 ; hao et al 2005 ) .
we build a library of stellar absorption spectra by applying pca technique on standard ssp models developed by bruzual & charlot ( 2003 , hereafter bc03 ) .
the first seven eigenspectra are used to model the starlight component by their linear combination . in order to appropriate model the starlight component ,
our template contains the seven eigenspectra , a power - law continuum , a feii complex of agn ( adopted from bg92 ) and a galactic extinction curve ( cardelli et al .
1989 ) . in order to avoid the distortion of low s / n
, a @xmath15 minimizing is performed over the rest wavelength range from @xmath163700 to @xmath166800 , except for the regions around the strong emission lines .
the removal of the stellar component is illustrated in figure 1 as well .
also plotted are the modeled feii blends and emission - line spectrum of the agn .
the flux of the feii blends measured between rest frame wavelength @xmath164434 and @xmath164684 is @xmath17 .
the detailed comparison between the observed and modeled spectrum at blue is shown in the insert plot a. it is clear that the modeled spectrum reproduces the observed absorption features rather well .
the agn emission lines are modeled by specfit task as described in wang et al .
( 2005 , 2006 ) .
briefly , each narrow line ( e.g. , h@xmath7 , h@xmath18 , [ nii ] ) is modeled by a gaussian component
. the intensity ratio of [ nii ] doublets is forced to equal to the theoretical prediction .
the spectra modeling is schemed in the insert plot b and c for h@xmath18 and h@xmath7 region , respectively .
the upper limit of [ oiii]@xmath165007 flux @xmath19 } ) < 1.2\times10^{-16}\ \rm{ergs\ s^{-1}\ cm^{-2}}$ ] is estimated by integration rather than fit because the feature is very faint .
in addition to these lines , the spectrum displays a broad emission feature at about 5880 . based on its asymmetric profile
, we identified this feature as hei@xmath165876 contaminated by broad naid@xmath205890 , 5896 emission .
we find ew(hei+naid)@xmath21 and ( hei+naid)/h@xmath22 .
the naid emission was detected in a few agns ( e.g. , veron - cetty et al .
2004 ; veron - cetty et al . 2006 and references therein , thompson 1991 ) .
the calculations indicated that the existence of large naid emission ( naid / h@xmath7=0.01 - 0.05 ) is related to models with high density ( @xmath23 ) and large column density ( @xmath24 ) ( thompson 1991 ) .
the measured line properties are summarized in table 1 .
the propagation of error is included in the uncertainties shown in parentheses .
all the uncertainties given in columns ( 2 ) and ( 3 ) are caused by profile modeling .
according to its spectra properties , we classify sdssj085338.27 + 033246.1 as a nls1 both because of the extremely faint [ oiii ] emission ( [ oiii]/[email protected] ) and because of the relatively narrow h@xmath18 broad component ( fwhm(h@xmath26)=@xmath27 ) .
moreover , the measured rfe is as large as @xmath28 , which is typical of nls1s .
we further calculate its black hole virial mass and eddington ratio in terms of the h@xmath7 broad component according to the equations in greene & ho ( 2005 ) : @xmath29 @xmath30 the intrinsic luminosity @xmath31 is used for calculations because of the large intrinsic extinction @xmath32 .
the extinction is calculated from the line ratio h@xmath33/h@xmath34 , assuming a galactic extinction curve and @xmath13 ( osterbrock 1989 ) . the inferred virial mass and eddington ratio is @xmath2 and @xmath1 , respectively .
it is clear that all these properties ( i.e. , small black hole mass , large rfe value and large eddington ratio ) indicate the object should be potentially classified as a nls1 . the radio radiation at 1.4ghz of the object was detected by nvss ( condon et al .
1998 ) and first survey ( becker et al . 1995 ) .
the map of the first survey shows an unresolved source with a flux @xmath34.04mjy and position @xmath35 , @xmath36 ( j2000 ) .
the discrepancy in position given by the optical and radio surveys is much less than 1 ( @xmath30.3 ) .
we determined the extinction - corrected continuum flux that is contributed by agn at rest wavelength @xmath37 to be @xmath38mjy . if the radio emission is mainly due to agn , the inferred radio loudness @xmath39 defined as @xmath40 ( kellermann et al .
1989 ) is @xmath41=1.40 corresponding to @xmath42=1.12 for 5ghz radio flux assuming a spectral shape @xmath43 from optical to radio band . ] . adopting the division @xmath44 between radio loud and quiet agn , sdssj085338.27 + 033246.1
is an object with intermediately strong radio emission .
we further calculate its radio luminosity @xmath45 .
this luminosity has same order of magnitude of the most radio - luminous starburst ( @xmath46 , smith et al .
it has long been known that the decimeter radio luminosity emitted from star - forming galaxy traces supernova rate of massive enough stars ( i.e. , @xmath47 ) .
the lifetime of these massive supernova progenitors is @xmath48yr .
the radio supernova itself and its remnant has lifetime @xmath49yr and @xmath50yr , respectively .
the average sfr over the past @xmath51yr therefore relates with the radio emission attributed to supernova as ( condon 1992 ) @xmath52 for the object , the estimated average sfr is @xmath53 ( see sect . 3.2 for details ) . using a salpeter imf ( salpeter 1955 )
, this corresponds to a @xmath54 .
the inferred luminosity is therefore @xmath55 and is about one order lower than the total radio luminosity .
the `` a - type star '' spectrum indicates the age of the stellar population is no more than a few@xmath56 yrs . to interpret the spectrum of the young stellar population , we extract a series of spectra of ssp models of bc03 with a solar metallicity and with a chabrier initial mass function .
the extracted spectra are matched with the observed spectrum one by one within the wavelength range 3500 - 5000 .
the spectrum red of @xmath165000 is not used because this part is obviously contaminated by an old population .
figure 2 illustrates the best match between the observed spectrum and a 0.1gyr old ssp model .
the model predicted mass of the starburst is @xmath57 .
this simple ssp model shows a past average sfr @xmath58 .
however , recent studies suggested that the agn and starburst activity perhaps not co - evolve simultaneously .
schmitt ( 2001 ) suggested that the starbursts may predominate over agns in their earlier phase .
yan et al .
( 2005 ) recently suggested that most of sdss post - starburst galaxies hold somewhat agn like activities .
the post - starburst is the phase in which the galaxy already ceased star formation at some recent epoch .
the derived @xmath59gyr post - starburst in sdssj085338.27 + 033246.1 is somewhat similar with the @xmath60gyr post - starburst that was obtained by brotherton et al .
( 1999 ) in unj1025 - 0040 .
[ oii]@xmath163727 line is widely used as an empirical indicator of ongoing sfr ( @xmath61yr ) for emission line galaxy surveys ( e.g. , gallagher et al .
1989 ; kennicutt 1992 ; hippelein et al .
its practicability led to several calibrations and a great deal discussions ( e.g. , kewley et al .
2004 ; kennicutt 1998 ; yan et al .
the extinction - corrected flux of [ oii]@xmath62 doublets is @xmath63 ( with its observed flux shown in table 1 ) , which yields a ratio @xmath64([oii]/[oiii])@xmath650.55 . adopting the ratio @xmath64([oiii]/h@xmath66)@xmath67 - 0.65
, the object is consequently located in the region that represents star formation ( kim et al .
2006 and references therein ) .
it means that the [ oii ] emission ought to be almost entirely attributed to star formation ( see below ) .
we use the calibration @xmath68,42}}}{16.73 - 1.75[\log(\mathrm{o / h})+12]}m_{\odot}\rm{yr^{-1}}\ ] ] ( kewley et al . 2004 ) to estimate the sfr , where @xmath69,42}$ ] is the luminosity of [ oii ] emission in units of @xmath70 .
the sfr is inferred to be round about 3.0 @xmath71 by adopting the assumption used in ho ( 2005 ) , i.e. , the metallicity is twice of solar corresponding to @xmath72 . here , we obtain a residual but quenched ongoing sfr as obtained in the optically selected qsos by ho ( 2005 ) .
the distance - independent birthrate parameter @xmath73 is defined as the ratio of the current sfr to the average past one ( kennicutt et al .
1994 ) . as a particular case ,
the object has extremely small value of @xmath74 , i.e. , has present sfr significantly lower than the past .
the traditional bpt diagram ( baldwin et al . 1981 ) is a powerful tool for diagnosing the origin of emission of narrow lines for emission - line galaxies .
the location of the object in the bpt diagram in which the line ratio [ oiii]/h@xmath18 is plotted against [ nii]/h@xmath7 is marked by solid square in figure 3 ( left panel ) .
the solid and dashed line shows the demarcation line between agn and starburst galaxy defined by kewley et al .
( 2004 ) and kauffmann et al .
( 2003 ) , respectively .
the figure shows clearly that the position of the object is marginally above the threshold used to define representative hii region .
the similar plot is shown in figure 3 ( right panel ) but for [ oiii]/h@xmath18 v.s . ( [ sii]@xmath166719 + 6731)/h@xmath7 . in this case , the position of the object falls into the region that represents starburst and is far below the demarcation line .
the combination of these two diagnosis suggests the observed narrow emission lines are mainly produced by star - formation .
what is the distribution in the bpt diagram for post - starburst agns is a further interesting question . here
it is interesting to give more discussions on the elusive agn - starburst connection .
wang et al . ( 2006 ) recently indicated that the well - documented e1 sequence most likely represents an evolutionary track of agn .
the track further implies that agn with high eddington ratio evolves to one with low eddington ratio .
this result also naturally implies that nls1s that occupy one extreme end of e1 should be associated with relatively young stellar populations .
apart from the evidence mentioned above , for the ir - selected qsos , canalizo & stockton ( 2001 ) found the ages of circumnuclear stellar populations range from current star formation to @xmath75myr old .
subsequent spectroscopy studies indicated that about 70%-100% ir - selected qsos are extremely or intermediately strong feii emitters ( e.g. , lipari et al .
2003 ; zheng et al .
moreover , zhou et al .
( 2005 ) found that more than half of the post - starburst type i agn fulfill the criterion for nls1 .
all these evidence appear to reveal that a young agn is most likely a strong feii emitter with relatively narrow h@xmath18 profile .
in contrast , old stellar populations ( @xmath76gyr ) are found to dominate the off - nuclear ( @xmath77 ) stellar population for optical - selected qsos ( nolan et al .
2001 and references there in ) .
these accumulating clues lead us to propose a possible evolutionary scenario that links both agn and starburst phenomena
. the massive starburst might be dominant in the earlier evolutionary phase .
then the starburst passively evolves to a - type stars when the plenty gas falls into the center of a galaxy under the gravitational attraction of smbh . at one stage about @xmath78yr after the beginning of the starburst , the smbh begin to accrete matter with small black hole mass and high eddington ratio .
the feedback of the agn suppresses circumnuclear star formation at the same time ( e.g. , kim et al .
the black hole mass grow substantially in the formed bulge in this short phase ( e.g. mathur et al .
2001 ; mathur & grupe 2005 ) .
the duration of the phase is commonly estimated as @xmath79-folding time scale @xmath80yr ( salpeter 1964 ) , where @xmath81 is the radiative efficiency .
the differential growth of agn and starburst allows the nls1s below the @xmath82 relation for inactive galaxies to approach normal @xmath82 relation ( mathur & grupe 2005 ; mathur et al .
2001 ; wandel 2002 ; bian & zhao 2004 ) .
the agn then shines with decreasing accretion rate and insignificantly increasing black hole mass because of the consumption of the gas .
the scenario implies that the agn and host co - evolve marginally around the common @xmath82 relation .
the luminous agn phase persists for typical agn lifetime which is usually believed to be @xmath83 yr with however large uncertainties ( see reviews in matini 2004 ) . during this phase , the young stellar population continuously ages and
its importance gradually fades .
the underlying old stellar population ( a few gyr ) finally becomes dominant in the emission of host of an old agn .
in fact , in addition to the old stellar populations , relatively old post - starbursts ( 0.1 - 2gyr ) are frequently detected in off - nuclear regions of a few powerful radio - loud agns ( e.g. tadhunter et al .
2005 and references therein ) .
after the luminous phase , agn then passively evolves to a stage with low luminosity , or low radiative efficiency . in summary
, this scenario could be described as following path : starburst @xmath84 nls1+post - starburst @xmath84 luminous agn with decreased eddington ratio+old stellar population @xmath84 less luminous agn .
we believe future studies of agn+post - starburst composite objects such as sdssj085338.27 + 033246.1 are important for developing a rational scenario to interpret the agn - starburst connection .
the physical properties of post - starburst agn sdssj085338.27 + 033246.1 are derived by analyzing its optical spectrum .
it allows us to obtain results as following .
the object can be identified as a nls1 ( ( [ oiii]/h@xmath85 , @xmath86 and [email protected] ) associated with a post - starburst stellar population as identified from the size of the balmer jump .
a simple ssp model indicates the starburst with a mass @xmath87 possibly took place @xmath88 yr ago .
the [ oii ] inferred sfr is as small as @xmath33.0@xmath89 which is much smaller than the past average one @xmath6 .
its locations in the bpt diagrams show the hii region significantly contributes to the emission of the narrow lines .
we further discuss its implications on the elusive agn - starburst connection .
we would like to thank the anonymous referee for very useful comments and important suggestions .
we thank d. w. xu , c. n. hao , j. s. deng and c. cao for valuable discussions .
we are grateful to todd a. boroson and richard f. green for providing us the feii template .
this work was supported by the national science foundation of china ( grant 105030005 and 10473013 ) .
this research has made use of the nasa / ipac extragalactic database , which is operated by jpl , caltech , under contact with the nasa .
the sdss archive data is created and distributed by the alfred p. sloan foundation .
adelman - mccarthy , j. k. , agueros , m. a. , allam , s. s. , et al . , 2006 , , 162 , 38 baldwin , j. a. , phillips , m. m. , & terlevich , r. , 1981 , pasp , 93 , 5 becker , r. h. , white , r. l. , & helfand , d. j. , 1995 , , 450 , 559 bain , w. , & zhao , y. , , 347 , 607 boroson , t. a. , & green , r. f. , 1992 , , 80 , 109 brotherton , m. s. , van breugel , wil . , stanford , s. a. , et al . , 1999 , , 520 , l87 bruzual . g. , & charlot . s. , 2003 , , 344 , 1000 canalizo , g. & stockton , a. , 2001 , , 555 , 719 cardelli , j. a. , clayton , g. c. , & mathis , j. s. , 1989 , , 345 , 245 cid fernandes , r. , heckman , t. , schmitt , h. , et al . , 2001 , , 588 , 81 condon , j. j. , ara&a , 30 , 575 condon , j. j. , cotton , w. d. , greisen , e. w. , et al . , 1998 , , 115 , 1693 gallagher , j. s. , hunter , d. a. , & bushouse , h. , 1989 , , 97 , 100 gebhardt , k. , bender , r. , bower , g. , et al .
2000 , , 539 , l13 gonzalez delgado , r. , 2002 , aspc , 258 , 101 greene , j. e. , & ho , l. c. , 2005 , , 630 , 122 greene , j. e. , & ho , l. c. , 2006 , , 641 , l21 hao , l. , strauss , m. a. , tremonti , c. a. , et al .
2005 , , 129 , 1783 heckman , t. m. , kauffmann , g. , brinchmann , j. , et al .
, 2004 . , 613 , 109 hippelein , h. , maier , c. , meisenheimer , k. , et al .
2003 , , 402 , 65 ho , l. c. , 2005 , , 629 , 680 kauffmann , g. , heckman , t. m. , tremonti , c. , et al . , 2003 , , 2003 , 346 , 1055 kellermann , k. i. , sramek , r. , & schmidt , m. , 1989 , , 98 , 1195 kennicutt , r. c. , 1992 , , 388 , 310 kennicutt , r. c. , tamblyn , p. , & congdon , c. e. , 1994 , , 435 , 22 kennicutt , r. c. , 1998 , ara&a , 36 , 189 kewley , l. j. , geller , m. j. , & jansen , r. a. , 2004 , , 127 , 2002 kim , m. , ho , l. c. , & i m , m. , 2006 , astro - ph/0601316 li , c. , wang , t. g. , zhou , h. y. , et al .
2005 , , 129 , 669 lipari , s. , terlevich , r. , diaz , r. j. , et al .
, 2003 , , 340 , 289 magorrian , j. , tremaine , s. , richstone , d. , et al . , 1998 , , 115 , 2285 martini , p. , 2004
, in coevolution of black holes and galaxies , ed . , l. c. ho ( cambridge univ . press ) , 170 mathur , s. , 2000 , , 314 , l17 mathur , s. , kuraszkiewicz , j. , & czerny , b. , 2001 , newa , 6 , 321 mathur , s. , & grupe , d. , 2005 , , 432 , 463 nolan , l. a. , dunlop , j. s. , kukula , m. j. , et al .
2001 , , 323 , 308 osterbrock , d. e. , 1989 , astrophysics of gaseous nebulae and active galactic nuclei , mill valley ca : university science books osterbrock , d. e. , & pogge , r. , 1985 , , 197 , 166 salpeter , e. e. , 1955 , , 121 , 161 salpeter , e. e. , 1964 , , 140 , 796 schmitt , h. r. , 2001 , , 122 , 2243 smith , d. a. , herter , t. , & haynes , m. p. , 1998
, , 494 , 150 tadhunter , c. , robinson , t. g. , gonzalez - delgado , r. m. , wills , k. , & morganti , r. , 2005 , , 356 , 480 thompson , k. l. , 1991 , , 374 , 496 tremaine , s. , gebhardt , k. , bender , r. , et al . , 2002 , , 574 , 740 veron - cetty , m. -p . , joly , m. , veron , p. , 2004
, , 417 , 515 veron - cetty , m. -p . , joly , m. , veron , p. , et al .
, 2006 , astro - ph/0602239 wandel , a. , 2002 , , 565 , 762 wang , j. , wei , j. y. , & he , x. t. , 2004 , chjaa , 4 , 415 wang , j. , wei , j. y. , & he , x. t. , 2005 , , 436 , 417 wang , j. , wei , j. y. , & he , x. t. , 2006 , , 638 , 106 yan , r. b. , newman , j. a. , faber , a. m. , et al .
2006 , astro - ph/0512446 york , d. g. , et al .
2000 , , 120 , 1579 zheng , x. z. , xia , x. y. , mao , s. , wu , h. , deng , z. g. , 2002 , , 124 , 18 zhou , h. y. , wang , t. g. , dong , x. b. , wang , j. , lu , h. , 2005 , mem , s. a. it 76 , 93 cccc line identification & flux ratio & fwhm + & & ( @xmath90 ) + h@xmath66 & 1.0 & @xmath91 + h@xmath26 & @xmath92 & @xmath93 + @xmath94\lambda$]5007 & @xmath95 & + h@xmath33 & @xmath96 & @xmath97 + h@xmath98 & @xmath99 & @xmath100 + @xmath101\lambda6583 $ ] & @xmath102 & @xmath103 + feii 4570 & @xmath104 & + @xmath105\lambda$]3727 & @xmath106 & + @xmath107\lambda$]6716 & @xmath108 & @xmath109 + @xmath107\lambda$]6731 & @xmath110 & + hei+naid & @xmath111 & + | we analyze the physical properties of post - starburst agn sdssj085338.27 + 033246.1 according to its optical spectrum and discuss its implications on agn s evolution .
the spectra pca method is developed to extract emission lines and absorption features from the total light spectrum .
the emission - line analysis indicates that the object can be classified as a nls1 with feii / h@xmath0 , large eddington ratio ( @xmath1 ) , small black hole mass ( @xmath2 ) and intermediately strong radio emission .
a simple ssp model indicates that the absorption features are reproduced by a @xmath3100 myr old starburst with a mass of @xmath4 rather well .
the current sfr @xmath5 inferred from the [ oii ] emission is much smaller than the past average sfr@xmath6 , however .
the line ratio diagnosis using the bpt diagrams indicates that the narrow emission lines are almost entirely emitted from hii regions .
we further discuss a possible evolutionary path that links agn and starburst phenomena . |
the subject of quantum magnetism dates back to 1920s .
e. ising@xcite suggested a simplest model of a magnet as a collection of @xmath8 spins which may acquire two values @xmath9 and interact with nearest neighbors on a lattice as @xmath10 and with an external magnetic field as @xmath11 .
to explain the properties of the model we have to calculate the partition function @xmath12 which yields the helmholtz free energy per site @xmath13 ( in what follows we set @xmath14 to simplify the notations ) . in one dimension
the problem was solved by e. ising .
later l. onsager solved the square - lattice ising model@xcite and we know the solution in two dimensions@xcite .
there is no solution of the ising model in three dimensions until now .
another version of interspin interaction was suggested by p. a. m. dirac and w. heisenberg .
the heisenberg exchange interaction reads @xmath15 where the pauli matrices @xmath16 are defined as @xmath17 denoting the halves of the pauli matrices as @xmath18 ( in what follows we set @xmath19 to simplify the notations ) we consider the following hamiltonian @xmath20 we note that the hamiltonian of the anisotropic @xmath21 heisenberg model ( [ 1.02 ] ) covers in the limiting cases some specific models like the ising model ( @xmath22 ) , the isotropic @xmath6 ( or @xmath1 or @xmath23 ) model ( @xmath24 , @xmath25 ) , the anisotropic @xmath6 model ( @xmath26 , @xmath25 ) , the isotropic ( @xmath27 ) heisenberg model ( @xmath28 ) , and the heisenberg - ising ( @xmath29 ) model ( @xmath30 , @xmath31 ) .
again we would like to calculate the partition function @xmath32 of the spin-1/2 model ( [ 1.02 ] ) .
unfortunately , this task is very complicated even in one dimension . due to h. bethe we know how to find the eigenstates of the spin-1/2 linear chain heisenberg model@xcite .
the famous bethe ansatz for the wave function has the form @xmath33 where the sum in the definition of coefficients @xmath34 , @xmath35 , runs over all @xmath36 permutations of the labels @xmath37 , @xmath38 is the image of @xmath39 under the permutation @xmath40 . for further details
see , e.g. , ref .
@xcite .
let us briefly recall the quantities of interest in the statistical mechanical studies of the spin models . as we have mentioned already the thermodynamic quantities like the entropy , the specific heat , the magnetization etc .
follow from the partition function @xmath41 , the sum runs over all states @xmath42 of the system with energy @xmath43 .
usually we are also interested in the equal - time spin correlation functions , e.g. @xmath44 , @xmath45 ; their nonzero limiting values , ( e.g. , @xmath46 ) may indicate the existence of long - range order in the system . within a linear response regime
we add to the hamiltonian @xmath47 a small perturbation @xmath48 , where the external field @xmath49 couples to the dynamical variable @xmath50 , and observe a response of a dynamical variable @xmath51 , @xmath52 with @xmath53\rangle_0 $ ] ( here @xmath54 is the heaviside step function ) .
the fourier - transform of the dynamic susceptibility @xmath55 , @xmath56 , is the quantity which can be measured experimentally .
we note that the real and imaginary parts of the dynamic susceptibility are connected via the dispersion ( or kramers - kronig ) relation . on the other hand ,
the imaginary part of the dynamic susceptibility can be expressed with the help of the fluctuation - dissipation theorem through another dynamic quantity , the dynamic structure factor .
thus , @xmath57 .
usually , the operator @xmath51 is constructed from the local operator of the considered system @xmath58 as follows : @xmath59 .
we can also rewrite the dynamic structure factor in the following forms @xmath60 sometimes it is convenient to make the following change in the first line in eq .
( [ 1.04 ] ) : @xmath61 , @xmath62 .
we also note that in the zero - temperature limit @xmath63 ( or @xmath64 ) the second line in eq .
( [ 1.04 ] ) becomes simpler , @xmath65 , @xmath66 . in what follows
we discuss mainly the dynamic properties of spin-1/2 @xmath6 chains ; just for this class of spin models application of the jordan - wigner fermionization approach is most fruitful .
we notice here that recently it has been found that cs@xmath67cocl@xmath68 is a good realization of the spin-1/2 @xmath1 chain@xcite and calculations of the dynamic quantities for the corresponding spin models might be important for the interpretation of the data from dynamic experiments@xcite . as an example of earlier studies we may mention dynamic experiments on the spin-1/2 @xmath1 chain compound prcl@xmath69@xcite .
the rest of this chapter is organized as follows .
at first we briefly introduce the jordan - wigner transformation ( sec .
[ secdk2 ] ) and concisely discuss some of its generalizations ( sec .
[ secdk3 ] ) . then we consider in detail the dynamic structure factors for the spin-1/2 isotropic @xmath6 chain in a transverse field distinguishing the quantities which probe two - fermion , four - fermion and many - fermion excitations ( sec .
[ secdk4 ] ) .
next we examine the dynamics for two slightly more complicated chains : the dimerized isotropic @xmath6 chain ( sec .
[ secdk5 ] ) and the @xmath6 chains with the dzyaloshinskii - moriya interaction ( sec .
[ secdk6 ] ) .
the results obtained for one - dimensional @xmath6 spin models do not involve any approximation .
this is not true in the two - dimensional case for which the jordan - wigner approach provides only approximate expressions for dynamic quantities .
we illustrate the jordan - wigner fermionization approach in two dimensions examining some dynamic quantities for the square - lattice spin-1/2 isotropic @xmath6 model ( sec .
[ secdk7 ] ) .
we end up with a brief summary ( sec .
[ secdk8 ] ) . to be specific , we consider the one - dimensional spin @xmath0 @xmath29 heisenberg chain with the hamiltonian @xmath70 we imply either periodic or open boundary conditions in eq .
( [ 2.01 ] ) . here
the spin operators @xmath71 satisfy the commutation relations @xmath72 = { \rm{i}}\delta_{ij}\epsilon_{\alpha\beta\gamma}s_i^{\gamma}$ ] , @xmath73 is the totally antisymmetric levi - civita tensor with @xmath74 .
in particular , @xmath75 = { \rm{i}}\delta_{ij}s_i^z$ ] etc . obviously , @xmath76 can be viewed as the halves of the pauli matrices ( [ 1.01 ] ) . after introducing the spin raising and lowering operators ( or the ladder operators ) @xmath77 ( @xmath78 , @xmath79 )
the hamiltonian ( [ 2.01 ] ) becomes @xmath80 we note that the spin raising and lowering operators satisfy commutation relations of fermi type at the same site , i.e. @xmath81 and of bose type at different sites @xmath82 = \left[s_n^-,s_m^-\right ] = \left[s_n^+,s_m^+\right]=0 , \;\;\ ; n\ne m. \label{2.04}\end{aligned}\ ] ] we may use the jordan - wigner transformation@xcite to introduce fermi operators according to the following formulas @xmath83 @xmath84 ( sometimes one can find in eqs . ( [ 2.05 ] ) , ( [ 2.06 ] ) instead of @xmath85 the identical expressions @xmath86 . )
really , the operators introduced always satisfy the fermi commutation relations @xmath87 ( to check this one has to note that @xmath88 and that @xmath89 . ) the inverse transformation to the one given by eqs .
( [ 2.05 ] ) , ( [ 2.06 ] ) reads @xmath90 @xmath91 moreover , the hamiltonian ( [ 2.02 ] ) in terms of the fermi operators ( [ 2.05 ] ) , ( [ 2.06 ] ) has the following form @xmath92 ( we use @xmath93 etc . ) . in the case of periodic boundary conditions implied for the spin hamiltonian ( [ 2.02 ] ) the transformed hamiltonian ( [ 2.10 ] ) obeys either periodic or antiperiodic boundary conditions depending on the parity of the number of fermions .
however , in what follows the calculated quantities in the thermodynamic limit @xmath94 will be insensitive to the boundary conditions implied ( for further details see ref .
@xcite ) . from eq .
( [ 2.10 ] ) it becomes clear that the spin-1/2 isotropic @xmath6 chain in a transverse ( @xmath2 ) magnetic field with the hamiltonian @xmath95 in the jordan - wigner picture is represented by the hamiltonian @xmath96 and therefore is an exactly solvable model@xcite .
moreover , the @xmath6 exchange interaction may be anisotropic ; then the intersite interaction has the form @xmath97 with @xmath98 , @xmath99 .
we can also consider an additional intersite interaction , the so - called dzyaloshinskii - moriya interaction , which does not spoil a simple fermionic hamiltonian@xcite @xmath100 moreover , within the jordan - wigner fermionization approach we can examine rigorously some types of multi - spin interactions@xcite , for example , @xmath101 within the frames of the jordan - wigner approach we can also generalize simple spin-1/2 @xmath6 chains assuming regularly alternating hamiltonian parameters@xcite or some types of random hamiltonian parameters@xcite and still face exactly solvable models . on the other hand , as can be easily seen from eq .
( [ 2.10 ] ) the ising interaction between @xmath2 spin components leads to interacting spinless fermions and as a result the advantages of fermionization are less evident .
( obviously , we can split the interaction term in the spirit of the hartree - fock approximation@xcite , however , the resulting theory will be only an approximate one .
on the other hand , in the low - energy limit we can bosonize the fermionic hamiltonian obtaining an exact low - energy effective theory@xcite . )
we can not examine rigorously within the jordan - wigner fermionization approach the case of the next - nearest - neighbor interaction since @xmath102 it is worthwhile to note here that recently the jordan - wigner fermionization approach has been applied to the spin-1/2 isotropic @xmath6 model on a diamond chain@xcite , however , the authors of that paper apparently missed some interaction terms in the fermionic hamiltonian and their statement about rigorous results for such a model is wrong .
finally we note that an external magnetic field directed along @xmath103 or @xmath104 axes has an enormously complicated form in the jordan - wigner picture .
the jordan - wigner fermionization is a powerful tool for the study of quantum spin chains . since the late 1980s there were several attempts to extend this approach to two ( and three ) dimensions@xcite as well as to spin values @xmath105@xcite . for a review on the two - dimensional jordan - wigner fermionization approach see also ref .
@xcite .
bearing in mind the jordan - wigner transformation in one dimension as a guideline we consider in the two - dimensional case the following relation between spin @xmath0 and fermi operators @xmath106 here @xmath107 , @xmath108 are the fermi operators , the operators @xmath109 defined according to ( [ 3.01 ] ) commute at different sites if the c - number matrix @xmath110 satisfies the relation @xmath111 there are many choices of the matrix @xmath110 which realize the two - dimensional jordan - wigner transformation . following y. r.
wang@xcite we use the cartesian coordinates @xmath112 to construct a complex number @xmath113 and then choose @xmath114 indeed , for such a choice eq .
( [ 3.02 ] ) is satisfied , @xmath115 .
another choice of the matrix @xmath110 has the following form@xcite @xmath116 here @xmath54 is the heaviside step function ( see also ref .
@xcite ) . after performing the jordan - wigner transformation ( [ 3.01 ] ) for the two - dimensional spin-1/2 @xmath29 heisenberg hamiltonian one
gets @xmath117 with @xmath118 ( we have used eq .
( [ 3.03 ] ) for @xmath119 ( [ 3.01 ] ) ) .
we need further approximations to proceed with statistical mechanics calculations for the hamiltonian ( [ 3.05 ] ) . within the mean - field description one assumes @xmath120 .
we expect such an approximation to be valid in the case of zero magnetic field . for the mean - field description in the case of nonzero magnetic field and
an analysis of the magnetization processes in the spin system see ref .
we also note that a more sophisticated ( self - consistent site - dependent ) mean - field treatment has been suggested as well@xcite . after adopting the mean - field approach
we face the problem of particles in a magnetic field with the flux per elementary plaquette @xmath121 .
we may change the gauge preserving the flux per elementary plaquette to make the hamiltonian more convenient for further calculations .
for example , for a square lattice we have @xmath122 in the one - dimensional case when either vertical or horizontal bonds vanish the hamiltonian ( [ 3.07 ] ) transforms into eq .
( [ 2.10 ] ) ( with @xmath123 ) .
recently a. kitaev has suggested a new exactly solvable two - dimensional quantum spin model@xcite .
this is a spin-1/2 model on a honeycomb lattice with interactions between different components of neighboring spins along differently directed bonds .
an alternative representation of the honeycomb lattice is a brick - wall lattice ( see fig .
[ fig01 ] ) .
the hamiltonian of the model reads @xmath124 @xmath39 and @xmath125 denote the column and row indices of the lattice .
we discuss in what follows a fermionic representation for the kitaev model@xcite .
let us perform the jordan - wigner transformation @xmath126 ( compare with eqs .
( [ 3.01 ] ) , ( [ 3.04 ] ) ) . as a result
we find that @xmath127 next we introduce the following operators @xmath128 in terms of these operators the hamiltonian reads as follows @xmath129 since @xmath130 are good quantum numbers the hamiltonian ( [ 3.12 ] ) corresponds to a model of spinless fermions with local static @xmath131 gauge fields .
thus , eq . ( [ 3.12 ] ) explains a hidden simple structure of the spin model ( [ 3.08 ] ) .
the generalizations of the jordan - wigner transformation for arbitrary spin values were discussed by several authors@xcite , however , these mappings have not yet provided a substantial break - through for difficult strongly correlated problems . [
[ spin-12-isotropic - xy - chain - in - a - transverse - field - dynamic - quantities - secdk4 ] ] spin-1/2 isotropic @xmath6 chain in a transverse field : dynamic quantities [ secdk4 ] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ we start with the simplest spin-1/2 @xmath6 model , the transverse @xmath1 chain , with the hamiltonian ( [ 2.11 ] ) . after performing the jordan - wigner transformation
we arrive at a tight - binding model for spinless fermions ( [ 2.12 ] ) and after performing the fourier transformation , @xmath132 ( @xmath133 if the number of fermions is odd or @xmath134 if the number of fermions is even , @xmath135 if @xmath8 is even or @xmath136 if @xmath8 is odd ) , the hamiltonian ( [ 2.12 ] ) becomes diagonal @xmath137 as it has been mentioned above , for the analytical calculations discussed below we may consider only periodic boundary conditions for the fermionic hamiltonian ( i.e. @xmath133 in eq .
( [ 4.01 ] ) ) .
we begin with the transverse dynamic structure factor @xmath138 ( [ 1.04])@xcite .
the calculation of the @xmath3 time - dependent spin correlation function is straightforward . after exploiting the jordan - wigner transformation
we have @xmath139 . here
@xmath140 and @xmath141 .
next we have to use the wick - bloch - de dominicis theorem , @xmath142 , and to calculate the elementary contractions introducing the fermi function @xmath143 , @xmath144 , @xmath145 . as a result , the final expression for the @xmath3 time - dependent spin correlation function reads @xmath146 plugging eq .
( [ 4.02 ] ) into eq . ( [ 1.04 ] ) we get the desired transverse dynamic structure factor @xmath147 where @xmath148 are the solutions of the equation @xmath149 . the @xmath3 dynamic structure factor ( [ 4.03 ] ) is governed exclusively by a two - fermion ( one particle and one hole ) excitation continuum .
the properties of the two - fermion excitation continuum were discussed by g. mller et al@xcite ; we present these results briefly below .
the boundaries of the two - fermion continuum in the plane wave - vector @xmath150
frequency @xmath151 ( we assume @xmath152 , @xmath153 ) are determined by the equations @xmath154 where @xmath155 . moreover , in the ground state we have to require in addition @xmath156 and @xmath157 , i.e. @xmath158 and @xmath159 .
we start with the zero - temperature case . in this case
the two - fermion excitation continuum exists as long as @xmath160 .
let us introduce the parameter @xmath161 and the following characteristic lines in the @xmath150@xmath151 plane @xmath162 the two - fermion dynamic quantities in the ground state may have non - zero values only within a restricted region of the @xmath150@xmath151 plane with the lower boundary @xmath163 and the upper boundary @xmath164 if @xmath165 or @xmath166 if @xmath167 .
obviously , the two - fermion dynamic quantities may have only three soft modes @xmath168 .
moreover , there is a middle boundary of the two - fermion excitation continuum @xmath169 if @xmath167 along which the two - fermion dynamic quantities exhibit a jump increasing their values by 2 .
finally , the two - fermion dynamic quantities show one - dimensional square - root van hove divergencies along the curve @xmath170 . in fig .
[ fig02 ] as temperature increases the lower boundary is smeared - out and finally disappears , the upper boundary becomes @xmath171 along which van hove singularities occur . in the high - temperature limit
the two - fermion dynamic structure factor becomes @xmath172-independent .
there are other dynamic quantities which probe the two - fermion excitation continuum .
let us consider the dimer operator @xmath173 that operator is related to a perturbation to the hamiltonian ( [ 2.11 ] ) which mimics dimerization , @xmath174 .
the dynamics of fluctuations of the dimer operator can be measured experimentally : the corresponding dimer dynamic structure factor is relevant to phonon - assisted optical absorption processes in magnetic - chain compounds@xcite .
the calculation of the time - dependent dimer - dimer correlation function repeats all steps discussed above while deriving ( [ 4.02 ] ) and ends up with @xmath175 inserting eq .
( [ 4.09 ] ) into eq . ( [ 1.04 ] ) we get the dimer dynamic structure factor @xmath176 we can also calculate the @xmath177 and @xmath178 dynamic structure factors @xmath179 in fig .
[ fig04 ] comparing eqs .
( [ 4.03 ] ) , ( [ 4.10 ] ) , ( [ 4.11 ] ) we immediately recognize that all these dynamic quantities are governed exclusively by the two - fermion excitation continuum ( for other two - fermion dynamic quantities see below and also refs .
@xcite ) and therefore all of them exhibit generic properties inherent in the two - fermion dynamic quantities .
however , they also exhibit some specific properties originating from additional factors in the integrands in eqs .
( [ 4.03 ] ) , ( [ 4.10 ] ) , ( [ 4.11 ] ) ( e.g. singularities may be suppressed etc . , compare fig .
[ fig03]b and fig . [ fig04]a ) .
in general , the dynamic structure factor governed by the two - fermion excitation continuum can be written in the following form @xmath182 all these quantities show generic properties ( spectral boundaries , soft modes , singularity structure ) and specific properties controlled by @xmath183 .
we proceed by considering more complicate dynamic quantities .
namely , consider a trimer operator@xcite @xmath184 that operator enters as a perturbation to the hamiltonian ( [ 2.11 ] ) which mimics trimerization , @xmath185 .
the dynamics of fluctuations of the trimer operator , although it can be analyzed rigorously , is less evident from the experimental point of view .
its importance , however , is justified as a quantity of intermediate complexity between the @xmath3 and the @xmath4 and @xmath5 dynamic quantities .
the calculation of the time - dependent trimer - trimer correlation function contains no new conceptual ideas but is somewhat tedious .
the final result for the time - dependent trimer correlation function reads @xmath186 with @xmath187 and @xmath188 .
obviously , the calculation of such an average as @xmath189 according to the wick - bloch - de dominicis theorem is rather complicated . substituting ( [ 4.14 ] ) into eq .
( [ 1.04 ] ) we obtain the following result for the trimer dynamic structure factor @xmath190 with @xmath191 for further details see ref . @xcite . as can be seen from eqs .
( [ 4.17 ] ) , ( [ 4.18 ] ) , ( [ 4.19 ] ) the trimer dynamic structure factor is governed both by the two - fermion ( one particle and one hole ) excitation continuum discussed above , the term @xmath192 , and by the four - fermion ( two particles and two holes ) excitation continuum , the term @xmath193 . the four - fermion excitation continuum is determined by the conditions @xmath194 where @xmath195 and @xmath196 . moreover , in the ground state we have to require in addition @xmath156 , @xmath197 , @xmath198 , @xmath199 , i.e. @xmath158 , @xmath200 , @xmath201 , @xmath202 .
we start with the zero - temperature case .
the lower boundary is given by one of the following curves @xmath203 depending on the value of @xmath172 , @xmath204 and the value of @xmath150 , @xmath205 as is shown in the left panel in fig . [ fig05 ] .
the boundary between the region @xmath206 ( where @xmath207 is the lower boundary ) and the region @xmath39 ( where @xmath208 is the lower boundary ) ( see the left panel in fig . [ fig05 ] ) is given by the formula @xmath209 where @xmath210 , @xmath211 ; @xmath212 , @xmath213 ; @xmath214 ; @xmath215 ; @xmath216 , @xmath217 ; @xmath218 ( for further details see ref .
the four - fermion contribution to dynamic quantities may exhibit soft modes @xmath219 .
next we pass to the upper boundary of the four - fermion excitation continuum which is given by one of the following curves @xmath220 depending on the value of @xmath172 , @xmath204 and the value of @xmath150 , @xmath205 as is shown in the right panel in fig . [ fig05 ] .
the boundary between the regions @xmath221 and @xmath222 is given by the curve @xmath223 .
[ fig06 ] we compare the ground - state two - fermion and four - fermion excitation continua for two values of the transverse field @xmath172 .
the four - fermion excitation continuum always contains the two - fermion excitation continuum .
the lower boundaries may coincide ( e.g. in the zero - field case the lower boundary is @xmath224 for both continua , panel a in fig .
[ fig06]a ) whereas the upper boundaries are different .
next we turn to the van hove singularities inherent in the four - fermion dynamic quantities . evidently , the quantity @xmath225 may exhibit van hove singularities characteristic to the three - dimensional density of states .
the lines of potential singularities are as follows @xmath226 the four - fermion dynamic quantities may exhibit cusp singularities ( akin to density - of - states effects in thee dimensions ) along these curves .
we illustrate potential singularities in the frequency profiles of @xmath227 ( [ 4.24 ] ) at different @xmath150 in fig .
[ fig07 ] . for nonzero temperatures
the lower boundary is smeared out and finally disappears .
the upper boundary is given by eq .
( [ 4.22 ] ) . in the high - temperature limit the properties of the four - fermion excitation continuum
become @xmath172-independent .
after discussing some generic properties of the four - fermion dynamic quantities ( inherent in any four - fermion dynamic quantity ) we illustrate some specific properties conditioned by the function @xmath228 ( [ 4.16 ] ) . in fig .
[ fig08 ] we display the trimer dynamic structure factor ( [ 4.17 ] ) .
the contributions of the two - fermion excitation continuum and the four - fermion excitation continuum to this quantity can be easily distinguished .
we may formally introduce the polymer operator @xmath229 ( evidently @xmath230 and @xmath231 ) .
now the dynamic polymer structure factor @xmath232 will involve @xmath233-fermion excitations with @xmath234 .
these quantities are of moderate complexity in comparison with @xmath235 and @xmath236 which are enormously complex ( see below ) .
we also note that in the limit @xmath237 @xmath238 the last term in eq .
( [ 4.27 ] ) is nonzero only if @xmath239 .
in passing , we note that the multimagnon continua of quantum spin chains have been discussed recently on general ground by t. barnes@xcite .
obviously , there is an essential difference in comparison with our case , since the jordan - wigner fermions obey the fermi statistics and this point has important consequences for the four - fermion excitation continuum considered in some detail above .
we pass to dynamic structure factors which are governed by many - fermion excitations .
let us recall that according to the jordan - wigner transformation we have @xmath240 where we have introduced the operators @xmath241 .
obviously the operators @xmath242 are linear combinations of the operators @xmath243 , @xmath244 in terms of which the hamiltonian is diagonal ( see eq .
( [ 4.01 ] ) ) .
consider now the @xmath4 time - dependent spin correlation function @xmath245 it contains a product of @xmath246 @xmath247 operators ( in contrast to @xmath248 which contains the product of only four @xmath247 operators ) .
therefore the calculation of @xmath4 and @xmath5 dynamic quantities ( which are governed by many - fermion excitations ) is essentially more complicated . at zero temperature
the @xmath4 and @xmath5 time - dependent correlation functions are extremely simple only when @xmath251@xcite .
consider for example the case @xmath252 when the ground state is completely polarized @xmath253 ( in spin language ) or completely empty @xmath254 ( in fermionic language ) .
owing to the simplicity of the ground state @xmath255 , @xmath256 .
therefore @xmath257 and as a result @xmath258 many results at @xmath63 refer to the asymptotic behavior of the @xmath4 or @xmath5 time - dependent spin correlation functions@xcite . from the paper by a. r. its et al we know the long - time asymptotic behavior at nonzero temperatures @xmath259 on the other hand ,
we can obtain the @xmath4 and @xmath5 dynamic quantities numerically@xcite .
consider the slightly more complicated inhomogeneous spin-1/2 anisotropic @xmath6 chain in a transverse field with the hamiltonian @xmath260 where @xmath261 to diagonalize a form bilinear in fermi operators like ( [ 4.34 ] ) we perform the linear canonical transformation @xmath262 the resulting hamiltonian reads as follows @xmath263 if the coefficients
@xmath264 , @xmath265 satisfy the set of equations @xmath266 further it may be convenient to introduce the linear combinations @xmath267 and @xmath268 which enter the relations @xmath269 we calculate the time - dependent spin correlation functions using the wick - bloch - de dominicis theorem . for example , @xmath270 the r.h.s . of eq .
( [ 4.40 ] ) may be compactly written as the pfaffian of the @xmath271 antisymmetric matrix @xmath272 similarly ( see eq .
( [ 4.29 ] ) ) , for the more complicated @xmath4 time - dependent spin correlation function we have @xmath273 i.e. @xmath274 can be written as a pfaffian of a @xmath275 antisymmetric matrix .
the elementary contractions involved in ( [ 4.41 ] ) , ( [ 4.42 ] ) read @xmath276 it is worthwhile to recall some properties of the pfaffians which are used for calculating them . in the first numerical studies the authors used the relation @xmath277 and computed numerically the determinants which gave the values of pfaffians according to ( [ 4.44 ] ) . on the other hand
, the pfaffian may be computed directly@xcite noting that @xmath278 and that @xmath279 we use the approach described to calculate the @xmath4 and @xmath5 dynamic structure factors for the spin-1/2 transverse @xmath1 chain numerically@xcite . to estimate the quality of the numerical procedure we compare our numerical findings with exact analytical results in the high - temperature limit and with exact asymptotics for finite temperatures in fig .
[ fig09 ] . knowing the time - dependent correlation functions we obtain the corresponding dynamic structure factors according to @xmath280 with @xmath281 . in practice
we consider chains of @xmath282 sites , take @xmath283 , @xmath284 up to 50 or up to 100 , and set @xmath285 ( see ref .
the results of our calculations for @xmath235 and @xmath236 are illustrated in figs .
[ fig10 ] , [ fig11 ] .
let us recall , the transverse dynamic structure factor @xmath138 probes two - particle excitations , i.e. it is governed by the excitations which are composed of two jordan - wigner spinless fermions .
the two - fermion excitation continuum has a sharp upper frequency cutoff at which @xmath138 may diverge . at @xmath63
it has also a sharp lower frequency cutoff which touches @xmath286 at @xmath287 ( soft modes ) .
@xmath138 is almost structureless ( apart from upper boundary singularities ) and exists only for @xmath160 . in the high - temperature limit @xmath249 @xmath138 becomes @xmath172-independent .
all these features are nicely seen in fig .
[ fig03 ] .
in contrast , the dynamic structure factors @xmath235 and @xmath236 are many - particle quantities in terms of the jordan - wigner spinless fermions . the frequency range of these quantities is not _ a priori _ restricted , however , in the low - temperature limit @xmath288 @xmath235 and @xmath236 are rather small ( but nonzero ) outside the two - fermion excitation continuum .
these quantities show washed - out excitation branches roughly following the boundary of the two - fermion excitation continuum .
[ although the results presented in figs .
[ fig10 ] , [ fig11 ] refer to the case @xmath289 ( the ferromagnetic sign of the @xmath1 exchange interaction ) the results for @xmath290 ( the antiferromagnetic sign of the @xmath1 exchange interaction ) follow by symmetry .
in fact , while changing the sign of @xmath1 exchange interaction , @xmath291 , we get @xmath235 , @xmath236 given by eq .
( [ 4.47 ] ) in which the wave - vector is changed @xmath292 . ] from the exact calculation in the strong - field zero - temperature limit ( [ 4.32 ] ) we know that @xmath235 , @xmath236 are proportional to @xmath293 , @xmath196 . in the high - temperature limit @xmath249 @xmath235 and @xmath236 become @xmath150-independent , but depend on @xmath172 .
all the features described can be seen in figs .
[ fig10 ] , [ fig11 ] .
it is instructive to compare our precise numerical findings in the low - temperature limit with the results for the ground - state dynamic structure factors obtained by bosonization@xcite . in the case
@xmath294 within the framework of the bosonization approach we have @xmath295 here @xmath296 is the velocity and @xmath297 , @xmath298 are the exponents which describe correctly the singularity at the lower continuum boundary @xmath299 as @xmath300 or @xmath301 . in the case of nonzero transverse fields @xmath239 the values of the fermi momentum and fermi velocity are changed . in fig .
[ fig12 ] to summarize , in this section we have discussed dynamic properties of the spin-1/2 transverse @xmath1 chain within the jordan - wigner fermionization approach . within this scheme
the spin hamiltonian corresponds to the hamiltonian of noninteracting spinless fermions .
the transverse dynamic structure factor corresponds to the fermionic density dynamic structure factor and probes the two - fermion excitation continuum .
there are more dynamic structure factors which probe the two - fermion excitation continuum , e.g. , the dimer dynamic structure factor .
all two - fermion dynamic quantities have common features ( spectral boundaries , potential soft modes and singularities ) and specific features .
there are also dynamic quantities which probe the four - fermion excitation continuum ; as an example of such a quantity we have discussed the trimer dynamic structure factor .
remarkably , the dynamic structure factors which are associated with the dynamics of fluctuations of the @xmath103 or @xmath104 spin components ( in contrast to the transverse dynamic structure factor which is associated with dynamics of fluctuations of the @xmath2 spin component ) are enormously complex within the jordan - wigner description since they probe many - fermion excitations .
nevertheless the two - fermion excitation continuum is still important for these dynamic quantities at low temperatures .
as we have observed in our numerics , most of the spectral weight is concentrated along the boundaries of the two - fermion excitation continuum ( it was also noted earlier for the @xmath29 heisenberg chain@xcite ) .
this is not the case in the high - temperature limit when these dynamic structure factors show gaussian ridges . in the next two sections we shall follow to what extent our observations
survive for more complicated spin-1/2 @xmath6 chains .
[ [ dimerized - spin-12-isotropic - xy - chain - in - a - transverse - field - secdk5 ] ] dimerized spin-1/2 isotropic @xmath6 chain in a transverse field [ secdk5 ] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ now we pass to the dimerized spin-1/2 @xmath1 chain in a transverse field .
the hamiltonian of the model reads @xmath302 where @xmath303 is the dimerization parameter ( @xmath304 ) . after performing consequently the fourier transformation , @xmath305 , @xmath306 , @xmath307 ( @xmath8 is even ) , and the bogolyubov transformation , @xmath308 , @xmath309 , @xmath310 , @xmath311 , the hamiltonian becomes diagonal , @xmath312 with the elementary excitation energy @xmath313 , @xmath314 ( for further details see refs .
@xcite ) .
again for the @xmath4 and @xmath5 dynamic structure factors exact analytical results are rather scarce . in the high - temperature limit only the autocorrelation functions
survive@xcite @xmath316 here the jacobian theta and eta functions are defined as follows @xmath317 and the complete elliptic integrals of the 1st and the 2nd kinds are given by @xmath318 moreover , @xmath319 . in the strong - field limit @xmath320 at @xmath63
we can repeat the calculation of the previous section to find , for example , for the @xmath4 time - dependent correlation function and the corresponding dynamic structure factor the following result @xmath321 let us discuss the dynamic quantities of the dimerized transverse @xmath1 chain .
we begin with the transverse dynamic structure factor which can be written as ( compare with eq .
( [ 4.12 ] ) ) @xmath322 as can be seen from eq . ( [ 5.08 ] ) the transverse dynamic structure probes two - fermion excitations . @xmath138
may have nonzero value within a restricted region of the @xmath150@xmath151 plane when there is such a wave - vector @xmath323 , @xmath155 that @xmath324 or @xmath325 .
moreover , at zero temperature there are additional restrictions arising from the fermi functions .
the lines of potential singularities follow from the analysis of the equations @xmath326 and @xmath327 .
the characteristic lines in the @xmath150@xmath151 plane which determine the behavior of the transverse dynamic structure factor were reported for the first time by j. h. taylor and g. mller@xcite .
we show the transverse dynamic structure factor at different temperatures . comparing fig .
[ fig13]b and fig . [ fig14 ]
one can see the van hove singularities and the effects of the fermi functions and the @xmath328- and @xmath329-functions . in figs .
[ fig15 ] , [ fig16 ] next we pass to the @xmath4 dynamic structure factor obtained numerically .
typically we consider chains of @xmath282 sites assume in ( [ 1.04 ] ) @xmath332 , @xmath125 up to 50 , consider @xmath333 up to @xmath334 and take @xmath335@xcite . in fig .
[ fig17 ] in contrast to the @xmath3 dynamic structure factor which is a two - fermion dynamic quantity , the @xmath4 dynamic structure factor is a many - particle dynamic quantity within the jordan - wigner fermionization approach .
therefore , nonzero values of @xmath235 far above the two - fermion excitation continua may be expected . however , as can be seen in fig .
[ fig17 ] the opposite is true : at low - temperatures @xmath235 shows several well - defined excitation branches which follow roughly the boundaries of the two - fermion excitation continua .
although we can describe the low - energy physics also using the bosonization treatment , high - frequency features can not be reproduced within such an approach .
finally we note that the dimerized @xmath1 chain does not show bound - state branches ; within the fermionization approach this may be related to the absence of interactions between fermions .
in contrast , a particle - hole bound state can be observed in the dimerized heisenberg chain@xcite . [ [ spin-12-xy - chains - with - the - dzyaloshinskii - moriya - interaction - secdk6 ] ] spin-1/2 @xmath6 chains with the dzyaloshinskii - moriya interaction [ secdk6 ] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ in this section we examine the effect of the dzyaloshinskii - moriya interaction ( actually , the @xmath2 component of the vector of the dzyaloshinskii - moriya interaction , see eq .
( [ 2.14 ] ) ) .
the hamiltonian of the transverse @xmath1 chain with the dzyaloshinskii - moriya interaction reads @xmath336 interestingly , the dzyaloshinskii - moriya interaction can be eliminated from the hamiltonian ( [ 6.01 ] ) resulting in renormalization of the isotropic @xmath6 exchange interaction@xcite . to see this , consider the following spin axes rotation @xmath337 after such a unitary transformation the hamiltonian ( [ 6.01 ] ) becomes @xmath338 using the unitary transformation ( [ 6.02 ] ) we can examine the effect of the dzyaloshinskii - moriya interaction using the dynamic quantities of the transverse @xmath1 chain without the dzyaloshinskii - moriya interaction discussed already in sec .
[ secdk4 ] .
first of all we note that the @xmath3 dynamic structure factor is given by eq .
( [ 4.03 ] ) , however , with @xmath339 .
the formulas determining the two - fermion excitation continua boundaries are still given by eqs .
( [ 4.05 ] ) , ( [ 4.06 ] ) , ( [ 4.07 ] ) but with @xmath340 instead of @xmath341 on the l.h.s . of these equations and in the definition of the parameter @xmath342 . exploiting ( [ 6.02 ] )
we find the following relations between the @xmath4 and @xmath5 dynamic structure factors of the model ( [ 6.01 ] ) ( l.h.s .
( [ 6.04 ] ) ) and the dynamic structure factors of the model ( [ 6.03 ] ) ( r.h.s . of eq . (
[ 6.04]))@xcite @xmath343 therefore , using eq .
( [ 4.30 ] ) we obtain for the model ( [ 6.01 ] ) @xmath344 i.e. in the high - temperature limit @xmath235 and @xmath236 are @xmath150-independent and display a single gaussian ridge at @xmath345 . for arbitrary values of temperature and transverse field we use eq .
( [ 6.04 ] ) and numerical results for the @xmath4 and @xmath5 dynamic structure factors of the transverse @xmath1 chain ( [ 6.03 ] ) ( see sec . [ secdk4 ] ) to reveal the effect of the dzyaloshinskii - moriya interaction .
some of our findings are plotted in fig .
[ fig18 ] we recall that in the low - temperature limit when @xmath289 and @xmath348 @xmath235 and @xmath236 are concentrated in the @xmath150@xmath151 plane along the curves ( [ 4.05 ] ) , ( [ 4.06 ] ) , ( [ 4.07 ] ) which determine the boundaries of the two - fermion excitation continuum @xmath351 , @xmath352 and @xmath353 ( see sec .
[ secdk4 ] , figs .
[ fig10 ] , [ fig11 ] ) .
[ for the antiferromagnetic sign of @xmath1 exchange interaction @xmath290 these dynamic quantities are concentrated along the curves @xmath354 , @xmath355 and @xmath356 as it follows from simple symmetry arguments . ] in the case when the dzyaloshinskii - moriya interaction is present , @xmath349 , the two - fermion excitation continuum splits into two continua ( see fig . [ fig18 ] ) , the ` left ' one with the boundaries @xmath357 , @xmath358 and @xmath359 and the ` right ' one with the boundaries @xmath360 , @xmath361 and @xmath362 .
( the ` left ' and the ` right ' continua are connected by symmetry operation . )
the larger @xmath363 is the larger is the splitting controlled by @xmath364 . at fixed @xmath349 and @xmath123 the spectral weight
is equally distributed between the left and the right continua ( panel b in fig .
[ fig18 ] ) . while @xmath365 increases from 0 to @xmath366 the spectral weight moves from one continuum to another continuum ( panel d in fig .
[ fig18 ] ) .
we note in passing that the discussed peculiarities of the @xmath4 dynamic structure factor may be used for an unambiguous determination of the dzyaloshinskii - moriya interaction in chain compounds for example , in resonance experiments@xcite . in the case of the anisotropic @xmath6 chain
the dzyaloshinskii - moriya interaction can not be eliminated by the transformation ( [ 6.02 ] ) .
now we face the hamiltonian @xmath367 with @xmath98 and @xmath99 .
this hamiltonian can be put into a diagonal form by performing the fourier transformation , @xmath368 , @xmath369 , and the bogolyubov transformation , @xmath370 , @xmath371 , @xmath372 , @xmath373 , @xmath374 .
the final result reads @xmath375 , @xmath376 .
we notice that the elementary excitation energy spectrum is gapless when @xmath377 and @xmath378 or when @xmath379 and @xmath380 .
the calculation of the transverse dynamic structure factor follows the lines explained in some detail in sec .
[ secdk4 ] and ends up with@xcite @xmath381 the transverse dynamic factor , as it follows from eq .
( [ 6.08 ] ) , is shown in panel c in fig .
[ fig19 ] for a typical set of parameters . where we show two - fermion excitation continua for a specific set of parameters @xmath382 , @xmath383 , @xmath384 , @xmath385 ) .
we begin with the high - temperature limit when the fermi factors are not essential ( left panels in fig .
[ fig20 ] ) .
the two - fermion dynamic structure factor may have nonzero values in the @xmath150@xmath151 plane if the equation @xmath386 has at least one solution @xmath387 , @xmath388 .
next , the lower and the upper boundaries are given by @xmath389 the two - fermion dynamic quantities may exhibit van hove singularities along the line @xmath390 where @xmath323 satisfies the equation @xmath391 if for the solution of eq .
( [ 6.10 ] ) we also have @xmath392 the two - fermion dynamic structure factor shows the well - known square - root singularity .
however , it may happen that @xmath393 but @xmath394 .
then a van hove singularity is characterized by the exponent 2/3 . that is really the case , for example , for @xmath395 , @xmath383 , @xmath384 , @xmath385 for @xmath396 at @xmath397 .
we have @xmath398 , @xmath399 at @xmath400 .
therefore in the @xmath401-vicinity of @xmath402 the two - fermion dynamic structure factor should be proportional to @xmath403 .
we mention that the singularity with this exponent is also present for @xmath348@xcite . in the zero - temperature case the effect of the fermi functions involved in the @xmath404-functions ( see eq .
( [ 6.08 ] ) ) becomes important ( right panels in fig .
[ fig20 ] ) . as a result the region of possible values of @xmath323
is contracted .
further details can be found in ref .
@xcite . finally , we mention the role of the @xmath405-functions ( see eq .
( [ 6.08 ] ) ) for the two - fermion dynamic structure factors which are responsible for some specific features of the transverse dynamic structure factor ( compare panels b , d , f in fig .
[ fig20 ] and panel c in fig .
[ fig19 ] ) .
we pass to the @xmath4 and @xmath406 dynamic structure factors ( see panels a and b in fig .
[ fig19 ] ) .
these dynamic structure factors are many - fermion dynamic quantities and although they are not restricted to some region in the @xmath150@xmath151 plane , they are rather small outside the two - fermion excitation continua ( compare panels a and b with panel c in fig . [ fig19 ] ) . in the low - temperature limit the @xmath4 and
@xmath406 dynamic structure factors show several washed - out excitation branches which are in correspondence with characteristic lines of the two - fermion excitation continua . in the high - temperature limit @xmath235 and @xmath407
become @xmath150-independent .
it should be stressed that the constant frequency or constant wave - vector scans of the dynamic structure factors clearly manifest the presence of the dzyaloshinskii - moriya interaction and some easily recognized features of these quantities can be used for determining the dzyaloshinskii - moriya interaction .
let us discuss what kind of results for spin models can be obtained in two dimensions after applying the jordan - wigner transformation ( sec .
[ secdk3 ] ) .
we consider the spin-1/2 isotropic @xmath6 model on a spatially anisotropic square lattice with the hamiltonian @xmath408 where @xmath341 and @xmath409 are the @xmath1 exchange interactions in the horizontal and vertical directions .
our aim is to calculate the transverse dynamic structure factor @xmath410 we notice that the transverse dynamic structure factor @xmath7 ( [ 7.02 ] ) for the spin model ( [ 7.01 ] ) is related to the density - density dynamic structure factor of hard - core bosons on a square - lattice @xcite .
we apply the two - dimensional jordan - wigner transformation ( [ 3.01 ] ) , ( [ 3.03 ] ) to the spin hamiltonian ( [ 7.01 ] ) .
moreover , we adopt the mean - field approach for the phase factors and change the gauge leaving the flux per elementary plaquette @xmath411 to be equal to @xmath412 . as a result
we arrive at the hamiltonian like ( [ 3.07 ] ) , i.e. @xmath413 the hamiltonian ( [ 7.03 ] ) contains the correct results in the one - dimensional limit when either @xmath414 or @xmath415 ( in the former case to recover the one - dimensional hamiltonian ( [ 2.12 ] ) ( with @xmath294 ) one has to perform in addition a gauge transformation @xmath416 , @xmath417 , @xmath418 ) .
the hamiltonian ( [ 7.03 ] ) can be diagonalized by performing 1 ) the fourier transformation , @xmath419 , @xmath420 , @xmath421 , @xmath422 , @xmath423 is even , which yields @xmath424 with @xmath425 and @xmath426 and 2 ) the bogolyubov transformation , @xmath427 , @xmath428 , which yields @xmath429 ( the prime near the sum in eq .
( [ 7.05 ] ) means that @xmath430 varies in the thermodynamic limit in the region @xmath431 , @xmath432 ) .
the calculation of the transverse dynamic structure factor repeats the steps elaborated in some detail for the one - dimensional case .
first , we use the wick - bloch - de dominicis theorem to obtain the @xmath3 time - dependent spin correlation function @xmath433 then we plug eq .
( [ 7.06 ] ) into eq .
( [ 7.02 ] ) to obtain the following expression for the @xmath3 dynamic structure factor @xmath434 one can easily convince oneself that eq .
( [ 7.07 ] ) contains the correct result in the one - dimensional limit ( [ 4.03 ] ) ( with @xmath294 ) when either @xmath435 or @xmath415 . in the two - dimensional case
( [ 7.07 ] ) is an approximate formula for the transverse dynamic structure factor of the spin-1/2 isotropic @xmath6 model on a spatially anisotropic square lattice .
we show frequency profiles of @xmath7 for a representative set of parameters .
formula ( [ 7.07 ] ) implies the interpretation of @xmath7 as a two - fermion excitation quantity . as can be seen from figs .
[ fig21 ] , [ fig22 ] @xmath7 exhibits several washed - out excitation branches which can be generated by two spinless fermions in accordance with ( [ 7.07 ] ) .
we begin with the low - temperature limit when only the fourth term in eq .
( [ 7.07 ] ) ( which contains @xmath436 ) survives .
consider , for example , two fermions with @xmath437 and @xmath438 with the energy of the pair @xmath439 or two fermions with @xmath440 and @xmath441 with the energy of the pair @xmath442 these modes are the well - known spin waves@xcite clearly present at low temperatures ( panels a and c in fig .
[ fig21 ] ) .
further , one can recognize the high frequency modes [ @xmath443 , @xmath444 with the dispersion relations @xmath445 another set of high - frequency modes [ @xmath446 , @xmath447 have the dispersion relations @xmath448 the low - frequency mode [ @xmath449 with the dispersion relation @xmath450 ( it is composed of two fermions , the energy of one of which equals zero ) forms the low - frequency cutoff at zero temperature . comparing left and right panels in figs .
[ fig21 ] and [ fig22 ] one can also see the modes which become visible only as temperature increases ( at zero temperature they are forbidden because of the fermi factors in eq .
( [ 7.07 ] ) ) .
putting @xmath451 , @xmath452 for @xmath453 and @xmath454 , @xmath455 for @xmath456 we get @xmath457 this excitation branch contains most of the spectral weight at high temperatures ( see panel d in fig .
[ fig21 ] and panel b in fig . [ fig22 ] ) .
the established modes ( [ 7.08 ] ) ( [ 7.15 ] ) manifest themselves as peaks , cusps or cutoffs in the frequency or wave - vector profiles of @xmath7 .
the frequency profiles depicted in fig .
[ fig22 ] may be almost symmetric or asymmetric , they may resemble @xmath303-peaks or result from two coalesced peaks , they may gradually disappear or may be abruptly cut off .
it is worthwhile to mention here some experimental studies on dynamic properties of two - dimensional quantum spin models , in particular , the neutron scattering experiments on cs@xmath67cucl@xmath68@xcite ( for a theory of dynamic correlations in the spin - liquid phase in cs@xmath67cucl@xmath68 see ref .
cs@xmath67cucl@xmath68 is a two - dimensional low - exchange quantum magnet .
it has a layered crystal structure ; in each layer the exchange paths form a triangle lattice with nonequivalent interactions along chains @xmath458 mev and along zig - zag bonds @xmath459 .
the interlayer coupling is small @xmath460 and it stabilizes the long - range magnetic order below @xmath461 k. the neutron scattering measurements in the spin - liquid phase ( i.e. above @xmath462 but below @xmath341 , @xmath463 when the two - dimensional magnetic layers are decoupled ) clearly indicate that the dynamic correlations are dominated by highly dispersive excitation continua which is a characteristic signature of fractionalization of spin-1 spin waves into pairs of deconfined spin-1/2 spinons .
linear spin - wave theory including one- and two - magnon processes can not describe the continuum scattering .
the proposed theories@xcite are based either on a quasi - one - dimensional approach ( that immediately introduces spinon language ) or on the explicitly two - dimensional resonating - valence - bond picture . as a final remark we recall that eq .
( [ 7.07 ] ) contains the exact result ( [ 4.03 ] ) in the one - dimensional limit . on the other hand , eq .
( [ 7.07 ] ) gives an approximate result in the two - dimensional case because of the mean - field description of the phase factors which arise after fermionization .
the adopted mean - field treatment neglects a complicated interaction between spinless fermions . in the case of the @xmath29 heisenberg model
the interaction between fermions is present even within the adopted mean - field procedure due to the interaction between @xmath2 spin components .
the quartic terms in the fermionic hamiltonian may be treated after making further approximation ( see references cited in ref .
@xcite and also ref .
@xcite ) . the jordan - wigner transformation which realizes a spin - to - fermion mapping
was suggested as a rigorous framework for the description of spin-1/2 @xmath6 chains in the early 1960s . in general , the jordan - wigner fermionization permits to map a system of interacting spins @xmath0 onto a system of spinless fermions .
it may happen that the spinless fermions are noninteracting . in this case
this approach reveals an exactly solvable spin model .
however , even for exactly solvable spin models not all ` simple ' quantities of interest in spin language remain simple in fermionic language .
for example , the @xmath2 spin component attached to the site @xmath39 , @xmath464 , becomes the product of two fermi operators attached to this site , @xmath465 .
in contrast , the local spin operators @xmath466 , @xmath467 , @xmath468 become nonlocal objects in fermionic description involving a string of sites @xmath469 ( see eqs .
( [ 2.08 ] ) , ( [ 2.09 ] ) ) .
this leads to some complications in studying the dynamics of fluctuations of these operators : the dynamics of fluctuations of operators which seem to be rather simple in spin language may be governed by many - particle correlations in fermionic language . as we have discussed in sections [ secdk4 ] , [ secdk5 ] , [ secdk6 ] , the jordan - wigner fermionization approach permits to establish a number of rigorous results for the dynamics of spin-1/2 @xmath6 chains . especially easy
are the cases of two- and four - fermion dynamic quantities which are amenable mostly to analytical calculations .
the case of many - fermion dynamic quantities is more complicated , however , these quantities can be examined numerically at very high precision . for more realistic spin-1/2 @xmath29 heisenberg chains the jordan - wigner fermionization approach leads to a system of interacting spinless fermions .
the simplest way to proceed in this case is to apply hartree - fock - like approximations@xcite .
if we are interested in low - energy physics only it might be helpful to apply the bosonization approach@xcite .
the results for one - dimensional quantum spin systems obtained using the jordan - wigner fermionization can be compared with the outcomes of alternative approaches : field - theoretic bosonization techniques@xcite valid in the low - energy limit ( see fig . [ fig12 ] ) , bethe ansatz calculations ( for calculation of dynamic structure factors of spin-1/2 @xmath29 chains see refs .
@xcite ) or exact diagonalization computations which , however , are restricted to small finite systems . for two - dimensional quantum spin models achievements
are rather modest . in this case
the jordan - wigner fermionization approach can provide an approximate theory ; the simplest one treats in the mean - field spirit the phase factors which arise after fermionization .
we end up with a brief comment about the experimental relevance of some of the dynamic quantities calculated .
they may be used for interpretation of the energy absorption in electron spin resonance ( esr ) experiments@xcite . consider an esr experiment in which the static magnetic field is directed along the @xmath2 axis and the electromagnetic wave with the polarization in @xmath470 direction ( say @xmath471 )
are applied to a magnetic system which is described as a spin-1/2 @xmath6 chain ( esr experiment in the standard faraday configuration ) . in such an esr experiment one
measures the intensity of the radiation absorption @xmath472 as a function of frequency @xmath473 of the electromagnetic wave . within the linear response theory
the absorption intensity is written as @xmath474 where @xmath475 is the imaginary part of the @xmath476 component of the dynamic susceptibility @xmath477 at zero wave - vector @xmath478 .
we notice that @xmath479 where the dynamic structure factor is defined by eq .
( [ 1.04 ] ) .
thus , the peculiarities of the dynamic structure factor @xmath480 caused , e.g. , by the @xmath6 exchange interaction anisotropy , dzyaloshinskii - moriya interaction or dimerization should manifest themselves in esr experiments .
the time - dependent spin correlation functions taken at the same site or at the neighboring sites manifest themselves in the spin - lattice relaxation rate @xmath481 measured by nuclear magnetic resonance ( nmr)@xcite .
the activity in the field of the jordan - wigner fermionization approach has expanded much over the last few decades . despite some limitations ,
the jordan - wigner fermionization approach has a wide range of applicability .
particularly attractive is that it allows one to handle complicated problems of low - dimensional quantum spin systems armed with relatively simple tools .
it thus seems quite likely that it will continue to be used successfully in the coming years .
the author would like to thank t. krokhmalskii , t. verkholyak , j. stolze , g. mller and h. bttner in collaboration with whom the study of the dynamics was performed .
he is grateful to t. krokhmalskii for preparing all figures for the paper , many interesting conversations and helpful comments and suggestions .
he thanks j. stolze and t. verkholyak for a critical reading of the manuscript .
nato support is acknowledged ( the grant reference number cbp.nukr.clg 982540 , project `` dynamic probes of low - dimensional quantum magnets '' ) .
the author acknowledges kind hospitality of the organizers of the 43rd karpacz winter school of theoretical physics `` condensed matter physics in the prime of xxi century : phenomena , materials , ideas , methods '' in ladek zdrj in february 2007 .
caux , f. h. l. essler and u. lw , _ phys .
rev . _ * b68 * , 134431 ( 2003 ) ; + d. v. dmitriev and v. ya .
krivnov , _ phys .
rev . _ * b70 * , 144414 ( 2004 ) ; + d. v. dmitriev and v. ya .
krivnov , _
pisma zh . eksp .
fiz . _ * 80 * , 349 ( 2004 ) .
m. suzuki , _ prog .
phys . _ * 46 * , 1337 ( 1971 ) ; + d. gottlieb and j. rssler , _ phys .
rev . _ * b60 * , 9232 ( 1999 ) ; + o. derzhko , j. richter and v. derzhko , _ ann .
( leipzig ) _ * 8 * , si-49 ( 1999 ) ; + i. titvinidze and g. i. japaridze , _ eur . phys .
j. _ * b32 * , 383 ( 2003 ) .
j. p. de lima and l. l. gonalves , _
j. magn .
mater . _ * 206 * , 135 ( 1999 ) ; + o. derzhko , j. richter and o. zaburannyi , _ physica _ * a285 * , 495 ( 2000 ) ; + o. derzhko , j. richter , t. krokhmalskii and o. zaburannyi , _ phys .
rev . _ * e69 * , 066112 ( 2004 ) ; + j. p. de lima , l. l. gonalves and t. f. a. alves , _ phys .
rev . _ * b75 * , 214406 ( 2007 ) .
i. affleck , field theory methods and quantum critical phenomena . in : _
fields , strings and critical phenomena _ , ed .
e. brzin and j. zinn - justin ( elsevier , amsterdam , 1989 ) pp .
563 - 640 ; + j. von delft and h. schoeller , _ ann .
( leipzig ) _ * 7 * , 225 ( 1998 ) ; + s. rao and d. sen , arxiv : cond - mat/0005492
. m. azzouz , _ phys .
rev . _ * b48 * , 6136 ( 1993 ) ; + m. azzouz , l. chen and s. moukouri , _ phys .
rev . _ * b50 * , 6233 ( 1994 ) ; + m. azzouz and c. bourbonnais , _ phys .
rev . _ * b53 * , 5090 ( 1996 ) ; + m. azzouz , b. dumoulin and a. benyoussef , _ phys .
* b55 * , r11957 ( 1997 ) .
g. misguich , th .
jolicoeur and s. m. girvin , _ phys .
lett . _ * 87 * , 097203 ( 2001 ) ; + d. c. cabra , m. d. grynberg , p. c. w. holdsworth and p. pujol , _ phys .
rev . _ * b65 * , 094418 ( 2002 ) ; + m .- c . chang and m .- f .
yang , _ phys .
rev . _ * b66 * , 184416 ( 2002 ) .
j. lorenzana and g. a. sawatzky , _ phys .
lett . _ * 74 * , 1867 ( 1995 ) ; + j. lorenzana and g. a. sawatzky , _ phys .
rev . _ * b52 * , 9576 ( 1995 ) ; + h. suzuura , h. yasuhara , a. furusaki , n. nagaosa and y. tokura , _ phys .
lett . _ * 76 * , 2579 ( 1996 ) ; + yongmin yu , g. mller and v. s. viswanath , _ phys . rev . _ * b54 * , 9242 ( 1996 ) ; + j. lorenzana and r. eder , _ phys .
rev . _ * b55 * , r3358 ( 1997 ) ; + r. werner , _ phys .
rev . _ * b63 * , 174416 ( 2001 ) .
b. m. mccoy , e. barouch and d. b. abraham , _ phys .
rev . _ * a4 * , 2331 ( 1971 ) ; + j. lajzerowicz and p. pfeuty , _ phys .
rev . _ * b11 * , 4560 ( 1975 ) ; + b. m. mccoy , j. h. h. perk and r. e. shrock , _ nucl .
phys . _ * b220 * , 35 ( 1983 ) ; + b. m. mccoy , j. h. h. perk and r. e. shrock , _ nucl .
phys . _ * b220 * , 269 ( 1983 ) ; + g. mller and r. e. shrock , _ phys . rev . _ * b29 * , 288 ( 1984 ) .
f. colomo , a. g. izergin , v. e. korepin and v. tognetti , _ phys .
lett . _ * a169 * , 243 ( 1992 ) ; + a. r. its , a. g. izergin , v. e. korepin and n. a. slavnov , _ phys .
lett . _ * 70 * , 1704 ( 1993 ) ; + a. r. its , a. g. izergin , v. e. korepin and n. a. slavnov , _ phys .
lett . _ * 70 * , 2357 ( 1993 ) ; + f. colomo , a. g. izergin and v. tognetti , _ j. phys . _ * a30 * , 361 ( 1997 ) .
m. oshikawa and i. affleck , _ phys .
lett . _ * 79 * , 2883 ( 1997 ) ; + d. n. aristov and s. v. maleyev , _ phys .
rev . _ * b62 * , r751 ( 2000 ) ; + o. derzhko , j. richter and o. zaburannyi , _ j. phys . : condens .
matter _ * 12 * , 8661 ( 2000 ) .
t. matsubara and h. matsuda , _ prog .
* 16 * , 569 ( 1956 ) ; + h. matsuda and t. matsubara , _ prog .
* 17 * , 19 ( 1957 ) ; + h. matsuda , _ prog .
phys . _ * 18 * , 357 ( 1957 ) ; + t. morita , _ prog .
phys . _ * 18 * , 462 ( 1957 ) .
o. derzhko and t. krokhmalskii , _ physica _ * b337 * , 357 ( 2003 ) ; + o. derzhko and t. krokhmalskii , _ acta physica polonica _ * b32 * , 3421 ( 2001 ) ; + o. derzhko and t. krokhmalskii , _ j. magn
. magn .
mater . _ * 232 - 245 * , 778 ( 2002 ) ; + o. derzhko and t. krokhmalskii , _ czechoslovak journal of physics _ * 55 * , 601 ( 2005 ) .
r. coldea , d. a. tennant , r. a. cowley , d. f. mcmorrow , b. dorner and z. tylczynski , _ phys .
lett . _ * 79 * , 151 ( 1997 ) ; + r. coldea , d. a. tennant , a. m. tsvelik and z. tylczynski , _ phys .
lett . _ * 86 * , 1335 ( 2001 ) ; + r. coldea , d. a. tennant , k. habicht , p. smeibidl , c. wolters and z. tylczynski , _ phys .
lett . _ * 88 * , 137203 ( 2002 ) ; + r. coldea , d. a. tennant and z. tylczynski , _ phys .
rev . _ * b68 * , 134424 ( 2003 ) .
m. bocquet , f. h. l. essler , a. m. tsvelik and a. o. gogolin , _ phys .
rev . _ * b64 * , 094425 ( 2001 ) ; + m. bocquet , _ phys .
rev . _ * b65 * , 184415 ( 2002 ) ; + c. h. chung , j. b. marston and r. h. mckenzie , _ j. phys .
: condens .
matter _ * 13 * , 5159 ( 2001 ) ; + c .- h .
chung , k. voelker and y. b. kim , _ phys .
rev . _ * b68 * , 094412 ( 2003 ) ; + yi zhou and x .- g .
wen , arxiv : cond - mat/0210662 .
caux and j. m. maillet , _ phys .
lett . _ * 95 * , 077201 ( 2005 ) ; + j .- s .
caux , r. hagemans and j. m. maillet , _ j. stat
: theor .
_ , p09003 ( 2005 ) ; + j .- s . caux and r. hagemans , _ j. stat
: theor .
_ , p12013 ( 2006 ) . | the jordan - wigner transformation is known as a powerful tool in condensed matter theory , especially in the theory of low - dimensional quantum spin systems .
the aim of this chapter is to review the application of the jordan - wigner fermionization technique for calculating dynamic quantities of low - dimensional quantum spin models . after a brief introduction of the jordan - wigner transformation for one - dimensional spin one - half systems and some of its extensions for higher dimensions and higher spin values we focus on the dynamic properties of several low - dimensional quantum spin models .
we start from the famous @xmath0 @xmath1 chain . as a first step we recall well - known results for dynamics of the @xmath2-spin - component fluctuation operator and then turn to the dynamics of the dimer and trimer fluctuation operators .
the dynamics of the trimer fluctuations involves both the two - fermion ( one particle and one hole ) and the four - fermion ( two particles and two holes ) excitations .
we discuss some properties of the two - fermion and four - fermion excitation continua .
the four - fermion dynamic quantities are of intermediate complexity between simple two - fermion ( like the @xmath3 dynamic structure factor ) and enormously complex multi - fermion ( like the @xmath4 or @xmath5 dynamic structure factors ) dynamic quantities .
further we discuss the effects of dimerization , anisotropy of @xmath6 interaction , and additional dzyaloshinskii - moriya interaction on various dynamic quantities .
finally we consider the dynamic transverse spin structure factor @xmath7 for the @xmath0 @xmath1 model on a spatially anisotropic square lattice which allows one to trace a one - to - two - dimensional crossover in dynamic quantities . |
in the primary visual area ( v1 ) of many mammals , most neurons respond to the stimulation of two eyes unevenly : they are either right or left eye dominated . in some species ,
right / left eye dominated neurons are segregated and form a system of alternating monocular regions known as the ocular dominance pattern ( odp ) ( wiesel and hubel , 1965 , 1969 ) . in others ,
odp is not observed ( see horton and hocking , 1996b for a comprehensive list of species ) .
odps , when observed , vary significantly between different species and even between different parts of v1 in the same animal .
most modeling studies of odp ( erwin et al . , 1995 ;
swindale , 1996 ) have addressed its development .
they succeeded in generating odps of realistic appearance .
however , several _ why _ rather than _ how _ questions remained unanswered .
for instance , ( 1 ) why , from functional point of view , do the odps exist ? ( 2 ) why do some mammalian species exhibit odps while others do not ( horton and hocking , 1996b ; livingstone , 1996 ) ? ( 3 ) why do the monocular regions have different appearance ( stripes as opposed to patches ) between different parts of v1 within the same animal ( levay et al . , 1985 ) ?
the question of functional significance of odps has been addressed theoretically using the wiring economy principle ( mitchison , 1991 ; chklovskii , 2000 ) .
the idea is that evolutionary pressure to keep the brain volume to a minimum requires making the wiring ( axons and dendrites ) as short as possible , while maintaining neuronal functional properties ( cajal , 1995 ; allman and kaas , 1974 ; cowey , 1979 ; cherniak , 1992 ; young , 1992 ; chklovskii et al .
, 2001 ; koulakov and chklovskii , 2001 )
. in many cases these functional properties are specified by the rules of establishing connections between neurons , or wiring rules .
the problem presented by the wiring economy principle is therefore to find , for given wiring rules , the spatial neuronal layout that minimizes the total connection length .
this approach allows to understand many features in cortical maps , such as orientation preference maps ( koulakov and chklovskii , 2001 ) , as evolutionary adaptations , which minimize the total cortical volume .
the goal of this study is to find the simplest model , which on one hand is supported by experimental evidence , and on the other encompasses most of od phenomenology .
the use of the simple model allows us to explore its parameter space completely and to give answers to the set of questions above .
we also evaluate the dependence of the odp period on the parameters of our model and compare it to the odp periodicity observed in macaque monkey .
we find that the experimentally observed variation of the period is in agreement with the wiring economy theory .
for the purposes of minimizing the cortical wiring we consider only intra - cortical connections since they constitute the majority of gray matter wiring ( levay and gilbert , 1976 ; peters and payne , 1993 ; ahmed et al . , 1994 ) .
we therefore disregard the thalamic afferents and other extra - cortical projections . in an attempt to make wiring economy argument
more quantitative , we propose a model describing the component of intracortical circuitry sensitive to od .
the principal elements of our model are therefore the connection rules between cortical neurons . to assess the sensitivity of the intracortical wiring to od we examine the connections in the cortical layer @xmath1 , where od is most strongly pronounced .
such sensitivity has been studied by katz et al .
they made three observations regarding the wiring rules : _ i ) _ neurons in the layer @xmath1 near the interface between two od columns arborize more in home rather than in the opposite eye column .
therefore neurons establish more connections with the neurons dominated by the same rather than by the opposite eye .
_ ii ) _ axons and dendrites of these neurons have a tendency to bend away from the interface between od columns .
this implies that not only they avoid penetration to the opposite od column but also they attempt to maintain sufficiently high number of connections in the home column . _
iii ) _ axons or dendrites penetrating through the opposite eye column to the next same eye column are _ never _ observed .
this means that retinotopy has little effect on connections in layer @xmath1 .
indeed the neurons on the edges of two same eye columns separated by one opposite eye column have on average receptive fields centered next to each other . if connections in @xmath1 were sensitive to the retinotopic coordinates , these two edges should be connected ( mitchison , 1991 ) .
however out of 21 cells examined katz et al .
( 1989 ) observed _ none _ producing axons reaching the next same eye domain .
the only possibility for such cells to be connected is due to the overlap between dendritic and axonic arbors of two cells separated by more than @xmath2 @xmath3 m .
such possibility is small because of the strong repulsion of the connections by the opposite eye column located between two cells ( observation _ i ) _ ) .
these three observations lay the basis of our model which we now describe .
the elementary unit of our model mimics the columnar organization of the cortex ( mountcastle , 1957 ) and uniformity of odp along the direction normal to the slab .
the elementary unit is therefore a microcolumn , which is defined as a box , spanning the cortex perpendicular to its surface , whose other two dimensions are smaller than the characteristic scale of odp ( @xmath4 ) , and yet large enough to include many neurons .
a possible choice of dimensions for such a microcolumn is _ thickness of cortex _
@xmath5 mm @xmath6 @xmath7 , in which case it includes about 310 cells in v1 ( rockel et al . , 1980 ) .
the microcolumn units are therefore arranged on a square lattice with 30@xmath3 period .
although the choice of the elementary unit size may seem arbitrary , the results of our calculation are independent of the choice .
the size of the unit is analogous to the integration step , which does not affect the value of an integral significantly if chosen to be small enough . ,
width=144 ] motivated by the second observation in layer @xmath1 listed above , i.e. that neurons maintain a fixed number of connections in the home od column , we make the following assumption about the connection rules
. each microcolumn unit must establish connections with @xmath8 distinct units dominated by the same eye and @xmath9 units dominated by the opposite eye .
these connection rules are illustrated in fig .
. only the relative values of @xmath8 and @xmath9 ( rather than absolute ) are important because of the arbitrariness in the definition of microcolumn .
thus , if a @xmath10 microcolumn receives @xmath11 projections from the same od column and @xmath12 projections from the opposite od column , a @xmath13 microcolumn receives four times less projections respectively , or @xmath14 and @xmath15 .
this is because with coarser units each projection is more effective : connecting to one @xmath13 unit implies connecting to four @xmath10 units .
both these implementations of the model produce the same od pattern , discretized in `` pixels '' of different size .
the important quantity , which is invariant with respect to the change of `` pixel''/microcolumn size , is the ratio between @xmath8 and @xmath9 ( equal to 10 in this example ) .
this is the first parameter of our model .
the second parameter of our model is the filling fraction of the units ( microcolumns ) dominated by the left eye afferents @xmath16 with respect to the total number of units , averaged over several odp periods .
this parameter is the fraction of the left eye dominated units @xmath16 by @xmath17 . for the majority of important cases
@xmath18 , however on the periphery of visual field one of the eyes ( ipsilateral ) is underrepresented .
therefore , its filling fraction is less than @xmath19 .
the third observation above implies that the component of connections sensitive to od is not sensitive to the retinotopy , and both numbers @xmath8 and @xmath9 do not depend on the position of the receptive field of the unit .
this may be due to significant scatter of the receptive field of the cells within cortical column on the scales of about 1 mm ( hubel and wiesel , 1974 ) .
the position of the receptive field of the microcolumn is therefore vaguely defined and can not affect od sensitive connections significantly . given these wiring rules
we look for an optimal layout of the microcolumn units which minimizes the total length of connections . to find the layout minimizing the total wirelength
we employ a combination of computational and analytical techniques . to make our choice of methods clear we first comment on the expected properties of the solution .
a possible solution of our model is the _ salt and pepper _ layout in which the units dominated by right and left eyes are uniformly intermixed . in this layout the units belonging to different eyes are not segregated , odp is not formed , and the local values of the filling fraction are equal to @xmath19 ( by local value we understand an average over a domain including many units yet small compared to the period of odp ) . it should be contrasted to the case when units dominated by the same eye fill in large domains i.e. form the odp . in the latter case the local values of the filling factor of each eye vary from 0 to 1 .
however one can imagine an intermediate situation when the local filling fraction varies from @xmath20 to @xmath21 , where the amplitude of variation @xmath22 .
this corresponds to the case of _ weak segregation _ into odp .
the weak segregation is found in squirrel monkey where odp has fuzzy appearance and until recently was suspected not to be formed ( horton and hocking , 1996 ) .
if @xmath23 , i.e. the local filling fraction varies from @xmath24 to @xmath25 , the odp s have sharp appearance .
using the general terminology from binary mixtures ( cite diblock copolimer paper ) we call this regime the _ strong segregation _ limit .
the methods useful in the strong segregation limit are not good in the weak segregation case and vice versa .
we use the simulated annealing to find the optimum phases for the strong and nearly strong segregation cases . having found the optimum phase in the strong segregation case to assess the period of odp we use the exact enumeration technique , which compares layouts belonging to the same phase with different periods .
the treatment of the weak segregation case requires the use of continuous variables and is done employing the perturbation theory .
below we describe these methods in more detail . the parameters of metropolis monte - carlo method ( metropolis et al . , 1953 )
are optimized to render most consistent results for multiple restarts .
we use square @xmath26 array of units with periodic boundary conditions .
the units are either left or right eye dominated . at each step
the algorithm attempts to change the dominance of one unit to the opposite .
the value of the average filling fraction @xmath27 is enforced by adding the following term to the total connection length : @xmath28 where @xmath29 and @xmath16 are the current values of the total wirelength and average filling fraction . such term in
the functional keeps the current value @xmath16 close to the required value @xmath27 . to map out the phase diagram the values of @xmath16 change from @xmath30 to @xmath31 in @xmath32 increments .
the values of @xmath8 and @xmath9 satisfy the condition @xmath33 and are changed in unit increments , i.e. have the following values : 12 , 18 ; 13 , 17 ; 14 , 16 ; 15 , 15 ; 16 , 14 ; etc .
the phases at the intermediate points are taken from the nearest points , where result is available . the monte - carlo temperature is gradually annealed from @xmath34 to @xmath35 ( @xmath36 is the total number of units ) in 5000 sweeps through the entire system ( @xmath37 steps ) .
the resulting layout is then examined and the phases visually identified . _ salt and pepper _ layout is relatively easy to study due to its uniformity , and can be solved exactly ( chklovskii and koulakov , 2000 ) . if a layout does not deviate significantly from _ salt and pepper _ ,
i.e. the weak segregation case takes place , it can also be solved exactly .
this implies that the wire length can be written as an explicit functional of density distribution of the units .
such functional was evaluated and optimized with respect to the density variations by chklovskii and koulakov ( 2000 ) .
the optimization shows that odps are formed for the values of parameter @xmath38 .
however the simulated annealing method can not distinguish weak segregated odp from _ salt and pepper _ for @xmath39 .
there are two reasons for the failure of simulated annealing to do so : * simulated annealing is performed at small but finite temperature that destroys weakly segregated odp .
* the units can be either completely right or left eye dominated .
this implies that od can change only sharply in the described annealing version .
this is useful for obtaining the strongly segregated phases , which occupy major part of the parameter space .
however , in the weak segregation limit the local od changes smoothly .
thus used version of simulated annealing performs poorly at @xmath40 .
we therefore replace the simulated annealing results by those from chklovskii and koulakov ( 2000 ) at small values of parameter @xmath41 ( see the phase diagram below ) . to evaluate the period of odp precisely , we first determine the phase ( _ salt and pepper _ , _ stripes _ , or _ patches _ ) for the given set of parameters @xmath42 and @xmath16 , using methods described above .
we then take a lattice containing a large number of units , which exceeds sufficiently the lattice used in simulated annealing .
this is possible because the method of determining period described below is much less time consuming than simulated annealing .
we then arrange the two types of units on the lattice , using odp determined by the simulated annealing , and vary the period of the pattern to find the period producing the minimum of the wire length .
below we describe the procedure for both _
stripes _ and _ patches _ in more details . _
i ) stripes _ to find the optimum period for stripes we use an array containing 300 by 300 units .
this array includes three periods of the stripes , which run parallel to one of the sides of the region .
each period therefore includes 100 units , containing @xmath43 left and @xmath44 right eye units , @xmath45 . by varying @xmath43
we accomplish the change in the filling fraction of the left ( ipsilateral ) eye , according to the formula : @xmath46 .
we consider a string of 100 units at the center of the array , which is representative of all the units in the configuration .
for each of the central units the computer program establishes connections , according to the connection rules .
most of the calculations are done for @xmath47 .
we check that results change for different @xmath48 in a predictable fashion ( see below , results , subsection [ sp ] ) .
stripes therefore have a fixed period in terms of number of units ( 100 ) . to find the optimal spatial period of the stripes we vary the shape of each elementary cell in the 300 by 300 array . thus , if the rectangular cell dimensions are @xmath49 perpendicular and @xmath50 parallel to the stripes , we vary both @xmath49 and @xmath50 , keeping the area of elementary cell @xmath51 constant . by doing
so we do not change the density of units , but vary the spatial od period , according to the formula @xmath52 . for each value of @xmath49 the cells are reconnected according to the connection rules .
special care is taken about exclusion of the boundary effects by making sure that none of the units on the edges of the array is connected to .
after the optimum period is found the period in terms of number of units is changed from 100 to another value , closer to the value of spatial period , to check for the absence of geometric artifacts , associated with distortions of elementary cells .
the change of the spatial period after this procedure is typically absent but in extreme cases does not exceed 3% . _ ii ) patches _ since our results indicate that a triangular crystal of _ patches _ is formed ( see fig .
[ phases]j ) , we consider an array in the shape of parallelogram commensurate with the triangular arrangement of _ patches_. the lattice sites in the array , representing units , are also arranged on a triangular lattice .
their positions are given by @xmath53 and @xmath54 , where @xmath55 and @xmath56 are integers varying between 1 and @xmath57 . here
@xmath58 is the period of odp to be optimized .
the centers of _ patches _ are located at points @xmath59 and @xmath60 .
each patch includes lattice sites at the distance from a center determined by the filling fraction of the ipsilateral eye : @xmath61 .
the units within / outside the patch are left / right eye dominated .
the units in the configuration are then represented by the central parallelogram : @xmath62 . for each of the units connections
are made according to the connection rules with @xmath47 .
the optimum period is obtained by varying parameter @xmath58 . to determine the experimental dependence of the odp period on the filling factor , the image of odp in macaque monkey ( horton and hocking , 1996a )
is converted into a digital format . in this format the image is represented by a set of pixels .
a pixel with coordinates @xmath63 and @xmath64 is represented by a number @xmath65 , equal to @xmath24 for the right eye dominated and @xmath25 for the left eye dominated area . for each position in the image
we then determine the local value of the average filling fraction of the ipsilateral eye and the value of local od period .
both these calculations are similar . to do the calculation at a certain point in the map , given by coordinates @xmath66 , we surround the corresponding pixel by a square , containing @xmath67 pixels ( black square in fig . [ monkey1 ] , @xmath68 mm ) .
the dimensions of the square are such that one hand it contains a few odp periods ( about 3 ) , which is needed for averaging , and on the other hand it is small compared to the characteristic dimensions over which the properties of odp change ( @xmath69 1 cm , see fig . [ monkey1 ] ) . to determine the filling fraction we average the scanned image over the square : for position @xmath66 in the map the local value of the average filling fraction is given by @xmath70 , width=297 ] to determine the local value of odp period we perform the fourier transform of the @xmath71 in the square . as a result
we obtain a set of numbers @xmath72 , representing the fourier transform amplitudes , defined on a @xmath67 set of wave vectors @xmath73 .
the spectral power , represented by @xmath74 , is shown in fig .
[ tf ] for one of the points in the pattern , corresponding to _
stripes_. it clearly has a bimodal appearance , indicating the average in the square direction of the stripes .
we then determine the average value of the wave vector , using the formula : @xmath75 ) .
the spectral power in the scale bar is in arbitrary units .
[ tf ] , width=297 ] the value of the mean odp period is then defined as @xmath76 this value for each pixel in the shaded area in fig .
[ monkey1 ] is shown in fig .
we start by finding optimal layouts for three illustrative examples of wiring rules with small numbers of connections , @xmath8 and @xmath9 .
we caution the reader that because of the small numbers of connections phase assignments may seem arbitrary .
these examples are chosen to illustrate our main results which will be confirmed both in the lattice model with large @xmath8 and @xmath9 later in this section and in the continuous model ( chklovskii and koulakov , 2001 ) .
for the first two examples we set equal numbers of left and right dominated neurons , @xmath18 . in the first example each neuron connects with equal numbers of the same - eye and other - eye neurons , @xmath77 . then the optimal layout is the `` chess board '' of left / right neurons , fig.[lat1]a .
this layout is a realization of the _ salt and pepper _ phase , fig.[phases]a , because each neuron has an equal number of left and right neurons among its immediate neighbors . to calculate the length of connections per neuron , @xmath58
, we notice that in this layout all neurons have the same pattern of connections . by considering one of them ,
fig.[lat1]a , we find that @xmath78 .
this layout is optimal because each neuron makes all of its connections with immediate neighbors . and @xmath77 .
( a ) a realization of the _ salt and pepper _ phase gives minimal wire length ( lattice constants per neuron ) .
( b ) a realization of the _ stripe _ phase is suboptimal ( ) .
[ lat1 ] , width=297 ] a suboptimal layout for the same wiring rules is illustrated by a realization of the _ stripe _ phase , fig.[lat1]b . in this layout
each neuron has the same pattern of connections up to a mirror reflection . by considering one of them ,
fig.[lat1]b , we find @xmath79 , greater than @xmath80 for the _ salt and pepper _ phase . here
each neuron has among its immediate neighbors only three other - eye neurons , while the wiring rules require connecting with four other - eye neurons .
a connection to a more distant neighbor is longer making the layout suboptimal .
we confirm the optimality of the _ salt and pepper _ phase for @xmath81 for large @xmath8 , @xmath9 both numerically and analytically . and
@xmath82 , @xmath83 . ( a ) a realization of the _ salt and pepper _ is suboptimal ( ) .
( b ) a realization of the _ stripe _ phase gives minimal wire length ( ) .
[ lat2 ] , width=297 ] in the second example each neuron connects with more same - eye than other - eye neurons : @xmath82 , @xmath83 .
then a realization of the _ salt and pepper _ phase , fig.[lat2]a is not optimal anymore .
the length of connections per neuron is @xmath84 , while the _ stripe _ phase , fig.[lat2]b gives @xmath80 .
the _ salt and pepper _ phase loses in wiring efficiency because there are not enough same - eye neurons among immediate neighbors and connections with more distant neighbors are needed .
the _ stripe _
phase , fig.[lat2]b rectifies this inefficiency by having each neuron make connections only with immediate neighbors . thus , clustering of same - eye neurons is advantageous if each neuron connects more with the same - eye than with the other - eye neurons . and
@xmath82 , @xmath83 .
realizations of the ( a ) _ salt and pepper _ ( ) and ( b ) _ stripes _ ( ) are suboptimal .
( c ) a realization of the _ l - patch _ phase gives minimal wire length ( ) .
[ lat3 ] , width=297 ] in the third example we use the same wiring rules ( @xmath82 , @xmath83 ) but take different numbers of left / right neurons , @xmath85 , @xmath86 . the realizations of the _ salt and pepper _ phase is shown in fig.[lat3]a and of the _ stripe _ phase in fig.[lat3]b . in these layouts ,
different neurons have different patterns of connections . to find the wiring length per neuron we average over different patterns and find for the _ salt and pepper _ phase @xmath87 and for the _ stripe _ phase @xmath88 .
a more efficient layout is the _
l - patch _
phase , fig.[lat3]c , where @xmath89 .
although we can not prove that the _ l - patch _ phase is optimal , this seems likely .
thus , the optimal shape of monocular regions depends on the relative numbers of left / right neurons . after giving some examples of odps with small numbers of connections @xmath8 and @xmath9
we discuss the opposite case of large numbers . as we show below in section [ sp ] , the shape of od columns in this case
does not depend on the absolute values of parameters @xmath8 and @xmath9 .
it is determined by the ratio @xmath42 and by the relative amount of ipsilateral neurons @xmath16 ( assuming that the left eye is ipsilateral ) .
depending on the values of parameters @xmath42 , and @xmath16 , optimal layout belongs to the one of the eight phases shown in fig .
[ phases ] , where ipsilateral and contralateral - eye dominated neurons are shown by black and white regions respectively .
these phases can be divided into three major classes . the first class is represented by the unsegregated _ salt and pepper _ layout , in which two types of neurons are uniformly intermixed ( figure [ phases]a ) .
the second class includes _ stripy _ layouts , shown in figures [ phases]c , e , g , i. the third class consists of _ patchy _ layouts , displayed in figures [ phases]d , f , h , g. . *
a * : _ salt and pepper _ ; * b * : _ stripes _ mixed with _ patches _ ; * c * and * d * : weakly segregated _ stripes _ and _ patches _ obtained by the perturbation theory ; * e * and * f * : weakly segregated _ stripes _ and _ patches _ obtained by simulated annealing ; * g * : modulated _ stripes _ ; * h * : elongated _ patches _ ; * i * : sharp _ stripes _ ; * j * : sharp _ patches _ ; [ phases ] , width=192 ] we distinguish several subclasses of _ stripy _ phases .
first , it is the sharp _ stripes _ ( figure [ phases]i ) , which consists of alternating lamellar monocular regions .
second , it is the weakly segregated _ stripes _ ( figure [ phases]c , e ) . in this odp
the variation of density of left / right eye dominated neurons is small .
this is an intermediate pattern between the unsegregated _ salt an pepper _ and the sharp _ stripe _ layouts .
this phase is therefore fragile and difficult to obtain numerically . in some cases
simulated annealing can produce such a phase , figure [ phases]e . in the other cases the weak segregated phase
can only be obtained by the perturbation theory , which can carefully account for a weak variation of neuronal density .
such case is shown in figure [ phases]c .
third , we also obtain _ stripy _ phases that show a tendency to become _ patches _
, by e.g. their longitudinal modulation , such as shown in figure [ phases]g .
similar subclasses exist among _ patchy _ layouts .
we obtain sharp , weakly segregated ( obtained from simulated annealing or perturbation theory ) , and elongated _ patches _ , which are shown in figures [ phases]j , f , d , and h respectively . finally , there are mixed phases containing both _ stripes _ and _ patches _ , such as in figure [ phases]b .
these odp s are shown on the phase diagram ( pd ) in figure [ phd ] .
the phase diagram shows the optimum phase ( minimizing the total wire length ) for given values of parameters @xmath42 and @xmath16 . the important feature of the pd is its left - right eye symmetry .
it is apparent from the symmetry of figure [ phd ] with respect to the line @xmath90 .
this is a consequence of the left - right eye symmetry of our model , implying that the connection rules , defined by numbers @xmath8 and @xmath9 are independent on whether a neuron is left or right - eye dominated . for this reason
the phase for @xmath91 can be obtained from the point with the same @xmath42 and the value of the filling fraction equal to @xmath92 .
this corresponds to the replacement of the white regions in figure [ phases ] by black and vice versa .
a similar correspondence takes place when one compares odp s in left and right hemisphere . and @xmath42 .
for the color key see figure [ phases ] .
[ phd ] , width=297 ] another important feature of the pd is the existence of the _ salt and pepper _ region around the line @xmath93 .
this implies that the difference between @xmath8 and @xmath9 is the driving force of segregation into odp .
the larger the difference , the sharper the odp becomes .
the area of the pd adjacent to @xmath18 is occupied by _
odps . at small values of the filling
fraction the phases are _
patchy_. a transition from _ stripes _ to _ patches _ occurs at @xmath94 almost independently on parameter @xmath42 .
an example of such transition for @xmath95 is shown in figure [ transition ] .
, width=297 ] the reasons for the transition for _ small _ values of @xmath96 are discussed in koulakov and chklovskii , 1999 . for _
larger_differences , when odp becomes sharp , the transition occurs due to the presence of surface contribution to the wire length . to demonstrate this we present the following argument , which is rigorously valid in the asymptotic limit of large number of connections to the same - eye neurons , i.e. @xmath97 . in this
limit connections to the same - eye neurons are the most abundant and therefore most costly , from wire length prospective .
hence , we can disregard connections to the opposite - eye neurons in the first approximation .
consider a unit near the interface between two od columns ( top unit in fig .
[ connections ] ) .
the connection arbor of this unit to the same od units , shown by empty circles in fig .
[ connections ] , is strongly biased toward the home column , since the unit has to maintain certain number of connections there .
this effect has been seen by katz et al .
, 1989 , in macaque striate cortex ( see also the discussion in the model section above ) . for units away from the interface the connection arbor within the same od column is close to a circle ( fig .
[ connections ] bottom unit ) .
thus , circular arbor renders the minimum wirelength in the absence of constraints , such as the interface between od columns . with the interface
present the connection arbor to the same eye neurons is therefore strongly deformed with respect to the optimum .
hence , the presence of the interface between the od columns implies an increase in the wirelength , and is therefore associated with a surface cost ( similar to malsburg , 1979 ) .
this surface cost drives the transition from _ stripes _ to _ patches_. indeed if @xmath98 the patchy phases have much shorter length of the surface compared to _ stripes_. this is because _ patches _ shrink when @xmath99 reducing their surface length , whereas _ stripes _ become narrower , keeping their surface length the same .
however , this is not true for @xmath18 where _ stripes _ have a shorter surface for numerical reasons .
therefore , when @xmath16 is decreased , the _ stripes _ should unavoidably condense into _ patches _ to minimize the surface cost .
this is similar to droplets of water assuming circular shape to minimize the surface energy .
, width=297 ] we conclude therefore that two important transitions occur in our model . *
the transition from unsegregated _ salt and pepper _ to weakly segregated and then sharp odp is driven by the difference between parameters @xmath8 and @xmath9 characterizing the intra - cortical circuitry .
* the transition from _ stripy _ to _ patchy _ odp is driven by the decreasing filling fraction of the ipsilateral eye and occurs at @xmath100 , if left eye is underrepresented .
our phase diagram in fig .
[ phd ] shows that the transition from _ stripes _ to _ patches _ occurs when @xmath101 for a wide range of @xmath42 .
this value will be compared now with the value of @xmath16 at which the transition occurs in the experiment , found from the relative area occupied by left eye dominated neurons .
the conclusion that the _ patch _ phase becomes optimal when contralateral eye dominates is , indeed , non - trivial , because there may be a system of alternating wide and narrow monocular stripes instead .
we test our conclusion on the data from macaque monkey first ( horton and hocking , 1996a ) .
the relative area occupied by the left / right eye depends on the location in v1 . in the parafoveal part of v1 both eyes
are represented equally , i.e. @xmath102 .
odp has a stripy appearance , in agreement with the phase diagram .
away from the foveal region contralateral eye becomes dominant .
the odp becomes patchy there ( levay et .
al . , 1985 ) , just as expected from the theoretical phase diagram .
we verify the location of the transition by using the following algorithm .
we find @xmath16 for each point of the pattern by calculating the relative area occupied by the left / right regions in a window centered on that point and a few od periods wide ( dashed lines in fig .
[ macaque ] ) .
then we draw a contour corresponding to @xmath103 , fig .
[ macaque ] .
we observe in fig .
[ macaque ] that stripes indeed become patchy at the black contour indicating @xmath104 . .
shown is a fragment of the macaque odp from horton and hocking ( 1996a ) .
areas dominated by the left eye are grey and neurons dominated by the left eye are white .
the black contour corresponds to @xmath103 averaged over a window , whose dimensions are shown by the dashed square ( @xmath68 mm ) .
the points of transition from _ stripes _ to _ patches _ coincides with the black contour .
[ macaque ] , width=297 ] in _ cebus _ monkey the odp has a similar transition ( rosa et al . , 1992 ) . for monkey co6l from rosa et al
, 1992 , we determine visually that along the horizontal meridian the transition occurs at the eccentricity of @xmath105 . according to the plot of the relative representations given in rosa et al .
, 1992 , @xmath16 changes in the range @xmath106 at these eccentricities .
our theoretical conclusion about a transition at @xmath103 falls into this interval . for the upper @xmath107 degree meridian of the same monkey
the transition occurs at the eccentricity of @xmath108 degrees or at filling fractions @xmath109 .
again , the predicted value belongs to this interval .
we conclude that these data are consistent with the results of our model . in cats
the odps have a patchy appearance ( anderson et al . , 1988 ; jones et al . , 1991 ) .
in this case our theory implies that one of the eyes should dominate . according to some reports ( shatz and stryker , 1978 ; crier et al . , 1998 )
the filling fraction of the contralateral eye in cat v1 is about @xmath31 in young animals ( before p22 ) .
this may lead to _ patches _ in cat v1 .
the strong contralateral bias disappears in older animals ( crier et al . , 1998 ) .
this is consistent with other reports ( anderson et al . ,
1988 ) that both eyes are represented almost equally in adult cats .
one of the general features of our model is that the period of od pattern becomes larger , when the total number of connections is increased .
indeed , the size of the connection arbors grows if both @xmath8 and @xmath9 are increased , given that the density of units ( @xmath110 ) is kept constant .
this is because one has to go further to find the necessary number of connections to satisfy the wiring rules . since dimensions of the connection arbors set up a characteristic scale for the od pattern , the period of the latter also increases .
this property of our model , which we call _ scalability _ , is discussed in this subsection .
let us define scalability in a mathematically rigorous fashion .
assume that one has found a minimum wire length configuration for certain set of parameters @xmath8 , @xmath9 , and @xmath16 ( or @xmath111 ) .
assume that both @xmath8 and @xmath9 are very large .
in this case the dimensions of connection arbors are much larger than the lattice spacing , and one can ignore the fine structure imposed on the connection arbors by the square lattice .
this is exactly the limit in which our model has some validity , both because realistic numbers of neuronal connections are large and because we would like to avoid artifacts introduced by the square lattice .
an example of connection arbors for a neuron for some set of @xmath8 and @xmath9 are shown in fig .
[ scalaba ] ( left ) .
this neuron and its connection arbors resemble the top neuron , marked by the star , in fig . [ connections ] .
the connection arbors in fig .
[ scalaba ] look like continuous circular pieces , due to the large @xmath8 and @xmath9 limit ( square lattice makes the boundaries of connection arbors look like staircases , whose steps are too small to show in the picture for large @xmath8 and @xmath9 ) .
imagine now a geometric transformation , in which the dimensions of the connection arbors of all of the neurons , as well as the od pattern itself , are blown up by the same scaling factor @xmath112 .
the two - dimensional density of the neurons must be preserved during this transformation .
the obtained new od pattern and the new connection arbors are shown schematically in fig .
[ scalaba ] ( right ) .
since the transformation is applied to the two - dimensional objects , and each of the dimensions was stretched by the factor @xmath113 , each neurons in the new configuration will receive @xmath114 and @xmath115 connections from the same and opposite eye neurons . the newly obtained configuration ( fig .
[ scalaba ] right ) will satisfy wiring rules with connection numbers given by @xmath114 and @xmath115 .
note that the filling fraction is not changed by this transformation .
it is @xmath18 in fig .
[ scalaba ] .
will this be the minimum wire length configuration for this set of parameters ? to prove that the new configuration minimizes the total wire length for the new set of parameters @xmath114 and @xmath115 we notice that the total wirelength for the new configuration is given by @xmath116 , where @xmath29 is the total wirelength for the original configuration ( fig .
[ scalaba ] left ) .
this is because each neuron now receives the number of connections increased by @xmath117 , and each connection is stretched by @xmath113 .
imagine now that one finds a new configuration , which has the total connection length @xmath118 .
let us take this more optimal configuration and shrink it by the factor of @xmath113 .
we obtain a configuration , satisfying wiring rules for the set @xmath8 and @xmath9 , whose total wirelength is @xmath119 . but this contradicts to our postulate that the original configuration in fig .
[ scalaba ] ( left ) is optimal for the set of parameters @xmath8 and @xmath9 .
thus the stretched configuration provides the minimum of the wirelength for the new set of parameters @xmath114 and @xmath115 .
this property is important , because once the solution for given @xmath8 and @xmath9 is found , one can obtain other solutions , with the set of parameters @xmath114 and @xmath115 , by stretching the original configuration by the factor of @xmath113 uniformly in all the directions . the important property which remains the same for all these related configurations is the ratio between the numbers of the same and other eye connections , @xmath42 .
thus , we conclude that this ratio determines the shapes of the od patterns , which is unchanged during the uniform stretching procedure . what is changed in the uniform stretching
is the odp period ?
since the period is proportional to the stretching parameter @xmath113 , while the total number of connections is proportional to @xmath117 , we conclude that the period is proportional to the square root of the total number of connections , when the ratio @xmath42 is kept constant .
this implies that @xmath120 here @xmath121 , where @xmath122 is the size of the microcolumn unit .
parameter @xmath123 has a meaning of the average axonal arbor diameter .
the quantity @xmath124 is the _ normalized od period _ , calculated in the units of the average axonal diameter .
this quantity is introduced here for easier comparison to the experiment .
notice that this quantity does not depend on the total number of connections .
the latter dependence is entirely absorbed by the parameter @xmath123 .
scalability is valid for the limit of large @xmath8 and @xmath9 , when square lattice effects can be ignored , and our model becomes continuous .
does scalability apply to our numerical solution , described in subsection [ exenum ] ? to check this we plot the ratio @xmath125 , obtained using methods described in [ exenum ] , for different values of the total number of connections @xmath126 in fig .
[ scalab ] . if eq .
( [ scalab_form ] ) is valid , this ratio should not depend on the total number of connections . as evident from fig .
[ scalab ] this property is indeed satisfied .
hence , below in this subsection we always present the results for @xmath127 , which are independent on the total number of connections , assuming that the latter dependence can be easily recovered using eq .
( [ scalab_form ] ) . ,
top neuron .
the stretched configuration is shown on the right .
[ scalaba ] , width=297 ] we now examine the dependence of normalized period @xmath128 [ see eq .
( [ scalab_form ] ) ] on the parameter @xmath42 , for @xmath90 , when we have to consider the stripe phase , according to subsection [ phdsection ] .
the results are shown in fig .
these results have been obtained using methods described in subsection [ exenum ] . in general , the normalized period increases with increasing parameter @xmath42 .
this increase in the od period can be understood considering the interplay between connections to the same and opposite eye units .
indeed , the presence of connections between the same eye units implies affinity between the same od neurons .
an increase in the relative number of such connections ( @xmath42 ) strengthens such affinity .
the od columns provide a neighborhood rich with the same eye neurons ; even more so , on average , for coarser domains .
thus stronger affinity between the same eye neurons ( @xmath42 ) leads to an increase in the size of od domains .
this effect is produced by wiring economy principle , since the latter provides a substrate for the affinity of connected neurons .
, width=297 ] we now examine fig .
[ fs ] in more detail .
the relative period diverges in the limit @xmath129 .
the divergence can be described by the asymptotic formula @xmath130 shown in fig .
[ fs ] by the dotted curve .
the divergence can be understood as follows .
imagine that the neurons do not have to connect to the neurons of the opposite od , i.e. parameter @xmath131 , @xmath132 , which corresponds to the extreme case @xmath129 . in this case
the optimum wire length configuration consists of only two large domains , dominated by left and right eye neurons , occupying a half of v1 each .
this is because of the notion of surface contribution , introduced in subsection [ phdsection ] . to minimize this interface contribution the system phase segregates into two large domains
thus , in the case @xmath131 odp has maximum possible period , spanning the entire v1 .
this explains the tendency of the period diverge in the limit @xmath133 ( @xmath134 ) in fig .
[ fs ] . what happens if @xmath133 ? since the neurons now have to connect to the opposite eye neurons , phase segregated configuration ( two large domains spanning the entire v1 ) is no longer optimum .
the system introduces more interfaces between od columns to shorten distances between opposite eye neurons .
more interfaces implies a reduction in the od period .
thus , finite @xmath42 leads to the finite od period .
this is reflected by the asymptotic dependence ( [ asymp_p ] ) and the dotted curve in fig . [ fs ] .
an interesting phenomenon observed in fig .
[ fs ] is the presence of a few discontinuous changes in the od period .
one such a change occurs at @xmath135 and consists in an abrupt increase in the od period by a factor of about @xmath136 .
another discontinuous transition occurs at @xmath137 and the corresponding increase in the period is by a factor of @xmath138 .
note that these transitions are truly discontinuous , i.e. they occur at discrete points .
they appear smooth in fig.[fs ] due to a sparse sampling ( the real data points are shown by dots ) .
note also that the quantity @xmath123 in eq .
( [ scalab_form ] ) changes negligibly in the interval between @xmath139 and @xmath140 , which implies that both od period @xmath141 and the normalized period @xmath142 change approximately by the same factor . such discontinuous changes in the od period in the weakly segregated regime , i.e. when the odp is not well defined , may be responsible for the coarsening of odp in strabismic squirrel monkeys ( see discussion for more details ) .
the dependence of the normalized period @xmath124 on the filling fraction of the left eye @xmath16 is shown in fig .
these results have been obtained using methods described in subsection [ exenum ] .
four dependencies are shown , for four values of the parameter @xmath42 marked on each curve . the general tendency for the period to grow with increasing parameter @xmath42 , described in the previous subsection , is evident in the figure .
, width=297 ] for small values of parameter @xmath42 the period increases when the filling fraction moves away from @xmath90 , as long as one stays within the same phase ( _ stripes _ or _ patches _ ) .
this behavior is seen for the two bottom curves in fig .
[ fr ] . in the transitional region between _ stripes _ and _ patches _ the od period experiences a discontinuity , marked by the dotted lines . for large values of @xmath42 ,
the dependence of the period on @xmath16 is opposite : the period decreases , as the filling factor deviates from @xmath19 , as demonstrated by the top curve in fig .
[ fr ] .
, width=297 ] we now compare this behavior of our model to the observations in the striate cortex of macaque monkey ( horton and hocking , 1996a ) , using fourier transform method described in subsection [ fourier ] . to make this comparison possible the following technical consideration
is taken into account .
the fourier transform method evaluates the average value of the spatial frequency of the odp @xmath143 .
the od period is then calculated by the formula @xmath144 .
for the _ stripe _ phase this procedure results in the value , which is close to the period of stripes . for _ patches _ it results in the distance between rows of patches , which is smaller than the period by the factor @xmath145 ( see fig .
[ rows ] ) .
thus , to make comparison to the experiment possible , the values in fig . [ fr ] corresponding to _ patches _ should be multiplied by the factor @xmath146 .
the result of this procedure is shown in fig .
[ fp ] by the gray line . ) .
the distinction between _ patch _ period and distance between rows should be taken into account for accurate comparison to the experimental observations .
[ rows ] , width=240 ] fig .
[ fp ] shows that the period observed in the experiment decreases when the filling factor of the ipsilateral eye deviates from @xmath19 .
this warrants the use of the top curve in fig .
[ fr ] to represent the theoretical result .
since , the shape of the theoretical dependence does not change much when @xmath147 , the parameter @xmath42 can not be established from the comparison of the theory to the experiment . to obtain the gray curve in fig .
[ fp ] the @xmath148 dependence in fig .
[ fr ] was multiplied by the fitting parameter @xmath149 mm .
this is the only fitting parameter used . as seen in fig .
[ fp ] , our theory describes the experimentally observed dependence quite well . .
the latter is the top curve in fig .
[ fr ] , with the sector of the data corresponding to _ patches _ corrected by the factor for compatibility with the fourier transform .
the only fitting parameter used is @xmath149 mm [ see eq .
( [ scalab_form ] ) ] . [ fp ] , width=297 ]
the widths of the ipsilateral and contralateral eye stripes in macaques has also been measured independently by tychsen and burkhalter ( 1997 ) . based on their results
one can evaluate the odp period and the filling fraction : @xmath150 here @xmath151 , @xmath152 , and @xmath153 are the filling fraction of the ipsilateral eye , and the ipsilateral / contralateral column widths respectively .
the dependence of the period on the filling fraction can therefore be established .
this dependence is shown in fig.[tychsen ] .
the best parabolic fit to the data in fig.[tychsen ] is given by : @xmath154.\ ] ] the coefficient @xmath155 is estimated using bootstrap ( efron and tibshirani , 1993 ) .
the expectation value of the coefficient is therefore below zero , as seen from fig.[tychsen ] .
the probability of the coefficient to be greater than zero is @xmath156 as evaluated by bootstrap procedure , which is used since the distribution of coefficients @xmath157 is non - gaussian .
this implies that with great degree of certainty one can assume that the period of odp does decrease with the filling fraction deviating from @xmath19 .
it should be noted that the value of coefficient @xmath157 can be obscured by the variability of odp period between different individuals , since data in fig.[tychsen ] are assembled from three monkeys ( four v1 s ) . to reduce the impact of inter - individual variability we then normalized the period for each area v1 by the mean value for each individual animal .
the value of the coefficient is then @xmath158 , with the probability of positive coefficient @xmath159 .
thus the decrease of the period with filling fraction is even more certain , when the inter - individual variability is accounted for .
the value of coefficient @xmath157 obtained from the theory is @xmath138 ( fig . [ fr ] , @xmath148 ) .
it is consistent with both measurements .
, width=297 ]
in this work we propose a model which can account for most of experimentally observed features of odps .
our model has two principal parameters .
the first parameter characterizes the intracortical circuitry .
it is the difference between the number of connections to the same and to the opposite od neurons .
our results suggest that this difference is the driving force of segregation into odps .
we argue therefore that the sensitivity of the intra - cortical connectivity to od provides a reason to formation of od columns ( see below ) . the second parameter is the fraction of neurons dominated by the ipsilateral eye .
this parameter determines the shape of monocular regions in odp . in the majority of the primary visual area of macaque and _
cebus _ monkeys this parameter is close to @xmath160 , which implies that both ipsi- and contralateral eyes are equally represented .
however , in the proximity of monocular crescent the ipsilateral eye becomes underrepresented .
this is because the inputs into the eye are blocked by the nose of the animal .
our theory suggests that the decrease in the filling fraction of the ipsilateral eye drives the transition in the odp structure from stripy ( zebra skin like ) to patchy ( similar to leopard skin ) . the transition occurs when the fraction of the ipsilateral eye dominated neurons approaches @xmath0 in both macaque and _ cebus _ monkeys ( see below ) .
we also analyze the dependence of od period on the parameters of our model and find satisfactory agreement with experimental data .
each neuron in our model establishes certain number of intra - cortical connections with neurons dominated by the same and the opposite eye . as suggested by experimental studies in macaque striate cortex
, neurons in layer @xmath1 typically make more connections with neurons of the same od ( katz et al . , 1989 ) .
thus , from wiring economy prospective , connections with neurons of the same od are more important than the opposite eye connections .
therefore , it is advantageous to form od columns , since they provide environment rich with the same od neurons , which results in shortening connections to the same eye neurons .
the wiring economy principle thus provides a natural reason for the existence of od patterns , i.e. answers the first question in the program listed in the introduction .
our model suggests that in primates with weakly defined od columns , such as squirrel monkey ( hubel et al . , 1976 ; livingstone , 1996 ; horton and hocking , 1996b ) and owl monkey ( livingstone , 1996 ) , the difference between these two components of intracortical connectivity is not large .
such difference may be increased in these animals by experimentally induced strabismus . indeed , strabismus reduces correlated activity between opposite od cortical neurons , therefore reducing their connectivity @xmath9 .
reduction in @xmath9 unbalanced by the corresponding reduction in @xmath8 increases the parameter @xmath42 and leads to sharpening of od columns , according to our phase diagram in fig .
[ phd ] . such sharpening
is indeed observed experimentally ( shatz et al . , 1977 ; lowel , 1994 ; livingstone , 1996 ) .
this phenomenon was also predicted theoretically by goodhill ( 1993 ) .
the two parameters of intra - cortical circuitry , @xmath8 and @xmath9 , represent in our model the interplay between two classes of processing performed by the visual cortex .
the first class includes the processing of the monocular image , for which connections to the same od neurons are necessary .
the second class includes various tasks related to stereopsis , which require comparison of two monocular images , relying on the connections between the opposite od neurons .
we proposed above that the function of od columns is to shorten the connections between the same eye neurons
. the impact of stereoscopic vision should therefore be the opposite : strong stereoscopy should make odp weaker .
inversely , weak stereoscopy induces sharp odps .
the latter statement is justified by the observations in animals with experimental strabismus .
however , one should be careful about this statement , since the relation between functional and anatomical properties may not be direct .
the situation in the animals with lateral eye positioning , such as mice , rats , tree shrews , etc . , is different . in such animals
the visual pathway is almost completely crossed , i.e. v1 in each hemisphere is strongly dominated by the contralateral eye [ drager , 1974 , 1975 , 1978 ; drager and olsen , 1980 ; gordon and stryker , 1996 ( mouse ) ; hubel , 1977 ( rat ) ; casagrande and harting , 1975 ; mully and fitzpatrick , 1992 ( tree shrew ) ; horton and hocking , 1996b ( other species ) ] . as suggested by antonini et al .
this implies that the odp contains only two large monocular columns , each spanning the whole hemisphere .
this can be interpreted as an od having very large period , spanning both striate cortices .
this picture can be fitted into the framework of our model .
indeed , we predict that if the number of connections to the other od neurons ( @xmath9 ) is very small the od has a very large period ( fig.[fs ] ) .
thus our model predicts that the number of connections received by each neuron from the neurons of the same od ( @xmath8 ) is much larger than the number of opposite eye connections ( @xmath9 ) in the species with lateral eye positioning .
this should include the cross - hemispheric projections .
this statement should have functional consequences . since @xmath9 is small , synthesis of images from two eyes is weaker .
therefore the animals have to find another strategy to implement stereopsis .
hooded rats for example use vertical head movements to estimate distances ( legg and lambert , 1990 ) .
our conclusion about small @xmath9 should also apply to the superior colliculus in these animals ( colonnese and constantine - paton , 2001 ) , as well as to the tectum in lower vertebrates ( schmidt and tieman , 1985 ) , in which cases the visual inputs cross over almost completely too .
to summarize , our model encompasses most of the phenomena related to the sharpness and observability of odp .
it relates the interspecies variability in the odp to the relative amount of binocular interaction occurring in different species .
thus , we predict , that in the animals with weakly segregated columns ( squirrel monkey ) @xmath40 ; in the animals with sharp columns ( macaque ) @xmath8 is much larger than @xmath9 [ confirmed by katz et al . , ( 1989 ) ] ; and , finally , in the animals with lateral eye positioning , @xmath9 should have a value , whose contribution to the wirelength can be neglected .
another consequence of a decreasing @xmath16 in macaque is a decrease in the odp period ( levay et al . , 1985 ) .
in fig . [ fp ]
we compare the result of our theory to the data from macaque monkey ( fourier transform applied to data from horton and hocking , 1996a ) . from this comparison
we conclude that , according to the wiring economy principle , parameter @xmath161 , or cells establish much more connection with the same od cells , than with the opposite .
this result of is consistent with the observations of od sensitive circuitry in the striate cortex of macaque by katz et al .
( 1989 ) .
we chose the regions proximal to the horizontal meridian for this comparison .
this is based on the assumption that cortical properties , such as @xmath8 and @xmath9 , change little along this meridian .
this assumption is in part supported by the fact that od periodicity changes little on the large segment of the meridian occupied by stripes , spanning the region between about 2 and 25 degree eccentricity ( notice very little scatter in fig .
[ fp ] around the point @xmath90 ) .
the changes in the period begin to occur when @xmath16 deviates from @xmath19 .
this is illustrated by fig .
other authors notice a decrease in the period when comparing vertical to horizontal meridian .
studies based of computer reconstructions report about 2 fold decrease in od periodicity comparing these areas ( levay et .
al . , 1985 )
, while others , based of flat - mounts ( horton and hocking , 1996a ) , report a more moderate change .
such variation can not be accounted for by a decrease in parameter @xmath16 alone , since @xmath16 is about @xmath19 on both meridians in close proximity to parafoveal region ( < 20 degrees of eccentricity ) .
our model suggests two possibilities based on the variation in the intracortical circuitry , described by @xmath8 and @xmath9 . since such differences in the circuitry
may also be responsible for variability of the od period between different animals , we discuss this possibility in the next subsection .
studies in macaque monkeys ( horton and hocking , 1996a ) reveal large inter - individual variability of the stripe period .
the stripe period is @xmath162 along the v1 border , after comparison of 6 animals .
two factors may contribute to this phenomenon in the framework of our model .
( 1 ) the basic diameter of axonal and dendritic arbors @xmath123 varies from animal to animal .
this could be due to changes in @xmath8 , @xmath9 , or neuronal density .
( 2 ) the ratio between monocular and binocular interactions @xmath42 varies .
the former reason is justified by eq . ( [ scalab_form ] ) .
the latter can be understood from fig .
simply speaking , monocular interactions ( @xmath8 ) favor formation of od columns , making them wider , in an effort to provide same od rich environment for all the neurons .
binocular interactions ( @xmath9 ) favor interfaces between columns , since interfaces bring opposite od neurons closer to each other .
they therefore decrease od period .
when @xmath42 increases the od period increases too ( fig .
this may occur when comparing different individuals .
similar consideration may apply to the experiments in strabismic animals ( lowel , 1994 ; livingstone , 1996 ) . since strabismus reduces correlations between eyes , its effect in our model
is to reduce parameter @xmath9 .
hence , the ratio @xmath42 is increased . according to our model ( fig .
[ fs ] ) this generally leads to an increase in the relative od period ( ratio of the basic od periodicity to the connection range @xmath123 ) .
this result is understood from the interplay between affinity between the same eye neurons ( @xmath8 ) , increasing the period , and the affinity between opposite od neurons ( @xmath9 ) , reducing od period . since the latter is reduced by strabismus
, the od period grows .
the degree of the period change depends on the decrease in the number of interocular connection , and is difficult to estimate .
a curious feature displayed by od period in our model is an abrupt increase at @xmath163 by a factor of about @xmath136 , cf .
this implies that close to point @xmath164 the od period may be very sensitive to developmental manipulations .
this finding may have correlate in squirrel monkey , for which the observed increase in period is by a factor @xmath165 ( horton and hocking , 1996b ) .
these data are obtained from comparison to a single strabismic animal .
the following scenario is possible , comparing squirrel monkey to the strabismus experiments in owl monkey ( livingstone , 1996 ) , in which no significant increase in periodicity is observed .
parameter @xmath42 in squirrel monkey passes the point @xmath164 due to strabismus , leading to the substantial increase in period . in owl monkey parameter
@xmath42 is above @xmath166 in normal animal .
strabismus therefore has little effect .
this scenario is consistent with sharper od columns in normal owl monkeys ( @xmath167 ) than in normal squirrel monkeys ( @xmath168 ) ( livingstone , 1996 ; horton and hocking , 1996b ) .
experimentally induced strabismus in cat leads to an increase in the od period by a factor of @xmath169 ( lowel , 1994 ; goodhill , 1993 ; see however jones et al . , 1996 ) .
our model suggests that parameter @xmath167 in cat , and the increase in the period is due to the smooth part of the dependence in fig .
[ fs ] , which may not be so substantial as in squirrel monkey .
the relevance of wiring economy principle to the neuronal spatial organization can be illustrated by the following thought experiment ( koulakov and chklovskii , 2001 ) .
imagine taking a cortical area and scrambling neurons in that area , while keeping all the connections between neurons the same .
since the circuit is unchanged , the functional properties of the neurons remain intact .
therefore , from the functional point of view , the scrambled region is identical to the original one .
in fact , the only difference caused by scrambling is in the length of neuronal connections .
therefore , it is hard ( if not impossible ) to justify the existence of systematic cortical maps , such as od pattern , without invoking the cost of making long neuronal connections .
although some theories of map formation may not explicitly mention the wiring optimization principle , it is present implicitly , usually in requiring the locality of intra - cortical connections .
how important is the constraint imposed by wiring minimization ? in principle
one can imagine development of an organism , which has 30% excess of wire with respect to the existing ones .
it turns out that the existence of such an organism is close to impossible .
indeed , imagine that an external object , such as a blood vessel , is introduced in certain area of the gray matter .
in this case some of the neuronal connection have to go around the vessel , therefore increasing in length .
if the nerve pulses are to be delivered at the original speed and/or intensity , the elongated axons and dendrites have to be made thicker , to increase the pulse propagation speed and decrease dendritic attenuation .
this leads to more obstacles on the way of other neuronal connections and so on .
thus , introduction of a new blood vessel leads to an infinite series of axonal and dendritic reconstructions .
the same is true about the extra connection volume , resulting from wasteful neuronal positioning .
such infinite series of reconstructions can diverge , which implies that the connection volume resulting from more and more reconstructions increases indefinitely . in this case
the new blood vessel can never be inserted without sacrificing significantly the brain function .
it turns out that mammalian brain has reached the verge of this so called _ wiring catastrophe _ ( chklovskii and stevens , 2001 ) , so that it gets increasingly more difficult to accommodate excess volume in the nerve tissues .
the wiring catastrophe occurs when the fraction of axons and dendrites in the cortical volume reaches @xmath170 .
electron microscopy studies of cortical slices show that the actual volume occupied by neuronal processes is about @xmath171 ( chklovskii and stevens , 2001 ) .
the brain therefore has approached the barrier imposed by wiring catastrophe .
further increase of the volume fraction of neuronal processed may deteriorate the brain function . as discussed in the previous subsection ,
wiring optimization is the only known way to relate neuronal layout ( as manifested in the odp ) to the statistics of neuronal connectivity .
models of the odp development that do not explicitly rely on wiring optimization invoke it implicitly , usually requiring the locality of intra - cortical connections . in his pioneering work ,
mitchison ( 1991 ) studied a question whether odp minimize the wiring volume relative to the _ salt and pepper _ layout .
he assumed that the inter- neuronal connectivity is determined both by ocular dominance and retinotopy with all neurons having the same connectivity rules .
he found that the answer to this question depends on the detailed assumptions about axonal branching rules .
in particular it depends on the value of axonal branching exponent @xmath157 .
he has shown that if all axonal segments have the same caliber ( @xmath172 ) , than odp s are indeed advantageous for certain range of ratios of same - eye to opposite - eye connections .
he also showed that if @xmath173 than the odp do not save wiring volume relative to _ salt and pepper_. however , existing data seems to suggest that axonal caliber branches with @xmath173 ( deschenes and landry 1980 , adal and barker , 1965 ) .
the case of axonal branching with the cross - sectional area conservation corresponds closely to our model because we require a separate connection for each neuron . the reason we find that odp minimize wiring length is because we drop the retinotopy requirement on inter - neuronal connection rules , an assumption supported by the experimental data ( katz ) .
although , effectively connections are roughly retinotopic , connection rules may vary from neuron to neuron thus providing some flexibility .
the advantage of our approach is its simplicity allowing us to map out a complete phase diagram and make experimentally testable predictions .
the full theory of the odp will require a detailed analysis of axonal branching which must rely on better knowledge of axonal branching rules .
jones et al .
( 1991 ) proposed an explanation for why odp have either stripy or patchy appearance .
they assumed that neurons are already segregated into the odp ( by considering units whose size equals the width of monocular regions ) and found that the difference between stripy and patchy appearances of the odp could be due to the boundary conditions , i.e. different shape of v1 in different species .
although the correlation between the shape of v1 and the odp layout is observed , the model of jones et al .
does not explain why peripheral representation of macaque v1 has patchy odp or why ocular dominance stripes run perpendicular to the long axis of v1 in some parts of v1 but not in others .
moreover , it is the local structure of odp that is likely to determine the shape of v1 and not the other way around .
therefore , unlike jones et al .
, our work proposes a unified theory of odp including _ salt and pepper _ , stripy and patchy layouts , and is based on local inter - neuronal connectivity rules . another model related to wiring length minimization is the elastic net model studied by goodhill and coworkers ( 1993 ) .
the original formulation of the model minimized the cost function which penalized for placing nearby neurons whose activity is not correlated , a choice justified by computational convenience .
later the elastic model was generalized by the introduction of a c measure .
maximization of c measure effectively corresponds to penalizing for placing correlated neurons far apart . unlike wiring optimization
the penalty does not increase beyond a distance called cortical interaction . because of this , elastic net often yields solutions where left and right eye neurons are completely segregated into left and right eye maps .
our wiring optimization models can be viewed as a sub - set of models described by c measure .
the advantage of our wiring optimization approach is that it has a transparent biological justification for the cost of placing neurons far from each other - the cost of wiring . because of this , wiring optimization is a natural choice for questions related to the anatomy of intra - cortical connections .
wiring optimization provides a link between neuronal connectivity and spatial layout .
thus , it leaves open the connection between connectivity and computational function . unlike most other models , which assume that neurons should be connected only if they are correlated , wiring optimization makes other assumptions about connectivity possible , for example connecting neurons with anti - correlated firing .
our theory relates functional requirements on the neuronal circuits to its structural properties . in particular
, our model relates the amounts of the neuronal intraocular and interocular interactions , and the filling fraction of ipsilateral neurons , to the structure of the ocular dominance pattern .
we conclude that the interspecies variability in the ocular dominance patterns may be explained by differences in the underlying cortical circuitry .
adal mn , barker d ( 1965 ) intramuscular branching of fusimotor fibers .
j physiol 177:288 - 299
. ahmed b , anderson jc , douglas rj , martin ka , nelson jc ( 1994 ) polyneuronal innervation of spiny stellate neurons in cat visual cortex .
j comp neurol 341:39 - 49 .
allman jm , kaas jh ( 1974 ) the organization of the second visual area ( v ii ) in the owl monkey : a second order transformation of the visual hemifield .
brain res 76:247 - 265 .
anderson pa , olavarria j , van sluyters rc ( 1988 ) the overall pattern of ocular dominance bands in cat visual cortex .
j neurosci 8:2183 - 2200 .
antonini a , fagiolini m , stryker mp ( 1999 ) anatomical correlates of functional plasticity in mouse visual cortex .
j neurosci 19:4388 - 4406 .
blasdel gg , fitzpatrick d ( 1984 ) physiological organization of layer 4 in macaque striate cortex .
j neurosci 4:880 - 895 .
cajal sry ( 1995 ) histology of the nervous system of man and vertebrates .
new york : oxford university press .
casagrande va , harting jk ( 1975 ) transneuronal transport of tritiated fucose and proline in the visual pathways of tree shrew tupaia glis .
brain res 96:367 - 372 .
cherniak c ( 1992 ) local optimization of neuron arbors .
biol cybern 66:503 - 510 .
chklovskii db , koulakov aa ( 2000 ) a wire length minimization approach to ocular dominance patterns in mammalian visual cortex .
physica a 284:318 - 334 .
chklovskii db ( 2000 ) binocular disparity can explain the orientation of ocular dominance stripes in primate primary visual area ( v1 ) .
vision res 40:1765 - 1773 .
colonnese mt , constantine - paton m ( 2001 ) chronic nmda receptor blockade from birth increases the sprouting capacity .
j neurosci 21:1557 - 1568 .
cowey a ( 1979 ) cortical maps and visual perception : the grindley memorial lecture .
q j exp psychol 31:1 - 17 .
crair mc , gillespie dc , stryker mp ( 1998 ) the role of visual experience in the development of columns in cat visual cortex .
science 279:566 - 570 .
deschenes m , landry p ( 1980 ) axonal branch diameter and spacing of nodes in the terminal arborization of identified thalamic and cortical neurons .
brain res 191:538 - 544 .
drager uc ( 1974 ) autoradiography of tritiated proline and fucose transported transneuronally from the eye to the visual cortex in pigmented and albino mice .
brain res 82:284 - 292 .
drager uc ( 1975 ) receptive fields of single cells and topography in mouse visual cortex .
j comp neurol 160:269 - 290 .
drager uc ( 1978 ) observations on monocular deprivation in mice .
j neurophysiol 41:28 - 42 .
drager uc , olsen jf ( 1980 ) origins of crossed and uncrossed retinal projections in pigmented and albino mice .
j comp neurol 191:383 - 412 .
efron b , tibshirani r ( 1993 ) an introduction to the bootstrap .
new york : chapman and hall .
erwin e , obermayer k , schulten k ( 1995 ) models of orientation and ocular dominance columns in the visual cortex : a. neural comput 7:425 - 468 .
goodhill gj ( 1993 ) topography and ocular dominance : a model exploring positive correlations .
biol cybern 69:109 - 118 .
gordon ja , stryker mp ( 1996 ) experience - dependent plasticity of binocular responses in the primary visual cortex of the mouse .
j neurosci 16:3274 - 3286 .
horton jc , hocking dr ( 1996 ) intrinsic variability of ocular dominance column periodicity in normal macaque monkeys .
j neurosci 16:7228 - 7239 .
horton jc , hocking dr ( 1996 ) anatomical demonstration of ocular dominance columns in striate cortex of the squirrel monkey .
j neurosci 16:5510 - 5522 .
hubel dh , wiesel tn ( 1965 ) binocular interaction in striate cortex of kittens reared with artificial squint .
j neurophysiol 28:1041 - 1059 .
hubel dh , wiesel tn ( 1969 ) visual area of the lateral suprasylvian gyrus ( clare - bishop area ) of the cat .
j physiol 202:251 - 260 .
hubel dh , wiesel tn ( 1974 ) uniformity of monkey striate cortex : a parallel relationship between field size , scatter , and magnification factor .
j comp neurol 158:295 - 305 .
hubel dh , wiesel tn ( 1977 ) ferrier lecture .
functional architecture of macaque monkey visual cortex .
proc r soc lond b biol sci 198:1 - 59 .
jones dg , van sluyters rc , murphy km ( 1991 ) a computational model for the overall pattern of ocular dominance .
j neurosci 11:3794 - 3808 .
jones dg , murphy km , van sluyters rc ( 1996 ) spacing of ocular dominance columns is not changed by monocular deprivation or strabismus .
invest ophtalmol vis sci 37:1964 .
katz lc , gilbert cd , wiesel tn ( 1989 ) local circuits and ocular dominance columns in monkey striate cortex .
j neurosci 9:1389 - 1399 .
koulakov aa , chklovskii db ( 2001 ) orientation preference patterns in mammalian visual cortex : a wire length minimization approach .
neuron 29:519 - 527 .
levay s , gilbert cd ( 1976 ) laminar patterns of geniculocortical projection in the cat .
brain res 113:1 - 19 .
levay s , connolly m , houde j , van essen dc ( 1985 ) the complete pattern of ocular dominance stripes in the striate cortex and visual field of the macaque monkey .
j neurosci 5:486 - 501 .
livingstone ms ( 1996 ) ocular dominance columns in new world monkeys .
j neurosci 16:2086 - 2096 .
lowel s ( 1994 ) ocular dominance column development : strabismus changes the spacing of adjacent columns in cat visual cortex . j neurosci 14:7451 - 7468 .
mitchison g ( 1991 ) neuronal branching patterns and the economy of cortical wiring .
proc r soc lond b biol sci 245:151 - 158 .
mountcastle vb ( 1957 ) j neurophysiol 20:408 - 434 .
muly ec , fitzpatrick d ( 1992 ) the morphological basis for binocular and on / off convergence in tree shrew striate cortex .
j neurosci 12:1319 - 1334 .
peters a , payne br , budd j ( 1994 ) a numerical analysis of the geniculocortical input to striate cortex in the monkey .
cereb cortex 4:215 - 229 .
rockel aj , hiorns rw , powell tp ( 1980 ) the basic uniformity in structure of the neocortex .
brain 103:221 - 244 .
rosa mg , gattass r , fiorani m , jr .
, soares jg ( 1992 ) laminar , columnar and topographic aspects of ocular dominance in the primary visual cortex of cebus monkeys .
exp brain res 88:249 - 264 .
schmidt jt , tieman sb ( 1985 ) eye - specific segregation of optic afferents in mammals , fish , and frogs : the role of activity .
cell mol neurobiol 5:5 - 34 .
shatz cj , lindstrom s , wiesel tn ( 1977 ) the distribution of afferents representing the right and left eyes in the cat s visual cortex .
j neurosci 16:5510 - 22 .
shatz cj , stryker mp ( 1978 ) ocular dominance in layer iv of the cat s visual cortex and the effects of monocular deprivation .
j physiol 281:267 - 283 .
swindale nv ( 1980 ) a model for the formation of ocular dominance stripes .
proc r soc lond b biol sci 208:243 - 264 .
swindale nv ( 1996 ) the development of topography in the visual cortex : a review of models .
network : computation in neural systems 7:161 - 247 .
tychsen l , burkhalter a ( 1997 ) nasotemporal asymmetries in v1 : ocular dominance columns of infant , adult , and strabismic macaque monkeys .
j comp neurol 388:32 - 46 .
von der malsburg c ( 1979 ) development of ocularity domains and growth behaviour of axon terminals .
biol cybern 32:49 - 62 .
wiesel tn , hubel dh , lam dm ( 1974 ) autoradiographic demonstration of ocular - dominance columns in the monkey striate cortex by means of transneuronal transport .
brain res 79:273 - 279 .
young mp ( 1992 ) objective analysis of the topological organization of the primate cortical visual system .
nature 358:152 - 155 . | we study a mathematical model for ocular dominance patterns ( odps ) in primary visual cortex . this model is based on the premise that odp is an adaptation to minimize the length of intra - cortical wiring .
thus we attempt to understand the existing odps by solving a wire length minimization problem .
we divide all the neurons into two classes : left- and right - eye dominated .
we find that segregation of neurons into monocular regions reduces wire length if the number of connections to the neurons of the same class ( intraocular ) differs from the number of interocular connections .
the shape of the regions depends on the relative fraction of neurons in the two classes .
we find that if both classes are almost equally represented , the optimal odp consists of interdigitating stripes . if one class is less numerous than the other , the optimal odp consists of patches of the less abundant class surrounded by the neurons of the other class .
we predict that the transition from stripes to patches occurs when the fraction of neurons dominated by the underrepresented eye is about @xmath0 .
this prediction agrees with the data in macaque and cebus monkeys .
we also study the dependence of the periodicity of odp on the parameters of our model . |
x - ray surveys have consistently indicated the existence of a moderately bright x - ray source near the galactic longitude 41.97@xmath4 and latitude @xmath5 . this source , which we refer to as x 1908 + 075 , was cataloged as 4u 1909 + 07 by uhuru , the first earth - orbiting mission dedicated to celestial x - ray astronomy ( forman et al .
subsequent determinations were cataloged as 1 m 1912 + 077 with mit / oso 7 ( markert et al . 1979 ) , 3a 1907 + 074 with ariel 5 ( warwick et al .
1981 ; bell burnell & chiappetti 1984 ) , 1h 1907 + 074 with heao a-1/lass ( wood et al .
1984 ) , gps 1908 + 075 with exosat / me ( warwick et al . 1988 ) .
there are large uncertainties in the position determinations of these survey instruments , as shown in fig .
the heao a-2 experiment on heao-1 scanned this region and detected x - ray emission of @xmath6 mcrab in the 1.520 kev band ( heasarc archival data ) .
the source was also detected with the modulation collimator experiment ( heao a-3 ) on heao-1 , with an 8@xmath7 detection in each of the two collimators for the energy range of 313 kev ( heasarc archival data ) .
the heao a-3 instrument yields a multiplicity of relatively precise error regions which appear as a grid of diamonds in fig [ pos ] .
the einstein imaging proportional counter ( ipc ) imaged a portion of this region and detected a source , 1e 1908.4 + 0730 , the brightest source in the center of the field , at a significance level of @xmath8 ( heasarc archival data ) .
the ipc flux is about 0.3 mcrab ( 0.53.5 kev ) , possibly highly attenuated by interstellar absorption ( see table 1 ) .
the position ( j2000 ) is @xmath9 and @xmath10 with a positional accuracy of @xmath11 .
this position overlaps one of the heao a-3 `` diamonds '' .
there are a few other 5-@xmath7 detections at the edge of the field of view ( @xmath12 square ) of the ipc , and these sources are located well beyond the celestial map provided in fig .
[ pos ] . furthermore
, none of these sources have positions that are consistent with the heao a-3 diamonds .
it is thus highly likely that the einstein / ipc detection represents the same source as the x - ray survey detections .
no optical counterpart has been identified within the error circle of 1e 1908.4 + 0730 .
+ + fig.1 positions of x 1908 + 075 and other x - ray sources detected previously , with reported error boxes superposed on an optical image taken from the digitized sky survey .
the degenerated positions determined with heao a-3 can be seen as a grid of diamonds .
the position of x 1908 + 075 used in the asm analysis was taken to be that of einstein / ipc source 1e 1908.4 + 0730 , represented as a small circle .
[ pos ] the x - ray spectra and flux available from the hard x - ray survey detections are roughly consistent with each other as summarized in table 1 .
the fluxes are about 5 mcrab in the 26 kev band and 714 mcrab in the 220 kev band .
the spectrum of exosat / me observation of 4u 1909 + 07 in the @xmath13@xmath14 kev band ( at the uhuru position ) indicates excessive absorption at photon energy below @xmath15 kev compared with that of the crab nebula ( heasarc archival data ) .
the observations from the ariel 5 spectrometer show that 3a 1907 + 074 has a hard spectrum which can be fitted with a power law of photon index @xmath16 , hydrogen column density of @xmath17 @xmath18 , and an iron line emission with equivalent width @xmath19 ev ( bell burnell & chiappetti 1984 ) .
none of these pointed observations were long enough to investigate the long - term variability of this source . in this paper , we use the data from the all sky monitor ( asm ) ( levine et al . 1996 ) on board _ rxte _ ( bradt , rothschild , & swank 1993 ) to confirm the position and to study the nature of x 1908 + 075 .
we show that the coded mask patterns for the asm , processed for the einstein / ipc position , yield a persistent @xmath1 mcrab x - ray source with a coherent @xmath20-day periodicity for over 3 years .
we then discuss the implications of this discovery .
the _ rxte_/asm has been monitoring the sky routinely since @xmath21 february .
it consists of three scanning shadow cameras , each consisting of a coded mask and a position - sensitive proportional counter .
the asm is well - suited for our purpose since each of its cameras has a 1.512 kev response , an intrinsic angular resolution of a few arc minutes , and a large field of view of @xmath22(fwhm ) .
in addition , the asm provides observations with relatively good sensitivity ( @xmath23@xmath24 mcrab in a day ) , frequent data sampling ( @xmath2530 times a day ) and long baseline ( thus far 3 years ) . for a 90-s exposure in each asm camera , a set of source intensities for the known sources in the fov
is derived from a @xmath26 fit for the overlapping shadow patterns of the x - ray sources and the x - ray background .
the intensity is computed in each of the three energy bands ( @xmath27@xmath28 , @xmath28@xmath29 and @xmath29@xmath30 kev ) and the results are normalized in units of the count rate for a source at the center of the field of view of camera # 1 . in these units ,
the @xmath27@xmath30 kev flux for the crab nebula is @xmath31 asm ct / s .
the estimated errors of the source intensities include the uncertainties due to counting statistics and a systematic error of @xmath32 determined empirically from the light curves of the crab nebula . a detailed description of the asm and the light curves can be found in levine et al . ( 1996 ) and levine ( 1998 ) .
the x - ray light curve from the asm observations of x 1908 + 075 ( 1996 march to 1999 january ) is shown in fig .
[ ltc_all ] .
the intensity points are 5.0-d average of the intensities derived from the individual 90-s dwells for the assumed position of the einstein / ipc source .
it is apparent that x 1908 + 075 has generally persisted at an intensity level of about 0.6 ct / s , or 8.0 mcrab in the 1.5 - 12 kev band for about three years .
the spatial response function of the asm implies that these photons arise within 12 arcmin of the einstein / ipc source position , thus confirming its association with x 1908 + 075 .
the average x - ray intensity of x 1908 + 075 is well below the detection threshold ( 30 mcrab ) for a single camera exposure of the asm .
the effort to determine an independent source position using asm data alone , e.g. , with the superposition of cross - correlation maps , would be limited by the appearance of systematic noise feature with amplitudes that are a substantial fraction of the average source intensity .
the average values of the hardness ratios of x 1908 + 075 are hr1=@xmath33 , hr2=@xmath34 respectively , where hr1 is defined as the ratio of the asm count rates in the @xmath28@xmath29 kev band to that of the @xmath27@xmath28 kev band , and hr2 is the ratio of the count rates in the @xmath29@xmath30 kev band relative to the @xmath28@xmath29 kev band .
x 1908 + 075 has a much harder x - ray spectrum than either the crab nebula ( hr1=0.9 , hr2=1.1 ) or cyg x-1 in the low - hard state ( hr1=1.0 , hr2=1.5 ) .
+ fig . 2 @xmath35/asm light curves of x 1908 + 075 for the period @xmath21 february to @xmath36 november .
the data are binned in @xmath37-day bins .
only @xmath38 detected data points are plotted .
mjd @xmath39 ( mjd = jd-2400000.5 ) corresponds to @xmath36 july @xmath15 . there is a persistent x - ray emission at @xmath40 ct / s ( 8 mcrab ) for 3 years .
[ ltc_all ] periodicities of x 1908 + 075 were sought by means of the lomb - scargle periodograms of both the light curves and the derived hardness ratios .
the lomb - scargle periodogram ( lomb 1976 ; scargle 1982 ; press et al .
1992 ) was used for the calculation instead of the fast fourier transform ( fft ) since the asm data points are unevenly spaced in time . with this periodogram , a maximum in the power occurs at the frequency which gives the least squares fit to a sinusoidal function .
the expected average power for random noise is @xmath41 .
we oversampled the power spectrum so that the frequencies are a factor of 4 more closely spaced than @xmath42 , where @xmath43 is the total duration of the data used .
the goal is to ensure the detection of a peak for a signal that is of marginal statistical significance and to best locate the peak .
the result is shown in fig .
[ power ] .
there is a distinct narrow peak in the power spectra at a frequency that is corresponding to @xmath44 d. the width of the power peak is 0.009-d , consistent with that of a pure sinusoidal wave .
the uncertainty of this period is estimated to be @xmath45 d , assuming a single signal with gaussian noise . with 3 years of source coverage
, the uneven data spacing would not degrade the uncertainty to any noticeable degree ( horne & balinuas 1986 ) . if we assume the noise follows an exponential distribution as expected for random noise , the false - alarm probability of this detection is less than @xmath46 .
in addition , there is a low - frequency peak corresponding to about @xmath47-d period . except for these two peaks
the overall power is flat with an average of about @xmath48 .
we have also conducted another independent periodicity search using the epoch folding technique ( leahy et al .
1983 ; davies 1990 ) . the data were folded modulo a trial period and then grouped by phases . the resulting deviation from uniform random noise can then be detected using a @xmath49 test . for periods between 0.017 days ,
a distinct peak indicating the maximum deviation was found around @xmath44 d with @xmath50 false alarm probability level .
this is consistent with the results of the lomb - scargle periodogram .
3 the lomb - scargle periodogram of the light curves of x 1908 + 075 .
the dotted line indicates the @xmath51 confidence level assuming an exponential distribution of the noise powers . there is a distinct peak at @xmath20-d period . the 180-d period is probably caused by scattered solar x - ray contamination .
[ power ] the data were then folded modulo the period of the best - fit value of @xmath52 days from the lomb - scargle periodogram to study the intensity variations with phase ( fig .
[ fold ] ) .
+ + fig .
4 folded light curves of x 1908 + 075 .
observational data points are represented as histograms .
phase zero is defined as the phase of the minimum x - ray flux .
the error bars represent one standard deviation .
the solid lines are the best - fit sinusoidal waves .
the most distinctive feature is the nearly sinusoidal modulation with large amplitude .
[ fold ] the most distinctive feature in the folded light curves is the nearly sinusoidal modulation of large amplitude .
we fitted a sinusoidal wave plus a constant to the folded light curves .
the best fit yields a @xmath53 amplitude of the sinusoidal wave in the @xmath27@xmath30 kev band , @xmath54 in the @xmath27@xmath28 kev band , @xmath55 in the @xmath28@xmath29 kev band , and @xmath56 in the @xmath29@xmath30 kev band .
the best - fit epoch for the minimum flux is : @xmath57 ) ( jd ) the persistence and coherence of this periodic variation were tested by subdividing the asm light curves into four consecutive @xmath58-d intervals and then searching for periodicities .
the same analyses described above were applied to each of the @xmath58-d light curves . in the calculated power density spectra ,
the signals are still visible as narrow peaks near @xmath59 d in the power density spectra even though they are no longer independently significant .
the best - fit amplitudes and phases of the sinusoidal waves are roughly consistent with each other ( see table [ tab2 ] for a summary ) .
we thus conclude that the modulation is persistent and coherent over the 3-year time interval .
the x - ray emission detected by the asm has been found to arise from the immediate region of the einstein / ipc source , to exhibit a hard spectrum and intensity similar to the historical survey detections , and to be persistent over @xmath60 years .
this together with the einstein ipc and heao a-3 positional correspondence strongly suggest that the position of 1e 1908.4 + 0730 is the accurate location of the sources seen in the historical x - ray surveys .
it is natural to link the observed @xmath20-day period of x 1908 + 075 to the orbital period of a binary .
the periods of x - rays from known sources are believed to be associated with one of the following ( white , nagase , & parmar 1995 ) : ( 1 ) rotational periods of neutron stars in x - ray pulsars , ranging between milliseconds to hundreds of seconds ; ( 2 ) orbital periods of binary systems , ranging from tens of minutes to tens of days , and possibly hundreds of days ; and ( 3 ) `` superorbital '' orbital periods from @xmath61 days to hundreds of days , which in some cases are believed to be the precession period of the accretion disk ( e.g. , levine & jernigan 1982 ) .
it is clear that the coherence and the timescale of the @xmath20-day x - ray oscillation in x 1908 + 075 can be best interpreted as the orbital period of a binary system .
we can then picture x 1908 + 075 as a binary system consisting of a compact object and a mass - donor companion star .
an x - ray binary with a low mass companion ( lmxb ) is unlikely the scenario for x 1908 + 075 , since the known orbital periods of most lmxbs ( white et al . 1995 ) are on the order of hours , much shorter than the detected 4.4 days .
there are only a few known lmxbs having orbital periods longer than a day , among them only cir x-1 ( @xmath62 d ) and her x-1 ( @xmath63 d , a possible lmxb ) exhibit orbital modulations in x - rays .
the modulation in x - rays of cir x-1 manifests itself as periodic flares and dips likely due to a highly eccentric orbit , while her x-1 is an eclipsing system .
such dramatic modulations are not evident in the orbital light curves of x 1908 + 075 .
on the other hand , most hmxbs have orbital periods on the orders of days .
the smooth and nearly sinusoidal modulations in x 1908 + 075 have a lot in common with a supergiant systems such as cyg x-1 .
we therefore suggest that x 1908 + 075 is a high mass x - ray binary ( hmxb ) .
we further suggest that the companion star is a supergiant rather than a be star for the following reasons .
( 1 ) the @xmath20-d orbital period fits in the range favored more by supergiant systems than by be - star systems , as the known orbital periods of the latter are much longer than @xmath20 days ( @xmath64-d ) ( white et al .
( 2 ) the known orbital light curves of be - star systems are all outburst - like rather than sinusoidal ( e.g. , exo 2030 + 375 ) , probably due to the enhanced accretion caused by the the compact object s periastron passage around the be star . ( 3 ) the orbital modulation of a supergiant system are known to be persistent while this is generally not true in be - star systems ( e.g. , 4u 0115 + 63 in asm light curves during 1996 march to 1998 november ) .
the smooth and roughly sinusoidal orbital modulation in the x - rays of x 1908 + 075 suggests that the system could be embedded in a strong stellar wind from the companion star ( e.g. , a supergiant o or b star ) , which absorbs and scatters the x - rays .
the x - ray modulation is then caused by changes in the optical depth along the line of sight to the x - ray source as a function of orbital phase . for
x 1908 + 075 , absorption of x - rays probably dominates over scattering , as the amplitude of the modulation in the @xmath27@xmath29 kev band is larger than in the @xmath29@xmath30 kev band . to explain a modulation with an amplitude of nearly @xmath65 in the 512 kev band
, the difference in the hydrogen column densities between phases 0 and 0.5 is expected to be more than @xmath66 @xmath18 for an absorption and scattering process .
the mechanism of the orbital modulation of x 1908 + 075 could be very similar to that of cyg x-1 , wherein hard state orbital modulation in the asm light curves can be well explained with absorption and scattering in a stellar wind ( wen et al .
another example of such a system could be the newly discovered x - ray pulsar xte j1855 - 026 ( corbet et al .
1999 ) . in this system
, the orbital period is proposed to be 6.1-d as detected by the asm .
the modulation is roughly sinusoidal with an amplitude about @xmath67 in the 1.512 kev band .
it has also been proposed as a supergiant wind accretion system .
the hydrogen column density @xmath68 can be used to estimate the optical extinction in the v band using the empirical relation between the @xmath68 and the color excess @xmath69 ( savage & mathis 1979 ) .
it is found that @xmath70 for @xmath71 @xmath18 , as indicated from the ariel 5 observation of 3a 1907 + 074 .
the optical objects within the einstein / ipc error circle have magnitudes @xmath72 in the palomar sky survey .
to estimate the distance to x 1908 + 075 , we first investigated the possible range of the absolute v - band magnitudes @xmath73 ( from lang 1992 ) known for supergiant stars in binaries with known orbital periods .
we found that @xmath74 . together with @xmath70 and @xmath75
, we obtained @xmath76 kpc as the lower - limit for the distance .
the x - ray luminosity , estimated from the asm count rate ( 1.512 kev ) , is then @xmath77 erg s@xmath78 if we assume the ariel 5 spectrum .
given the @xmath79 value , a more likely scenario is a distance @xmath80 7 kpc estimated from the formula provided by allen ( 1973 ) , where the relation between the extinction of the star light near the galactic plane and the distance is estimated based on the average properties of interstellar absorbing clouds and the grains between them .
we obtained similar distance by comparing directly the @xmath81 value of x 1908 + 075 with that of another galactic plane source grs 1915 + 105 , which is @xmath82 away from x 1908 + 075 and is believed to have a kinetic distance of 12.5 kpc and @xmath83 @xmath84 ( chaty et al . 1996 ) . at the distance of 7 kpc , the estimated x - ray luminosity of x 1908 + 075 is @xmath85 erg s@xmath78 .
finally , the detected @xmath47-d period seems to be caused by scattered solar x - ray contamination .
we found from the asm light curve that three major flares happened at the times when the sun was near the field of view of this source , while flares with smaller amplitudes are evident when the sun is @xmath86 away from the source position .
the latter may be caused by the back - scattering of the solar x - rays off the earth s atmosphere and the instrument .
a persistent hard x - ray emission ( @xmath1 mcrab at 1.512 kev ) of x 1908 + 075 has been detected with the _ rxte_/asm at the einstein / ipc position of @xmath87 , @xmath88 ( j@xmath89 for a 3-year period .
a @xmath20-d period has been dicovered from the asm light curves .
the modulation is nearly sinusoidal , persistent and coherent for 3 years .
the amplitude of the best - fit sinusoidal wave is @xmath53 in the @xmath27@xmath30 kev band .
our results support the simple interpretation that x 1908 + 075 is the same source as 4u 1909 + 07 , 3a 1907 + 074 , 1h 1907 + 074 , gps 1908 + 075 , 1e 1908.4 + 0730 , as well as the source detected by heao a-3 .
we argue that the 4.4-d period is the orbital period of a high mass binary system .
we also suggest that this system consists of a compact object and a supergiant star that produces a strong stellar wind .
the x - ray modulation is then caused by changes in the optical depth along the line of sight to the x - ray source as a function of orbital phase .
allen , c. w. 1973 , astrophysical quantities , 3rd ed .
( london : university of london , athlone press ) , 263 bell burnell , s. j. & chiappetti , l. 1984 , a&as , 56 , 415 bradt , h. v. , rothschild , r. e. & swank , j. h. 1993 , a&as , 97 , 355 chaty , s. , mirabel , i. f. , duc , p. a. , wink , j. e. & rodriguez , l. f. 1996 , , 310 , 825 corbet , r. h. d. , marshall , f. e. , peele , a. g. & takeshima , t. 1999 , , 517 , 956 davies , s. e. 1990 , , 244 , 93 forman , w. , jones , c. , cominsky , l. , julien , p. , murray , s. , peters , g. , tananbaum , h. , & giacconi , r. 1978 , , 38 , 357 horne , j. h. & baliunas , s. l. 1986 , , 302 , 757 lang , k. r. 1992 , astrophysical data i. planets and stars ( springer - verlag new york ) , 144 leahy , d. a , darbro , w. , elsner , r. f. , weisskopf , m. c. , sutherland , p. g. , kahn , s. & grindlay , j. e. 1983 , , 266 , 160 levine , a. m. & jernigan , j. g. 1982 , , 262 , 294 levine , a. m. , bradt , h. , cui , w. , jernigan , j. g. , morgan , e. h. , remillard , r. , shirey , r. e. , & smith , d. a. 1996 , apj , 469 , l33 levine , a. m. 1998 , in nuclear physics b ( proc . suppl . ) , 69/1 - 3 , 196 lomb , n. r. 1976 , ap & ss , 39 , 447 markert , t. h. , et al . 1979 , , 39 , 573 press , w. h. , teukolsky , s. a. , vetterling , w. t. , flannery , b. p. 1992
, numerical recipes , 2nd ed .
( cambridge : cambridge university press ) , 569 savage , b. d. & mathis , j. s. 1979 , , 17 , 73 scargle , j. d. 1982 , , 263 , 835 warwick , r. s. , et al .
1981 , , 197 , 865 warwick , r. s. , norton , a. j. , turner , m. j. l. , watson , m. g. , & willingale , r. 1988 , , 232 , 551 wen , l. , cui , w. , levine , a , m. & bradt , h. v. 1999 , apj , in press ( astro - ph/9906110 ) white , n. e. , nagase , f. & parmar , a. n. 1995 , in x - ray binaries , ed .
lewin , w. h. g.,van paradijs , j. , & van den heuvel , e. p. j. ( cambridge : cambridge university press ) , 1 wood , k. s. , et al .
1984 , , 56 , 507 lccc rxte / asm & 1996-present & 1.5 - 12 & 7.7 @xmath2 2.0 exosat / me & 1983 & 2 - 6 & @xmath91 einstein / ipc & 1980 & 0.5 - 3.5 & 0.27@xmath2 0.04 heao a2 & 1977 & 1.5 - 20 & 13.1 @xmath2 0.1 ariel 5/ssi & 1977 & 2 - 10 & 8 @xmath2 4 uhuru & 19711973 & 2 - 6 & 4.9 @xmath2 0.5 ccccc 0250 & 31.0 @xmath92 7.4 & 0.046 @xmath2 0.034 & 0.9 + 250500 & 25.9 @xmath92 5.4 & -0.011 @xmath2 0.032 & 0.8 + 500750 & 22.7 @xmath92 10.0 & -0.065 @xmath92 0.077 & 2.0 + 7501000 & 30.0 @xmath92 5.6 & 0.029@xmath2 0.028 & 0.9 + | x 1908 + 075 is an optically unidentified and highly absorbed x - ray source that appears in early surveys such as uhuru , oso-7 , ariel v , heao-1 , and the exosat galactic plane survey .
these surveys measured a source intensity in the range of 212 mcrab at 210 kev , and the position was localized to @xmath0 .
we use the rossi x - ray timing explorer ( _ rxte _ ) all sky monitor ( asm ) to confirm our expectation that a particular einstein ipc detection ( 1e 1908.4 + 0730 ) provides the correct position for x 1908 + 075 .
the analysis of the coded mask shadows from the asm for the position of 1e 1908.4 + 0730 yields a persistent intensity @xmath1 mcrab ( 1.512 kev ) over a 3 year interval beginning in 1996 february .
furthermore , we detect a period of 4.400 @xmath2 0.001 days with a false alarm probability @xmath3 .
the folded light curve is roughly sinusoidal , with an amplitude that is 22% of the mean flux .
the x - ray period may be attributed to the scattering and absorption of x - rays through a stellar wind combined with the orbital motion in a binary system .
we suggest that x 1908 + 075 is an x - ray binary with a high mass companion star . |
it is not an exaggeration to say that @xmath0-bosons along with their charged partners , @xmath13-bosons , have been at the focus of attention in high - energy physics over more than the past two decades .
the properties of these particles , which are mediators of weak interaction between leptons and quarks , were studied in detail in experiments at the lep and slc electron - positron colliders . in a series of experiments performed between 1989 and 1995 at cern @xcite ,
enormous statistics of observations of @xmath0-boson production and decays were collected . in all , approximately @xmath14 events associated with leptonic modes of @xmath0-boson decay and more than @xmath15 events of transformation of these unstable particles into hadrons were detected .
the results of these investigations made it possible to determine , among other things , the mass of the @xmath0-boson , @xmath16 , and its total decay width , @xmath17 , to an extremely high degree of precision @xcite .
many other parameters of the process @xmath18 were also analyzed both in the vicinity of the @xmath19-resonance and far from it @xcite .
the precision of experimental data reached in recent years has given a motivation to theoretical physicists for performing formidable work on calculating radiative corrections to the processes involving @xmath0-boson production and decay ( see , for example , @xcite and references therein ) . not only does @xmath0-boson physics provide a precision test for standard model predictions concerning particle interactions , but it is also a unique tool for seeking manifestations of new physics ( beyond the standard model ) , which will replace sooner or later the generally accepted @xmath20 scheme . despite the impressive successes of the standard model in adequately describing a formidable set of experimental data , there is no doubt among the physics community that the modern model of fundamental interactions can not be conclusive because of a number of fundamental theoretical drawbacks .
these include , first , an enormous number ( more than 19 ! ) of independent input parameters coupling constants , the masses of quark and leptons , the parameters of the cabibbo - kobayashi - maskawa mixing matrix , etc ; second , the instability of the mass of the as - yet - undiscovered higgs boson with respect to quadratically divergent radiative corrections and the allied hierarchy problem ; and , third , the isolated character of gravitational interactions , which do not fit in the existing quantum model .
this is not the whole story , however .
paradoxically as it is , the main drawback of the standard model at the present time is that it is in perfect agreement with all experimental data accumulated thus far , providing no hint as to the nature of new physics that will have to replace it .
this kind of `` recess '' in the experimental development of physics theory caused a flash of creative activity in theoretical physics in attempts at guessing vague outlines of a new theory of everything . by no means
do grand unification theories , supersymmetry , supergravity , superstrings , supermembranes , and @xmath21-theory exhaust the list of ideas that require an experimental verification .
the questions that a future theory should answer include that of the origin of particle masses , that of why our space - time is four - dimensional , that of why several generations of leptons exist , that of how many flavors quarks have , and that of how many neutrino flavors exist .
even today , physicists try to find answers to many of these questions by analyzing available experimental data . for example , the total invisible width of @xmath0-bosons , @xmath22 , with respect to the decays of these particles through channels that are rather difficult to detect plays an important role in @xmath0-boson physics . in the experiments performed at the lep accelerator ,
the invisible width @xmath23 was determined indirectly as the difference of the total decay width of the @xmath0-boson and its total width with respect to all observed decays . in the standard model , the invisible decay width of the @xmath0-boson
is interpreted as the partial width with respect to decays to neutrinos of various flavors : @xmath24 this quantity can be considered as some kind of a counter of neutrino generations , and its precise measurement is of particular interest for new physics ( see , for example , @xcite ) .
it is intriguing that the present - day experimental value of the number of neutrinos from the invisible decay width of the @xmath0-boson , @xmath25 @xcite , differs somewhat from the result expected in the standard model , @xmath26 .
exotic channels of @xmath0-boson decay frequently provide unique tests for some fundamental principles of modern physics .
for example , analysis of the forbidden ( in a vacuum ) @xmath0-boson decays to photons ( @xmath27 ) and gluons ( @xmath28 ) makes it possible to establish a quantitative measure of the principle of quantum indistinguishability of integer - spin particles bose symmetry for photons and gluons @xcite . according to the landau - yang theorem @xcite
, a massive vector particle can not decay to massless vector states . in an external electromagnetic field , however , the reactions @xmath27 and @xmath28 become possible @xcite . the problem of searches for new possible manifestations of supersymmetry is also tightly related to @xmath0-boson physics .
there is still a hope for finding relatively light superparticles produced in rare channels of @xmath0-boson decay . as an example
, one can consider the reaction @xmath29 , in which a pair of light gluinos , @xmath30 , having a mass of about @xmath31 gev appear .
this process , which is extensively discussed in the literature ( see @xcite and references therein ) , is of special interest in view of prospects for observing new - physics manifestations at next - generation @xmath32-colliders .
projects of experiments with @xmath0-bosons at the large hadron collider(lhc ) also caused interest in studying the properties of these particles in the presence of a quark - gluon plasma , which is quite an unusual environment of @xcite .
in view of the aforesaid , it is of particular interest to discuss some other new phenomena in @xmath0-boson physics that manifest themselves under unusual conditions . in this study , we investigate changes in partial decay widths of the @xmath0-boson in the presence of strong electromagnetic fields .
interest in such investigations is explained both by astrophysical applications and by applications in the physics of relativistic particles channeling through single crystals ( so - called channeling of particles @xcite ) .
it is well known that the strength of electric fields generated by the axes and planes of single crystals may reach formidable values of ( @xmath33v / m ) @xcite over macroscopic distances . at the same time
, it is well known that astrophysics studies exotic objects ( neutron stars , white dwarfs ) that , at the latest stages of their evolution , can undergo a strong compression , and this leads to a significant increase in the magnetic field strength inside them ( @xmath34 g @xcite ) . at such enormous values of the external field strength ,
the physics of quantum processes changes significantly .
external fields frequently remove the forbiddance of specific reactions whose occurrence is impossible in a vacuum .
the first investigations into the physics of relativistic particles in the presence of a strong external field were performed within quantum electrodynamics ( see , for example , @xcite ) .
later , the technique of calculations that was developed there was also used in non - abelian gauge theories @xcite .
respective calculations rely on the method of exact solutions to the wave equations for charged particles , which makes it possible to take into account the interaction with the electromagnetic field beyond standard perturbation theory . in this case , the wave functions for all charged particles and their propagators are modified in quite a unwieldy manner , depending on the configuration of an external field @xcite .
although the resulting expressions are cumbersome and although the calculations are quite involved , the ultimate results , which are valid at arbitrarily high values of the external - electromagnetic - field strength , carry information that can not be obtained by perturbative methods , and this is an obvious advantage of the method . from this point of view , investigation of quantum processes in superstrong fields provides a unique possibility of analyzing the self - consistency of the physical theory globally in the case where an expansion in a small parameter is impossible .
the method of calculations that is used below is based on the crossed - field model , which was successfully applied in the previous studies of the present author to studying photino - pair production @xcite and to analyzing the decays of @xmath35-bosons in an external field @xcite .
in the leading order of perturbation theory in the coupling constants @xmath36 and @xmath37 , the matrix element of @xmath0-boson decay to a pair of charged leptons , @xmath38 , is given by @xmath39 the presence of an external electromagnetic field is taken into account via choosing specific wave functions @xmath40 and @xmath41 for the charged leptons @xmath38 .
this approach , which , in particle physics , is referred to as the furry representation , yields , for the probability of @xmath0-boson decay , an expression in which electromagnetic interactions are taken into account exactly , beyond standard perturbation theory in the electromagnetic coupling constant @xmath42 .
the explicit form of the wave functions for charged particles in an external electromagnetic field depends on the field configuration . in the present study
, we restrict ourselves to the crossed - field configuration , in which case the strength tensor @xmath43 has the form @xmath44 it is assumed in this case that the external - electromagnetic - field potential @xmath45 is chosen in the gauge @xmath46 here , the unit constant @xmath47-vector @xmath48 determines the spatial field configuration and satisfies the conditions @xmath49 a crossed field is a particular case of the field of a plane electromagnetic wave having an isotropic wave @xmath47-vector @xmath50 .
the invariant parameter @xmath51 characterizes the crossed - field strength and has the dimensions of energy .
recall that , in an arbitrary reference frame , a crossed field is a superposition of constant electric and magnetic fields whose strength vectors are orthogonal and are equal in absolute value : @xmath52 .
moreover , the two relativistic invariants of the electromagnetic field are zero : @xmath53 in the crossed - field model , the choice of wave @xmath47-vector @xmath54 is quite arbitrarily .
it is only necessary that this vector be isotropic , have dimensions of energy , and satisfy the conditions @xmath55 in particular , the wave @xmath47-vector @xmath56 can be expressed in terms of the electromagnetic - field - strength tensor and the constant electromagnetic - potential @xmath47-vector @xmath57 as @xmath58 in dealing with the problem of @xmath0-boson decay , it is advisable to express the wave vector @xmath56 in terms of the @xmath0-boson 4-momentum @xmath59 and the electromagnetic - field - strength tensor @xmath60 ( see next section ) as @xmath61 in an arbitrary constant uniform electromagnetic field , the probability of the decay @xmath62 is a function of three invariant dimensionless parameters , @xmath63 .
these parameters are determined by the strength tensor of an external macroscopic field as @xmath64 in the crossed - field model , the parameter @xmath6 has the simplest form @xmath65 the square of the wave vector ( [ wavevector ] ) having the form @xmath66 in a crossed field , the two parameters ( [ invarianta ] ) and ( [ invariantb ] ) of the external - field strength vanish by virtue of the equality in ( [ crossed ] ) , @xmath67 ; therefore , the right - hand side of eq.([k2 ] ) also vanishes . for a constant uniform electromagnetic field of general form
, the condition that the wave vector in ( [ wavevector ] ) is isotropic is not satisfied : @xmath68 . even in this case , however , we can use the semiclassical crossed - field model , which ensures a satisfactory description of the decay probability in an electromagnetic field of general form , provided that @xmath69 and @xmath70 .
the closer to zero the right - hand side of eq.([k2 ] ) and the higher the precision to which the condition of isotropicity is satisfied , the more precise the results that are obtained in this approximation .
the above constraints are obviously satisfied in the region of relatively weak electromagnetic fields for @xmath71 and @xmath72 .
thus , the decay probability @xmath73 in a crossed field will correspond to the first leading term @xmath74 in the semiclassical expansion of the total decay probability @xmath75 in a power series in the vectors of the external - field strengths at small values of the parameters @xmath76 and @xmath77 ; that is , @xmath78 this relation between the probabilities of quantum processes in a crossed field and a constant electromagnetic field of general form was obtained for the first time within quantum electrodynamics @xcite .
the wave functions for charged fermions in the field of a plane electromagnetic wave were obtained as far back as 1937 by d.m.volkov ( see , for example , @xcite ) . in the case of a crossed field , these exact solutions to the dirac equations for leptons @xmath38 in an external field have the form @xmath79\times \nonumber\\ \times\biggl\{1-\frac{e(kx)}{4(pk)}(f_{\mu\lambda}\gamma^\mu \gamma^\lambda ) \biggr\ } \frac{u(p ) } { \sqrt{2p_0 v}};\end{aligned}\ ] ] @xmath80\times \nonumber\\ \times\biggl\{1+\frac{e(kx)}{4(p'k)}(f_{\mu\lambda}\gamma^\mu \gamma^\lambda ) \biggr\ } \frac{u^c(p ' ) } { \sqrt{2p'_0 v}};\end{aligned}\ ] ] the spin part of these wave functions normalized to a three - dimensional spatial volume @xmath81 is determined by the constant dirac bispinors @xmath82 and @xmath83 .
the antilepton wave function @xmath84 differs from expression ( [ lepton - ell- ] ) only by the reversal of the sign of the electric charge @xmath85 : @xmath86 . in the expression for the matrix element of the decay in ( [ sz - matrix ] )
, the antilepton wave function appears in the charge - conjugate form ( [ lepton - ell+ ] ) , which can be derived on the basis of the dirac - conjugate bispinor @xmath87 with the aid of the charge - conjugation matrix @xmath88 ; that is , @xmath89 as for the @xmath0-boson , its wave function does not change in an electromagnetic field because it is electrically neutral : @xmath90 the spin states of the @xmath0-boson are characterized by the complex polarization @xmath47-vector @xmath91 satisfying standard conditions for massive vector fields ; that is , @xmath92 we substitute the particle wave functions ( [ lepton - ell- ] ) , ( [ lepton - ell+ ] ) , and ( [ z - function ] ) into expression ( [ sz - matrix ] ) for the @xmath93-matrix element and perform integration of @xmath94 over the phase space of final leptons . upon summation over the spin states of the lepton - antilepton pair and averaging over polarizations of the @xmath0-boson ,
we arrive at @xmath95 \phi_1(z)-\nonumber\\ -\frac{2\varkappa^{2/3}}{\left[u(1-u)\right]^{1/3 } } \biggl[1 - 2u+2u^2+(1+\lambda ) \frac{m_\ell^2}{m_z^2}\biggr ] \phi'(z ) \biggr\}.\end{aligned}\ ] ] the probability of @xmath0-boson decay is expressed in terms of the airy functions @xmath96 and @xmath97 ( for necessary details concerning these special mathematical functions , see the appendix ) depending on the argument @xmath98^{2/3}}.\ ] ] the couplings of the @xmath0-boson to the charged leptons of all three generations , @xmath99 and @xmath100 , can be expressed in terms of the weinberg angle @xmath101 and the electroweak coupling constant @xmath102 as @xmath103 the dimensionless parameter @xmath104 in expression ( [ z - decay ] ) is given by @xmath105 at @xmath106 , it is very close to unity ( @xmath107 ) .
the fact that the law of energy - momentum conservation for particles involved in the decay @xmath108 does not have a conventional form , @xmath109 is a distinctive feature of processes occurring in an external electromagnetic field .
along with the @xmath0-boson and final - lepton energy - momentum @xmath47-vectors @xmath110 , @xmath111 , and @xmath112 , expression ( [ law ] ) also involves the wave vector ( [ wavevector ] ) determining the energy contribution of the external field .
it should be noted that , in this case , we are dealing with asymptotic momenta that charged particles possess at rather large distances , where the effect of the external field is negligible , rather than with their dynamical @xmath47-momenta in a crossed field .
this is the meaning that the parameters @xmath113 and @xmath114 appearing in the volkov solutions ( [ lepton - ell- ] ) and ( [ lepton - ell+ ] ) to the dirac equation in a crossed field have .
it is indeed extremely difficult to observe the kinematics of the decay @xmath108 in the field itself because the charged - lepton @xmath47-momenta are not integrals of the motion .
as soon as the leptons escape from the region of the external field , their energies and momenta do not change any longer , so that one can measure them experimentally . in this case , the @xmath47-momenta of all particles satisfy the ordinary kinematical conditions @xmath115 and this makes it possible to represent the isotropic vector @xmath56 in the form @xmath116 where the parameter @xmath117 characterizes the energy imbalance of the reaction @xmath108 in the external field , @xmath118 expression ( [ law ] ) arises in a natural way as the argument of the four - dimensional dirac delta function in calculating the @xmath93-matrix element ( [ sz - matrix ] ) . in performing integration over the phase space of final leptons , there also occurs summation of the external - field contributions characterized by various values of the parameter @xmath117 .
this is the circumstance that distinguishes the kinematics of @xmath0-boson decay in an external field from a similar process involving a real photon , @xmath119 , where the @xmath47-momentum @xmath54 is fixed and is independent of the final - lepton momenta .
the physical meaning of the above formulas can most easily be understood in the @xmath0-boson rest frame , where the law of energy - momentum conservation in ( [ law ] ) can be expressed in terms of the following @xmath47-vectors : @xmath120 we chose the axes of the three - dimensional cartesian coordinate system in such a way that the vectors of the electric- and magnetic - field strength , @xmath121 and @xmath122 , are directed along the @xmath123 and @xmath124 axes , respectively .
the direction of the @xmath125-momentum @xmath126 formed by the spatial components of the wave @xmath47-vector @xmath56 ( [ k3 ] ) then coincides with the positive direction of the @xmath127 axis , which is parallel to the vector product @xmath128}$ ] . in the cartesian coordinate system chosen above
, the orientation of the lepton ( @xmath129 ) and antilepton ( @xmath130 ) @xmath125-momenta @xmath131 and @xmath132 with respect to the @xmath127 axis can be described by two azimuthal angles @xmath133 and @xmath134 ( @xmath135 ) .
figure [ vectors ] shows schematically the three - dimensional configuration formed by all of the aforementioned vectors .
= 12.cm if the decay @xmath108 occurs in the vacuum and if the external electromagnetic field is inoperative , the lepton and antilepton have momenta that are equal in absolute value but are oppositely directed and which can be expressed in terms of their masses and the @xmath0-boson mass . the total energy taken away by the lepton - antilepton pair is shared between the particles in equal parts and
its total amount is equal to the rest energy of the @xmath0-boson , @xmath136 in a crossed field , each of the equalities in ( [ f=0 ] ) is in general violated , and the energy @xmath137 that is taken away in the decay by the lepton is not equal to the antilepton energy @xmath138 .
only fulfillment of the following conservation laws is guaranteed : @xmath139 the vectors @xmath131 , @xmath132 , and @xmath126 lie in the same plane , but , in this case , it is not mandatory that the lepton and the antilepton fly apart along the same straight line , @xmath140 .
the deficit in the ordinary law of energy - momentum conservation is compensated in an external field by the isotropic wave vector @xmath56 ( [ wavevector ] ) .
formula ( [ scm1 ] ) represents the three - dimensional form of the conservation law of the covariant projections of the @xmath47-momenta of all particles on to the direction of the wave vector @xmath141 , while relation ( [ scm2 ] ) is the conservation law for the ordinary three - dimensional projections of the particle momenta on to the axis that is orthogonal to the vector @xmath126 and which lies in the plane spanned by the @xmath125-vectors @xmath131 , @xmath132 , and @xmath126 .
we can now interpret physically the variable @xmath142 appearing in the integrand on the right - hand side of ( [ z - decay ] ) and in the argument of the airy functions in ( [ z ] ) .
this variable is associated with the angles @xmath133 and @xmath134 of divergence of the lepton - antilepton pair with respect to the @xmath127 axis in the rest frame of the initial @xmath0-boson : @xmath143 if the decay @xmath108 is unaffected by an external field , then , by virtue of relations ( [ f=0 ] ) , the variable @xmath142 can be represented in the form @xmath144={1\over
2}\left[1+\left(1-{4 m_\ell^2\over m_z^2}\right)^{1/2}\cos\vartheta_2\right].\ ] ] the interval of its values admissible from the point of view of decay kinematics in a vacuum is determined by the inequality @xmath145\ ] ] the limiting transition @xmath146 in ( [ z - decay ] ) reproduces exactly well - known results for the ordinary probability of the decay @xmath108 without any field .
the kinematically allowed region of values of the variable @xmath142 ( [ u - limits ] ) corresponds to negative values of the argument of the airy functions in ( [ z ] ) ; therefore , the limiting transition to zero values of the external - electromagnetic - field strength is implemented by means of the formal substitution of the heaviside step @xmath147-function for the airy function @xmath97 : @xmath148 in the limit @xmath146 , the differential probability of @xmath0-boson decay with respect to the variable @xmath142 does not depend on the direction of divergence of leptons and has the form @xmath149\ ] ] as a matter of fact , integration with respect to the invariant variable @xmath142 now reduces to multiplying expression ( [ dz0width - du ] ) by the phase space of this variable , @xmath150 . as a result , we obtain the well - known formula for the leptonic - decay width of the @xmath0-boson in a vacuum , @xmath151 where we have introduced the notation @xmath152 in a crossed electromagnetic field , the kinematics of @xmath0-boson decay changes , first of all , owing to the increase in the phase space of possible states of final leptons @xmath153 . in the three - dimensional space of the momenta @xmath154 of one of the leptons
for example , @xmath129 the kinematically allowed region of the decay @xmath108 in a vacuum [ see ( [ f=0 ] ) ] can be represented as a three - dimensional sphere , each point of this sphere corresponding to one of the final states of this lepton . in an external electromagnetic field ,
all possible final states of the lepton @xmath129 already fill some volume in momentum space . at a fixed value of the azimuthal angle @xmath133
, the momentum of the lepton @xmath129 can take any values from the interval @xmath155 the energy @xmath137 of this particle will be bounded by the inequality @xmath156 the variables @xmath137 and @xmath133 determine unambiguously decay kinematics in a crossed field .
the energy @xmath138 and the azimuthal angle @xmath134 of the antilepton @xmath130 can be expressed in terms of these variables with the aid of the kinematical relations ( [ scm1 ] ) and ( [ scm2 ] ) and the standard relativistic relations @xmath157 after some simple algebra , we obtain @xmath158 an analysis of the resulting formulas shows that , as a rule , the leptons @xmath153 diverge at arbitrary angles @xmath133 and @xmath134 with respect to the @xmath127 axis .
this is true for almost all of the azimuthal angles , with the exception of the case where the lepton and antilepton move in the direction parallel to the @xmath127 axis . in the case of @xmath159 and @xmath160 , the lepton and antilepton energies
@xmath137 and @xmath138 , respectively , can be expressed in terms of the emission angles @xmath133 and @xmath134 as @xmath161 = 12.cm these relations were obtained in the relativistic approximation , where @xmath162 , @xmath163 and where the parameter @xmath164 ( [ ell - const ] ) can be disregarded .
we can now draw qualitative conclusions on a typical behavior of the energies of the lepton - antilepton pair almost over the entire azimuthal - angle range @xmath165 .
it was indicated above that , at a fixed value of the emission angle ( @xmath166 ) of the lepton @xmath129 , its energy can vary from the rest energy to the maximum value @xmath167 ( [ emax ] ) .
concurrently , the emission angle @xmath134 of the antilepton @xmath130 can take an arbitrary value between zero and @xmath168 . as for its energy @xmath138
, one can see that , as the energy @xmath137 increases , @xmath138 first decreases from the value @xmath169 to its local minimum @xmath170 whereupon it increases monotonically to infinity as the angle @xmath134 tends to zero , while the lepton energy @xmath137 tends to its maximum value @xmath167 .
the above relationship between the energies of the lepton - antilepton pair and the directions of lepton and antilepton emission with respect to the @xmath127 axis is illustrated by the graphs in figs .
[ epsilon-1 ] and [ epsilon-2 ] .
it should be noted that the minimum value of the antilepton energy @xmath138 ( [ e2min ] ) is reached at the emission angle of @xmath171 , in which case the energy of the other particle , @xmath129,is @xmath172 = 12.cm if the lepton escapes in the direction parallel to the @xmath127 axis , expressions ( [ e1 ] ) and ( [ e2 ] ) are inapplicable , so that one must use the exact kinematical relations ( [ scm1 ] ) and ( [ scm2 ] ) or their consequences ( [ e2exact ] ) and ( [ angle2exact ] ) . in this case ,
two situations are possible that in which the leptons fly apart in opposite directions and that in which the leptons move in the same direction . in the first case ( at @xmath173 and @xmath174 or at @xmath175 and @xmath176 ) , the energy of the lepton moving along the positive direction of the @xmath127 axis can take any value from the rest energy @xmath177 to infinity .
the energy of the second lepton , which flies in the opposite direction , is virtually constant it ranges between the initial value @xmath178 ( [ e0 ] ) and its maximum value @xmath179 which is very close to @xmath178 . the case where both particles @xmath130 and @xmath129 produced in @xmath0-boson decay move in the same direction opposite to the positive direction of the @xmath127 axis ( @xmath180 ) is also possible . in this case , their energies @xmath137 and @xmath138 can vary from the rest energy to the limiting value @xmath178 ( [ e0 ] ) . in addition , the following relation can be obtained for this case from ( [ scm1 ] ) in the relativistic approximation ( for @xmath181 ) : @xmath182 thus , one can see that , in relation to the kinematics of the process @xmath108 in a vacuum [ see([f=0 ] ) ] , the physics of @xmath0-boson decay in an external electromagnetic field is richer in the number of possible final lepton states . from the mathematical point of view , this manifests itself in the fact that all values of the invariant variable @xmath142 ( [ u ] ) from the domain @xmath183 are now admissible .
moreover , it is noteworthy that , in a crossed field , the physical meaning of this variable changes because it now characterizes two independent kinematical quantities simultaneously - for example , @xmath137 and @xmath133 . disregarding the cases in which leptons fly apart in the direction parallel to the @xmath127 axis , we can express the variable @xmath142 in terms of the azimuthal angles @xmath133 and @xmath134 alone by using formulas ( [ e1 ] ) and ( [ e2 ] ) .
the result is @xmath184 = 12.cm an analysis of the integrand on the right - hand side of ( [ z - decay ] ) shows that , in an external field , not all of the directions of divergence of the lepton and antilepton are equiprobable .
the graphs in fig .
[ dzdu ] that represent the dependence of the differential decay width of the @xmath0-boson on the invariant variable @xmath142 corroborate that , even in a relatively weak external field , there are noticeable deviations from a uniformly distributed probability of the leptonic decay mode in a vacuum [ see ( [ dz0width - du ] ) ] . in the case where leptons produced in the process
@xmath108 fly apart at specific angles @xmath185 in diametrically opposite directions @xmath186 , the invariant variable @xmath142 takes values of @xmath187 . in this region , which is kinematically allowed for @xmath0-boson decay in a vacuum , the deviations of the differential decay width @xmath188 from the corresponding static value of @xmath189 mev are in fact unobservable .
if one of the leptons flies away at a specific angle @xmath190 with respect to the @xmath127 axis and at an energy close to the maximum value @xmath191 in ( [ emax ] ) and if the second lepton moves along the direction nearly parallel to this axis ( in the positive direction ) , the variable @xmath142 takes values in the region around @xmath192 or @xmath193 . in this case
, one can observe oscillations of the differential decay width of the @xmath0-boson that are characterized by a growing amplitude , which increases and decreases the static vacuum value in ( [ dz0width - du ] ) by more than 15% .
similar oscillations of probabilities of quantum processes in an external electromagnetic field are characteristic of all particle transformations allowed in a vacuum and were described in detail in the literature @xcite .
the integral representation of the partial width of the @xmath0-boson with respect to its decay to a pair of charged leptons in an external electromagnetic field can be derived from expression ( [ z - decay ] ) for the probability by going over to the rest frame ( @xmath194 ) and by making the change of the integration variable @xmath195 .
the result is @xmath196 \phi_1(z)-\nonumber\\ & -&2\left(\frac{\varkappa^2}{x}\right)^{1/3 } \biggl [ 1 - 2x+\delta_\ell^2(1+\lambda)\biggr ] \phi'(z ) \biggr\},\end{aligned}\ ] ] where the argument of the airy functions is now determined by the expression @xmath197 let us now consider the asymptotic estimates obtained from this formula at various values of the external - field - strength parameter @xmath6 ( [ kappa ] ) . in weak electromagnetic fields such that @xmath198 , the partial decay width of the @xmath0-boson can be written as the sum of two terms , @xmath199 the first term in this expression is the @xmath0-boson width with respect to the leptonic decay @xmath200 in a vacuum [ see expression ( [ zll - freewidth ] ) above ] . the experimental value of this quantity is known to a precision of 0.1% : @xmath201 @xcite .
the second term characterizes the effect of the external electromagnetic field . in the leading order of the expansion in the small parameters @xmath6 and @xmath164
, this term can be represented in the form @xmath202\end{aligned}\ ] ] similar calculations of the probability of @xmath0-boson decay in a magnetic field were performed in @xcite , where the contribution of the virtual electron - positron loop to the amplitude for elastic @xmath0-boson scattering was considered , its imaginary part being directly related to the @xmath108 decay width being studied .
it is worth noting , however , that the results of the present calculations [ expressions ( [ z - width ] ) and ( [ correction ] ) ] differ somewhat from their counter parts in @xcite .
for the parameter of the external - electromagnetic - field strength , we will now consider the region specified by the condition @xmath203 .
we use the fact that the masses of the leptons of all three generations are much smaller than the @xmath0-boson mass and go over to the limit @xmath181 in formulas ( [ z - width ] ) and ( [ zx ] ) .
we can then calculate analytically the integral of the airy function on the right - hand side of ( [ z - width ] ) at arbitrary values of the external - field - strength parameter @xmath6 .
the final expression for the leptonic - decay width of the @xmath0-boson is written in terms of the bessel functions @xmath204 carrying the noninteger indices @xmath205 and @xmath206 and depending on the argument @xmath207 .
we now consider a universal function @xmath208 that describes the degree of influence of the external field on the leptonic mode of @xmath0-boson decay : @xmath209 = 15.cm in the massless - lepton approximation ( @xmath210 ) , this function , which is a relative width of the @xmath0-boson with respect to the decay @xmath200 in an electromagnetic field , can be represented in the form @xmath211,\ ] ] where the following notation has been introduced : @xmath212,\ ] ] @xmath213,\ ] ] @xmath214.\end{aligned}\ ] ] = 15.cm the graphs representing the dependence of the relative leptonic - decay width ( [ ratio ] ) normalized to unity at zero external field on the invariant field - strength parameter @xmath6 ( [ invariantk ] ) are on display in figs .
[ zr1 ] , [ zr2 ] , [ zr3 ] . in relatively weak fields ( such that @xmath215 ) , there arise oscillations of the probability of the decay @xmath108 , their amplitude increasing quadratically with the parameter @xmath6 . in this region
, one can use the asymptotic expansion for the bessel functions and approximate formula ( [ ratio - k ] ) by the expression = 15.cm @xmath216 it is obvious from figs . [ zr1 ] and
[ zr2 ] that the maximum deviation of the decay width oscillating in an external field from its vacuum value in ( [ zll - freewidth ] ) does not exceed 2% .
a further increase in the parameter of the external - electromagnetic - field strength to values in the region @xmath217 leads to a complete disappearance of oscillations and a monotonic decrease in the relative width @xmath208 down to the absolute minimum @xmath218 in the vicinity of the point @xmath219 ( see fig . [ zr3 ] ) . in the region of strong fields ( such that @xmath220 )
, the relative width of the @xmath0-boson with respect to its leptonic decay mode grows monotonically with the field strength . in the region
@xmath221 , this growth can be approximated by the asymptotic expression . @xmath222
the results obtained in the preceding section admit a straight forward generalization to the case of the decay @xmath1 , where , by @xmath223 , we denote an arbitrarily charged fermion , @xmath224 standing for the absolute value of its electric charge . as is well known , the charge is @xmath225 for the up quarks ( @xmath226 ) and @xmath227 for the down quarks ( @xmath228 ) .
it is necessary to consider that , for the @xmath229 vertex , the coupling constants @xmath36 and @xmath37 ( [ sz - matrix ] ) also depend on the absolute value of the electric charge ( @xmath224 ) . in the glashow - weinberg - salam model of electroweak interactions , they are determined by the relations @xmath230 upon making the formal substitution @xmath231 or @xmath232 in ( [ z - decay ] ) , the differential @xmath233 decay width as a function of the invariant variables @xmath6 ( [ invariantk ] ) and @xmath142 ( [ u ] ) assumes the form @xmath234\phi_1(z_f)-\\ \nonumber -2\left[\frac{q_f^2\varkappa^2}{u(1-u)}\right]^{1/3 } \left[(1-c_f)(1 - 2u+2u^2)+\delta_f^2\right ] \phi'(z_f)\biggr\}.\end{aligned}\ ] ] in this expression , use is made of the same notation as in ( [ z ] ) and ( [ ell - const ] ) : @xmath235 @xmath236^{2/3}}.\ ] ] the limiting transition to an infinitely weak external field ( @xmath146 ) leads to reproducing the well - known results for the partial widths of the @xmath0-boson with respect to its decay to light quarks ; that is , @xmath237.\ ] ] @xmath238.\ ] ] these relations are obtained from ( [ zll - freewidth ] ) upon the formal substitution @xmath239 and @xmath240 and the multiplication by the common factor @xmath241 ( number of color quark states ) .
the inclusion of radiative corrections leads to various modifications of expressions ( [ zup ] ) and ( [ zdown ] ) , whose explicit form depends on the choice of input parameters in electroweak theory and renormalization scheme for divergent loop diagrams @xcite .
for example , sirlin , who was the author of one of the pioneering studies in these realms @xcite , proposed the parametrization @xmath242 which has been widely used since then . here , @xmath243
includes all radiative corrections . for the sake of simplicity ,
we restrict ourselves to the case of massless quarks ( @xmath244 ) and , without going in to details of the calculation of radiative corrections , try to analyze the effect of external electromagnetic fields on the hadronic decay modes of the @xmath0-boson . without allowance for the external - field effect , the width @xmath245 agrees with the experimentally measured value of @xmath246 mev @xcite to a high degree of precision .
we use this approximation in the region of relatively strong electromagnetic fields ( @xmath247 ) and estimate the width of the @xmath0-boson with respect to the hadronic decay mode as @xmath248 = 15.cm where the function @xmath208 is determined by expressions ( [ ratio ] ) ( [ r3 ] ) .
figure [ z - hadrons ] shows the results of numerical calculations performed for the total width of the @xmath0-boson with respect to hadronic decay modes in an external field on the basis of expression ( [ zhadrons ] ) with allowance for the production of only five quarks ( @xmath249 and @xmath77 ) . in this region ,
the contribution of the heavy @xmath11 quark is insignificant .
this contribution will be considered individually in the next section .
one can see from fig .
[ z - hadrons ] that , in just the same way as the width with respect to leptonic decay modes ( see fig .
[ zr3 ] ) , the width with respect to the hadronic modes of @xmath0-boson decay in an intense external field behaves nonmonotonically .
as the parameter @xmath6 increases , the width @xmath250 first decreases , reaching the absolute minimum of @xmath251 mev at the point @xmath252 , where upon it begins growing . in a weak electromagnetic field
( @xmath253 ) , the gradual decrease in the hadronic - decay width @xmath250 is accompanied by quite complicated oscillations .
in a vacuum , @xmath0-boson decay to @xmath11 quarks via the process @xmath254 is forbidden by the energy - conservation law because the mass of the heavy @xmath11 quark is rather high , @xmath255 gev @xcite .
the presence of an external electromagnetic field removes this forbiddance , opening new channels of @xmath0-boson decay . in principle
, the production of arbitrary charged fermions whose masses satisfy the relation @xmath256 becomes possible .
the missing energy for these reactions comes from the external field , and this resembles tunneling through a potential barrier in nonrelativistic quantum mechanics . at small values of the electromagnetic - field strength ( such that @xmath257 ) , the @xmath1 decay width is exponentially small : = 15.cm @xmath258.\ ] ] this formula was derived from expression ( [ zff - width ] ) upon integration with respect to the invariant variable @xmath142 ( [ u ] ) in the approximation @xmath259 ( [ z_f ] ) . substituting the @xmath11-quark charge @xmath225 and performing summation over three color states , we obtain an asymptotic estimate for the partial width of the @xmath0-boson with respect to its decay to @xmath11-quarks in a relatively weak electromagnetic field ; that is , @xmath260 } { 4\pi\sqrt{6}\left(4 m_t^2 - m_z^2\right)\sqrt{8 m_t^2 + m_z^2 } } \times \nonumber\\ \varkappa\ \exp\left[-\frac{\left(4 m_t^2 - m_z^2 \right)^{3/2 } } { 2\varkappa m_z^3}\right].\end{aligned}\ ] ] the results of numerical calculations show that the above approximate formula for @xmath261 leads to an error not greater than 4% for all values of the external - field - strength parameter from the region @xmath262 .
as the parameter @xmath6 grows , the deviation of the results obtained according to the asymptotic formula ( [ z - top ] ) from the precise value of the partial width of the @xmath0-boson with respect to the decay @xmath254 in an external field increases [ see fig.[z - topgraf ] , where the dotted curve corresponds to the calculation according to formula ( [ z - top ] ) , while the solid curve represents the results of a numerical integration of expression ( [ zff - width ] ) ] .
= 15.cm the @xmath11-quark contribution to the total decay width of the @xmath0-boson remains negligible ( less than 1% ) up to strength - parameter values of about @xmath263 .
figure [ z - total-10 ] displays the total decay width of the @xmath0-boson as a function of the parameter @xmath6 in the region of strong fields ( @xmath264 ) .
the solid curve corresponds to the total decay width of the @xmath0-boson with allowance for the @xmath11-quark contribution , while the dotted curve represents the same quantity calculated without taking into account the @xmath11-quark contribution .
one can clearly see a trend toward the growth of both the absolute value of the total decay width of the @xmath0-boson and the fraction of the @xmath11-quark contribution in this decay width . as the external - field strength increases ,
the process @xmath254 , which is forbidden in a vacuum , becomes dominant in superstrong electromagnetic fields . by way of example
, it can be indicated that , in the @xmath0-boson decay width at @xmath265 , the fraction associated with @xmath11-quarks is as large as 50% .
this circumstance clearly demonstrates how an external electromagnetic field can change drastically the standard physics of quantum processes in a vacuum and serve as some kind of a catalyst for a number of new and nontrivial phenomena .
the effect of a strong electromagnetic field on the probability of @xmath0-boson decay has been investigated .
the present calculations have been performed within the crossed - field model , which makes it possible to trace characteristic variations in the modes of @xmath0-boson decay to known leptons and quarks .
we have found that , in the region of relatively strong fields ( @xmath266 ) , all partial decay widths @xmath267 decrease in the same manner by about @xmath268 to @xmath269 .
this decrease can be described by a universal function @xmath208 [ see ( [ ratio])([r3 ] ) ] depending only on the energy - momentum of the @xmath0-boson and on the external - electromagnetic - field strength .
this circumstance becomes quite obvious if we consider that all known fermions observed in @xmath0-boson decays have masses at least an order of magnitude smaller than the @xmath0-boson mass .
therefore , it comes as no surprise that , in the region of strong electromagnetic fields , the massless - fermion approximation works well , also making it possible to calculate the dependence of the total decay width @xmath4 of the @xmath0-boson on the external - field - strength parameter @xmath6 ( [ invariantk ] ) .
the results of the present calculations have revealed that , in just the same way as in the case of the @xmath35-boson @xcite , the total decay width of the @xmath0-boson in a strong electromagnetic field depends nonmonotonically on the strength parameter @xmath6 , this dependence featuring the point of a local minimum , @xmath270 gev , at @xmath7 .
the graph in fig .
[ z - total ] , which represents the behavior of @xmath4 in the region @xmath271 , illustrates this observation .
= 15.cm for the sake of comparison , it can be indicated that the analogous @xmath35-boson - decay process @xmath272 in an external field is also characterized by a nonmonotonic dependence of the decay width on the parameter @xmath6 with the point of a local minimum at @xmath273 @xcite . upon rescaling this value to the @xmath0-boson energy scale
, it corresponds to @xmath274 .
thus , the probabilities for the decays of gauge bosons in an external field , @xmath275 and @xmath272 , take their minimum values at nearly the same field strength .
this agreement is not accidental since the two reactions in question proceed under similar kinematical conditions and since the masses of the @xmath13 and @xmath0 bosons are almost indistinguishable in order of magnitude .
it is also intriguing that , in the region around @xmath276 , the presence or the absence of the initial - gauge - boson electric charge is immaterial because the change in the decay width is due primarily to the increase in the phase space of final particles . in the region of superstrong fields ( @xmath277 ) ,
a sizable contribution to the total decay width of the @xmath0-boson comes from the @xmath11-quark - production process @xmath254 , which is forbidden in a vacuum by conservation laws . as the external - field strength increases , this reaction becomes dominant in relation to other modes of @xmath0-boson decay , this being due to a very large mass of the @xmath11-quark .
the external field serves as some kind of a catalyst for super heavy - particle - production processes , which can not occur under ordinary conditions . in view of the aforesaid , it would be of interest to discuss searches for new ( as - yet - undiscovered ) hypothetical particles ( supersymmetry and so on ) among new @xmath0-boson decay modes that arise in an external electromagnetic field .
although external - field strengths necessary for directly observing such reactions are not yet available in experiments ( see , for example , @xcite ) , these investigations are of course very interesting from the point of view of fundamental principles of physical theory .
in the present study , we have employed the special mathematical functions @xmath278 and @xmath97 generically referred to as airy functions .
the function @xmath279 is a particular solution to a second - order linear differential equation for specific initial conditions : @xmath280 it has the well - known integral representation @xmath281 two other functions @xmath96 and @xmath97 can be expressed in terms of this function as @xmath282 at small values of the argument @xmath127 ( @xmath283 ) , the airy function @xmath279 can be calculated on the basis of the following expansion in a numerical series : @xmath284 at large values of the argument ( @xmath285 ) , the airy function can be evaluated by using the asymptotic expressions @xmath286 aleph collab .
( r. barate , _ et al .
_ ) , eur . phys .
j. c * 14 * , 1 ( 2000 ) + delphi collab .
( p. abreu , _
_ ) , eur . phys .
j. c * 16 * , 371 ( 2000 ) + l3 collab .
( m. acciarri , _ et al .
_ ) , eur . phys .
j. c * 16 * , 1 ( 2000 ) + opal collab .
( g. abbiendi , _ et al .
_ ) , eur .
j. c * 19 * , 587 ( 2001 ) .
et al . _ , ( particle data group ) , j. phys .
g * 33 * , 1 ( 2006 ) .
aleph collab . ,
delphi collab .
, l3 collab . ,
opal collab . ,
sld collab .
, lep electroweak working group , sld electroweak , heavy flavour groups , phys.rept . * 427 * , 257 ( 2006 ) [ hep - ex/0509008 ] m.i .
vysotsky , v.a .
novikov , l.b .
okun , and a.n .
rozanov , usp .
nauk * 166 * , 539 ( 1996 ) [ phys .
* 39 * , 503 ( 1996 ) ] ; [ hep - ph/9606253 ] ; + v.a .
novikov , l.b .
okun , a.n .
rozanov , and m.i .
vysotsky , rep .
prog . phys . *
62 * , 1275 ( 1999 ) [ hep - ph/9906465 ] .
r.harlander , acta phys .
b * 29 * , 2961 ( 1998 ) [ hep - ph/9806524 ] .
d. bardin , m. grunewald , g. passarino , hep - ph/9902452 .
g. degrassi , p. gambino , nucl.phys . b * 567 * , 3 ( 2000 ) [ hep - ph/9905472 ] m. carena , a. gouvea , a. freitas , m. schmitt , phys.rev . d * 68 * 113007 ( 2003 ) [ hep - ph/0308053 ] .
ignatiev , g.c .
joshi , m. matsuda , mod .
lett . a * 11 * 871 ( 1996 ) [ hep - ph/9406398 ]
landau , dokl .
nauk sssr * 60 * , 207 ( 1948 ) ; + c.n .
yang , phys .
, * 77 * , 242 ( 1950 ) .
tinsley , phys.rev . d * 65 * , 013008 ( 2002 ) [ hep - ph/0106142 ]
. z. luo , phys.rev .
d * 67 * , 115007 ( 2003 ) [ hep - ph/0301051 ] .
kapusta , s.m.h .
wong , phys.rev .
d * 62 * , 037301 ( 2000 ) [ hep - ph/0002192 ] .
baier , v.m .
katkov , and v.m .
strakhovenko , usp .
nauk * 159 * , 455 ( 1989 ) [ sov .
usp . * 32 * , 972 ( 1989 ) ] .
moller , cern report 94 - 05 .
s. shapiro and s. teukolsky , _ black holes , white dwarfs and neutron stars : the physics of compact objects _ ( wiley , newyork , 1983 ; mir , moscow , 1985 ) .
a.a . sokolov and i.m .
ternov , _ relativistic electron _
( nauka , moscow , 1983 ) [ inrussian ] .
ritus and a.i .
nikishov , _ quantum electrodynamics of phenomena in intense fields _ , trudy fian sssr , vol.111 ( nauka , moscow , 1979 ) [ in russian ] .
ternov , v.r .
khalilov , and v.n .
rodionov , _ interaction of charged particles with strong electromagnetic fields _ ( izd - vo mgu , moscow,1982 ) [ in russian ] .
borisov , a.s .
vshivtsev , v.ch .
zhukovsky , and p.a .
eminov , usp .
nauk * 167 * , 241 ( 1997 ) [ phys .
usp . * 40 * , 229 ( 1997 ) ] .
kurilin , nuovo cimento , * 112 a * , 977 ( 1999 ) [ hep - ph/0210194 ] .
kurilin and a.i .
ternov , yad .
fiz . * 63 * , 1944 ( 2000 ) [ phys . at
. nucl . * 63 * , 1855 ( 2000 ) ] ; pisma zh .
* 63 * , 305 ( 1996 ) [ jetp lett . * 63 * , 311 ( 1996 ) ] .
kurilin , yad . fiz . *
67 * , 2116 ( 2004 ) [ phys . at
nucl . * 67*,2095 ( 2004 ) ] ; [ hep - ph/0709.0335 ] .
berestetskii , e.m .
lifshitz , and l.p .
pitaevskii , _ quantum electrodynamics _
( nauka , moscow , 1989 ; pergamon , oxford , 1982 ) .
borisov , v.ch .
zhukovsky , m.yu .
knizhnikov , and a.n .
prokopenya , izv .
30 * ( 5 ) , 44 ( 1987 ) .
a. sirlin , phys .
d * 22 * , 971 ( 1980 ) m. abramowitz and i.a .
stegun , _ handbook of mathematical functions _
( dover , newyork , 1964 ; nauka , moscow , 1979 ) [ in russian ] . | the probability of @xmath0-boson decay to a pair of charged fermions in a strong electromagnetic field , @xmath1 , is calculated . on the basis of a method that employs exact solutions to relativistic wave equations for charged particles , an analytic expression for the partial decay width @xmath2
is obtained at an arbitrary value of the parameter @xmath3 , which characterizes the external - field strength . the total @xmath0-boson decay width in an intense electromagnetic field , @xmath4 , is calculated by summing these results over all known generations of charged leptons and quarks .
it is found that , in the region of relatively weak fields ( @xmath5 ) , the field - induced corrections to the standard @xmath0-boson decay width in a vacuum do not exceed 2% . as @xmath6 increases , the total decay width @xmath4 develops oscillations against the background of its gradual decrease to the absolute - minimum point . at @xmath7 ,
the total @xmath0-boson decay width reaches the minimum value of @xmath8 gev , which is smaller than the @xmath0-boson decay width in a vacuum by more than 10% . in the region of superstrong fields ( @xmath9 )
, @xmath4 grows monotonically with increasing external - field strength . in the region @xmath10 , the @xmath11-quark - production process @xmath12 , which is forbidden in the absence of an external field , begins contributing significantly to the total decay width of the @xmath0- boson . |
high energy radiation can be detected directly by converting photons into electrical signals , or indirectly using scintilators and photodiodes .
direct conversion detectors are expected to have higher signal - to - noise ration , higher spatial and temporal resolution , even on a single photon counting level .
this could lead to spectrally resolved computed tomography and advanced medical x - ray imaging @xcite .
an efficient detector optimization is a challenging technological task .
the single crystal growth is one of them , high mobility - relaxation time product ( @xmath1 ) is another one and least but not the last one is different mobility and trapping cross - section of photo - generated holes .
the latter one contributes to significant degradation of temporal resolution .
a wide range of materials @xcite provide good sensitivity , but poor temporal performance is achieved .
the low mobility of holes , compared to electron mobility , have been solved by coplanar grid ( cpg ) detection @xcite in cdznte detectors , semiconductor analogue of frisch grids in gas ion chambers .
as the technology of majority of proposed materials needs yet to mature , the choice of gaas , crystalline cdte or cdznte is mandatory for counting and high energy resolution applications .
the problem with different electron and hole mobilities is overcome using the coplanar grid electrodes and it has been proven also to be efficient way to increase specral resolution @xcite .
the coplanar - grid as well as other electron - only detection techniques @xcite are effective in overcoming some of the material problems of cdznte and , consequently , have led to efficient gamma - ray detectors with good energy resolution while operating at room temperature @xcite .
the main limitations yet remain to be high quality crystalline material and polarization effects .
their reduction has been addressed both experimentally and theoretically .
atomistic simulations @xcite are a powerful method for exploring crystalline defects at a resolution unattainable by experimental techniques @xcite .
determination of electric field shape precisely due to relatively complicated electrodes structure and application of the bias and intergrid voltage simultaneously is experimentally also difficult task .
though exact solutions are often tedious , simplified models have been also proposed .
a small part of pixel detector has been considered @xcite and spectral response modeled .
polarization and carrier trapping was solved by simplified set of kinetic equations @xcite in cdznte detectors , the effect of transport parameters inhomogenities ( charge density , mobility , potential fluctuations ) and hot carriers were discussed @xcite and plasmons in solid state radiation detectors were studied @xcite . the laser induced transient current technique ( l tct ) provides information about charge collection dynamics and detector polarization @xcite .
it has been measured on pixel @xcite and coplanar grid detectors @xcite .
though it is important for further detector optimization , the theoretical treatment is usually given only in simplified experimental geometries and basic formulation of schockley - ramo theorem @xcite .
the coplanar grid detectors have been modeled also only in simplified geometry and spectral response has been calculated only @xcite .
sparse realistic detector geometries were considered for x - ray spectra simulations @xcite , however , no temporal response has been calculated .
we provide here an insight on carrier transport in a radiation coplanar grid detector based on cdznte , valid also for other semiconductors used in direct conversion detectors . we determine influence of bulk semiconductor parameters on a temporal and spectral resolution in real detector geometry . the requirement on efficient charge collection
is discussed with respect to effects of depletion width , diffusion and detector shielding .
tct waveforms compared to numerical model allow evaluation of the electric field profile inside coplanar grid detector with high spatial resolution unattainable experimentally .
the spectral resolution improvement is quantitatively determined in a coplanar grid detector and the differential current sensing @xcite is shown to overcome problems of low hole mobility , high hole trapping cross - section and random nature of x-/gamma ray photon absorption within a detector .
we also aim to show simplified analytically based considerations to understand detector operation , which is useful for qualitative optimization without need of elaborate numerical calculations .
the l tct response has been modeled by drift transport mechanism of photo - generated carriers . the above band gap photo - generation is assumed .
holes have been neglected for their small mobility @xmath2 @xmath3v@xmath4s@xmath4 and short life time @xcite .
for this reason we assume electrons only @xcite and collection times much shorter than electron life time . in l
tct , holes are swept to cathode in subnanosecond times , hence , their influence is negligible . the current response caused by moving charge between electrodes
have been calculated using shockley - ramo theorem .
the foundation of shockley - ramo theorem is based on known electrostatic potential in two cases .
the first one consists of real electrostatic potential distribution within a detector .
the second is a potential distribution of a weighting potential . here , the electrode of interest is biased to 1 v and all other electrodes are at 0 v. the corresponding electric field determines the charge response on a given electrode and the time derivative determines the current response @xcite .
the full real potential and corresponding electric field determines the trajectory of photo - generated carriers .
we assume only small enough concentration of photo - generated carriers that they do not spread by diffusion or electrostatic repulsion on the length scale of detector size before they are all swept by anode .
this is clearly well fulfilled since the diffusion time broadening is about 15 ns and carriers move about 1 @xmath5s from cathode to anode .
the limitations of this assumption are discussed comparing the model and experimental data .
tct current response is modeled by assuming 1.4 mm wide laser spot size on partially transparent cathode .
we trace 300 electron trajectories equally spaced on a cathode center .
the motion of each electron ( electron packet ) has been calculated by integrating their drift velocity .
the current response is then calculated for each of 300 trajectories and summed up .
the resulting summed current response has been compared to experimental data .
we also assume positively charged depletion width below cathode and negatively charged space region below anode , as commonly observed in cpg .
these space charge regions give rise to current decrease ( increase ) at the time of 0 ( 1 @xmath5s ) .
the experiments done by hard x - ray and @xmath6-ray transient current technique ( x tct ) has been modeled by solution of poisson equation and , in the second computation step , electron trajectories have been obtained by integrating their velocities ( as in l tct curves ) . however , here , the initial position of photo - generated carriers have been determined in a monte - carlo loop .
random position of photo - generated carriers is a model of large absorption length of x / gamma - ray radiation , hence the model holds for more than 100 s ev energetic photons . here
, we have not summed all trajectories with a random initial position .
instead , the current response has been calculated for each trajectory separately and the total collected charge has been calculated by integrating current response in time . the electron current response has been considered only since holes mobility is @xmath7 lower @xcite and holes have higher trapping cross - section .
we have traced 1500 trajectories and total collected charge from electron motion has been statistically analyzed .
the electrostatic potential has been calculated by solving poisson s equation @xmath8 by finite element method .
the detector geometry has been chosen to properly model a redlen inc
. state - of - the - art cdznte cpg detector .
the mesh triangulation has been used to properly describe size effects of rectangular metal anode grid .
the anode behaves as a set of parallel charged wires when charge is far from the anode compared to anode characteristic size ( anode strip width ) . and
, the anode behaves like infinitely large planar electrode if the charge is close enough to anode compared to characteristic anode size .
we consider the latter case as a measure of sufficiently dense triangulation .
this regime can be analytically treated as well , and pronounces itself as a constant electric field and linear - in - distance electrostatic potential .
the strong electric field at both electrodes causes velocity saturation .
the field maximal values are @xmath9 v / cm , giving drift speed up to @xmath10 cm / s assuming electron mobility is @xmath11@xmath12v@xmath4s@xmath4 .
the saturation velocity in cdznte is about @xmath13 cm / s @xcite therefore the velocity field dependence is expected to show non - linear behavior .
the velocity saturation @xcite is taken into account by @xmath14 .
the modeled l tct waveforms have been compared to experimental data measured on coplanar grid cdznte detector .
the detector , purchased from redlen inc . , has dimensions @xmath15 mm@xmath16 .
the details of our ltct setup @xcite and data proccessing are descibed elswhere @xcite .
the detector cathode was irradiated by optical pulses ( @xmath17 ns as fwhm , 200 ps rise and fall time ) using the laser diode ( 660 nm , 300 mw ) that is powered by an ultra - fast pulse generator .
pulse output energy was 0.4 nj and repetition rate was 100 hz . a collimation and imaging optics and a translation stage is used in the setup to adjust the laser diode beam onto the examined spot of @xmath17 mm@xmath18 on the cathode .
neutral density disc filter is placed in the beam line for optimal intensity attenuation of the laser pulse .
negative bias @xmath19 is applied to the cpg detector cathode ( up to 1700 v ) and supplementary intergrid voltage @xmath20 between the collecting and non collecting anode grid is introduced by an additional adjustable low noise voltage supply ( @xmath21 v ) . ,
4.0 and 16.0 @xmath22 @xmath23 are distinguished by red , orange and violet color .
solid ( dashed ) curves in ( a , c , e ) show numerical simulation for intergrid bias @xmath24 v ( 150 v ) .
solid ( dashed ) curves in ( b , d , f ) show numerical simulation for negative cathode bias @xmath25 v ( 1000 v ) .
the insets ( e , f ) depict detail of the current response around @xmath26s .
the space charge width below cathode is @xmath27 .
the simulation is shown for ( a , b , e , f ) collecting grid and ( c , d ) non - collecting grid . the i@xmath28 trend ( a , c , e ) is studied for negative v@xmath29=1700 v and the v@xmath29 trend ( b , d , f )
is studied for v@xmath30=150 v.,width=340 ]
the effect of doping is shown in fig .
[ fig1 ] for positive space charge @xmath31 , 4.0 and 16@xmath22 @xmath23 .
the current response in l
tct is modeled for collecting grid ( cg ) in fig .
[ fig1 ] ( a , b , e , f ) and for non - collecting grid ( ncg ) in fig . [ fig1 ] ( c , d ) .
the current response can be split into three characteristic intervals .
the first one is given by charge deceleration in a positive space charge below cathode .
the second regime is at intermediate times when the charge moves through the part of the detector without any fixed space charge .
the third regime is a time of charge collection at the anode .
it can be seen from numerical simulation that the peak current response at the collection time increases with decreasing doping .
the initial current peak is decreasing with decreasing doping and ideally disappears .
the initial current peak can give rise to a spurious signal contributing to a low energy spectral shoulder in @xmath6-ray spectra .
the anode intergrid bias is shown for two values 0 and 150 v in fig .
[ fig1 ] ( a ) and ( c ) .
we note that the intergrid bias does not influence the initial current response , as is shown in the inset fig .
[ fig1 ] ( e ) , in contrast to cathode bias , inset ( f ) .
this is caused by the local effects of intergrid bias on a small area around anode .
the cathode bias @xmath19 ( cathode to ncg bias ) influences mean field within a detector , thus changing current response for all times from the photo - carrier generation to their complete collection on cg .
the black , red and green curves correspond to depletion width 0.64 , 0.80 and 0.85 dw , respectively .
the numerical simulation is shown for ( a ) collecting and ( b ) non - collecting grid .
the inset in ( a ) depicts detail of current evolution at @xmath26s . the intergrid bias @xmath32 v and cathode bias negative @xmath33 v.,width=340 ] the cathode depletion width dependence of the l tct waveform is shown in fig . [ fig2 ] .
the cathode is a planar electrode and the constant current is expected for uncharged bulk semiconductor .
the space charge ( fixed or mobile ) causes deceleration of carriers hence decreasing current .
the equilibrium depletion width ( @xmath34 ) is given by @xmath35 , where @xmath36 is doping density ( or fixed space charge ) , @xmath37 is a voltage drop within a depletion width and @xmath38 is permitivity .
the width of the depletion region varies from this simple model when carrier trapping and de - trapping is assumed , or more than one trapping level is located in the vicinity of the fermi level .
it can be seen from fig .
[ fig2 ] that the peak current increases with decreasing depletion width .
this effect is due to the reduced total voltage drop in a depletion region and consequently higher voltage drop at the anode .
the latter causing higher electric field and higher current .
we note here , that this is in contrast to doping density . the current response of a ncg
is shown in fig .
( 2 ) ( b ) .
the current response is , similarly to cg , stronger for thinner depletion width .
@xmath23 ) . diffusion peak broadening @xmath39 0 , 10 , 20 and 30 ns has been assumed .
the effect of diffusion is simulated for ( a ) collecting and ( b ) non - collecting grid . the intergrid bias @xmath32 v and cathode bias @xmath33 v.,width=340 ] the effect of diffusion is shown in fig . [ fig3 ] .
diffusion takes important role for long enough transient times @xmath40 when @xmath41 t is comparable or larger than drift current temporal width . for this reason
we show only the current peak modification at the charge collection on anode .
the temporal broadening can be estimated from @xmath42 , where @xmath43 is electron diffusion constant related to electron mobility @xmath44 , temperature @xmath45 and mean drift velocity @xmath46 , which can be estimated from the transient time @xmath40 and detector height @xmath47 ; @xmath48 .
the current response for cg and ncg is shown for four diffusion broadening times ( 0 , 10 , 20 , 30 ns ) .
the current response is calculated for @xmath49 @xmath23 and space charge width 0.64 of the equilibrium depletion width .
we study here the effect of electrostatic shielding on inner potential distribution within a detector .
the detector shielding has been assumed to cover 2/3 of detector height .
this coverage maximizes the shielding effect yet reduces a risk of material and surrounding air dielectric breakdown caused by strong electric field between cathode ( top shielding edge ) and anode . the critical electric field
@xmath50 kv / mm in air provides possibility to apply @xmath51 kv between shielded cathode and anode ( shielding - anode gap 3 mm for 2/3 of detector height coverage ) .
we show three trial electron trajectories in fig .
[ fig9 ] for ( a ) unshielded and ( b ) shielded detector .
the electron trajectories are almost parallel for unshielded detector and they tend to get localized to one trajectory in the shielded case . the current response is shown in fig .
the initial current peak decreases to about 40% of the unshielded value and the collection peak current rises by @xmath52 .
both effects improve detector functionality .
the first peak reduction will lead to lower spurious signal and the stronger collection peak to better signal - to - noise ratio .
the measured l tct waveforms in a coplanar grid detector on a cg and ncg anode are shown in fig .
the data show good qualitative agreement , see fig .
[ fig8 ] , and in many aspects also quantitative comparison with simulated data can be made , especially concerning relative intensities of current peaks at the beginning and at the end of charge collection .
the initially decreasing current is caused by decelerating electrons in a positive space charge below the cathode .
the second collection current peak is stronger than the first one especially due to the strong electric field between cg and ncg .
this current originates also in a negative space charge below anode .
the current peak at the time of charge collection @xmath53 ns is broadened by electron diffusion .
the amount of broadening estimated from the numerical simulation is @xmath5410 ns .
v and three inter - grid biases @xmath24 v , 20 v , 100 v depicted by black , red and green curves , respectively.,width=264 ]
here we study importance of the measurement technique to detect photo - generated carriers .
an example of three current waveforms detected on cg , ncg and differentially measured signal between cg and ncg is shown in fig .
the advantage of differential sensing is lack of any signal before charge reaches coplanar grid anode .
there is no influence of the positive space charge at cathode , since such a current response is equal for cg and ncg .
we note that the differential sensing does not bring significant improvement for l tct , where the differential current waveform can be formed by data post - processing , as has been done in numerical simulation in fig .
[ fig5 ] to illustrate the method .
the advantage is more pronounced when @xmath6-ray photons are detected . here
, the absorption events occur randomly within a detector volume and they contribute to signal broadening . the @xmath6-ray photon detection is modeled in fig . [ fig6 ] .
voltage drop at the cathode ( positive space charge @xmath49
@xmath23 ) and @xmath55 voltage drop at the anode ( negative space charge @xmath56 @xmath23 ) .
the current response measured at the collecting grid ( solid blue curve ) , non - collecting grid ( dashed red curve ) and differential signal ( solid black curve ) .
the applied biases are @xmath32 v and negative @xmath25 v.,width=226 ] we have traced 1500 trajectories of randomly generated photo - excited electrons and the total electron - induced charge has been calculated .
an example of 50 random trajectories is shown in fig .
[ fig6 ] ( d ) .
we note , that if holes are considered the total collected charge is @xmath57 , however , we assume low hole mobility and high trapping cross - section for holes .
this leads to much weaker current response , or , the current response on very long time scales in comparison to electron induced tct waveforms .
for this reason , electron traced from below cathode gives larger induced charge than electrons traced from the anode vicinity .
the statistics of all monte - carlo simulated absorption events are shown in fig .
[ fig6 ] ( a , b , c ) for cg , ncg and for differential sensing , respectively .
the cg and ncg current waveforms have rectangular like shape caused by equal distribution of absorption events within a detector .
the mean collected charge on ncg is shifted towards zero because ideally no charge is collected by ncg .
we consider here negative space charge at the anode .
this space charge screens the electric field of neutral bulk detector ( no charge collected by ncg ) and it allows certain electron trajectories to be collected by ncg . -ray photon absorption events .
doping density , depletion widths and applied biases are the same as in fig .
the absorption probability is assumed to be equal throughout the detector volume .
currents induced by holes are neglected .
the histograms ( a , b , c ) depict statistics of collected charge for all 1500 absorption events .
the statistics is shown for the case of current measured on ( a ) collecting , ( b ) non - collecting grid and ( c ) differential signal between collecting and non - collecting grid .
equipotential lines ( blue to red contour lines ) , detector area ( solid black lines ) and 50 selected random absorption events are shown in ( d ) .
, width=340 ] in contrast to cg current response , the differential sensing , see fig .
[ fig6 ] ( c ) shows the same mean charge collection , however , the full width at half maximum is 10@xmath0 smaller .
this is caused by the signal insensitivity to the position of random - in - nature absorption events of @xmath6-ray photons .
we also note that there is no spurious signal at 0 ns , fig .
[ fig5 ] , caused by positive space charge hence reduced low energy spectral shoulder @xcite is expected beside the compton scattering contribution .
the effect of holes is also reduced regardless of assumptions made in our simulations .
holes are attracted towards cathode giving even weaker response on a cg anode .
general considerations to understand current waveforms follow from shochley - ramo theorem @xmath58 . decreasing / increasing current shows on decreasing / increasing electric field in a positive / negative space charge at cathode / anode . the space charge at anode forms weak current response at @xmath59 0v and it causes current shoulder for early time of charge collection .
the dominant signal at the charge collection time for @xmath60 0 is caused by accelerated electrons in a spatially modulated electric field by inter - grid bias @xmath20 .
more specifically , we have shown that the cathode bias increases the peak current response .
it also increases the initial signal caused by positive space charge at the cathode [ fig1 ] ( b ) .
this can lead to spurious signal leading to low energy spectral shoulder in x ray spectra . at the same time , since the charge is swept faster through detector bulk , there is lower probability for carrier trapping and detector polarization .
the lowest doping density is still the ultimate option here because it worsens the charge collection efficiency .
the effect of diffusion is twofold .
first , the higher is diffusion / mobility , the faster is a response time at a given detector size .
second , as absorption efficiency increases with detector size , diffusion lowers the signal .
hence , for fast response and strong current signal , thinner detector is acceptable .
we also point out that thinner detector volume also eliminates carrier trapping .
the optimal detector thickness and coplanar grid dimensions can be established for differential signal sensing .
the final peak width in the case of differential sensing is caused by space charge at the anode , grid size ( signal increases for denser anode grid ) and by carrier diffusion .
low diffusion rate also leads to better signal - to - noise ratio .
the diffusion length on a distance of detector size should be smaller than anode intergrid spacing .
if diffusion length is comparable or larger than the anode intergrid spacing , the cg and ncg diffusion dominated current will be equal , and , differential signal will be negligible . the shielding has been shown to improve the current waveforms in a solely drift model , as shown experimentally @xcite .
it improves signal to noise ratio , deceases the initial spurious signal caused by positive space charge at the cathode and eliminates noise from the external detector environment .
contrary to that , it slows the detector response time by about @xmath61 and directs carriers into one single conducting channel inside detector .
the channeling effect leads to increased diffusion and electron repulsion even for intermediate times before charge collection .
next , further simplifications of the proposed numerical model are discussed .
it is shown that many current waveform characteristics can be explained separately one by one using simple physical considerations .
these are based on electrostatic potential spatial distribution of the space charge in bulk semiconductor and assuming proper scaling of charge - to - electrode distance .
if charge is close to planar cathode with respect to the cathode dimension , the electric field can be considered constant and electrostatic potential linearly scales with distance from cathode . in the presence of space charge below cathode and charge - cathode proximity , the field scales linearly and electrostatic potential quadratically with the charge - cathode distance .
the schockley - ramo theorem than gives quadratic current waveform .
the situation at the anode has to be split into two regimes .
when electron is far from anode compared to the anode dimension , the electric field can be tought as that of charged wire .
the intergrid potential difference will be negligible and the electron effectively moves in the effective electric field given by mean potential on cg and ncg anode .
this mean potential influences the transient time by the order of 100d / l % , where d is anode width and l is a detector thickness .
when electron approaches anode on a distance much smaller than anode width , the situation becomes equal to the one at cathode in close charge - cathode proximity .
the iterative schemes can be applied to describe analytically field profiles through whole detector volume .
deviations from these assumptions lead to another model parameters .
we have studied negative space charge below anode as one example and discussed diffusion due to localized carriers in one narrow potential minimum when shielding is applied as a second example of these deviations . the first additional parameter is easily taken into account in our drift model , the second one exceeds our model limitations .
the proper treatment of the latter case has to be solved by coupled set of drift - diffusion , continuity and poisson equation .
the numerical hardness of such simulation however hinders basic physical considerations as discussed above .
we have provided simple drift - based model to understand laser induced transient current waveforms .
the model has been compared to experimental data where qualitatively full current waveform has been reproduced .
the capability of the proposed model has been shown to provide even quantitative agreement within limitations given mainly by diffusion and electron repulsion .
we have simplified the model even more based on our experience from the numerical simulations .
the simplified analytical considerations have been shown to provide a fruitful insight into detector operation .
the current waveforms were studied under varied doping density , positive and negative space charge , cathode and intergrid bias , diffusion length and detector shielding .
the trends on how current waveforms are influenced by these parameters provide numerically insight on a detector optimization .
a mutual influence of these parameters has been discussed and their possible contradiction was pointed out . finally , we have quantified increased efficiency of the commonly employed differential sensing in hard x - ray and @xmath6-ray detectors .
the more elaborate extension of the work has been discussed in terms of nonlinear set of coupled drift - diffusion , continuity and poissons equation .
the presented work will be useful for quick insight on a radiation detector operation and optimization for even interdisciplinary scientific community involving broader physics and medical physics .
financial support from the technological agency of the czech republic under project te01020445 , and the grant agency of the czech republic under contract p102/16/23165s are gratefully acknowledged .
38ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty @noop * * , ( ) link:\doibase 10.1088/0957 - 0233/17/4/r01 [ * * , ( ) ] link:\doibase 10.3390/s110505112 [ * * , ( ) ] link:\doibase { 10.1209/0295 - 5075/96/60010 } [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1063/1.2776373 [ * * , ( ) ] link:\doibase 10.1063/1.1502922 [ * * , ( ) ] link:\doibase 10.1103/physrevb.86.245203 [ * * , ( ) ] link:\doibase 10.1103/physrevb.85.115206 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1088/0022 - 3727/42/17/175101 [ * * , ( ) ] link:\doibase 10.1103/physrevb.77.035205 [ * * , ( ) ] link:\doibase 10.1103/physrevb.77.195207 [ * * , ( ) ] link:\doibase 10.1209/0295 - 5075/23/2/004 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1063/1.4963353 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1063/1.2755931 [ * * , ( ) ] link:\doibase 10.1016/j.apradiso.2014.09.003 [ * * , ( ) ] link:\doibase 10.1016/j.vacuum.2015.09.014 [ * * , ( ) ] in @noop _ _ , , vol .
, ( , ) pp .
, link:\doibase 10.1016/j.nima.2006.01.072 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop _ _
( , ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1209/0295 - 5075/110/42002 [ * * , ( ) ] link:\doibase 10.1016/j.cap.2015.04.027 [ * * , ( ) ] | we have modeled laser - induced transient current waveforms in radiation coplanar grid detectors .
poisson s equation has been solved by finite element method and currents induced by photo - generated charge were obtained using shockley - ramo theorem . the spectral response on a radiation flux
has been modeled by monte - carlo simulations .
we show 10@xmath0 improved spectral resolution of coplanar grid detector using differential signal sensing .
we model the current waveform dependence on doping , depletion width , diffusion and detector shielding and their mutual dependence is discussed in terms of detector optimization .
the numerical simulations are successfully compared to experimental data and further model simplifications are proposed . the space charge below electrodes and a non - homogeneous electric field on a coplanar grid anode
are found to be the dominant contributions to laser - induced transient current waveforms . |
noncommutative quantum field theories show in most cases a phenomenon called _ uv / ir - mixing _
@xcite which seems to prevent the perturbative renormalisation .
there is an enormous number of articles on this problem , most of them performing one - loop calculations extrapolated to higher order . a systematic analysis of noncommutative ( massive ) field theories at _ any loop order _ was performed by chepelev and roiban @xcite .
they calculated the integral of an arbitrary feynman graph using the parametric integral representation and expressed the result in terms of determinants involving the incidence matrix and the intersection matrix .
they succeeded to evaluate the leading contribution to the determinants in terms of topological properties of _ ribbon graphs _ wrapped around riemann surfaces . in this way a power - counting theorem was established which led to the identification of two power - counting non - renormalisable classes of ribbon graphs .
rings_-type class consists of graphs with classically divergent subgraphs wrapped around the same handle of the riemann surface .
com_-type class consists of planar graphs with external legs ending at several disconnected boundary components with the momentum flow into a boundary component being identically zero . except for models with enough symmetry , noncommutative field theories are not renormalisable by standard techniques .
one may speculate that the reason is the too nave way of performing the various limits .
namely , a field theory is a dynamical system with infinitely many degrees of freedom defined by a certain limiting procedure of a system with finitely many degrees of freedom .
one may perform the limits formally to the path integral and evaluate it by feynman graphs which are often meaningless .
it is the art of renormalisation to give a meaning to these graphs .
this approach works well in the commutative case , but in the noncommutative situation it seems to be not successful .
a procedure which deals more carefully with the limits is the renormalisation group approach due to wilson @xcite , which was adapted by polchinski to a very efficient renormalisability proof of commutative @xmath1-theory in four dimensions @xcite .
there are already some attempts @xcite to use polchinski s method to renormalise noncommutative field theories .
we are , however , not convinced that the claimed results ( uv - renormalisability ) are so easy to obtain .
the main argument in @xcite is that the polchinski equation is a one - loop equation so that the authors simply compute an integral having exactly one loop .
it is , however , not true that nothing new happens at higher loop order .
for instance , all one - loop graphs can be drawn on a genus - zero riemann surface .
the entire complexity of riemann surfaces of higher genus as discussed by chepelev and roiban @xcite shows up at higher loop order and is completely ignored by the authors of @xcite .
as we show in this paper , the same discussion of riemann surfaces is necessary in the renormalisation group approach , too .
the mentioned complexity is due to the phase factors described by the intersection matrix which result in convergent but not absolutely convergent momentum integrals . as such it is very difficult to access this complexity by polchinski s procedure @xcite which is based on taking _ norms _ of the contributions .
moreover , chepelev and roiban established the link between power - counting and the topology of the riemann surface via the parametric integral representation , which is based on gauian integrations .
these are not available for polchinski s method where we deal with cut - off integrals . in conclusion , we believe it is extremely difficult ( if not impossible ) to use the exact renormalisation group equation for noncommutative field theories in momentum space . the best one can hope is to restrict oneself to limiting cases where e.g. the non - planar graphs are suppressed @xcite .
even this restricted model has rich topological features .
fortunately , there exists a base @xmath2 for the algebra under consideration and its treatment by renormalisation group methods , see @xcite . ] where the @xmath3-product is reduced to an ordinary product of ( infinite ) matrices , @xmath4 , see @xcite .
the interaction @xmath5 can then be written as @xmath6 where @xmath7 is now an infinite matrix ( with entries of rapid decay ) .
the price for the simplification of the interaction is that the kinetic matrix , or rather its inverse , the propagator , becomes very complicated .
however , in polchinski s approach the propagator is anyway made complicated when multiplying it with the smooth cut - off function @xmath8 $ ] .
the parameter @xmath9 is an energy scale which varies between the renormalisation scale @xmath10 and the initial scale @xmath11 .
introducing in the bilinear ( kinetic ) part of the action the cut - off function and replacing the @xmath1-interaction by a @xmath9-dependent effective action @xmath12 $ ] , the philosophy is to determine @xmath12 $ ] such that the generating functional @xmath13 $ ] is actually independent of @xmath9 . in this paper
we provide the prerequisites to investigate the renormalisation of general non - local matrix models .
we prove a power - counting theorem for the effective action @xmath12 $ ] by solving ( better : estimating ) the polchinski equation perturbatively .
our derivation and solution of the matrix polchinski equation combines the original ideas of @xcite with some of the improvements made in @xcite . in particular , we follow @xcite to obtain @xmath14-independent estimations for the interaction coefficients .
the polchinski equation for matrix models can be visualised by ribbon graphs .
the power - counting degree of divergence of a ribbon graph depends on the topological data of the graph and on two scaling dimensions of the cut - off propagator . in this way
, suitable scaling dimensions provide a simple criterion to decide whether a non - local matrix model has the chance of resulting in a renormalisable or not
. however , having the right scaling dimensions is not sufficient for the renormalisability of a model , because a divergent interaction is parametrised by an infinite number of matrix indices .
thus , a renormalisable model needs further structures which relate these infinitely many interaction coefficients to a finite number of base couplings . nevertheless , the dimension criterion proven in this paper is of great value .
for instance , it discards immediately the standard @xmath1-models on noncommutative @xmath0 , @xmath15 , in the matrix base .
these models have the wrong scaling dimension , which is nothing but the manifestation of the old uv / ir - mixing problem @xcite .
looking closer at the origin of the wrong scaling dimensions it is not difficult to find a deformation of the free action which has the chance to be a renormalisable model . in the matrix base of the noncommutative @xmath0 ,
the laplace operator becomes a tri - diagonal band matrix .
the main diagonal behaves nicely , but the two adjacent diagonals are `` too big '' and compensate the desired behaviour of the main diagonal . making the adjacent diagonals `` smaller ''
one preserves the properties of the main diagonal and obtains the good scaling dimensions required for a renormalisable model .
the deformation of the adjacent diagonals corresponds to the inclusion of a harmonic oscillator potential in the free field action .
we treat in @xcite the @xmath1-model on noncommutative @xmath16 within the wilson - polchinski approach in more detail .
we prove that this model is renormalisable when adding the harmonic oscillator potential .
remarkably , the model remains renormalisable when scaling the oscillator potential in a certain way to zero with the removal @xmath17 of the cut - off .
we prove in @xcite that the @xmath1-model on noncommutative @xmath18 is renormalisable to all orders by imposing normalisation conditions for the physical mass , the field amplitude , the frequency of the harmonic oscillator potential and the coupling constant .
in particular , the harmonic oscillator potential can not be removed from the model .
it gives the explicit solution of the uv / ir - duality which suggests that noncommutativity relevant at short distances goes hand in hand with a different physics at very large distances .
the oscillator potential makes the @xmath1-action covariant with respect to a duality transformation @xcite between positions and momenta .
we stress that the power - counting theorem proven in this paper was indispensable to have from the start the right @xmath1-model for the renormalisation proof @xcite .
many noncommutative field theories have a matrix formulation , too .
we think of fuzzy spaces and @xmath19-deformed models .
our general power - counting theorem can play an important rle in the renormalisation proof of these examples .
we consider a @xmath1-matrix model with a general ( non - diagonal ) kinetic term , @xmath20 & = \mathcal{v}_d \big ( \sum_{m , n , k , l } \frac{1}{2 } g_{mn;kl } \phi_{mn } \phi_{kl } + \sum_{m , n , k , l } \frac{\lambda}{4 ! } \phi_{mn } \phi_{nk } \phi_{kl } \phi_{lm } \big)\ ; , \label{actiong}\end{aligned}\ ] ] where @xmath21 . for the noncommutative @xmath0 , @xmath22
even , we have @xmath23 .
the factor @xmath24 is the volume of an elementary cell .
the choice of @xmath1 is no restriction but for us the most natural one because we are interested in four - dimensional models .
standard matrix models are given by @xmath25 for reviews on matrix models and their applications we refer to @xcite .
the idea to apply renormalisation group techniques to matrix models is also not new @xcite .
the difference of our approach is that we will not demand that the action can be written as the trace of a polynomial in the field , that is , we allow for matrix - valued kinetic terms .
the only restriction we are imposing is @xmath26 the restriction ( [ restindex ] ) is due to the fact that the action comes from a trace .
it is verified for the noncommutative @xmath0 due to the @xmath27-symmetry of both the interaction and the kinetic term .
the kinetic matrix @xmath28 contains the entire information about the differential calculus , including the underlying ( riemannian ) geometry , and the masses of the model .
more important than the kinetic matrix @xmath29 will be its inverse , the propagator @xmath30 defined by @xmath31 due to ( [ restindex ] ) we have the same index restrictions for the propagator : @xmath32 let us introduce a notion of locality : a matrix model is called if @xmath33 for some function @xmath34 , otherwise .
[ def - local ] we add sources @xmath35 to the action ( [ actiong ] ) and define a ( euclidean ) quantum field theory by the generating functional ( partition function ) @xmath36&= \int \big(\prod_{a , b } d \phi_{ab}\big ) \,\exp\big(-s[\phi]- \mathcal{v}_d \sum_{m , n } \phi_{mn } j_{nm}\big)\;. \label{pathintm}\end{aligned}\ ] ] according to polchinski s derivation of the exact renormalisation group equation we now consider a ( at first sight ) different problem than ( [ pathintm ] ) . via a cut - off function @xmath37
$ ] , which is smooth in @xmath9 and satisfies @xmath38=1 $ ] , we modify the weight of a matrix index @xmath39 as a function of a certain _ scale _ @xmath9 : @xmath40&= \int \big(\prod_{a , b } d \phi_{ab}\big ) \,\exp\big(-s[\phi , j,\lambda]\big)\ ; , \label{zjl } \\ * s[\phi , j,\lambda ] & = \mathcal{v}_d \big ( \sum_{m , n , k , l } \frac{1}{2 } \phi_{mn } g^k_{mn;kl } ( \lambda)\ , \phi_{kl } + l[\phi,\lambda ] \nonumber \\ * & + \sum_{m , n , k , l } \phi_{mn } f_{mn;kl}[\lambda ] j_{kl } + \sum_{m , n , k , l } \frac{1}{2 } j_{mn } e_{mn;kl}[\lambda ] j_{kl } + c[\lambda ] \big)\ ; , \\ * g^k_{mn;kl}(\lambda ) & : = \big(\prod_{i \in m , n , k , l } k[i,\lambda]^{-1 } \big ) g_{mn;kl } \ ; , \label{pathintp}\end{aligned}\ ] ] with @xmath41=0 $ ] . accordingly , we define @xmath42\big ) \delta_{nm;lk}\;. \label{gkdk1 } \end{aligned}\ ] ] for indices @xmath43 we would write the cut - off function as a product @xmath37=\prod_{i=1}^{\frac{d}{2 } } k\big(\frac{m^i}{(\mathcal{v}_d)^{\frac{2}{d } } \lambda^2}\big)$ ] where @xmath44 is a smooth function on @xmath45 with @xmath46 for @xmath47 and @xmath48 for @xmath49 . in the limit @xmath50 , the partition function ( [ zjl ] )
vanishes unless @xmath51 for @xmath52 , thus implementing a cut - off of the measure @xmath53 in ( [ zjl ] ) .
all other formulae involve positive powers of @xmath54 which multiply through the cut - off propagator ( [ gkdk1 ] ) the appearing matrix indices . in the limit @xmath55 ,
@xmath37 $ ] has finite support in @xmath39 so that all infinite - sized matrices are reduced to finite ones .
the function @xmath56 $ ] is the vacuum energy and the matrices @xmath57 and @xmath58 , which are not necessary in the commutative case , must be introduced because the propagator @xmath30 is non - local .
it is , in general , not possible to separate the support of the sources @xmath35 from the support of the @xmath9-variation of @xmath59 . due to
@xmath38=1 $ ] we formally obtain ( [ pathintm ] ) for @xmath60 in ( [ pathintp ] ) if we set @xmath61 & = \sum_{m , n , k , l } \frac{\lambda}{4 ! } \phi_{mn } \phi_{nk } \phi_{kl } \phi_{lm}\ ; , \hspace*{-10em } \nonumber \\ * c[\infty]&=0\ ; , & e_{mn;kl}[\infty ] & = 0\ ; , & f_{mn;kl}[\infty ] & = \delta_{ml}\delta_{nk}\;. \label{initinfty}\end{aligned}\ ] ] however , we shall expect divergences in the partition function which require a renormalisation , i.e. additional ( divergent ) counterterms in @xmath62 $ ] . in the feynman graph solution of the partition function one
carefully adapts these counterterms so that all divergences disappear .
if such an adaptation is possible with a _ finite number _ of _ local _ counterterms , the model is considered as perturbatively renormalisable .
following polchinski @xcite we proceed differently to prove renormalisability .
we first ask ourselves how to choose @xmath63 in order to make @xmath13 $ ] _ independent _ of @xmath9 .
after straightforward calculation one finds the answer @xmath64 & = 0\qquad \text{iff } \\
\lambda \frac{\partial l[\phi,\lambda]}{\partial \lambda } & = \sum_{m , n , k , l } \frac{1}{2 } \lambda \frac{\partial \delta^k_{nm;lk}(\lambda)}{\partial \lambda } \bigg ( \frac{\partial l[\phi,\lambda]}{\partial \phi_{mn}}\frac{\partial l[\phi,\lambda]}{\partial \phi_{kl } } - \frac{1}{\mathcal{v}_d } \big[\frac{\partial^2 l[\phi,\lambda]}{\partial \phi_{mn}\,\partial \phi_{kl}}\big]_\phi \bigg ) \ ; , \label{poll } \\ \lambda\frac{\partial f_{mn;kl}[\lambda]}{\partial \lambda } & = -\sum_{m',n',k',l ' } g^k_{mn;m'n ' } ( \lambda ) \ , \lambda \frac{\partial \delta^k_{n'm';l'k'}(\lambda)}{\partial \lambda } f_{k'l';kl}[\lambda]\ ; , \label{polf } \\
\lambda \frac{\partial e_{mn;kl}[\lambda]}{\partial \lambda } & = -\sum_{m',n',k',l ' } f^t_{mn;m'n ' } [ \lambda ] \ ,
\lambda \frac{\partial \delta^k_{n'm';l'k'}(\lambda)}{\partial \lambda } f_{k'l';kl}[\lambda]\ ; , \label{pole } \\
\lambda \frac{\partial c[\lambda]}{\partial \lambda } & = \frac{1}{\mathcal{v}_d } \sum_{m , n } \lambda \frac{\partial}{\partial \lambda } \ln \big ( k[m,\lambda ] k[n,\lambda ] \big ) \nonumber \\ * & -\frac{1}{2\mathcal{v}_d } \sum_{m , n , k , l } \lambda \frac{\partial \delta^k_{nm;lk}(\lambda)}{\partial \lambda } \frac{\partial^2 l[\phi,\lambda]}{\partial \phi_{mn}\,\partial \phi_{kl}}\big|_{\phi=0}\ ; , \label{polc}\end{aligned}\ ] ] where @xmath65\big]_\phi:=f[\phi]-f[0]$ ] .
navely we would integrate ( [ poll])([polc ] ) for the initial conditions ( [ initinfty ] ) .
technically , this would be achieved by imposing the conditions ( [ initinfty ] ) not at @xmath66 but at some finite scale @xmath67 , followed by taking the limit @xmath68 .
this is easily done for ( [ polf])([polc ] ) : @xmath69 & = \sum_{m',n ' } g^k_{mn;m'n'}(\lambda ) \delta^k_{n'm';kl}(\lambda_0 ) \ ; , \\
e_{mn;kl}[\lambda ] & = \sum_{m',n',k',l ' } \delta^k_{mn;m'n ' } ( \lambda_0 ) \ ,
\big(g^k_{n'm';l'k'}(\lambda ) - g^k_{n'm';l'k'}(\lambda_0 ) \big ) \delta^k_{k'l';kl}(\lambda_0)\ ; , \\ c[\lambda ] & = \frac{2}{\mathcal{v}_d } \ln\big ( \prod_{m } k[m,\lambda ] k^{-1}[m,\lambda_0 ] \big ) \nonumber \\ * & + \frac{1}{2\mathcal{v}_d } \int_{\lambda}^{\lambda_0 } \!\!d\lambda'\ ; \sum_{m , n , k , l } \frac{\partial \delta^k_{nm;lk}(\lambda')}{\partial \lambda ' } \frac{\partial^2 l[\phi,\lambda']}{\partial \phi_{mn}\,\partial \phi_{kl}}\big|_{\phi=0}\;.\end{aligned}\ ] ] at @xmath67 the functions @xmath70 become independent of @xmath14 and satisfy , in particular , ( [ initinfty ] ) in the limit @xmath68 .
the partition function @xmath13 $ ] is evaluated by feynman graphs with vertices given by the taylor expansion coefficients @xmath71 : = \frac{1}{n ! } \big(\frac{\partial^n l[\phi,\lambda]}{\partial \phi_{m_1n_1}\ , \partial \phi_{m_2n_2}\ , \dots \partial \phi_{m_nn_n } } \big)_{\phi=0 } \label{ltaylor}\end{aligned}\ ] ] connected with each other by internal lines @xmath72 and to sources @xmath35 by external lines @xmath73 . as @xmath37 $ ] has finite support in @xmath39 for finite @xmath9 , the summation variables in the above feynman graphs are via the propagator @xmath72 restricted to a finite set .
thus , loop summations are finite , provided that the interaction coefficients @xmath74 $ ] are bounded . in other words
, for the renormalisation of a non - local matrix model it is necessary to prove that the differential equation ( [ poll ] ) admits a regular solution .
as pointed out in the introduction , to obtain a physically reasonable quantum field theory one has additionally to prove that there is a regular solution of ( [ poll ] ) which depends on a _ finite number of initial conditions _ only .
this requirement is difficult to fulfil because there is , a priori , an infinite number of degrees of freedom given by the taylor expansion coefficients ( [ ltaylor ] ) .
this is the reason for the fact that renormalisable ( four - dimensional ) quantum field theories are rare .
we are going to integrate ( [ poll ] ) between a certain renormalisation scale @xmath10 and the initial scale @xmath14 .
we assume that @xmath75 can be decomposed into parts @xmath76 which for @xmath77 scale homogeneously : @xmath78 } { \partial \lambda}\big| \leq \lambda^{r_{i}}\ ; p^{q_i}\big[\ln \frac{\lambda}{\lambda_r } \big]\;. \label{difftoy}\end{aligned}\ ] ] here , @xmath79 \geq 0 $ ] stands for some polynomial of degree @xmath19 in @xmath80 .
clearly , @xmath79 $ ] , for @xmath81 , can be further bound by a polynomial with non - negative coefficients . as usual
we define homogeneous parts @xmath82 $ ] in ( [ difftoy ] ) with @xmath83 are called , with @xmath84 and with @xmath85 .
there are two possibilities for the integration , either from @xmath14 down to @xmath9 or from @xmath10 up to @xmath9 , corresponding to the identities @xmath86 \nonumber \\ * & = l_{m_1n_1;\dots;m_nn_n}^{(i)}[\lambda_0 ] - \int_{\lambda}^{\lambda_0 } \frac{d\lambda'}{\lambda ' } \big ( \lambda'\frac{\partial}{\partial \lambda'}l_{m_1n_1;\dots;m_nn_n}^{(i)}[\lambda ' ] \big ) \label{int - down } \\ & = l_{m_1n_1;\dots;m_nn_n}^{(i)}[\lambda_r ] + \int_{\lambda_r}^{\lambda } \frac{d\lambda'}{\lambda ' } \big ( \lambda'\frac{\partial}{\partial \lambda'}l_{m_1n_1;\dots;m_nn_n}^{(i)}[\lambda ' ] \big ) \;. \label{int - up}\end{aligned}\ ] ] one has @xmath87
\displaystyle \frac{1}{q{+}1 } \big(\ln \frac{x}{x_r}\big)^{q+1 } + \text{const } \qquad & \text{for } r=0 \;. \end{array}\right . \label{intlog}\end{aligned}\ ] ] at the end we are interested in the limit @xmath68 .
this requires that positive powers of @xmath14 must be prevented in the estimations .
for @xmath84 we we can safely take the direction ( [ int - down ] ) of integration and then , because all coefficients are positive , the limit @xmath17 in the integral of ( [ int - down ] ) .
thus , @xmath88 \big| & \leq \big|l_{m_1n_1;\dots;m_nn_n}^{(i)}[\lambda_0 ] \big| + \int_{\lambda}^{\infty } \frac{d\lambda'}{\lambda ' } \big| \lambda'\frac{\partial}{\partial \lambda'}l_{m_1n_1;\dots;m_nn_n}^{(i)}[\lambda ' ] \big| \nonumber \\ * & \leq \big|l_{m_1n_1;\dots;m_nn_n}^{(i)}[\lambda_0 ] \big| + \lambda^{-|r_i| } p^{q_i}\big[\ln\frac{\lambda}{\lambda_r}\big]\;,\qquad \text{for } r_i<0\;. \label{intdown}\end{aligned}\ ] ] here , @xmath89 is a new polynomial of degree @xmath90 with non - negative coefficients .
now , the limit @xmath17 carried out later requires that @xmath91 \big|$ ] in ( [ intdown ] ) is bounded , i.e. @xmath91 \big| < \frac{c}{\lambda_0^{s_i}}$ ] , with @xmath92 . as the resulting estimation ( [ intdown ] ) is further iterated , @xmath93 must be sufficiently large .
we do not investigate this question in detail and simply note that it is safe to require @xmath94 \big| < \lambda_0^{-|r_i| } p^{q_i}\big[\ln\frac{\lambda_0}{\lambda_r}\big]\;,\qquad \text{for } r_i<0\ ; , \label{intdown0}\end{aligned}\ ] ] for the boundary condition . in the other case @xmath95 , the integration direction ( [ int - down ] )
will produce divergences in @xmath17 .
thus , we have to choose the other direction ( [ int - up ] ) .
the integration ( [ intlog ] ) produces alternating signs , but these can be ignored in the maximisation .
the only contribution from the lower bound @xmath10 in the integral of ( [ int - up ] ) is the term with @xmath96 in ( [ intlog ] ) .
there , we can obviously ignore it in the difference @xmath97 .
we thus obtain from ( [ intlog ] ) the estimation @xmath88
\big| \leq \big|l_{m_1n_1;\dots;m_nn_n}^{(i)}[\lambda_r ] \big| + \left\ { \begin{array}{ll } \lambda^{r_i}p^{q_i}\big[\ln\frac{\lambda}{\lambda_r}\big ] \qquad & \text{for } r_i > 0 \\ p^{q_i+1}\big[\ln\frac{\lambda}{\lambda_r}\big ] \qquad & \text{for } r_i = 0\;. \end{array}\right . \label{intup}\end{aligned}\ ] ] the reduction from @xmath98 $ ] in polchinski s original work @xcite to @xmath99 $ ] is due to @xcite .
we can summarise these considerations as follows : [ def - lem1 ] let @xmath100\big|$ ] be bounded by ( [ difftoy ] ) , @xmath101 } { \partial \lambda}\big| \leq \lambda^{r_{i } } p^{q_i}\big [ \ln \frac{\lambda}{\lambda_r}\big ] \;.
\label{difftoy1}\end{aligned}\ ] ] the integration of ( [ difftoy1 ] ) is for irrelevant interactions performed from @xmath14 down to @xmath9 starting from an initial condition bounded by @xmath91 \big| < \lambda_0^{-|r_i| } p^{q_i}\big[\ln\frac{\lambda_0}{\lambda_r}\big]$ ] . for relevant and marginal interactions we have to integrate ( [ difftoy1 ] ) from @xmath10 up to @xmath9 , starting from an initial condition @xmath102
< \infty$ ] . under these conventions we have @xmath88\big| \leq \lambda^{r_{i } }
p^{q_i+1}\big [ \ln \frac{\lambda}{\lambda_r}\big]\;. \label{intresult}\end{aligned}\ ] ] let us give a few comments : * the stability ( [ difftoy1 ] ) versus ( [ intresult ] ) of the estimation will be very useful in the iteration process .
* integrations according to the direction ( [ int - up ] ) , which entail an initial condition @xmath102 $ ] , are expensive for renormalisation , because each such condition ( even the choice @xmath102=0 $ ] ) corresponds to a normalisation experiment . in order to have a meaningful theory , there has to be only a finite number of required normalisation experiments .
initial data at @xmath14 do not correspond to normalisation conditions , because the interaction at @xmath103 is experimentally not accessible . moreover , unless artificially kept alive is the initial @xmath1-interaction in two - dimensional models @xcite . ] , an irrelevant coupling scales away for @xmath17 via its own dynamics .
the property @xmath104=0 $ ] for an irrelevant coupling is , therefore , a result and no condition .
* there might be cases where the direction ( [ int - up ] ) for @xmath84 gives convergence for @xmath68 nevertheless .
this corresponds to the over - subtractions in the bphz renormalisation scheme .
we shall not exploit this possibility .
unless there are further correlations between functions with different indices , specifying @xmath102 $ ] means to impose an _ infinite number _ of normalisation conditions ( because of @xmath105 ) .
hence , a non - local matrix model with relevant and/or marginal interactions can only be renormalisable if some additional structures exist which relate all divergent functions to a finite number of relevant / marginal base interactions .
such a distinguished property depends crucially on the model .
presumably , the class of models where such a reduction is possible is rather small .
it can not be the purpose of this paper to analyse these reductions . instead , our strategy is to find the general power - counting behaviour of a non - local matrix model which limits the class of divergent functions among which the reduction has to be studied in detail .
for example , we will find that under very general conditions on the propagator all non - planar graphs ( as defined below ) are irrelevant .
such a result is already an enormous gain for the detailed investigation of a model .
thus , our strategy is to integrate the polchinski equation ( [ poll ] ) perturbatively between two scales @xmath10 and @xmath14 for a self - determined choice of the boundary condition according to definition / lemma [ def - lem1 ] .
the resulting normalisation condition for relevant and marginal interactions will not be the correct choice for a renormalisable model .
nevertheless , the resulting estimation ( [ intresult ] ) is compatible with a more careful treatment .
taking the example @xcite of the @xmath1-model on noncommutative @xmath18 , we would replace * almost all of the relevant functions with bound @xmath106 $ ] in ( [ intresult ] ) by irrelevant functions with bound @xmath107 $ ] , and * almost all marginal functions with bound @xmath108 $ ] in ( [ intresult ] ) by irrelevant functions with bound @xmath109 $ ] , for some reference scale @xmath110 .
we can symbolise the expansion coefficients @xmath75 as @xmath111 the big circle stands for a possibly very complex interior and the outer ( dotted ) double lines stand for the valences produced by differentiation ( [ ltaylor ] ) with respect to the @xmath112 fields @xmath113 .
the arrows are merely added for bookkeeping purposes in the proof of the power - counting theorem .
since we work with real fields , i.e. @xmath114 , the expansion coefficients @xmath75 have to be unoriented .
the situation is different for complex fields where @xmath115 leads to an orientation of the lines . in this case
we would draw both arrows at the double line either incoming or outgoing . the graphical interpretation of the polchinski equation ( [ poll ] )
is found when differentiating it with respect to the fields @xmath113 : @xmath116 combinatorical factors are not shown and a symmetrisation in all indices @xmath117 has to be performed .
on the rhs of ( [ pollgraph ] ) the two valences @xmath118 and @xmath119 of the subgraphs are connected to the ends of a _ ribbon _ which symbolises the differentiated propagator @xmath120 . for local matrix models in the sense of definition
[ def - local ] we can regard the ribbon as a product of single lines with interaction given by @xmath34 . for non - local matrix models
there is an exchange of indices within the entire ribbon
. we can regard ( [ poll ] ) as a formal construction scheme for @xmath12 $ ] if we introduce a grading @xmath12=\sum_{v=1}^\infty \lambda^v l^{(v)}[\phi,\lambda]$ ] and additionally impose a cut - off in @xmath112 for @xmath121 , i.e @xmath122=0 \qquad \text{for } n > n_0\;.\end{aligned}\ ] ] in order to obtain a @xmath1-model we choose @xmath123 and the grading as the degree @xmath124 in the coupling constant @xmath125 .
we conclude from ( [ poll ] ) that @xmath126 is independent of @xmath9 so that it is identified with the original @xmath127-interaction in ( [ actiong ] ) : @xmath128 \hspace*{-3em } & \nonumber \\ * = \frac{1}{4 ! \,6 } \big ( & \delta_{n_1 m_2 } \delta_{n_2 m_3 } \delta_{n_3 m_4 } \delta_{n_4 m_1 } + \delta_{n_1 m_3 } \delta_{n_3 m_4 } \delta_{n_4 m_2 } \delta_{n_2 m_1 } \nonumber \\
+ & \delta_{n_1 m_4 } \delta_{n_4 m_2 } \delta_{n_2 m_3 } \delta_{n_3 m_1 } + \delta_{n_1 m_4 } \delta_{n_4 m_3 } \delta_{n_3 m_2 } \delta_{n_2 m_1 } \nonumber \\
+ & \delta_{n_1 m_3 } \delta_{n_3 m_2 } \delta_{n_2 m_4 } \delta_{n_4 m_1 } + \delta_{n_1 m_2 } \delta_{n_2 m_4 } \delta_{n_4 m_3 } \delta_{n_3 m_1 } \big)\;. \label{l4start}\end{aligned}\ ] ] to the first term on the rhs of ( [ l4start ] ) we associate the graph @xmath129 the graphs for the other five terms are obtained by permutation of indices . as mentioned before
, a complex @xmath1-model would be given by oriented propagators @xmath130 and examples for vertices are @xmath131 the consequence is that many graphs of the real @xmath1-model are now excluded .
we can thus obtain the complex @xmath1-model from the real one by deleting the impossible graphs .
the iteration of ( [ pollgraph ] ) with starting point ( [ l4graph ] ) leads to _ ribbon graphs_. the first examples of the iteration are @xmath132 we can obviously build very complicated ribbon graphs with crossings of lines which can not be drawn any more in a plane .
a general ribbon graph can , however , be drawn on a _ riemann surface _ of some _ genus _ @xmath133 .
in fact , a ribbon graph _ defines _ the riemann surfaces topologically through the _ euler characteristic _ @xmath134 .
we have to regard here the external lines of the ribbon graph as amputated ( or closed ) , which means to directly connect the single lines @xmath135 with @xmath136 for each external leg @xmath117 .
a few examples may help to understand this procedure : @xmath137 \parbox{45mm}{\begin{picture}(30,25 ) \put(0,0){\epsfig{scale=.9,file = gt3,bb=71 612 204 681 } } \put(0,11.5){\mbox{\scriptsize$n_1 $ } } \put(3,17.5){\mbox{\scriptsize$m_1 $ } } \put(15,11.5){\mbox{\scriptsize$m_2 $ } } \put(17,17.5){\mbox{\scriptsize$n_2 $ } } \end{picture } } \quad \rightarrow \quad \parbox{35mm}{\begin{picture}(30,25 ) \put(0,0){\epsfig{scale=.9,file = gt4,bb=71 612 178 681 } } \end{picture } } \qquad \begin{array}{r@{\,}l@{\qquad}r@{\,}ll } \tilde{l } & = 1 & b & = 1 \\ i & = 3 & n & = 2 \\ v & = 2 & v^e & = 1 \\ g & = 1 & \iota & = 1 \end{array } \label{gt3 } \\
\parbox{40mm}{\begin{picture}(30,25 ) \put(0,0){\epsfig{scale=.9,file = gt5,bb=71 625 187 684 } } \put(15.5,19){\mbox{\scriptsize$n_1 $ } } \put(22.5,16){\mbox{\scriptsize$m_1 $ } } \put(6,19){\mbox{\scriptsize$n_2 $ } } \put(12.5,16){\mbox{\scriptsize$m_2 $ } } \put(0,7.5){\mbox{\scriptsize$n_3 $ } } \put(2,13){\mbox{\scriptsize$m_3 $ } } \put(5,4){\mbox{\scriptsize$m_4 $ } } \put(12.5,1){\mbox{\scriptsize$n_4 $ } } \put(15,4){\mbox{\scriptsize$m_5 $ } } \put(22,1){\mbox{\scriptsize$n_5 $ } } \put(25,4){\mbox{\scriptsize$m_6 $ } } \put(32,1){\mbox{\scriptsize$n_6 $ } } \end{picture } } \quad \rightarrow \quad \parbox{35mm}{\begin{picture}(30,10 ) \put(0,0){\epsfig{scale=.9,file = gt6,bb=71 651 161 684 } } \end{picture } } \qquad \begin{array}{r@{\,}l@{\qquad}r@{\,}ll } \tilde{l } & = 2 & b & = 1 \\
i & = 3 & n & = 6 \\ v & = 3 & v^e & = 3 \\ g & = 0 & \iota & = 0 \end{array } \label{gt5}\end{aligned}\ ] ] the genus is computed from the number @xmath138 of single - line loops of the closed graph , the number @xmath139 of internal ( double ) lines and the number @xmath124 of vertices of the graph according to @xmath140 there can be several possibilities to draw the graph and its riemann surface , but @xmath141 and thus @xmath133 remain unchanged . indeed , the polchinski equation ( [ poll ] ) interpreted as in ( [ pollgraph ] ) tells us which external legs of the vertices are connected .
it is completely irrelevant how the ribbons are drawn between these legs . in particular
, there is no distinction between overcrossings and undercrossings .
there are two types of loops in ( amputated ) ribbon graphs : * some of them carry at least one external leg . they are called _ boundary components _ ( or holes of the riemann surface ) .
their number is @xmath142 . *
some of them do not carry any external leg .
they are called _
inner loops_. their number is @xmath143 .
boundary components consist of a concatenation of _ trajectories _ from an incoming index @xmath136 to an outgoing index @xmath144 . in the example ( [ gt1 ] )
the inner boundary component consists of the single trajectory @xmath145 whereas the outer boundary component is made of two trajectories @xmath146 and @xmath147 .
we let @xmath148 $ ] be the outgoing index to @xmath149 and @xmath150 $ ] be the incoming index to @xmath144 . we have to introduce a few additional notations for ribbon graphs .
external vertex _ is a vertex which has at least one external leg .
we denote by @xmath151 the total number of external vertices . for the arrangement of external legs at an external vertex
there are the following possibilities : @xmath152 we call the first three types of external vertices _ simple vertices_. they provide one starting point and one end point of trajectories through a ribbon graph .
the fourth vertex in ( [ vertexe ] ) is called _ composed vertex_. it has two starting points and two end points of trajectories .
a composed vertex can be decomposed by pulling the two propagators with attached external lines apart : @xmath153 in this way a given graph with composed vertices is decomposed into @xmath154 _ segments_. the external vertices of the segments are either true external vertices or the halves of a composed vertex .
if composed vertices occur in loops , their decomposition does not always increase the number of segments .
we need the following [ defiota ] the @xmath155 of a graph is the maximal number of decompositions of composed vertices which keep the graph connected .
it follows immediately that if @xmath156 is the number of composed vertices of a graph and @xmath154 the number of segments obtained by decomposing all composed vertices we have @xmath157 in order to evaluate @xmath74 $ ] by connection and contraction of subgraphs according to ( [ pollgraph ] ) we need estimations for _ index summations _ of ribbon graphs .
namely , our strategy is to apply the summations in ( [ pollgraph ] ) either to the propagator or the subgraph only and to maximise the other object over the summation indices .
we agree to fix all starting points of trajectories and sum over the end points of trajectories .
however , due to ( [ restindexd ] ) and ( [ l4start ] ) not all summations are independent : the sum of outgoing indices equals for each segment the sum of incoming indices
. since there are @xmath158 ( end points of ) trajectories in a ribbon graph , there are @xmath159 independent index summations .
the inequality ( [ numsum ] ) also holds for the restriction to each segment if @xmath151 includes the number of halves of composed vertices belonging to the segment .
we let @xmath160 be the set of @xmath161 end points of trajectories in a graph over which we are going to sum , keeping the starting points of these trajectories fixed .
we define @xmath162=\text{const } } \qquad \text{if } \mathcal{e}^{s } = \{m_1,m_2,\dots , m_{s}\}\;.\end{aligned}\ ] ] taking the example of the graph ( [ gt1 ] ) , we can due to @xmath163 apply up to two index summations , i.e. a summation over at most two of the end points of trajectories @xmath164 , where the corresponding incoming indices @xmath165=n_5 $ ] , @xmath166=n_3 $ ] and @xmath167=n_1 $ ] are kept fixed . for the example of the graph ( [ gt3 ] ) we can due to @xmath168 apply at most one index summation , either over @xmath169 for fixed @xmath170=n_2 $ ] or over @xmath165=n_1 $ ] . for @xmath171 we would consider @xmath172
note that for given @xmath173 the other outgoing index is determined to @xmath174 through index conservation at propagators ( [ restindexd ] ) and vertices ( [ l4graph ] ) .
it is part of the proof to show that the index summation ( [ indsumexample ] ) is bounded independently of the incoming indices @xmath175 .
we first have to transform the polchinski equation ( [ poll ] ) into a dimensionless form .
it is important here that in the class of models we consider there is always a dimensionful parameter , @xmath176 which instead of @xmath9 can be used to absorb the mass dimensions .
the effective action @xmath12 $ ] has total mass dimension @xmath22 , a field @xmath7 has dimension @xmath177 and the dimension of the coupling constant for the @xmath178 interaction is @xmath179 .
we thus decompose @xmath12 $ ] according to the number of fields and the order in the coupling constant : @xmath180 = \sum_{v=1}^\infty \sum_{n=2}^{2v+2 } \frac{1}{n ! } \sum_{m_i , n_i } \big(\frac{\lambda}{\mu^{4-d}}\big)^v \!\ !
a^{(v)}_{m_1n_1;\dots;m_nn_n}[\lambda ] \big(\frac{\phi_{m_1n_1}}{\mu^{\frac{d-2}{2}}}\big ) \cdots \big(\frac{\phi_{m_nn_n}}{\mu^{\frac{d-2}{2}}}\big ) \;. \label{ltaylor1}\end{aligned}\ ] ] the functions @xmath181 $ ] are assumed to be symmetric in their indices @xmath117 .
inserted into ( [ poll ] ) we get @xmath182 \nonumber \\ * & = \sum_{m , n , k , l } \frac{1}{2 } q_{nm;lk}(\lambda ) \bigg\ { \sum_{n_1=2}^n \sum_{v_1=1}^{v-1 } a^{(v_1)}_{m_1n_1;\dots;m_{n_1 - 1}n_{n_1 - 1};mn}[\lambda ] a^{(v - v_1)}_{m_{n_1}n_{n_1};\dots;m_{n}n_{n};kl}[\lambda ] \nonumber \\*[-1ex ] & \hspace*{15em } + \big(\binom{n}{n_1{-}1 } -1\big ) \text { permutations } \nonumber \\ * & \hspace*{8em } - a^{(v)}_{m_1n_1;\dots;m_{n}n_{n};mn;kl}[\lambda]\bigg\}\ ; , \label{poll2}\end{aligned}\ ] ] where @xmath183 the permutations refer to the possibilities to choose @xmath184 of the pairs of indices @xmath185 which label the external legs of the first @xmath186-function .
the cut - off function @xmath59 in ( [ gkdk1 ] ) has to be chosen such that for finite @xmath9 there is a finite number of indices @xmath187 with @xmath188 . by suitable normalisation
we can achieve that the volume of the support of @xmath189 with respect to a chosen index scales as @xmath190 : @xmath191\big| \leq c_d \big(\frac{\lambda}{\mu}\big)^d\ ; , \label{volumefactor}\end{aligned}\ ] ] for some constant @xmath192 independent of @xmath9 .
for such a normalisation we define two exponents @xmath193 by @xmath194 in ( [ est1 ] ) the index @xmath195 is kept constant for the summation over @xmath196 .
it is convenient to encode the dimension @xmath22 in a further exponent @xmath197 which describes the product of ( [ volumefactor ] ) with ( [ est0 ] ) : @xmath198 \big| & \leq c_2 \big(\frac{\lambda}{\mu}\big)^{\delta_2 } \;.
\label{est2}\end{aligned}\ ] ] we have obviously @xmath199 and @xmath200 . a non - local matrix model defined by the cut - off propagator @xmath201 given by ( [ gkdk1 ] ) and
( [ q ] ) and the normalisation ( [ volumefactor ] ) of the cut - off function is called if @xmath202 , otherwise .
[ def - regular ] the three exponents @xmath203 play an essential rle in the power - counting theorem which yields the @xmath9-scaling of a homogeneous part @xmath204 $ ] of the interaction coefficients @xmath205 \nonumber \\ * & = \sum_{1 \leq v^e \leq v}\ ; \sum_{1 \leq b \leq
n } \ ; \sum_{0 \leq g \leq 1+\frac{v}{2}-\frac{n}{4}-\frac{b}{2 } } \ ; \sum_{0 \leq \iota\leq b-1 } a^{(v , v^e , b , g,\iota ) } _ { m_1n_1;\dots;m_nn_n}[\lambda]\big|_{2 \leq n \leq 2v+2}\;. \label{hompart}\end{aligned}\ ] ] it is important that the sums over the graphical ( topological ) data @xmath206 in ( [ hompart ] ) are finite .
we are going to prove [ thm1 ] the homogeneous parts @xmath207 $ ] of the coefficients of the effective action describing a @xmath1-matrix model with initial interaction ( [ l4start ] ) and cut - off propagator characterised by the three exponents @xmath203 are for @xmath208 and @xmath209 bounded by @xmath210 \big| & \leq \big(\frac{\lambda}{\mu}\big)^{\delta_2 ( v-\frac{n}{2}+2 - 2g - b ) } \big(\frac{\mu}{\lambda}\big)^{\delta_1(v - v^e-\iota+2g+b-1+s ) } \nonumber \\ * & \times \big(\frac{\mu}{\lambda}\big)^{\delta_0(v^e+\iota-1-s ) } \,p^{2v-\frac{n}{2}}\big[\ln \frac{\lambda}{\lambda_r}\big]\ ; , \label{annorm1}\end{aligned}\ ] ] provided that for all @xmath211 and @xmath212 the initial conditions for relevant / marginal ( irrelevant ) @xmath213 $ ] are imposed at @xmath10 ( @xmath14 ) , respectively , according to definition / lemma [ def - lem1 ] . the bound ( [ annorm1 ] ) is independent of the unsummed indices @xmath214 .
we have @xmath215\equiv 0 $ ] for @xmath216 or @xmath217 . the proof will be given in section [ appd ] .
we remark that @xmath218 is the number of inner loops of a graph .
the power - counting estimation ( [ annorm1 ] ) does not make any reference to the initial scale @xmath14 @xcite so that we can safely take the limit @xmath17 .
in this way we have constructed a regular solution of the polchinski equation ( [ poll ] ) associated with the non - local matrix model .
however , this solution remains useless unless it can be achieved by a _
finite number _ of integrations from @xmath10 to @xmath9 depending on a finite number of initial conditions at @xmath10 .
we refer to the remarks following definition / lemma [ def - lem1 ]
. a first step would be to achieve regular scaling dimensions : for regular matrix models according to definition [ def - regular ] we have independently of the segmentation index and the numbers of external vertices @xmath219 \big| \leq \big(\frac{\lambda}{\mu}\big)^{\omega - d(2g+b-1 ) } p^{2v-\frac{n}{2}}\big[\ln \frac{\lambda}{\lambda_r}\big]\ ; , \label{cor1eq}\end{aligned}\ ] ] where @xmath220 is the classical power - counting degree of divergence .
[ cor1 ] we have derived the relation ( [ cor1eq ] ) with respect to the classical power - counting degree of divergence only for @xmath1-matrix models , but it is plausible that it also holds for more general interactions .
we provide here the proof of theorem [ thm1 ] , which is quite long and technical .
the proof amounts to study all possible connections of two external legs of either different graphs or the same graph .
it will be essential how the legs to connect are situated with respect to the remaining part of the graph .
there are the following arrangements of the external legs at the distinguished vertex one ( or two ) of which we are going to connect : @xmath221 a big oval stands for other parts of the graph the specification of which is not necessary for the proof .
dotted lines entering and leaving the oval stand for the set of all external legs different from the external legs of the distinguished vertex to contract .
if two or three internal lines are connected to the oval this does not necessarily mean that these two lines are part of an inner loop .
we are going to integrate the polchinski equation ( [ poll2 ] ) by induction upward in @xmath124 and for constant @xmath124 downward in @xmath112 . due to the grading @xmath222 ,
the differential equation ( [ poll2 ] ) is actually constructive .
we consider in section [ tree - contr ] the connection of two smaller graphs of @xmath223 and @xmath224 vertices and external legs and in sections [ self - contr-1 ] and [ self - contr-2 ] the self - contraction of a graph with @xmath225 vertices and external legs .
these graphs are further characterised by @xmath226 external vertices , boundary components , genera and segmentation indices , respectively . since the sums in ( [ hompart ] ) and the number of arrangements of legs in ( [ as ] ) are finite
, it is sufficient to regard the contraction of subgraphs individually .
that is , we consider individual subgraphs @xmath227 the contraction of which produces an individual graph @xmath228 .
we also ignore the problem of making the graphs symmetric in the indices @xmath117 of the external legs .
at the very end we project the sum of graphs @xmath228 to homogeneous degree @xmath229 . to these homogeneous parts there contributes according to ( [ hompart ] ) a finite number of contractions of @xmath230 .
we thus get the bound ( [ annorm1 ] ) if we can prove it for any individual contraction .
the theorem is certainly correct for the initial @xmath1-interaction ( [ l4start ] ) which due to ( [ ltaylor1 ] ) gives @xmath231| \leq 1 $ ] .
we start with the first term on the rhs of ( [ poll2 ] ) which describes the connection of two smaller subgraphs @xmath227 of @xmath232 vertices and @xmath233 external legs via a propagator .
the total graph @xmath228 for a tree - contraction has @xmath234 because two loops of the subgraphs are merged to a new loop in the total graph .
it follows from ( [ euler ] ) that for tree - contractions we always have additivity of the genus , @xmath235 as an example for a contraction between graphs in the first line of ( [ as ] ) let us consider @xmath236 where @xmath237 and @xmath238 stand for the set of all other outgoing and incoming indices via external legs at the remaining part of the left subgraph @xmath239 and similarly for @xmath240 and @xmath241 for the right subgraph @xmath242 .
the two boundary components to which the contracted vertices belong are joint in the total graph , i.e. @xmath243 .
moreover , we obviously have @xmath244 and @xmath245 .
the graph ( [ aa1 ] ) determines the @xmath9-scaling @xmath246 \nonumber \\ * & = \frac{1}{2 } \sum_{m , l } a^{(v_1,v^e_1,b_1,g_1,\iota_1)\gamma_1 } _ { m_1n_1;m_2n_2 ; \sigma m \,\sigma n;m m_1}[\lambda ] \ , q_{m_1m;ll_1}(\lambda)\ , a^{(v_2,v^e_2,b_2,g_2,\iota_2)\gamma_2 } _ { l_1l;\sigma k \,\sigma l ; k_2l_2;k_1l_1}[\lambda ] \;.\end{aligned}\ ] ] due to the conservation of the total amount of indices in @xmath239 and @xmath242 by induction hypothesis ( [ annorm1 ] ) , both @xmath247 are completely fixed by the other external indices so that from the sum over @xmath39 and @xmath248 there survives a single term only .
then , because of the relation @xmath249 from the propagator @xmath250 , see ( [ est0 ] ) , it follows that the total amount of indices for @xmath251 is conserved as well .
let @xmath252 and @xmath253 be the numbers of external vertices and segmentation indices on the segments of the subgraphs @xmath230 on which the contracted vertices are situated .
the induction hypothesis ( [ annorm1 ] ) gives us the bound if these segments carry @xmath254 index summations .
the new segment of the total graph @xmath228 created by connecting the boundary components of @xmath230 carries @xmath255 external vertices and @xmath256 segmentation indices and therefore admits up to @xmath257 index summations . in ( [ aa1 ] )
that additional index summation will be the @xmath169-summation .
due to ( [ numsum ] ) ( for segments ) there has to be an external leg on each segment the outgoing index of which is not allowed to be summed .
if on the @xmath242-part of the contracted segment there is an unsummed external leg , we can choose @xmath39 as that particular index in @xmath239 .
in this case we take in the propagator the maximum over @xmath258 and sum the part @xmath242 for given @xmath248 over those indices which belong to @xmath160 .
the result is bounded independently of @xmath248 and all other incoming indices .
next , we sum over the indices in @xmath160 which belong to @xmath239 , regarding @xmath39 as an unsummed index .
there is the possibility of an @xmath169-summation applied to the propagator in the last step , with @xmath259 kept fixed , for which the bound is given by ( [ est1 ] ) . in this case we therefore get @xmath260 \big|\nonumber \\ * & \qquad \leq \frac{1}{2 } \big ( \sum_{\mathcal{e}_1^{s_1 } } \big|a^{(v_1,v^e_1,b_1,g_1,\iota_1)\gamma_1 } _ { m_1n_1;m_2n_2 ; \sigma m
\,\sigma n;m m_1}[\lambda]\big| \big ) \big(\max_{l_1 } \sum_{m_1}\max_{m , l } \big|q_{m_1m;l l_1}(\lambda)\big|\big ) \nonumber \\ * & \qquad\qquad\qquad \times \big ( \sum_{\mathcal{e}_2^{s_2 } } \big|a^{(v_2,v^e_2,b_2,g_2,\iota_2)\gamma_2 } _ { l_1l;\sigma k \,\sigma l ; k_2l_2;k_1l_1 } [ \lambda]\big|\big ) \nonumber \\ & \qquad \leq \frac{1}{2 } c_1 \big(\frac{\lambda}{\mu}\big)^ { \delta_2(v-\frac{n+2}{2}+4 - 2g-(b+1 ) ) } \big(\frac{\mu}{\lambda}\big)^{\delta_1(1 + v - v^e-\iota + 2g+ ( b+1)-2 + ( s-1 ) ) } \nonumber \\ * & \qquad\qquad \qquad \times \big(\frac{\mu}{\lambda}\big)^{\delta_0 ( v^e+\iota-2-(s-1 ) ) } p^{2v-\frac{n+2}{2}}\big[\ln \frac{\lambda}{\lambda_r}\big]\;. \label{e0}\end{aligned}\ ] ] we have used the induction hypothesis ( [ annorm1 ] ) for the subgraphs as well as ( [ est1 ] ) for the propagator and have inserted @xmath261 , @xmath262 , @xmath263 , @xmath264 , @xmath265 , @xmath266 and @xmath267 , because there is an additional summation over @xmath169 which belongs to @xmath268 but not to @xmath269 . if @xmath270 we take instead the unsummed propagator and replace in ( [ e0 ] ) one factor ( [ est1 ] ) by ( [ est0 ] ) as well as @xmath271 by @xmath161 .
the total exponents of @xmath228 remain unchanged .
next , let there be no unsummed external leg on the contracted segment of @xmath242 viewed from @xmath228 .
now , we can not directly use the induction hypothesis . on the other hand , for a given index configuration of @xmath242 and the propagator ,
the index @xmath272 is not an independent summation index : @xmath273 see also ( [ nk ] ) .
if @xmath274 there must be an unsummed outgoing index on the contracted segment of @xmath239 .
we can thus realise the @xmath272-summation as a summation over @xmath39 in @xmath239 for fixed index configuration of @xmath242 and @xmath275 .
this @xmath39-summation is applied together summation over the @xmath239-indices of @xmath160 to @xmath239 as the first step , taking again the maximum of the propagator over @xmath258 . in the second step we sum over the restriction of @xmath160 to @xmath242 and the propagator .
it is obvious that the estimation ( [ e0 ] ) remains unchanged , in particular , @xmath276 . if @xmath169 is the only unsummed index we realise the @xmath272-summation as a summation of the propagator over @xmath248 .
here , one has to take into account that the subgraph @xmath242 is bounded independently of the incoming index @xmath248 .
again we get the same exponents as in ( [ e0 ] )
. we can summarise ( [ e0 ] ) and its discussed modification to @xmath277 \big| \nonumber \\ * & \leq \big(\frac{\lambda}{\mu}\big)^{\delta_2(v-\frac{n}{2}+2 - 2g - b ) } \big(\frac{\mu}{\lambda}\big)^{\delta_1(v - v^e-\iota+2g+b-1+s ) } \big(\frac{\mu}{\lambda}\big)^{\delta_0(v^e+\iota-1-s ) } \nonumber \\*[-0.5ex ] & \qquad \times p^{2v-\frac{n}{2}-1}\big[\ln \frac{\lambda}{\lambda_r}\big]\;. \label{eall}\end{aligned}\ ] ] for the choice of the boundary conditions according to definition / lemma [ def - lem1 ] , the @xmath9-integration increases ( again according to definition / lemma [ def - lem1 ] ) the degree of the polynomial in @xmath278 by @xmath279 . hence , we have extended ( [ annorm1 ] ) to a bigger degree @xmath124 for contractions of type ( [ aa1 ] ) . in particular , the bound is ( by induction starting with ( [ est1 ] ) , which represents the third graph in ( [ examplesofgraphs ] ) ) independent of the incoming indices @xmath280 .
the verification of ( [ annorm1 ] ) for any contraction between graphs of the first line in ( [ as ] ) is performed in a similar manner .
taking the same subgraphs as in ( [ aa1 ] ) , but with a contraction of other legs , the discussion is in fact a little easier because there are no trajectories going through both subgraphs : @xmath281 the contractions @xmath282 are treated in the same way .
the point is that the summation indices of the propagator ( @xmath283 for the upper graph and @xmath284 for the lower graph in ( [ aa2 ] ) ) are fixed by index conservation for the subgraphs . in the same way
one also discusses any contraction between the second and third graph in ( [ as ] ) .
let us now contract the left graph in the second line of ( [ as ] ) with any graph of the first line of ( [ as ] ) , e.g. @xmath285 the number of boundary components is reduced by @xmath279 , giving @xmath265 .
we clearly have @xmath245 , but there is now one external vertex less on which we can apply an index summation , @xmath286 .
at the same time we need the index summation from the subgraph , because in the @xmath9-scaling @xmath287 \nonumber \\ * & = \frac{1}{2 } \sum_{m , n , k } a^{(v_1,v^e_1,b_1 , g_1 , \iota_1)\gamma_1}_{mn;\sigma m \ , \sigma n}[\lambda ] \ , q_{nm;k_1k}(\lambda)\ , a^{(v_2,v^e_2,b_2,g_2 , \iota_2)\gamma_2 } _
{ kk_1;k_1l_1;\sigma k \ , \sigma l}[\lambda ] \label{aa3l}\end{aligned}\ ] ] there is now one undetermined summation index : @xmath288 first , let there be an additional unsummed external leg on the segment of @xmath289 in @xmath239 .
then , the induction hypothesis ( [ annorm1 ] ) gives the bound for a summation over @xmath39 .
we thus fix @xmath284 and all indices of @xmath242 in the first step and realise a possible @xmath290-summation due to @xmath291 as an @xmath39-summation , which is applied together with the summation over the @xmath239-indices of @xmath160 , after maximising the propagator over @xmath292 .
the result is independent of @xmath195 .
we thus restrict the @xmath195-summation to the propagator , see ( [ est1 ] ) , and apply the remaining @xmath160-summations to @xmath242 , where @xmath196 remains unsummed .
we have @xmath293 and get the estimation @xmath294 \big|\nonumber \\ * & \leq \frac{1}{2 } \big(\sum_{m,\mathcal{e}_1^{s_1 } } \big| a^{(v_1,v^e_1,b_1,g_1 , \iota_1)\gamma_1 } _ { mn;\sigma m \ , \sigma n}[\lambda ] \big|\big ) \big ( \max_{k}\sum_{n } \max_{m , k_1 } \big|q_{nm;k_1k}(\lambda ) \big|\big ) \nonumber \\ * & \qquad\qquad \times \big ( \sum_{\mathcal{e}_2^{s_2}\not\ni k_1 } \big| a^{(v_2,v^e_2,b_2,g_2 , \iota_2)\gamma_2}_{kk_1 ; k_1l_1;\sigma k \ , \sigma l } [ \lambda]\big| \big ) \nonumber \\ & \leq \frac{1}{2 } c_1 \big ( \frac{\lambda}{\mu}\big)^{\delta_2(v-\frac{n+2}{2}+4 -2g-(b+1 ) ) } \big(\frac{\mu}{\lambda}\big)^{\delta_1(1 + v-(v^e+1)-\iota + 2 g + ( b+1)-2+s ) } \nonumber \\ * & \qquad\qquad \times \big(\frac{\mu}{\lambda}\big)^{\delta_0 ( ( v^e+1)+\iota-2-s ) } p^{2v-\frac{n+2}{2}}\big[\ln \frac{\lambda}{\lambda_r}\big]\;. \label{aa3le1}\end{aligned}\ ] ] if @xmath295 we do not need the @xmath39-summation on @xmath239 .
again we have @xmath296 and ( [ aa3le1 ] ) remains unchanged .
here , we may allow for index summations at all other external legs on the segment of @xmath289 in @xmath239 .
if there is no unsummed external leg on the segment of @xmath289 in @xmath239 , we must realise the @xmath290-summation as follows : we proceed as before up to the step where we sum the propagator for given @xmath196 over @xmath195 . for each term in this sum we have @xmath297 .
we thus achieve a different @xmath290 if for given @xmath298 we start from a different @xmath196 .
since the result of the summations over @xmath239 and the propagator is independent of @xmath196 , see ( [ est1 ] ) , we realise the @xmath290-summation as a sum over @xmath196 restricted to @xmath242 .
we now get the same exponents as in ( [ aa3le1 ] ) also for this case . according to definition / lemma [ def - lem1 ]
, the @xmath9-integration extends for contractions of type ( [ aa3 ] ) the bound ( [ annorm1 ] ) to a bigger order @xmath124 .
the contraction of the other leg of the right vertex @xmath299 is easier to discuss because the @xmath290-summation is directly applied to @xmath242 .
taking the second vertex of the first line of ( [ as ] ) instead , we have two contractions which are identical to ( [ aa3 ] ) and ( [ aa3a ] ) and a third one with contractions as in the first and last graphs of ( [ aa1a ] ) where @xmath239 and @xmath242 form different segments in @xmath228 .
this case is much easier because there is no trajectory involving both subgraphs .
moreover , contracting the last instead of the first vertex of the second line of ( [ as ] ) gives the same estimates if the two propagators between the vertex and the oval belong to the same segment : @xmath300 the only modification to ( [ aa3le1 ] ) and its variants is to replace @xmath301 by @xmath124 and @xmath155 by @xmath302 , because the total number of external vertices is unchanged whereas the total segmentation index is reduced by @xmath279 .
if we contract the second vertex of the last line in ( [ as ] ) in such a way that the contracted indices @xmath289 belong to different segments of @xmath239 , e.g.@xmath303 \parbox{75 mm } { \begin{picture}(75,32 ) \put(0,0){\epsfig{scale=.9,file = aa6,bb=71 600 317 684 } } \put(0,6.5){\mbox{\scriptsize$\sigma n$ } } \put(2,13.5){\mbox{\scriptsize$\sigma m$ } } \put(28.5,28){\mbox{\scriptsize$\sigma n'$ } } \put(38,26){\mbox{\scriptsize$\sigma m'$ } } \put(34.5,7){\mbox{\scriptsize$n_1 $ } } \put(36.5,13){\mbox{\scriptsize$m_1 $ } } \put(23.5,12.5){\mbox{\scriptsize$n_3 $ } } \put(15.5,14.5){\mbox{\scriptsize$m_3 $ } } \put(16.5,27){\mbox{\scriptsize$n_2 $ } } \put(23.5,25){\mbox{\scriptsize$m_2 $ } } \put(62.5,1){\mbox{\scriptsize$l_1 $ } } \put(56,3){\mbox{\scriptsize$k_1 $ } } \put(81.5,9){\mbox{\scriptsize$\sigma k$ } } \put(83.5,17.5){\mbox{\scriptsize$\sigma l$ } } \put(47,6.5){\mbox{\scriptsize$m$ } } \put(47,12.5){\mbox{\scriptsize$n$ } } \put(56.5,12.5){\mbox{\scriptsize$k$ } } \end{picture } } \label{aa5}\end{aligned}\ ] ] they are actually determined by index conservation for the segments .
the entire discussion of these examples is therefore similar to the graph ( [ aa1 ] ) with bound ( [ e0 ] ) and its modifications .
note that we have @xmath244 and @xmath245 in ( [ aa5 ] ) .
accordingly , we can replace in all previous examples a vertex of the first line of ( [ as ] ) by the composed vertex under the condition that the two contracted trajectories at the composed vertex belong to different segments .
it remains to study the contraction @xmath304 where two contraction indices ( @xmath39 or @xmath195 and @xmath196 or @xmath248 ) are undetermined .
we have @xmath305 and @xmath245 .
we first assume that at least one of the boundary components of @xmath230 to contract carries more than one external vertex .
in this case we have @xmath243 .
there has to be at least one unsummed external vertex on the segment , say on @xmath242 .
we fix the indices of @xmath242 as well as @xmath195 in the first step , take in the propagator the maximum over @xmath258 and sum over the @xmath239-indices of @xmath160 . here
, @xmath39 can be regarded as an unsummed index .
we take the maximum of @xmath239 over @xmath195 so that the @xmath195-summation restricts to the propagator only .
we take in the summed propagator the maximum over @xmath196 so that the remaining @xmath196-summation is applied together with the summation over the @xmath242-indices of @xmath160 .
we thus need @xmath306 summations and the bound ( [ est1 ] ) for the propagator : @xmath307 \big|\nonumber \\ * & \leq \frac{1}{2 } \big(\max_n \sum_{\mathcal{e}_1^{s_1 } } \big| a^{(v_1,v^e_1,b_1,g_1 , \iota_1)\gamma_1 } _ { mn;\sigma m \ , \sigma n}[\lambda ] \big|\big ) \big ( \max_{k}\sum_{n } \max_{m , l } \big|q_{nm;lk}(\lambda ) \big|\big ) \nonumber \\ * & \qquad\qquad \times \big ( \sum_{k,\mathcal{e}_2^{s_2 } } \big| a^{(v_2,v^e_2,b_2,g_2 , \iota_2)\gamma_2}_{kl;\sigma k \ , \sigma l } [ \lambda]\big| \big ) \nonumber \\ & \leq \frac{1}{2 } c_1 \big ( \frac{\lambda}{\mu}\big)^{\delta_2(v-\frac{n+2}{2}+4 -2g-(b+1 ) ) } \big(\frac{\mu}{\lambda}\big)^{\delta_1(1 + v-(v^e+2)-\iota+2g+(b+1)-2+(s+1 ) ) } \nonumber \\ * & \qquad\qquad \times \big(\frac{\mu}{\lambda}\big)^{\delta_0 ( ( v^e+2)+\iota-2-(s+1 ) ) } p^{2v-\frac{n+2}{2}}\big[\ln \frac{\lambda}{\lambda_r}\big]\;. \label{aa4e}\end{aligned}\ ] ] finally , we have to consider the case where the only external vertices of both boundary components of @xmath230 to contract are just the contracted vertices . in this case
the contraction removes these two boundary components at expense of a completely inner loop , giving @xmath308 .
the differences @xmath309 and @xmath310 are fixed by the remaining indices of @xmath230 . for given @xmath39 we may thus take the maximum of @xmath242 over @xmath248 and realise the @xmath248-summation as a summation ( [ est1 ] ) over the propagator .
we thus exhaust all differences @xmath311 . in order to exhaust all values of @xmath39
we take the maximum of @xmath239 over @xmath289 and multiply the result by a volume factor ( [ volumefactor ] ) .
we thus replace in ( [ aa4e ] ) @xmath312 and @xmath313 , and combine one factor ( [ est0 ] ) and a volume factor ( [ volumefactor ] ) to ( [ est2 ] ) .
we thus get the same total exponents as in ( [ annorm1 ] ) so that the @xmath9-integration extends ( [ annorm1 ] ) to a bigger order @xmath124 for all contractions represented by ( [ aa4 ] ) .
the contractions @xmath314 \parbox{73 mm } { \begin{picture}(73,30 ) \put(0,0.1){\epsfig{scale=.9,file = aa8a , bb=71 608 283 684 } } \put(1,7){\mbox{\scriptsize$\sigma n$ } } \put(1.5,19.5){\mbox{\scriptsize$\sigma m$ } } \put(19.5,10.5){\mbox{\scriptsize$n_1 $ } } \put(21.5,16){\mbox{\scriptsize$m_1 $ } } \put(71,6){\mbox{\scriptsize$\sigma k$ } } \put(72,19.5){\mbox{\scriptsize$\sigma l$ } } \put(50.5,10){\mbox{\scriptsize$k_1 $ } } \put(52.5,16){\mbox{\scriptsize$l_1 $ } } \put(32,10){\mbox{\scriptsize$m$ } } \put(32,16){\mbox{\scriptsize$n$ } } \put(41.5,15.5){\mbox{\scriptsize$k$ } } \put(41.5,10){\mbox{\scriptsize$l$ } } \end{picture } } \label{aa8}\end{aligned}\ ] ] are treated in the same way as ( [ aa4 ] ) , now with the two unknown summation indices taken into account by a reduction of @xmath315 . in particular , there is also the situation where @xmath289 and @xmath316 are the only external legs of their boundary components before the contraction . in this case
the number of boundary components drops by @xmath317 , which requires a volume factor in order to realise the sum over the starting point of the inner loop .
thus , ( [ annorm1 ] ) is proven for any contractions produced by the first ( bilinear ) term on the rhs of ( [ poll2 ] ) .
it remains to verify the scaling formula ( [ annorm1 ] ) for the second term ( the last line ) on the rhs of the polchinski equation ( [ poll2 ] ) , which describes self - contractions of graphs . the graphical data for the subgraph will obtain a subscript 1 , such as the number of external vertices @xmath318 , the segmentation index @xmath319 and the set @xmath320 of summation indices .
we always have @xmath321 and @xmath322 .
we first consider contractions of external lines at the same vertex , for which we have the possibilities shown in ( [ as ] ) .
the very first vertex leads to two different self - contractions : @xmath323 \big| & = \sum_l |q_{m_1 l;ln_2}(\lambda)|\ , \delta_{n_1m_2 } \nonumber \\[-8 mm ] & & & + \sum_l |q_{ln_1;m_2l}(\lambda)|\ , \delta_{n_2m_1 } \ ; , \label{a2planar } \\
\parbox{40mm}{\begin{picture}(30,25 ) \put(0,0){\epsfig{scale=.9,file = a12b , bb=71 626 158 684 } } \put(0,6.5){\mbox{\scriptsize$n_1 $ } } \put(3,12.5){\mbox{\scriptsize$m_1 $ } } \put(15,6.5){\mbox{\scriptsize$m_2 $ } } \put(17,12.5){\mbox{\scriptsize$n_2 $ } } \end{picture } } & & \big|\lambda \frac{\partial}{\partial \lambda } a^{(1,1,2,0,1)}_{m_1n_1;m_2n_2}[\lambda]\big| & = |q_{m_1 n_2;m_2 n_1}(\lambda)|\;. \label{a2nonplanar}\end{aligned}\ ] ] for the planar contraction ( [ a2planar ] ) we estimate the @xmath248-summation by a volume factor so that we obtain ( [ annorm1 ] ) from ( [ est2 ] ) . for the non - planar graph ( [ a2nonplanar ] )
we obtain ( [ annorm1 ] ) for @xmath324 directly from ( [ est0 ] ) . according to ( [ numsum ] )
we can apply one index summation which yields ( [ annorm1 ] ) via ( [ est1 ] ) . for the second graph in the first line of ( [ as ] ) we first investigate the contraction @xmath325 the number of loops of the amputated graph
is increased by @xmath279 , @xmath326 , so that due to ( [ euler ] ) and @xmath327 we get @xmath328 .
the graph ( [ al1 ] ) determines the @xmath9-variation @xmath329 & = -\frac{1}{2 } \sum_{k , l } q_{m_1k;ln_1}(\lambda)\ , a^{(v_1,v^e_1,b_1,g_1,\iota_1 ) \gamma_1}_{m_1n_1;n_1l;\sigma m\ , \sigma n;km_1 } [ \lambda]\;,\end{aligned}\ ] ] with one of the indices @xmath316 being undetermined .
first , let there be at least one further external leg on the same boundary component as @xmath330 . in this case
the number of boundary components is increased by @xmath279 , @xmath331 . if there is an unsummed index on the segment of @xmath316 we can realise the @xmath196-summation in @xmath228 as a summation in @xmath239 after taking in the propagator the maximum over @xmath316 .
we thus have @xmath332 and consequently @xmath333 \big| \nonumber \\ * & \leq \frac{1}{2 } \big(\max_{k , l } \big|q_{m_1k;ln_1}(\lambda)\big|\big ) \big ( \sum_{k,\mathcal{e}_1^{s } } \big|a^{(v , v^e_1,b_1,g_1,\iota_1)\gamma_1}_{m_1n_1 ; n_1l;\sigma m\ , \sigma n;km_1 } [ \lambda]\big ) \nonumber \\ & \leq \frac{1}{2 } c_0 \big(\frac{\lambda}{\mu_0}\big)^{\delta_2 ( v-\frac{n+2}{2}+2 - 2g-(b-1 ) ) } \big(\frac{\mu}{\lambda}\big)^{\delta_1(v - v^e-\iota + 2 g+(b-1)-1+(s+1 ) ) } \nonumber \\ * & \qquad\qquad \times \big(\frac{\mu}{\lambda}\big)^{\delta_0(1+v^e+\iota -1-(s+1 ) ) } p^{2v-\frac{n+2}{2}}\big[\ln \frac{\lambda}{\lambda_r}\big]\;. \label{al1e}\end{aligned}\ ] ] we can sum the contracting propagator over @xmath169 for fixed @xmath334 , which amounts to replace one factor ( [ est0 ] ) by ( [ est1 ] ) compensated by @xmath335 replacing @xmath336 .
if @xmath196 can not be a summation index in @xmath239 then @xmath169 must be unsummed in @xmath228 .
we first apply the summation over @xmath337 $ ] for given @xmath248 in @xmath239 .
the result is independent of @xmath248 so that , for given @xmath196 , the @xmath248-summation can be restricted to the contracting propagator maximised over @xmath338 .
finally , the remaining @xmath268-summations are applied .
we have to replace in ( [ al1e ] ) @xmath339 by @xmath161 and one factor ( [ est0 ] ) by ( [ est1 ] ) .
finally , let there be no further external leg on the same boundary component as @xmath330 .
now the number of boundary components remains constant , @xmath340 . since @xmath341 is a constant , the required summation over e.g. @xmath196 has to be estimated by a volume factor ( [ volumefactor ] ) .
we thus replace in ( [ al1e ] ) @xmath342 and @xmath343 and combine one factor ( [ est0 ] ) and the volume factor to ( [ est2 ] ) . in summary , we extend after @xmath9-integration the scaling law ( [ annorm1 ] ) for the same degree @xmath124 to a reduced number @xmath112 of external lines . next
, we study the following contraction of the second graph in ( [ as ] ) which gives rise to an inner loop : @xmath344 it describes the @xmath9-variation @xmath345 & = -\frac{1}{2 } \big(\sum_{l } q_{n_1l;ln_1}(\lambda)\big)\ , a^{(v_1,v^e_1,b_1,g_1,\iota_1)\gamma_1 } _ { m_1n_1;n_1l;ln_1;\sigma m\,\sigma n } [ \lambda]\;.\end{aligned}\ ] ] the number of loops of the amputated graph is increased by @xmath279 and the number of boundary components remains constant , giving @xmath328 and @xmath340 .
note that @xmath346 is independent of @xmath248 so that the @xmath248-summation acts on the propagator only .
we estimate the @xmath248-summed propagator by ( [ est2 ] ) for the product of ( [ est0 ] ) with a volume factor ( [ volumefactor ] ) . the factor ( [ est2 ] )
compensates the decrease @xmath347 , all other exponents remain unchanged when passing from @xmath239 to @xmath228 .
now the @xmath9-integration extends the scaling law ( [ annorm1 ] ) to a reduced @xmath112 .
the third graph in the first line of ( [ as ] ) leads to the contracted graph @xmath348 there is one additional loop of the amputated graph , giving @xmath328 .
we have @xmath340 if there are further external legs on the boundary component of @xmath195 and @xmath349 if no further external leg exists on the contracted boundary component . very similar to ( [ al2 ] ) , the @xmath248-summation is restricted to the propagator maximised over @xmath195 , giving a factor ( [ est2 ] ) which compensates @xmath347 in the first exponent of ( [ annorm1 ] ) .
for @xmath340 the @xmath195-summation in ( [ al3 ] ) is provided by the subgraph @xmath239 , where the additional summation @xmath332 compared with @xmath228 compensates the change @xmath350 of external vertices in the second and third exponent of ( [ annorm1 ] ) . on the other hand , if @xmath351 we have @xmath335 and the summation over @xmath195 has to come from a volume factor ( [ volumefactor ] ) combined with one factor ( [ est0 ] ) to ( [ est2 ] ) .
this verifies ( [ annorm1 ] ) for the contraction ( [ al3 ] ) . the last case for which contractions of two external lines at the same vertex are to investigate is the last vertex in the second line of ( [ as ] ) . as before in the proof for tree - contractions , we have to distinguish whether the composed vertex under consideration appears inside a tree , in a loop but together with further composed vertices , or in a loop but as the single composed vertex . in the first case we have to analyse the graph @xmath352 before the contraction
, the indices @xmath187 were all located on the same loop of the amputated graph and the same boundary component .
after the contraction they are split into two loops , @xmath328 .
the number of boundary components is increased by @xmath279 if both resulting boundary components of @xmath353 and @xmath354 carry further external legs , @xmath331 .
we have @xmath340 if only one of the resulting boundary components of @xmath353 or @xmath354 carries further external legs and @xmath349 if there are no further external legs on these boundary components .
we clearly have @xmath355 and @xmath356 .
due to index conservation for segments , either @xmath196 or @xmath195 is an unknown summation index , and either @xmath248 or @xmath39 .
we first consider the case @xmath331 . in both segments of @xmath239 to contract there must be at least one unsummed outgoing index , which we can choose to be different from the vertex to contract .
we thus take in the propagator the maximum ( [ est0 ] ) over all indices and restrict the required index summations over @xmath357 to the segments of the subgraphs .
this means that we have @xmath358 summations , which compensates the change of the numbers of boundary components @xmath359 , external legs @xmath322 and external vertices
@xmath350 : @xmath360 \big| \nonumber \\ * & \leq \frac{1}{2 } \big ( \max_{m , n , k , l } \big|q_{nm;lk}(\lambda)\big|\big ) \big(\max_{l , n } \sum_{k , m,\mathcal{e}^{s } } \big|a^{(v_1,v^e_1,b_1,g_1,\iota_1)\gamma_1 } _ { \sigma m\,\sigma n;mn;\sigma m'\ , \sigma ' n;kl } [ \lambda]\big| \big ) \nonumber \\ & \leq \frac{1}{2 } c_0 \big(\frac{\lambda}{\mu}\big)^{\delta_2 ( v-\frac{n+2}{2}+2 - 2g-(b-1 ) ) } \big(\frac{\mu}{\lambda}\big)^{\delta_1 ( v-(v^e+1)-\iota+2g+(b-1)-1+(s+2 ) ) } \nonumber \\ * & \qquad \times \big(\frac{\mu}{\lambda}\big)^{\delta_0 ( 1+(v^e+1)+\iota-1-(s+2 ) ) } p^{2v-\frac{n+2}{2}}\big[\ln \frac{\lambda}{\lambda_r}\big]\;. \label{al4e}\end{aligned}\ ] ] we immediately confirm ( [ annorm1 ] ) . alternatively , instead of consuming a @xmath239-summation to get the @xmath196-summation we can also sum the propagator for maximised @xmath353 and given @xmath196 over @xmath195
. compared with ( [ al4e ] ) we have to replace @xmath361 by @xmath339 and one factor ( [ est0 ] ) by ( [ est1 ] ) , ending up in the same exponents . next , we investigate the case @xmath340 where , for example , the restriction of the boundary component to the left segment does not carry another external leg than @xmath258 . the summation over @xmath39 in @xmath239
is now provided by a volume factor , which means that in ( [ al4e ] ) we have to replace @xmath361 by @xmath339 , @xmath362 by @xmath142 and one factor ( [ est0 ] ) by ( [ est2 ] ) .
all exponents match again ( [ annorm1 ] ) .
finally , let us look at the possibility @xmath349 where the indices @xmath187 to contract were the only external indices of the boundary component .
we thus combine two volume factors ( [ volumefactor ] ) and two factors ( [ est0 ] ) to two factors ( [ est2 ] ) , compensating @xmath363 and @xmath364 .
after @xmath9-integration we extend ( [ annorm1 ] ) to a reduced @xmath112 . the case that the two sides of the composed vertex to contract are connected but belong to different segments , e.g. @xmath365 is similar to treat concerning index summations , but for the interpretation of the genus there is a different situation possible . in the amputated subgraph @xmath239 the indices @xmath289 and
@xmath316 may be situated on _ different _ loops and thus different boundary components .
the contraction joins in this case the two loops , @xmath366 , which results due to ( [ euler ] ) in @xmath367 and @xmath349 .
there is at least one additional external leg on each of the boundary components of @xmath289 and @xmath316 before the contraction , because in order to close the loop we have to pass through the vertex @xmath368 .
now we have to replace in ( [ al4e ] ) @xmath362 by @xmath369 and @xmath133 by @xmath370 , confirming ( [ annorm1 ] ) also in this case .
if all indices @xmath187 are on the same loop in @xmath239 , the contraction splits it into two and the entire discussion of ( [ al4 ] ) can be used without modification to the present example .
it remains the case @xmath371 where the two halves of the composed vertex to contract belong to the same segment .
three of the indices @xmath187 are now summation indices .
we have @xmath372 and @xmath356 .
let first the indices @xmath289 on one hand and @xmath316 on the other hand be situated on different loops of the amputated graph @xmath239 .
these are joint by the contraction , yielding @xmath367 . if there remain further external legs on the contracted loop we have @xmath349 , otherwise @xmath373 .
we start with @xmath349 . due to the segmentation index present in @xmath239 ,
the induction hypothesis for @xmath239 gives us the bound for two additional summations over @xmath283 not present in @xmath228 .
the third summation is provided by the propagator via ( [ est1 ] ) .
assuming @xmath374 \neq l , n$ ] in @xmath239 we first take in the contracting propagator the maximum over @xmath258 , then sum the @xmath289-boundary component over @xmath39 and those indices of @xmath160 which belong to the @xmath289-boundary component , followed by the summation of the propagator over @xmath195 for given @xmath196 .
finally , we sum @xmath239 over the remaining indices of @xmath160 and over @xmath196 : @xmath375 \big| \nonumber \\ * & \leq \frac{1}{2 } \big ( \max_k \sum_{n } \max_{l , m } \big|q_{mn;kl}(\lambda)\big|\big ) \sum_{k , m,\mathcal{e}^{s } } \big|a^{(v_1,v^e_1,b_1,g_1,\iota_1)\gamma_1 } _ { \sigma m\,\sigma n;mn;kl } [ \lambda]\big|\big ) \nonumber \\ & \leq \frac{1}{2 } c_1 \big(\frac{\lambda}{\mu}\big)^{\delta_2 ( v-\frac{n+2}{2}+2- 2(g-1)-(b+1 ) ) } \nonumber \\ * & \qquad\qquad \times \big(\frac{\mu}{\lambda}\big)^{\delta_1(1 + v-(v^e+1)-(\iota+1)+2(g-1)+(b+1)-1+(s+2 ) ) } \nonumber \\ * & \qquad\qquad \times \big(\frac{\mu}{\lambda}\big)^{\delta_0 ( ( v^e+1)+(\iota+1)-1-(s+2 ) ) } p^{2v-\frac{n+2}{2}}\big[\ln \frac{\lambda}{\lambda_r}\big]\;. \label{al5e}\end{aligned}\ ] ] it is essential that the summation over @xmath376 $ ] is independent .
if there are no external legs on the contracted loop , @xmath373 , then we have in @xmath239 either @xmath377=n,\;\mathfrak{i}[k]=l$ ] or @xmath377=l,\;\mathfrak{i}[k]=n$ ] . in the first case we would first fix @xmath284 and maximise the propagator over @xmath258 .
now the @xmath39-summation restricts to @xmath239 with bound independent of @xmath195 .
thus , the @xmath195-summation for given @xmath196 restricts to the propagator and delivers a factor ( [ est1 ] ) , independent of @xmath196 .
however , since @xmath376\equiv k - l= n - m$ ] is already exhausted in @xmath239 , the remaining @xmath196-summation has to come from a volume factor .
we thus make in ( [ al5e ] ) the replacements @xmath378 , @xmath379 and combine one factor ( [ est0 ] ) with a volume factor ( [ volumefactor ] ) to ( [ est2 ] ) .
the exponents match again ( [ annorm1 ] ) .
next , we investigate the situation where all indices @xmath187 are located on the same loop of the amputated subgraph @xmath239 . in this case
the contraction to @xmath228 splits that loop into two so that we have @xmath328 . as before we have @xmath331 if both split loops contain further external legs , @xmath340 if only one of the split loops contains further external legs , and @xmath349 if the split loops do not contain further external legs .
the discussion is similar as for ( [ al4 ] ) , the difference is that three of @xmath187 are now summations indices , which is taken into account by the replacement of @xmath155 in ( [ al4e ] ) by @xmath302 .
we thus finish the verification of ( [ annorm1 ] ) for self - contractions of a vertex .
it remains to check ( [ annorm1 ] ) for contractions of different vertices of the same graph .
the external lines of the two vertices are arranged according to ( [ as ] ) .
we start with two vertices of the type shown as the second graph in ( [ as ] ) . one possible contraction of their external lines is @xmath380 assuming that the vertices to contract are located on the same segment in @xmath239 .
one of the indices @xmath258 is a summation index .
we first consider the case that the two vertices to contract are located on the same loop of the amputated graph @xmath239 .
the contraction to @xmath228 splits that loop into two , giving @xmath328 .
we have @xmath331 if the trajectory starting at @xmath248 does not leave @xmath239 ( and @xmath228 ) in @xmath39 , whereas @xmath340 if @xmath258 are on the same trajectory in @xmath239 .
in case of @xmath331 we keep @xmath377 $ ] in @xmath239 fixed , take in the propagator the maximum over @xmath258 and restrict the @xmath39-summation to @xmath239 . due to @xmath381 , @xmath382 and @xmath359 we have in the case that @xmath169 remains unsummed @xmath383\big| \nonumber \\ * & \leq \frac{1}{2 } \big ( \max_{m , l , m_1,l_1 } \big|q_{m_1m;ll_1 } ( \lambda)\big|\big)\ , \big ( \sum_{m,\mathcal{e}^{s } } \big| a^{(v_1,v^e_1,b_1,g_1,\iota_1)\gamma_1 } _ { k_2l_2;k_1l_1;l_1 l;m m_1;m_1n_1;m_2n_2 ; \sigma m\ , \sigma n}[\lambda ] \big|\big ) \nonumber \\ * & \leq \frac{1}{2 } c_0 \big(\frac{\lambda}{\mu}\big)^{\delta_2 ( v-\frac{n+2}{2}+2 - 2g-(b-1 ) ) } \big(\frac{\mu}{\lambda}\big)^{\delta_1 ( v - v^e-\iota+2g+(b-1)+1+(s+1 ) ) } \nonumber \\ * & \qquad\qquad \times \big(\frac{\mu}{\lambda}\big)^{\delta_0 ( 1+v^e+\iota-1-(s+1 ) ) } p^{2v-\frac{n+2}{2}}\big[\ln \frac{\lambda}{\lambda_r}\big]\;. \label{ab1e}\end{aligned}\ ] ] summing additionally over @xmath169 we replace in ( [ ab1e ] ) one factor ( [ est0 ] ) by ( [ est1 ] ) .
it is clear that this reproduces the exponents of ( [ annorm1 ] ) correctly .
if @xmath384 $ ] in @xmath239 , we have to realise the @xmath39-summation by a volume factor .
we thus replace in ( [ ab1e ] ) @xmath342 , @xmath312 and combine ( [ volumefactor ] ) with one factor ( [ est0 ] ) to ( [ est2 ] ) .
finally , the two vertices to contract in ( [ ab1 ] ) may be located on different loops of the amputated graph @xmath239 .
they are joint by the contraction to @xmath228 , giving @xmath367 , and because the newly created loop obviously has external legs , we have @xmath349 . as separated loops in @xmath239
, @xmath248 can not be the incoming index of the trajectory through @xmath39 .
therefore , the @xmath39-summation gives the same bound as the rhs of ( [ ab1e ] ) , now with @xmath362 replaced by @xmath369 and @xmath133 by @xmath370 .
we have thus extended ( [ annorm1 ] ) to a reduced @xmath112 for all types of contractions ( [ ab1 ] ) . if the vertices to contract are located on different segments in @xmath239 , e.g. @xmath385 both indices @xmath258 are determined by index conservation for segments .
we can thus save an index summation compared with ( [ ab1e ] ) and replace there and in its discussed modifications @xmath339 by @xmath161 and @xmath382 by @xmath386 .
since the @xmath39-summation is not required , there is effectively an additional summation possible in agreement with ( [ numsum ] ) .
it is not possible that @xmath39 and @xmath248 are located on the same trajectory in @xmath239 so that either @xmath387 or @xmath388 .
let us make a few more comments on the segmentation index .
it is essential that the contraction joins separated segments .
for instance , the contraction @xmath389 does not increase the segmentation index , because in agreement with definition [ defiota ] the number of segments remains constant .
the graph on the left has @xmath390 , and the internal indices @xmath258 are determined by the external ones .
the graph on the right has @xmath390 as well , and now one of the indices @xmath284 becomes a summation index .
having several composed vertices in the middle link does not change the segmentation index : @xmath391 it makes , however , a difference if the two composed vertices are situated on different links : @xmath392 here , the segmentation index increases from @xmath390 on the left to @xmath393 on the right , in agreement with definition [ defiota ] .
the case @xmath394 is completely identical to ( [ ab1 ] ) . in the contraction
@xmath395 the summation index @xmath195 or @xmath248 is provided by the propagator , replacing in ( [ ab1e ] ) and its modifications @xmath339 by @xmath161 and one factor ( [ est0 ] ) by ( [ est1 ] ) .
it is not possible that @xmath195 and @xmath248 are located on the same trajectory in @xmath239 so that either @xmath387 or @xmath388 . in order to treat the contraction @xmath396 one has to use that the summation over @xmath169 can due to @xmath397 be transferred as a @xmath196-summation of @xmath239 .
the summation over the undetermined index @xmath39 is applied in the last step . finally , @xmath398 are similar to the @xmath155-increased variant ( [ ab2 ] ) .
the contraction @xmath399 is an example for a realisation of ( [ iota ] ) where @xmath400 is increased by @xmath317 and @xmath154 by @xmath279 , giving again a segmentation index increased by @xmath279 .
it is obvious that the discussion of contractions involving the second and third or two of the third vertices of the first line in ( [ as ] ) is analogous .
let us now study loop contractions which involve the first graph in the second line of ( [ as ] ) , assuming first that the vertices are situated on the same segment : @xmath401 we thus have @xmath355 and @xmath356 .
two of the summation indices @xmath402 are undetermined .
let first the two vertices to contract be located on the same loop of the amputated subgraph @xmath239 .
the contraction splits that loop into two , giving @xmath328 .
next question concerns the number of boundary components .
we have @xmath331 if there are further external legs on the loop through @xmath353 and @xmath340 if @xmath384 $ ] in @xmath239 .
we start with @xmath331 . in general
, the induction hypothesis provides us with bounds for summations over @xmath39 and @xmath196 , because @xmath403 $ ] . if @xmath169 is an unsummed index we thus have @xmath404 \big|\nonumber \\ * & \leq \frac{1}{2 } \big ( \max_{m , l , k , m_1 } \big|q_{m_1m;lk } ( \lambda)\big|\big ) \big ( \sum_{m , k,\mathcal{e}^{s } } \big| a^{(v_1,v^e_1,b_1,g_1,\iota_1)\gamma_1 } _ { kl;nm_1;m_1n_1 ; \sigma m\ , \sigma n}[\lambda ]
\big|\big ) \nonumber \\ * & \leq \frac{1}{2 } c_0 \big(\frac{\lambda}{\mu}\big)^{\delta_2 ( v-\frac{n+2}{2}+2 - 2g-(b-1 ) ) } \big(\frac{\mu}{\lambda}\big)^{\delta_1 ( v-(v^e+1)-\iota + 2 g+(b-1)+1+(s+2 ) ) } \nonumber \\ * & \qquad\qquad \times \big(\frac{\mu}{\lambda}\big)^{\delta_0 ( 1+(v^e+1)+\iota-1-(s+2 ) ) } p^{2v-\frac{n+2}{2}}\big[\ln \frac{\lambda}{\lambda_r}\big]\;. \label{ab5e}\end{aligned}\ ] ] now an additional summation over @xmath169 can immediately be taken into account by replacing the maximised propagator ( [ est0 ] ) by the summed propagator ( [ est1 ] ) , in agreement with @xmath361 replaced by @xmath339 .
the @xmath169-summation is applied before the @xmath196-summation is carried out .
these considerations require an unsummed outgoing index on the contracted segment of @xmath239 .
if this is not the case then @xmath169 has to be the unsummed outgoing index .
now the @xmath248-summation for given @xmath39 has to be restricted to the propagator and delivers a factor ( [ est1 ] ) .
the exponents match again ( [ annorm1 ] ) .
next , for @xmath405 $ ] in @xmath239 we can not use a summation over @xmath39 in @xmath239 in order to account for the undetermined contraction index , because the incoming index @xmath248 would change simultaneously .
instead we have to use a volume factor ( [ volumefactor ] ) combined with one factor ( [ est0 ] ) to ( [ est2 ] )
. additionally we have to replace in ( [ ab5e ] ) @xmath339 by @xmath161 and @xmath362 by @xmath142 .
second , the two vertices to contract may be located on different loops of the amputated graph @xmath239 .
they are joint by the contraction , giving @xmath406 . because the loop carries at least the external leg @xmath407
, we necessarily have @xmath349 .
now , @xmath408 $ ] in @xmath239 so that we use summations over @xmath283 in @xmath239 , giving the same balance ( [ ab5e ] ) for the exponents , with @xmath363 and @xmath409 .
the discussion is identical for the contraction @xmath410 which in the case that @xmath316 belong to the same segment in @xmath228 has two undetermined summation indices as well .
we thus proceed as in ( [ ab5e ] ) and its discussed modification and only have to replace @xmath301 by @xmath151 and @xmath155 by @xmath302 .
if @xmath316 are situated on different segments in @xmath228 , e.g. @xmath411 there is only one undetermined summation index , which is reflected in the analogue of ( [ ab5e ] ) by the fact that the segmentation index remains unchanged , @xmath355 .
note that in the right graph of ( [ ab6b ] ) we either have @xmath412 or @xmath413 .
of course we get the same estimations if the segment of @xmath239 with external lines @xmath414 are connected by several composed vertices to the part of @xmath239 with external lines @xmath298 . the contractions @xmath415 are a little easier because the contracting propagator does not have outgoing indices which for certain summations had to be transferred to the subgraph @xmath239 . if @xmath416 $ ] in @xmath239 , the @xmath196-summation for given @xmath417 can be restricted to @xmath239 after maximising the propagator over all indices .
since the result for @xmath239 is independent of the starting point @xmath195 , the @xmath196-summation can be regarded as a summation over all differences @xmath418 .
the final summation over all pairs @xmath354 with fixed difference @xmath418 is provided by a volume factor ( [ volumefactor ] ) combined with ( [ est0 ] ) to ( [ est2 ] ) .
the balance of exponents is identical to ( [ ab5e ] ) and its discussed variants .
it is clear that the analogue of ( [ ab6b ] ) with the left vertex connected as in ( [ ab5b ] ) is similar to treat .
next , we discuss the variant of ( [ ab5 ] ) where the two vertices to contract belong to different segments in the subgraph @xmath239 : @xmath419 now only one of the indices @xmath402 is an undetermined summation index , with @xmath196 being the most natural choice .
we therefore get a bound for the @xmath9-scaling analogous to ( [ ab5e ] ) but with @xmath155 replaced by @xmath420 , reflecting the increase of the segmentation index @xmath421 .
there is now an additional index summation possible , here via ( [ est1 ] ) over the index @xmath169 .
note that we have either @xmath412 or @xmath413 .
the discussion of the variants of ( [ ab5a ] ) with the right vertex taken as the second one in the last line of ( [ as ] ) and/or the left vertex arranged as in ( [ ab5b ] ) is very similar .
it remains to investigate contractions between two of the vertices in the second line of ( [ as ] ) .
we discuss in detail the contraction @xmath422 all variants are similar as described between ( [ ab5 ] ) and ( [ ab5a ] ) .
three of the four summation indices @xmath187 in ( [ ab8 ] ) are undetermined .
we clearly have @xmath423 and @xmath355 .
we first consider the case where the four indices @xmath187 are located on the same loop of the amputated subgraph @xmath239 .
the contraction will split that loop into two , giving @xmath328 .
there are three possibilities for the change of the number of boundary components after the contraction .
first , if on both paths of trajectories in @xmath239 from @xmath195 to @xmath196 and from @xmath248 to @xmath39 there are further external legs , we have @xmath331 .
second , if on one of these paths there is no further external leg , we have @xmath340 .
third , if both paths contain no further external legs , i.e. @xmath39 and @xmath196 are the outgoing indices of the trajectories starting at @xmath248 and @xmath195 , respectively , we have @xmath349 .
we start with @xmath331 . then , @xmath374 $ ] and @xmath377 $ ] are fixed as external indices so that the induction hypothesis for @xmath239 provides the bounds for two summations over @xmath357 .
we first apply a possible summation to the outgoing index of the trajectory starting at @xmath248 .
the result is maximised independently from @xmath248 so that we can restrict the @xmath248-summation to the propagator , maximised over @xmath354 with @xmath39 being fixed . finally , we apply the summations over @xmath357 and all remaining @xmath160-summations to @xmath239 .
we thus obtain @xmath424 \big|\nonumber \\ * & \leq \frac{1}{2 } \big ( \max_m \sum_l \max_{n , k } \big|q_{nm;lk } ( \lambda)\big|\big ) \big ( \sum_{m , k,\mathcal{e}^{s } } \big| a^{(v_1,v^e_1,b_1,g_1,\iota_1)\gamma_1}_{mn ; \sigma m\ , \sigma n;kl}[\lambda ] \big|\big ) \nonumber \\ * & \leq \frac{1}{2 } c_1 \big(\frac{\lambda}{\mu}\big)^{\delta_2 ( v-\frac{n+2}{2}+2- 2g-(b-1 ) ) } \big(\frac{\mu}{\lambda}\big)^{\delta_1(1 + v-(v^e+2)+\iota+1 + 2g+(b-1)-1+(s+2 ) ) } \nonumber \\ * & \qquad\qquad \times \big(\frac{\mu}{\lambda}\big)^{\delta_0 ( ( v^e+2)+\iota-1-(s+2 ) ) } p^{2v-\frac{n+2}{2}}\big[\ln\frac{\lambda}{\lambda_r}\big]\;. \label{ab8e}\end{aligned}\ ] ] the @xmath9-integration verifies ( [ annorm1 ] ) in the topological situation under consideration .
next , we discuss the case @xmath340 , assuming e.g. @xmath384 $ ] in @xmath239 .
we maximise the propagator over @xmath354 for given @xmath248 so that the @xmath39-summation can be restricted to @xmath239 .
next , we apply the @xmath160-summations and the @xmath196-summation to @xmath239 , still for given @xmath248 . the final @xmath248-summation counts the number of graphs with different @xmath248 , giving the bound ( [ est0 ] ) of the propagator times a volume factor . in any case
the required modifications of ( [ ab8e ] ) , in particular @xmath342 , lead to the correct exponents of ( [ annorm1 ] ) .
if @xmath349 , i.e. @xmath384 $ ] and @xmath416 $ ] , we take in the propagator the maximum over @xmath284 so that for given @xmath248 the @xmath39-summation can be restricted to @xmath239 .
the result of that summation is bounded independently of @xmath248 .
thus , each summand only fixes @xmath425 , and the remaining freedom for the summation indices is exhausted by two volume factors and the bound ( [ est0 ] ) for the propagator .
we thus replace in ( [ ab8e ] ) @xmath426 , @xmath427 and one factor ( [ est1 ] ) by ( [ est0 ] )
. then two factors ( [ est0 ] ) are merged with two volume factors ( [ volumefactor ] ) to give two factors ( [ est2 ] ) .
finally , we have to consider the case where @xmath289 are located on a different loop of the amputated subgraph @xmath239 than @xmath316 .
the contraction joins these loops , giving @xmath367 . if the resulting loop carries at least one external leg we have @xmath349 , whereas we get @xmath373 if the resulting loop does not carry any external legs .
we first consider the case that there is a further external leg on the @xmath428-loop in @xmath239 .
we take in the propagator the maximum over @xmath354 and sum the subgraph for given @xmath429 over @xmath196 and possibly the outgoing index of the @xmath195-trajectory .
the result is independent of @xmath429 .
next , we sum the propagator for given @xmath39 over @xmath248 and finally apply the remaining @xmath160-summations and the summation over @xmath39 to @xmath239 .
we get the same estimates as in ( [ ab8e ] ) with @xmath362 replaced by @xmath369 and @xmath133 by @xmath370 .
if there are no further external legs on the contracted loop we would maximise the propagator over @xmath354 , then sum @xmath239 over @xmath196 for given @xmath248 , next sum the propagator over @xmath248 for given @xmath39 . for each resulting pair
@xmath316 the remaining @xmath39-summation leaves @xmath430 constant .
we thus have to use a volume factor in order to exhaust the freedom of @xmath430 , combining one factor ( [ est0 ] ) and the volume factor ( [ volumefactor ] ) to ( [ est2 ] ) .
we thus confirm ( [ annorm1 ] ) for any contraction of the form ( [ ab8 ] ) .
it is obvious that all examples not discussed in detail are treated in the same manner .
we conclude that ( [ annorm1 ] ) provides the correct bounds for the interaction coefficients of @xmath1-matrix model with cut - off propagator described by the three exponents @xmath203 .
by solving the polchinski equation perturbatively we have derived a power - counting theorem for non - local matrix models with arbitrary propagator .
our main motivation for the renormalisation group investigation of non - local matrix models was to tackle the renormalisation problem of field theories on noncommutative @xmath0 from a different perspective .
the momentum integrals leading to the parametric integral representation are not absolutely convergent ; nevertheless one exchanges the order of integration . in momentum space
one can therefore not exclude the possibility that the uv / ir - mixing is due to the mathematically questionable exchange of the order of integration .
the renormalisation group approach to noncommutative field theories in matrix formulation avoids these problems .
we work with cut - off propagators leading to finite sums and take absolute values of the interaction coefficients throughout .
oscillating phases never appear ; they are not required for convergence of certain graphs .
our power - counting theorem provides a necessary condition for renormalisability : the two scaling exponents @xmath193 of the cut - off propagator have to be large enough relative to the dimension of the underlying space . in @xcite we determine these exponents for @xmath1-theory on noncommutative @xmath0 , @xmath15 : the propagator for the real scalar field on noncommutative @xmath0 , @xmath15 , is characterised by the scaling exponents @xmath432 and @xmath433 . adding a harmonic oscillator potential to the action one achieves @xmath202 .
the weak decay @xmath434 of the propagator leads to divergences in @xmath435 of arbitrarily high degree .
the appearance of unbounded degrees of divergences in field theories on noncommutative @xmath18 is often related to the so - called uv / ir - mixing @xcite .
we learn from the power - counting theorem ( theorem [ thm1 ] ) that similar effects will show up in any matrix model in which the propagator decays too slowly with @xmath9 .
this means that the correlation between distant modes is too strong , i.e. the model is too non - local .
we are grateful to jos gracia - bonda and edwin langmann for discussions concerning the integral representation of the @xmath3-product and its matrix base .
we would like to thank thomas krajewski for advertising the polchinski equation to us and volkmar putz for accompanying the study of polchinski s original proof .
we are grateful to christoph for indicating to us a way to reduce in our original power - counting estimation the polynomial in @xmath436 to a polynomial in @xmath437 , thus permitting immediately the limit @xmath17 .
we would also like to thank manfred schweda and his group for enjoyable collaboration .
we are indebted to the erwin schrdinger institute in vienna , the max - planck - institute for mathematics in the sciences in leipzig and the institute for theoretical physics of the university of vienna for the generous support of our collaboration .
s. minwalla , m. van raamsdonk and n. seiberg , `` noncommutative perturbative dynamics , '' jhep * 0002 * ( 2000 ) 020 [ arxiv : hep - th/9912072 ] . i. chepelev and r. roiban , `` renormalization of quantum field theories on noncommutative @xmath438 .
i : scalars , '' jhep * 0005 * ( 2000 ) 037 [ arxiv : hep - th/9911098 ] . i. chepelev and r. roiban , `` convergence theorem for non - commutative feynman graphs and renormalization , '' jhep * 0103 * ( 2001 ) 001 [ arxiv : hep - th/0008090 ] . c. becchi , s. giusto and c. imbimbo , `` the wilson - polchinski renormalization group equation in the planar limit , '' nucl .
b * 633 * ( 2002 ) 250 [ arxiv : hep - th/0202155 ] .
k. g. wilson and j. b. kogut , `` the renormalization group and the epsilon expansion , '' phys .
* 12 * ( 1974 ) 75 .
j. polchinski , `` renormalization and effective lagrangians , '' nucl .
b * 231 * ( 1984 ) 269 .
l. griguolo and m. pietroni , `` wilsonian renormalization group and the non - commutative ir / uv connection , '' jhep * 0105 * ( 2001 ) 032 [ arxiv : hep - th/0104217 ] . c. becchi , s. giusto and c. imbimbo , `` the renormalization of non - commutative field theories in the limit of large non - commutativity , '' nucl .
b * 664 * ( 2003 ) 371 [ arxiv : hep - th/0304159 ] .
e. nicholson , `` perturbative wilsonian formalism for noncommutative gauge theories in the matrix representation , '' arxiv : hep - th/0305044 . j. m. gracia - bonda and j. c. vrilly , `` algebras of distributions suitable for phase space quantum mechanics . 1 , '' j. math .
* 29 * ( 1988 ) 869 .
g. keller , c. kopper and m. salmhofer , `` perturbative renormalization and effective lagrangians in @xmath439 , '' helv .
acta * 65 * ( 1992 ) 32 .
w. zimmermann , `` reduction in the number of coupling parameters , '' commun .
* 97 * ( 1985 ) 211 .
h. grosse and r. wulkenhaar , ` renormalisation of @xmath1-theory on noncommutative @xmath16 in the matrix base,' jhep * 0312 * ( 2003 ) 019 [ arxiv : hep - th/0307017 ] .
h. grosse and r. wulkenhaar , `` renormalisation of @xmath1-theory on noncommutative @xmath18 in the matrix base , '' arxiv : hep - th/0401128 , to appear in commun .
e. langmann and r. j. szabo , `` duality in scalar field theory on noncommutative phase spaces , '' phys .
b * 533 * ( 2002 ) 168 [ arxiv : hep - th/0202039 ] .
r. dijkgraaf , `` intersection theory , integrable hierarchies and topological field theory , '' arxiv : hep - th/9201003 .
di francesco , p. ginsparg and j. zinn - justin , `` 2-d gravity and random matrices , '' phys .
* 254 * ( 1995 ) 1 [ arxiv : hep - th/9306153 ] .
e. brezin and j. zinn - justin , `` renormalization group approach to matrix models , '' phys .
b * 288 * ( 1992 ) 54 [ arxiv : hep - th/9206035 ] . | solving the exact renormalisation group equation la wilson - polchinski perturbatively , we derive a power - counting theorem for general matrix models with arbitrarily non - local propagators .
the power - counting degree is determined by two scaling dimensions of the cut - off propagator and various topological data of ribbon graphs . as a necessary condition for the renormalisability of a model , the two scaling dimensions have to be large enough relative to the dimension of the underlying space . in
order to have a renormalisable model one needs additional locality properties typically arising from orthogonal polynomials which relate the relevant and marginal interaction coefficients to a finite number of base couplings .
the main application of our power - counting theorem is the renormalisation of field theories on noncommutative @xmath0 in matrix formulation . |
in recent years , the advent of spintronics has motivated the search for methods of generating spin currents with little or no dissipation .
one class of proposals involves designing an adiabatic pump in which the cyclic variation of some control parameters results in the transfer of spin across an otherwise insulating structure@xcite .
such a spin pump has been realized in quantum dot structures@xcite .
a second class of proposals involves using the spin hall effect to generate a spin current using an electric field @xcite .
interest in this approach has been stimulated by the experimental observation of spin accumulation induced by the spin hall effect in doped gaas structures@xcite . in these experiments
the spin current is accompanied by a dissipative charge current .
this motivated murakami , nagaosa and zhang@xcite to propose an interesting class spin hall insulator " materials which are band insulators that have , according to a kubo formula , a large spin hall conductivity .
however , the spin current which flows in the bulk of these materials is not a transport current , and can not be simply measured or extracted . a crucial ingredient for the generation of transport currents
is the existence of gapless extended edge states .
such states are generically not present in the spin hall insulators@xcite .
motivated by the spin hall insulator proposal , we introduced a model of graphene in which the symmetry allowed spin orbit interactions lead to a quantum spin hall effect@xcite .
a related phase has been proposed for @xmath1 in the presence of a uniform strain gradient@xcite .
this phase is characterized by a bulk excitation gap and gapless edge excitations . in the special case where the spin @xmath2 is conserved
, this phase can be viewed as two copies of the quantum hall state introduced by haldane@xcite .
the phase persists , however , in the presence of spin nonconserving interactions as well as disorder@xcite .
time reversal symmetry protects the gapless edge states when electron interactions are weak , though strong interactions can open an energy gap at the edge accompanied by time reversal symmetry breaking@xcite .
we argued that the quantum spin hall phase is distinguished from a band insulator by a @xmath0 topological index@xcite , which is a property of the bulk system defined on a torus .
we suggested a formula for this index in terms of the bloch wavefunctions .
however , the physical meaning of this formula and its relation to the edge states was not explicit . when placed on a cylinder ( or equivalently a corbino disk ) , the quantum spin hall system defines a kind of adiabatic pump as a function of the magnetic flux threading the cylinder . in the case where @xmath2 is conserved , advancing the flux by one flux quantum results in the transfer of spin @xmath3 from one end of the cylinder to the other
this is a spin pump , whose operation is analogous to a charge pump which could be constructed with a quantum hall state on a cylinder . as envisioned by thouless and co workers in the 1980s@xcite , the adiabatic charge pumping process
is characterized by a topological invariant - the chern number - which is an integer that determines the quantized charge that is pumped in the course of a cycle .
equivalently , the chern number provides a topological classification of the two dimensional quantum hall state@xcite .
when @xmath2 is conserved , similar ideas can be used to describe a quantized adiabatic spin pump@xcite .
a local conservation law is essential for this type of topological pumping process . for a finite system with closed ends , the eigenstates before and after a complete cycle must be distinct .
this means that two energy levels must cross in the course of the cycle . in the case of the charge pump ,
that level crossing is protected by local charge conservation because the two states differ in their eigenvalue of the charge at each end . in the absence of a conservation law there
will , in general , be no level crossings , and the system will be in the same state before and after the cycle . unlike charge , spin does not obey a fundamental conservation law , so unless spin non conserving processes can be made very small it is not possible to define an adiabatic spin pump that works analogously to the thouless charge pump . nonetheless , in ref .
we argued that _ time reversal symmetry _ introduces a conservation law which allows for a topological pumping process .
specifically , we showed that for a quantum spin hall state on a cylinder the eigenstates before and after adiabatic flux insertion are orthogonal , and can not be connected by any local time reversal invariant operator .
when a second flux is added , however , the system returns to its original ground state . in this sense , the quantum spin hall effect defines a @xmath0 pump " . in one cycle
there is no charge transferred between the two ends .
since spin is not conserved , the two states can not be distinguished by a spin quantum number ( though the expectation value of the spin at the end changes by a non quantized amount ) .
the question therefore arises : what is it that is pumped between the ends of the cylinder ? in this paper we examine this issue carefully and introduce a class of one dimensional models that exhibit a similar pumping behavior that is protected by time reversal symmetry .
@xmath0 spin pump " is analogous to the quantum spin hall effect in the same sense that the charge pump is analogous to the ordinary quantum hall effect .
we introduce the concept of the _ time reversal polarization _ , a @xmath0 quantity which signals whether or not a time reversal invariant one dimensional system has a kramers degeneracy associated with its ends .
we show that the change in the time reversal polarization in the course of an adiabatic cycle is related to a @xmath0 topological invariant which distinguishes a @xmath0 spin pump from a trivial cycle of an insulator which pumps nothing .
this @xmath0 invariant is equivalent to the invariant introduced in ref . to characterize the quantum spin hall effect .
the present work , however , provides an alternative formulation which clarifies both the physical and mathematical meaning of the invariant .
we study a family of one dimensional hamiltonians which have a bulk energy gap and a length which is much larger than the exponential attenuation length associated with that gap .
we suppose the hamiltonian depends continuously on a pumping parameter " @xmath4 , satisfying the following properties : @xmath5 = h[t ] , \label{cond1}\ ] ] @xmath6 = \theta h[t ] \theta^{-1 } , \label{cond2}\ ] ] where @xmath7 is the time reversal operator . in the case that the one dimensional system corresponds to a two dimensional system on a cylinder @xmath8 may be viewed as the magnetic flux threading the cylinder in units of the flux quantum . in the course of the cycle time reversal symmetry
is broken .
however , the second constraint ensures that the system passes through _ two _ distinct points @xmath9 and @xmath10 at which the hamiltonian is time reversal invariant . condition ( [ cond2 ] ) may be relaxed somewhat , but it is essential that there exist two distinct time reversal invariant points @xmath11 and @xmath12 where ( [ cond2 ] ) is locally valid .
the existence of two such points plays a crucial role in the topological classification of the pumping cycle .
in particular , we will show that pumping cycles in which @xmath13 $ ] and @xmath14 $ ] have different time reversal polarization are topologically distinct from trivial cycles .
we will begin in section ii by introducing a simple one dimensional tight binding model which exemplifies the @xmath0 spin pump .
this model is closely related to a model of a spin pump that was recently introduced by shindou@xcite which may be applicable to certain spin 1/2 quantum spin chains , such as cu - benzonate and yb@xmath15as@xmath16 .
this tight binding model incorporates spin non conserving spin orbit interactions and provides a concrete illustration of the @xmath0 pumping effect .
in section iii we provide a general formulation of the time reversal polarization for non interacting electrons .
our discussion closely parallels the theory of charge polarization@xcite , in which the charge polarization is related to the berry s phase of bloch wavefunctions .
we show how the change in the time reversal polarization defines a @xmath0 topological invariant characterizing the pumping cycle . in section
iv we argue that the notion of time reversal polarization and the topological classification that follows from it can be generalized to interacting systems .
we describe an interacting version of our 1d model using abelian bosonization .
this provides a different formulation of the time reversal polarization , which is well defined in the presence of interactions . in section
v we conclude by addressing two issues . in va
we discuss the implications of the time reversal polarization for the quantum spin hall effect .
we argue that the two dimensional quantum spin hall phase is a distinct phase from a band insulator even in the presence of electron electron interactions .
we then prove that this phase either has gapless edge excitations or exhibits a ground state degeneracy associated with time reversal symmetry breaking at the edge .
we also comment on a proposal by sheng , weng , sheng and haldane@xcite to classify the quantum spin hall effect in terms of a chern number matrix .
in vb we ask whether the @xmath0 spin pump we have defined can actually pump spin .
despite the fact that the isolated pump returns to its original state after two cycles , we argue that when connected weakly to leads , the @xmath0 pump _ does _ pump spin , although the amount of spin pumped in each cycle is not quantized . in the appendix we relate different mathematical formulations of the @xmath0 topological invariant .
we begin by showing that , like the chern invariant , the @xmath0 invariant can be interpreted as an obstruction to globally defining wavefunctions , _ provided _ a constraint relating time reversed wavefunctions is enforced .
we then prove that the invariant derived in this paper is equivalent to the one introduced in ref . .
in this section we introduce a one dimensional tight binding model of the @xmath0 spin pump .
this model is closely related to a model introduced by shindou as an adiabatic spin pump@xcite .
shindou considered an antiferromagnetic spin 1/2 heisenberg chain to which two perturbations which open a gap in the excitation spectrum are added .
the first term is a staggered magnetic field @xmath17 which locks the spins into a neel ordered state .
the second is a staggered component to the exchange interaction @xmath18 , which leads to a dimerized state .
interestingly , shindou suggested that this model may be relevant to certain @xmath19 quantum spin chains , such as cu - benzonate and yb@xmath15as@xmath16 , in which spins reside at two crystallographically inequivalent sites .
he argued that in these systems @xmath17 can be controlled by applying a _ uniform _ magnetic field , and @xmath18 can be controlled with a uniform electric field .
shindou showed that a cycle in which @xmath18 and @xmath17 are adiabatically varied defines a topological spin pump , which transfers @xmath20 in each cycle .
the topological quantization of shindou s pump requires the conservation of @xmath2 . in general , however , @xmath2 non conserving processes are allowed by symmetry .
in particular the dzyaloshinskii - moriya interaction , @xmath21 ) , is allowed , and will inevitably lead to the violation of @xmath2 conservation .
we will argue , however , that provided the system retains time reversal invariance when @xmath22 this system remains a @xmath0 spin pump even in the presence of the dzyaloshinskii - moriya interaction .
we study a non interacting electron version of the shindou model , where in addition to the spin degree of freedom we allow charge fluctuations .
consider a one dimensional tight binding model with a staggered magnetic field , a staggered bond modulation as well as a time reversal invariant spin orbit interaction , @xmath23 where @xmath24 @xmath25 @xmath26 and @xmath27 here @xmath28 is an arbitrary vector characterizing the spin orbit interaction .
this term explicitly violates the conservation of @xmath2 , playing a role similar to the dzyaloshinskii - moriya interaction in shindou s model .
we consider an adiabatic cycle in which @xmath29 since @xmath30 is odd under time reversal , while @xmath31 is even , condition ( [ cond2 ] ) is clearly satisfied . at @xmath32 and @xmath33
the hamiltonian is time reversal invariant . in fig .
1(a ) we depict groundstates in the strong coupling limit at representative points along the cycle . at @xmath34 and @xmath35 @xmath30
dominates and locks the spins into a neel ordered state . at @xmath32 and @xmath36 ,
@xmath31 dominates , and the system is dimerized with singlet pairs of electrons occupying alternate bonds .
importantly , the groundstate at @xmath36 is distinguished from the groundstate at @xmath32 by the presence of _ unpaired _ spins at each end . when @xmath37 @xmath2 is conserved , and this model describes a spin pump . in this case
@xmath38 can be decomposed into two independent periodic potentials which lock the densities of the up and down spin particles and slide in opposite directions as a function of @xmath4 .
as @xmath4 evolves from @xmath39 to @xmath40 , the periodic potentials slide by one lattice constant .
provided there is space " for the added spin at the ends , spin @xmath41 will accumulate at the end following each cycle .
we wish to understand how this spin pump is modified when @xmath42 , so that @xmath2 is not conserved . in fig .
1(b ) we plot the single particle energy levels for a @xmath43 site chain as a function of @xmath4 for nonzero @xmath44 .
the bulk energy gap can be clearly seen with continuum states above and below .
the energy levels that cross the gap are end states .
each line consists of two states which are localized at opposite ends .
the crossing of the end states at @xmath36 will play a critical role in what follows .
when @xmath37 the degeneracy at @xmath33 is protected by spin conservation because the two states at each edge have @xmath45 .
nonzero @xmath44 does _ not _ lift the degeneracy provided the hamiltonian remains time reversal invariant at @xmath33 .
the two end states form a kramers doublet whose degeneracy can not be broken by any time reversal invariant perturbation .
because of the level crossing at @xmath36 it is clear that a system which starts in the ground state at @xmath32 will be in an excited state at @xmath46 .
however , since the end states merge with the continuum the excitation will not be localized near the edge , and bulk particle - hole pairs will be excited .
this is because there is no space " to put the excitations at the ends . in section
vb we will discuss the effect of connecting this pump to reservoirs which allow the end states to be
emptied " without exciting bulk particle - hole pairs .
for the purpose of this section , however , we will study the operation of an isolated pump by adding several sites at the ends of the chain for which @xmath30 and @xmath31 vanish .
this introduces additional midgap states localized at each end , allowing the cycle to proceed without generating bulk excitations .
1(c ) shows the energy levels as a function of @xmath4 with the extra sites added at each end .
there are now several midgap states at each end .
since all of the midgap states are localized at one end or the other , the low energy excitations of the system can be factorized as a product of excitations at each of the two ends . in fig .
1(d ) we plot for @xmath47 the energies of the lowest few many body eigenstates associated with a single end , obtained by considering particle hole excitations built from the single particle states localized at that end .
though this picture was computed for non interacting electrons , it is clear that the kramers degeneracy of the groundstate at @xmath36 and @xmath48 will be robust to the addition of electron electron interactions .
the first excited state at @xmath49 , @xmath40 and @xmath50 in fig .
1(d ) is four fold degenerate ( the middle level coming into that point is doubly degenerate ) .
this degeneracy , however , is an artifact of non interacting electrons .
the degeneracy is present because there are four ways of making particle hole excitations with two pairs of kramers degenerate states .
electron electron interactions , however , will in general split this degeneracy , as shown in the inset , so there will be no level crossing at @xmath46 .
we thus conclude that when the isolated pump starts in its ground state at @xmath32 , it arrives in an excited state after one complete cycle at @xmath46 . after a second cycle , however , at @xmath51
the system returns to its original state .
for this reason , we call it a @xmath0 pump " .
it is possible that by coupling to other degrees of freedom an _ inelastic _
process ( such as emitting a phonon ) could cause the excited state to relax back to the ground state .
nonetheless , there is an important distinction between this adiabatic process which generates an excited state and one that does not . in section
vb we will return to this issue when we discuss connecting the pump to leads .
the nontrivial operation of a single cycle depends critically on the time reversal symmetry at @xmath36 .
breaking time reversal symmetry at that point leads to an avoided crossing of the energy levels , so that the system returns adiabatically to its original state at @xmath46 . from the point of view of the end states ,
the non trivial pumping effect arises because there exist kramers degenerate end states at @xmath52 , but not at @xmath49 . in the next section we show that this property is determined by the topological structure of the _ bulk _ hamiltonian , @xmath53 .
in this section we introduce the time reversal polarization for non interacting electrons and show that changes in it define a topological invariant .
our discussion will parallel the theory of charge polarization in insulators@xcite . in order to establish this connection and to define our notation
we will therefore begin by reviewing that theory , which relates the charge polarization to the average center of wannier orbitals , which in turn are related to the berry s phase of the bloch wavefunctions .
we next consider the role of kramers degeneracy in time reversal invariant systems and define a corresponding time reversal polarization in terms of the difference between the wannier centers of kramers degenerate bands .
finally , we show that the _ change _ in the time reversal polarization between @xmath32 and @xmath36 of the pumping cycle defines a @xmath0 topological invariant which distinguishes a nontrivial @xmath0 pump from a trivial cycle . consider a one dimensional system with lattice constant @xmath54 , length @xmath55 with periodic boundary conditions and @xmath56 occupied bands .
the normalized eigenstates for the @xmath57th band can then be written in terms of cell periodic bloch functions as @xmath58 we may define wannier functions associated with each unit cell associated with lattice vector @xmath59 as @xmath60 the wannier functions are not unique because they depend on a gauge choice for @xmath61 .
in addition to changing the phases of the individual wavefunctions , the wavefunctions can be mixed by a general @xmath62 transformation of the form @xmath63 after this transformation , @xmath61 need no longer be the individual eigenstates of the hamiltonian , but rather should be interpreted as basis vectors spanning the space spanned by the @xmath56 occupied eigenstates .
the slater determinant of the @xmath56 wavefunctions is unchanged up to a phase .
marzari and vanderbilt@xcite have provided a prescription for choosing @xmath64 to optimally localize the wannier wavefunctions . here , however , we are concerned with the total charge polarization , which is insensitive to the details of @xmath64 .
the polarization is given by the sum over all of the bands of the center of charge of the wannier states associated with @xmath65 , and may be written@xcite @xmath66 where the @xmath67 berry s connection is given by @xmath68 the integral is over the brillouin zone from @xmath69 to @xmath70 .
if we require that the wavefunction @xmath71 be defined continuously in the reduced zone scheme , so that @xmath72 then @xmath73 , and the integral may be considered to be on a _ closed _ loop , despite the fact that @xmath74 is discontinuous from @xmath75 to @xmath70@xcite . under @xmath62 a transformation which preserves this continuity @xmath76 is invariant up to a lattice constant . for a transformation in which the @xmath67 phase of @xmath77 advances by @xmath78 when @xmath79 advances around the brillouin zone @xmath80 .
this reflects the fact that the polarization can only be defined up to a lattice vector .
_ changes _ in the polarization induced by a continuous change in the hamiltonian @xmath81 $ ] are , however , well defined .
thus , if the wave functions @xmath82 are defined _ continuously _ between @xmath83 and @xmath84 for all @xmath79 in the brillouin zone , then we may write @xmath85-p_\rho[t_1 ] = { 1\over { 2\pi}}\left [ \oint_{c_2}dk\ { \cal a}(t , k ) -\oint_{c_1}dk \ { \cal a}(t , k ) \right],\ ] ] where @xmath86 is the loop @xmath87 to @xmath70 for fixed @xmath88 . using stokes theorem , this can be written as an integral of the berry curvature @xmath89 over the surface @xmath90 of the cylinder spanned by @xmath79 and @xmath4 bounded by @xmath91 and @xmath92 : @xmath85-p_\rho[t_1 ] = { 1\over { 2\pi } } \int_{\tau_{12 } } dt dk\ { \cal f}(t , k ) .
\label{prho2}\ ] ] for a periodic cycle @xmath93=h[t]$ ] , the change in the polarization over one cycle , @xmath94 is given by the integral in ( [ prho2 ] ) over the entire torus defined by @xmath4 and @xmath79 .
this quantity is an integer and defines the first chern number associated with the wavefunction @xmath82 on the torus .
the chern number characterizes the charge pumped in each cycle . for a cycle which satisfies the time reversal constraint in eq .
( [ cond2 ] ) , @xmath95 , so the chern number is equal to zero .
consider now a time reversal invariant system .
the time reversal operator has the form @xmath96 where @xmath97 is spin operator and @xmath98 is complex conjugation .
since @xmath99 for spin 1/2 electrons , it follows from kramers theorem that every bloch state at wavevector @xmath79 is degenerate with a time reversed bloch state .
therefore , the energy bands come in pairs , which are degenerate at the two time reversal invariant points @xmath100 and @xmath70 , as shown in fig .
2 . note that in the presence of spin orbit interactions these bands can _ not _ be labeled with spin quantum numbers . in section iiia
we related the charge polarization as the sum of the wannier centers of all of the bands .
kramer s theorem guarantees , however that the wannier states come in kramer s degenerate pairs , in which each pair has the same center .
the idea is therefore to keep track of the center of _ one _ of the degenerate wannier states per pair by defining a partial polarization " .
this will contain more information than eq .
( [ prho ] ) , which is the sum over both states . for simplicity
we assume that there are no degeneracies other than those required by time reversal symmetry .
therefore , the @xmath56 eigenstates may be divided into @xmath101 pairs which satisfy @xmath102 where @xmath103 .
the second equation follows from the first , along with the property @xmath99 .
as shown in fig .
2 these bands are defined continuously at the degeneracy points @xmath104 . this representation is not invariant under the general @xmath62 transformation ( [ u2n ] ) . however , that invariance will be restored below .
we define wannier states associated with these two sets of bands along with the corresponding wannier centers . by analogy with ( [ prho ] )
the partial polarization associated with one of the categories @xmath105 or @xmath106 may then be written @xmath107 where @xmath108 the partial polarizations are clearly invariant ( up to a lattice translation ) under changes in the phases of @xmath109 and @xmath110 .
however , they appear to depend on the arbitrary choice of the labels @xmath111 and @xmath106 assigned to each band .
we now show that the partial polarizations ( [ ps ] ) can be written in a form that is invariant under a general @xmath62 transformation of the form ( [ u2n ] ) . to make this invariance explicit for @xmath112 we treat the portions of the integral for positive and negative @xmath79 separately , @xmath113.\ ] ] for the second term we use the time reversal constraint ( [ timerev ] ) along with the fact that @xmath114 to write @xmath115
it then follows that @xmath116 .
\label{pi1}\ ] ] the first term is expressed in terms of the berry s connection @xmath117 . however , since the path of integration is not closed , the second term is necessary to preserve gauge invariance . the second term , can be rewritten in a suggestive manner by introducing the @xmath62 matrix which relates the time reversed wavefunctions , @xmath118 in the representation ( [ timerev ] ) @xmath119 is a direct product of two by two matrices with @xmath120 and @xmath121 on the off diagonal . at @xmath122 and @xmath123 @xmath119
antisymmetric_. an antisymmetric matrix may be characterized by its pfaffian , whose square is equal to the determinant .
we then find that @xmath124 \over { { \rm pf}[w(0 ) ] } } = \exp[i \sum_\alpha ( \chi_{\pi,\alpha}-\chi_{0,\alpha})].\ ] ] thus , the second term in ( [ pi1 ] ) can be expressed in terms of @xmath125 $ ] .
this leads to @xmath126 \over { { \rm pf}[w(0)]}}\right)\right ] .
\label{pi2}\ ] ] using the identity @xmath127={\rm det}[x]{\rm pf}[a]$ ] it can be shown that under the @xmath62 transformation ( [ u2n ] ) @xmath128 \rightarrow { \rm pf}[w ] { \rm det}[u]$ ] . both terms in ( [ pi2 ] ) are thus clearly @xmath129 invariant . moreover , under a @xmath67 transformation the two terms compensate one another , so @xmath130 is @xmath62 invariant . like the charge polarization ( [ prho ] ) ,
@xmath130 is only defined modulo a lattice vector .
this is reflected in the ambiguity of the imaginary part of the log in ( [ pi2 ] ) as well as the dependence of gauge transformations where the phase of @xmath61 advances by @xmath131 for @xmath132 .
a similar calculation can be performed for @xmath133 , and it is clear from time reversal symmetry that @xmath134 modulo an integer , reflecting the kramer s pairing of the wannier states . from ( [ prho ] ) and ( [ ps ] ) the charge polarization is given by the sum of the two partial polarizations : @xmath135
we now define the _ time reversal polarization _ as the difference : @xmath136 this then has the form , @xmath137 \over { { \rm pf}[w(0)]}}\right)\right ] .
\label{pth1}\ ] ] this may be written more compactly in terms of @xmath119 as @xmath138 - 2 { \rm log } \left({{\rm pf}[w(\pi)]\over{{\rm pf}[w(0)]}}\right)\right ] .
\label{pth2}\ ] ] this can further be simplified by noting that the first term gives the winding of the @xmath67 phase of @xmath119 between @xmath39 and @xmath70 .
thus , @xmath139 - 2 { \rm log}\left({{\rm pf}[w(\pi ) ] \over { { \rm pf}[w(0)]}}\right)\right ] .
\label{pth3}\ ] ] since @xmath140 = { \rm pf}[w]^2 $ ] this quantity is an integer , and the due to the ambiguity of the log , this integer is only defined modulo 2 .
even and odd integers are distinct , however , and determine whether @xmath141 $ ] is on the same branch or opposite branch of @xmath142}$ ] at @xmath122 and @xmath70 .
an alternative way of writing it is thus , @xmath143}\over{{\rm pf}[w(0 ) ] } } { \sqrt{{\rm det}[w(\pi)]}\over{{\rm pf}[w(\pi ) ] } } , \label{pth4}\ ] ] where the branches of @xmath144}$ ] are chosen such that the branch chosen at @xmath122 evolves continuously along the path of integration in ( [ pth3 ] ) into the branch chosen at @xmath123 - eliminating the ambiguity of the square root .
( [ pth1]-[pth4 ] ) are among the principle results of this paper , and can be regarded as a generalization accounting for time reversal symmetry of the berry s phase formulation of the charge polarization@xcite .
the @xmath0 time reversal polarization @xmath145 defines two distinct polarization states . in the next section
we will argue that the value of @xmath145 is related to the presence or the absence of a kramers degenerate state at the end of a finite system . as is the case for @xmath76 ,
the value of @xmath145 is not meaningful by itself , because a gauge transformation @xmath146 changes its value .
equivalently , the presence or absence of a kramers degeneracy at the end can not be determined from the state in the bulk , since it will depend on how the crystal is terminated . nonetheless , the two values of @xmath145 are topologically distinct in the sense that the value of @xmath145 can not be altered by a continuous change in the hamiltonian which preserves time reversal symmetry .
however , in the next section we will argue that an adiabatic change in the hamiltonian which preserves time reversal symmetry at the end points - but not in between - leads to a well defined change in @xmath145 .
this change defines a topological classification of distinct pumping procedures . in the previous subsection we focused on a time reversal invariant hamiltonian , which occurs at @xmath32 and @xmath33 in our pumping cycle .
we now consider the continuous evolution of the hamiltonian through the cycle and show that the change in the time reversal polarization which occurs in _ half _ the cycle defines a @xmath0 topological invariant which distinguishes a @xmath0 spin pump from a trivial cycle .
this physical meaning of this invariant is easiest to see pictorially by considering the shift in the wannier centers in the course of one cycle .
3(a ) depicts the centers of the occupied wannier orbitals as a function of @xmath4 . at @xmath147 and @xmath40
, time reversal symmetry requires that the wannier states come in time reversed pairs .
however , in going from @xmath32 to @xmath36 the wannier states switch partners " . in this process , the time reversal polarization , which tracks the difference between the positions of the time reversed wannier states changes by one .
in addition , this switching results in the appearance of an unpaired occupied wannier state at each end .
since the wannier states come in pairs , there must be twofold kramers degeneracy associated with each end - resulting in a total degeneracy of four .
when the system evolves from @xmath36 to @xmath46 , there is another switch , and the time reversal polarization returns to its original value .
however , since @xmath81 = \theta h[t - t]\theta^{-1}$ ] , the system with open ends does not return to its original state at @xmath32 but its ends are in an excited state because of the level crossing at @xmath36 .
we now relate the occurrence of this non trivial pumping cycle to a topological property of the _ bulk ground state _ as a function of @xmath4 .
we thus consider the change in the time reversal polarization , @xmath148 between @xmath32 and @xmath36 .
note that though @xmath145 is not gauge invariant , the _ change _ in @xmath145 is gauge invariant .
this difference @xmath149 defines a @xmath0 topological invariant which characterizes the mapping from the torus defined by @xmath79 and @xmath4 to the wavefunctions : @xmath82 . from ( [ pth4 ] )
we may write this invariant as @xmath150}\over { { \rm pf}[w(\gamma_i)]}}. \label{delta2}\ ] ] here @xmath151 are the four time reversal invariant points " on the torus shown in fig
the branches of the square root are chosen as in ( [ pth4 ] ) by continuously evolving @xmath152}$ ] along the paths @xmath153 and @xmath154 . in order to apply this formula ,
it is crucial for the wavefunctions to be defined _ continuously _ on the torus .
it is always possible to find such smoothly defined wavefunctions via a transformation of the form ( [ u2n ] ) because the chern number - which is the obstruction to doing so - is equal to zero . in the appendix
we will relate different mathematical formulations of this invariant .
we will first show that it can be interpreted as an obstruction to defining continuous wavefunctions provided an additional constraint relating the wavefunctions at time reversed points is enforced .
this leads to a different formula for the invariant , which can be expressed in terms of the berry s curvature @xmath155 and the berry s connection @xmath156 .
we will then prove that ( [ delta2 ] ) is equivalent to the formula for the invariant introduced in ref . .
the preceding discussion has focused on non interacting electrons .
an important question is therefore whether or not these ideas apply to interacting systems .
the presence or absence of a groundstate kramers degeneracy associated with the ends of a finite interacting time reversal invariant system is clearly a well posed yes or no question .
this suggests that the time reversal polarization is a well defined quantity - at least for non fractionalized phases for which the groundstate with periodic boundary conditions is non degenerate .
therefore , we believe the topological distinction of the @xmath0 pump is still present with interactions . calculating the time reversal polarization for interacting electrons
is more subtle than for non interacting electrons .
one possible approach would involve characterizing the _ entanglement entropy _ , as in ref .
, which is sensitive to the presence or absence of end states . in this section
we adopt a simpler approach by studying an interacting version of the model introduced in section ii using abelian bosonization .
we find that bosonization provides a natural description of the time reversal polarization .
we begin with a continuum version of ( [ ham ] ) described by the hamiltonian density , @xmath157 here @xmath158 is a four component field , where the left and right moving fields @xmath159 are specified by the eigenvalues of @xmath160 and the spin @xmath161 by @xmath162 .
we now bosonize according to @xmath163 where @xmath164 is a short distance cutoff .
define charge / spin variables so that @xmath165 , and charge / current variables ( with @xmath166 ) as @xmath167 .
these obey @xmath168 = i ( \pi/2 ) \delta_{\mu\mu'}\delta(x - x')$ ] .
the bosonized hamiltonian then has the form ( [ ham ] ) with @xmath169,\ ] ] @xmath170 @xmath171 and @xmath172 in the absence of the spin orbit term , the spin sector of this hamiltonian ( when @xmath173 is pinned at @xmath174 ) is equivalent shindou s model@xcite .
this hamiltonian describes an insulating phase in which both @xmath173 and @xmath175 are pinned .
first focus on the case @xmath22 where the hamiltonian is time reversal invariant .
if we choose a gauge such that @xmath176 , the behavior of these operators under time reversal can be deduced : @xmath177 the time reversal invariance of the hamiltonian when @xmath22 can easily be verified .
it is now straightforward to consider time reversal invariant interaction terms , such as @xmath178 , @xmath179 , @xmath180 , @xmath181 , @xmath182 , etc .
provided these interaction terms ( as well @xmath44 defined above ) are not too large , the system will retain its bulk gap and be in a phase in which @xmath173 and @xmath175 are pinned .
we now identify the time reversal polarization with @xmath183 the apparent dependence of @xmath145 on the spin quantization axis is an artifact of abelian bosonization .
in fact , @xmath145 is @xmath184 invariant .
this can be seen by noting that global spin rotations are generated by @xmath185 and @xmath186 .
the latter obeys @xmath187 = \pm\pi s^\pm$ ] , so that @xmath188=0 $ ] .
it can further be seen that even in the presence of spin nonconserving terms in @xmath44 as well as the interaction terms discussed above , @xmath189=0 $ ] . since @xmath190 there are two distinct possible values for the time reversal polarization : @xmath191 or @xmath192 .
thus @xmath145 can be used to classify time reversal invariant insulating states .
consider a finite system with ends .
we now argue that the value of @xmath193 determines the presence or absence of kramers degenerate states at the ends .
the end of a one dimensional system at @xmath194 must be characterized by a boundary condition for @xmath195 .
time reversal symmetry limits the possible values to @xmath196 .
the value of @xmath57 , however , depends on how the lattice is terminated .
first suppose that @xmath197 . then , when @xmath198 , the pinning of @xmath175 in the bulk is not consistent with the boundary condition .
the closest it can be is @xmath199 .
thus , near the end there must be a kink of @xmath200 in @xmath175 at the end .
time reversal symmetry requires these two possibilities to be degenerate , so there is a kramers degeneracy of two at the end . on the other hand , when @xmath201 , the bulk energy gap is consistent " with the boundary condition , allowing for @xmath202 everywhere .
the groundstate in this case is unique .
we thus conclude that bosonization provides an alternative approach for formulating the time reversal polarization in terms of @xmath175 , just as it allows for a formulation of the charge polarization @xmath203 .
this suggests that the topological distinction of the @xmath0 spin pump remains in the presence of electron interactions .
the quantum spin hall phase introduced in ref . is a phase of a two dimensional electron system . in a manner analogous to laughlin s construction for the quantum hall effect@xcite ,
this phase , when compactified onto a cylinder , defines a @xmath0 pump of the sort studied in this paper . in this section
we outline the implications of the present work for the quantum spin hall effect .
we begin by relating the @xmath0 index introduced in section iii to the index which distinguishes the quantum spin hall phase from a band insulator .
we then discuss the presence or absence of gapless edge states in the quantum spin hall effect . finally , we comment on an alternative topological characterization of the quantum spin hall effect in terms of a chern number matrix " that has recently been proposed by sheng et al .
@xcite . for non interacting electrons ,
the electronic phase of a two dimensional system with a bulk gap is characterized by the wavefunctions defined on the brillouin zone torus , @xmath204 .
the relationship between the one dimensional @xmath0 pump and the two dimensional quantum spin hall effect can be established by the identification of @xmath205 with @xmath206 .
( [ cond2 ] ) then reflects the time reversal invariance of the two dimensional hamiltonian . as we prove in the appendix , the @xmath0 topological index introduced in ref . is equivalent to the @xmath0 index characterizing the pump .
the considerations of this paper provide a natural physical interpretation of this index in terms of the change in the time reversal polarization in half of the cycle .
in addition , our observation that the time reversal polarization is related to the presence or absence of a kramers degeneracy at the end suggests that the @xmath0 classification of time reversal invariant two dimensional ground states transcends the non interacting model for which it was derived .
this means that the quantum spin hall effect describes a phase that is distinct from a band insulator even in the presence of electron electron interactions . in the regular quantum hall effect ,
the topological structure of the bulk state guarantees the existence of gapless edge excitations .
the non trivial @xmath0 invariant , however , does _ not _ guarantee gapless edge states . as shown by wu
et al.@xcite and xu et al.@xcite , when the interactions at the edge are sufficiently strong the edge can undergo a transition which opens a gap .
the considerations of this paper allow us to prove that _ either _ there are gapless edge excitations _ or _ there is a ground state degeneracy at the edge associated with the breaking of time reversal symmetry . to establish this proposition ,
consider the many particle eigenstates of the quantum spin hall phase on a cylinder as a function of the magnetic flux through the cylinder .
when the radius @xmath59 of the cylinder large , then the @xmath207 change in the energy of the many particle eigenstates when the one half a flux is inserted will be much less than any energy gap .
but in the quantum spin hall state , the non trivial @xmath0 index requires the ground state to have a kramers degeneracy at either @xmath208 or @xmath209 , but not both .
thus there are two possibilities as schematically illustrated in fig .
4 . either there are edge states with energy @xmath207 which become gapless for @xmath210 or the ground state is degenerate for @xmath210 and split by at most @xmath207 by the magnetic flux .
this required ground state degeneracy distinguishes the quantum spin hall phase from that of a band insulator . unlike a band insulator , the quantum spin hall state in a system with edges can
_ not _ have a unique ground state with a gap for all excitations .
we now comment on a different topological classification of the quantum spin hall effect proposed by sheng et al . @xcite these authors defined a matrix of chern numbers by considering a system with a generalized class of periodic boundary conditions . specifically , they considered boundary conditions of the form @xmath211 , where @xmath212 define the periodicity and @xmath213 are _ independent _ phase twists for the up and down spins .
they then characterized the topological classes of the groundstate wavefunction as a function of these phase twists , and defined a matrix of chern numbers , @xmath214 .
sheng et al . argued that this classification contains more information than the @xmath0 classification because it distinguishes quantum spin hall states which belong to the same @xmath0 class .
this can be illustrated by looking at the continuum version of the graphene model introduced in ref .
, described by the hamiltonian @xmath215\psi.\ ] ] here , in the notation of ref .
, @xmath216 describes the sublattice of the honeycomb lattice , @xmath217 describes the two inequivalent valleys at the corners of the brillouin zone and @xmath218 describes the spin .
when @xmath219 is nonzero the system is in a quantum spin hall phase and belongs to the nontrivial @xmath0 class .
sheng et al .
@xcite argued that the _ sign _ of @xmath219 defines two distinct phases which are distinguished by the matrix of chern numbers .
when @xmath218 is conserved this is certainly correct , and the chern number matrix can be viewed as independent chern numbers for the up and down spins .
however , when spatial symmetries are relaxed and spin is not conserved this distinction is no longer meaningful .
the two phases discussed above are in fact the _ same _ phase because they can be continuously transformed into one another without closing the gap . specifically , consider the more general spin orbit interaction which preserves the energy gap : @xmath220 when the unit vector @xmath221 is continuously varied from @xmath222 to @xmath223 the two phases " are connected .
of course , the process of connecting these phases requires the breaking of the @xmath224 lattice symmetry of graphene .
but in general , disorder will break all spatial symmetries , so one can not rely on a spatial symmetry to protect a topological property .
this presents a conundrum because the chern matrix formulation distinguishes the two states with distinct topological integers - even when the @xmath224 symmetry is explicitly violated .
what happens to these integers when the continuous path in eq .
( [ path ] ) is adiabatically followed ?
the answer is that somewhere along the path the energy gap must vanish at the _ edge _ where the twisted spin boundary condition is imposed@xcite .
the spin phase twist imposed by sheng et al .
can be decomposed into a @xmath67 part @xmath225 and a spin " part @xmath226 .
the spin phase twist @xmath175 is fundamentally different from @xmath173 when the bulk hamiltonian does not commute with the @xmath2 .
the boundary where the spin phase twist is imposed is physically different from the rest of the system , and the spectrum of the hamiltonian will in general be different for different values of @xmath175 .
in contrast , the location of the charge phase twist @xmath173 introduced by niu and thouless @xcite , can be moved around by performing a local gauge transformation without changing the spectrum . since the vector potential can be spread out over the circumference @xmath227 of the torus the change in the spectrum due changing @xmath173 can be at most @xmath207 .
in contrast , if the spin phase twist is spread out over the circumference , changing @xmath175 changes the hamiltonian by an amount of order @xmath192 .
the spectrum need not be close to that of the physical hamiltonian . we conclude that the additional topological structure implied by the chern number matrix is a property of the _ boundary _ where the twisted phase condition is imposed rather than a property of the bulk two dimensional phase .
the bulk quantum spin hall effect is classified by the @xmath0 invariant alone .
is the @xmath0 pump we introduced a spin pump ?
since an isolated @xmath0 pump returns to its original state after two cycles , the simple answer to this question is no .
however , any functioning pump must be connected to reservoirs into which the pump can pump . in this section
we briefly consider the effect of connecting the @xmath0 spin pump to reservoirs .
we conclude that the @xmath0 pump _ does _ pump spin , though the spin pumped per cycle is not quantized .
moreover , we argue that when the coupling to the reservoirs is weak the @xmath0 topological structure of the pump is essential for a nonzero spin to be pumped . for stronger coupling , however , the @xmath0 structure is not essential .
we consider a simple case where the reservoirs can be described by non interacting electrons with vanishing spin orbit interaction .
we first suppose the coupling to the reservoir is weak , so that the level width @xmath228 induced in the pump due the coupling is small compared to the energy gap @xmath229 . however , we require the coupling @xmath228 to be large compared to the pumping rate , as well as any inelastic scattering rate for the end states . in the limit @xmath230
the eigenstates of the pump maintain their integrity , though coupling to the reservoirs allows transitions between different states .
as illustrated in fig .
5 , there is a point in every cycle @xmath231 , where the groundstate of the pump becomes degenerate .
this degeneracy is due to the end states , which are in proximity to the reservoir .
for @xmath232 the pump is in an _ excited _ state .
coupling to the reservoir , however , allows the pump to relax back to its groundstate .
this relaxation , however , must involve a process in the reservoir which is _ odd _ under time reversal .
generically this will involve changing the spin of the reservoir .
the spin added to the reservoir need not be quantized .
it could even be equal to zero , but generically , it will be of order @xmath41 . in this weak coupling limit
it is clear that the @xmath0 structure of the pumping cycle is essential because it guarantees the level crossing in the end states .
if the end states did not cross , then there would be no transitions , and the spin in the reservoirs would be unchanged after a complete cycle . however , finite coupling between to leads relaxes this requirement .
suppose that the time reversal symmetry is weakly broken at @xmath36 , so that there is a small anticrossing of magnitude @xmath233 . in this case the @xmath0 character of the cycle is lost .
but if @xmath234 , then the states have no way of knowing " about the anticrossing , and the pump proceeds as if @xmath235
. this reflects the fact that spin can be introduced into a reservoir which is connected to an insulating material when the insulator is deformed through a periodic cycle .
the spin injected can be expressed in terms of the unitary reflection matrix @xmath236 for electrons at the fermi energy in the reservoir@xcite , which in general depends on the hamiltonian @xmath53 of the insulator , @xmath237.\ ] ] in general , this quantity is non zero .
the difficulty is coming up with a cycle in an insulating material for which @xmath238 is not very small .
the @xmath0 pump accomplishes this by guaranteeing that there is a resonance in the reflection matrix , which occurs when the kramers degenerate end state appears .
note that this resonance need not involve charge transfer between the reservoir and the insulator , for it could be a _
kondo resonance _ where the kramers degenerate end state becomes entangled with the reservoir electrons .
it should be emphasized that the spin added to the reservoir is not a property of the bulk hamiltonian of the pump , but rather it depends on how the pump is connected to the reservoir .
the spin transferred to the reservoirs at the two ends of the pump need not be related .
thus , one can not view the spin as being pumped along the length of the pump .
however , the presence of the end state resonance , which follows from the change in the time reversal polarization , _ is _ a property of the bulk insulating state . in this sense ,
the @xmath0 pump is a pump for spin .
it is a pleasure to thank f.d.m .
haldane , e.j .
mele and t. pantev for helpful discussions .
this work was supported by the national science foundation grant dmr-00 - 79909 .
in this appendix we relate different mathematical formulations of the @xmath0 invariant @xmath229 .
our starting point is eq .
( [ delta1 ] ) along with eqs .
( [ pth1]-[pth4 ] ) which express the invariant in terms of the change in the time reversal polarization between @xmath32 and @xmath36 .
we will first show that @xmath229 can be interpreted as an obstruction to defining wavefunctions continuously , provided a time reversal constraint is enforced .
this will lead to a formula for @xmath229 in terms of the berry curvature and berry connection , which will be shown to be equivalent to ( [ delta1 ] ) .
we will then show that ( [ delta1 ] ) is equivalent to the @xmath0 invariant proposed for the quantum spin hall effect in ref . .
we will use a notation appropriate for the @xmath0 pumping problem and consider bloch wavefunctions defined continuously on the torus defined by @xmath239 and @xmath240 .
for the two dimensional quantum spin hall effect we should identify @xmath241 with @xmath79 and @xmath242 with @xmath243 .
it is well known that a nonzero value of the chern invariant is an obstruction to smoothly defining the wavefunction throughout the entire torus@xcite .
instead , wavefunctions must be defined on overlapping patches " which are related to each other by a gauge transformation called a transition function " . the chern number is then related to the winding number of the phase of the transition function around a non contractable path . for the problem studied in this paper ,
the chern number is zero , so there is no obstruction to finding a transformation of the form ( [ u2n ] ) which makes the wave functions smoothly defined on a single patch .
however , we will show in this section that if we enforce the _ time reversal constraint _ @xmath244
then a nonzero value of the @xmath0 invariant is an obstruction in a manner precisely analogous to the chern number .
this constraint means that the gauges for the wavefunctions at @xmath245 are not independent . at the four time reversal
invariant points @xmath246 , the allowed transformations of the form ( [ u2n ] ) are restricted to be _
symplectic _ , @xmath247 .
that a nonzero value of the @xmath0 invariant @xmath229 is inconsistent with this constraint is easy to see because it implies that @xmath248 = 1 $ ] for all @xmath79 and @xmath4 and @xmath249 = 1 $ ] , so eq . ( [ delta2 ] ) trivially gives @xmath250 .
we will now relate the @xmath0 invariant to the winding of the phase of transition functions relating the wavefunctions on different patches .
in addition to establishing the connection between the @xmath0 invariant and the chern invariant , this approach will derive a formula for the @xmath0 invariant which expresses it in terms of the berry s connection and curvature .
the similarity between the @xmath0 invariant and the chern invariant has been emphasized by haldane@xcite .
the formulation of the @xmath0 invariant as an obstruction has also been discussed by roy@xcite , though that work did not establish a formula for the invariant .
suppose that we have wavefunctions obeying ( [ tconst ] ) defined smoothly on two patches in the torus labeled a and b in fig .
6 . in patch
a the wavefunctions @xmath251 are smoothly defined everywhere in the upper left and lower right quadrants of fig .
6 , while for patch b @xmath252 are defined in the upper right and lower left quadrants . in the overlapping regions these different wavefunctions are related by a @xmath62 transition matrix @xmath253 where @xmath254 and @xmath57 are shorthand of @xmath255 and @xmath256 .
consider the change in the @xmath67 phase of @xmath257 around the closed loop @xmath258 in fig .
6 . @xmath259 .
\label{d1}\ ] ] this will clearly be an integer because it is equal to the winding number of the phase of @xmath260 $ ] around the loop @xmath258 .
if @xmath261 is non zero and can not be eliminated by a gauge transformation , then there is an obstruction to smoothly defining the wavefunctions on a single patch . in what follows ,
we show that @xmath262 is precisely equal to the @xmath0 invariant defined in this paper . from ( [ d1 ] ) , we may write @xmath263 where @xmath264 and likewise for @xmath265 .
since @xmath266 is smoothly defined in the interior of @xmath267 we may write it in terms of the berry s flux , @xmath268 this can not be done for @xmath269 which is not necessarily defined continuously inside @xmath267 .
however , it can be related to the berry s flux through @xmath270 , @xmath271 combining these we thus find the winding number for the transition function can be expressed as an integral involving the berry connection and the berry curvature . @xmath272 \
{ \rm mod } \ 2 .
\label{d2}\ ] ] the patch labels can be safely removed because @xmath155 is gauge invariant , and the line integral is gauge invariant modulo 2 .
it is essential that the time reversal constraint ( [ tconst ] ) be enforced for this equation to have meaning . if not ,
then a gauge transformation on patch b can change the line integral by 1 , making the formula vacuous .
when ( [ tconst ] ) is enforced , an _ odd _ value of @xmath261 can not be gauged away because the phases of @xmath273 and @xmath274 can not be independently changed .
thus , @xmath275 presents an obstruction to defining wavefunctions on a single patch .
we now show that this winding number is precisely the same as the invariant @xmath229 . to this end
, we rewrite @xmath229 in terms of the _ partial polarization _ @xmath130 defined in [ ps ] . using @xmath276
we have @xmath277 when ( [ tconst ] ) is enforced eq .
( [ pi2 ] ) for the partial polarization implies that @xmath278 eq .
( [ prho ] ) shows that @xmath279 combining the two terms thus establishes that @xmath280 .
the two terms in ( [ d2 ] ) thus acquire physical meaning : the line integral gives twice the change in the partial polarization between @xmath32 and @xmath36 , while the surface integral gives the change in the total polarization . in ref .
, the @xmath0 invariant was introduced by considering the matrix elements of the time reversal operator , @xmath281 this should be contrasted with the matrix @xmath282 introduced in section iii , which can be generalized as a function of @xmath4 to be @xmath283 at the four time reversal invariant points @xmath284 , @xmath285 and @xmath286 coincide , but in general they are different .
@xmath285 is unitary with @xmath287|=1 $ ] , while @xmath286 is not unitary .
since @xmath99 , @xmath286 is antisymmetric .
the pfaffian of @xmath254 is therefore defined for all @xmath79 and @xmath4 . in ref .
we argued that the @xmath0 invariant could be determined by counting the number of zeros of the pfaffian in half the torus , modulo 2 . to establish the equivalence of this with ( [ delta1 ] ) , we begin by rewriting the time reversal polarization @xmath145 in terms of @xmath288 $ ] .
the key observation to be made is that @xmath289 = { { \rm pf}[m(k , t ) ] \over { { \rm pf}[m(-k ,- t)]^*}},\ ] ] which can be proved by noting that @xmath290 and using the identity @xmath291 = { \rm det}[x ] { \rm pf}[a]$ ] . introducing @xmath292 $ ] for @xmath293
it follows that @xmath294 = { \rm log}\ p(k ) + { \rm log}\ p(-k)$ ] . thus we may rewrite ( [ pth3 ] ) as @xmath295 - 2 { \rm log } \left ( { p(\pi ) \over p(0 ) } \right ) \right ] , \label{pth5}\ ] ] where we have used the coincidence of @xmath296 and @xmath254 at @xmath122 and @xmath70 .
this may be simplified further by changing variables @xmath297 in the middle term and writing the last term as an integral from @xmath39 to @xmath70 .
this gives @xmath298 \ \ { \rm mod}\ 2,\ ] ] where the integral is now over the closed loop @xmath32 , @xmath299 .
this expression is only defined modulo 2 because of the ambiguity of the imaginary part of the log in ( [ pth5 ] ) .
thus , we have established that @xmath145 is given by the phase winding of the pfaffian , @xmath300 around the 1d brillouin zone modulo 2 . while this quantity is not gauge invariant , the change in it due to continuous evolution between @xmath301 and @xmath302 is gauge invariant .
this defines the @xmath0 topological invariant , which , as in ref . , may be written @xmath303 \
\ { \rm mod}\ 2,\ ] ] where @xmath304 is the boundary of half the torus defined by @xmath299 and @xmath305 ( see fig .
if @xmath306 $ ] has point zeros , then this quantity counts the number of zeros in @xmath307 modulo 2 . | we introduce and analyze a class of one dimensional insulating hamiltonians which , when adiabatically varied in an appropriate closed cycle , define a @xmath0 pump " . for an isolated system a single closed cycle of the pump
changes the expectation value of the spin at each end even when spin orbit interactions violate the conservation of spin . a second cycle , however returns the system to its original state .
when coupled to leads , we show that the @xmath0 pump functions as a spin pump in a sense which we define , and transmits a finite , though non quantized spin in each cycle . we show that the @xmath0 pump is characterized by a @xmath0 topological invariant that is analogous to the chern invariant that characterizes a topological charge pump .
the @xmath0 pump is closely related to the quantum spin hall effect , which is characterized by a related @xmath0 invariant .
this work presents an alternative formulation which clarifies both the physical and mathematical meaning of that invariant . a crucial role
is played by time reversal symmetry , and we introduce the concept of the time reversal polarization , which characterizes time reversal invariant hamiltonians and signals the presence or absence of kramers degenerate end states . for non interacting electrons we derive a formula for the time reversal polarization which is analogous to the berry s phase formulation of the charge polarization . for interacting electrons , we show that abelian bosonization provides a simple formulation of the time reversal polarization .
we discuss implications for the quantum spin hall effect , and argue in particular that the @xmath0 classification of the quantum spin hall effect is valid in the presence of electron electron interactions . |
the average 210 kev x ray spectrum of low / medium - redshift quasars observed with exosat ( comastri et al .
1992 ; lawson et al . 1992 ) , ginga ( williams et al . 1992 ; lawson & turner 1997 ) and asca ( reeves et al . 1997 ; george et al .
2000 ) is well represented by a power law with a slope similar to that of lower luminosity seyfert 1 galaxies ( @xmath8 1.92.0 , nandra & pounds 1994 ) and significant dispersion around the mean value @xmath9 0.20.3 .
the typical imprints of reprocessing gas ( i.e. the reflection `` hump '' and the k@xmath0 emission line ) , which characterize the spectra of seyfert 1 galaxies ( nandra & pounds 1994 ; nandra et al .
1997 ) are however not present in high luminosity objects .
+ indeed , only a small number of quasars have a spectrum more complex than a simple power law ( plus , sometimes , a soft excess at low energies ) .
this fact has been tentatively ascribed to different physical conditions of the accreting gas in the innermost regions surrounding high - luminosity quasars with respect to seyfert galaxies .
iwasawa & taniguchi ( 1993 ) suggested the existence ( based on ginga data ) of an x ray `` baldwin '' effect , whereby the equivalent width of the iron k@xmath0 line decreases with luminosity .
nandra et al .
( 1997 , 1999 ) have recently confirmed and extended this result using asca data . both the profile and strenght of the k@xmath0 line change with luminosity . in particular ,
for luminosities of 10@xmath10 erg s@xmath11 , ionized iron lines have been detected in a few radio - quiet quasars ( rqqs ) ( nandra et al .
1996 ; yamashita et al . 1997 ; george et al .
2000 ) . at higher luminosities
there is no evidence for any emission line at all ( nandra et al .
1999 ; vignali et al .
unfortunately , high - luminosity quasars are usually found at relatively high - redshift and therefore are rather weak in x rays , making the study of spectral features extremely difficult .
+ high luminosity quasars differs from seyfert 1s also in the observed shape of the x ray continuum .
recent studies have suggested that high - redshift rqqs ( vignali et al . 1999 ) have flatter x ray spectral slopes than lower redshift objects ( george et al .
2000 ) . however , it is not clear whether the spectral flattening is a function of redshift , implying a flattening of the continuum towards high energies , or rather it depends on the luminosity , thus suggesting a different emission mechanism at high luminosities .
+ luminous radio - quiet quasars in the local universe are rare . in this
regard , pds 456 is an exception and could provide hints on the nature of some of the properties described above , which are of great relevance to the physics of agns .
pds 456 is a bright ( b=14.7 ) nearby ( z = 0.184 , torres et al .
1997 ) radio - quiet quasar ( s@xmath12 = 22.7 mjy , condon et al . 1998
; radio - loudness r@xmath13 = @xmath140.7 , reeves et al .
2000 ) close to the galactic plane ( l = [email protected] , [email protected] ; n@xmath16 @xmath17 2 @xmath2
10@xmath18 @xmath4 , dickey & lockman 1990 ) .
the bolometric luminosity ( l @xmath6 10@xmath19 erg s@xmath11 ; reeves et al . 2000 ) makes pds 456 more luminous than the nearby ( z = 0.158 ) radio - loud quasar 3c 273 , without being jet - dominated .
the asca and rxte quasi - simultaneous observations discussed by reeves et al .
( 2000 ) reveal the presence of a deep edge like feature at e @xmath20 8 kev which is likely to be due to highly ionized iron atoms .
the 210 kev flux doubled its intensity during a strong outburst which lasted for about 4 hours . in order to further investigate the peculiar properties of this quasar we have carried out a medium deep observation of pds 456 with bepposax .
the results of the analysis are presented in @xmath21 2 , compared with a re - analysis of the asca observation in @xmath21 3 and discussed in @xmath21 4 . throughout the paper a friedmann cosmology with h@xmath22 = 50 km s@xmath11 mpc@xmath11 and q@xmath22 = 0.0 is assumed .
the italian - dutch satellite bepposax ( boella et al . 1997a ) carries four co - aligned narrow - field instruments ( hereafter nfi ) , two of which are gas scintillation proportional counters with imaging capabilities : the low energy concentrator spectrometer ( lecs , 0.110 kev , parmar et al .
1997 ) and the medium energy concentrator spectrometer ( mecs , 1.510 kev , boella et al .
the remaining two instruments are the high pressure gas scintillation proportional counter ( hpgspc , 4120 kev , manzo et al .
1997 ) and the phoswich detector system ( pds , 13200 kev , frontera et al .
the quasar was observed by bepposax on 1998 august 19 - 20 .
data presented in this paper were obtained only with the imaging spectrometers , as pds 456 has not been detected in the other two instruments .
standard reduction techniques and screening criteria were applied in order to produce useful linearized and equalized event files .
a total of 62.7 ks and 29.7 ks have been accumulated for mecs and lecs instruments , respectively .
spectra have been extracted from circular regions of radius 4@xmath23 and 6@xmath23 around the source centroid for the mecs and lecs , respectively .
background spectra have been extracted from both blank - sky event files and from source - free regions in the target field - of - view .
no apparent difference between the two spectra have been revealed .
the background contributes to about 10 % and 25 % to the mecs and lecs count rates , the total net counts being 6.35@xmath24 @xmath2 10@xmath25 in the 1.510 kev band and 2.12@xmath26 @xmath2 10@xmath25 in the 0.44 kev energy range , respectively .
+ the xspec ( version 10 , arnaud 1996 ) program has been extensively used to perform spectral analysis . in this paper errors
are quoted at 90 % confidence level for one interesting parameter ( @xmath27 = 2.71 , avni 1976 ) , and energies are reported in the source rest frame . in all our models , for both neutral and ionized absorbers , we assume solar abundances as tabulated in anders & grevesse ( 1989 ) .
pds 456 lies fairly close to the galactic plane ( b @xmath6 11@xmath15 ) .
the galactic h i column density derived by radio measurements is about 22.4 @xmath2 10@xmath18 @xmath4 ( dickey & lockman 1990 ; stark et al .
we have checked whether molecular hydrogen , traced by co emission , is present towards the pds 456 direction , with negative result ( lebrun & huang 1984 ; dame et al . 1987 ) .
therefore a value of 3 @xmath2 10@xmath18 @xmath4 ( which is similar to the value found from the near - infrared observation of pds 456 by simpson et al .
1999 ) for the galactic column density has been employed in all the spectral fits .
+ the combined bepposax mecs and lecs spectra were first fitted with a single power law plus galactic absorption .
the complexity of the spectrum is evident from the quality of the fit ( @xmath28/dof = 253/105 ) and from the data - to - model ratio ( fig . 1 ) .
this shows systematic deviations over the entire lecs+mecs band .
in particular , an excess of counts below @xmath6 1 kev , as well as a strong deficit above 78 kev are clearly present , suggesting the presence of additional components at those energies .
furthermore , the monotonic rising of the model / counts ratio between @xmath29 and @xmath6 5 kev suggests that additional absorption , exceeding the galactic value , is required .
lcccccccc & @xmath30 & @xmath31 & @xmath32 & @xmath33 & @xmath34/logu@xmath35 & @xmath36 & @xmath37 & @xmath38/dof + & & ( 10@xmath7 @xmath4 ) & ( 10@xmath3 @xmath4 ) & & & ( kev ) & & + * a * & & & & 0.94@xmath39 & & 8.77@xmath40 & 1.23@xmath41 & 143/103 + * b * & 3.95@xmath42 & 2.40@xmath43 & & 1.29@xmath44 & & 8.82@xmath45 & 0.98@xmath46 & 97/100 + * c * & 3.99@xmath47 & 2.77@xmath48 & & 1.61@xmath49 & 4230@xmath50 & & & 94.2/100 + * d * & 3.95@xmath51 & 2.67@xmath52 & 4.47@xmath53 & 1.44@xmath54 & 3.90@xmath55 & & & 94.7/100 + @xmath56 in units of erg cm s@xmath11 + @xmath35 u = n@xmath57/n@xmath58 is the ionization parameter , defined as the ratio of the ionizing photons density at the surface of the cloud ( n@xmath57 = q/(4@xmath59r@xmath60c ) ) and the electron density of the gas . to cure the strong residuals at energies above 7 kev
( see fig .
1 ) we added an absorption edge to the power law model ( model * a * in table 1 ) .
the improvement in the fit is highly significant ( @xmath27/ @xmath61dof = 110/2 ) .
the best - fit energy of the edge is consistent at 90 % with that of the photoelectric k - edge of .
if this identification is correct , then the measured optical depth ( see table 1 ) implies that a high column density ( @xmath62 @xmath4 ) of highly ionized matter is reprocessing the primary x ray continuum of pds 456 along our line of sight . though the fit is largely improved by the addition of the absorption edge , the @xmath38 is still unacceptably high ( 143/103 ) and the residuals continue to show deviations at e @xmath63 1 kev .
moreover , the spectral index is very flat ( @xmath64 = 0.94@xmath39 ) . an acceptable fit and a more reasonable value for the power law slope ( @xmath64 @xmath17 1.3 )
is obtained adding to model a two additional spectral components ( model * b * ) : ( a ) an intrinsic neutral absorber obscuring the hard nuclear power law continuum , and ( b ) a soft power law , absorbed by the galactic column density only .
the best - fitting soft power law is very steep ( though only poorly constrained : table 1 ) , indicating the presence of a luminous source of soft photons ( l@xmath65 @xmath17 5.2 @xmath2 10@xmath66 erg s@xmath11 ) in the nuclear environment of pds 456 . the column density of neutral gas absorbing the hard x ray power law largely exceedes the galactic value along the line of sight to pds 456 , suggesting a `` type-2-like '' orientation for this high luminosity radio - quiet quasar .
the phenomenological model b provides a good fit to the observed 0.410 kev spectrum ( fig . 2 ) and good constraints to the various spectral parameters ( table 1 , fig . 3 and fig .
4 ) . to test the uniqueness of this model and
to investigate whether the observed soft excess of photons , compared to the absorbed nuclear power law , is indeed due to genuine emission and not instead to an artifact due to additional complexity in the geometrical or physical status of the matter covering our line of sight to pds 456 , a few different spectral models were tested .
more specifically , the soft x
ray power law was eliminated and the full - covering neutral absorber was either replaced by a partial covering absorber , allowing a fraction of the nuclear radiation to escape unabsorbed , or by a mildly ionized absorber transparent at energies lower than @xmath67 kev .
none of these models was able to provide good quality fits to our data .
we also replaced the soft power law with thermal emission from a collisionally ionized plasma , and refitted the data .
this did not modify the best - fit parameters of the remaining spectral components , nor improved further the quality of the fit , compared to model b. moreover , a significant contribution to the soft x
rays from starburst emission is unlikely .
indeed , given a 60100 @xmath68 m luminosity of 610 @xmath2 10@xmath69 erg @xmath4 s@xmath11 ( depending on the assumed ir spectral slope ) and relation [ 2 ] in david et al .
( 1992 ) , only about 5 % of the soft x
ray emission may be due to star - forming emission .
we then conclude that pds 456 shows genuine soft x
ray emission below @xmath70 kev .
such a component is likely to be related to the strong uv bump present in the multiwavelength energy distribution ( reeves et al . 2000 ) .
the energy and optical depth of the absorption edge in model b strongly suggest the presence of highly ionized gas absorbing and/or reflecting the primary radiation . to verify this hypothesys
we then replaced the photoelectric edge in model b , with either an ionized reflector ( parameterized by the model pexriv in xspec , magdziarz & zdziarski 1995 : model * c * ) or an ionized absorber ( using the cloudy - ferland 1996 - based model described in nicastro et al . 2000a , and adopting the mathews & ferland 1987 parameterization for the agn continuum : model * d * ) , and refitted the data . in both cases we left the soft and the hard , reflected / absorbed , power law free to vary independently .
both parameterizations gave acceptable @xmath28 .
the best - fit values of @xmath71 and @xmath72 were consistent with each other between the two parameterizations .
the hard reflected / absorbed continuum ( @xmath64 @xmath6 1.41.6 ) is still much flatter than the soft x ray power law . finally , the best - fit ionization parameters are consistent with each others , and their values correspond to a very high ionization state of the matter reflecting / absorbing the nuclear radiation , with iron mainly distributed among species xxi and xxvii . both models
provide then an acceptable description of our data .
however , according to matt , fabian & ross ( 1993 ) , such a high ionized reflector should produce strong ( ew @xmath6 300500 ev ) fluorescence iron k@xmath0 emission lines at 6.76.97 kev .
these lines are not observed in our data , the 90 % upper limits on their equivalent width being of 120 , 115 and 80 ev for neutral , he like and h like ionic species respectively .
the corresponding 3@xmath73 limits lie in the range 150180 ev .
we then conclude that , while statistically acceptable , the model including an ionized reflector is not fully consistent with the bepposax data of pds 456 .
k@xmath0 emission from highly ionized iron is instead not expected to be strong in the case of an ionized absorber .
the predicted strength of these lines depends on the fraction of solid angle covered by the absorber ( as seen by the central source ) , and on the dynamics of the gas ( see 4 ) , but does not exceed few tens of ev for the strongest line , fully compatible with our data .
a spectrum transmitted by a high column ( @xmath74 @xmath17 4.5 @xmath2 10@xmath3 @xmath4 ) of highly ionized matter would then self - consistently account for both the observed deep absorption fexxiv - xxvi k edge and the absence of k@xmath0 line emission from the same iron species ( model d , table 1 ) .
the best - fit spectrum gives a 210 kev flux of about 5.7 @xmath2 10@xmath75 erg s@xmath11 @xmath4 , which corresponds to an intrinsic 210 kev luminosity of about 1.2 @xmath2 10@xmath69 erg s@xmath11 . at this flux level and given the relatively short exposure time , the source has not been detected at high energies by the pds instrument ( which spends about half of the time to monitor the background ) .
in fact , a 3@xmath73 detection would have been possible only for a power law harder that @xmath64 @xmath17 1.3 ( guainazzi & matteuzzi 1997 ) .
the present upper limit provides therefore only a loose contraint on the high energy spectrum .
pds 456 was observed by the asca satellite ( tanaka et al .
1994 ) on 1998 march 7 during the ao6 phase .
the focal plane instruments consist of two solid - state imaging spectrometers ( sis , gendreau 1995 ) and two gas scintillation imaging spectrometers ( gis , makishima et al .
1996 ) , characterized by a good spectral resolution ( about 2 % and @xmath6 8 % at 5.9 kev , respectively ) and broad - band ( @xmath6 0.610 kev ) capabilities .
the observations were performed in faint mode and then corrected for dark frame error and echo uncertainties as suggested by otani & dotani ( 1994 ) .
the data were screened with the version 1.4b of the xselect package with standard criteria .
the final observing time is about 41 ks for both sis and gis detectors .
source counts were extracted from circles of 6@xmath23 radius for gis and [email protected] for sis centered on the source , and background spectra were extracted from source - free regions from the same ccd chip for sis and from the same field of view for gis . the source count rates ( after background - subtraction ) are 6.90@xmath76 @xmath2 10@xmath25 and 6.13@xmath76 @xmath2 10@xmath25 counts per second for sis and gis , respectively .
sis and gis data were fitted simultaneously allowing the relative normalizations to be free to vary , to account for residual discrepancies in the absolute flux calibration .
the residuals of a single power law fit ( @xmath77 , fig .
5 ) are similar to those of bepposax and fully consistent with the results reported by reeves & turner ( 2000 ) in their analysis of the same asca data .
in particular , a soft excess and an edge appear prominently . no emission line has been detected .
the 90 % upper limits for the neutral , he like and h like lines are in the 4570 ev range while , the 3@xmath73 upper limits are of the order of 100130 ev .
the presence of an ionized edge is fully confirmed by asca data ( @xmath27/@xmath61dof = 47/2 with respect to a single power law model ) .
the edge spectral parameters are consistent with the bepposax observation but less constrained due to the lower asca effective area at e @xmath20 7 kev . at low - energies , the spectrum clearly shows a soft component .
the addition of a power law improves the fit .
looking for consistency between the bepposax and the asca data of pds 456 , we then fitted our chosen parameterization of the bepposax data ( model d ) to the asca data . the edge is fully accounted for by this model .
both the ionization parameter and the warm absorbing column density are in agreement with bepposax values .
the asca data require a steeper spectral slope ( @xmath78 @xmath17 1.62.0 ) and a moderately lower neutral column density , but consistent , within the errors , with the bepposax results .
asca and bepposax observations of pds 456 were taken about 5 months ( observer frame ) apart .
the flux level of the source was lower during the asca observation ( f@xmath79 @xmath17 3.6 @xmath2 10@xmath75 erg @xmath4 s@xmath11 ) than during the bepposax one ( f@xmath79 @xmath17 5.7 @xmath2 10@xmath75 erg @xmath4 s@xmath11 ) .
a 50 % flux variation over 3 days has been detected by rxte during a quasi simultaneous observation with asca .
the rxte range of fluxes encompass the asca and bepposax measurements .
in addition , a strong flare with a doubling time of about 17 ksec is present in the rxte observation ( reeves et al . 2000 ) .
the underlying continuum of the asca / rxte spectrum ( @xmath8 2.32.4 ) is steeper than the slope obtained from the present analysis of bepposax and asca data and from the independent analysis of the asca spectrum discussed in reeves & turner ( 2000 ) .
the origin of such a discrepancy is unclear .
intercalibration errors between rxte and asca and/or a significant steepening of the continuum spectrum above 10 kev may provide a plausible explanation .
the presence of a prominent ionized iron k edge , with similar values for the best - fit energy and optical depth in rxte and bepposax data , represents the most striking observed feature in pds 456 and unambiguosly points to a highly ionized nuclear environment in this quasar .
the edge parameters are equally well reproduced by either reflection off or transmission through a highly ionized gas . in the first case the ionization status inferred from the edge energy would imply a strong ( @xmath6 300500 ev ; matt et al .
1993 ) ionized ( 6.76.97 kev ) iron line .
if the edge originates in a high column density ( n@xmath16 a few 10@xmath3 @xmath4 ) warm gas , no strong iron line is expected ( see below ) .
the bepposax and asca upper limits on the line intensity derived from the present analysis would then favour a transmission model . on the other hand ,
the tentative detection of a broad iron line in the rxte data ( @xmath80 1 kev , ew @xmath6 350 ev ) would instead favour the reflection scenario ( reeves et al .
2000 ) . to further investigate the properties of the high column density , highly ionized absorber
we have carried out more detailed calculations using the photoionization models described in nicastro , fiore and matt ( 1999 ) and nicastro et al .
( 2000a ) , which includes photoelectric and resonant absorption as well as gas emission .
the best - fit value of the ionization parameter implies that the iron atoms are equally distributed among the 3 highest ionization states : ( @xmath6 30 % ) , ( @xmath6 40 % ) and ( i.e. fully ionized , @xmath6 30 % ) . in a spherical configuration ,
the net intensity of the corresponding emission lines depends ( a ) on the fraction of solid angle ( as seen by the central source ) covered by the absorber / emitter , and ( b ) the ratio between the outflowing ( @xmath81 ) and the turbolence ( @xmath82 ) gas velocities .
we then ran our models for different values of the covering factor between 0.1 and 1 . to maximize the net contribute of line emission , compared to absorption , we used @xmath81/@xmath82 = 5 , which guarantees a peak - to - peak separation of absorption and emission lines by the same transition greater than 3 times the width of these features ( nicastro et al .
. the rather uncertain knowledge of the dynamical status of the gas prevents us from a more detailed treatment .
the results indicate that the total equivalent width of the line blend ( including recombination , fluorescent and intercombination lines ) never exceeds @xmath6100 ev , a value lower than the upper limits derived from asca and bepposax data , leading further support to the transmission model .
finally , the 0.1 - 2 kev continuum - plus - line fluxes predicted by these models allowed us to estimate the maximum contribution of gas emission to the soft component measured in both the bepposax and asca data to be not larger than @xmath6 20 % .
the most important results obtained from the bepposax observation of the luminous quasar pds 456 can be summarized as follows : * the bepposax observation confirms the presence of a deep ionized k edge discovered by rxte at @xmath6 8.59 kev implying the presence of highly ionized gas around the central source . + * the absence of significant iron line emission in both the bepposax and asca data favours a model where the hard x ray continuum is absorbed by a high column density of highly ionized gas .
further spectral complexity is present at lower energies , requiring the presence of an additional cold ( or weakly ionized ) absorber . + * the high luminosity of pds 456 ( @xmath83 @xmath6 10@xmath19 erg s@xmath11 ) and the properties of the absorbing medium which partially hides the x ray source make this object a good candidate to test the emission mechanisms and reprocessing in quasars .
+ further investigations with chandra and xmm would be extremely helpful to assess the nature of the x ray emission ( both in the soft and in hard x rays domain ) and the ionization state of the absorbing matter .
moreover , x ray spectroscopy of others bright , relatively nearby quasars would be very important to check whether pds 456 is an exceptional object or the imprints of highly ionized high column density gas are common in luminous quasars
. a better knowledge of the x ray properties of bright agn would also allow to probe the physics of accretion processes at high luminosities .
we thank all the people who , at all levels , have made possible the bepposax mission .
we also thank all the members of the asca team who operate the satellite and maintain the software and database . this work has made use of the nasa / ipac extragalactic database ( ned ) which is operated by the jet propulsion laboratory , caltech , under contract with the national aereonautics and space administration , of data obtained through the high energy astrophysics science archive research center online service , provided by the goddard space flight center and of the simbad database , operated at cds , strasbourg , france .
the authors would like to thank the referee , s. komossa , for her suggestions who improved the quality of the paper .
financial support from italian space agency under the contract ars98119 is ackowledged by c. v. , a. c and g.g.c .
p. this work was partly supported by the italian ministry for university and research ( murst ) under grant cofin98 - 02 - 32 .
aknowledge the nasa grant nag5 - 9216 .
nicastro f. , elvis m. , fiore f. , matt g. , savaglio s. , 2000b , proceedings of the conference `` x ray astronomy 99 : stellar endpoints , agns and the diffuse x ray background ''
otani c. , dotani t. , 1994 , asca newsl . 2 , 25 | bepposax and asca observations of the nearby ( z = 0.184 ) , high luminosity , radio - quiet quasar pds 456 are presented .
the x ray spectrum is characterized by a prominent ionized edge at 89 kev ( originally discovered by rxte , reeves et al .
2000 ) and by a soft excess below 1.5 kev .
the lack of any significant iron k@xmath0 emission line suggests for the edge an origin from line of sight material rather than from reflection from a highly ionized accretion disc .
the hard x ray continuum is indeed well modelled by transmission through a highly - ionized medium with a large column density ( @xmath1 4.5 @xmath2 10@xmath3 @xmath4 ) plus an additional cold absorber with a lower column density ( @xmath5 @xmath6 2.7 @xmath2 10@xmath7 @xmath4 ) . |
the deuteron , the only @xmath5 nucleus , provides the simplist microscopic test of the _ conventional nuclear model _ , a framework in which nuclei and nuclear interactions are explained as baryons interacting through the exchange of mesons . with improved nucleon - nucleon force models from the 1990s @xcite , and advances in our understanding of relativistic bound state techniques , more accurate calculations of deuteron structure are possible . during the 1990s
there have also been revolutionary improvements in our experimental knowledge of deuteron electromagnetic structure .
the start of experiments at the thomas jefferson national accelerator facility ( jlab ) has now made available continuous high energy beams , with high currents and large polarization , along with new detector systems .
several experiments have now significantly extended the energy and momentum transfer range of deuteron electromagnetic studies , including @xmath0 and @xmath6 for elastic @xmath3 scattering , and photodisintegration cross sections and polarizations .
existing experimental proposals promise to continue this trend .
other laboratories have also made several important measurements , generally at lower momentum transfer . in this context
, a review of the deuteron electromagnetic studies , examining the current status of the agreement between experiments and theory , is appropriate .
we attempt to cover our current knowledge of the deuteron electromagnetic structure , focussing on the recent jlab results , and prospects for the future .
we do not consider experiments that use the deuteron as a neutron target , for example , or for studies of the ( extended ) gerasimov - drell - hearn sum rule , deep inelastic scattering , or baryon resonance production in nuclei .
our interest is on experiments that probe the conventional picture of a nucleus as composed of baryons and mesons , and that probe how far models with these effective degrees of freedom can be extended .
table [ dexptab ] is a summary of some of the jlab exeriments that fit this description , and that we will review in the sections below . the high precision , large momentum transfer measurements may be sensivitive to effects not incorporated in the conventional nuclear model .
it seems self - evident that probes of short distances , well below the size of the nucleon , should require explicit consideration of the quark substructure of the nucleons .
our review suggests that evidence for the appearance of these effects seems to depend on the nature of the reaction . in elastic scattering , where only the @xmath7 chanel is expicitly excited , a successful description is obtained using a relativistic description of the @xmath7 channel together with a minor modification of the short - range structure of the nucleon current ( see sec . [
sec : elasticscat ] ) .
in photodisintegration by 4 gev photons , where hundreds of @xmath8 channels are explicitly excited , an efficient explanation seems to require the explicit use of quark degrees of freedom ( see sec .
[ photodis ] ) .
@lll experiment & reaction / observables & status + & _ elastic scattering _ & + 91 - 026 & @xmath0 & paper published @xcite + & @xmath1 & analysis in progress + 94 - 018 & @xmath0 & paper published @xcite + & @xmath6 & paper published @xcite + & _ electrodisintegration _ & + 89 - 028 & recoil proton polarization & analysis in progress + 94 - 004 & in - plane response functions & analysis in progress + 94 - 102 & high momentum structure & awaiting beam time + 00 - 103 & threshold @xmath9 & proposal + & _ photodisintegration _ & + 89 - 012 & cross sections & paper published @xcite + 89 - 019 & @xmath10 , @xmath11 , @xmath12 & paper published @xcite + 96 - 003 & cross sections & paper published @xcite + 93 - 017 & cross sections & analysis in progress + 99 - 008 & cross sections & analysis in progress + 00 - 007 & @xmath10 , @xmath11 , @xmath12 & awaiting beam time + 00 - 107 & @xmath10 , @xmath11 , @xmath12 & awaiting beam time + this review begins with a survey of deuteron wave functions , and then discusses the deuteron form factors , threshold electrodisintegration , and high energy deuteron photodisintegration .
we also call attention to recent reviews by garon and van orden @xcite , and by sick @xcite .
these reviews contain a discussion of the static properties of the deuteron and a survey of recent models of the nucleon form factors , two topics we have decided to omit from this work .
both also have an extensive discussion of the deuteron form factors .
calculations of deuteron form factors and photo and electrodisintegration to the @xmath7 final state require a deuteron wave function , the final state @xmath7 scattering amplitude ( if the transition is inelastic ) , and the current operator , all of which should be consistently determined from the underlying dynamics .
deuteron wave functions used in the _ conventional nuclear model _ will be reviewed in this section .
the nonrelativistic @xmath7 wave function of the deuteron can be written in terms of two scalar wave functions . in coordinate space
the full wave function is @xmath13 where @xmath14 are the spherical harmonics normalized to unity on the unit sphere , @xmath15 and @xmath16 are the reduced @xmath17 and @xmath18-state wave functions , and the @xmath19 distinguishes this from other ( relativistic ) components of the wave function to be described below .
the spin part of the wave function is @xmath20 introducing the familar compact notation for matrix operations on each of the two nucleon subspaces 1 and 2 @xmath21 where @xmath0 is any @xmath22 operator , we can show that @xmath23 these identities permit us to write the wave function ( [ rspacewf ] ) in a convenient operator form @xcite : @xmath24\;\chi^{1m}_{_{12 } } \label{rspacewf2}\end{aligned}\ ] ] in momentum space the deuteron wave function becomes @xmath25\;\chi^{1m}_{_{12 } } \ , .
\label{ft1}\end{aligned}\ ] ] we use the same notation for both coordinate and momentum space wave functions . if @xmath26 and @xmath27 , then @xmath28 note the appearance of the factors @xmath29 , a feature of the symmetric definition ( [ ft1 ] ) .
the normalization condition @xmath30 implies @xmath31=\int_0^\infty p^2\,dp \left[u^2(p ) + w^2(p)\right]\ ] ] the @xmath18-state probability , @xmath32 is an interesting measure of the strength of the tensor component of the @xmath7 force , even though it is a model dependent quantity with no unique measurable value @xcite .
the best nonrelativistic wave functions are calculated from the schrdinger equation using a potential adjusted to fit the @xmath7 scattering data for lab energies from 0 to 350 mev .
the quality of realistic potentials have improved steadily , and now the best potentials give fits to the @xmath7 data with a @xmath33/d.o.f @xmath34 .
the paris potential @xcite was among the first potentials to be determined from such realistic fits , and it has since been replaced by the argonne v18 potential ( denoted by av18 ) @xcite , the nijmegen potentials @xcite , and most recently by the cd bonn potentials @xcite . the @xmath17 and @xmath18-state wave functions determined from three of these models are shown in figs .
[ uwr ] and [ uwp ] .
these figures also show @xmath17 and @xmath18 wave functions from two relativistic models to be discussed shortly . in the right panel of the second figure we plot the dimensionless ratios @xmath35 and
@xmath36 , where the scaling functions , in units of gev@xmath37 , are @xmath38 here @xmath39 is the deutreron binding energy and @xmath40 the nucleon mass ( we used @xmath41 mev@xmath42 ) , @xmath43 gev@xmath42 , and @xmath44 gev@xmath42 .
we emphasize that these scaling functions _ have absolutely no theoretical significance _ and were introduced merely to remove the most rapid momentum dependence so that the percentage difference between models can be more easily read from the ratio graph .
we conclude that the five models shown are almost identical ( i.e.variations of less than 10% ) for momenta below about 400 mev , and that they vary by less than a factor of 2 as the momenta reaches 1 gev ( except near the zeros ) .
the definition of the relativistic deuteron wave function depends in large part on the formalism used to treat relativity . in formalisms based on hamiltonian dynamics ( discussed in sec .
[ hexamples ] ) the wave function in the deuteron rest frame can be taken to be identical to the nonrelativistic wave function , and no further discussion is necessary until the wave function in a moving frame is needed . in formalisms based on the bethe - salpeter equation @xcite , the covariant spectator equation @xcite , or on some other quasi - potential equation @xcite
, the wave functions usually have additional components which do not vanish in the rest frame . in the relativistic spectator formalism
@xcite , where one of the two bound nucleons is off - mass shell , the wave function is a sum of a positive energy component and a negative energy component @xcite @xmath45 where @xmath46 and @xmath47 are nucleon spinors for the off - shell particle ( particle 2 in this case ) with dirac index @xmath48 .
the positive energy part has the same structure as the nonrelativistic wave function , with an @xmath17 and @xmath18-state component .
the new negative energy wave function has the form @xmath49\,\chi^{1m}_{ab}\ , , \label{pwf}\end{aligned}\ ] ] where @xmath50 and @xmath51 are two additional @xmath52-state components of the wave function , and @xmath53 is the nuclear spin 0 wave function .
the names of the @xmath52-state wave functions follow from the fact that @xmath54 couples to the spin triplet ( @xmath55 ) and @xmath56 to the spin singlet ( @xmath57 ) wave function .
the equivalence of the two forms given in eq .
( [ pwf ] ) follows from identities like those given in eq .
( [ identity ] ) . two model relativistic @xmath17 and @xmath18-state wave functions were shown in figs .
[ uwr ] and [ uwp ] .
both models are based on relativistic one boson exchange model developed in ref .
model iib is a revised version of the model of the same name originally described in ref .
@xcite and model w16 is one of a family of models with varying amounts of off - shell sigma coupling that were introduced in connection the relativistic calculations of the triton binding energy described in ref .
these models are described further in a number of conference talks @xcite .
the relativistic @xmath52-state components are small , but can make important contributions to the deuteron magnetic form factor . as figs .
[ uwr ] and [ uwp ] show , the large @xmath17 and @xmath18-state components of these relativistic wave functions are very close to their nonrelativistic counterparts .
because of the very small value of the electromagnetic fine structure constant ( @xmath58 ) , elastic electron
deuteron scattering is described to high precision by assuming that the electron exchanges a single virtual photon when scattering from the deuteron . in this one - photon exchange approximation
@xcite elastic scattering is fully described by three deuteron form factors @xcite . in its most general form
, the relativistic deuteron current can be written @xcite @xmath59-g_3(q^2)\frac{(\xi'^*\cdot q)(\xi\cdot q)}{2m_d^2}\right\ } \,(d^\mu+d'^\mu ) \nonumber\\ & & + g_m(q^2)\;[\xi^\mu(\xi'^*\cdot q ) -\xi'^{*\mu}(\xi\cdot q)]\biggr)\ , , \label{deutcurrent}\end{aligned}\ ] ] where the form factors @xmath60 , @xmath61 , are all functions of @xmath62 , the square of the four - momentum transferred by the electron , with @xmath63 and @xmath64 .
[ in most of the following discussion we will suppress the explicit @xmath62 dependence of the form factors . ] in practice , @xmath65 and @xmath66 are replaced by a more physical choice of form factors @xmath67 with @xmath68 . at @xmath69 , the form factors @xmath70 , @xmath71 , and @xmath72 give the charge , magnetic and quadrupole moments of the deuteron @xmath73 the form factors can also be related to the helicity amplitudes of the deuteron current ( where helicity is the projection of the spin in the direction of the particle three - momentum ) . in the breit frame ( where the energy transfer @xmath74 is zero ) the polarizations of the incoming ( @xmath75 ) and outgoing ( @xmath76 ) deuteron are @xmath77 ( where the phases for the incoming deuteron follow the conventions of jacob and wick @xcite for particle 2 ) and the virtual photon polarization is @xmath78 hence , denoting the helicity amplitudes by @xmath79 , the three independent amplitudes are @xmath80 where @xmath81 .
the scattering amplitude in the one - photon approximation is @xmath82\,\left < d'|j_\mu|d\right > \ , , \label{matrix}\end{aligned}\ ] ] where @xmath83 and @xmath84 are electron spinors with @xmath85 ( @xmath86 ) the momentum and helicity of the incoming ( outgoing ) electron , respectively . squaring ( [ matrix ] ) , summing over the final spins , and averaging over initial spins
give the following result for the unpolarized differential cross section @xmath87 = { { d\sigma}\over{d\omega } } \bigg|_{ns}\,s(q^2,\theta ) \label{edxsect}\ ] ] where @xmath88 is defined by this relation , and @xmath89 is the cross section for scattering from a particle without internal structure ( @xmath90 is the mott cross section ) , and @xmath91 , @xmath92 , and @xmath93 are the electron scattering angle , the incident and final electron energies , and the solid angle of the scattered electron , all in the lab system .
the structure functions @xmath0 and @xmath1 depend on the three electromagnetic form factors @xmath94 where the definitions of @xmath95 , @xmath96 , and @xmath97 should be clear from the context . while cross section measurements can determine @xmath0 , @xmath1 , and @xmath71 , separating the charge @xmath70 and quadrupole @xmath72 form factors requires polarization measurements .
the polarization of the outgoing deuteron can be measured in a second , analyzing scattering .
the cross section for the double scattering process can be written @xcite @xmath98\ , , \label{eq : ds}\end{aligned}\ ] ] where @xmath99 is the polarization of the incoming electron beam , @xmath100 the angle between the two scattering planes ( defined in the same way as the @xmath101 shown in fig . [ fig_cross ] ) , and @xmath102 and the @xmath103 are the vector and tensor analyzing powers of the second scattering .
although there is a @xmath104 component to the vector polarization , the term is omitted from eq .
( [ eq : ds ] ) as there is no longitudinal vector analyzing power ; without spin precession , this term can not be determined .
the polarization quantities @xmath105 and @xmath106 ( sometimes denoted @xmath107 , but we will reserve capital letters for target asymmetries ) are functions of the form factors and the electron scattering angle @xmath108^{1/2}g_m(g_c+ { \textstyle{1\over3}}\eta g_q)\,\tan{\textstyle{1\over2}}\theta\nonumber\\ s\,p_z=\;\;\,{\textstyle{2\over3}}\eta\bigl[(1+\eta)(1+\eta\sin^2 { \textstyle{1\over2}}\theta)\bigr]^{1/2}g^2_m \tan { \textstyle{1\over2}}\theta\,\sec{\textstyle{1\over2}}\theta\nonumber\\ - \sqrt{2 } s\ , t_{20 } = { \textstyle{8\over3 } } \,\eta\ , g_c g_q + { \textstyle{8\over9}}\,\eta^2 g_q^2 + { \textstyle{1\over3 } } \eta \bigl[1 + 2(1+\eta)\tan^2 { \textstyle{1\over2}}\theta\bigr ] g_m^2 \nonumber\\ \sqrt{3}\ , s \,t_{21 } = 2 \,\eta \bigl[\eta + \eta^2\sin^2 { \textstyle{1\over2}}\theta\bigr]^{1/2 } g_m g_q\ , \sec { \textstyle{1\over2}}\theta \nonumber\\ -\sqrt{3 } s \,t_{22 } = { \textstyle{1\over2 } } \,\eta \,g_m^2\end{aligned}\ ] ] the same combinations of form factors occur in the tensor polarized target asymmetry as in the recoil deuteron tensor polarization . of these quantities , @xmath109 has been most extensively measured ; it does not require a polarized beam or a measurement of the out of plane angle @xmath100 . for measurements of @xmath0 and @xmath2 at forward electron scattering angles ,
the @xmath71 terms are very small , and one may approximate @xmath0 and @xmath2 by @xmath110 introducing @xmath111 gives @xmath112 the minimum of @xmath113 is reached for @xmath114 .
the node in the charge form factor , @xmath115 , occurs when @xmath116 , and @xmath117 , giving @xmath118 .
this approximation also makes it clear that @xmath119 largely depends on the deuteron structure , rather than the nucleon electromagnetic form factors . in the nonrelativistic limit ( to be discussed shortly ) , both @xmath70 and @xmath72 are a product of the nucleon isoscalar electric form factor multiplied by the _ body form factor _ , which is an integral over products of the deuteron wave functions weighted by spherical bessel functions .
hence , in this approximation , the nucleon electric form factor cancels in the ratio @xmath120 , and @xmath119 depends only on the deuteron wave function .
we note that the relations above between the form factors and the observables are model independent , so it is possible to extract form factors from the data and compare directly to theoretical calculations .
the most complete form factor determination appeared recently in ref .
@xcite ( see also the analysis in ref .
we will discuss the data below in sec .
[ thyandexp ] , after we have reviewed the experiments and the theory .
the initial measurements of elastic @xmath3 scattering were by mcintyre and hofstadter in the mid 1950s @xcite .
since then many experiments have run at several laboratories ; the fits of ref .
@xcite include 269 cross sections from 19 references , dating from 1960 to the presentan important feature of the recent fits of the world data is that the measured cross sections were refit rather than using extracted structure functions or form factors .
this is necessary since most extractions of @xmath0 ( @xmath1 ) used corrections for contributions of @xmath1 ( @xmath0 ) to their cross sections from earlier data . in some cases alternative definitions ( or incorrect formulas )
have been given .
a minor point is the definition of @xmath121 ; in some cases the recoil factor @xmath122 is included , while in our definition eq .
( [ mott ] ) it is not .
the magnetic form factor @xmath71 can be in units of @xmath123 ( our convention ) , @xmath124 , or dimensionless , with magnitude of 1.714 , 0.857 , or 1.0 , respectively , at @xmath62 = 0 , and leading to modified coefficients in eq .
( [ aandb ] ) . buchanan and yearian
@xcite have an alternate definition of @xmath70 , and @xmath72 , with @xmath125 .
benaksas @xcite and galster @xcite include an extra factor @xmath126 in the magnetic terms in @xmath0 and @xmath1 , which changes the @xmath62 dependence of the magnetic form factor , though not its value at @xmath62 @xmath127 0 .
ganichot @xcite and grossette @xcite both use a factor of @xmath128 , rather than @xmath129 , in their definitions of the mott cross section .
cramer @xcite give a dimensionally incorrect formula for their @xmath130 ( @xmath127 @xmath131 ) , with explicitly stated energy factors of @xmath132 , as opposed to our factor of @xmath133 . .
polarization experiments are much more difficult .
the first results were published in 1984 and there are now only 20 published @xmath6 data points , and 19 points for other polarization observables .
the fits of ref .
@xcite include a slightly larger data base , with 340 points for momentum transfers up to @xmath134 of about 1.6 gev ; this misses only a handful of the largest momentum transfer slac and jlab data points .
forward - angle cross section measurements suffice to determine @xmath0 , both because @xmath1 is small and because of the @xmath135 dependence .
the magnetic form factor @xmath71 is determined from large angle measurements of @xmath1 , since the @xmath0 contribution vanishes as @xmath91 @xmath136 180@xmath137 . with @xmath138 and @xmath139
, we obtain the following relations at @xmath91 @xmath127 180@xmath140 : @xmath141 one can see that the beam energies needed for high @xmath62 measurements of @xmath1 are quite low , with @xmath142 @xmath127 0.65 , 1.02 , 1.35 , 1.67 , and 1.97 gev corresponding to @xmath62 @xmath127 1 , 2 , 3 , 4 , and 5 gev@xmath42 , respectively .
note that throughout this review we use @xmath134 @xmath127 @xmath143 to avoid confusion with the magnitude of the three momentum transfer @xmath144 , and we use units of gev and gev@xmath42 , not @xmath145 or @xmath146 .
accurate measurements require that @xmath62 be known accurately since @xmath0 and @xmath1 vary rapidly with @xmath62 .
energy or angle offsets of a few times 10@xmath147 could lead to @xmath62 being off by up to 0.5% . for both @xmath0 and @xmath1 , this leads to offsets that increase with @xmath62 , reaching about 2% at @xmath62 @xmath127 1 gev@xmath42 and 4% at @xmath62 @xmath127 6 gev@xmath42 . while cross section measurements can determine @xmath0 , @xmath1 , and @xmath71 , separating the charge @xmath70 and quadrupole @xmath72 form factors requires polarization measurements , most often @xmath6 .
coincidence detection of the scattered electron and deuteron , which suppresses the background and allows experiments to be performed with moderate resolution , is a common technique .
several experiments have measured the structure function @xmath0 at small @xmath134 . of particular note are the high precision , 1 - 2% measurements from monterey @xcite , mainz @xcite , and saclay als @xcite .
the only measurements at moderately large @xmath134 are from slac e101 @xcite , bonn @xcite and cea @xcite , plus the two recent jlab experiments in halls a @xcite and c @xcite .
data for several experiments are shown in fig .
[ adata ] and summarized in table [ tab : edelast ] ; see refs .
@xcite , @xcite or @xcite for more extensive listings of data .
.[tab : edelast ] some measurements of @xmath0 .
symbols are given for data shown in the figures .
[ cols="<,<,^,^ , < " , ] @xmath148have larger errors and are consistent with the other data sets .
[ adata ] reveals an unfortunate history of certain measurements not agreeing to within the stated uncertainties .
for example , at low @xmath134 the monterey and mainz data overlap well , but the overlap of mainz and saclay als data indicates problems .
the four largest @xmath134 mainz points used rosenbluth separations , with @xmath0 largely determined from forward angles of 50@xmath140 , 60@xmath140 , 80@xmath140 , and 90@xmath140 at 298.9 mev .
saclay @xmath0 data were extracted from measured cross sections using previous @xmath1 data .
the closest corresponding saclay points , for the same scattering angles at a beam energy of 300 mev , have cross section about 7% smaller ; the difference is beyond the quoted experimental uncertainties .
significant differences such as this are often obscured by semilog plots or not plotting all data sets .
the body of data , aside from the lowest @xmath134 orsay point , suggests the correctness of the saclay measurments .
theoretical predictions span the range between the two data sets , and do not help to determine which is correct .
thus , a new high precision experiment in this @xmath62 range appears desirable .
the agreement between data from cea , slac e101 , and bonn near 1 gev was also unsatisfactory . in discussing these measurements , we will compare to the trend of the data as determined by the saclay and jlab measurements .
the cea data have large uncertainties , and are systematically low by about 1@xmath149 .
this experiment measured scattered electrons in a shower counter and deuterons in a spectrometer that used a quadrupole magnet with a stopper blocking out the central weak field region .
in such a case it is difficult to determine the solid angle precisely , and this uncertaintly might introduce systematic errors into this data . alternatively , since the spectra were not significantly wider than the elastic peak , it has been suggested that over - subtraction of background was a problem .
however , the background rates were determined to be consistent with expected rates from random coincidences and target cell walls .
bonn measured coincidence cross sections at large electron scattering angles , @xmath150 @xmath151 80@xmath137 - 140@xmath137 .
using forward angle data from slac e101 , cea , and orsay , bonn determined @xmath0 and @xmath1 .
slightly inconsistent results from the other experiments led to a small uncertainty on the bonn determination of @xmath0 .
thus , it is only the largest @xmath62 point , for which there was only the large angle bonn data , that has very significant disagreement with other determinations of @xmath0 .
finally , the lowest @xmath62 slac point is high .
the disagreements between the cea , slac e101 , and bonn data were part of the motivation for two jlab experiments that determined a. hall a experiment e91 - 026 @xcite measured @xmath0 for @xmath62 from 0.7 to 6.0 gev@xmath42 .
hall c experiment e94 - 018 @xcite measured @xmath0 in the same kinematics as its @xmath6 points , from 0.7 to 1.8 gev@xmath42 .
the main advantages of these experiments over previous work include the continuous beam , large luminosities , and modern spectrometers . the hall a measurements @xcite used @xmath152 100 @xmath153a beams on a 15 cm cryogenic ld@xmath154 target , to achieve a luminosity of approximately 5 @xmath155 10@xmath156/cm@xmath42/s , and two approximately 6 msr spectrometers .
the hall c measurements used the hms spectrometer along with the deuteron channel built to measure @xmath6 with the recoil polarimeter polder .
a feature of this system is that the solid angles of the two spectrometers were well matched , to within a few percent . in the overlap region ,
the two jlab experiments show better precision than the earlier data and generally good agreement ; comparisons of theory to data should focus on these results , rather than the older data .
however , these measurements also show a significant disagreement with each other .
uncertainties in each experiment are dominated by systematics of approximately 5 - 6% , with statistical precisions near 1% .
the hall c data are systematically larger than the hall a data by just over 2@xmath149 , slightly over 10% , and there appears to be a tendency of the data sets to diverge with increasing momentum transfer
. this discrepancy will be decreased by a few percent , but not eliminated , by a correction @xcite to a lower , more accurate , beam energy in hall c during the experiment .
it is unclear if the discrepancy can be further resolved .
an important experimental point is the use in these experiments , and in many earlier ones , of @xmath158 elastic scattering to calibrate the solid angle acceptance ; a fit to the world @xmath158 cross section data is often used @xcite .
however , recent high precision polarimetry results @xcite imply that @xmath159 is significantly smaller than previously believed , with @xmath159 dropping nearly linearly for @xmath62 from about 0.5 to 5.6 gev@xmath42 . refitting
the world cross section data , with the jlab data for the form factor ratio , decreases @xmath160 but enhances @xmath161 by about 2% @xcite .
the new fits imply that the @xmath158 cross section is generally a few percent larger than would have been calculated previously , less than the systematic uncertainties of most experiments , and too small to affect comparisons of measurements of the @xmath3 cross sections and @xmath0 .
the effects on the theoretical deuteron form factor predictions will be addressed below . in summary ,
the structure function @xmath0 is reasonably well determined up to @xmath62 @xmath127 6 gev@xmath42 , if one neglects several poorer data points .
there remain regions in which there are up to about 10% systematic discrepancies between data of different experiments ; the resolution of these problems is at present unclear .
@llccl experiment & @xmath134 ( gev ) & symbol & # of & year and + & & & points & reference + stanford mark ii & 0.10 - 0.13@xmath162 & @xmath148 & 2 & 1964 @xcite + stanford mark iii & 0.48 - 0.68 & @xmath163 & 4 & 1965 @xcite + orsay & 0.20 - 0.28 & @xmath148 & 3 & 1966 @xcite + orsay & 0.34 - 0.44 & @xmath148 & 3 & 1966 @xcite + stanford mark iii & 0.44 - 0.63 & @xmath148 & 5 & 1967 @xcite + orsay & 0.14 - 0.48 & @xmath148 & 4 & 1972 @xcite + naval research lab & 0.11 & @xmath148 & 1 & 1980 @xcite + mainz & 0.25 - 0.39 & @xmath164 & 4 & 1981 @xcite + bonn & 0.71 - 1.14 & @xmath165 & 5 & 1985 @xcite + saclay als & 0.51 - 1.04 & @xmath166 & 13 & 1985 @xcite + slac npas ne4 & 1.10 - 1.66 & @xmath167 & 9 & 1987 @xcite + jlab hall a & 0.7 - 1.4 & @xmath168 & 6 & unpublished + @xmath148these data sets are not shown ( @xmath1 must be inferred from the publication ) . the highest @xmath62 measurements of the @xmath1 structure function come from slac npas experiment ne4 @xcite , which covered the @xmath62 range of 1.20 to 2.77 gev@xmath42 .
these measurements extended the range of previous data from saclay @xcite ( which went to 1.1 gev@xmath42 ) , and from bonn @xcite ( which went to 1.3 gev@xmath42 , and gave the results for @xmath0 discussed above ) .
there is good overlap in all but a few of the earliest @xmath1 measurements .
measurements of @xmath1 were taken as part of e91 - 026 for @xmath62 @xmath127 0.7 to 1.4 gev@xmath42 , but analysis is not yet final .
a summary of the world data is shown in table [ t20tab ] .
the first polarization measurements were from an argonne / bates recoil polarimeter experiment @xcite and a novosibirsk vepp-2 experiment @xcite using a polarized gas jet target . in the gas jet experiment , a polarimeter measured the gas polarization after it passed through the interaction region .
there were three second generation experiments .
an argonne / novosibirsk vepp-3 measurement @xcite pioneered the use of storage cells , increased the internal target density about a factor of 15 over the gas jet alone , and pushed out to 0.58 gev , near the minimum in @xmath6 . because the polarization of the gas varies in the cell , due to wall and beam interactions , it was decided to normalize the gas polarization by setting the lowest @xmath134 datum , at 0.39 gev , to theory where the uncertainties are small .
such internal targets in storage rings are now common . a bonn polarized target experiment @xcite had large uncertainties .
at bates , the ahead deuteron polarimeter was used @xcite to determine @xmath6 in the range just past the minimum of @xmath6 to just past the node in @xmath70 .
a continuation of the novosibirsk experiment had large uncertainties @xcite , and was never published .
note that , to facilitate comparison between different experiments , the data are often `` corrected '' to an electron scattering angle of 70@xmath140 , but this adjustment and the uncertainty it introduces are small . over the past several years , internal target experiments at nikhef @xcite and novosibirsk @xcite have improved the precision of the lower @xmath169 data , over a range of @xmath170 0.3 - 0.8 gev .
the improvements in novosibirsk include higher luminosity resulting from an improved atomic beam source and a modified beam tune that allows use of a higher impedance storage cell .
jlab hall c e94 - 018 @xcite used the recoil polarimeter polder to measure to the highest @xmath62 , 1.72 gev@xmath42 .
the overlap of the data is good , but apparent systematic shifts can be seen , as the nikhef and bates measurements are more negative than the jlab and novosibirsk measurements ; note that this is not a difference between polarized targets and recoil polarimeters .
the issue of determing at what @xmath62 @xmath70 @xmath127 0 is affected by this difference .
the bates data @xcite suggest a larger @xmath62 than do the novosibirsk @xcite and jlab @xcite data .
the fits of ref .
@xcite do not include the unpublished novosibirsk data @xcite , and average between the bates and jlab points .
we do not discuss the data for @xmath171 and @xmath172 .
because of their dependence on @xmath71 , they have not been as useful in providing new information as has @xmath6 . to test time reversal invariance ,
one measurement of the _ induced _ vector polarization was made @xcite .
the observed result was consistent with zero .
it is instructive to see how the deuteron form factors are related to the free nucleon form factors and the deuteron wave function in the nonrelativistic limit .
because the deuteron is an isoscalar target , only the isoscalar nucleon form factors @xmath173 will contribute to the form factors . in the nonrelativistic theory , _ without exchange currents or @xmath174 corrections _ , the deuteron form factors are @xmath175\ , , \label{body}\end{aligned}\ ] ] where the _ body form factors _ @xmath176 , @xmath177 , @xmath178 , and @xmath179 are all functions of @xmath62 .
if we choose to evaluate these in the breit frame , defined by @xmath180 then the relativistic and nonrelativistic momentum transfers are identical , @xmath181 , and the relativistic nucleon form factors can be used without corrections . note that , in this nonrelativistic limit , only the nucleon electric form factors contribute to the deuteron charge and quadrupole structure , while both nucleon form factors contribute to the deuteron magnetic structure .
the nonrelativistic formulae for the body form factors @xmath18 involve overlaps of the wave functions , weighted by spherical bessel functions @xmath182
j_0\left(\tau \right)\nonumber\\ d_q(q^2 ) & = & { { 3}\over{\sqrt{2}\eta } } \int_0^{\infty } dr\ ; w(r ) \left[u(r ) - { { w(r)}\over{\sqrt{8 } } } \right ] j_2\left(\tau \right)\nonumber\\ d_m(q^2 ) & = & \int_0^{\infty } dr \,\bigl [ 2\,u^2(r ) - w^2(r ) \bigr ] j_0(\tau ) + \bigl[\sqrt{2}\,u(r)w(r ) + w^2(r)\bigr ] j_2(\tau)\nonumber\\ d_e(q^2 ) & = & { { 3}\over{2 } } \int_0^{\infty } dr\ , w^2(r)\ , \bigl [ j_0(\tau ) + j_2(\tau ) \bigr ] \label{bodyff}\end{aligned}\ ] ] where @xmath183 . at @xmath69 ,
the body form factors become @xmath184=1\nonumber\\ d_q(0 ) & = & { { m_d^2}\over{\sqrt{50 } } } \int_0^{\infty } r^2\,dr\ ; w(r ) \left[u(r ) - { { w(r)}\over{\sqrt{8 } } } \right ] \nonumber\\ d_m(0 ) & = & \int_0^{\infty } dr \,\bigl [ 2\,u^2(r ) - w^2(r ) \bigr]=2 - 3p_d \nonumber\\ d_e(0 ) & = & { { 3}\over{2 } } \int_0^{\infty } dr\ , w^2(r)= { { 3}\over{2 } } p_d\end{aligned}\ ] ] giving the nonrelativistic predictions @xmath185 with @xmath186 = 0.8798 the isoscalar nucleon magnetic moment .
the experimental value of the deuteron magnetic moment ( in these units ) is 1.7139 , leading to a predicted @xmath18-state probability of @xmath187 .
however this estimate can not be taken too seriously because the magnetic moment is _ very sensitive to relativistic corrections and interaction currents _ which can easily alter this result significantly .
these contributions will be reviewed qualitatively later in this review .
the study of deuteron form factors is complicated by the fact that they are a _ product _ of the _ nucleon isoscalar _ form factors , @xmath188 , and the _ body _ form factors , @xmath18 .
the dependence of the deuteron form factors on older models of the nucleon form factors is well discussed in ref .
a year ago the model of mergell , meissner and drechsel @xcite ( referred to as mmd ) gave a good fit , and could have been adopted as a standard .
figure [ ges - study ] shows the mmd isoscalar electric and magnetic form factors divided by the familiar dipole form factor @xmath189 ( with @xmath62 in gev@xmath42 ) .
note that the mmd model does not differ by more than 20% from the dipole over the entire @xmath62 range , suggesting that the dipole approximation works very well ( on the scale of the experimental errors see below ) .
however , recent measurements of the proton charge form factor are producing a surprising result , and at the time this review was being completed the picture was begining to change
. the recent jlab measurements of both the neutron and proton charge form factors now suggest that the isoscalar charge form factor may be well approximated by @xmath190 where @xmath191 is the dipole form factor and @xmath192 .
this jlab model is a sum of the old galster @xcite fit for the nucleon charge form factor ( supported by the recent measurements @xcite ) and a linear approximation to the new jlab @xmath193 data @xcite ( from which the charge form factor is obtained by assuming that @xmath194 ) . figure
[ ges - study ] shows that this form factor differs significantly from the dipole ( and also the previously favored mmd model ) , and may have a significant affect on the theoretical interpretation of the data .
this will be discussed in sec .
[ thyandexp ] below .
[ the @xmath195 ratios shown in the figure will be discussed in sec .
[ hexamples ] below . ] dividing the individual factors @xmath95 , @xmath96 , and @xmath97 [ introduced in eq . ( [ aandb ] ) ] by @xmath196 gives reduced quantities that are ( except for the weak dependence on the ratio of @xmath197 ) independent of the choice of nucleon form factor .
the contribution of these reduced quantities , which we denote by @xmath198 , @xmath199 , and @xmath200 , to the total @xmath201 is shown in fig .
[ a - study ] .
the figure shows that the contribution of the magnetic term , @xmath200 , to the total @xmath202 is small for @xmath203 gev@xmath42 ( for most of the @xmath62 range it is less than a few percent , reaching 10% at @xmath204 and also near 4 gev@xmath42 ) . above @xmath62 of 4 gev@xmath42
it is larger , and very model dependent .
this justifies the observation that the @xmath0 structure function can be well approximated by @xmath205 , as stated eariler in eq .
( [ tildeaapprox ] ) .
[ note that the new jlab data for @xmath206 , discussed briefly above , may enhance the magnetic contributions to @xmath0 above 4 gev@xmath42 , but will not change these conclusions qualitatively ] .
how well does this simple nonrelativistic theory explain the data ? the high @xmath62 data for
@xmath0 provide the most stringent test . in fig .
[ ahighq2 ] we compare the data for @xmath0 with calculations using the five nonrelativistic wave functions shown in figs .
[ uwr ] and [ uwp ] .
the calculations use eq .
( [ body ] ) with mmd isoscalar nucleon form factors and nonrelativistic body form factors given in eq .
( [ bodyff ] ) . in the right panel
the data and models have been divided by the `` fit '' described in eq .
( [ scale ] ) below .
it is easy to see that the nonrelativistic models _ are a factor 4 to 8 smaller than the data for @xmath207 gev@xmath42_. furthermore , since the difference between different deuteron models is substantially smaller than this discrepancy , it is unlikely that any _ realistic _ nonrelativistic model can be found that will agree with the data .
if the nucleon isoscalar charge form factor were larger than the mmd model by a factor of 2 to 3 it might explain the data , but this is also unlikely since the variation between nucleon form factor models is substantially smaller than this .
[ if we use the fit eq .
( [ jlabges ] ) to the jlab @xmath208 measurements the discrepancy will be even larger .
] we are forced to conclude that these high @xmath62 measurements _ can not be explained by nonrelativistic physics and present very strong evidence for the presence of interaction currents , relativistic effects , or possibly new physics . _ a detailed comparison of the nonrelativistic models with the three deuteron form factors , @xmath70 , @xmath71 , and @xmath72 is given in fig .
[ gfactors ] .
the functions used to scale the data and theory in the right - hand panels of the figure are @xmath209 where @xmath62 is in gev@xmath42 and @xmath210 . while some of the factors in these expressions are theoretically motivated ( note the presence of the dipole form of the nucleon form factor ) we do not attach _ any _ theoretical significance to these functional forms ; they merely provide a reasonably simple way to scale out the rapid exponential decreases from the form factors .
figure [ gfactors ] shows that the nonrelativistic models do a good job of predicting the form factors to a momentum transfer @xmath211 gev , beyond which departures from the data and variations of the models make the agreement increasingly unsatisfactory .
point seriously ; kinematic factors make it difficult to extract this point accurately and it is only one standard deviation from the theory .
the large @xmath72 and small @xmath70 values for the points at 0.55 gev and 0.58 gev result from the @xmath6 data points , from @xcite and @xcite respectively , being about 1 standard deviation more negative than calculations and overlapping the negative limit for @xmath6 of @xmath212 .
note that the tabulated uncertainty of @xmath72 at @xmath213 0.55 gev in ref .
@xcite should be asymmetric , + 0.075/@xmath2140.713 , ( as shown in the fig .
[ gfactors ] ) . ]
however , careful comparison reveals that there are still ( small ) discrepancies between the data and the nonrelativistic theory , even at low @xmath62 . the data and curves from the lower panel of fig .
[ ahighq2 ] are shown on an expanded logarithmic scale in fig .
[ alowq2 ] . in the lowest @xmath62 range from about 0.15 to 0.4 gev@xmath42
the data lie _ below _ the nonrelativistic theory , and are larger than the nonrelativistic theory only for @xmath62 above 1 gev@xmath42 .
the very low @xmath62 discrepancy seems to be due in part ( but not entirely ) to the columb distortion corrections that have been used recently to explain the deuteron radius @xcite .
we will discuss these corrections in sec .
[ nucsec ] below .
before we turn to a detailed discussion of the possible explanations for the failure of nonrelativistic models to explain the form factors at high @xmath62 , we discuss the low momentum transfer results from the perspective of effective theories .
the recent development of effective field theory provides a powerful method for theoretical study of low @xmath62 physics .
we will briefly review these results here , and return to the discussion of the high @xmath62 results in the next section .
effective field theory techniques exploit the fact that the physics at low energies @xmath215 ( or large distances @xmath216 ) can not be sensitive to the _ details _ of the interactions at very high energies @xmath217 ( or short distances ) . for example
, a low energy long wavelength probe may detect the presence of a small scattering center , but can not resolve its structure ( much as the far - field of a collection of electric charges depends on only one parameter , the total charge ) .
the parameters that depend on the short - range physics may be very important , but they can not be calculated and must be determined by a fit to the data .
effective field theory works best if the distance scales of the ( unknown ) short - range physics and the ( known ) long - range physics are clearly separated .
then for energies well below the scale of the short - range physics ( which we take to be @xmath218 ) , the short - range physics is treated systematically by expanding in powers of @xmath219 . in applications to the @xmath7 system , two scales have been discussed .
the so - called `` pionless theory '' chooses @xmath220 , and therefore requires no theory of the @xmath221 interaction .
this approach can work only at _ very _ low energies .
the chiral theory chooses @xmath222 and attempts to describe @xmath7 scattering up to the @xmath223 mass scale using the known pion - nucleon interaction as given by chiral symmetry .
( more precisely , if the magnitude of the center of mass relative momentum @xmath224 , the nucleon lab kinetic energy will @xmath225 , which is @xmath226 mev for the pionless theory , and @xmath227 mev for chiral perturbation theory . )
the effective range theory introduced by bethe @xcite is an early version of what we now call the pionless effective theory .
weinberg @xcite first applied modern chiral perturbation theory to @xmath7 scattering .
he proposed making a chiral expansion of the @xmath7 potential , and then inserting this potential into the schrdinger equation .
later kaplan , savage , and wise ( ksw ) @xcite criticized the consistency of this approach , and introduced an alternative organizational scheme , sometimes referred to as @xmath134 counting , in which the pion interaction is to be included as a perturbative correction ( as opposed to including it as part of the potential , and counting it to all orders , as proposed by weinberg ) .
ksw applied this method to calculation of the deuteron form factors @xcite .
it is now known that the tensor part of the one pion exchange interaction is too strong to be treated perturbatively , and recent work has focused on how to include the singular parts of one pion exchange in the most effective manner @xcite . in the following discussion we review the recent results from phillips , rupak , and savage ( prs ) , who give a nice account of the calculations of the deuteron form factors in a pionless theory @xcite . the effective lagrangian density for a pionless effective theory of the @xmath7 interaction in any channel ( the coupled @xmath228 for example ) is @xmath229 \nonumber\\ & + \cdots \ , , \label{eft1}\end{aligned}\ ] ] where @xmath230 is a ( nonrelativistic ) nucleon field operator and @xmath231 , @xmath232 , and the general coefficient @xmath233 ( which fixes the strength of the terms with @xmath234 derivatives ) are determined from data . the coefficients @xmath233 parameterize the strength and shape of the short range interaction .
the scattering amplitude predicted by ( [ eft1 ] ) is a sum of bubble diagrams which can be regularized using the ksw dimensional regularization scheme with power law divergence subtraction @xcite . in lowest order ( lo )
this bubble sum is @xmath235 where @xmath236 is the magnitude of the nucleon three - momentum in the c.m .
system , @xmath237 is regular at the pole @xmath238 , @xmath239 with @xmath39 the deuteron binding energy , @xmath240 is related to the asympotic normalization of the deuteron wave function , and the dependence of @xmath231 on the ( arbitrary ) renormalization point @xmath153 is dictated by the requirement that the overall result be _ independent _ of @xmath153 .
the lo result is @xmath241 in terms of the effective range expansion @xmath242 with @xmath202 the scattering length and @xmath243 the effective range , the lo calculation gives @xmath244 contributions from the next to leading order ( nlo ) term @xmath232 changes the relations in ( [ zeq ] ) and ( [ req ] ) ; in particular , the wave function renormalization constant @xmath240 begins to differ from unity and the effective range @xmath243 to differ from zero .
prs point out that the most stable results are obtained by constraining @xmath231 and @xmath232 to give the experimental values of the deuteron parameters @xmath245 fm and @xmath246 instead of @xmath247 and @xmath248 fm .
this is because the asympotic deuteron wave function is fixed by @xmath247 and @xmath240 @xmath249 and _ it is the wave function and not the scattering _ that largely determines the deuteron form factors and other electromagnetic observables .
@xmath250 using this approach , the lo charge form factor is given entirely by the asymptotic wave function ( [ asywave ] ) @xmath251 where @xmath252 is the magnitude of the three momentum transferred by the electron , and working in the breit frame ( where the differences between relativistic and nonrelativistic theory is a minimum ) is also equal to @xmath143 .
prs show that expansion of the charge form factor up to nnlo terms is @xmath253 \nonumber\\ & - { { 1}\over{6}}\,r_n^2\,q^2\,g_c^{(0)}(q^2 ) + \cdots\end{aligned}\ ] ] and because the wave function is correctly normalized , there are no wave function effects beyond nlo ( the second term ) . at nnlo effects from the finite nucleon size , @xmath254 ,
similarly , the lo quadrupole form factor obtained by prs is @xmath255\ ] ] with @xmath256 the asymptotic @xmath257-state ratio , and the lo quadrupole moment , @xmath258 , equal to @xmath259 @xmath127 0.335 fm@xmath42 .
the expansion of @xmath72 to nnlo is then @xmath261 \nonumber\\ & - { { 1}\over{6 } } \,r_n^2\ , q^2 g_q^{(0)}(q^2 ) + \cdots\end{aligned}\ ] ] note that the quadrupole moment at nlo includes a contribution @xmath262 from a four - nucleon - one - photon contact term , not determined by @xmath7 scattering , and is used to fit the experimental value of @xmath263 .
prs suggest that the absence of this piece of short distance physics in conventional calculations may explain their underprediction of the quadrupole moment .
the finite size of the nucleon again comes in at nnlo . with parameters largely set by other data , the deuteron charge , quadrupole , and magnetic form factors
are well predicted up to about @xmath213 0.2 gev , as shown in fig .
[ fig : eft ] . the approach seems to converge well , but beyond nnlo more parameters enter , and there is less predictive power .
the great strength of the pionless effective theory is that strips away complexity , revealing the essential physics required to understand the low @xmath134 results , and showing ( for example ) the central importance of the asympotic s - state normalization @xmath240 .
however , as expected , it clearly does not work for @xmath134 much beyond 0.4 to 0.5 gev .
the theory with pions ( sometimes referred to as a `` pionful '' theory ) will work to higher @xmath62 @xcite .
removal of divergences from these theories is under active study .
we now return to discussion of the reasons for the failure of nonrelativistic theory at high @xmath62 . in sec .
[ comparenr ] we showed that the naive nonrelativistic theory can not explain the deuteron form factor data for @xmath264 gev . in this section
we classify the possible explanations for this failure , preparing the way for detailed discussions to follow in secs.[nucsec ] and [ quarksec ] .
the differences between the data and the nonrelativistic theory can only be explained by a combination of the following effects * interaction ( or meson exchange ) currents ; * relativistic effects ; or * new ( quark ) physics .
the only possibilities excluded from this list are variations in models of the nucleon form factors , or model dependence of the deuteron wave functions . in the previous section we argued that _ neither _ the current uncertainty in our knowledge of the nucleon form factors , _ nor _ the model dependence of the nonrelativistic deuteron wave functions are sufficient to provide an explanation for the discrepencies .
possible interaction currents that might account for the discrepency are shown in fig .
because the deuteron is an isoscalar system , the familiar large @xmath265 exchange currents are `` filtered '' out and only @xmath266 exchange currents can contribute to the form factors .
the @xmath266 currents tend to be smaller and of a more subtle origin .
the nucleon @xmath267-graphs , fig .
[ mec]c , and the recoil corrections , fig .
[ mec]d , are both of relativistic origin .
( the recoil graphs will give a large , incorrect answer unless they are renormalized @xcite . )
the two - meson exchange currents should be omitted unless the force also contains these forces .
the famous @xmath268 exchange current is very sensitive to the choice of @xmath268 form factor , which is hard to estimate and could easily be a placeholder for new physics arising from quark degrees of freedom . in most calculations based on meson theory ,
the two pion exchange ( tpe ) forces and currents arising from crossed boxes are excluded , and , except for the @xmath268 current ( which we will regard as new physics ) , the exchange currents are of relativistic origin .
additional relativistic effects arise from boosts of the wave functions , the currents , and the potentials , which can be calculated in closed form or expanded in powers of @xmath174 , depended on the method used . at low @xmath62 calculations
may be done using effective field theories ( discussed in sec .
[ sec : eft ] ) in which a small parameter is identified , and the most general ( i.e. exact ) theory is expanded in a power series in this small parameter . in these calculations
relativistic effects are automatically included ( at least in principle ) through the power series in @xmath174 .
_ hence any improvement on nonrelativistic theory using nucleon degrees of freedom leads us to relativistic theory . _
alternatively , one may seek to explain the discrepancy using quark degrees of freedom ( new physics ) .
when two nucleons overlap , their quarks can intermingle , leading to the creation of new @xmath7 channels with different quantum numbers ( states with nucleon isobars , or even , perhaps , so - called `` hidden color '' states ) .
these models require that assumptions be made about the behavior of qcd in the nonperturbative domain , and are difficult to construct , motivate , and constrain . at very high momentum transfers it may be possible to estimate the interactions using perturbative qcd ( pqcd ) .
very little has been done using other approaches firmly based in qcd , such as lattice gauge theory or skrymions ( but see ref .
@xcite ) .
we are thus led to two different alternatives for explanation of the failure of nonrelativistic models . in one approach
the nucleon ( hadron ) degrees of freedom are retained , and relativistic methods are developed that treat boost and interaction current corrections consistently . in another approach ,
quark degrees of freedom are used to describe the short range physics , and techniques for handling a multiquark system in a nonperturbative ( or perturbative ) limit are developed .
these two approaches will be reviewed in the next two sections .
while the discussion appears to be focused on the deuteron form factors , it is actually more general , and will be applied later to the treatment of deuteron photodisintegration .
are these two approaches really different ?
superficially , of course , the answer must be : yes !
however , qcd tells us that all physical states must be color singlets , and a basis of states that describes any color singlet state can be constucted from _ either _ quarks ( and gluons ) _ or _ hadrons ( this would not be true if colored states were physical ) .
so at a deeper level it appears that either approach ( hadrons or quarks ) should work , and the best choice is the system that can describe the relevant physics more compactly .
further discussion of this issue is clearly beyond the scope of this review .
this long section is divided into six parts as follows : ( i ) introduction , ( ii ) overview of propagator dynamics , ( iii ) choice of propagator and kernel , ( iv ) examples of propagator dynamics , ( v ) overview of hamiltonian dynamics , and ( vi ) examples of hamiltonian dynamics .
the inhomogeneous lorentz group , or the poincar group , is described by 10 generators : three pure rotations , three pure boosts , and four pure translations .
if we require the interactions to be local and manifestly covariant under the poincar group , we are led to a local relativistic quantum field theory with particle production and annihilation @xcite . in this case
the poincar transformations of all matrix elements can be shown to depend only on the kinematics ( i.e. they depend only on the masses and spins of the external particles ) .
the disadvantage is that the number of particles is not conserved .
if perturbation theory can be used , this approach is very successful , but in the nonperturbative regime of strong coupling meson theory it leads to an infinite set of coupled equations that can not be solved in closed form .
numerical , nonperturbative solutions of field theory can be obtained in euclidean space for a few special cases @xcite .
methods that limit the intermediate states to a _ fixed number of particles _ ( two nucleons in this case ) are more tractable , and all modern calculations are based on the choices depicted in the decision tree shown in fig .
[ decision ] . in deciding which method to use , if is first necessary to decide whether or not to allow _ antiparticle , or negative energy _
nucleons to propagate as part of the virtual intermediate state .
since nucleons are heavy and composite , so that their antiparticle states are very far from the region of interest , some physicists believe that intermediate states should be built only from positive energy nucleons , and that all negative energy effects ( if any ) should be included in the interaction .
these methods are referred to collectively as _ hamiltonian dynamics _ and are represented by the left hand branch in the figure .
unfortunately , it turns out that this choice precludes the possibility of retaining the properties of locality and manifest covariance enjoyed by field theory .
alternatively , in order to keep the locality and manifest covariance of the original field theory , other physicists are willing to allow negative energy states into the propagators .
these methods , represented by the right - hand branch of the figure , are referred to collectively as _
propagator dynamics_. including negative energy states tends to make calculations technically more difficult and harder to interpret physically , and those who advocate the use of hamiltonian dynamics do not believe the advantages of exact covariance justify the work it requires .
unfortunately , these two methods are so fundamentally different that many physicists do not realize that the limitations of one may not apply to the other .
for example , for some choices of propagator dynamics all 10 of the generators of the poincar group will depend only on the kinematics , and the poincar transformations of _ all amplitudes can be done exactly_. with hamiltonian dynamics this is not the case ; some of the 10 generators must depend on the interaction , and transformation of matrix elements under these `` dynamical '' transformations must be calculated .
comparison of the two methods is therefore very difficult ; the language and issues of each are very different and one can be easily misled by the different appearance of the results .
we can not discuss these issues in detail in this review , and refer the reader to two recent references that survey the subject @xcite . here
we will give a short review of some recent calculations , and explain these differences as we go along .
propagator calculations all start from the field theory description of two ( in this case ) interacting particles .
while some may prefer to express the field theory as a path integral , it is also possible to adopt a more intuitive approach and imagine expanding the path intergral as a sum of feynman diagrams ( ignoring issues of convergence for the moment ) . in order to generate the deuteron bound state , which produces as a pole in the scattering matrix ,
it is necessary to sum an infinite class of diagrams , written as @xmath269 where @xmath270 is the _ kernel _ being iterated , @xmath271 the two body propagator , @xmath272 the scattering amplitude , and the other quantities are defined below .
this sum is obtained in closed form by solving the integral equation @xmath273 if the series ( [ sumdiag ] ) is compared to a geometric series @xmath274 , then the solution to the integral equation ( [ rel1 ] ) can be compared to the sum of the geometric series @xmath275 .
the geometric series converges only when @xmath276 , but its unique analytic continuation , @xmath275 , is valid for all @xmath277 .
similiarly , it is assumed that the solution to ( [ rel1 ] ) is valid even when the series ( [ sumdiag ] ) diverges . and
just as the geometric series has a pole at @xmath278 , the solution to ( [ rel1 ] ) will have a pole at @xmath279 , the square of the deuteron mass .
the amplitudes @xmath280 , @xmath281 , and @xmath282 are all matrices in the @xmath7 spin - isospin space , and are functions of the four - momenta @xmath283 and @xmath284 , with @xmath285 and @xmath286 the momenta of the two particles ( labeled in fig .
[ feynmanladder ] ) .
the dimension of the volume integration is @xmath287 , normally either 3 + 1=4 ( 3 space + one time dimensions ) for the bethe - salpeter method , or 3 + 0=3 for the quasipotential methods described below . if eq .
( [ rel1 ] ) has a homogenous solution at some external four momentum @xmath288 , the scattering matrix will have an @xmath289 channel pole ( represented in fig .
[ feynmanbound ] ) , signifying the existence of a deuteron bound state .
the _ vertex _ function for the deuteron bound state satisfies the equation @xmath290 with covariant normalization condition @xmath291 the covariant bound state wave function is defined by @xmath292 one of the advantages of the propagator approach is that the construction of the current operator is comparatively straightforward .
it follows ( at least in principle ) from summing all electromagnetic interactions with all the consituents everywhere in the ladder sum . for bound states described by the bethe - salpeter or spectator formalisms ( see the discussion below ) there are two diagrams , illustrated in fig .
[ feynmancurrent ] , that can be written @xmath293 in the first term , @xmath294 is the sum of the neutron and proton currents [ recall eq .
( [ nnff ] ) ] and we have chosen particle 1 to interact with the photon ( always possible because of the antisymmetry of the wave function ) .
the interaction current is @xmath295 , and assumes a comparatively simple form if the kernel is a sum of single particle exchanges .
this case is illustrated in fig .
[ feynmancurrent ] .
current conservation , @xmath296 follows automatically @xcite from the bound state equation ( [ bound1 ] ) if the nucleon and interaction currents satisfy the following two - body ward - takahashi ( wt ) identities @xmath297\,\bigl\{g^{-1}(k;p_0)-g^{-1}(k';p'_0)\bigr\ } \nonumber\\ q_\mu \;i^\mu(k'_1,k'_2;k_1,k_2)&&= { \cal v}(k',k;p'_0)\,{\textstyle{1\over2}}[1+\tau_3 ] - { \textstyle{1\over2}}[1+\tau_3]\,{\cal v}(k',k;p_0)\ , .
\label{current3 } \end{aligned}\ ] ] note the appearance of @xmath298 $ ] , the isoscalar charge operator in isospin space .
the @xmath299 identity is the two - body version of the familiar one - body wt identity @xmath300 with @xmath301 ( 0 ) for the proton ( neutron ) and the undressed nucleon propagator normalized to @xmath302 .
note that the constraint on the interaction current is _ not _ zero ( and hence the interaction current is _ not _ zero ) if the kernel depends on the isospin or the total four - momentum @xmath303 . to fully specify a propagator dynamics , one must choose a propagator , @xmath281 , a kernel , @xmath280 , and current operators @xmath294 and @xmath304 .
four different progagators have been used in the study of the deuteron form factors .
the bethe - salpeter ( bs ) equation @xcite uses a fully off shell propagator for two nucleons @xmath305
\left [ e_p^2-\left({w\over2}-p_0\right)^2\right]}}\ , , \end{aligned}\ ] ] where @xmath306 is the ( off - shell ) positive energy projection operator and the right hand expression is the propagator for identical particles in the rest frame ( with @xmath307 , and @xmath308 used as a shorthand for the four - vector @xmath52 in the rest frame ) .
this choice of relativistic equation was the first to be introduced and is perhaps the best known .
it retains the full integration over all components of the relative four - momentum @xmath236 , and all of the off - shell degrees of freedom ( 2 for spin @xmath155 2 for `` @xmath223-spin '' , where @xmath309 are positive energy @xmath83 spinor states and @xmath310 are negative energy @xmath311 spinor states ) of both of the propagating nucleons , for a total of @xmath312 spin degrees of freedom .
the equation has inelastic cuts arising from the production of the exchanged mesons ( when energetically possible ) and additional singularities when the nucleons are off - shell .
these can be removed by transforming the equation to euclidean space .
however , the bs equation , when used in _ ladder _ approximation , does not have the correct one - body limit .
numerical comparisons of solutions obtained from the sum of _ all _ ladder and crossed ladder exchanges with ladder solutions of the bs equation , carried out for scalar theories , have shown that the ladder sum is inaccurate and that the one - body limit requires inclusion of crossed exchanges @xcite .
the bs equation has been solved in ladder approximation by tjon and his collaborators @xcite , and used to calculate the deuteron form factors @xcite .
the fits to the @xmath7 phase shifts originally obtained from these works are unsatisfactory by today s standards .
the spectator ( or gross ) equation ( denoted by @xmath17 ) @xcite restricts one of the two nucleons to its positive energy mass - shell .
if particle one is on - shell , the spectator propagator is @xmath313 where @xmath314 .
this has the effect of fixing the relative energy in terms of the relative three - momentum so as to maintain covariance and reduce the four dimensional integration to three dimensions [ @xmath315 in eq .
( [ rel1 ] ) ] .
identical particles are treated by properly ( anti)symmetrizing the kernel .
the restriction of one of the particles to its positive energy mass - shell also removes the @xmath310 states of one of the nucleons , reducing the number of spin degrees of freedom to 2@xmath1554=8 .
a primary motivation and justification for this approach is that it has the correct one - body limit , and the three body generalization satisfies the cluster property @xcite .
the equation also has a nice nonrelativistic limit that can be easily interpreted .
numerical studies of scalar field theories @xcite show that the exact ladder and crossed ladder sum is better approximated by the ladder approximation to this equation than it is by the ladder approximation to the bs equation .
the method can be extended to include gauge invariant electromagnetic interactions @xcite .
its principle drawback is that the kernel has unphysical singularities which can only be removed by an _
ad - hoc _ prescription .
results from this method will be reviewed in the next section .
the internal momentum integration can also be restricted to three dimensions in such a way that , for equal mass particles in the rest frame , the relative energy is zero and the particles are equally off shell .
the blankenbecler - sugar - logunov - tavkhelidze ( bslt ) equation @xcite can be defined so that the two propagating particles are on their positive energy mass - shell , reducing the number of spin-@xmath223-spin variables to 2@xmath1552=4 . however
, this equation does not satisfy the cluster property .
the approach of phillips , wallace , and mandelsweig @xcite , which we denote by pwm , also puts the particles equally off - shell , but includes all negative energy contributions . setting @xmath315 in eq .
( [ rel1 ] ) , the pwm propagator for equal mass particles in the c.m .
system is @xmath316 with @xmath312 spin degrees of freedom .
this propagator differs from bslt primarily by the presence of the additional @xmath70 term [ which contributes the last two terms in the curly brackets involving the `` mixed '' @xmath317 projection operators ] that includes contributions from crossed graphs approximately , and correctly builds in the one body limit .
the retarded kernel to be used with this propagator , in ladder approximation , is @xmath318 perhaps the principle obstacle to implementing this method is that construction of current operators is problematic , and manifest poincar invariance is lost ( but wallace @xcite has recently shown how to compute boosts for scalar particles exactly ) .
calculations using this method will be described in next section .
we now turn to a description of two examples of propagator dynamics .
_ van orden , devine , and gross _ [ vog ] .
the spectator equation has been used to successfully describe @xmath7 scattering and the deuteron bound state @xcite , and this work uses these results to describe the deuteron form factors @xcite . the relativistic kernel used to describe
the @xmath7 system consists of the exchange of 6 mesons [ @xmath319 , @xmath320 , @xmath149 , @xmath321 ( or @xmath322 ) , @xmath223 and @xmath323 .
the model includes a form factor for the off - shell nucleon @xcite , giving a `` dressed '' single nucleon propagator of the form @xmath324 where @xmath325 is one of the parameters of the model .
coupling constants and form factor masses ( 13 parameters in all ) are determined by a fit to the data and the deuteron wave functions are extracted @xcite . to insure current conservation
, the one - nucleon current must satisfy the ward - takahashi identity @xmath326 and this requires an off - shell modification of the single nucleon current .
the solution used by vog is @xmath327 where @xmath328 but is otherwise undefined [ in the applications described below , @xmath329 where @xmath191 is the dipole form factor of eq .
( [ dipole ] ) ] , and @xmath330 + { h(p'^2)\over h(p^2 ) } \left[{m^2-p^2\over p'^2-p^2 } \right]\nonumber\\ g_0(p',p)=\left({h(p^2)\over h(p'^2)}-{h(p'^2)\over h(p^2)}\right ) { 4m^2\over p'^2-p^2}\ , .
\end{aligned}\ ] ] while the on - shell form of the current ( [ onej ] ) is fixed by the nucleon form factors , and the functions @xmath331 and @xmath332 are fixed by the wt identity ( [ wt ] ) , other aspects of the off - shell extrapolation of the current ( [ onej ] ) are _ not unique_. using this one - nucleon current , and recalling that there are no currents of the type shown in figs . [ mec](a ) or ( b ) [ we postpone discussion of the @xmath268 current ] , it was shown @xcite that the full two body current to use with the spectator equation is given by the diagrams shown in fig .
[ scurrent ] .
these diagrams are manifestly covariant , and _ automatically include effects from z - graphs or retardation _ illustrated in figs .
[ mec](c ) or ( d ) .
they are referred to as the complete impulse approximation ( cia ) to distinguish them from the relativistic impulse approximation ( ria ) , an approximate current used in earlier calculations @xcite .
the ria is obtained by multiplying diagram [ scurrent](a ) by two , and is very close to the cia . _
phillips , wallace , divine , and mandelsweig _ [ pwm ] .
this work is based on the mandelsweig and wallace equation @xcite , supplemented by contributions from the crossed graphs @xcite , as described above .
it is sometimes referred to as the equal - time approach .
a feature of this equation is that it includes the full strength of the @xmath267-graphs ; the @xmath317 contributions shown in eq .
( [ pweq ] ) are roughly twice as strong as the @xmath267-graph contributions included in the spectator equation .
the pwm propagator is also explicitly symmetric , a convenience when applied to identical particles . in the published work reviewed here @xcite ,
the deuteron is described by a one boson exchange force using the parameters of the bonn - b potential with the exception of the @xmath149 meson coupling , which is adjusted to give the correct deuteron binding energy .
lorentz invariance is broken by the approximation ; the boosts of the deuteron wave functions from their rest frames are treated approximately .
the pwm current is a modification of eq .
( [ current1 ] ) . in the present work retardation effects [ like those illustrated in fig .
[ mec](d ) ] are omitted from the current operator ; only one body terms and @xmath267-graph contributions are included .
we now turn to a discussion of the other major approach detailed in fig .
[ decision ] : hamiltonian dynamics .
approaches based on hamiltonian dynamics start from a very different point than propagator dynamics , and this is one reason it is difficult to compare the two . while propagator dynamics starts from field theory ( which can be described as a quantum mechanics with an arbitrary number of particles ) , hamiltonian dynamics starts from quantum mechanics with a fixed number of particles . for a detailed review ,
see ref .
@xcite .
quantum mechanics begins with a hilbert space of states defined on a fixed space - like surface in four - dimensional space - time .
the various options for choosing this space - like surface were classified by dirac in 1949 @xcite .
the _ instant - form _ corresponds to choosing to construct states at a fixed time @xmath333 , and is the choice usually made in elementary treatments .
alternatively , _
front - form _ quantum mechanics constructs states on a fixed - light front , customarily defined to be @xmath334 .
( we use units in which the speed of light , @xmath335 , is unity . ) more generally , the light - front may be chosen in any direction defined by @xmath336 , with @xmath337 and @xmath338 . finally , _
point - form _ quantum mechanics constructs states on a forward hyperboloid , with @xmath339 ; @xmath340 [ the limiting cases of @xmath341 gives the instant - form ( with @xmath342 ) , and @xmath343 the front - form ] .
these three surfaces are shown pictorially in fig .
[ qmsurfaces ] .
[ while the point @xmath344 , @xmath345 is not on the hyperbolid , all distances between points on the hyperbolid are space - like . ]
the poincar transformations are symmetries that leave all probabilities unchanged ; they must be unitary transformations ( with hermitian generators ) on the space of quantum states .
the 10 generators of the full poincar group are the hamiltonian , @xmath346 , generator of time translations , three - momenta , @xmath347 , generators of spatial translations , angular momenta , @xmath348 , generators of rotations , and @xmath349 , generators of boosts .
they satisfy the following commutation relations : @xmath350=[h , j^i]=[p^i , p^j]=0\ , , \qquad[j^i , x^j]=i\epsilon_{ijk } x^k\ , , { \rm for}\ ; x^i = j^i , p^i , k^i\nonumber\\ \fl&&[k^i , k^j]=-i\epsilon_{ijk}j^k\ , , \qquad[k^i , p^j]=-i\delta_{ij}h\ , , \qquad[k^i , h]=-ip^i \label{ccr1}\end{aligned}\ ] ] for each of the forms of quantum mechanics there is a subgroup of the poincar transformations that leave the states invariant on the fixed surface associated with that form .
this is the kinematic subgroup , and the transformations in this subgroup will not depend on the dynamics ( since the dynamics describe how the states change away for the fixed surface ) . in the _ instant - form _ , space translations and rotations clearly leave the surface @xmath344 invariant .
generators of these transformations form a subgroup of the poincar group , with commutation relations @xmath351=0\ , , \quad[j^i , j^j]=i\epsilon_{ijk } j^k\ , , \quad[j^i , p^j]=i\epsilon_{ijk } p^k \ , .\label{ccr2}\end{aligned}\ ] ] transformations of states under these transformations will not depend on the dynamics .
the hamiltonian @xmath346 carries the states away from the initial fixed @xmath344 surface , and contains the dynamics .
the other three generators ( the boosts ) will also , in general , depend on the dynamics because their commutators involve @xmath346 .
the _ front - form _ surface @xmath352 is left invariant by translations in the @xmath353 , @xmath354 , and @xmath355 directions [ the generator of translations in the @xmath355 direction is @xmath356 because @xmath357 .
it is also left invariant by rotations and boosts in the @xmath277 direction , and by the generalized boosts @xmath358 and @xmath359 .
these 7 generators form a subgroup of the poincar group , with commutation relations @xmath360=[p^i , h^+]=[j^z , h^+]=[e^i , e^j]=[e^i , h^+]=[j^z , k^z]= [ k^z , p^i]=0 \nonumber\\ \fl&&[j^z , p^i]=i\epsilon_{ij}p^{j}\ , , \quad[j^z , e^i]=i\epsilon_{ij}e^{j}\ , , \quad [ e^i , p^j]=-i\delta_{ij}h^+\nonumber\\ \fl&&[k^z , h^+]=-ih^+ \ , , \quad[k^z , e^i]=-ie^i\ , , \label{ccr3}\end{aligned}\ ] ] where @xmath361 and @xmath362 , and @xmath363 , @xmath364 , and @xmath365 .
the fact that the front - form kinematic subgroup includes _ seven _ generators , including the boost @xmath366 and generalized boosts @xmath367 , makes the front - form popular . but a principle motivation for using the front - form is that it is a natural choice at very high momentum , where the interactions single out a preferred direction ( the beam direction ) and the dynamics evolves along the light - front in that direction .
the disadvantage is that the generators that contain dynamical quantities are @xmath368 and @xmath348 , and this means that angular momentum conservation must be treated as a dynamical constraint .
finally , the _ point - form _ hyperbolid is left invariant by the homogeneous lorentz group itself , with commutation relations @xmath369=i\epsilon_{ijk } j^k\ , , \quad[j^i , k^j]=i\epsilon_{ijk}k^k\ , , \quad [ k^i , k^j]=-i\epsilon_{ijk}j^k\ , .
\label{ccr4}\end{aligned}\ ] ] the hamiltonian and the momentum operators @xmath347 all carry the dynamical information .
we see that each of the forms of quantum mechanics has a different set of kinematic generators , and in no case are they all kinematic .
practitioners of hamiltonian dynamics sometimes speak as if it were impossible to treat the full poincar group kinematically .
this is true only in the context of hamiltonian dynamics ; _ all _ of the generators are kinematic in the bs or spectator forms of progagator dynamics .
dynamics is introduced whenever the states are propagated away from the surface on which they are initially defined . as in normal quantum mechanics
, the deuteron will be an eigenstate that propagates in `` time '' without loss of probability ; it will be an eigenstate of the generalized hamiltonian . in the instant - form ,
the rest state @xmath370 is an eigenstate of the momentum operators @xmath371 and the bound state equation in the rest frame is @xmath372 in the front - form the rest state @xmath373 is an eigenstate of the operators @xmath374 and @xmath356 @xmath375 and the dynamical bound state equation is @xmath376 finally , in point - form the rest frame eigenfunction must satisfy the four dynamical equations @xmath377 where @xmath378 and @xmath379 . in applications , the dynamical equations ( [ instant ] ) , ( [ front ] ) and the @xmath380 component of ( [ point ] )
can all be taken to be the nonrelativistic schrdinger equation in the rest frame , so the same nonrelativistic phenomenology can be used for any of these forms of mechanics . to complete the calculation of the deuteron form factors using hamiltonian dynamics one must choose a current operator that conserves current , and construct the proper matrix elements of this operator between deuteron wave functions .
this will be discussed next .
the steps taken to construct the current and calculate the form factors depend on the form of quantum mechanics used , and the taste of the investigator involved . here
we briefly describe recent work by five groups .
_ forest , schiavilla , and riska _ [ fsr ] : based on the work of schiavilla and riska @xcite , forest and schiavilla @xcite have done an instant - form calculation of the deuteron form factors . the original work of ref .
@xcite used on - shell matrix elements of the one body charge and current operators @xmath381u(p , s)\label{1body1}\end{aligned}\ ] ] expanded in powers of @xmath382 . here
@xmath383 and @xmath384 are the dirac and pauli form factors , usually replaced by the familiar charge and magnetic form factors @xmath385 in the recent unpublished work of ref .
@xcite the calculations have been done in momentum space , where the one body current operators have been evaluated without making any @xmath386 expansions , the relativistic kinetic energy @xmath387 has been used in place of the usual nonrelativistic expansion @xmath388 [ with the parameters of the av18 potential refitted ] , and the boost corrections to the deuteron wave functions have been included .
this work also includes two - body charge operators from @xmath319 and @xmath223 exchange using methods developed by riska and collaborators @xcite .
_ arenhvel , ritz , and wilbois _ [ arw ] : this recent calculation @xcite does a systematic @xmath382 expansion of relativistic effects that arise from the one body current operator and from contributions from meson exchanges .
the current operator ( [ 1body1 ] ) is approximated by @xmath389 & $ \mu=0$\cr f_1\,{({\bf p}'+{\bf p})^i\over2 m } + g_m\ , { i\,[\,\sigma\times { \bf q}\,]^i\over2 m } + { \cal o}\left[\left({v / c}\right)^3\right ] & $ \mu = i$ } \label{1body}\end{aligned}\ ] ] where @xmath390 or @xmath391 ( with appropriate @xmath383 and @xmath384 ) , @xmath149 is the operator in the nuclear spin space , and @xmath392 is the _
three_-momentum transferred by the electron .
this charge operator is correct to order @xmath174 , and the @xmath393 contribution to the current operator is given in ref .
the @xmath394 correction term to the charge operator is referred to as the darwin - foldy term .
the @xmath395 is the spin - orbit term .
there are ambiguities in all calculations based on expansions in powers of @xmath174 .
one ambiguity arises from the fact that the square of the three - momentum , @xmath396 , depends on the frame in which it is evaluated . in the breit frame , @xmath397 , while in the center of mass of the final deuteron ( the frame preferred by arw ) , @xmath398 ( recall that @xmath320 was defined in sec .
[ dffdef ] ) .
a second ambiguity surrounds the choice of @xmath399 versus @xmath383 . some experts @xcite advocate using @xmath399 ( because it is the correct charge operator ) in place of @xmath383 .
the difference between @xmath383 and @xmath399 is of higher order .
these ambiguities introduce theoretical uncertainty into any calculation .
the size of this uncertainty depends on both the value of @xmath62 and the choice of nucleon form factors ; for example the difference between using @xmath383 or @xmath399 can be inferred from the ratios shown in fig .
[ ges - study ] and is large for the recently measured jlab form factors and small for the mmd form factors .
uncertainties of this kind do not arise if the calculation is done covariantly , or to all orders in @xmath174 .
arw also include boost corrections originally derived to lowest order in @xmath382 by krajcik and foldy @xcite .
the boost corrections can be written as an operation on the wave function of the form @xmath400 where @xmath401 with @xmath402\cdot { \bf p}\over8m^2 } \label{boost1 } \end{aligned}\ ] ] the boost associated with the kinetic energy and the spin and @xmath403 the boost associated with the potential . in ( [ boost1 ] ) , @xmath404 and @xmath405 are the relative coordinate and relative momentum of the nucleon pair , and @xmath406 is the three - momentum of the moving deuteron .
arw use the values of @xmath403 worked out by friar @xcite , and also include relativistic effects from retardation , isobar currents , and meson exchange . to evaluate the latter a meson exchange model
is needed , and arw use the interactions and parameters of the bonn obepq potentials [ only results from the obepq b potential are presented in the next section , although ref .
@xcite includes results from all three obepq potentials ] .
friar has emphasized that relativistic effects can be moved in and out of the wave functions and currents by unitary transformations @xcite , so that all of these effects are ambiguous unless fully defined by the theory .
effects due to pair currents or recoil corrections , shown in fig . [ mec](c ) and ( d ) , do not appear to be included .
arw state that their calculations should be good only up to @xmath407 1.2 gev@xmath42 .
corrections to the charge operator to order @xmath174 obtained from instant - form dynamics and from the spectator form of propagator dynamics have been compared @xcite . in the cases
studied , the same total result was obtained from the sum of _ all _ of the corrections , but the individual terms in the sum were found to have a very different form even when they appeared to come from the same physical effects . _ carbonell and karmanov _ [ ck ] : in this front - form calculation @xcite the direction of the light - front [ denoted by @xmath408 where @xmath409 is treated as an unphysical degree of freedom .
wave functions and amplitudes may depend on @xmath410 but only those components of scattering matrix elements independent of @xmath410 will be physical . it is argued that this approach will give an explicitly covariant front - form mechanics @xcite . when applied to the deuteron form factors there are 11 spin invariants , three that are physical and 8 that depend on @xmath410 and are unphysical . in an exact calculation the 8 unphysical invariants would be zero , but in approximate calculations , such as that carried out in ref .
@xcite , they will not be zero .
the deuteron form factors can be extracted from the three physical invariants by projecting them from the general result , as derived in ref .
@xcite . for the choice @xmath411 ( corresponding to choosing the front - form surface @xmath352 ) this method shows that the charge and quadrupole form factors can be extracted from the @xmath412 component of the current ( in common with other treatments ) , but also shows that the magnetic form factor _ can not be obtained only from this component _ and requires a different projection ( and includes contributions from contact terms ) .
the rules for a general graph technique for calculating amplitudes in this formalism are given in ref .
@xcite . using this method
the deuteron wave function will in general have 6 components , only three of which have been found to be numerically large .
in addition to the familiar @xmath17 and @xmath18-state components , the third large component is proportional to a new scalar function @xmath413 , and adds the term @xmath414 to the deuteron wave function displayed in eq .
( [ ft1 ] ) ( to obtain this form we renormalized the expression in ref .
@xcite so that @xmath415 and used the transformations in ref .
@xcite ) . in ref .
@xcite @xmath413 is calculated perturbatively using the bonn potential from ref .
@xcite without change of parameters .
they find that @xmath413 is the largest of the three components for all momenta greater than 500 mev , and believe that the perturbative estimate is accurate to about 20% .
the physical meaning of the @xmath413 contribution has been studied in threshold deuteron electrodisintegration , where it contributes about 50% of important pair term contributions . _
lev , pace , and salm _ [ lps ] : the lps @xcite calculation is a recent version of a series of light - front calculations that have assumed the light - front is fixed ( at @xmath334 ) . in the past , calculations with fixed light fronts have run into a problem with the loss of angular momentum conservation , and before we review the lps results we will discuss this issue . in calculating form factors with fixed light fronts
it has been conventional to choose a coordinate system where @xmath416 and @xmath417 .
current conservation is then satisfied if only one component of the current ( @xmath412 ) is non zero .
consider the matrix elements of the deuteron current , @xmath418 , where @xmath419 ( @xmath420 ) are the helicities of the outgoing ( incoming ) deuterons .
one consequence of the loss of _ manifest _ rotational invariance is that there are _ four _ independent matrix elements of the @xmath412 current related by the constraint @xmath421 this is a dynamical constraint often referred to as the `` angular condition '' @xcite .
the deuteron form factors can be extracted from _ any _ choice of three of the matrix elements @xmath418 , and if condition ( [ angular ] ) is not satisfied each choice will yield different results .
the form factors will not be uniquely determined unless the angular condition is satisfied . to avoid ( or solve ) this problem , lps work in the breit frame , where @xmath422 and @xmath423 .
current conservation then requires that @xmath424 .
a current operator that satisfies these conditions was constructed in ref .
@xcite . for elastic scattering
this operator has the form @xmath425 where @xmath426 is the free ( one body ) current operator , and @xmath427 and @xmath428 are rotations by @xmath429 about the @xmath353 axis , @xmath427 in the vector space and @xmath428 in the spinor space .
note that the definition insures that @xmath424 as required by current conservation . using this current
, lps have calculated the deuteron quadrupole moment , @xmath263 , to 2% accuracy @xcite .
the calculation shown below in sec .
[ thyandexp ] uses mmd nucleon form factors and the nijmegen ii deuteron wave functions . _
allen , klink , and polyzou _ [ akp ] : the deuteron form factors have also been recently calculated using the point - form of quantum mechanics @xcite .
here there is no difficulity in writing down manifestly covariant matrix elements , but there is some ambiguity in deciding how to impose current conservation .
akp work in the breit frame , choose a one body impulse current to describe the @xmath430 , and 2 components of the current , and introduce a two body current @xmath431 ( which need not be calculated ) to insure current conservation .
new effects come from the way the wave functions are constructed in point form ( `` velocity '' states are constructed ) , and from the fact that momentum is now a dynamical generator , so that the momentum transferred to the nucleon _ inside _ the deuteron is not equal to the momentum transferred to the deuteron as a whole .
they argue that the momentum transferred to each nucleon inside the deuteron is @xmath432 at momentum transfers @xmath433 gev@xmath42 this is a 25% increase , and leads to a large suppression of the form factors .
this explains part of the decrease in the size of the form factors predicted by this model .
the results reported below use the mmd nucleon form factors and the av18 @xmath7 potentials .
we now turn to a brief review of methods using quark degrees of freedom .
calculations based on quark degrees of freedom must confront the fact that the deuteron is at least a six quark system .
since the six quarks are identical ( because of internal symmetries ) the system must be antisymmetrized , and it is not clear that the nucleon should retain its identity when in the presence of another nucleon .
how does the clustering of the six quarks into the two three - quark nucleons appear at large distance scales ? how do we treat the confining forces in the presence of so many quarks ?
the approach to these issues depends on whether of not @xmath62 is large enough to justify the use of perturbative qcd ( pqcd ) . at modest @xmath62
the momentum transferred by the gluons is small and the qcd coupling is too large for perturbative methods to be useful . in the nonperturbative regime calculations
must be based on models .
many papers have been written addressing these issues , and a complete review is beyond the scope of our discussion . here
we mention only two contributions that give the flavor of the discussion .
maltman and isgur @xcite studied the ground state of six quarks interacting through a @xmath434 potential previously used to explain the spectrum of excited nucleons , and found that there was a natural tendency for the quarks to cluster into two groups of three ( i.e. nucleons ) .
they obtained a reasonable description of the deuteron , and confirmed that the short range @xmath7 repulsion could be largely understood in terms of quark exchange .
later , de forest and mulders @xcite , using a very simple model , considered the effect of antisymmetrization on the structure of the form factor .
their calculations suggest that the zeros seen in form factors could be a consequence of antisymmetration alone .
they also show that the factorized form of the impulse approximation obtained from nonrelativistic ( and some relativistic ) theories , which gives the deuteron form factors as a product of a nucleon form factor and a nuclear ( or body ) form factor @xmath435 may not be a good description in the presence of antisymmetration .
when quarks are exchanged between nucleons it is no longer possible to separate the _ nucleon _ structure from the _ nuclear _ structure .
consideration of the quark exchange diagram shown in fig .
[ ffexchange ] suggests a factorization formula of the form @xmath436 ^ 2)\right]^2\times d_2(q^2)\ , .
\label{exff}\ ] ] because nucleons are composite and identical , either of the forms ( [ fac1 ] ) or ( [ exff ] ) ( or yet other relation ) might hold , and there is no clearly correct way to isolate the structure of the nucleon from the structure of the bound state . in model calculations these issues can be handled by separating the problem into two regions : at large separations ( @xmath437 ) it is assumed that the system separates into two nucleons interacting through one pion exchange , and at small distances ( @xmath438 ) the system is assumed to coalesce into a six - quark bag with all the quarks treated on an equal footing . in this review we report the results of a calculation by dijk and bakker @xcite , where references to other calculations of this type can also be found ( see also the work of buchmann , yamauchi , and faessler @xcite ) .
_ dijk and bakker _ [ db ] : this calculation is based on the quark compound bag model introduced by simonov @xcite . here
the six - quark wave function is assumed to be the sum of a hadronic part and a quark part .
the hadronic part is a fully antisymmetrized product of two three - quark wave functions , each with the quantum numbers of a nucleon , and a relative @xmath7 wave function @xmath439 @xmath440\ , , \ ] ] where @xmath441 is the effective internucleon separation .
the quark part is a sum of eigenstates @xmath442 of a confined 6-quark system @xmath443 where the confined 6-quark states are zero outside of a confining radius @xmath444 , which is a parameter of the calculation . in the applications , only one term @xmath445 needs to be included in the sum ( [ cbstates ] ) .
the dynamical quantities determined by the calculation are the @xmath7 wave function @xmath446 and the spectroscopic coefficient @xmath447 which is a function of the energy @xmath142 .
the @xmath7 scattering phase shifts and mixing parameters are determined by replacing the spectroscopic coefficients by boundary conditions on the surface @xmath444 and integrating the schrdinger equation for @xmath448 .
the paris potential @xcite is used to describe the @xmath7 interaction in the peripheral region and is set to zero in the inner region .
two models were developed ; in this review we report results from the fits to the arndt single energy 1986 @xcite solutions , which db denote qbc86 .
this fit finds @xmath449 fm .
calculation of the form factors requires an assumption about the form factor of the internal compound bag part of the wave function .
they use @xmath450 with @xmath451 gev obtained from a fit to the @xmath0 and @xmath1 structure functions .
results from this model are reported in sec .
[ thyandexp ] below .
if one believes the momentum transfer is high enough , perturbative qcd ( pqcd ) may be used to study the deuteron form factor and reactions .
here it is assumed that the problem naturally factors into a hard scattering process in which the momentum transfer is distributed more or less equally to all of the six quarks , _ preceeded and followed _ by soft , nonperturbative scattering that sets the scale of the interaction but does not strongly influence its @xmath62 behavior .
the @xmath62 behavior is therefore determined by the hard scattering , which can be calculated perturbatively . the formalism and method
are reviewed in the seminal papers by brodsky and farrar @xcite and lepage and brodsky @xcite .
these calculations of the form factor are all based on the diagrams shown in fig .
[ pqcddiagrams ] . in the hard scattering ,
the momentum transfer @xmath134 is distributed to the six quarks through the five hard gluon exchanges , the last of which carries a momentum of approximately @xmath452 .
if the spin factors are included with the quark propagators , the only large @xmath62 dependence comes from the @xmath453 of each gluon propagator , giving the _ counting rule _ for the hard scattering part @xmath454 @xmath455^{-(n_c-1)}\ , , \label{counting}\end{aligned}\ ] ] where @xmath456 is the number of constituent quarks ( 6 for the deuteron ) and @xmath457 the number of gluon propagators .
this leads immediately to the prediction that the leading contribution to the deuteron form factor should go like @xmath458 , or that @xmath459 .
the argument also shows that the perturbative result can not be expected to set in until @xmath460 to 1 gev , somewhere in the region of @xmath62 from 9 to 36 gev@xmath42 .
( all agree that pqcd must give correct predictions at sufficiently high @xmath62 , but how large this @xmath62 must be is a topic of considerable controversy @xcite . )
note that this simple argument does not set the scale of the form factor ; estimates can be obtained from detailed evaluation of more than 300,000 diagrams that contribute to the hard scattering @xcite .
it turns out that this leading twist pqcd estimate is 10@xmath461 10@xmath462 times smaller than the measured deuteron form factor , implying large soft contributions to the form factor , in agreement with @xcite , and suggesting that pqcd should not be used as an explanation for the form factor .
the calculation is extremely complicated and a confirmation , or refutation , is desirable .
perturbative qcd also predicts the _ spin dependence _ of the hard scattering , and these predictions provide a more stringent test of the onset of pqcd .
these spin dependent predictions have implications for the individual deuteron form factors , and these were first presented in ref .
@xcite , and further developed in refs .
application of these rules to hadronic form factors in general shows that * hadrons with an _ even _ number of quarks will be dominated by the _ longitudinal _ ( charge ) currents , while these with an _ odd _ number of constituents by _ transverse _ ( magnet ) currents .
hence , the dominant form factors at large @xmath62 should be the nucleon magnetic form factors and the deuteron charge ( or quadruple ) form factors . *
the dominant form factors at large @xmath62 are those that conserve helicity . when applied to the deuteron , these rules lead to the conclusion that the helicity amplitude @xmath463 [ c.f .
( [ dhelicity ] ) ] dominates at large @xmath62 ; the others are smaller by at least a power of @xmath134 . in particular
, this implies that @xmath464 , where @xmath465 is the mass scale above which the nonleading terms can be neglected .
hence @xmath466 unfortunately , this argument does not allow one to estimate the ratio @xmath467 , since @xmath1 is controlled by a different , independent helicity amplitude . in ref .
@xcite an attempt was made to improve on the constraint ( [ t20prediction ] ) .
these authors used the front - form , and evaluated the current in the light - front breit frame where the plus component of the momentum transfer @xmath468 . in this frame
all three deuteron form factors may be written in terms of matrix elements of the @xmath469 component of the current @xmath470 where @xmath471^{-1}$ ] .
perturbative qcd predicts that @xmath472 will dominate at large @xmath62 , and if this happens at a scale @xmath473 it follows from ( [ frontff ] ) that the form factors go in the ratio of @xmath474 .
this leads to a prediction for @xmath467 and to a prediction for @xmath475 that differs from ( [ t20prediction ] ) at moderate @xmath62 . however , rotational invariance is not manifest in the light front , and there are _ four _ nonzero components of the @xmath412 current corresponding to deuteron helicity combinations of @xmath476 , @xmath477 , @xmath478 and @xmath479 that are related by the angular condition discussed in sec .
[ hexamples ] above .
carlson has recently shown @xcite that the angular condition places strong constraints on the possible subleading behavior of the helicity amplitudes .
perturbative qcd predicts that the subleading amplitudes will go like @xmath480 where @xmath202 , @xmath481 , and @xmath335 are dimensionless constants of the order of unity , and @xmath465 is the scale at which pqcd begins working for the deuteron [ the @xmath281s in eq .
( [ brodg ] ) are identical to the @xmath482s in eq .
( [ angular ] ) ] .
assuming that @xmath483 ( but making no assumption about the size of @xmath465 ) the angular condition in leading order becomes : @xmath484 solution of this equation therefore requires that @xmath485 and @xmath486 . under these conditions the relations ( [ frontff ] ) again
produce _ only _ the result ( [ t20prediction ] ) . in this section
we compare theory with experiment and draw conclusions from this comparison .
our major conclusions will be restated and summarized again in sec .
[ conclusions ] below .
figure [ gc - relcorr ] shows how coulomb distortion of the incoming and outgoing @xmath3 plane waves effects the very low @xmath62 data ( extracted from ref .
@xcite ) for the charge form factor , @xmath70 .
the figure also compares this data with theory .
corrections for coulomb distortion change the deuteron radius from an apparant 2.113 fm ( as measured in @xmath3 scattering ) to 2.130 fm ( after the correction ) @xcite . to remove the distortions from the data of ref .
@xcite , we adjust @xmath70 by @xcite @xmath487 note that this decreases @xmath70 at very small @xmath134 giving a larger deuteron radius , but increases @xmath70 where the data have been extracted .
the figure shows both the uncorrected and the coulomb corrected data normalized to the nonrelativistic av18 calculation with the mmd nucleon form factor .
note that the difference between the coulomb corrected and uncorrected data is about half of the experimental error at @xmath488 0.5 gev .
the figure also shows the size of relativistic and interaction current corrections that arise from the instant - form calculation arw of ref .
@xcite , the front - form calculation lps of ref .
@xcite , the point form calculation akp of ref .
@xcite , and the cia and ria approximations from ref .
these calculations were discussed in sec .
[ pexamples ] and [ hexamples ] above . at the scale of the current experimental accuracy ( a few percent ) ,
the relativistic treatments _ differ noticeably_. they also differ from the eft calculation ( shown previously in fig .
[ fig : eft ] ) which drops sharply below the data for @xmath489 0.2 gev .
it is important that these calculations be systematically compared and the different physical content of these approaches be isolated and understood . in particular
, it would be very interesting to know why the covariant cia and ria have more positive corrections than those obtained from the hamiltonian forms of dynamics . @llll model & dynamics & description & consistent current + vog @xcite & propagator & spectator & yes + pwm @xcite & propagator & modified mandelsweig - wallace & no + fsr @xcite & hamiltonian & instant - form ; no @xmath382 expansion & yes + arw @xcite & hamiltonian & instant - form ; with @xmath382 expansion & yes + ck @xcite & hamiltonian & front - form ; dynamical light - front & no + lps @xcite & hamiltonian & front - form ; fixed light - front & no + akp @xcite & hamiltonian & point - form & no + db @xcite & nonrelativistic & quark - cluster & yes + the high @xmath62 predictions for the eight models reviewed in secs .
[ pexamples ] , [ hexamples ] , and [ npqcd ] are shown in figs .
[ alowq27][rabt20 ] .
the models are summarized in table [ tab : theories ] .
these calculations give very different results .
figure [ alowq27 ] shows the predictions for @xmath490 , with the model dependent @xmath268 exchange current intentionally omitted from all of the calculations .
all of the models except the akp point - form calculation give a reasonable description of @xmath0 out to @xmath491 gev@xmath42 , beyond which they begin to depart strongly from each other and the data .
taking into account that the @xmath268 exchange current _ could be added to any of these models , and that this contribution tends to increase a above @xmath491 gev@xmath42 _
( for the sign of the @xmath268 coupling constant used in the discussion in the following paragraph ) , four models seem to have the right general behavior : the vog , fsr , arw and the quark model of db ( but there are no results for this model beyond @xmath492 gev@xmath42 ) .
ironically , none of the models favored by the high @xmath62 data does as well at low @xmath62 as the three `` unfavored '' models shown in the right panels ( unless the platchkov @xcite data are systematically too low ) .
another possibility suggested by effective field theory @xcite is that the asymptotic normalization of the relativistic deuteron wave functions is incorrect , and that a small adjustment in @xmath7 parameters to insure a good value for this constant would correct the problem .
figure [ jlaba ] shows the effect of the new jlab measurements of the nucleon form factors on predictions for @xmath490 .
these new form factors will _ decrease _ predictions for @xmath0 for momentum transfers in a region around @xmath492 gev@xmath42 ( by a factor of 2 for the model shown ) , increasing the descrepancy between predictions and the data .
[ however , it may improve the prediction for those models ( pwm , ck , and lps ) that are currently too large in this region . ]
the figure also shows how the @xmath268 exchange current could increase predictions at large @xmath62 .
the difference between the solid and long dashed lines is due largely to the effect of the @xmath268 exchange current ( but is also do in small part to the fact that the cia result is slightly smaller that the ria ) .
unfortunately , the size of the @xmath268 exchange current is very sensitive to the @xmath268 form factor , as discussed in ref .
@xcite , and could even be too small to see at these momentum transfers ( if the current estimates of the @xmath268 form factor are too large ) .
this is our reason for insisting that this contribution should be viewed as new physics not readily predictable within a meson model .
finally , fig .
[ rabt20 ] shows the predictions for the structure functions @xmath0 , @xmath1 , and @xmath2 for the eight models discussed .
the lps calculation shows a large descrepency with the @xmath2 data , but the most striking feature of these plots is the _ large model dependence _ of the predictions for @xmath493 .
the magnetic structure function provides the most stringent test , and the predictions are comparatively free of the @xmath268 exchange current ( which gives only a small contribution to @xmath1 ) .
examination of the figure shows that the @xmath1 predictions of the pwm , arw , akp , ck models fare the worst . in all , taking the predictions for the three structure functions together , the best results are obtained with the fsr , vog , and db models .
we conclude our comparison with theory by extending our discussion somewhat beyond the limits of previously published papers . to make the point clearly ,
focus on the vog calculation using the spectator equation , and recall that current conservation _ required _ that the single nucleon current in this approach , eq .
( [ onej ] ) , include a new form factor , @xmath494 .
this form factor must satisfy the constraint @xmath328 , but is otherwise _
completely unspecified_. in the published vog calculation @xcite and in all of the plots shown so far , this form factor was taken to be the standard dipole , @xmath191 .
figure [ fig : f3 ] shows how agreement between theory and experiment can be significantly improved by choosing a different form for the unknown form factor @xmath495 .
this figure shows three theoretical predictions : ( i ) the `` standard '' vog ria prediction with @xmath329 and _ no _ @xmath268 exchange current , ( ii ) another model with _ no _ @xmath268 exchange current , but with a tripole @xmath495 of the form @xmath496 and ( iii ) a model with @xmath329 and a dipole @xmath268 form factor @xmath497 in both form factors @xmath62 is in gev@xmath42 .
these form factors are shown in the bottom left panel of the figure .
figure [ fig : f3 ] shows that a good agreement between theory and experiment can be obtained with the tripole @xmath495 without any @xmath268 exchange current , and that to some extent the @xmath268 exchange current can be substituted for a hard @xmath495 ( although @xmath495 is better than @xmath498 at improving all three observables simultaneously ) .
we see that we have , in some sense , achived the goals of a theory of elastic @xmath3 scattering based on nucleon degrees of freedom . _ with small adjustments of unknown form factors associated with short range physics
, the nn theory can describe all three form factors quite well_. it seems likely that any nucleon model with a _ consistent and complete _ description of the current ( c.f.table [ tab : theories ] ) can do as well .
the reasonable results obtained from the fsr and db models are probably due to the fact that they have consistent , complete currents not based on an expansion in powers of @xmath174 ( which must fail at high @xmath62 ) . from the discussions in sec .
[ thyandexp ] above , it is clearly of interest to extend measurements of @xmath0 to higher @xmath62 .
an @xmath3 coincidence experiment is straightforward , but prohibitive timewise with present accelerators .
the proposed 12 gev jlab upgrade allows one to take advantage of the approximate @xmath499 scaling of @xmath90 at constant @xmath62 and high energy @xcite .
a large acceptance spectrometer such as mad would be very helpful .
depending on the details of the upgrade , a one month experiment could provide data to @xmath62 of 8 gev@xmath42
. it might also be desirable to do a new , high precision experiment at _ low _ @xmath62 .
the goal of this experiment would be to resolve the discrepancy between the data sets of refs .
@xcite and @xcite , and to check the low @xmath62 limit of the relativistic calculations .
these measurements require little time , but do require excellent control of systematic uncertainties , at the level of 1 2% , if they are to be meaningful .
a hasty examination of fig . [ rabt20 ] might lead one to believe that the problems with @xmath1 are mainly theoretical , and that there is no need for new data .
we believe this attitude would be inappropriate for two reasons .
first , some of the calculations shown in the figures are in early stages of development , and will improve before any new data are available .
the number of calculations is a reflection of the challenge that the high @xmath62 data present , and a reflection of interest in , and knowledge of relativistic methods that is emerging from the study of these measurements .
second , and most important , measurements of @xmath1 vary by 5 orders of magnitude .
the existence and position of the first minimum has not yet been firmly established , and the location and existence of a possible second minimum is unknown .
these minima in @xmath1 ( if they exist ) result from cancellations of various physical effects and provide a very precise test of any theory . as indicated above , measurements of @xmath1 do not require high energy , but do require large scattering angles , as close as possible to 180@xmath137 .
at 180@xmath137 , beam energies of 1 to 2 gev cover a @xmath62 range from 1.4 gev@xmath42 to 5 gev@xmath42 .
the slac ne4 experiment was a heroic effort , run with @xmath3 coincidences .
energy resolution was limited by thick targets , 20 - 40 cm long .
there was a large but manageable background of @xmath500 in the @xmath501 spectrometer , about 3 or 4 to 1 , and an extremely large proton background from @xmath502 in the deuteron channel - up to two events per beam pulse were seen in the worst kinematics . since ne4 ran at beam currents of up to 50 ma instantaneous , with a duty factor of about 0.3 @xmath155 10@xmath147 , corresponding to an average current of about 15 @xmath153a , the jlab continuous beam structure essentially eliminates random coincidence backgrounds such as these .
background coincidence reactions included @xmath503 and @xmath504 at 180@xmath140 , with the photon producing an electron detected in the electron channel .
the spectrum of these photoreactions ends near the elastic peak , allowing the background contributions to be fit and determined .
a jlab experiment would run with both larger spectrometer acceptance and higher luminosity to increase rates , but with a shorter target to reduce these backgrounds .
the proposed configuration @xcite would use the hall a septum magnets to detect the forward - going deuterons at angles of 3 - 6@xmath140 , along with special electron channels to detect scattered electrons at about 160 - 170@xmath140 .
based on the slac ne4 cross sections , a one month experiment can map out @xmath1 to about 6 gev@xmath42 .
extending @xmath6 or other polarization observables to higher @xmath62 is quite difficult @xcite .
there are three obvious possibilites .
recoil tensor polarimetry requires a well calibrated polarimeter with a large figure of merit , but no such device exists .
for example , polder , used in the jlab hall c @xmath6 experiment , relied on the @xmath505 reaction , for which the figure of merit decreases at larger energies , leading to a practical upper limit in @xmath62 , that was reached in e94 - 018 .
hypomme @xcite is promising , but not well enough calibrated .
the combination of polarized electron beams with recoil vector polarimeters is an untested possibility @xcite . with @xmath0 , @xmath1 , and
@xmath71 known , @xmath104 is calculable and can calibrate the polarimeter analyzing power , while @xmath506 determines the form factor combination @xmath507 .
( the ratio of the two polarization components depends on this combination of form factors , times kinematic factors and divided by @xmath71 . ) since the jlab polarized source can provide 50
100 @xmath153a beams , there is no luminosity problem .
the difficulty with this measurement is that the polarization components are expected to be small in the @xmath62 range of interest , @xmath151 0.01 , and @xmath508 @xmath509 5 10 . a one month measurement for one @xmath62 of @xmath151 2 - 2.5 gev@xmath42 can determine the _ polarization components _ well , but
if these are small as expected , the _ form factors _ can only be extracted with factor of two uncertainties .
an alternative is to use asymmetries from a polarized target .
however , the reduced currents that can be used with polarized targets require a large acceptance detector such as clas , to make up for the lack of luminosity . also , although current polarized targets have moderately large deuteron vector polarizations , tensor polarizations are small .
furthermore , the asymmetry varies as @xmath510 , with @xmath511 the polarization direction , so it is desirable to have complete azimuthal coverage , with @xmath511 in the direction of @xmath512 at one azimuthal angle , rather than being purely transverse .
extensive beam time would be needed , either as an external polarized target experiments at jlab , or as an internal polarized target experiment at hermes .
we note that a series of moderate @xmath62 measurements are planned with the mit bates blast detector @xcite , for @xmath62 from 0.1 to 0.9 gev@xmath42 .
proposed next generation colliders , such as epic / erhic , are promising due to large planned luminosities ; for this experiment the lower proposed c.m .
collision energies are desirable for ensuring exclusivity .
the spin direction of the polarized deuteron beam must be controllable . in collider kinematics ,
the scattered electron and deuteron energies are close to their respective beam energies and are slow functions of @xmath62 , while the scattering angles for fixed @xmath62 vary slowly with the beam energies .
thus , if an experiment is possible , it would attempt large azimuthal coverage of coincidence @xmath513 elastic scattering , with the outgoing particles at angles from a few to about 20@xmath137 from the beam line .
comparison of theory and experiment leads to the following conclusions : * nonrelativistic quantum mechanics ( without exchange currents or relativistic effects ) is ruled out by the @xmath490 data at high @xmath62 .
reasonable variations in nucleon form factors or uncertainties in the nonrelativistic wave functions can not remove the discrepancies .
* in some relativistic approaches using @xmath7 degrees of freedom only , short range physics not calculable within the model ( @xmath495 or @xmath498 , for example ) can be adjusted to give good agreement with all the data . *
different ways of calculating relativistic effects ( or meson exchange currents ) can give results that differ substantially from each other . even at low @xmath62 , where all calculations are constrained ,
these differences are larger than errors in the data .
this is not understood , but may be due to the failure of some models to use realistic currents . *
the deuteron form factors provide no evidence for the onset of perturbative qcd , but quark cluster models could explain the data .
study of the experimental situation leads to the following conclusions : * a good database of @xmath0 , @xmath1 , and @xmath6 measurements has been obtained ; while discrepancies exist they are generally not large enough to affect the theoretical interpretation . *
the minimum of @xmath1 is very sensitive to details of the models , and improved measurements of @xmath1 for @xmath62 in the region 1.5 - 4 gev@xmath42 are particularly compelling .
it is important to accurately map out the zero in the @xmath1 structure function .
* detailed disagreements between theories and different data sets suggests the need for precision studies at low @xmath62 .
as the deuteron has no excited bound states , inelastic scattering experiments have largely consisted either of ( i ) measurements in which the final state mass , w , is very close to @xmath514 ( referred to as _ threshold electrodisintegration _ even when @xmath62 is very large because the final state is close to the @xmath7 scattering threshold ) , ( ii ) measurements near the quasifree peak ( defined by the condition that the `` spectator '' nucleon remain at rest ) , or ( iii ) deep inelastic scattering in which both @xmath62 and @xmath308 become very large .
we will not discuss deep inelastic scattering in this review . for processes at modest energies near the quasielastic peak
, a rough estimate of the cross section can be made using the _ unrealistic _ plane wave impulse approximation ( pwia ) in which all final state interactions are ignored . denoting the momentum of the outgoing struck nucleon by @xmath515 , the cross section in pwia is proportional to @xmath516 where @xmath517 is some combination of the squares of the electric and magnetic form factors of the nucleon , and @xmath518 is an average of the square of the momentum space wave function of the deuteron with internal relative momentum @xmath405 .
were the pwia realistic , eq .
( [ pwia ] ) shows that inelastic scattering in quasielastic kinematics would provide a direct measure of the ( square ) of the deuteron wave function .
while the pwia is overly simplistic , it does illustrate ( correctly ) one of the central justifications for quasielastic measurements . within the context of a more realistic dynamical theory , one can use response function separations and polarization observables to enhance the sensitivity to various model dependent _ nonobservables _ , such as momentum distributions , meson - exchange currents , and medium modifications .
one strong recent interest has been to choose kinematics in which the unobserved nucleon has a large momentum ; the plane wave approximation shows that this configuration enhances sensitivity to initial - state short range correlations ( i.e. the wave function ) and possible quark effects .
a number of these experiments have been carried out at various accelerators , but no experiments at jlab have yet reported results .
thus , an experimental review of this topic is unwarrented at this time .
however , because photodisintegration , electrodisintegration , and threshold electrodisintegration are closely related theoretically , we present the theoretical background for these processes below . in this subsection
the cross section and polarization observables for electrodisintegration and photodisintegration to an @xmath519 final - state are reviewed briefly .
( we do not discuss pion and meson production . ) the electroproduction cross sections will be obtained first , and photoproduction will then be treated as a special case .
the most general decomposition of the @xmath520 coincidence cross section was first discussed by donnelly and raskin @xcite . here
we will follow the later work of ref .
the cross section can be shown to have the form ( eq . ( 95 ) of ref .
@xcite ) @xmath521 \nonumber\\ & & + s_{lt}\left[\cos\phi \tilde{r}^{({\rm i})}_{lt } + \sin\phi\tilde{r}^{({\rm ii})}_{lt}\right ] + 2 h \ , s_{t'}\tilde{r}^{({\rm ii})}_{t^{\prime } } \nonumber\\ & & + 2hs_{lt'}\left[\sin\phi\tilde{r}^{({\rm i})}_{lt^{\prime } } + \cos\phi\tilde{r}^{({\rm ii})}_{lt^{\prime}}\right]\biggl\}\ , , \label{crosssec}\end{aligned}\ ] ] where , in the lab frame , @xmath522 , and the nine response functions are functions of @xmath64 , @xmath523 , and the scattering angle @xmath524 of the final - state proton ( measured in a coincidence experiment and integrated over in an inclusive measurement ) .
the ejectile plane is tilted at an angle @xmath101 with respect to the electron scattering plane , as illustrated in fig .
[ fig_cross ] .
the mott cross section and other variables are as defined in eq .
( [ mott ] ) .
the electron kinematic factors are @xmath525 with @xmath526 .
one of the virtues of eq .
( [ crosssec ] ) is that the response functions @xmath527 are _ covariant _ , and hence ( [ crosssec ] ) it can be used to describe the cross section in either the c.m . of the outgoing @xmath519 pair or the laboratory frame ,
provided we use the appropriate form of @xmath528 : @xmath529 where @xmath530 is the lab recoil factor @xmath531 with @xmath308 the invariant mass of the outgoing pair . in any frame the nine response functions of ( [ crosssec ] ) are related to the components of the deuteron response tensor @xmath532 @xmath533 with x = i or ii , and @xmath534 all of these quantities are written in the helicity basis , with @xmath535 the helicity of the virtual photon , @xmath536 and @xmath537 the helicities of particles 1 and 2 in the final - state , and @xmath538 the helicity of the initial deuteron .
the matrix element of the helicity - basis current operator between helicity states is represented by @xmath539 . in cases
where the deuteron target might be polarized _ only _ in the @xmath540 direction , and where _ only _ the polarization of the outgoing particle 1 might be measured , the spin density matrices for particle 1 in the final state is given by @xmath541 and that of the deuteron in the initial state by @xmath542 , where @xmath543 where @xmath544 is the @xmath540 component of the spin - one matrix , @xmath545 is the vector polarization of the deuteron target , @xmath105 is the direction of the polarization of outgoing particle 1 , measured with respect to the @xmath546 coordinate system shown in fig .
[ fig_cross ] , and @xmath547 are the pauli matrices .
note that only those response functions of type i ( denoted by the superscript ) are nonzero if all of the hadrons are unpolarized ; type ii response functions require ( for the cases considered here ) measurement of the polarization of the outgoing nucleon .
further details and additional cases can be found in ref .
@xcite .
the familiar unpolarized inclusive cross section is easily obtained by integrating ( [ crosssec ] ) in the c.m . and summing over electron polarizations .
the result is @xmath548 with @xmath549 for real photons the longitudinal components are absent , and the cross section simplifies ( there is no electron scattering plane and no electron kinematics ) .
the most general polarization of the incoming photon , @xmath39 , is therefore a superposition of the circular polarization states @xmath550 , which we write as @xmath551 with @xmath552 .
the expansion coefficients @xmath553 can therefore be written in terms of only three independent parameters @xmath554 with @xmath555 .
the coincidence cross section for a polarized photon beam is then @xmath556 where @xmath557 and @xmath558 are the fractions of right circular and linear photon polarizations , respectively , and @xmath559 depends on the frame @xmath560 with @xmath530 defined in eq .
( [ defrecoil ] ) .
we will return to the cross section ( [ photonxsec ] ) in sec .
[ photodis ] .
it is important to appreciate that these formulae for the cross section are _ exact _ relativistic results ( subject only to the one photon exchange approximation ) .
all of our ignorance is confined to the hadronic matrix elements of the current @xmath561 and the structure functions ( [ rrela ] ) that are products of these currents . in much of the older literature , particularly for studies of the @xmath562 reaction from nuclei with mass number @xmath563 @xcite ,
the cross section is written @xmath564 where @xmath565 is a kinematic factor , @xmath566 is cross section for scattering of an electron from an `` off - shell '' proton , and @xmath567 is the proton spectal function ( which gives the probability of finding a proton with momentum @xmath236 and separation energy @xmath568 in the target nucleus ) .
the proton momentum distribution is obtained by integrating the spectral function over the separation energy @xmath569 some early experiements focused on `` measuring '' the momentum distribution and the spectral function . while this picture has a nice physical interpretation [ it is motivated by the pwia , eq .
( [ pwia ] ) ] , and presenting data this way is sometimes useful , particularly in the early phases of the program , it is important to realize that the individual structure functions that enter the exact cross section ( [ crosssec ] ) are , in general , _ independent _ functions which are _ not _ proportional to each other , and that therefore eq .
( [ spectral ] ) is _ only an approximation _ to the cross section @xcite .
attempts to refine the definitions of @xmath566 and @xmath567 can have limited value at best , and at worst can lead to many unproductive debates about the precise definition of the spectral function .
calculation of the hadronic current matrix elements ( [ currents ] ) is complicated by requirement that the current be conserved , @xmath570 . for elastic scattering , where the initial and final states are identical , invariance under time inversion
usually guarantees that even simple approximations to the dynamics will satisfy this constraint . but
building in current conservation for inelastic processes usually requires consistent treatment of both final - state interactions and interaction currents . the failure of approximate calculations ( and the pwia in particular ) to satisfy current conservation
is often seen as a serious obstacle .
some _ ad hoc _ prescription of the kind introduced by de forest @xcite , is needed .
we propose the simple prescription introduced recently in the study of deep inelastic scattering @xcite .
suppose the _
current is composed of two parts @xmath571 . in general , _ neither _ of these two parts will satisfy current conservation _ alone _ ; that is @xmath572 for each @xmath361 . however , since the exact current satisfies current conservation , @xmath573 .
we propose replacing each of the individual terms in the current by @xmath574 this procedure is covariant , guarantees that each component conserves current ( so that one can be calculated without knowing the other ) , and that their sum is unchanged : @xmath575 .
perhaps the best argument can be found in ref .
@xcite where it was shown ( for a very simple case ) that the born term defined in this way dominates the final - state interaction term in the deep inelastic limit , resolving a long standing puzzle .
finally , note that @xmath576 ( where @xmath577 are the virtual photon polarization vectors satisfying @xmath578 ) so that the response tensor ( [ resptens ] ) is _ unaffected _ by the redefinition ( [ redef ] ) !
we now turn to a brief discussion of threshold electrodisintegration .
threshold deuteron electrodisintegration measures the @xmath579 reaction in kinematics in which the proton and neutron , rather than remaining bound , are unbound with a few mev of relative kinetic energy in their center of mass system .
if the final - state energy is low enough , the final state will be dominated by transitions to the @xmath580 final state , and will be a pure @xmath581 , @xmath582 , @xmath583 transition , similar to the @xmath584 transition . this transition is a companion to the @xmath1 structure function ; both are magnetic transitions and both are filters for exchange currents with only one isospin ( @xmath585 is @xmath586 and @xmath587 is @xmath588 ) . to see the similarity , compare the top right panel of fig .
[ rabt20 ] with the threshold measurements shown in fig .
[ thresh ] . both have a similar shape , and in both cases the uncertainties in the theoretical predicitions are large .
the similarity of these two processes ( elastic and threshold inelastic ) also holds for the theory .
these two processes can be used to separately determine the precise details of the @xmath266 and @xmath265 exchange currents .
once the exchange currents are fixed , they can be used to predict the results of @xmath520 over a wide kinematic region . any theoretical approach that works for the form factors should also work equally well for threshold electrodisintegration , yet very few of the groups who have calculated form factors
have also calculated the threshold process .
this may be due in part to the fact that the @xmath265 interaction currents are larger than the @xmath266 interaction currents , making the threshold electrodisintegration calculation more difficult than the elastic calculation .
a more definitive test of the various relativistic approaches discussed in the previous sections will be possible once the elastic calculations are extended to the threshold inelastic process .
previous threshold electrodisintegration experiments have reported an _ average _
cross section that can be obtained theoretically by integrating the relative @xmath589 energy in the final state , @xmath590 , from 0 to 3 , or ( in some cases ) 0 to 10 mev .
the unbound @xmath580 final state dominates at threshold ( because the @xmath591-@xmath592 scattering state is orthogonal to the deuteron state at threshold ) , but above threshold there are contributions from the @xmath591-@xmath592 scattering state , and eventually from the @xmath7 @xmath593 scattering states as well . to emphasize this magnetic transition ,
data have been taken at large electron scattering angles , mostly 155@xmath137 or 180@xmath137 .
threshold electrodisintegration provides strong evidence ( perhaps the best we have ) for the existence of isovector exchange currents @xcite .
the impulse approximation calculation of the transition to the @xmath580 final state has a zero arising from the negative interference between the @xmath594 and @xmath595 pieces of the transition that lead to a minimum at @xmath62 near 0.5 gev@xmath596 .
this minimum is not seen in the data , and theoretical calculations of the @xmath265 exchange current contribution fill in the minimum and explain the data .
experiments have measured @xmath9 threshold electrodisintegration to @xmath62 about 1.6 gev@xmath42 with better than 1 mev ( @xmath149 ) resolution , integrating up to @xmath590 @xmath127 3 mev @xcite , and to nearly 3 gev@xmath42 with 12 - 20 mev ( fwhm ) resolution , and integrating up to @xmath590 @xmath127 10 mev @xcite . where @xmath410 ( @xmath597 ) is the energy transfer ( scattered electron energy ) .
the slac articles @xcite and first bates article @xcite use instead @xmath598 , where @xmath599 is the total @xmath519 kinetic energy in their c.m .
frame , @xmath600 .
( a typographical error at one point in @xcite misidentifies @xmath599 as the energy of a nucleon , rather than the two nucleons . ) these cross sections are related by the jacobian @xmath601 which numerically ranges from about 1.3 to 2.3 for the data that we present .
the articles showing the `` @xmath602 '' cross sections appear to plot the saclay cross sections as published , as `` @xmath603 '' cross sections , rather than converted . ]
figure [ thresh ] shows the smooth rapid fall off of threshold electrodisintegration cross sections for @xmath62 up to about 1.2 gev@xmath42 , and the change in slope for the higher @xmath62 slac and bates data .
there is good agreement between the various measurements , including those not shown , considering the change in scattering angle ( 155@xmath140 for saclay , 160@xmath140 for bates , 180@xmath140 for slac ) and integration region for @xmath590 ( 0 - 3 mev for saclay and bates data shown , 0 - 10 mev for slac ) .
the figure also shows results from seven theoretical calculations .
the oldest calculation shown in the figure is from mathiot @xcite .
this model includes one and two body currents based on @xmath319 and @xmath223 exchange , and also contributions to the two - body current operator from the exitation of the @xmath604 .
mathiot confirms that the one body current ( the impulse approximation ) produces a sharp minimum in the cross section at about @xmath605 0.5 gev@xmath42 ( in agreement with the recent schiavilla-1 body curve shown in the figure and in complete disagreement with the data ) and that the @xmath319 exchange current fills this in , shifting the minimum to about 1 gev@xmath42 .
the @xmath223 exchange is also important , shifting the mimimum to @xmath407 1.4 gev@xmath42 .
contributions from the electroexication of a @xmath604 are smaller , at least below @xmath606 gev@xmath42 .
figure [ thresh ] shows that this calculation breaks down above @xmath607 gev@xmath42 , probably because it does not include many of the contributions included in modern calculations .
more recent calculations , also based on hadronic degrees of freedom , are from leidemann , schmidtt , and arenhvel [ lsa ] @xcite and schiavilla .
( the calculations by schiavilla shown on the figure are based on the work of ref .
@xcite , but use the more recent argonne av18 potential .
these curves were also published in the review @xcite . ) in both of these cases the details are largely unpublished and the inelastic calculations are not as up - to - date as the corresponding elastic calculations recently done by the same groups , so it is premature to draw definite conclusions . the work of smejkal , truhlk , and gller [ stg ] @xcite obtains exchange current contributions from the @xmath608 system using a chiral lagrangian , and does a good job describing the data out to 2 gev@xmath42 .
predictions from two quark cluster models , the early model of yamauchi , yamamoto , and wakamatsu [ yyw ] @xcite , and the more recent model of lu and cheng [ lc ] @xcite , are also shown in fig .
[ thresh ] .
these calculations both tend to have too much structure in the region of the shoulder , but do show that quark cluster models have the ability to describe the exchange currents needed to account for the data .
the recent improvements in relativistic theory discussed in sec .
[ nucsec ] will lead to a new generation of calculations that will rely on threshold electrodisintegration to provide details about the nature of the @xmath265 exchange currents .
the most precise constraint on these currents comes from the @xmath609 transition , and this part of the transition is partly obscured by the poor energy resolution of the existing high @xmath62 measurements .
a new and improved experiment at jlab with higher resolution would allow the threshold @xmath609 process to be better extracted , with a better resulting determination of the isovector exchange currents .
it is also important to determine whether or not there is a minimum near 1.2 gev@xmath42 .
this present indication of a minimum might be an artefact of the end of the saclay data @xcite vs. the start of the slac data @xcite , along with systematic uncertainties of these and the bates measurements @xcite .
a high statistics high resolution measurement at large @xmath62 is feasible .
measurements have been proposed @xcite with 1.5 mev resolution at a scattering angle of 160@xmath137 .
the experiment would use hall a with the hrs spectrometer vacuum coupled to the scattering chamber , and a special cryotarget and collimators to enhance resolution and reduce backgrounds .
measurements were proposed for 6 points from 1 to 3.7 gev@xmath42 .
we note that slac ne4 simultaneously measured the threshold electrodisintegration along with the elastic structure function @xmath1 . because the large @xmath62 threshold inelastic cross section is typically an order of magnitude larger , the @xmath610 measurements see essentially only the inelastic processes , and @xmath611 is needed to determine the elastic scattering .
if it is possible to maintain large solid angle for the elastic scattering , and good resolution for the threshold electrodisintegration , both data sets can be obtained simultaneously .
deuteron photodisintegration was first investigated in the early 1930s , in order to understand the structure of the neutron .
after the discovery of the neutron by james chadwick , attention turned to its mass and structure .
was the neutron a fundamental particle , like the proton and electron , or was it a bound state of the electron and proton , different from the hydrogen atom ? if it was a bound state of the proton and electron , how were the electrons confined into the small nuclear volume ?
conflicting experimental evidence on the neutron mass prevented resolution of the issue until 1934 , when chadwick and maurice goldhaber used deuteron photodisintegration @xcite to determine that the neutron mass was slightly heavier than that of the hydrogen atom .
thus , the neutron , being heavier than the proton plus electron , was a fundamental particle , and there was no longer any basis for thinking electrons could be present in nuclei @xcite . subsequently , deuteron photodisintegration cross sections have served as a standard test case for nuclear theory .
the effects , for example , of the @xmath604 resonance in cross sections for beam energies near 300 mev are pronounced , but until recently large discrepancies between different experimental data sets made precise tests of theories difficult @xcite .
the 1960s-1970s saw the start of polarization measurements .
the earliest data were intermediate energy measurements of the induced proton polarization @xcite and low energy measurements of the induced neutron polarization @xcite .
large induced polarizations were observed @xcite soon afterward , particularly for energies above the @xmath604 resonance and for center of mass angles near 90@xmath140 . the combination of more extensive confirming measurements @xcite for @xmath612 about 350 - 700 mev , which could not be reproduced theoretically , and interest in dibaryons led to much excitement about deuteron photodisintegration in the late 1970s and early 1980s .
there were many serious theoretical efforts , numerological studies involving inclusion of dibaryon resonances , and extensive experimental studies of cross sections and polarization observables .
since _ both _ the recent deuteron form factor measurements _ and _ the recent high energy deuteron photodisintegration measurements have been made with 4 gev electron beams , it is sometimes assumed that the same theory should work for both . in this review
we emphasize that this need not be the case .
the kinematics of elastic electron - deuteron scattering and deuteron photodisintegration are very different , and the physics being explored by these two measurements is also very different . the implications of this remarkable feature of electronuclear physics , often not fully appreciated , will be discussed briefly in this section .
@llll & @xmath613 & & @xmath614 + @xmath615 & @xmath616 ( 939 ) & @xmath617 & @xmath618 + @xmath619 & @xmath620 , @xmath621 , @xmath622 , @xmath162 & @xmath623 & @xmath624 , @xmath625 , @xmath626 + & @xmath627 , @xmath628 , @xmath629 , & & + & @xmath630 , @xmath631 , @xmath632 & & + @xmath633 & @xmath634 , @xmath635 , @xmath636 & @xmath637 & @xmath638 , @xmath639 , @xmath640 , + & & & @xmath641 , @xmath642 + @xmath643 & @xmath644 channels & @xmath645 & @xmath646 channels + @lclr band & mass range & members & number of + & & & channels + @xmath647 & 1878 & @xmath648 & 1 + @xmath649 & 2171 & @xmath650 & 1 + @xmath651 & 2464 2579 & @xmath652 , @xmath653 , @xmath654 & 13 + @xmath655 & @xmath162 2858 2872 @xmath162 & @xmath656 , @xmath657 , @xmath658 & 17 + @xmath659 & @xmath162 3155 3280 @xmath162 & @xmath660 , @xmath653 , @xmath661 , @xmath662 & 86 + @xmath663 & @xmath162 3452 3652 @xmath162 & @xmath664 , @xmath665 , @xmath666 , @xmath667 , & + & & @xmath668 , @xmath669 & 66 + @xmath670 & @xmath162 3832 3860 @xmath162 & @xmath671 , @xmath672 , @xmath673 , @xmath674 , & 52 + @xmath675 & @xmath162 4046 4440
@xmath162 & @xmath676 , @xmath677 , @xmath678 , @xmath679 , & + & & @xmath680 , @xmath681 , @xmath682 & 50 + the kinematics of elastic scattering and photodisintegration are compared in fig.[fig : dgpkin ] , which shows @xmath683 as a function of the photon ( real or virtual ) energy @xmath684 where @xmath685 and @xmath686 for real photons .
the mass of the final excited state increases rapidly as @xmath353 decreases below its maximum allowed value of @xmath687 . for any energy @xmath74 or any @xmath62
, elastic @xmath3 scattering leaves the @xmath589 system bound , with no internal excitation energy added to the two nucleons . for quasifree scattering ( @xmath688 ) the mass of the final @xmath589 system grows with @xmath74 , and as @xmath353 decreases below 1 the mass grows more rapidly with @xmath74 . as @xmath353 @xmath136 0 , we approach the real photon limit .
real photons produce the maximum value of @xmath308 of any given beam energy . with each 1 gev of beam energy @xmath689 increases by approximately 500 mev , driving the final state deep into the resonance region .
the well established nucleon resonances , all of which can be excited by 4 gev photons , are listed in table [ tab : thresh ] , and the bands of thresholds at which these resonances are excited , either singly or in pairs , are listed in table [ tab : channels ] and shown in fig .
[ fig : dgpkin ] .
[ the fermi momentum of the struck nucleon , and the widths of the resonances , will average these thresholds over a wider kinematic region than shown . ] by @xmath612 = 1.2 gev , the final - state mass is already reaching @xmath308 = 2 gev , the nominal onset of deep inelastic scattering ( dis ) from a single nucleon if we assume one of the nucleons in the deuteron remains at rest . at @xmath612
= 4 gev , the final - state mass is approximately 4.5 gev , and at least 286 thresholds for the production of pairs of baryon resonances have been crossed ( and there are probably more from unseen or weakly established resonances ) .
a photon energy of 4 gev corresponds to @xmath519 scattering with a laboratory kinetic energy of about 8 gev ! ( see also the kinematic argument given by holt @xcite . )
it is clearly very difficult ( if not impossible ) to construct a theory of high energy photoproduction in which all of the 24 established baryon resonances and their corresponding 286 production thresholds are treated microscopically .
by contrast , elastic electron deuteron scattering requires a microscopic treatment of only _
one channel_. all of the 286 channels also contribute to elastic scattering , of course , but in this case they are not explicitly excited , and can probably be well described by slowly varying short - range terms included in a meson exchange ( or potential ) model . in photodisintegration , _ each of these channels is excited explicitly_. as we shall see below , it is sufficiently difficult to construct an adequate theory in the region of the @xmath604 resonance , so it is difficult to imagine this program being extended to a realistic treatment of many resonances , including intermediate states in which multiple mesons are present .
it would instead appear that an alternate framework that _ averages over the effects of many hadronic states _ is needed .
the alternatives are to use a glauber - like approach , or to borrow from our knowledge of dis and build models that rely on the underlying quark degrees of freedom .
we will return to these issues in our review of the theory in sec .
[ phototheory ] below .
the photodisintegration cross section was previously given in sec .
[ photodis1 ] . including the polarization observables defined by the density matrices of eq .
( [ density ] ) , and using the notation of ref .
@xcite , the structure functions are @xmath690 where the coordinate system , shown in fig . [ dgpcoord ] ( a simplified and relabeled version of that given in fig . [ fig_cross ] ) is constructed from the incident photon direction @xmath691 and the outgoing proton direction @xmath692 .
substituting the expansions ( [ polex ] ) into the cross section formula ( [ photonxsec ] ) , gives , in a notation suggested by ref .
@xcite , @xmath693 \label{gdcross}\end{aligned}\ ] ] where @xmath694 is the differential cross section for unpolarized photons , and explicit expressions for the asymmetry parameters are given in table [ tab : polpara ] .
note that the observables @xmath695 and @xmath10 are the real and imaginary parts of the same combination of amplitudes , so that an experimental measurement of both of these observables will fully determine this linear combination of amplitudes .
for the induced polarization instead of @xmath10 .
] this is not true for any other pair of observables shown in the table . in the c.m .
system @xmath696 where @xmath697 \ , .\end{aligned}\ ] ] the @xmath698s are defined in table [ tab : fs ] @lll amplitude & ref .
@xcite@xmath699 & ref .
@xcite + @xmath700 & @xmath701 & @xmath702 + @xmath703 & @xmath704 & @xmath705 + @xmath706 & @xmath707 & @xmath708 + @xmath709 & @xmath710 & @xmath711 + @xmath712 & @xmath713 & @xmath714 + @xmath715 & @xmath716 & @xmath717 + the _ single _ polarization observables are the induced proton polarization @xmath10 , the linearly polarized photon asymmetry @xmath718 , and the vector polarized ( along @xmath719 ) target asymmetry @xmath720 . the quantities @xmath10 and @xmath718 are defined by @xmath721 where @xmath722 refers to @xmath723 , and in the expression for @xmath718 the photon is polarized either in plane along the @xmath724 direction ( @xmath101=0 , denoted @xmath725 ) or out of plane along the @xmath719 direction ( @xmath726 , denoted @xmath727 ) .
note that the opposite phase convention for @xmath718 ( @xmath728 - @xmath729 ) is often used , and is used in the figures below .
double _ polarization transfer observables are @xmath730 and @xmath731 ( for linearly polarized photons ) , and @xmath732 ( for circularly polarized photons ) , with the subscripts on @xmath733 and @xmath734 giving the polarization direction in the final state .
there have been many attempts to understand low energy deuteron photodisintegration using a conventional meson - baryon framework .
since the first band of nucleon resonances is not excited until about 400 mev photon energy ( recall fig .
[ fig : dgpkin ] ) it makes sense to describe the process below 400 mev using a model of coupled @xmath7 , @xmath735 and @xmath736 channels .
laget @xcite showed that the prominant shoulder in the total cross section at @xmath737 = 300 mev can be largely explained by the mechanism in which a @xmath604 is photoproduced at a nucleon followed by the reabsorption of its decaying pion by the other nucleon .
he also examined many other mechanisms , including rescattering up to second order , but did not do a full calculation of the final - state interaction .
later leideman and arenhvel @xcite treated the @xmath7 , @xmath735 and @xmath738 as coupled channels and included final - state effects to all orders .
tanabe and ohta @xcite followed with a more complete treatment of the final state which is consistent with three - body unitarity . in a number of conference talks , lee @xcite reported on coupled channel calculations using @xmath616 , @xmath604 and the @xmath620 ( roper ) resonances , which he extended to @xmath739 gev .
his work suggests that final - state interactions significantly enhance the cross section for photon energies above 1 gev .
the recent calculations by schwamb and arenhvel and collaborators @xcite include the @xmath7 , @xmath735 and @xmath740 channels , and also contributions from meson retardation , meson exchange currents , and the meson dressing of the nucleon lines required by unitarity .
all parameters are fixed from nucleon - nucleon scattering and photoreactions such as @xmath604 excitation from the nucleon , so no new parameters are introduced into the calculations of the deuteron photodisintegration process itself .
they obtain a reasonable description of @xmath7 scattering up to lab energies of 800 mev , particularly for the important @xmath741 partial wave , and emphasize that the consistent inclusion of retardation effects improves their results for photodisintegration . in another work it was shown that inclusion of the @xmath742 and @xmath743 resonances @xcite seems to have only a small effect below about 400 mev .
these resonances enhance the total cross section by only 14% at 680 mev , although effects on double polarization observables can be more significant .
hence there is some justification to limiting the theory of low energy photodisintegration to the channels considered in ref .
@xcite . by comparison to the detailed and careful treatments developed for lower energies , calculations specifically designed to describe higher energy photodisintegration are less complete .
an unpublished bonn calculation @xcite includes pole diagrams generated from @xmath319 , @xmath223 , @xmath320 , and @xmath410 exchange , plus the 17 well - established nucleon and @xmath604 resonances with mass less than 2 gev and @xmath482 @xmath744 5/2 ( listed in table [ tab : thresh ] ) .
the calculation uses resonance parameters taken from the particle data group .
nagornyi and collaborators @xcite have introduced a covariant model based on the sum of pole diagrams .
the latest version @xcite gives the photodisintegration amplitude as a sum of contributions from only 4 feynman diagrams : three pole diagrams coming from the coupling of the photon to the three external legs of covariant @xmath745 vertex plus a contact interaction designed to maintain gauge invariance .
this model divides the @xmath745 vertex into `` soft '' and hard parts , with the hard part designed to reproduce the pqcd counting rules and its strength determined by a fit to the data at 1 gev .
the model has no final - state interactions or explicit nucleon resonance contributions .
there is also a relativistic calculation of photodisintgration in born approximation using the bethe - salpeter formalism @xcite .
it is found that the cross section is a factor of 2 to 10 times too small , and that relativistic effects are large .
all of these models are rather crude , and taken together it is not clear what one should conclude from them .
the calculations each appear to emphasize some aspects of what a comprehensive meson - baryon theory of photodisintegration should entail .
perhaps we can say that conventional calculations that neglect final - state interactions seriously underestimate the cross section , but may be corrected in an approximate manner by introducing diagrams with poles in the @xmath289 channel .
s - channel pole diagrams can be regarded as a crude approximation to the missing final - state interactions .
more generally , using pole diagrams with form factors that have the correct behavior at high momentum transfer may also insure that meson - baryon theories of deuteron photodisintegration will also have the correct high energy behavior @xcite .
perturbative qcd , discussed briefly in sec . [ pertqcd ] , provides explicit , testable predictions for the cross section and polarization observables . for the case of elastic scattering , the high energy ( virtual ) photon had to share its mommentum equally with all of the constitutents , leading to the typical diagram shown in fig .
[ pqcddiagrams ] .
photo ( or electro ) disintegration differs in that the momentum will not be distributed to _ all _ six quarks unless momentum is transferred to _ each _ nucleon , and this requires non - forward scattering , i.e. the angle @xmath746 between the three - momentum of the outgoing proton and the photon ( in the c.m .
system ) must not be 0 or @xmath319 . more precisely ,
if @xmath737 is the energy of the photon in the lab system , the square of the momentum transferred to each nucleon @xmath747 ( where @xmath748 1,2 ) can be written @xmath749 where @xmath289 , the square of the c.m .
energy , is @xmath750 and the limits in ( [ t1t2 ] ) are the result for @xmath751 . if @xmath752 the momentum transferred to each nucleon is balanced and the pqcd result is reached most rapidly . a `` typical '' pqcd diagram leading to large angle scattering
is illustrated in fig .
[ fig : dg - pqcd ] . in general
we can not expect the cross section to follow pqcd unless the minimum of the two momentum transfers @xmath753 and @xmath754 is larger than some value @xmath755 at which pqcd holds .
if the above conditions are satisfied , then the cross section and polarization observables should satisfy results predicted by pqcd : the constituent counting rules ( ccr ) and hadron helicity conservation ( hhc ) .
the ccr @xcite predicts the energy dependence of scattering cross sections at fixed center of mass angle @xmath756 where @xmath287 is the total number of pointlike particles in the initial and final states of the reaction .
for deuteron photodisintegration , @xmath757 ( there is one photon and 6 + 6=12 quarks ) , and because of eq .
( [ t1t2 ] ) @xmath758 may be replaced by @xmath289 , as is commonly done . for elastic @xmath3 scattering @xmath759 because there are electrons in both the initial and final state , and @xmath760 , the momentum transfered by the electron . in this case @xmath761 , and recalling eq .
( [ mott ] ) we have @xmath762 recovering the pqcd prediction @xmath459 .
the second rule from pqcd is that the total helicity of the incoming and outgoing hadrons should be conserved @xcite @xmath763 where @xmath764 ( @xmath765 ) are the helicities of the initial ( final ) hadrons . in sec .
[ pertqcd ] we discussed the implication of this rule for the deuteron form factors .
hhc makes predictions for many spin observables , particularly vector polarizations @xcite .
here it predicts that the amplitudes of table [ tab : fs ] will have the following behavior at large @xmath758 @xmath766 where we have used the fact that each helicity flip is suppressed by a power of @xmath758 .
the implication of hhc for the observables given in table [ tab : polpara ] is summarized in table [ tab : dgppol ] .
the limits at @xmath746 @xmath127 90@xmath140 on some of the observables require assumptions about relations between the helicity conserving variables @xcite .
@lll observable & hhc limit & approach + & & to hhc + @xmath767 & @xmath768 & @xmath769 + @xmath10 & 0 & @xmath770 + @xmath720 & 0 & @xmath770 + @xmath718 & @xmath771 at 90@xmath140 & @xmath769 + @xmath730 & 0 & @xmath770 + @xmath11 & 0 & @xmath770 + @xmath12 & @xmath772 / ( d\sigma / d\omega ) \;\rightarrow 0 $ ] at 90@xmath140 & @xmath769 + @xmath773 & 0 & @xmath770 + @xmath774 & @xmath775 at 90@xmath140 & @xmath769 + in an attempt to include some of the soft physics omitted from pqcd , and to extend the region of applicability of pqcd down to lower momentum transfers , brodsky and hiller introduced the idea of reduced nuclear amplitudes ( rna ) @xcite
. in this model the gluon exchanges that contribute to identifiable subprocesses ( such as nucleon form factors ) are collected together and their contributions replaced by experimentally determined nucleon form factors .
it is hoped that the resulting expressions will correctly include much of the missing soft physics , and will therefore be valid to lower momentum transfers than the original pqcd expressions from which they were obtained .
when applied to deuteron photodisintegration , the cross section is written @xmath776 where the phase space factor of @xmath777 comes for a careful reduction of the phase space factors in eqs .
( [ resptens ] ) , ( [ sigma0 ] ) , and ( [ t1t2 ] ) , instead of @xmath778 . ]
@xmath779 is the reduced nuclear amplitude , @xmath780 and @xmath781 are the proton and neutron form factors with @xmath782 and @xmath783 the average momentum transferred to the proton and neutron @xmath784 and the square of the transverse momentum is @xmath785 the power of the `` extra '' factor of @xmath786 is fixed once the phase space and nucleon form factors have been taken into account . note that this model does not attempt to normalize @xmath787 .
the rna form given in eq .
( [ rna ] ) is somewhat arbitrary , particularly in the specification of the form of @xmath788 .. this choice of @xmath789 leads to an increased energy dependence and worse agreement with the data . ]
radyushkin @xcite has argued that the elementary process not accounted for by the nucleon form factors ( i.e. the absorption of a hard photon followed by exchange of a hard gluon with another quark ) should include nonperturbative contributions .
the effect is to replace the @xmath790 factor in eq .
( [ rna ] ) by a smooth function @xmath791 , which is assumed to vary slowly in energy and angle . in the fits
described below , @xmath792 will be taken to be a constant adjusted to fit the data , implying that @xmath793 instead of @xmath794 . alternatively , if the quark exchange mechanism shown in fig .
[ fig : dg - pqcd ] is to be taken seriously , a more detailed calculation is possible .
this is the motivation for the work of frankfurt , miller , sargsian , and strikman @xcite where the quark exchange diagram ( which the authors refer to as a quark rescattering diagram ) is calculated in front - form dynamics using model wave functions for the nucleons and the deuteron .
the matrix element is written as a convolution of an elementary quark exchange interaction with the initial and final nucleon wave functions .
the final nucleons are free and the distribution of the initial nucleons is given by the deuteron wave function .
since the photon momentum is shared by the proton and neutron , there is little sensitivity to the high momentum part of the deuteron wave function .
the elementary interaction is a quark exchange between the two nucleons , with the photon absorbed by one quark which then gives up its momentum through a hard gluon exchange with another quark .
the authors show that this can be replaced approximately by the wide angle @xmath519 scattering cross section ( also dominated by quark exchange ) .
the final formula obtained from these arguments is @xmath795 where @xmath796 was used in the fits to the data discussed below .
the authors propose using experimental data for the @xmath519 cross section , but since data does not exist for the actual kinematic conditions needed , it must be extrapolated , and predictions for photodisintegration are given as a band corresponding to the uncertainties introduced by the extrapolations .
the authors believe that their predictions should be valid for @xmath612 @xmath152 2.5 gev , and nucleon momentum transfers @xmath797 and @xmath798 @xmath152 2 gev@xmath42 .
the high energy approaches described above all focus on the region where both @xmath782 and @xmath782 are large ( where perturbative arguments can serve as the foundation for the treatment ) . alternatively , we may ask what to expect when one of these momentum transfers is small ( but @xmath289 is still large ) .
the authors in refs .
@xcite develop a model ( which they refer to as the `` quark - gluon string model '' ) based on a reggie generalization of the nucleon exchange born term . here
the exchanged nucleon is replaced by a nucleon reggie trajectory that represents the sum of a tower of exchanged nucleon resonances ( or the exchange of three quarks dressed by an arbitrary number of gluons ) .
the energy dependence of the predicted cross section is @xmath799 where @xmath800 is the nucleon reggie trajectory , with @xmath801 , where @xmath758 is in units of gev@xmath42 .
recent work @xcite emphasizes the importance of the nonlinear term in the regge trajectory .
the model is intended to work for @xmath802 gev , with @xmath803 less than about 1 gev@xmath42 .
we now turn to a review of the experimental data and to a comparison between theory and experiment .
the world data set for cross sections ( except for the most recent experiments ) has been presented several times @xcite , and we will not review the normalization problems of older cross section data sets .
table [ tab : dgppoldata ] presents an extensive list of the published polarization data .
about 70 publications , starting in 1960 , present about 1200 data points for photodisintegration and the time reversed radiative capture reaction .
table [ tab : dgppoldata ] generally lists photon lab energy and proton c.m .
angle ( neutron c.m .
angle for the @xmath804 data ) .
all of the radiative capture experiments have measured @xmath805 , and for these experiments we usually give the neutron beam kinetic energy and outging photon c.m .
angle ; for the iucf experiment we give their reported proton c.m . angle . matching c.m .
energies leads to @xmath806 or @xmath807 at high energies .
comparison of these data indicates serious problems with backgrounds and/or estimates of systematic uncertainties in a number of cases , as will become clear in figures in the sections below .
we review some of the lower energy data in the next subsection .
high energy experiments , with photon energy above @xmath1511 gev , are covered in the following subsection . in this section
we review selected experiments with beam energies from about pion production threshold to several hundred mev .
tagged photon facilities , with their improved knowledge of incident beam flux , have allowed significantly improved cross section measurements in this region since the 1980s .
of particular note are extensive recent data sets from legs and mainz .
figures [ fig : dgpobs300 ] and [ fig : dgpobs410 ] show angular distributions for the cross section , and @xmath718 , @xmath720 , and @xmath10 at photon energies near 300 and 450 mev , respectively .
the legs tagged , backscattered , and linearly - polarized photon beam was used to determined cross sections and @xmath718 @xcite .
five independent measurements used three detector systems , two targets , and two different laser frequencies .
data were taken for @xmath612 = 100 - 315 mev with eight laboratory angles from 15 - 155@xmath140 .
cross section statistical uncertainties range from a few percent to about 15% , with systematic uncertainties of 5% .
one observation in the legs data is of pion contamination that may have been missed in earlier experiments , leading to increased cross sections .
the mami mainz experiment @xcite used the glasgow photon tagger along with the large solid angle detector daphne to determine cross sections in the ranges @xmath612 = 100 - 800 mev and @xmath746 = 30 - 160@xmath140 .
data were binned by 20 mev in energy and 10@xmath140 in angle .
statistical uncertainties ranged from a few percent at lower energies to about 25% at the highest energies .
systematic errors were also energy dependent , ranging from a few percent to several percent .
the @xmath718 asymmetry @xcite was obtained by using a coherent bremsstrahlung radiator .
agreement between the mainz and legs cross section results is generally better than 10% .
the @xmath718 asymmetries also agree well with each other and with earlier results from yerevan @xcite .
measurements from kharkov @xcite generally agree , except for a tendency to be slightly smaller at many beam energies .
the vector polarized target asymmetry @xmath720 was measured at tokyo @xcite and at bonn @xcite .
the two measurements generally agree , with the data appearing to follow , very roughly , a @xmath808 dependence at each energy , as do the calculations of @xcite .
the induced polarization @xmath10 has been measured at a number of laboraties , with significant amounts of lower energy data from stanford @xcite , tokyo @xcite , and kharkov .
there were several experiments at kharkov , including initial measurements @xcite , high - energy measurements @xcite , simultaneous measurements of @xmath718 , @xmath730 , and @xmath10 @xcite , and lower - energy measurements @xcite .
the kharkov data do not have a desirable level of consistency .
the simultaneous measurements of @xmath718 , @xmath730 , and @xmath10 were taken as single - arm data , away from the photon endpoint , and may suffer from backgrounds .
polarizations below about 300 mev , including the intermediate energy neutron measurements @xcite , appear to be well reproduced by theories .
there are numerous data , particularly at @xmath746 @xmath127 45@xmath140 , 78@xmath140 , 90@xmath140 , and 120@xmath140 , but there are few energies at which there are good angular distributions .
the conclusion that polarizations are large , close to @xmath809 , and peak at about 500 mev near 90@xmath140 is beyond dispute .
figures [ fig : dgpobs300 ] and [ fig : dgpobs410 ] show the good agreement of the recent mainz @xcite and older bonn @xcite calculations with the cross section and @xmath718 asymmetry .
theory seems to agree with @xmath810 better than @xmath811 @xcite .
the agreement is better at the lower energy , and the newer mainz calculation is generally in better agreement .
however , there is difficulty , particularly at the higher energy , with @xmath720 and @xmath10 , both imaginary parts of the interference of amplitudes .
the large induced polarizations above the @xmath604 resonance have remained a puzzle for almost 30 years , and are still not fully explained by the newest theories .
the bonn calculation was in sufficiently good qualitative agreement with all observables ( @xmath812 , @xmath718 , @xmath720 , _ and _ @xmath10 ) for energies up several hundred mev for the authors to consider this @xmath10 puzzle solved . however , detailed examination of fig .
[ fig : dgpobs410 ] shows that neither the shape nor strength of the angular distribution is accurately reproduced ; this will become clearer when we examine the energy dependence of @xmath10 at @xmath746 = 90@xmath140 below .
figure [ fig : dgpsigmas ] shows the published high energy photodisintegration data , from experiments ne8 @xcite and ne17 @xcite at slac , and e89 - 012 @xcite and e96 - 003 @xcite at jlab .
these experiments determine cross sections for @xmath746 @xmath151 36@xmath140 , 52@xmath140 , 69@xmath140,and 89@xmath140 at energies from about 0.7 to 5.5 gev ; there are also some backward angle data up to 1.8 gev from ne8 .
these data overlap well ; the experiments , while all run by essentially the same collaboration , used three spectrometers in two experimental halls at two laboratories .
there is also good overlap , variations of less than about 20% , with the highest energy mainz tagged photon data @xcite , and with older untagged data @xcite .
tagged photon measurements at low energies provide an accurate measure of beam flux , and along with the measured proton angle and energy , can determine a missing mass that allows background rejection . at high energies ,
smaller cross sections can not be determined with the reduced flux of tagged photons .
the bremstrahlung endpoint technique was used for all of the slac and jlab measurements shown in fig .
[ fig : dgpsigmas ] . in the endpoint technique ,
the measured proton momentum vector determines the incident photon energy and neutron kinematics , _ assuming _ the reaction is two body photodisintegration .
low momentum protons are cut from the analysis to prevent contamination from final states such as @xmath813 , while high momentum protons are cut to eliminate the larger uncertainty in the photon flux close to the photon endpoint .
backgrounds are determined by radiator out and empty target measurements , and subtracted .
events in the region beyond the endpoint can be used to check the subtraction .
the increase in time required for this subtraction makes it prohibitive for the highest energy measurements ; for these the electrodisintegration background is calculated @xcite and subtracted . to determine the cross section ,
the incident photon flux is calculated using the method of ref .
thick radiator corrections are typically about 15% for a radiator thickness of 6% of a radiation length .
the main feature of the cross section data above about 1 gev is the @xmath794 ( @xmath814 ) fall off of the cross sections @xmath815 ( @xmath812 ) at @xmath746 = 90@xmath140 and 69@xmath140 , in agreement with the ccr and thus with perturbative qcd expectations .
in contrast , the cross sections at the forward angles 36@xmath140 and 52@xmath140 fall off more slowly , with approximate @xmath816 scaling at lower energies until the onset of the @xmath794 behavior at about 4 and 3 gev beam energy , respectively . at each angle
, the onset of the @xmath794 behavior corresponds to @xmath789 @xmath151 1 gev . the rna and the radyushkin estimates in fig .
[ fig : dgpsigmas ] were normalized to the datum at 89@xmath137 and @xmath817 gev , fixing their one free parameter .
the rna is then almost a factor of 2 too large at 36@xmath137 , and also much too large at lower energies , requiring that the soft physics missing from the rna interfere _ destructively _ with the leading terms .
suggested angular dependences of @xmath818 ) @xcite would increase the rna curve further at the forward angles , worsening agreeement with the data .
in contrast , the radyushkin estimate gives a somewhat better account of both the angular and energy dependence ( even though it only goes asympotically as @xmath814 ) , confirming that phase space and nucleon form factors are all that is needed to account for much of the kinematic variation of the cross section .
while the apparent onset of scaling at the forward angles suggests this agreement is starting to break down , we conclude the present data are insufficient to uniquely fix the asymptotic energy dependence of the cross section .
the cross section data are also reasonably well reproduced by the model of frankfurt , miller , strikman and sargsian [ fmss ] @xcite and the quark gluon string ( qgs ) model @xcite .
the predictions of fmss are uncertain because ( a ) the high energy @xmath7 scattering data has an uncertain energy dependence reflecting the experimental errors , and ( b ) the extrapolation of the @xmath7 data required for the predictions introduces further errors .
these two uncertainties combine to give the jagged region shown in fig .
[ fig : dgpsigmas ] .
the qgs model describes the forward angle data up to 4 gev reasonably well , even for values of @xmath803 exceeding the nominal limits of the model .
the newer work @xcite predicts that the angular distributions will become increasingly symmetric at higher energies ; older estimates @xcite had predicted an increasing _
two experiments currently have unpublished data for cross sections at photon energies up to 2.5 gev .
jlab hall b e93 - 017 @xcite used the clas with tagged photons to determine nearly complete angular distributions .
the preliminary data agree well with earlier measurements , but are much more comprehensive than previous measurements in this energy range .
hall a e99 - 008 @xcite has taken angular distributions at eight angles with @xmath746 @xmath127 30@xmath140 143@xmath140 , at energies of 1.67 , 1.95 , and 2.50 gev .
[ fig : ad ] shows a sample angular distribution at approximately 1.6 gev .
there are only three sets of polarization data for deuteron photodisintegration at energies near and above 1 gev .
the induced polarization @xmath10 was measured at kharkov @xcite , the @xmath718 asymmetry was measured at yerevan @xcite , and @xmath10 and the polarization transfers @xmath11 and @xmath12 were measured at jlab @xcite .
data for the energy dependence of these four observables at a fixed @xmath752 are compared with theory in figs .
[ fig:89019py ] , [ fig:89019cxcz ] , and [ fig : sigfig ] .
the kharkov measurements of @xmath10 at @xmath746 @xmath127 90@xmath140 and 120@xmath140 extend up to about 1.1 gev ( see fig .
[ fig:89019py ] ) .
these experiments were very difficult .
the small duty factor at kharkov increases instantaneous background rates a factor of about 20,000 over those at jlab .
these large backgrounds made it necessary to use multiple spark chambers to track particle trajectories .
it was also difficult to calibrate the polarimeter .
calibrations of a polarimeter are best done by measuring its analyzing power using the known @xmath819 elastic scattering reaction @xcite , but kharkov had no polarized beam .
the kharkov measurements relied on a single elastic @xmath158 point to check false asymmetries in their polarimeter , and used analyzing powers from the literature . finally , the polarimeter had a rear trigger scintillator ; any inefficiencies in the scintillator would lead to false asymmetries .
in contrast , the recent jlab experiment @xcite had little background , used @xmath820 calibrations to determine false asymmetries and analyzing powers at each kinematic setting , and had no rear trigger scintillator . given the clear disagreement of the kharkov data with the recent jlab data , and noting that one of us ( rg ) is a spokesperson of the jlab experiment , we conclude that the highest energy set of karkov data should not be trusted .
the induced polarizations shown in fig . [ fig:89019py ] confirm our comments above concerning the bonn calculation .
the large negative polarization near 500 mev is not reproduced , and the calculation is qualitatively incorrect at higher energies . the imaginary part of the amplitude appears to be a problem in these meson - baryon calculations ; presumably this arises from an inadequate treatment of resonances . taken together ,
these recoil polarization data only weakly confirm the predictions of hhc , which predicts that @xmath10 and @xmath11 should approach zero as @xmath821 , and that ( with additional assumptions @xcite about relations between the helicity conserving amplitudes at @xmath746 = 90@xmath140 ) @xmath822 and @xmath823 as @xmath821 .
the highest energy polarization measurements of @xmath10 show that it is consistent with vanishing at energies above about 1 gev , the same energy at which the @xmath794 cross section scaling begins .
[ in the radyushkin model @xmath10 should be zero because the amplitudes are all real if there is no gluon exchange @xcite . ]
similiarly , the polarization transfer observables @xmath11 and @xmath12 both appear to peak near 1 gev , and decrease at higher energies .
however , @xmath11 does not appear to vanish sufficiently rapidly ; the data might be inconsistent with hhc .
so @xmath10 and @xmath12 ( and perhaps @xmath11 ) seem to have close to the correct behavior , but @xmath718 does not .
the highest energy @xmath718 asymmetry measurements from yerevan , with data up to about 1.6 gev , give the immediate impression that there is a minimum near 1.2 gev and that above this energy the asymmetry is tending to increase towards 1 , although the data are also statistically consistent with a constant value of about 0.3 .
in either case the trend is clearly not consistent with @xmath824 . in models based on meson - baryon degrees of freedom ,
the data indicate that the combined effects of resonances plus final - state interactions are small .
calculations of @xmath10 in meson baryon theories @xcite indicate that @xmath10 at higher energies arises largely from resonance - background interference , with a small contribution from final - state interactions .
calculations generally indicate that the @xmath604 resonance generates a large polarization , though only perhaps about 50% of the magnitude seen in the experimental data at @xmath746 @xmath127 90@xmath140 .
the roper and @xmath742 have small effects , while the @xmath743 has a large effect @xcite .
the @xmath825 , included only by the bonn group @xcite , also generates a large polarization .
( of the 17 resonances included in the bonn calculation @xcite , only those mentioned above had large effects on @xmath10 . ) as discussed above , it is hard to imagine that a theoretically acceptable high energy model based on hadronic degrees of freedom will be constructed in the near future . still , a modern relativistic calculation based on hadronic degrees of freedom would nonetheless be desirable . a calculation of @xmath10 has been done in the model of fmss @xcite . since the helicity amplitudes in the nucleon - nucleon scattering for this center of mass energy range are not uniquely determined , some modelling was needed .
the calculation showns that @xmath10 is generally very small , going from a negative value at low energies to positive values at several gev beam energy , and is consistent with the trend of the experimental results .
the polarization observable @xmath695 is also expected to be small , and opposite in sign to @xmath10 , while @xmath826 is expected to vanish .
the rapid falloff of photodisintegration cross sections with energy makes extension of the measurements difficult . as beam energy increases from 4 to 5 , 6 , and 7 gev ,
the @xmath794 dependence reduces cross sections by factors of 6 , 30 , and 115 , respectively .
only a few experiments are possible without the proposed 12 gev upgrade to jlab . as indicated in table [ dexptab ] , there are two approved experiments to continue the hall a recoil polarization measurements to additional angles , and to beam energies near 3 gev
. measurements of the @xmath718 asymmetry are possible in jlab hall b , over a range of angles and energies well over 1 gev . in the longer term ,
the jlab 12 gev upgrade offers additional possibilities . the luminosity increase planned for
hall b would allow precise polarization measurements to continue above 2 gev .
the proposed mad spectrometer for hall a @xcite would give about a factor of 5 improvement in solid angle , and if it has the low backgrounds characteristic of the current hrs spectrometers , cross section measurements are possible up to 7 gev and polarization measurements are possible up to 4 gev .
review of deuteron photodisintegration suggests the following : * a microscopic meson - baryon theory of deuteron photodisintegration must describe the @xmath7 interaction at high energies , including pion production and the contributions of hundreds of @xmath827 channels .
it is unlikely that such a theory will be constructed in the foreseeable future .
the data might indicate that the effect of many resonances is to increase the cross section and decrease the polarization observables ( by averaging over may phases ) .
this suggests that it might be possible to construct an effective theory based on hadronic degrees of freedom . * for @xmath788 greater than about 1 gev@xmath42 ,
cross sections appear to follow the constituent counting rules , but it is expected that an absolute pqcd calculation of the size of the cross section would give a result much too small .
similar observations may be made for other photoreactions , and it remains to be seen how this behavior arises , and if there is a general explanation for it . * the energy dependence of the photodisintegration cross sections has been shown to be potentially misleading indicator of the success of pqcd .
models with asymptotic behavior which differ from pqcd fit the data as well or better than pqcd . further theoretical
development and experimental tests of nonperturbative quark models would be desirable .
the new high energy measurements of the deuteron form factors and the deuteron photodisintegration observables have motivated much theoretical work .
in particular : * conventional meson theory works well in cases where all of the _ active _ hadronic channels that can contribute to a process are included in the calculations .
this has been done for the deuteron form factors ( where only the @xmath7 channel is active ) , but not for high energy deuteron photodisintegration where 100 s of @xmath8 channels are active . at high energy
any successful meson theory must include relativistic effects .
* new approaches , probably using quark degrees of freedom , are needed for high energy deuteron photodisintegration . while photodisintegration ( as well as other reactions ) seem to follow the scaling laws , theoretical estimates using pqcd give cross sections orders of magnitude too small .
the data do not support hadronic helicity conservation . thus scaling
is no longer seen as sufficient evidence for the applicability of pqcd .
there is good evidence that , in this energy region , the deuteron is undergoing a transition from a region in which conventional hadronic degrees of freedom describe the physics to a region in which quark degrees of freedom are more appropriate .
we thank h. arenhvel , b. l. g. bakker , j. carbonell , w. h. klink , d. phillips , r. schiavilla , and g. salm and for supplying numerical values of their deuteron form factor calculations , i. sick for values of the coulomb distortion corrections , and m. sargsian and m. schwamb for supplying numerical values for their photodisintegration calculations .
it is also a pleasure to thank a. afanasev , c. carlson , m. garon , t .- s . .
lee , k. mccormick , g g petratos , j. w. van orden , a. radyushkin , and g. warren for several helpful discussions , and s. strauch for assisting with photodisintegration figures .
this work was supported in part by the us department of energy .
the southeastern universities research association ( sura ) operates the thomas jefferson national accelerator facility under doe contract de - ac05 - 84er40150 .
rg acknowledges the support of the national science foundation from grant phy-00 - 98642 to rutgers university .
999 buck w w and gross f friar j l lancombe m , loiseau b , richard j m , vinh mau r , ct j , pirs p and de tourreil r wiringa r b , stoks v g j and schiavilla r stoks v g j , klomp r a m , terheggen c p f and de swart j j machleidt r , sammarruca f and song y machleidt r salpeter e e and bethe h a gross f ; ; logunov a a and tavkhelidze a n ; blankenbecler r and sugar r phillips d r , wallace s j , and devine n k gross f , van orden j w and holinde k stadler a and gross f gross f 1998 _ proceedings of the workshop on electronuclear physics with internal targets and the blast detector _ , may 2830 , mit experimental limits on possible two photon exchange are discussed in rekalo m p , tomasi - gustafsson e and prout d gross f arnold r e , carlson c e and gross f donnelly t w and raskin a s jones h jacob m and wick g c abbott d mcintyre j a and hofstadter r buchanan c d and yearian r benaksas d galster s ganichot d grosset@xmath828te b , drickey d , and lehmann p cramer r schulze m e dmitriev v f wojtsekhowski b b [ ] gilman r boden b the i garon m popov s g 1995 _ proceedings of the 8th international symposium on polarization phenomena in nuclear physics , bloomington , indiana , 1994 _ , aip conf .
339 ( aip , new york ) ferro - luzzi m bouwhuis m nikolenko d m prepost r , simonds r m and wiik b h gross f 1991 _ modern topics in electron scattering _
( frois b and sick i ed )
p 219 jackson a j , lande a and riska d o thompson r h and heller h nyman e m and riska d o 1986 _ proceedings of weak and electromagnetic interactions in nuclei , heidelberg 1986 _ p 63 weinberg s 1995 _ the quantum theory of fields i _ ( cambridge university press ) simonov yu a and tjon j a ; nieuwenhuis t and tjon j a ; nieuwenhuis t 1995 ph.d .
thesis university of utrecht ; savkli c , tjon j a and gross f keister b d and polyzou w n gross f and riska d o tjon j a 1995 _ xivth international conference on few body problem in physics , williamsburg , virginia , 1994 _ aip conf . proc .
334 ( aip , new york ) p 177 fleischer j and tjon j a ; ; hummel e and tjon j a ; ; mandelzweig v b and wallace s j ; wallace s j and mandelzweig v b wallace s 2001 talk at the _ int workshop on theories of nuclear forces and few - nucleon systems _ van orden j w , devine n and gross f dirac p a m schiavilla r and riska d forest j and schiavilla r 2001 ( private communication to be published ) schiavilla r , pandharipande v r and riska d o ; + carlson j , riska r o , schiavilla r and wiringa r b arenhvel h , ritz f and wilbois t ritz f , gller h , wilbois t and arenhvel h krajcik r a and foldy l l friar j l friar j l gross f 1978 proceedings of the _ 8th international conference on few body problems _
( springer - verlag lec .
notes in physics # 82 , zingl h ed ) , p. 46 . [ the factor @xmath829 in eq .
( 12 ) should read @xmath830 . ]
carbonell j and karmanov v a karmanov v a and smirnov a v ; carbonell j , desplanques b , karmanov v a and mathiot j -f machleidt r , holinde k and elster ch lev f m , pace e and salm g grach
i l and kondratyuk l a ; chung p l , coester f , keister b d and polyzou w n lev f m , pace e and salm g lev f m , pace e and salm g allen t w , klink w h and polyzou w n maltman k and isgur n maltman k and isgur n de forest t and mulders p j dijk h and bakker b l g buchmann a , yamauchi y and faessler a simonov yu a arndt r a brodsky s j and farrar g r ; lepage g p and brodsky s j ; brodsky s j and lepage g p isgur n and llewellyn - smith c radyushkin a v farrar g r , huleihel k and zhang h carlson c e and gross f brodsky s j and hiller j r kobushkin a and syamtomov a cao jun and wu hui - fang carlson c e ( private communication ) petratos g g , _ deuteron and helium form factor measurments at jefferson lab _ , contribution to the jefferson lab 12 gev upgrade ; ( private communication ) garon m gilman r 1988 _ jefferson lab workshop on physics & instrumentation with 6 - 12 gev beams _
p 275 kox s ( unpublished ) tomasi - gustafsson e ladygin v p tomasi - gustafsson e arnold r e , carlson c e and gross f zhou z mit bates blast experiment 00 - 03 hoekert j , riska d o , gari m and huffman a simon g g bernheim m auffret s lee k s schmitt w m arnold r g frodyma m mathiot j f leidemann w , schmitt k - m and arenhvel h carlson j and schiavilla r smejkal j , truhlk e and gller h yamauchi y , yamamoto r and wakamatsu m lu l c and cheng t s jourdan j , warren g , jefferson lab proposal 00 - 103 laget j m leidemann w and arenhvel h tanabe h and ohta k lee t - s h ; lee t - s h 1991 _ argonne national laboratory preprint _
phy-6886-th-91 ; lee t - s h 1991 _ argonne national laboratory preprint _
phy-6843-th-91 schwamb m and arenhvel h 2001 _ preprint _ nucl - th/0105033 ; schwamb m and arenhvel h ; schwamb m and arenhvel h ; schwamb m , arenhvel h , wilhelm p , and wilbois th schwamb m , arenhvel h , wilhelm p kang y , erbs p , pfeil w , and rollnik h 1990 _ abstracts of the particle and nuclear intersections conference _ , ( mit , cambridge , ma ) ; kang y 1993 ph.d .
thesis bonn nagornyi s i , kasatkin yu a , and kirichenko i k [ ] dieperink
a e l and nagornyi s i kazakov k yu and shulga d v _ preprint _ nucl - th/0101059 . kazakov k yu and shimovsky s eh carlson carl e matveev v carlson carl e and chachkhunashvili m afanasev a ( private communication ) ; sivers d ( private communication ) brodsky s j and hiller j r ; radyushkin a ( private communication ) frankfurt l l , miller g a , sargsian , m m and strikman m i frankfurt l l , miller g
a , sargsian m m , and strikman m i kondratyuk l a grishina v yu white c tiator l and wright l e ; wright l e and tiator l matthews j l and owens r o napolitano j freedman s j belz j e myers h ching r and schaerf c baba k rossi p , de sanctis e , jlab proposal e93 - 017 gilman r , holt r , meziani z e jefferson lab proposal e99 - 008 ; schulte e to be published | recent measurements of the deuteron electromagnetic structure functions @xmath0 , @xmath1 , and
@xmath2 extracted from high energy elastic @xmath3 scattering , and the cross sections and asymmetries extracted from high energy photodisintegration @xmath4 , are reviewed and compared to theory .
the theoretical calculations range from nonrelativistic and relativistic models using the traditional meson and baryon degrees of freedom , to effective field theories , to models based on the underlying quark and gluon degrees of freedom of qcd , including nonperturbative quark cluster models and perturbative qcd .
we review what has been learned from these experiments , and discuss why elastic @xmath3 scattering and photodisintegration seem to require very different theoretical approaches , even though they are closely related experimentally . |
johansen s ( 1988 , 1991 ) likelihood ratio ( lr ) test for cointegration rank is a very popular econometric technique .
however , it is rarely applied to systems of more than three or four variables . on the other hand
, there exist many applications involving much larger systems .
for example , davis ( 2003 ) discusses a possibility of applying the test to the data on seven aggregated and individual commodity prices to test lewbel s ( 1996 ) generalization of the hicks - leontief composite commodity theorem . in a recent study of exchange rate predictability , engel , mark , and west ( 2015 )
contemplate a possibility of determining the cointegration rank of a system of seventeen oecd exchange rates .
banerjee , marcellino , and osbat ( 2004 ) emphasize the importance of testing for no cross - sectional cointegration in panel cointegration analysis ( see breitung and pesaran ( 2008 ) and choi ( 2015 ) ) , and the cross - sectional dimension of modern macroeconomic panels can easily be as large as forty .
the main reason why the lr test is rarely used in the analysis of relatively large systems is its poor finite sample performance . even for small systems ,
the test based on the asymptotic critical values does not perform well ( see johansen ( 2002 ) ) . for large systems , the size distortions become overwhelming , leading to severe over - rejection of the null in favour of too much cointegration as shown in many simulation studies , including ho and sorensen ( 1996 ) and gonzalo and pitarakis ( 1995 , 1999 ) . in this paper , we study the asymptotic behavior of the sample canonical correlations that the lr statistic is based on , when the number of observations and the system s dimensionality go to infinity simultaneously and proportionally . we show that the empirical distribution of the squared sample canonical correlations almost surely converges to the so - called _
wachter distribution _ which also arises , albeit with different parameters , as the limit of the empirical distribution of the squared sample canonical correlations between two independent high - dimensional white noises ( see wachter ( 1980 ) ) .
our analytical findings explain the observed over - rejection of the null hypothesis by the lr test , shed new light on the workings and limitations of the bartlett - type correction approach to the problem ( see johansen ( 2002 ) ) , and lead us to propose a very simple graphical device , similar to the scree plot , for a preliminary analysis of the validity of cointegration hypotheses in large vector autoregressions .
the basic framework for our analysis is standard . consider a @xmath0-dimensional var in the error correction form @xmath1 where @xmath2 and @xmath3 are vectors of deterministic terms and zero - mean gaussian errors with unconstrained covariance matrix , respectively .
the lr statistic for the test of the null hypothesis of no more than @xmath4 cointegrating relationships between the @xmath0 elements of @xmath5 against the alternative of more than @xmath4 such relationships is given by @xmath6 where @xmath7 is the sample size , and @xmath8 are the squared sample canonical correlation coefficients between residuals in the regressions of @xmath9 and @xmath10 on the lagged differences @xmath11 @xmath12 and the deterministic terms . in the absence of the lagged differences and deterministic terms ,
the @xmath13 s are the eigenvalues of @xmath14 where @xmath15 and @xmath16 are the sample covariance matrices of @xmath9 and @xmath17 respectively , while @xmath18 is the cross sample covariance matrix .
more substantively , @xmath19 is the largest possible squared sample correlation coefficient between arbitrary linear combinations of the entries of @xmath9 and the entries of @xmath17 @xmath20 is the largest squared correlation among linear combinations restricted to be orthogonal to those yielding @xmath19 , and so on ( see muirhead ( 1982 ) , ch .
johansen ( 1991 ) shows that the asymptotic distribution of @xmath21 under the asymptotic regime where @xmath22 while @xmath0 remains fixed , can be expressed in terms of the eigenvalues of a matrix whose entries are explicit functions of a @xmath23-dimensional brownian motion .
unfortunately , for relatively large @xmath0 , this asymptotics does not produce good finite sample approximations , as evidenced by the over - rejection phenomenon mentioned above .
therefore , in this paper , we consider a _ simultaneous _ asymptotic regime @xmath24 where both @xmath0 and @xmath7 diverge to infinity so that @xmath25 , \label{asymptotic regime}\ ] ] while @xmath0 remains no larger than @xmath7 .
our monte carlo analysis shows that the corresponding asymptotic approximations are relatively accurate even for such small sample sizes as @xmath26 and @xmath27 . the basic specification for the data generating process ( [ model ] ) that we consider has @xmath28 .
in the next section , we discuss extensions to more general vars with low - rank @xmath29 matrices and additional common factor terms .
we also explain there that our main results hold independently from whether a deterministic vector @xmath2 with fixed or slowly - growing dimension is present or absent from the var .
our study focuses on the behavior of the empirical distribution function ( d.f . ) of the squared sample canonical correlations,@xmath30 where @xmath31 denotes the indicator function .
we find that , under the null of @xmath4 cointegrating relationships , as @xmath24 while @xmath32 almost surely ( a.s . ) , @xmath33 where @xmath34 denotes the weak convergence of d.f.s ( see billingsley ( 1995 ) , p.191 ) , and @xmath35 denotes the _
wachter _ d.f . with parameters
@xmath36 and @xmath37 .
the _ wachter distribution _ was derived by wachter ( 1980 ) as the limit of the empirical distribution of the eigenvalues of the multivariate beta matrix of growing dimension and degrees of freedom .
it has a simple density , which is introduced in the next section , and , for @xmath38 and/or @xmath39 point masses at zero and/or one , respectively .
weak convergence ( [ convergence ] ) and the fact that the squared sample canonical correlations are no larger than unity imply the a.s . convergence of averages @xmath40 for any @xmath41 which is bounded and continuous on @xmath42 $ ] . by definition ,
the likelihood ratio statistic scaled by @xmath43 has this form ( with omitted first @xmath4 summands ) , where @xmath44 is continuous but unbounded function .
therefore , ( [ convergence ] ) can guarantee an a.s .
asymptotic lower bound for the scaled lr statistic . for the lr statistic scaled by @xmath45 we have , almost surely,@xmath46 in contrast , we show that , under the ( standard ) asymptotic regime where@xmath22 while @xmath0 is held fixed@xmath47 @xmath48 concentrates around @xmath49 for relatively large @xmath0 . )
, our weak convergence results only guarantee that @xmath50 is a lower bound , but we conjecture that it is also the limit of the scaled lr statistic as first @xmath22 and then @xmath51 .
this conjecture is supported by monte carlo evidence . ]
a direct calculation reveals that @xmath49 is smaller than the lower bound ( [ bound ] ) , for all @xmath52 , with the gap growing as @xmath53 increases .
that is , the standard asymptotic distribution of the lr statistic is centered at a too low level , especially for relatively large @xmath0 .
this explains the tendency of the asymptotic lr test to over - reject the null .
the reason for the poor centering delivered by the standard asymptotic approximation is that it classifies terms @xmath54 in the asymptotic expansion of the likelihood ratio statistic as @xmath55 when @xmath0 is relatively large , such terms can substantially contribute to the finite sample distribution of the statistic , but will be ignored as asymptotically negligible .
in contrast , the _ simultaneous _ _ asymptotics _ classifies all terms @xmath54 as @xmath56 they are not ignored asymptotically , which improves the centering of the simultaneous asymptotic approximation relative to the standard one .
it is possible to use bound ( [ bound ] ) , with @xmath53 replaced by @xmath57 , to construct a bartlett - type correction factor for the standard lr test .
as we show below , for @xmath58 the value of such a theoretical correction factor is very close to the simulation - based factor described in johansen , hansen and fachin ( 2005 ) .
however , for larger @xmath57 , the values diverge , which may be caused by the fact that johansen , hansen and fachin s ( 2005 ) simulations do not consider combinations of @xmath0 and @xmath7 with @xmath59 and the functional form that they use to fit the simulated correction factors does not work well uniformly in @xmath57 . the weak convergence result ( [ convergence ] )
can be put to a more direct use by comparing the quantiles of the empirical distribution of the squared sample canonical correlations with the quantiles of the limiting wachter distribution . under the null
, the former quantiles plotted against the latter ones should form a 45@xmath60 line , asymptotically .
deviations of such a wachter quantile - quantile plot from the line indicate violations of the null .
creating wachter plots requires practically no additional computations beyond those needed to compute the lr statistic , and we propose to use this simple graphical device for a preliminary analysis of cointegration in large vars . to the best of our knowledge ,
our study is the first to derive the limit of the empirical d.f . of the squared sample canonical correlations between random walk @xmath10 and its innovations @xmath9 .
wachter ( 1980 ) shows that @xmath35 is the weak limit of the empirical d.f . of the squared sample canonical correlations between @xmath61- and @xmath62-dimensional independent gaussian white noises with the size of the sample @xmath63
when @xmath64 so that @xmath65 and @xmath66 .
yang and pan ( 2012 ) show that wachter s ( 1980 ) result holds without the gaussianity assumption for i.i.d .
data with finite second moments .
our proofs do not rely on those previous results .
the values of parameters @xmath36 and @xmath37 in ( [ convergence ] ) imply that the limiting d.f .
for the case of @xmath7 observations of @xmath0-dimensional random walk and its innovations , that we consider in this paper , is the same as the limiting d.f . for the case of @xmath67 observations of two independent white noises - one @xmath0-dimensional and the other @xmath68-dimensional .
it is tempting to think that there exists a deep connection between the two cases , even though we were unable to uncover it so far .
our paper opens up a new direction for the asymptotic analysis of panel var cointegration tests based on the sample canonical correlations .
one such test is developed in larsson and lyhagen ( 2007 ) .
it generalizes larsson , lyhagen , and lothgren ( 2001 ) and groen and kleibergen ( 2003 ) by allowing for cross - unit cointegration , which is important from the empirical perspective .
larsson and lyhagen ( 2007 ) are reluctant to recommend their test for large vars and suggest that for the analysis of relatively large panels it may be better to rely on tighter parameterized models , such as that of bai and ng ( 2004 ) . in the recent review of the panel cointegration literature , choi ( 2015 )
expresses a related concern that , with the large number of cross - sectional units , larsson and lyhagen s test may not work well even with the bartlett s correction .
we speculate that the larsson - lyhagen test , as well as johansen s lr test , based on the _ simultaneous _ asymptotics would work well in panels with comparable cross - sectional and temporal dimensions .
the results of this paper can be used to describe only the appropriate centering of the corresponding test statistics .
the next step would be to derive the _ simultaneous _ asymptotic distribution of scaled deviations of such statistics from the centering values .
we conjecture that the _ simultaneous _ asymptotic distribution of @xmath21 is gaussian , as is often the case for averages of regular functions of eigenvalues of large random matrices ( see bai and silverstein ( 2010 ) and paul and aue ( 2014 ) ) .
we are currently undertaking work to validate this conjecture .
the rest of this paper is structured as follows . in section 2
, we prove the convergence of @xmath69 to the _ wachter _ d.f . and
use this result to derive the asymptotic lower bound for @xmath70 section 3 derives the sequential limit of the empirical d.f . of the squared sample canonical correlations as , first @xmath22 and then @xmath51 .
it then uses differences between the obtained sequential asymptotic limit and the simultaneous limit derived in section 2 to explain the over - rejection phenomenon , and to design a theoretical bartlett - type correction factor for the lr statistic in high - dimensional vars .
section 4 contains a monte carlo study that confirms good finite sample properties of the wachter asymptotic approximation .
it also illustrates the proposed wachter quantile - quantile plot technique using a relatively high - dimensional macroeconomic panel .
section 5 concludes and points out directions for future research .
all proofs are given in the appendix .
consider the following basic version of ( [ model])@xmath71 with @xmath72-dimensional vector of deterministic regressors @xmath2 .
let @xmath73 and @xmath74 be the vectors of residuals from the ols regressions of @xmath9 on @xmath75 and @xmath10 on @xmath75 respectively .
define@xmath76 and let @xmath8 be the eigenvalues of @xmath77 the main goal of this section is to establish the a.s .
weak convergence of the empirical d.f . of the @xmath13 s to the wachter d.f .
, under the null of @xmath4 cointegrating relationships , when @xmath24 .
the wachter distribution with d.f . @xmath35 and parameters @xmath78 has density@xmath79 on @xmath80 \subseteq\left [ 0,1\right ] $ ] with@xmath81 and atoms of size @xmath82 at zero , and @xmath83 at unity .
we shall assume that model ( [ econometricians model ] ) may be misspecified in the sense that the true data generating process is described by the following generalization of ( [ model])@xmath84 where @xmath85 @xmath86 are still i.i.d .
@xmath87 with arbitrary @xmath88 @xmath89 but @xmath90 is not necessarily unity , and @xmath91 is a @xmath92-dimensional vector of deterministic or stochastic variables that does not necessarily coincide with @xmath2 .
for example , some of the components of @xmath91 may be common factors not observed and not modelled by the econometrician .
further , we do not put any restrictions on the roots of the characteristic polynomial associated with ( [ general model ] ) . in particular ,
explosive behavior and seasonal unit roots are allowed .
finally , no constraints on @xmath93 and the initial values @xmath94 , apart from the asymptotic requirements on @xmath92 and @xmath90 as spelled out in the following theorem , are imposed .
[ main]suppose that the data are generated by ( [ general model ] ) , and let @xmath95 @xmath96 be the eigenvalues of @xmath14 where @xmath97 are as defined in ( [ sss ] ) .
further , let @xmath69 be the empirical d.f . of the @xmath13 s , and let @xmath98 $ ] .
if@xmath99 as @xmath24 while @xmath0 remains no larger than @xmath7 , then , almost surely,@xmath100 condition ( [ wistles ] ) requires the number @xmath72 of deterministic regressors in the econometrician s model ( [ econometricians model ] ) , the dimensionality @xmath92 of @xmath91 , the number @xmath4 of the cointegrating relationships under the null , the order @xmath90 of the data generating var , and the dimensionality of the union of the column spaces of the matrix coefficients on further lags in ( [ general model ] ) to be either fixed or growing less than proportionally to the dimensionality @xmath0 or , equivalently , to the sample size @xmath7 .
this condition rules out situations where some or all lags which are omitted from the econometrician s model ( [ econometricians model ] ) have full rank coefficients @xmath29 .
the simplest special situation where ( [ wistles ] ) is clearly satisfied corresponds to the pure random walk data @xmath101 .
the reason why the limit of the empirical d.f .
@xmath102 does not change when the data generating process ( [ general model ] ) changes so that ( [ wistles ] ) remains true is that the corresponding changes in the matrix @xmath103 have rank that is less than proportional to @xmath0 ( and to @xmath7 ) . by the so - called rank inequality ( theorem a43 in bai and silverstein ( 2010 ) ) , the lvy distance between the empirical d.f . of eigenvalues corresponding to versions of @xmath103 that differ by a matrix of rank @xmath104
is no larger than @xmath105 which converges to zero as @xmath106 since the lvy distance metrizes the weak convergence ( see billingsley ( 1995 ) , problem 14.5 ) , the limiting d.f . is not affected . for further details , see the proof of theorem [ main ] in the appendix . in standard cases where
@xmath2 is represented by @xmath107 it is customary to impose restrictions on @xmath108 so that there is no quadratic trend in @xmath5 ( see johansen ( 1995 ) , ch .
then , the lr test of the null of @xmath4 cointegrating relationships is based on the eigenvalues of @xmath109 defined similarly to @xmath103 by replacing @xmath10 with @xmath110 and regressing @xmath9 and @xmath110 on constant only to obtain @xmath73 and @xmath111 the empirical distribution function of so modified eigenvalues still converges to @xmath112 because the difference between matrices @xmath113 and @xmath103 has small rank .
figure [ illustrationmain ] shows quantile plots of the wachter distribution with parameters @xmath114 and @xmath115 for different values of @xmath53 . for @xmath116
the dimensionality of the data constitutes 20% of the sample size .
the corresponding wachter limit of @xmath117is supported on @xmath118 $ ] .
in particular , we expect @xmath19 be larger than @xmath119 for large @xmath120 even in the absence of any cointegrating relationships . for @xmath121
the upper boundary of support of the wachter limit is unity .
this accords with gonzalo and pitarakis ( 1995 , lemma 2.3.1 ) finding that as @xmath122 @xmath123 for @xmath124 the wachter limit has mass @xmath125 at unity .
illustrationmain.eps wachter ( 1980 ) derives @xmath35 as the weak limit of the empirical d.f .
of eigenvalues of the @xmath0-dimensional beta matrix @xmath126 with @xmath127 degrees of freedom as @xmath128 so that @xmath129 and @xmath130 the eigenvalues of multivariate beta matrices are related to many important concepts in multivariate statistics , including canonical correlations , multiple discriminant ratios , and manova . in particular , the squared sample canonical correlations between @xmath61- and @xmath62-dimensional independent gaussian samples of size @xmath131 are jointly distributed as the eigenvalues of @xmath132 where @xmath133 and @xmath134 .
therefore , their empirical d.f .
weakly converges to @xmath135 with @xmath136 and @xmath137 as mentioned above .
since the squared canonical correlations in theorem [ main ] are between random walk and its innovations rather than independent white noises , the convergence to the wachter distribution came to us as a pleasant surprise . in the context of multiple discriminant analysis ,
wachter ( 1976b ) proposes to use a quantile - quantile ( qq ) plot , where the multiple discriminant ratios are plotted against quantiles of @xmath138 , as a simple graphical method that helps one recognize hopeless from promising analyses at an early stage .
a plot that clearly deviates from the 45@xmath60 line suggests that the data are at odds with the null hypothesis of the homogeneous population , and a further analysis of the heterogeneity is useful . nowadays ,
such qq plots are called _ wachter plots _ ( see johnstone ( 2001 ) ) .
theorem [ main ] implies that the wachter plot can be used as a simple preliminary assessment of cointegration hypotheses in large vars .
as an illustration , figure [ wachterplot ] shows a wachter plot of the simulated sample squared canonical correlations corresponding to a @xmath139-dimensional var(1 ) model ( [ econometricians model ] ) with @xmath140 so that there are three white noise and seventeen random walk components of @xmath5 .
no deterministic terms are included .
we set @xmath141 so that @xmath142 .
the graph clearly shows three canonical correlations that destroy the 45@xmath60 line fit , so that the null hypothesis of no cointegration is compromised .
wachterplot.eps theorem [ main ] does not provide any explanation to the fact that exactly three canonical correlations deviate from the 45@xmath60 line in figure [ wachterplot ] .
to interpret deviations of the wachter plots from the 45@xmath60 line , it is desirable to investigate behavior of @xmath143 under various alternatives .
so far , we were able to obtain a clear result only for the extreme alternative , where @xmath5 is a vector of independent white noises .
under such an alternative , @xmath144 we plan to publish a full proof of this and some related results elsewhere .
interestingly , for @xmath121 the wachter limits ( [ 5again ] ) and ( [ white noise ] ) corresponding to random walk and white noise nulls , respectively , coincide . hence , as @xmath53 approaches
@xmath145 not only the largest sample canonical correlation converges to one and the lr test breaks down , but also the wachter plot looses the ability to differentiate between opposite cointegration hypotheses . for smaller values of @xmath146
however , the wachter limits ( [ 5again ] ) and ( [ white noise ] ) become well separated .
we provide monte carlo analysis of the behavior of @xmath147 under some alternative hypotheses in section 4 below .
the almost sure weak convergence of @xmath69 established in theorem [ main ] implies the almost sure convergence of bounded continuous functionals of @xmath148 an example of such a functional is the scaled pillai - bartlett statistic for the null of no more than @xmath4 cointegrating relationships ( see gonzalo and pitarakis ( 1995))@xmath149 which is asymptotically equivalent to the lr statistic under the standard asymptotic regime where @xmath0 is fixed and @xmath22 . since , by definition , @xmath150 , $ ] we have@xmath151 where @xmath41 is the bounded continuous function@xmath152{cc}0 & \text{for } \lambda<0\\ \lambda & \text{for } \lambda\in\left [ 0,1\right ] \\ 1 & \text{for } \lambda>1 .
\end{array } \right . .\
] ] as long as @xmath153 as @xmath154 the second term on the right hand side of ( [ pbrepresentation ] ) converges to zero
. therefore , theorem [ main ] implies that @xmath155 almost surely converges to @xmath156 a direct calculation based on ( [ densityw ] ) , which we report in the supplementary appendix , yields the following corollary .
[ pb]under the assumptions of theorem [ main ] , as @xmath157 almost surely,@xmath158 a similar analysis of the lr statistic is less straightforward because@xmath159 and @xmath160 is unbounded on @xmath161 .$ ] in fact , for @xmath162 @xmath21 is ill - defined because a non - negligible proportion of the squared sample canonical correlations exactly equal unity .
however for @xmath163 we can obtain the almost sure asymptotic lower bound on @xmath164 note that for such @xmath146 the upper bound of the support of @xmath165 equals @xmath166 .
let @xmath167{cc}0 & \text{for } \lambda<0\\ \log(1-\lambda ) & \text{for } \lambda\in\left [ 0,b_{+}\right ] \\
\log(1-b_{+ } ) & \text{for } \lambda >
b_{+}. \end{array } \right . \label{truncated log}\ ] ] clearly , @xmath168 is a bounded continuous function and@xmath169 hence , we have the following a.s .
lower bound on @xmath170 ( the corresponding calculations are reported in the supplementary appendix ) .
[ lrsim]under the assumptions of theorem [ main ] , for @xmath163 as @xmath154 almost surely,@xmath171 [ conjectureremark]we conjecture that the lower bound reported in the corollary is , in fact , the a.s
. limit of @xmath164 to prove this conjecture , one needs to show that @xmath172 is almost surely bounded away from unity so that the unboundedness of @xmath173 is not consequential .
we leave this as an important topic for future research
. corollaries [ pb ] and [ lrsim ] suggest appropriate centering points for pb and lr statistics scaled by @xmath174 for relatively large and comparable @xmath0 and @xmath175 as we show in the next section , the standard asymptotic distribution of the scaled pb and lr statistics are likely to concentrate around very different points when @xmath0 becomes large .
as will be seen below , this difference sheds new light on the over - rejection phenomenon discussed above and on the workings and limitations of the bartlett correction for the lr statistic . to study the concentration of the standard asymptotic distributions of the scaled pb and lr statistics
as @xmath0 grows , we will consider the _ sequential _ asymptotic regime , where first @xmath176 and then @xmath51 .
to obtain useful results under the sequential asymptotics , we shall study eigenvalues of the scaled matrix@xmath177 note that under the simultaneous asymptotic regime @xmath154 the asymptotic behavior of the scaled and unscaled eigenvalues is the same up to the factor @xmath178 .
however , as first @xmath22 while @xmath0 remains fixed , the unscaled eigenvalues converge to zero , while scaled ones do not .
we shall denote the empirical d.f . of eigenvalues of the scaled matrix as @xmath179 .
without loss of generality , we focus on the case of simple data generating process @xmath180 and on the situation , where the econometrician does not include any deterministic regressors in his or her model , that is @xmath181 . there is no loss of generality in such simplifications because , as follows from lemma [ misspec ] and the rank inequality used in the proof of lemma [ rank ] in the appendix , the lvy distance between the versions of @xmath179 that correspond to the simplified and the general cases is bounded from above by a fixed multiple of @xmath182 .
we shall assume that the latter expression goes to zero as @xmath183 therefore , whatever the sequential asymptotic limit of @xmath184 is under the above simplification , it must also be the sequential asymptotic limit under the general case . for simplicity , in the rest of this section , we shall assume that @xmath185 and will consider statistics @xmath186 rather than more general @xmath70 under the above simplifications , johansen s ( 1988 , 1991 ) results imply that , as @xmath22 while @xmath0 is held fixed , the eigenvalues of the scaled matrix ( [ rawmatrix ] ) jointly converge in distribution to the eigenvalues of@xmath187 where @xmath188 is a @xmath0-dimensional brownian motion .
we denote the eigenvalues of ( [ johansenlimit ] ) as @xmath189 and their empirical d.f . as @xmath190
it is not unreasonable to expect that , as @xmath51 , @xmath191 becomes close to the limit of the empirical distribution of eigenvalues of ( [ rawmatrix ] ) under a simultaneous , rather than sequential , asymptotic regime @xmath192 where @xmath193 is close to zero .
we shall denote such a limit as @xmath194 this expectation turns out to be correct in the sense that the following theorem holds .
[ levy ] let @xmath195 be the weak limit as @xmath196 of @xmath197 then , as @xmath198 @xmath199 weakly converges to @xmath200 in probability .
importantly , the weak limit @xmath195 is not the wachter d.f .
instead , the following proposition holds .
[ mpproposition ] @xmath195 corresponds to a distribution supported on @xmath201 $ ] with @xmath202 and having density@xmath203 a reader familiar with large random matrix theory ( see bai and silverstein ( 2010 ) ) might recognize that @xmath195 is the cumulative distribution function of the continuous part of a special case of the marchenko - pastur distribution ( marchenko and pastur ( 1967 ) ) .
the general marchenko - pastur distribution has density@xmath204 over @xmath201 $ ] with @xmath205 and a point mass @xmath206 at zero .
density ( [ mp density ] ) is two times @xmath207 with @xmath208 and @xmath209 the multiplication by two is needed because the mass @xmath210 at zero is not a part of the distribution @xmath211 .
recall that , as @xmath22 while @xmath0 remains fixed , the lr statistic converges in distribution to @xmath0 times the trace of matrix ( [ johansenlimit]):@xmath212 on the other hand , according to theorem [ levy ] , for any @xmath213 and all sufficiently large @xmath214@xmath215 a direct calculation , which we report in the supplementary appendix , shows that @xmath216 hence , we have the following corollary . [
table reduction]as first @xmath176 and then @xmath217 the lower probability bound on @xmath218 is unity in the following sense . as @xmath22
while @xmath0 is held fixed , @xmath218 converges in distribution to @xmath219 further , for any @xmath220 and all sufficiently large @xmath214 the probability that @xmath221 is no smaller than @xmath222 is no smaller than @xmath223 the reason why we only claim the lower bound on @xmath224 is that theorem [ levy ] is silent about the behavior of the individual eigenvalues @xmath189 the largest of which may , in principle , quickly diverge to infinity .
we suspect that @xmath49 is not just the lower bound , but also the probability limit of @xmath225 , so that the sequential probability limit of @xmath218 is unity .
verification of this conjecture requires more work , similar to that discussed in remark [ conjectureremark ] .
corollary [ table reduction ] is consistent with the numerical finding of johansen , hansen and fachin ( 2005 , table 2 ) that , as @xmath7 becomes large while @xmath0 is being fixed , the sample mean of the lr statistic is well approximated by a polynomial @xmath226 ( see also johansen ( 1988 ) and gonzalo and pitarakis ( 1995 ) ) .
the value of @xmath227 depends on how many deterministic regressors are included in the var .
our theoretical result captures only the ` highest order ' sequential asymptotic behavior of the lr statistic , which remains ( bounded below by ) @xmath228 independent on the number of the deterministic regressors .
another piece of numerical support for @xmath228 being not only the lower bound but also the first order sequential asymptotic approximation to the lr statistic is provided by the tables of the asymptotic critical values for johansen s lr test ( see , for example , mackinnon , haug and michelis ( 1999 ) ) .
the critical values in such tables become uncomfortably large for @xmath229 .
of course , the reason for such an unpleasant growth is that those critical values are of order @xmath228 .
the transformation@xmath230 makes the lr statistic ` well - behaved ' under the sequential asymptotics .
the division by @xmath0 reduces the ` second order behavior ' to @xmath231 while subtracting @xmath68 eliminates the remaining explosive ` highest order term ' .
we report the corresponding transformed 95% critical values alongside the original ones in table [ cv ] .
the transformed critical values resemble 97 - 99 percentiles of @xmath232 .
since the lr test is one - sided , the resemblance is coincidental .
however , we do expect the sequential asymptotic distribution of the transformed lr statistic ( as well as its simultaneous asymptotic distribution ) to be normal ( possibly with non - zero mean and non - unit variance ) .
our expectation is based on the fact that @xmath233 behaves as the eigenvalue average ( see ( [ tlimit ] ) ) , which is a special case of the so - called linear spectral statistic .
the asymptotic normality of linear spectral statistics for relatively simple classes of high - dimensional random matrices is a well established result in the large random matrix theory ( see bai and silverstein ( 2010 ) ) . extending it to the linear spectral statistics of matrices of form ( [ johansenlimit ] )
is left as an important direction for future research .
[ c]l|l|l@xmath0 & unadjusted cv & cv/@xmath234 + & & + & & + & & + & & + & & + & & + & & + & & + & & + & & + & & + & & + in this subsection , let us assume that the following conjecture holds .
[ conjecture]the simultaneous and sequential asymptotic lower bounds for the scaled lr statistics derived in corollaries [ lrsim ] and [ table reduction ] represent the corresponding simultaneous and sequential asymptotic limits . specifically , for @xmath235,@xmath236@xmath237 figure [ overrejection ] plots the right hand side of ( [ simlim ] ) against the value of @xmath238 as demonstrated by the monte carlo analysis of the next section , in finite samples with comparable values of @xmath0 and @xmath7 , simultaneous asymptotics
provides a better approximation to the finite sample behavior of the lr statistic than the sequential asymptotics . therefore , ` typical ' finite sample values of the lr statistic are concentrated around the solid line in figure [ overrejection ] , and above the dashed line , which represents the points of concentration of the ` standard ' asymptotic critical values for the lr test . in other words ,
the standard asymptotic distribution of the lr statistic is centered at a too low level .
this leads to the over - rejection of the null of no cointegration .
overrejection1.eps gonzalo and pitarakis ( 1995 ) propose an interesting approach to address the problem .
using monte carlo , they find that , in contrast to the lr test , the pillai - bartlett test based on the pb statistic under - rejects the null .
therefore , they propose to test cointegration hypotheses using the average of the lr and pb statistics . according to corollary [ pb ] , under the simultaneous asymptotics @xmath239 almost surely .
this convergence holds independent on whether conjecture [ conjecture ] is true or not .
the fact that @xmath240 is smaller than one , explains the under - rejection of the test based on the pb statistic .
more interestingly , the average of the simultaneous asymptotic limits of the lr and pb statistics ( divided by @xmath228 ) turns out to be numerically close to one , and hence to the point of the concentration of the standard critical values ( divided by @xmath228 ) , at least for @xmath241 figure [ pbaver ] shows such an average .
this explains the much better performance of the ( lr+pb)/2 test relative to the lr test in gonzalo and pitarakis ( 1995 ) monte carlo experiments .
pbaverage.eps a more systematic and popular approach to addressing the over - rejection problem is based on the bartlett - type correction of the lr statistic .
it was explored in much detail in various important studies , including johansen ( 2002 ) .
the idea is to scale the lr statistic so that its finite sample distribution better fits the asymptotic distribution of the unscaled statistic .
specifically , let @xmath242 be the mean of the asymptotic distribution under the fixed-@xmath0 , large-@xmath7 asymptotic regime .
then , if the finite sample mean , @xmath243 , satisfies@xmath244 the scaled statistic is defined as @xmath245 by construction , the match between the scaled mean and the original asymptotic mean is improved by an order of magnitude .
although , as shown by jensen and wood ( 1997 ) in the context of unit root testing , the match between higher moments does not improve by an order of magnitude , it may become substantially better ( see nielsen ( 1997 ) ) .
a theoretical analysis of the adjustment factor @xmath246 can be rather involved .
in general , @xmath247 will depend not only on @xmath0 , but also on all the parameters of the var .
however , for gaussian var(1 ) without deterministic terms , under the null of no cointegration , @xmath247 depends only on @xmath0 . for @xmath248 the exact expression for @xmath247
was derived in larsson ( 1998 ) .
given the difficulty of the theoretical analysis of @xmath249 johansen ( 2002 ) proposes to numerically evaluate the bartlett correction factor @xmath250 by simulation .
johansen , hansen and fachin ( 2005 ) simulate @xmath251 for various values of @xmath252 and @xmath253 and fit a function of the form@xmath254 \right\}\ ] ] to the obtained results .
for relatively large values of @xmath255 the term @xmath256 $ ] in the above expression is small .
when it is ignored , the fitted function becomes particularly simple:@xmath257 our simultaneous and sequential asymptotic results shed light on the workings of @xmath258 given that conjecture [ conjecture ] holds,@xmath259 therefore , for non - negligible @xmath260 we expect @xmath251 to be well approximated by @xmath261 where @xmath262 is the finite sample analog of @xmath263 figure [ bartlettnew ] superimposes the graphs of @xmath264 and @xmath265 as functions of @xmath266 for @xmath267 there is a strikingly good match between the two curves , with the maximum distance between them @xmath268 .
for @xmath269 the quality of the match quickly deteriorates .
this can be explained by the fact that all @xmath120-pairs used in johansen , hansen and fachin s ( 2005 ) simulations are such that @xmath270 .
bartlettnew.eps further , the good match between @xmath265 and @xmath264 observed for @xmath270 would be impossible had johansen , hansen and fachin s ( 2005 ) specified the bartlett correction factor as a linear function of @xmath57 .
note that the standard theoretical choice for the bartlett correction factor , @xmath246 from ( [ bartlett theory ] ) , can be viewed as a linear function of @xmath57 with a slope possibly varying with @xmath0 .
this is obvious when @xmath271 is represented as @xmath272 with @xmath273 .
figure [ bartlettnew ] shows that such theoretical correction factors can not work well uniformly with respect to @xmath57 .
uniformly good correction factors must include terms @xmath54 with @xmath274 . under the fixed-@xmath214 large-@xmath7 asymptotics , such terms are of lower order than @xmath275 but under the simultaneous asymptotics , they are of order @xmath276 .
although the bartlett - type correction approach may deliver good results for high - dimensional systems with carefully chosen correction factor , we believe that tests based on the simultaneous asymptotics of the appropriately scaled and centered lr statistic would be preferable for relatively large @xmath0 .
in this section , we describe results of small - scale monte carlo experiments that assess the finite sample quality of the wachter asymptotic approximation . in addition
, we illustrate the wachter qq plot technique using a macroeconomic dataset of relatively high dimensions .
first , we generate pure random walk data with zero starting values for @xmath277 and @xmath278 throughout this section , the analysis is based on 1000 monte carlo replications .
the generated random walk data are ten - dimensional so that there are ten corresponding squared sample canonical correlations , @xmath13 .
figure [ mcboxplot ] shows the tukey boxplots summarizing the mc distribution of each of the @xmath95 @xmath279 ( sorted in the ascending order throughout this section)@xmath280 the boxplots are superimposed with the quantile function of the wachter limit with @xmath142 for the left panel and @xmath281 for the right panel .
precisely , for @xmath282 we show the value the @xmath283 quantile of the wachter limit . for @xmath284
these are the 5-th,15-th, ... ,95-th quantiles of @xmath285 even for such small values of @xmath0 and @xmath255 the theoretical quantiles track the location of the mc distribution of the empirical quantiles very well .
the smallest sample canonical correlation is an exception .
its distribution lies mostly below the corresponding theoretical quantile .
the dispersion of the mc distributions around the theoretical quantile is quite large for the chosen small values of @xmath0 and @xmath175 to see how such a dispersion changes when @xmath0 and @xmath7 increase while @xmath57 remains fixed , we generated pure random walk data with @xmath286 and @xmath287 for @xmath288 , and with @xmath289 and @xmath290 for @xmath291 instead of reporting the tukey boxplots , we plot only the 5-th and 95-th percentiles of the mc distributions of the @xmath95 @xmath292 against @xmath293 quantiles of the corresponding wachter limit .
the plots are shown on figure [ mcqqplot ] .
mcboxplot.eps mcqqplot.eps we see that the [ 5%,95% ] ranges of the mc distributions of @xmath294 are still considerably large for @xmath295 these ranges become much smaller for @xmath296 interestingly , the distribution of @xmath19 remains below the wachter limit even for @xmath296 this does not contradict our theoretical results because a weak limit of the empirical distribution of @xmath13 s is not affected by an arbitrary change in a finite ( or slowly growing ) number of them .
in fact , we find it somewhat surprising that only the distribution of @xmath19 is not well - alligned with the derived theoretical limit . our proofs are based on several low rank alterations of the matrix @xmath103 , and there is nothing in them that guarantee that only one eigenvalue of @xmath103 behaves in a special way . in future work , it would be interesting to investigate the behavior of @xmath19 and other extreme eigenvalues of @xmath103 theoretically .
next , we explore the effect of the deterministic regressor on the quality of the wachter approximation .
we generate data with and without constant in the data generating process ( [ general model ] ) .
that is , we consider two cases : @xmath297 and @xmath298 the coefficient @xmath299 on @xmath91 is a @xmath300 vector independent across different mc replications .
we also consider two models ( [ econometricians model ] ) contemplated by the econometrician : one with @xmath301 , and the other with @xmath302 if @xmath303 the econometrican s model is misspecified .
figure [ mcmisspec ] shows the wachter plots similar to those reported in figure [ mcqqplot ] .
the dimensions of the data are @xmath304 and @xmath305 if the data generating process ( dgp ) contains constant ( @xmath297 ) , but the econometrician does not include it in his or her model , then the largest @xmath306 @xmath307 start to significantly deviate from the 45@xmath60 line on the wachter plot ( lower right panel ) . if the econometrician s model is over - specified ( lower left panel ) , there are no dramatic deviations from the line .
mcqqdeterministic.eps our next monte carlo experiment simulates data that are not random walk . instead ,
the data are stationary var(1 ) with zero mean , zero initial value , and @xmath308 we consider three cases of @xmath309 @xmath310 and @xmath311 figure [ mcalternatives ] shows the wachter plots with solid lines representing 5th and 95-th percentiles of the mc distributions of @xmath312 plotted against the @xmath313 quantiles of the corresponding wachter limit . the dashed line correspond to the null case where the data are pure random walk ( shown for comparison ) .
mcqqalternatives.eps the lower panel of the figure corresponds to the most persistent alternative with @xmath314 samples with @xmath304 seem to be too small to generate substantial differences in the behavior of wachter plots under the null and under such persistent alternatives .
the less persistent alternative with @xmath315 is easily discriminated against by the wachter plot for @xmath288 ( left panel ) .
the discrimination power of the plot for @xmath316 ( central panel ) is weaker . for @xmath317
there is still some discrimination power left , but the location of the wachter plot under alternative switches the side relative to the 45@xmath60 line .
the plots easily discriminate against white noise ( @xmath318 ) alternative for @xmath142 and @xmath116 but not for @xmath319 in accordance to the result that we announced above , and plan to publish elsewhere , the wachter limit for @xmath281 approximates equally well the empirical distribution of the squared sample canonical correlations based on random walk and on white noise data .
results reported in figure [ mcalternatives ] indicate that for relatively small @xmath0 and @xmath260 wachter plots can be effective in discriminating against alternatives to the null of no cointegration , where the cointegrating linear combinations of the data are not very persistent .
further , tests of no cointegration hypothesis that may be developed using simultaneous asymptotics would probably need to be two - sided .
it is because the location of the wachter plot under the alternative may switch sides relative to the 45@xmath60 depending on the persistence of the data under the alternative .
finally , cases with @xmath53 close to 1/2 must be analyzed with much care . for such cases ,
the behavior of the sample canonical correlations become similar under extremely different random walk and white noise data generating processes .
furthermore , the largest sample canonical correlations are close to unity , which can result in an unstable behavior of the lr statistic .
our final mc experiment studies the finite sample behavior of the scaled lr statistic @xmath320 we simulate pure random walk data with @xmath26 and @xmath321 and @xmath7 varying so that @xmath57 equals 1/10,2/10, ... ,5/10 .
corollary [ lrsim ] shows that the simultaneous asymptotic lower bound on @xmath218 has form@xmath322 figure [ mclr ] shows the tukey boxplots of the mc distributions of @xmath218 corresponding to @xmath323 with @xmath26 ( left panel ) , and @xmath321 ( right panel ) .
the boxplots are superimposed with the plot of the line representing the above displayed formula for the lower bound ( with @xmath53 replaced by @xmath57 ) . for the case
@xmath324 we also show ( horizontal dashed line ) the 95% asymptotic critical value ( scaled by @xmath325 ) of the standard johansen trace test taken from mackinnon et al ( 1999 , table ii ) . for @xmath326 critical values for the standard test
are not available , and we show the dashed horizontal line at unit height instead . this is the sequential asymptotic lower bound on @xmath327 as established in corollary [ table reduction ] .
mclr.eps the reported results support our conjecture that the simultaneous asymptotic lower bound ( [ simultaneous lower bound ] ) is , in fact , the simultaneous asymptotic limit of @xmath218 for @xmath328 interestingly , the bound is located near the center of the mc distribution of the scaled lr statistic even for the case @xmath319 the left panel of figure [ mclr ] illustrates the over - rejection phenomenon .
the horizontal dashed line that corresponds to the 95% critical value of the standard test is just above the interquartile range of the mc distribution of @xmath224 for @xmath329 is below this range for @xmath330 and is below all 1000 mc replications of the scaled lr statistic for @xmath331 .
although the lower bound ( [ simultaneous lower bound ] ) seems to provide a very good centering point for the scaled lr statistic , the mc distribution of this statistic is quite dispersed around such a center for @xmath332 as discussed above , we suspect that the scaled statistic centered by ( [ simultaneous lower bound ] ) and appropriately rescaled has gaussian simultaneous asymptotic distribution .
optimistically , the tukey plots on figure [ mclr ] , that correspond to @xmath235 , look reasonably symmetric although some skewness is present for the left panel where @xmath332 our first example uses @xmath333 quarterly observations ( 1973q2 - 1998q4 , with the initial observation 1973q1 ) on bilateral us dollar log nominal exchange rates for @xmath334 oecd countries : australia , austria , belgium , canada , denmark , finland , france , germany , japan , italy , korea , netherlands , norway , spain , sweden , switzerland , and the united kingdom .
the data are as in engel , mark , and west ( 2015 ) , and were downloaded from charles engel s website at http : // www.ssc.wisc.edu / cengel /.
that data are available for a longer time period up to 2008q1 , but we have chosen to use only the early sample that does not include the euro period .
engel , mark , and west ( 2015 ) point out that log nominal exchange rates are well modelled by random walk , but may be cointegrated , which can be utilized to improve individual exchange rate forecasts relative to the random walk forecast benchmark . they propose to estimate the common stochastic trends in the exchange rates by extracting a few factors from the panel . in principle , the number of factors to extract can be determined using johansen s test for cointegrating rank , but engel , mark , and west ( 2015 ) do not exploit this possibility , referring to ho and sorensen ( 1996 ) that indicates poor performance of the test for large @xmath335 figure [ erfigure ] shows the wachter plot for the log nominal exchange rate data .
the squared sample canonical correlations are computed as the eigenvalues of @xmath14 where @xmath97 are defined as in ( [ sss ] ) with @xmath73 and @xmath74 being the demeaned changes and the lagged levels of the log exchange rates , respectively .
the dashed lines correspond to the 5-th and 95-th percentiles of the mc distribution of the squared canonical correlation coefficients under the null of no cointegration . precisely , we generated data from model ( [ econometricians model ] ) with @xmath336 @xmath333 , @xmath337 @xmath338 and @xmath108 being i.i.d .
@xmath300 vectors across the mc repetitions .
log exchange rates for 1973q1 was used as the initial value of the generated series .
erfigure.eps the figure shows a mild evidence for cointegration in the data with the largest five @xmath13 s being close to the corresponding 95-th percentiles of the mc distributions .
if we interpret this as the existence of five cointegrating relationships in the data , we would be lead to conclude that there are twelve stochastic trends .
recall , however , that the ability of the wachter plot to differentiate against highly persistent cointegration alternatives with @xmath339 is very low , so there well may be many more cointegrating relationships in the data .
whatever such relationships are , the deviations from the corresponding long - run equilibrium are probably highly persistent as no dramatic deviations from the 45@xmath60 line are present in the wachter plot . very different wachter plots ( shown in figure [ ipcpi ] ) correspond to the log industrial production ( ip ) index data and the log consumer price index ( cpi ) data for the same countries plus the us .
these data are still the same as in engel , mark , and west ( 2015 ) .
we used the long sample 1973q2:2008q1 @xmath340 because the ip and cpi data are not affected by the introduction of the euro to the same degree as the exchange rate data . for
the cpi data , we included both intercept and trend in model ( [ econometricians model ] ) for the first differences because the level data seem to be quadratically trending .
the plots clearly indicate that the ip and cpi data are either stationary or cointegrated with potentially many cointegrating relationships , short run deviations from which are not very persistent .
in this paper , we consider the simultaneous , large-@xmath214 large-@xmath7 , asymptotic behavior of the squared sample canonical correlations between @xmath0-dimensional random walk and its innovations .
we find that the empirical distribution of these squared sample canonical correlations almost surely weakly converges to the so - called _ wachter distribution _ with parameters that depend only on the limit of @xmath57 as @xmath106 in contrast , under the sequential asymptotics , when first @xmath22 and then @xmath198 we establish the convergence in probability to the so - called marchenko - pastur distribution .
the differences between the limiting distributions allow us to explain from a theoretical point of view the tendency of the lr test for cointegration to severely over - reject the null when the dimensionality of the data is relatively large .
furthermore , we derive a simple analytic formula for the bartlett - type correction factor in systems with relatively large @xmath57 ratio . we propose a quick graphical method , the wachter plot , for a preliminary analysis of cointegration in large - dimensional systems .
the monte carlo analysis shows that the quantiles of the wachter distribution constitute very good centering points for the finite sample distributions of the corresponding squared sample canonical correlations .
the quality of the centering is excellent even for such small @xmath0 and @xmath7 as @xmath26 and @xmath341 however , for such small values of @xmath0 and @xmath255 the empirical distribution of the squared sample canonical correlation can considerably fluctuate around the wachter limit . as @xmath0 increases to 100 , the fluctuations become numerically very small .
our analysis leaves many open questions .
first , it is very important to study the fluctuations of the empirical distribution around the wachter limit .
we conjecture that linear combinations of reasonably smooth functions of the squared sample canonical correlations , including the @xmath342 used by the lr statistic , will be asymptotically gaussian after appropriate centering and scaling .
the centering can be derived from the results obtained in this paper .
a proof of the asymptotic gaussianity would require different methods from those used here .
we are currently investigating this research direction .
further , it would be important to remove the gaussianity assumption on the data .
we believe that the existence of the finite fourth moments is a sufficient condition for the validity of the wachter limit .
next , it would be interesting to study the simultaneous asymptotic behavior of a few of the largest sample canonical correlations .
this may lead to a modification of johansen s maximum eigenvalue test .
another interesting direction of research is to study situations where the number of cointegrating relationships under the null is growing proportionally with @xmath0 and @xmath175 the simultaneous asymptotics of the empirical distribution of the squared sample canonical correlations under various alternatives , as well under the null in var(k ) with @xmath343 , also deserves further study .
still another , totally different , research direction is to investigate the quality of bootstrap when @xmath0 is large .
our own very preliminary analysis indicates that the currently available non - parametric bootstrap procedures ( see , for example , cavaliere , rahbek , and taylor ( 2012 ) ) do not work well for @xmath57 as large as , say , 1/3 .
however , further analysis is needed before we can claim any specific results .
we hope that this paper opens up an interesting and broad area for future research .
let @xmath344 and @xmath345 be distribution functions that may depend on @xmath0 and @xmath7 and are possibly random .
we shall call them asymptotically equivalent if the a.s .
weak convergence @xmath346 to some non - random d.f .
@xmath347 implies similar a.s .
weak convergence for @xmath348 and vice versa .
let @xmath349 and @xmath350 with @xmath351 be , possibly random , matrices that may depend on @xmath0 and @xmath7 such that @xmath349 and @xmath350 are a.s .
positive definite for @xmath352 below , we shall often refer to the following auxiliary lemma .
[ rank]if , almost surely , @xmath353 as @xmath24 for @xmath354 then @xmath344 and @xmath355 are asymptoically equivalent , where @xmath344 and @xmath356 are the empirical d.f .
of eigenvalues of @xmath357 and @xmath358 respectively .
let @xmath359 the a.s .
convergence @xmath353 implies the a.s .
convergence @xmath360 on the other hand , by the rank inequality ( theorem a43 in bai and silverstein ( 2010 ) ) , @xmath361 where @xmath362 is the lvy distance between @xmath344 and @xmath363 recall that the lvy distance metrizes the weak convergence .
therefore , the almost sure convergence @xmath364 yields the asymptotic equivalence of @xmath344 and @xmath365 now , let @xmath366 and @xmath367 and let@xmath368 since @xmath73 and @xmath369 which enter the definition ( [ sss ] ) of @xmath97 , are the residuals in the regressions of @xmath9 on @xmath2 and @xmath10 on @xmath2 , respectively , we have@xmath370 by assumption , @xmath371 as @xmath154 so that by lemma [ rank ] , @xmath102 is asymptotically equivalent to the empirical d.f .
of eigenvalues of @xmath372 therefore , we may and will replace @xmath73 and @xmath74 in the definitions ( [ sss ] ) of @xmath97 by @xmath9 and @xmath17 respectively , without loss of generality .
furthermore , scaling @xmath97 by @xmath7 does not change the product @xmath373 and thus , in the rest of the proof , we shall work with@xmath374 next , we show that , still without loss of generality , we may replace the data generated process ( [ general model ] ) by pure random walk with zero initial value .
indeed , let @xmath375,$ ] where @xmath376 are arbitrary and @xmath5 with @xmath377 are generated by ( [ general model ] ) .
further , let @xmath378 be zero vectors , @xmath379 for @xmath380 and @xmath381.$ ] a proof of this lemma is given in the supplementary appendix .
it is based on the representation of @xmath5 as a function of the initial values , @xmath383 and @xmath384 ( see theorem 2.1 in johansen ( 1995 ) ) , and requires only elementary algebraic manipulations .
lemmas [ misspec ] and [ rank ] together with the assumption ( [ wistles ] ) imply that replacing @xmath385 and @xmath10 in ( [ sssnew ] ) by @xmath386 and @xmath387 respectively , does not change the weak limit of @xmath388 hence , in the rest of the proof of theorem [ main ] , without loss of generality , we shall assume that the data are generated by@xmath389 assuming that @xmath13 s are the eigenvalues of @xmath103 with @xmath97 satisfying ( [ sssnew ] ) and ( [ rw dgp ] ) , we can interpret them as the squared sample canonical correlations between lagged values of a random walk @xmath10 and its current innovations @xmath3 . since the sample canonical correlations are invariant with respect to the multiplication of the data by any invertible matrix , we assume without loss of generality that the variance of @xmath3 equals @xmath390 further , we assume that @xmath7 is even . the case of odd @xmath7
can be analyzed similarly , and we omit it to save space .
let @xmath391 $ ] and let @xmath392 be the upper - triangular matrix with ones above the main diagonal and zeros on the diagonal .
then @xmath393 , $ ] so that @xmath394 we shall show that the empirical d.f . of the @xmath13 s
, @xmath395 is asymptotically equivalent to the empirical d.f .
@xmath396 of eigenvalues of @xmath397 where @xmath398 @xmath399 is a diagonal matrix,@xmath400 and @xmath401 is a block - diagonal matrix,@xmath402 here @xmath403 is the 2-dimensional identity matrix , and @xmath404 are defined as follows .
let @xmath405 then for @xmath406@xmath407{cc}\cos j\theta & -\sin j\theta\\ \sin j\theta & \cos j\theta \end{array } \right ) , \ ] ] whereas @xmath408 @xmath409 let @xmath411 be the circulant matrix ( see golub and van loan ( 1996 , p.201 ) ) with the first column @xmath412 direct calculations show that @xmath413 and @xmath414 where @xmath415 is the @xmath416-th column of @xmath417 and @xmath418 is the vector of ones . using these identities , it is straightforward to verify that@xmath419 now , let us define @xmath420 using identities ( [ sssu ] ) for @xmath97 and lemma [ rank ] , we conclude that @xmath102 is asymptotically equivalent to @xmath421 , where @xmath422 is the empirical d.f . of the eigenvalues of @xmath423
further , ( [ identity1 ] ) and ( [ identity2 ] ) yield@xmath424 applying lemma [ rank ] one more time , we obtain the asymptotic equivalence of @xmath425 and @xmath426 where @xmath427 is the empirical d.f .
of the eigenvalues of @xmath428 with@xmath429 as is well known ( see , for example , golub and van loan ( 1996 ) , chapter 4.7.7 ) , @xmath430 circulant matrices can be expressed in terms of the discrete fourier transform matrices@xmath431 with @xmath405 precisely,@xmath432 where @xmath433 and the star superscript denotes transposition and complex conjugation . for the @xmath434-th diagonal elements of @xmath435 and @xmath436
we have@xmath437 note that @xmath438 for @xmath439 if @xmath7 is even , as we assumed above , then there are @xmath440 pairs @xmath441 , and there is one pair @xmath442 that correspond to@xmath443 define a permutation matrix @xmath444 so that the equal diagonal elements of @xmath445 are grouped in adjacent pairs .
precisely , let @xmath446 , where@xmath447{l}1\text { if } t=2s-1\text { for } s=1, ... ,t/2\\
1\text { if } t=2\left ( t - s+2\right ) \operatorname{mod}t\text { for } s = t/2 + 1, ... ,t\\ 0\text { otherwise}\end{array } \right.\ ] ] and let @xmath448 be the unitary matrix@xmath449{cc}i_{2 } & 0\\ 0 & i_{t/2}\otimes z \end{array } \right ) \text { with } z=\frac{1}{\sqrt{2}}\left ( \begin{array } [ c]{cc}1 & 1\\ \mathrm{i } & -\mathrm{i}\end{array } \right ) , \ ] ] where @xmath450 denotes the kronecker product .
further , let @xmath451 . as is easy to check
, @xmath452 is an orthogonal matrix .
furthermore,@xmath453 where @xmath399 and @xmath401 are as defined in ( [ d1 ] ) and ( [ d2 ] ) . combining this with ( [ c2d2 ] ) and using lemma [ rank ] once again
, we obtain the asymptotic equivalence of @xmath454 and @xmath455 where @xmath456 is the empirical d.f . of the eigenvalues of @xmath457 with@xmath458 because of the rotational invariance of the gaussian distribution , the distributions of @xmath459 and @xmath383 are the same .
hence , @xmath460 is asymptotically equivalent to @xmath461 and thus , @xmath462 is asymptotically equivalent to @xmath143.@xmath463 our proof of the almost sure weak convergence of @xmath462 to the wachter distribution consists of showing that the stieltjes transform of @xmath410,@xmath464 almost surely converges pointwise in @xmath465 to the stieltjes transform @xmath466 of the wachter distribution . to establish such a convergence , we show that , if @xmath62 is a limit of @xmath467 along any _ subsequence _ of @xmath154 then it must satisfy a system of equations with unique solution given by @xmath466 . the almost sure convergence of @xmath410 ( and thus , also of @xmath143 )
to the wachter distribution follows then from the continuity theorem for the stieltjes transforms ( see , for example , corollary 1 in geronimo and hill ( 2003 ) ) .
we shall write @xmath468 for the stieltjes transform @xmath467 to simplify notation .
let @xmath469 then by definition ( [ mhat def ] ) , @xmath468 must satisfy the following equations@xmath470 , \label{equation2}\\ \hat{m } & = \frac{1}{p}\operatorname*{tr}\left [ d\tilde{m}^{-1}\right ] .\label{equation4}\ ] ] let us study the above traces in detail .
define@xmath471 , \text { } j=1, ... ,t/2.\ ] ]
we now show that the traces in ( [ equation2 ] ) and ( [ equation4 ] ) can be expressed as functions of the terms having form @xmath472 where @xmath473 is independent from @xmath474 then , we argue that @xmath475 i_{2}\ ] ] a.s .
converge to zero , and use this fact to derive equations that the limit of @xmath476 if it exists , must satisfy .
first , consider ( [ equation2 ] ) . note that@xmath477
= \frac{1}{p}{\displaystyle\sum_{j=1}^{t/2 } } \operatorname*{tr}\left [ \varepsilon_{(j)}^{\prime}m^{-1}\varepsilon _ { ( j)}\right ] .\label{germ of eq2}\ ] ] let us introduce new notation:@xmath478@xmath479@xmath480 in addition , let@xmath481 a straightforward algebra that involves multiple use of the sherman - morrison - woodbury formula ( see golub and van loan ( 1996 ) , p.50 ) @xmath482 and the identity@xmath483 establishes the following equality@xmath484\omega_{j}[v_{j},u_{j}^{\prime}]^{\prime},\label{stem of eq2}\ ] ] where@xmath485{cc}\frac{1}{1-z}i_{2}+v_{j } & \frac{1}{1-z}r_{j}\delta_{2j}^{\prime}+u_{j}^{\prime}\\ \frac{1}{1-z}r_{j}\delta_{2j}+u_{j } & \frac{z}{1-z}r_{j}i_{2}-s_{j}+w_{j}\end{array } \right ) ^{-1}.\ ] ] a derivation of ( [ stem of eq2 ] ) can be found in the supplementary appendix .
let us define @xmath486 , \text { } \hat{u}=\frac{1}{t}\operatorname*{tr}\left [ d^{-1}c^{\prime}m^{-1}\right ] , \\
\hat{v } & = \frac{1}{t}\operatorname*{tr}\left [ m^{-1}\right ] , \text { and } \\
\hat{w } & = \frac{1}{t}\operatorname*{tr}\left [ d^{-1}c^{\prime}m^{-1}cd^{-1}\right ] .\end{aligned}\ ] ] we have the following lemma , where @xmath487 denotes the spectral norm .
its proof is given in the supplementary appendix . [ the second equation]there exists @xmath491 such that , for any @xmath492 with zero real part , @xmath493 , and the imaginary part satisfying @xmath494 we have@xmath495@xmath496 and @xmath497 , as @xmath106 consider a @xmath498 matrix @xmath499 that is obtained from @xmath500 by replacing @xmath501 and @xmath502 in ( [ stem of eq2 ] ) with @xmath503 and @xmath504 respectively .
we have@xmath505\hat{\omega}_{j}[\hat { v}i_{2},\hat{u}i_{2}]^{\prime},\ ] ] where@xmath506{cc}\frac{1}{1-z}i_{2}+\hat{v}i_{2 } & \frac{1}{1-z}r_{j}\delta_{2j}^{\prime}+\hat{u}i_{2}\\ \frac{1}{1-z}r_{j}\delta_{2j}+\hat{u}i_{2 } & \frac{z}{1-z}r_{j}i_{2}+(\hat { w}-\hat{s})i_{2}\end{array } \right ) ^{-1}.\ ] ] a simple algebra and the identity @xmath507 yield@xmath508{cc}\frac{z}{1-z}r_{j}i_{2}+(\hat{w}-\hat{s})i_{2 } & -\frac{1}{1-z}r_{j}\delta_{2j}^{\prime}-\hat{u}i_{2}\\ -\frac{1}{1-z}r_{j}\delta_{2j}-\hat{u}i_{2 } & \frac{1}{1-z}i_{2}+\hat{v}i_{2}\end{array } \right ) , \label{omegatilde}\ ] ] and@xmath509 by definition , @xmath510 in the proof of lemma [ rigour1 ]
, we show that the norms @xmath511 @xmath512 and @xmath513 almost surely remain bounded as @xmath106 hence , @xmath514 @xmath515 @xmath516 and @xmath517 are also almost surely bounded .
further , by definition,@xmath518 where @xmath519 is an orthogonal matrix , so that @xmath520 and @xmath521 are clearly bounded uniformly in @xmath522 therefore , the norm of matrix @xmath523 almost surely remains bounded as @xmath24 , uniformly in @xmath522 regarding @xmath524 which appear in the denominator on the right hand side of ( [ omega ] ) , the supplementary appendix establishes the following result .
the above results imply that , for @xmath492 with @xmath493 and @xmath494 @xmath526 almost surely remains bounded as @xmath24 , uniformly in @xmath522 therefore , by lemma [ rigour1],@xmath527 where @xmath528 as @xmath154 uniformly in @xmath522 a straightforward algebra reveals that@xmath529 using this in equations ( [ approximation1 ] ) and ( [ germ of eq2 ] ) , we obtain@xmath530 where , in the latter expression , the term corresponding to @xmath531 is included in the @xmath532 term to take into account the special definition of @xmath533 as follows from lemma [ small delta lemma ] and the boundedness of @xmath534 and @xmath535 the derivative @xmath536 almost surely remains bounded by absolute value as @xmath154 uniformly in @xmath537 .$ ] therefore@xmath538 the statement of proposition [ the second equation ] now follows by noting that the latter integral is one quarter of the integral over @xmath539 .\square$ ] a similar analysis of equation ( [ equation4 ] ) gives us another proposition , describing @xmath468 as function of @xmath540 and @xmath541 where@xmath542 , \text { } \tilde{u}=\frac{1}{t}\operatorname*{tr}\left [ a^{-1}c\tilde{m}^{-1}\right ] , \\ \tilde{v } & = \frac{1}{t}\operatorname*{tr}\left [ \tilde{m}^{-1}\right ] , \text { and }
\\ \tilde{w } & = \frac{1}{t}\operatorname*{tr}\left [ a^{-1}c\tilde{m}^{-1}c^{\prime}a^{-1}\right ] .\end{aligned}\ ] ] we omit the proof because it is very similar to that of proposition [ the second equation ] .
although we now have two asymptotic equations for @xmath476 ( [ asyeq1 ] ) and ( [ asyeq2 ] ) , they contain many unknowns : @xmath545 and the corresponding variables with tildes .
the following result establishes simple relationships between the unknowns with hats and tildes .
the identity @xmath547 is established by the following sequence of equalities@xmath548 the relationship @xmath549 is obtained as follows@xmath550 the identity @xmath551 is obtained similarly to @xmath549 by interchanging the roles of @xmath552 and @xmath553.@xmath463 the identities ( [ connection identities ] ) imply the following equality@xmath554 we denote the reciprocal of the common value of the right and left hand sides of this equality as @xmath555 a direct calculation shows that@xmath556 and the asymptotic relationships ( [ asyeq1 ] ) and ( [ asyeq2 ] ) can be written in the following form@xmath557{l}\hat{m}=\frac{1}{2\pi c}\int_{0}^{2\pi}\hat{h}\left ( z,\varphi\right ) \left ( \left ( z\tilde{v}-4\sin^{2}\varphi\right ) \hat{v}-\hat{u}^{2}\right ) \mathrm{d}\varphi+o(1)\\ \hat{m}=\frac{1}{2\pi c}\int_{0}^{2\pi}\hat{h}\left ( z,\varphi\right ) \left ( \left ( z\hat{v}-1\right ) \tilde{v}-\hat{u}^{2}\right ) \mathrm{d}\varphi+o(1 ) \end{array } \right . .\label{system of two}\ ] ] this can be viewed as an asymptotic system of two equations with four unknowns : @xmath558 and @xmath559 we shall now complete the system by establishing the other two asymptotic relationships connecting these unknowns .
multiplying both sides of the identity@xmath560 by @xmath561 taking trace , dividing by @xmath0 , and rearranging terms , we obtain@xmath562 .\label{equation1}\ ] ] next , we analyze ( [ equation1 ] ) similarly to the above analysis of ( [ equation2 ] ) .
that is , first , we note that@xmath563 = \frac { 1}{p}{\displaystyle\sum_{j=1}^{t/2 } } \operatorname*{tr}\left [ \delta_{2j}^{\prime}\varepsilon_{(j)}^{\prime}d^{-1}c^{\prime}m^{-1}\varepsilon_{(j)}\right ] .\label{germ of eq1}\ ] ] then elementary algebra , based on the sherman - morrison - woodbury formula ( [ identity ] ) , yields@xmath564 \omega_{j}\left [ v_{j},u_{j}^{\prime } \right ] ^{\prime}\right ) \label{stem of eq1}\\ & + r_{j}\left ( r_{j}i_{2}+s_{j}\right ) ^{-1}\left ( u_{j}-\left [ u_{j},w_{j}\right ] \omega_{j}\left [ v_{j},u_{j}^{\prime}\right ] ^{\prime } \right ) .\nonumber\end{aligned}\ ] ] multiplying both sides of ( [ stem of eq1 ] ) by @xmath565 and replacing @xmath566 and @xmath502 by @xmath567 and @xmath504 respectively , yields an asymptotic approximation to @xmath568 which can be used in ( [ germ of eq1 ] ) and ( [ equation1 ] ) to produce the following result .
its proof , as well as the proof of ( [ stem of eq1 ] ) , are given in the supplementary appendix .
one might think that the remaining asymptotic relationship can be obtained by using the identity @xmath570 which parallels ( [ another identity ] ) .
unfortunately , following this idea delivers a relationship equivalent to ( [ asyeq3 ] ) .
therefore , instead of using ( [ parallel identity ] ) , we consider the identity@xmath571 = \frac{1}{p}\operatorname*{tr}\left [ dd^{-1}c^{\prime}m^{-1}\right ] , \label{last identity}\ ] ] which yields@xmath572 = \frac{1}{p}{\displaystyle\sum_{j=1}^{t/2 } } \operatorname*{tr}\left [ \delta_{1j}\varepsilon_{(j)}^{\prime}d^{-1}c^{\prime}m^{-1}\varepsilon_{(j)}\right ] .\label{germ of eq4}\ ] ] then , we proceed as in the above analysis of ( [ germ of eq1 ] ) and ( [ germ of eq2 ] ) to obtain the remaining asymptotic relationship .
the proof of the following proposition is given in the supplementary appendix .
summing up the results in propositions [ the second equation ] , [ the fourth equation ] , [ the first equation ] , and [ the last equation ] , the unknowns @xmath574 and @xmath575 must satisfy the following system of asymptotic equations@xmath557{l}\hat{m}=\frac{1}{2\pi c}\int_{0}^{2\pi}\hat{h}\left ( z,\varphi\right ) \left ( \left ( z\tilde{v}-4\sin^{2}\varphi\right ) \hat{v}-\hat{u}^{2}\right ) \mathrm{d}\varphi+o(1)\\ \hat{m}=\frac{1}{2\pi c}\int_{0}^{2\pi}\hat{h}\left ( z,\varphi\right ) \left ( \left ( z\hat{v}-1\right ) \tilde{v}-\hat{u}^{2}\right ) \mathrm{d}\varphi+o(1)\\ 1+z\hat{m}=\frac{1}{2\pi c}\int_{0}^{2\pi}\hat{h}\left ( z,\varphi\right ) \left ( 2\hat{u}\sin^{2}\varphi+z\tilde{v}\hat{v}-\hat{u}^{2}\right ) \mathrm{d}\varphi+o(1)\\ 0=\frac{1}{2\pi c}\int_{0}^{2\pi}\hat{h}\left ( z,\varphi\right ) \left ( 4\hat{v}\sin^{2}\varphi+2\hat{u}\right ) \mathrm{d}\varphi+o(1 ) \end{array } \right . .\label{asymptotic system}\ ] ] recall that the unknowns @xmath574 and @xmath575 in the asymptotic relationships ( [ asymptotic system ] ) depend on @xmath576 the definition ( [ mhat def ] ) of @xmath468 implies that @xmath577 is bounded by @xmath578 further , as shown in the proof of proposition [ the second equation ] , @xmath575 and @xmath579 are a.s .
bounded by absolute value , and it can be similarly shown that @xmath580 is a.s .
bounded by absolute value .
therefore , there exist a subsequence of @xmath120 along which @xmath574 and @xmath575 a.s .
converge to some limits @xmath581 and @xmath582 these limits must satisfy a non - asymptotic system of equations@xmath557{l}m=\frac{1}{2\pi c}\int_{0}^{2\pi}h\left ( z,\varphi\right ) \left ( \left ( zy-4\sin^{2}\varphi\right ) v - u^{2}\right ) \mathrm{d}\varphi\\ m=\frac{1}{2\pi c}\int_{0}^{2\pi}h\left ( z,\varphi\right ) \left ( \left ( zv-1\right ) y - u^{2}\right ) \mathrm{d}\varphi\\ 1+zm=\frac{1}{2\pi c}\int_{0}^{2\pi}h\left ( z,\varphi\right ) \left ( 2u\sin^{2}\varphi+zvy - u^{2}\right ) \mathrm{d}\varphi\\ 0=\frac{1}{2\pi c}\int_{0}^{2\pi}h\left ( z,\varphi\right ) \left ( 2v\sin ^{2}\varphi+u\right ) \mathrm{d}\varphi \end{array } \right .
, \label{the system}\ ] ] where@xmath583 ^{-1}.\ ] ] let us consider , until further notice , only such @xmath492 that have zero real part , @xmath493 , and the imaginary part satisfying @xmath494 for some @xmath491 .
let us solve system ( [ the system ] ) for @xmath584 adding two times the last equation to the first one , and subtracting the second equation we obtain@xmath585 note that @xmath586 otherwise , from the second equation of ( [ the system ] ) , we have @xmath587 which can not be true because @xmath468 is the stieltjes transform of the empirical distribution of the squared canonical correlations , all of which lie between zero and one . indeed ,
clearly , for any @xmath588 and @xmath492 with @xmath589@xmath590 therefore , @xmath591 and @xmath468 can not converge to @xmath592 . since @xmath593 ( [ zero ] ) yields @xmath594 with @xmath595 and @xmath596 ( if one of them equals zero , the other equals zero too , and @xmath592 by the second equation of ( [ the system ] ) , which is impossible ) . since @xmath597 the last equation implies that @xmath598 as well .
further , subtracting from the third equation the sum of @xmath492 times the second and @xmath599 times the last equation , and using ( [ reduction1a ] ) , we obtain@xmath600 this equation , together with the second equation of ( [ the system ] ) yield@xmath601 next , for the integrand in the last equation of ( [ the system ] ) , we have@xmath602 this assumes that @xmath603 if not , then @xmath604 ^{-1}\ ] ] would not depend on @xmath605 and the last equation of ( [ the system ] ) would imply that @xmath606 the latter equation and the equality @xmath607 would yield @xmath608 which when combined with the second equation of ( [ the system ] ) would give us @xmath609 which can not be true because @xmath610 being a limit of @xmath468 , must satisfy @xmath611 for @xmath612 finally , elementary calculations given in the supplementary appendix show that@xmath615 where @xmath616 $ ] . using ( [ fact ] ) , ( [ germ of red2 ] ) , and the definition of @xmath617 , we obtain the following relationship@xmath618 that holds as long as@xmath619 .\ ] ] the latter inclusion holds because otherwise @xmath617 is not a bounded function of @xmath620 which would contradict lemma [ small delta lemma ] . equation ( [ possibility1 ] ) can not hold because otherwise , ( [ reduction4 ] ) would imply that @xmath624 which is impossible as argued above . equation ( [ possibility2 ] )
taken together with ( [ reduction3 ] ) implies that@xmath625 which was ruled out above .
this leaves us with ( [ possibility3 ] ) , so that , using ( [ reduction4 ] ) , we get@xmath626 for @xmath627 with @xmath589 the imaginary part of the right hand side of ( [ second possibility ] ) is negative when ` @xmath628 ' is used in front of the square root .
here we choose the branch of the square root , with the cut along the positive real semi - axis , which has positive imaginary part .
since @xmath629 can not be negative , we conclude that@xmath630 but the right hand side of the above equality is the value of the limit of the stieltjes transforms of the eigenvalues of the multivariate beta matrix @xmath631 as @xmath106 this can be verified directly by using the formula for such a limit , given for example in theorem 1.6 of bai , hu , pan and zhou ( 2015 ) . as follows from wachter ( 1980 ) , the weak limit of the empirical distribution of the eigenvalues of the multivariate beta matrix @xmath632 as @xmath24 equals @xmath633 .
equation ( [ the limit ] ) shows that , for @xmath492 with @xmath493 and @xmath494 any converging subsequence of @xmath468 converges to the same limit .
hence , @xmath468 a.s .
converges for all @xmath492 with @xmath493 and @xmath634 note that @xmath468 is a sequence of bounded analytic functions in the domain @xmath635 where @xmath636 is an arbitrary positive number .
therefore , by vitaly s convergence theorem ( see titchmarsh ( 1939 ) , p.168 ) @xmath468 a.s .
converges to @xmath610 described by ( [ the limit ] ) , for any @xmath637 the almost sure convergence of @xmath462 ( and thus , also of @xmath69 ) to the wachter distribution follows from the continuity theorem for the stieltjes transforms ( see , for example , corollary 1 in geronimo and hill ( 2003 ) ) .
first , let us show that the weak limit @xmath195 of @xmath638 as @xmath196 exists and equals the continuous part of the marchenko - pastur distribution with density ( [ mp density ] ) . by definition and theorem [ main ] , @xmath639 is the ( scaled ) wachter d.f .
@xmath640 therefore , by ( [ densityw ] ) and ( [ support boundaries ] ) , the density , @xmath641 , and the boundaries of the support , @xmath642 , $ ] of the distribution @xmath643 equal @xmath644@xmath645 as @xmath646 @xmath647 where @xmath648 as in ( [ mp boundaries ] ) , and @xmath641 converges to the density given by ( [ mp density ] ) .
this implies the weak convergence of @xmath638 to @xmath195 with @xmath211 supported on @xmath201 $ ] and having density ( [ mp density ] ) . to establish the theorem
, it remains to show that , as @xmath51 , @xmath649 weakly converges to @xmath650 in probability .
recall that the weak convergence is metrized by the lvy distance @xmath651 .
we need to show that for any @xmath652 there exists @xmath653 such that ( s.t . ) for all @xmath654@xmath655 let @xmath656 be so small that@xmath657 for any @xmath214 let @xmath658 be the smallest even integer satisfying @xmath659 that is,@xmath660 for any @xmath661 by the triangle inequality , we have@xmath662 where @xmath663 and @xmath664 denote the empirical distributions of eigenvalues of@xmath665 with @xmath666 and @xmath667 respectively . by theorem [ main ] , @xmath668 a.s .
converges to zero as @xmath183 therefore , for all sufficiently large @xmath0 , we have@xmath669 further , as shown by johansen ( 1988 , 1991 ) , for any @xmath214 as @xmath670 the eigenvalues of ( [ rawmatrixagain ] ) with @xmath671 jointly converge in distribution to those of@xmath672 therefore , for any @xmath0 and all sufficiently large @xmath673 we have@xmath674 let us denote the sum of @xmath675 @xmath676 and @xmath677 as @xmath678 by ( [ triangle ] ) , we have@xmath679 inequalities ( [ levy1 ] ) , ( [ levy2 ] ) , and ( [ levy4 ] ) show that for any @xmath652 there exists @xmath680 such that ( s.t . ) for any positive @xmath681 there is a @xmath682 s.t .
for any @xmath683 there is a @xmath684 s.t . for any @xmath685@xmath686 the subscripts in @xmath687 @xmath682 and @xmath684 signify dependence on the value of the corresponding parameter
. inequalities ( [ levy124 ] ) and ( [ two terms ] ) would establish ( [ need to show ] ) as long as we are able to show that for any @xmath652 there exists @xmath688 s.t . for any positive @xmath689
there is a @xmath690 s.t . for any @xmath691 _ and any _
@xmath692 there exists @xmath693
s.t.@xmath694 let us denote @xmath695 where @xmath383 is a @xmath696 matrix with i.i.d .
@xmath697 entries , as defined in section 2 .
we shall assume that , as @xmath120 change , @xmath698 represents @xmath696 sections of a fixed infinite array of i.i.d .
standard normal random variables .
consider@xmath699 so defined matrix @xmath700 is identical to the real symmetric matrix @xmath701 the above definition is formulated in terms of @xmath698 to clarify that @xmath700 depends on @xmath7 not only via the term @xmath702 but also through @xmath703 and @xmath704 note that @xmath705 and @xmath664 are the empirical distributions of eigenvalues of @xmath706 and @xmath707 respectively .
the following lemma is established in the supplementary appendix .
[ alpha12 ] for any @xmath708 there exists @xmath709 s.t . for any positive @xmath710 ,
there is a @xmath690 s.t . for any @xmath711 and any @xmath692 there
exists @xmath693 s.t . with probability larger than @xmath712 @xmath713 can be represented as the sum of two real symmetric matrices @xmath714 and @xmath104,@xmath715 where @xmath716 @xmath717 , and @xmath718 is an absolute constant .
finally , let @xmath719 be the empirical distribution of eigenvalues of @xmath720 then , by theorem a45 ( norm inequality ) of bai and silverstein ( 2010),@xmath721 whereas by their theorem a43 ( rank inequality),@xmath722 therefore , by lemma [ alpha12 ] and the triangle inequality , for any @xmath708 there exists @xmath709 s.t . for any positive @xmath723 ,
there is a @xmath690 s.t . for any @xmath691 and any @xmath692 there exists @xmath724 s.t.@xmath725 for
@xmath726 this inequality implies ( [ levy3 ] ) with @xmath727 combining ( [ levy3 ] ) with ( [ levy124 ] ) yields ( [ need to show ] ) , which completes the proof .
gonzalo , j. and j - y pitarakis ( 1999 ) dimensionality effect in cointegration analysis , cointegration , causality , and forecasting .
a festschrift in honour of clive wj granger , oxford university press , oxford , pp .
212 - 229 ho , m.s . , and b.e . sorensen ( 1996 ) finding cointegration rank in high dimensional systems using the johansen test : an illustration using data based mote carlo simulations , _ review of economics and statistics _ 78 , 726 - 732 .
wachter , k. w. ( 1976b ) . probability plotting of multiple discriminant ratios , " _ proceedings of the social statistics section of the american statistical association , part ii _ 830 - 833 .
prindle , weber and schmidt , boston . | johansen s ( 1988 , 1991 ) likelihood ratio test for cointegration rank of a gaussian var depends only on the squared sample canonical correlations between current changes and past levels of a simple transformation of the data .
we study the asymptotic behavior of the empirical distribution of those squared canonical correlations when the number of observations and the dimensionality of the var diverge to infinity simultaneously and proportionally .
we find that the distribution almost surely weakly converges to the so - called _
wachter distribution_. this finding provides a theoretical explanation for the observed tendency of johansen s test to find spurious cointegration .
it also sheds light on the workings and limitations of the bartlett correction approach to the over - rejection problem .
we propose a simple graphical device , similar to the scree plot , for a preliminary assessment of cointegration in high - dimensional vars . |
the phenomenon of ferrimagnetism is related to the counteraction of opposite magnetic moments with unequal magnitudes located on different sublattices .
ferrimagnetic materials have , under certain conditions , a compensation temperature at which the resultant magnetization vanishes below its critical temperature @xcite .
recently , it has been both experimentally and theoretically shown that the coercive field exhibits a rapid increase at the compensation point @xcite .
it is obvious that such kind of point has a technological importance @xcite , because at this point only a small driving field is required to change the sign of the resultant magnetization . due to the recent developments in experimental techniques ,
scientists begin to synthesize new classes of molecular - based magnets @xcite .
for instance , it has been shown that the saturation magnetization , chemical analysis and infrared spectrum analysis of @xmath0 , where @xmath1 is tetracyanoethylene , are consistent with a ferrimagnet with spin-3/2 at the vanadium site and a spin of 1/2 at the tcne sites with @xmath2 @xcite . in this regard ,
it is possible to mention that the theoretical models referring the mixed systems are of great importance since they are well adopted to study and to provide deeper understanding of certain type of ferrimagnetism @xcite . from the theoretical point of view , a great deal number of studies
have been realized to get a clear idea about the magnetic properties of mixed spin-1/2 and spin-3/2 ferrimagnetic ising systems . in order to have a general overview about it ,
it is beneficial to classify the studies in two categories based on the investigation of equilibrium and nonequilibrium phase transition properties of such type of mixed spin systems . in the former group , static or equilibrium properties of these type of systems
have been analyzed within the several frameworks such as exact @xcite , effective field theory with correlations @xcite , bethe lattice @xcite , exact star - triangle mapping transformation @xcite , high temperature series expansion method @xcite , multisublattice green - function technique @xcite , oguchi approximation @xcite as well as monte carlo simulation @xcite .
it is underlined in the some studies noted above that when the system includes only the nearest - neighbor interaction between spins and the single - ion anisotropy , the temperature variation of resultant magnetization does not exhibit a compensation behavior . in contrary to this , when the next - nearest neighbor interaction between spins-1/2 takes into account and exceeds a minimum value which depends upon the other system parameters , the ferrimagnetic system reveals a compensation treatment which can not be observed in single - spin ising systems .
magnetically interacting system under the influence of a magnetic field varying sinusoidally in time exhibits two important striking phenomena : nonequilibrium phase transitions and dynamic hysteresis behavior .
nowadays , these types of nonequilibrium systems are in the center of scientists because they have exotic , unusual and interesting behaviors . for example , the universality classes of the ising model and its variations under a time dependent driving field are different from its equilibrium counterparts @xcite .
it is possible to emphasize that nonequilibrium phase transitions originate due to a competition between time scales of the relaxation time of the system and oscillating period of the external applied field . for the high temperatures and high amplitudes of the periodically varying magnetic field , the simple kinetic ferromagnetic system exists in dynamically disordered phase where the time dependent magnetization oscillates around value of zero and is able to follow the external applied magnetic field with some delay , whereas it oscillates around a non - zero value which indicates a dynamically ordered phase for low temperatures and small magnetic field amplitudes @xcite .
the physical mechanism described briefly above points out the existence of a dynamic phase transition ( dpt ) @xcite .
dpts and hysteresis behaviors can also be observed experimentally .
for example , by benefiting from surface magneto - optic kerr effect ( moke ) , dynamic scaling of magnetic hysteresis in ultrathin ferromagnetic fe / au(001 ) films has been studied , and it is reported that the dispersion of hysteresis loop area of studied system obeys to a power law behavior@xcite .
a comprehensive study , which includes the hysteresis loop measurement of well - characterized ultrathin fe films grown on flat and stepped w(110 ) surfaces , has been done by using moke , and prominent experimental observations are reported in ref .
in addition to these pioneering works mentioned briefly above , to the best of our knowledge , there exist a number of experimental studies regarding the nonequilibrium properties of different types of magnetic materials such as co films on a cu(001 ) surface @xcite , polycrystalline @xmath3 films @xcite , epitaxial fe / gaas(001 ) thin films @xcite , @xmath4 @xcite , finemet thin films with composition @xmath5 @xcite , @xmath6_{3}}$ ] magnetic multilayers with strong perpendicular anisotropy @xcite as well as assembly of paramagnetic colloids @xcite . based upon the detailed experimental investigations , it has been discovered that experimental nonequilibrium dynamics of considered real magnetic systems strongly resemble the dynamic behavior predicted from theoretical calculations of a kinetic ising model . from this point of view , it is possible to see that there exists an impressive evidence of qualitative consistency between theoretical and experimental investigations . on the other hand , in the latter group
there exists a limited number of nonequilibrium studies concerning the influences of time varying magnetic field on the mixed spin-1/2 and spin-3/2 ising ferrimagnetic model .
for instance , thermal and magnetic properties of a mixed ising ferrimagnetic model consisting of spin-1/2 and spin-3/2 on a square lattice have been analyzed by making use of glauber - type stochastic process @xcite .
it has been reported that the studied system always exhibits a dynamic tricritical point in amplitude of external applied field and temperature plane , but it does not show in the single - ion anisotropy and temperature plane for low values of amplitude of field @xcite . following the same methodology ,
a similar study has been done to shed some light on what happens when an oscillating magnetic field is applied to the mixed spin-1/2 and spin-3/2 ising model on alternate layers of hexagonal lattice .
it has been found that depending on the hamiltonian parameters , the system presents dynamic multicritical as well as compensation behaviors @xcite .
however , the aforementioned studies are mainly based on molecular field theory .
it is a well known fact that , in molecular field theory , spin fluctuations are ignored and the obtained results do not have any microscopic information details of system .
from this point view , in order to obtain the true dynamics of a mixed spin-1/2 and spin-3/2 ising ferrimagnetic system on a square lattice under the presence of a time dependent oscillating magnetic field , we intend to use of monte carlo simulation technique which takes into account the thermal fluctuations , and in this way , non - artificial results can be obtained .
the outline of the paper is as follows : in section [ formulation ] we briefly present our model .
section [ discussion ] is dedicated to the results and discussion , and finally section [ conclusions ] contains our conclusions .
we consider a two - dimensional kinetic ising ferrimagnetic system with mixed spins of @xmath7 and @xmath8 defined on a square lattice , and the system is exposed to a time dependent magnetic field source .
the hamiltonian describing our model is given by @xmath9 where the @xmath10 , and @xmath11 are the ising spins on the sites of the sublattices a and b , respectively .
first and second sums in eq .
( [ eq1 ] ) are over the nearest- and next - nearest neighbor pairs of spins , respectively .
we assume @xmath12 such that the exchange interaction between nearest neighbours is antiferromagnetic .
@xmath13 is the exchange interaction parameter between pairs of next - nearest neighbors of spins located on sublattice a , and @xmath14 is single ion - anisotropy term which affects only @xmath8 spins located on sublattice b. the time varying sinusoidal magnetic field is as following @xmath15 here , @xmath16 and @xmath17 are amplitude and angular frequency of the external field , respectively .
the period of the oscillating magnetic field is given by @xmath18 .
the linear dimension of the lattice is selected as @xmath19 through all simulations , and monte carlo simulation based on metropolis algorithm @xcite is applied to the kinetic mixed ising ferrimagnetic system on a @xmath20 square lattice with periodic boundary conditions in all directions .
configurations were generated by selecting the sites sequentially through the lattice and making single - spin - flip attempts , which were accepted or rejected according to the metropolis algorithm .
data were generated over 50 independent samples realizations by running the simulations for 60000 mc steps per site after discarding the first 20000 steps .
this amount of transient steps is found to be sufficient for thermalization for the whole range of the parameter sets .
error bars are found by using jacknife method @xcite . because the calculated errors are usually smaller than the sizes of the symbols in the obtained figures ,
they have not been given in this study .
the instantaneous values of the sublattice magnetizations @xmath21 and @xmath22 , and also the total magnetization @xmath23 at the time t are defined as @xmath24 by benefiting from the instantaneous magnetizations over a full period of oscillating magnetic field , we obtain the dynamic order parameters as follows @xmath25 where @xmath26 , @xmath27 and @xmath28 denote the dynamic order parameters corresponding to the sublattices @xmath29 and @xmath30 , and the overall lattice , respectively . to determine the dynamic compensation temperature @xmath31 from the computed magnetization data , the intersection point of the absolute values of the dynamic sublattice magnetizations was found using @xmath32 with @xmath33 , where @xmath34 is the dynamic critical temperature .
we also calculate the time average of the cooperative part of energy of the kinetic mixed ising ferrimagnetic system over a full cycle of the magnetic field as follow @xcite @xmath35 thus , the specific heat of the system is defined as @xmath36 where @xmath37 represents the temperature .
we should mention here that dpt points separating the dynamically ordered and disordered phases are determined by benefiting from the peaks of heat capacities .
we also verified that the peak positions of heat capacities do not significantly alter when larger @xmath38 is selected .
in this section , we will focus our attention on the nonequilibrium dynamics of the mixed spin-1/2 and spin-3/2 ising ferrimagnetic system under a time dependent magnetic field .
first of all , we will discuss the dynamic nature of the system when the system includes only the nearest - neighbor interaction between spins , the single - ion anisotropy and external applied field .
next , we will give and argue the global dynamic phase diagrams including the both dynamic critical and compensation temperatures in the case of the existence the next - nearest neighbor interaction between spins-1/2 located on sublattice a. before we discuss the dpt features of the considered system , we should notice that the situation of @xmath39 indicates the equilibrium case , and our monte carlo simulation findings for this value of @xmath40 are completely in accordance with the recently published work @xcite where the equilibrium properties of the present system were analyzed by following a numerical methodology of heat - bath monte carlo algorithm .
the considered system exhibits three types of magnetic behaviors depending on the hamiltonian parameters .
these are dynamically ferrimagnetic ( _ i _ ) , ferromagnetic ( _ f _ ) and paramagnetic ( _ p _ ) phases , respectively .
in the first type of phase , namely in _
i _ phase , @xmath41 , and , the time dependent sublattice magnetizations , @xmath42 and @xmath43 oscillate with time around a non - zero value whereas they alternate around a non - zero value and @xmath44 in the second type of phase , namely in _
f _ phase . in _
phase which corresponds to the third type of phase , @xmath44 and @xmath42 and @xmath43 oscillate around zero value , and they are delayed with respect to the external applied magnetic field . keeping in this mind , we illustrate the dynamic phase diagrams in a @xmath45 plane with three oscillation periods @xmath46 and @xmath47 and for some selected values of the applied field amplitudes @xmath48 in figs . [ fig1](a)-(c ) .
one of the main findings is that dpt temperature decreases as the value of applied field amplitude increases .
the physical mechanism underlying this observation can be much better understood by following a simple way : if one keeps the system in one well of a landau type double well potential , a certain amount of energy coming from magnetic field is necessary to achieve a dynamic symmetry breaking .
if the amplitude of the applied field is less than the required amount then the system oscillates in one well .
in this situation , the magnetization does not change its sign .
in other words , the system oscillates around a non - zero value corresponding to a dynamically ordered phase . as the temperature increases , the height of the barrier between the two wells decreases . as a result of this ,
the less amount of magnetic field is necessary to push the system from one well to another and hence the magnetization can change its sign for this amount of magnetic field . consequently
, the time averaged magnetization over a full cycle of the oscillating magnetic field becomes zero .
for the relatively high oscillation period values , dynamic magnetizations corresponding to the instantaneous sublattice order parameters can respond to the oscillating magnetic field with some delay , whereas a competition occurs between the period @xmath49 of the field and the relaxation time of the system as the period of the external magnetic field decreases .
hence , the dynamic magnetizations can not respond to the external magnetic field due to the increasing phase lag between the field and the time dependent magnetization .
this mechanism makes the occurrence of the dpt difficult for the considered system .
another important observation is that an unexpected sharp dip occurs between dynamically ordered and disordered phases in the @xmath45 plane with increasing value of the amplitude of the external applied field , and our results show that observation such kind of treatment explicitly depends upon the value of @xmath49 of field . furthermore , it is necessary to state that for both large negative and positive values of single - ion anisotropies , the phase transition points saturate a certain temperature regions , and they tend to shift to the lower temperature regions with increasing amplitude and period of the external applied field . in figs .
2(a)-(b ) , we depict the effect of the single - ion anisotropy on the thermal variations of dynamic order parameters corresponding to the phase diagram illustrated in fig . [ fig1](a ) for value of @xmath50 .
it is clear from the figures that the treatments of the thermal variations of sublattices as well as total magnetizations curves sensitively depend upon the value of single - ion anisotropy , for selected values of hamiltonian parameters . in the bulk ferrimagnetism of nel , it is possible to classify the thermal variation of the total magnetization curve in certain categories @xcite . according to this nomenclature , for @xmath51 , the considered system clearly points out a q - type behavior , where the magnetizations of system begin to decrease gradually starting from their saturation values with increasing thermal agitation , and then they vanish at the dpt point .
one can easily see that , in the range @xmath52 , the magnetizations tend to fall prominently from their saturation values , and the system undergoes a second order dpt as temperature increases .
in addition to these , when @xmath53 , the system exhibits a l - type behavior at which the total magnetization shows a temperature induced maximum which definitively depends on the value of single - ion anisotropy as well as other hamiltonian parameters . based on the above simulation observations , it is possible to make an inference that the studied system has three types of dynamic magnetic behavior .
it is also worthy of note that even though the magnitudes of spins are different from each other , both @xmath26 and @xmath27 exhibit a dpt at the same critical temperature , which is a result of the nearest - neighbor exchange coupling @xmath54 .
[ fig2 ] the influences of the applied field amplitude @xmath40 on the thermal variations of total and sublattice magnetizations as well as dynamic heat capacity of the system are plotted in figs . [ fig3](a)-(c ) corresponding to the phase diagram constructed in fig .
[ fig1](a ) with value of single - ion anisotropy @xmath55 in fig .
[ fig3](a ) , total magnetization curves of the system are shown .
as seen in this figure , magnetization curves exhibit q - type behavior and dynamic critical temperatures decreases with increasing @xmath40 values .
on the other hand , dynamic heat capacity curves which are depicted in fig .
[ fig3](c ) show a sharp peak behavior indicating the phase transition temperature .
moreover , one can readily see that increasing value of the applied field amplitude gives rise to decrease the maximum of the dynamic heat capacity curves .
we should note here that such kinds of dynamic heat capacity behaviors have been found in a ferrimagnetic core - shell nanoparticle composed of a spin-@xmath56 ferromagnetic core which is surrounded by a spin-1 ferromagnetic shell layer under the presence of a time dependent magnetic field @xcite . in fig .
[ fig4 ] , we investigate the effect of the applied field period on the dpt features of the system .
phase diagrams in fig .
[ fig4](a ) are constructed for a value of the applied field amplitude @xmath57 it is possible to say that dpt points are depressed with increasing applied field period especially in the high values of the single - ion anisotropy .
the physics behind of these findings are identical to those emphasized in fig .
therefore , we will not discuss these interpretations here . instead of this , in fig .
[ fig4](b ) we will give the influence of applied field period on the thermal variations of sublattice magnetizations for a considered value of single - ion anisotropy @xmath58 corresponding to the phase diagram illustrated in fig . [ fig4](a ) .
it is found that as the thermal agitation increases starting from zero , the values of sublattice magnetizations begin to gradually decrease and the system undergoes a dpt at the critical temperature which sensitively depends on value of the @xmath49 . according to our simulation results , the kinetic mixed spin-1/2 and spin-3/2 ising ferrimagnetic system including only nearest - neighbor interaction between spins ,
the single - ion anisotropy and external applied field does not display a dynamic multicritical behavior for a wide range of hamiltonian parameters used in here , in contrary to the previously published molecular field investigation where dynamic first order phase transitions and dynamic tricritical points have been reported for the same model @xcite . in the following analysis , in order to shed some light on the effect of the next - nearest neighbor interaction between spins-1/2 of the studied system , we give the dynamic phase boundaries in @xmath59 plane for some considered values of applied field amplitudes with @xmath60 in figs .
[ fig5](a - c ) .
the phase boundaries containing both dynamic critical and compensation temperatures are plotted for three values of single - ion anisotropies @xmath61 and @xmath62 , respectively . at first sight
, one can clearly see that the dynamic compensation temperatures do not emerge until @xmath63 reaches a certain amount of value . after the aforementioned value of @xmath63 , with an increment in @xmath63 does not lead to change the location of dynamic compensation point for fixed values of @xmath64 and @xmath40 . in contrary to this behavior , in accordance with the expectations , as the value of @xmath63 gets bigger starting from zero , the much more thermal energy is necessary to reveal a dpt .
on the other side , both the dynamic critical and compensation temperatures strongly depend upon the selected values of @xmath64 and @xmath40 .
an increase in the value of @xmath40 gives rise to shift the dynamic critical and compensation points to lower temperatures and also allows the system to display a dynamic compensation behavior at the relatively low value of @xmath63 .
additionally , it is possible to make an inference that with increasing value of single - ion anisotropy , the region where the dynamic compensation behavior occurs shifts to upward for considered hamiltonian parameters .
this situation can be well understood by comparing the figs .
[ fig5](a ) , ( b ) and ( c ) with each other .
effects of the next - nearest neighbor interactions on the thermal variations of total and sublattice magnetizations as well as on dynamic heat capacity of the studied system for some selected values of @xmath65 with @xmath60 corresponding to the phase diagram shown in fig .
[ fig5](a ) are seen in figs .
[ fig6](a - c ) .
we give the total magnetization curves for changing value of @xmath63 in fig . [
fig6](a ) .
these curves explicitly refer the existence of a dynamic compensation behaviors , and they also exhibit a n - type magnetic behavior . as we discussed before , varying value of @xmath63 does not give rise to cause a change in value of dynamic compensation point in temperature plane . besides , in order to show how the dynamic compensation phenomenon rises , thermal variations of the absolute values of sublattice magnetizations are given in fig . [
fig6](b ) .
it can be said that an increase in value of @xmath63 leads to the existence of a dynamically stronger ferromagnetic interaction between @xmath66 spins . in this way
, the @xmath66 spins can remain ordered at relatively higher temperatures . when the thermal energy increases starting from zero , the values of sublattice magnetizations begin to gradually decrease from their saturation values until both sublattices magnetizations are equail in magnitude at a certain temperatures at which dynamic compensation point emerges below the dynamic critical transition temperature .
@xmath63 dependencies of dynamic heat capacities are presented in fig . [
fig6](c ) .
it is obvious from the figure that the dynamic heat capacity curves exhibit a sharp peak at the transition temperature , and when the value of @xmath63 increases , the dynamically ordered phase region gets wider , in other words , the location of the sharp peak slides to a higher value in the temperature plane .
the shape of the heat capacity curves are also nearly same for selected values of hamiltonian parameters . as a final investigation
, we give and discuss the influences of varying value of applied field amplitude on the thermal variations of total and sublattices magnetizations exhibiting dynamic compensation as well as critical temperatures in figs .
[ fig7](a - b ) corresponding to the dynamic phase boundary seen in fig . [ fig5](a ) for @xmath67 it is possible to mention that both dynamic compensation and critical temperatures strongly depend on value of the external applied field amplitude and they tend to shift to a lower region in temperature plane , and compensation behavior disappears with increasing value of @xmath40 . in order to demonstrate the detailed magnetic behavior of system , we give the dynamic phase boundaries in ( @xmath68 planes for three values of the applied field periods such as @xmath46 and @xmath47 for selected values of single - ion anisotropy @xmath69 and next - nearest neighbor interaction @xmath70 in fig . [ fig7](c ) .
based on the calculated phase diagrams , it can be said that the decreasing ( increasing ) applied field period has no effect on the dynamic behavior of compensation behavior whereas it affects prominently the dynamic critical temperature of the system such that the dynamically ordered phase region gets wider ( narrower ) .
in conclusion , it has been carried out a detailed monte carlo investigation based on standard single - spin flip metropolis algorithm to determine the true dpt properties of a mixed spin-1/2 and spin-3/2 ising ferrimagnetic system under a time varying magnetic field .
a complete picture of global dynamic phase diagrams separating the dynamically disordered and ordered phases has been constructed by benefiting from the peaks of thermal variations of dynamic heat capacities in order to have a better understanding of the physical background underlying of the considered system .
the most important observations reported in the present study can be briefly summarized as follows : * when the considered system only includes the nearest - neighbor interaction , single - ion anisotropy and a time dependent sinusoidally oscillating magnetic field , it does not point out a dynamic compensation point .
* stationary state solutions of the system strongly depend on the selected system parameters . as discussed in detail in previous section , with increasing values of amplitude and period of the external applied field ,
the dynamic phase boundaries tend to shift to the lower temperature regions in related planes , and a sharp dip occurs between dynamically ordered and disordered phases . * in contrary to the previously published investigation for the same model where dynamic first - order phase transitions and tricritical points have been reported@xcite , it has not been found any evidence of the dynamic first - order phase transitions in our present work .
the reason is most likely the fact that the method we used completely takes into account the thermal fluctuations , which allows us to obtain non - artificial results .
* when the next - nearest neighbor interaction between spins-1/2 is included and exceeded a characteristic value which sensitively depends on value of the single - ion anisotropy and amplitude of the external applied field , the system exhibits a dynamic compensation behavior below its critical temperature . according to the our simulation results , the changing value of applied field period has no effect on the location of dynamic compensation point .
finally , we should note that it is possible to improve the obtained results by making use of a more realistic system such as heisenberg type of hamiltonian . from the theoretical point of view
, such an interesting problem may be subject of a future work in order to provide deeper understanding of ferrimagnetic materials under a time dependent alternating magnetic field source .
the numerical calculations reported in this paper were performed at tbitak ulakbim ( turkish agency ) , high performance and grid computing center ( truba resources ) . 10 natexlab#1#1bibnamefont # 1#1bibfnamefont # 1#1citenamefont # 1#1url # 1`#1`urlprefix[2]#2 [ 2][]#2 l. nel , ann .
( paris ) * 3 * , 137 ( 1948 ) .
p. hansen , j. appl . phys . * 62 * , 216 ( 1987 ) . g.m .
buendia , and e. machado , phys .
b * 61 * , 14686 ( 2000 ) .
m. mansuripur , j. appl . phys . * 61 * , 1580 ( 1987 ) .
shieh , and m.h .
kryder , appl .
. lett . * 49 * , 473 ( 1986 ) .
manriquez , g.t .
yee , r.s .
mclean , a.j .
epstein , and j.s .
miller , science * 252 * , 1415 ( 1991 ) .
morin , p. zhou , c. hahm , and a.j .
epstein , j. appl .
* 73 * , 5648 ( 1993 ) .
g. du , j. joo , and a.j .
epstein , j. appl .
* 73 * , 6566 ( 1993 ) .
c. domb , adv .
* 9 * , 149 ( 1960 ) .
l. l. gonalves , phys .
* 32 * , 248 ( 1985 ) .
t. kaneyoshi , m. jaur and p. tomczak , j. phys . : condens .
matter * 4 * , l653 ( 1992 ) ; * 5 * , 5331 ( 1993 ) .
t. kaneyoshi , physica a * 215 * , 378 ( 1995 ) .
n. benayad , a. dakhama , a. klmper and j. zittartz , ann .
physik * 5 * , 387 ( 1996 ) .
a. bobk , and m. juriin , j. magn .
mang . mater . * 163 * , 292 ( 1996 ) .
a. bobk , and d. horvth , phys .
b * 213 * , 459 ( 1999 ) .
w. jiang , g.-0 .
wei , and z .- h .
xin , physica a * 293 * , 455 ( 2001 ) . y .-
liang , g .- z .
wei , and g .- l .
song , phys .
b * 241 * , 1916 ( 2004 ) .
m. aouzi , m. el hafidi , and e.m .
sakhaf , physica a * 345 * , 575 ( 2005 )
. i. essaoudi , k. brner , a. ainane , and m. saber , physica a * 385 * , 208 ( 2007 ) .
liang , g .- z .
j . xu and g .- l .
song , comm .
. phys . * 53 * , 957 ( 2010 ) .
a. yigit , and e. albayrak , j. magn .
magn . mater . *
329 * , 125 ( 2013 ) .
a. yigit , and e. albayrak , j. magn .
magn . mater . *
392 * , 4216 ( 2013 ) . c. ekiz , j. magn .
magn . mater . *
293 * , 913 ( 2005 ) . c. ekiz ,
physica a * 353 * , 286 ( 2005 ) .
e. albayrak , and a. alci , physica a * 345 * , 48 ( 2005 ) .
e. albayrak , chin .
b * 21 * , 067501 ( 2012 ) .
m. jaur , and j. streka , physica a * 358 * , 393 ( 2005 ) .
bowers , and b.y .
yousif , phys .
a * 96 * , 49 ( 1983 ) .
hunter , r.c.l .
jenkins , and c.j .
tinsley , j. phys .
* 23 * , 4547 ( 1990 ) .
j. li , g. wei , and a. du , j. magn .
magn . mater .
* 269 * , 410 ( 2004 ) .
a. bobk , z. feckov , and m. ukovic , j. magn .
* 323 * , 813 ( 2011 ) .
buendia , and r. cardona , phys .
b 59 , 6784 ( 1999 ) . b.k .
chakrabarti , m. acharyya , rev .
phys . * 71 * , 847 ( 1999 ) .
m. acharyya , int .
c * 16 * , 1631 ( 2005 ) .
h. park , m. pleimling , phys .
* 109 * , 175703 ( 2012 ) .
t. tom , m.j .
de oliveira , phys .
a * 41 * , 4251 ( 1990 ) .
w.s . lo , r.a .
pelcovits , phys .
a * 42 * , 7471 ( 1990 ) .
he and g .- c .
wang , phys .
* 70 * , 2336 ( 1993 ) .
suen and j.l .
erskine , phys .
lett . * 78 * , 3567 ( 1997 ) .
q. jiang , h .-
yang , and g .- c .
wang , phys .
b * 52 * , 14911 ( 1995 )
choi , w.y .
lee , a. samad , and j.a.c .
bland , phys .
b * 60 * , 11906 ( 1999 ) .
w.y . lee , b .- ch .
choi , y.b .
xu , and j.a.c
. bland , phys .
b * 60 * , 10216 ( 1999 ) .
a.rosales-rivera , j.m .
ferreira , and f.c .
montenegro , j. magn .
magn . mater . * 226 - 230 * 1309 ( 2001 ) .
l. santi , r.l .
sommer , a. magni , g. durin , f. colaiori , and s. zapperi , ieee trans . mag . * 39 * , 2666 ( 2003 ) .
robb , y.h .
xu , o. hellwig , j. mccord , a. berger , m.a .
novotny , and p.a .
rikvold , phys .
b * 78 * , 134422 ( 2008 ) .
p. tierno , t.h .
johansen , and j.m .
sancho , phys .
e * 87 * , 062301 ( 2013 ) .
glauber , j. math .
phys . * 4 * , 294 ( 1963 ) .
b. deviren , m. keskin , and o. canko , j. magn .
magn . mater . *
321 * , 458 ( 2009 ) .
m. keskin , b. deviren and o. canko , ieee trans . mag . * 45 * , 2640 ( 2009 ) . k. binder , _ monte carlo methods in statistical physics _
( springer , berlin , 1979 ) .
newman and g.t.barkema , _ monte carlo methods in statistical physics _
( clarendon press , oxford 2001 ) .
m. acharyya , phys .
e * 56 * , 2407 ( 1997 ) .
y. yksel , e. vatansever , and h. polat , j. phys . : condens .
matter * 24 * , 436004 ( 2012 ) .
e. vatansever , and h. polat , j. magn .
magn . mater .
* 343 * , 221 ( 2013 ) . | nonequilibrium phase transition properties of a mixed ising ferrimagnetic model consisting of spin-1/2 and spin-3/2 on a square lattice under the existence of a time dependent oscillating magnetic field have been investigated by making use of monte carlo simulations with single - spin flip metropolis algorithm . a complete picture of dynamic phase boundary and magnetization profiles have been illustrated and the conditions of a dynamic compensation behavior have been discussed in detail . according to our simulation results ,
the considered system does not point out a dynamic compensation behavior , when it only includes the nearest - neighbor interaction , single - ion anisotropy and an oscillating magnetic field source . as the next - nearest - neighbor interaction between the spins-1/2 takes into account and
exceeds a characteristic value which sensitively depends upon values of single - ion anisotropy and only of amplitude of external magnetic field , a dynamic compensation behavior occurs in the system .
finally , it is reported that it has not been found any evidence of dynamically first - order phase transition between dynamically ordered and disordered phases , which conflicts with the recently published molecular field investigation , for a wide range of selected system parameters . |
since the introduction of correlation functions into modern cosmology by e.g. , totsuji & kihara ( 1969 ) and peebles ( 1973 ) , such functions have proved to be useful in a large variety of situations .
a few typical applications are the study of the distribution of galaxies and clusters of galaxies in the universe ( e.g. , connolly et al .
2002 ; maddox , efstathiou & sutherland 1996 ; dalton et al .
1994 ; croft , dalton , & efstathiou 1999 ; lilje & efstathiou 1988 ; frieman & gaztaaga 1999 ; jing & brner 1998 ) , the analysis of the gravitational lensing shear ( e.g. , bernardeau , van waerbeke , & mellier 2003 ; takada & jain 2003 ) , or measurements of non - gaussianity in the cosmic microwave background ( e.g. , eriksen , banday , & grski 2002 ; kogut et al .
1996 ) .
unfortunately , higher order @xmath0-point correlation functions are also notorious for being computationally expensive . in general the evaluation of an @xmath0-point
correlation function scales as @xmath1 , and therefore the required cpu time soon becomes too large to handle .
to remedy this situation several `` fast '' algorithms have been proposed , in particular for analyzing discrete particle data sets ( such as galaxy catalogs ) .
one commonly used method is to aggregate particles on small scales , and treat each of these aggregations as a single particle ( e.g. , davis et al .
1985 ; kaiser 1986 ) .
this method effectively corresponds to smoothing the data set , and is only valid when the distance between the aggregated particles is much smaller than the scale of interest .
it is therefore of limited use if the primary interest lies in small scales , which often is the case for cmb data .
the main motivation behind this work is analysis of new high - resolution cmb maps . in this case , as in several other applications , the data are not consisting of point sets ( e.g. , positions of galaxies ) but in values of a function ( like temperature , polarization , or shear ) given for all pixel positions on a map covering part of or the full sky , where some fixed pixelization scheme has been used to divide the area of concern into pixels .
further , when analyzing such maps , the main interest is usually in comparing a statistic ( e.g. , a correlation function ) estimated on a map of the real sky with a large monte carlo set of the same statistic estimated on several thousand random realizations based on some theory , e.g. , gaussian fluctuations with a given power spectrum .
hence , one must estimate the correlation function for a set of many thousand similar maps with the same pixelization to do the analysis of interest . the @xmath2d - tree approach ( e.g. , moore et al .
2001 ) is based on a similar idea to the idea of aggregating particles on small scales .
it organizes nearby particles ( or pixels ) into a hierarchy of bounding boxes , and uses this hierarchy for rapidly discarding uninteresting pixel sets .
however , it is not obvious that the @xmath2d - tree approach is well suited for estimating correlation functions from a pixelized map ( which completely fills the @xmath2-dimensional space ) , unless bin widths much greater than the pixel size is desired .
( normally one sets the bin width roughly equal to the pixel size in order to obtain optimal resolution . )
a third approach relies on the fourier - transform , and this approach drastically speeds up the algorithms whenever the fast fourier transform ( fft ) is applicable ( for a two - point application , see , e.g. , szapudi , prunet , & colombi [ 2001a ] ) . however , this method is not very attractive for higher - order correlation functions . in the three - point case one
replaces an @xmath3 algorithm with a very _ small _ prefactor ( multiply three numbers together ) with an @xmath4 algorithm with an extremely _ large _ prefactor ( compute several wigner 3-@xmath5 symbols ; see , e.g. , gangui et al . [ 1994 ] ) .
four - point functions are obviously even more expensive .
thus , the advantage of fourier - methods for estimating high - order correlation functions is so far unclear . in this paper
we describe a new set of algorithms which allows computation of any ( small ) subset of the general @xmath0-point correlation functions from a pixelized map of the whole celestial sphere , or a part of the celestial sphere , with up to several hundred thousand pixels .
these algorithms are extremely well suited for monte carlo studies , since we organize the processes so that the very substantial amount of initialization cpu time is spent only once for the whole ensemble of sky maps , and not for each individual map . for ensembles of more than about one thousand realizations
, most of the cpu time is therefore spent on adding pixel values together , not on geometric computations .
the general idea behind these algorithms is to replace cpu - intensive inverse cosine operations by much less cpu - intensive ` if`-tests , at the cost of increasing the memory requirements .
first we compute the distances between any two pixels in the map and store this information in a set of tables optimized for fast searches ( and compression , if desirable ) .
these tables can be compared with a set of compasses , in the sense that each column of a table draws out a circle on the sphere of a given radius , centered on a given pixel .
then , by searching through two different columns for equal entries , we generate triangles with the desired size and shape , and , if necessary , add two or more such triangles to produce @xmath0-point multiplets .
in fact , our method is equivalent to ruler and compass construction of triangles . a similar algorithm for estimating @xmath0-point correlation functions for point sets ( e.g. , galaxies ) in three - dimensional space has recently been developed by barriga & gaztaaga ( 2002 )
however , there are several important differences between the two methods .
first , our methods are designed particularly for monte carlo simulations , in that we organize the information so that initialization costs are paid only once .
second , we use the `` compasses '' to determine all three edges in the triangle , while barriga & gaztaaga ( 2002 ) only use them for constraining two of the edges .
finally , our methods are particularly designed to take advantage of the healpix nested pixelization scheme ( grski , hivon , & wandelt 1999 ) , although they can ( at a considerable efficiency cost ) be generalized to any pixelization .
a first application of a preliminary version of these algorithms to estimate the three- and four - point correlation functions of the @xmath6-dmr maps was presented by eriksen et al .
( 2002 ) , and a thorough application of the algorithms to the wmap ( bennett et al .
2003 ) data will shortly be published ( eriksen et al .
2004 , in preparation ) .
the examples given at the end of this paper are in fact chosen to correspond with that analysis .
an @xmath0-point correlation function is defined as the average product of @xmath0 temperatures can of course be replaced by any scalar quantity . ] , as measured in a fixed relative configuration on the sky , @xmath7 here @xmath8 span an @xmath0-point polygon with geometric parameters , @xmath9 , on the sky , and @xmath10 is the field value in the direction given by the unit vector @xmath11 . in all common applications we assume isotropy , and
then the @xmath0-point correlation functions are only dependent on the shape and size of the @xmath0-point polygon , and not its particular position or orientation on the sky .
the smallest number of parameters which uniquely determines such a polygon is @xmath12 . on a pixelized map ,
the @xmath0-point correlation functions are estimated as simple averages over all pixel multiplets ( pairs , triplets , quadruples etc . ) satisfying the geometric constraints of the particular configuration , @xmath13 the pixel weights , @xmath14 , may be adjusted to account for , e.g. , noise or boundary effects .
note also that we always , but often implicitly , bin our correlation functions with a given bin width , @xmath15 . a pixel pair of angular separation @xmath16
is then defined to belong the the @xmath17th bin if @xmath18 . as seen from equation ( [ eq : def_comp_npt ] )
, the evaluation of an @xmath0-point correlation function is in principle very simple .
however , before we can multiply and add the individual pixel values , we need to determine all the correct pixel multiplets . that is less than trivial , and constitutes the main theme of this paper .
our starting point is the brute force method for estimating @xmath0-point correlation functions .
if we want to estimate the three - point correlation function for only a subset of all possible configurations , e.g. , for all equilateral triangles , the brute force method is to go through the set of all pixel triplets in the map ( resulting in an @xmath19 $ ] algorithm ) , and compute the three distances between each pixel pair .
each of the three distances is computed by taking the inverse cosine of the the dot product of the two pixel unit vectors . knowing that one inverse cosine operation takes about 40 times as many cpu cycles as a multiplication , we see that not only does the brute force algorithm scale as @xmath3 , but its prefactor is also very high . however , the most time consuming shortcoming of the brute - force approach is that we have to go through every single pixel triplet existing in the map , even though we may only be interested in a very small subset of all configurations .
thus , most of the cpu time is wasted on computing uninteresting information .
our first task is to produce `` compasses '' for picking out the right pixels .
having decided on a number of external parameters , such as pixel resolution ( parameterized by @xmath20 of healpix ) , separation bin width , @xmath21 , and total number of pixels in the data set , @xmath22 ( usually by introducing a mask ) , we make `` compasses '' by going through the set of all pixel pairs , compute the distance between the two pixels in each pair , and determine which separation bin the pair belongs to .
we then insert the pair into a special table ( in the following called a two - point table ) for that bin , a table which is optimized for fast searches ; the @xmath23th column of the @xmath2th table contains all pixels with a distance in the interval @xmath24 from pixel number @xmath23 , sorted according to increasing pixel number . in other words ,
the @xmath23th column of the @xmath2th table is a `` compass '' that draws out a circle on the healpix sphere with radius @xmath25 and width @xmath15 , centered on pixel number @xmath23 .
these two - point tables are computed once and for all , and stored in individual files for later use . already at this point
we can estimate the two - point correlation function with unprecedented efficiency , essentially performing just one addition for each pixel pair . the particular algorithm for this
is described in [
sec : calculations ] . however , our main goal is to estimate higher - order @xmath0-point correlation functions , and then we have to find all triangles , quadrilaterals etc . on the sky of some
given geometric configuration .
here it is worth noting that once we are able to construct triangles , all higher - order structures may be generated simply by putting triangles together .
therefore we here focus on how to construct triangles .
as noted previously , construction of triangles is performed by the use of `` compasses '' : suppose that we want to find all triangles with edge lengths of @xmath26 , @xmath27 and @xmath28 .
we then pick out one pair of pixels , @xmath23 and @xmath5 , from the two - point @xmath26-table , and look up the @xmath23th column in the @xmath27-table and the @xmath5th column in the @xmath28-table .
any common entries in these two tables will have the correct distances from the two base pixels , and together the three points span a triangle of the desired size and shape .
this search is fully equivalent to drawing out circles on the sky , and looking for intersections .
the correspondence between the geometric and algorithmic views is illustrated in figure [ fig : gen_triangles ] .
once this process has been performed for all base pairs in the @xmath26-table , we have also found all triangles of the correct shape and size . at this point
we may therefore forget all about expensive geometric computations , and concentrate on multiplying and adding the pixel triplets together as fast as possible .
let us now go through each step of the underlying algorithms in detail , paying particular attention to real - world problems , such as how to deal with limited physical ram and cpu - time .
the first step in our @xmath0-point correlation algorithm is to construct a set of two - point distance tables .
each of these tables corresponds to a set of `` compasses '' with fixed radius .
for our idea to be of practical use , the two - point tables must meet two requirements : first of all , each table must be small enough to comfortably fit into the physical memory of the computer .
second , the tables must facilitate fast searches for `` neighboring '' pixels ( defined as pixels within a ring of a given radius and width ) for any given center pixel .
since we always bin our correlation functions with a given bin width @xmath15 , the first requirement is met by splitting the full @xmath29 distance matrix up according to these bins , simply by letting the bin width be small enough .
thus , all pixel pairs belonging to the same bin are collected into one table , which later is stored as a separate file .
the next question is how to organize these tables .
the important point is to be able to efficiently draw out a circle around any given center pixel .
that is , given any position on the sky , we must be able to easily look up all pixels with a given distance from that position .
for this purpose we choose a simple two - dimensional array , in which each column corresponds to one single ring on the sky , centered on the pixel @xmath30 , listed in the zeroth row .
thus , all other pixels , @xmath31 , in that column satisfies the relation , @xmath32 , where @xmath33 are unit vectors and @xmath34 are the angular limits of the bin .
we will later be searching for identical entries in two different columns in order to generate triangles , and therefore we already at this point sort the neighbors of each column according to pixel number .
this facilitates @xmath35 searches , as opposed to @xmath36 for unsorted columns . in most experiments
we introduce a mask , either in order to exclude noise - contaminated pixels , or simply to limit our region of interest .
whatever reason , we incorporate such a mask already at the construction level , by only including accepted pixels in the tables .
thus , the zeroth row is nothing but a list of all accepted pixels , and all pixels in the table are guaranteed to be accepted by the mask . not only does this remove many redundant ` if`-tests later on , but more importantly , it makes sure that the algorithms and disk usage scale as @xmath37 , rather than @xmath38 , where @xmath22 is the number of _ accepted _ pixels . for small regions in a high - resolution map ,
the difference is crucial .
the remaining details are mostly of cosmetic nature .
first , since the two - point tables are simple two - dimensional arrays , there will be a small amount of unused space in many columns .
this extra space is padded with -1 s , distinguishing those entries from valid pixel numbers .
secondly , we also store all external parameters ( @xmath20 , @xmath39 , ` ordering ` , bin number etc . ) together with the two - point table in an external file , making each file an independent entity , in the spirit of object - oriented programming . although the two - point tables are quite simple to describe , they are surprisingly complicated to generate , the fundamental problem being their size .
suppose that we have a map with 100000 pixels , and we want to generate a complete set of two - point tables .
doing it straight - forwardly ( i.e. , by keeping all information stored in memory at once ) , we would need at least 100000@xmath40 of memory , and few current computers have that much physical ram available . on the other hand , this number is not prohibitive in terms of disk storage .
the question is therefore how to generate the two - point tables .
once this operation is completed , we can store them on a hard drive , compress them if desirable , and then read only a few at a time when actually estimating the correlation functions .
we utilize the healpix pixelization with _ nested _ pixel ordering .
this pixelization has the highly desirable property that any low - resolution pixel may be divided into four high - resolution pixels , where the pixel numbers of the high - resolution pixels are given by @xmath41 , @xmath23 being the pixel number of the low - resolution pixel and @xmath42 . in short
the generation process can be described as a three - step operation : first we generate a set of tables with low resolution , choosing a resolution low enough to keep all information simultaneously in the computer s physical memory .
second , each table is expanded to a higher resolution by replacing the low - resolution pixel numbers by the high - resolution pixel numbers .
since each mother pixel has four children , the output array is four times as wide and deep as the input array .
all pairs in the output array have distances in approximately the correct range , and therefore the output two - point table is almost a valid high - resolution table .
however , since the boundary between two rings is smoothed when the resolution is increased , one may find that a number of pixel pairs actually belong to the neighboring bins after expansion .
this has to be accounted for , and to do so we sort the pixel pairs in the the high - resolution output table according to actual bin number , and output these as incomplete two - point tables . finally , after subjecting all low - resolution tables to this operation , we complete the expansion process by collecting all partial tables of the same output bin number into complete two - point tables .
these two last expansion steps are repeated as many times as necessary to obtain the desired resolution . in the following three subsections ,
we describe each of these steps in more detail .
the first step is illustrated in figure [ fig : inmemgen ] while the two last steps are illustrated in figure [ fig : expansion ] .
assume first that we actually do have enough memory available to simultaneously store all @xmath43 pixel pairs in the computer s ram , i.e. , that there are no memory restrictions .
in that case the procedure is quite simple .
first we allocate an @xmath29 array , and run through the set of all pixel pairs . for each pair
we compute the angular distance @xmath44 between the two pixels @xmath23 and @xmath5 , and mark both the @xmath5th row of the @xmath23th column and the @xmath23th row of the @xmath5th column by the integer part of @xmath45 .
once this process is completed , the desired two - point tables can easily be extracted from the main array the @xmath2th table is found by scanning each column for the number @xmath2 , while storing the row numbers of these entries in the output array .
note that the output two - point table is automatically sorted by this method , maintaining the @xmath37 scaling of the algorithm ( rather than @xmath46 $ ] as would be the case if explicit sorting was necessary ) .
as is indicated in figure [ fig : inmemgen ] , we do not count the center pixel as a valid neighbor to itself in the zeroth bin . or speaking in terms of correlation functions , we never let a pixel be multiplied with itself . for most experiments
the noise in each pixel is quite high , but , fortunately , it is also usually independent from pixel to pixel . in order to eliminate this noise bias , we exclude powers of single pixel values from the correlation functions .
this is a general problem for all quadratic estimators .
let us also make one note on how to compute the binned distance between two pixels efficiently : the standard method is to take the dot product of the two pixel center vectors , compute the inverse cosine of that product , divide the angular distance by the bin width , and keep the integer part .
this method is quite slow because of the inverse cosine operation which requires about 40 times the cpu time of a multiplication .
when using the healpix nested pixelization , a faster alternative is available by means of a linear search : one simply allocates an array , ` binlim ` , of length @xmath47 which contains the cosine of the bin limits , such that @xmath48 } = \cos([i-1]d\theta)$ ] and @xmath49 } = \cos(n_{\textrm{bin}}d\theta)$ ]
. then the correct bin number , ` bin ` , may be found by searching for the array position defined by @xmath50 } \le \cos\theta_{ij } < \texttt{binlim[bin+1]}$ ] .
the crucial point here is that the bin position must be stored between consecutive searches : since the pixels @xmath5 and @xmath51 are very close to each other on the sky in the nested healpix pixelization , the distance between @xmath23 and @xmath51 will be almost the same as the distance between @xmath23 and @xmath5 .
therefore one does not have to move more than a few steps to find the correct bin for each pixel .
for the same reason the search should _ not _ be implemented by a quicksearch - algorithm , but rather with a straight - forward linear method . while the former is superior for general searches , the latter has considerably less overhead per step , and is therefore much faster if we only want to move a few steps , as is the case here .
the above trick is in fact very useful in most two - point correlation function algorithms , and if the nested healpix scheme is employed , one can often reduce the total cpu time by 60 - 70% .
another advantage is that the bin limits may be set arbitrarily , and therefore the method seamlessly accommodates for alternative binning schemes . in particular ,
gauss - legendre binning ( ie . , placing the bins at the roots of the legendre polynomials ) is often useful if one wants to integrate the two - point correlation function to find the power spectrum ( szapudi et al .
1991b ) .
the output from the above process is a completely valid set of two - point tables , but , unless massive amounts of memory is available , they only describe a low resolution map
. a computer with 1 gb of ram can typically only handle up to 15000 pixels through the above description , and a super - computer with shared memory of @xmath52 gb up to about 100000 pixels .
usually we therefore have to expand the structure set segment by segment in order to obtain the desired resolution and pixel number . as mentioned above
, increasing the resolution of a healpix map is particularly simple in the nested pixelization each mother pixel is replaced by four children pixels having pixel numbers given by @xmath41 , where @xmath23 is the pixel number of the mother pixel and @xmath42 .
the expansion in resolution is therefore easily performed by substituting each entry of the original two - point table by a @xmath53 sub - array ( the zeroth row has to be treated individually , though ) . however , since pairs that originate from the same mother pixel may belong to different high - resolution bins , we can not simply substitute the low - resolution pixel numbers by the above method , and hope that the resulting array will be a valid high - resolution two - point table .
usually we also want to increase the number of bins during such an expansion , considering that a higher resolution supports a smaller bin width .
for these two reasons we have to compute all angular distances once again to determine to which bins the various pairs belong . based on this information
we extract a set of partial two - point tables , similar to what was done in the first , low - resolution step , and store the resulting partial tables on file , labeled by two numbers , ` bin_from ` and ` bin_to ` .
the output from the above process is a set of arrays characterized by the set of parameters \{@xmath54 , @xmath55 , @xmath56 , @xmath57 , ` bin_from ` , ` bin_to`}. the final task is to collect all partial tables with the same ` bin_to ` ( containing pixel pairs in the same output bin , but originating from different low - resolution bins ) into complete two - point tables .
this operation obviously amounts to merging the partial array columns , while maintaining the already sorted relationships , and inserting the final result into a large output array .
since the partial tables are already sorted according to pixel number , no additional sorting is required .
this is actually true for all the above steps no explicit sorting is necessary to generate sorted output two - point tables .
at this point we have obtained a set of high - resolution two - point tables stored on disk files . and already at this point we may estimate the two - point correlation function extremely efficiently .
however , our main goal is to study the higher - order correlation functions , and we therefore proceed to the construction of triangles by means of the two - point tables , before turning to applications .
our algorithm for constructing triangles on the healpix sphere is fully equivalent to ruler and compass construction of triangles ( i.e. , where an equilateral triangle is constructed by first drawing a base line , then from a given first vertex on this line drawing a circle with the given length as radius , and then drawing a new circle with the same radius from the second vertex at the intersection between the line and the first circle , finding the third vertex of the triangle at the intersection between the two circles ) , the only difference is that the physical compass is replaced with the two - point tables . wanting to find all _ equilateral _ triangles having a given pixel @xmath23 as one vertex and edges given by a given bin number @xmath2 , the first step is to look up the @xmath23th column of the @xmath2th two - point table .
all entries in this column are located in the correct distance from pixel @xmath23 .
next , we pick an arbitrary pixel , @xmath5 , from this column . together
these two pixels constitute the base line of the triangle .
now we scan the @xmath23th and the @xmath5th column simultaneously , searching for identical entries .
any common entries in the two columns will be located in the correct distance from both pixel @xmath23 and @xmath5 , and therefore the three pixels together span an equilateral triangle .
once we find such a triplet we store it in an auxiliary array for later use .
we then keep scanning until the bottom of one column is reached .
note that since the columns are sorted according to pixel number , the searches scale as @xmath35 , rather than as @xmath36 in the unsorted case . for our ranges of pixel numbers in the maps and bin widths
, we typically find between two and six valid pixels in each such scan . if we are only interested in scalar valued symmetric three - point correlation functions , all of these triplets are acceptable , and in such cases we store all of them for later use .
however , if we want to estimate for instance the four - point correlation function , asymmetric three - point correlation functions or spin-2 field correlation functions ( for polarization maps ) , we need to know the orientation of the triangle , or in other words , we need to know whether we traverse the triangle clockwise or counter - clockwise when listing the pixels as \{@xmath23 , @xmath5 , @xmath58}. in these cases we introduce the convention that the two first pixels form the base line and the third pixel is positioned _ above _ the base line . in other words , @xmath59 .
the above procedure results in a @xmath60 array , in which each column is an acceptable pixel triplet for the current geometric configuration ( in practice we organize the the array slightly differently , to accommodate for fast summing over the third pixel see [ sec : calculations ] for details ) . at this point we can estimate the three - point correlation function by multiplying the three pixels together , and sum it all up . however , if we want to estimate the four - point correlation function we need to go one step further . we find the simplest example of a more complicated geometry when we want to estimate the rhombic four - point correlation function .
this configuration consists of two equilateral triangles `` glued '' together on one edge .
we can therefore use our list of equilateral triangles to find the set of all such quadruples .
the idea here is simply that the two triangles must have the same base line , but in reversed order ( since they must have opposite orientation ) .
thus , we must search for pairs of triangles for which @xmath61 and @xmath62 , where @xmath63 is the first vertex of the first triangle , @xmath64 is the second vertex of the first triangle , etc .
the four required pixels are then given by the two common vertices and the two third pixels .
this time we store the sets of four pixels in a @xmath65 array , listed so that the two first entries of each column forms the base line , the third pixel lies above the base line and the fourth lies below the base line . note that any @xmath0-point polygon may be built up by triangles , and the same ideas may therefore be used for computing any @xmath0-point correlation function . in the above examples we have focused on polygons for which all edges have the same length .
however , the same algorithm can easily be used to also produce general triangles , and thereby general quadrilaterals .
suppose , e.g. , that we want to find all triangles with edges given by the bin numbers @xmath66 .
in that case we would use the @xmath2th two - point table to find pixel pairs , @xmath23 and @xmath5 , which forms the base line of the triangle .
then we would simultaneously scan the @xmath23th column of the @xmath58th two - point table and the @xmath5th column of the @xmath67th two - point table to find the desired third pixels .
we start with the case where we want to estimate the two - point correlation function for separation given by bin number @xmath2 , for which we already have constructed a two - point table as described above .
if the number of different pairs with separation in this bin is @xmath68 , then the two - point correlation function is estimated by @xmath69 ) \cdot \sum_{j < i } \delta t(\texttt{table}[j , i ] ) .
\label{eq : comp_twopt}\ ] ] explicitly , we sum up all field values in each column , and multiply that sum with the zeroth row value .
it is difficult to imagine a more efficient method for estimating the two - point correlation function than this ; we basically perform only one addition for each pixel pair .
note also that we only sum over pairs for which @xmath70 in order to avoid double counting , and thus the total cpu time is reduced by a factor of 2 . for monte
carlo studies the initialization cost is paid once only , and in such cases the above equation is where the massive amounts of cpu time are spent . although our method intrinsically is quite slow , being an @xmath37 method , it is for this reason probably as fast as a two - point correlation function algorithm will ever be , without introducing either fourier - methods or large - scale smoothing .
let us now take a look at the higher - ordered correlation functions , and suppose that for any given configuration we have already computed an @xmath71 array containing all pixel multiplets satisfying the binned geometric requirements . here
@xmath0 is the order of the correlation function and @xmath17 is the number of pixel multiplets .
with such a table the correlation function is estimated by @xmath72 ) \cdots \delta t(\texttt{table}[n , i ] ) .
\label{eq : comp_npt}\ ] ] when the bin width is very small , there is typically only one polygon corresponding to each base line , and in such cases it is not possible to accelerate equation ( [ eq : comp_npt ] ) .
however , when the bin width is comparable to or larger than the pixel size , one often finds that for two given base line pixels there may be several acceptable third pixels .
if we want to estimate the three - point correlation function in such cases , we may organize our arrays a little differently : rather than listing each triplet individually , we may first list the two base line pixels , then the number of third pixels , and finally list all those third pixels . obviously , a one - dimensional array is more suitable for this organization than the original two - dimensional one , since the number of third pixels may vary strongly from set to set .
if the average number of third pixels is larger than about 1.5 ( as often is the case in real - life applications ) , this linear organization will both result in a smaller triplet array , and speed up the computations , since equation ( [ eq : comp_npt ] ) is replaced by @xmath73 if @xmath74 , we here have to perform two multiplications and two additions for each base pair , a total of four operations , while in the `` unrolled '' organization , the corresponding numbers are six multiplications and two additions .
thus the total cpu time is roughly halved by introducing the linear organization .
the same trick is even more powerful when estimating four - point correlation functions , since we in that case can sum over both the upper and the lower groups of third pixels .
having , e.g. , two pixels in both groups ( or in other words , having four different quadruples ) , would require twelve multiplications and four additions for each base pair in the `` unrolled '' organization , but only three multiplications and three additions in the linear one .
note also that this linear organization is very simple to construct , it falls out naturally from the method we use to construct the triangles in the first place where we fix our interest on two base line pixels , and scan through the two corresponding columns for third pixels .
thus , producing the linear organization is only a matter of storing the results from the scans appropriately .
before benchmarking the algorithms , we briefly discuss a few ideas of interest to anyone who wants to use these algorithms in practice , and which are implemented in our software .
first we show how to optimize the cache usage in the case of monte carlo studies , and then we briefly comment on how to parallelize these algorithms .
finally , we show how to compress the two - point tables in order to save disk space . from the previous sections it should be clear that these @xmath0-point algorithms are very close to being optimal for monte carlo simulations . in applications where the initialization cost is small compared to the computational costs ,
there is not much left to gain from an algorithmic point of view we have removed all geometry from the computations , and all that is left is to multiply field values together as fast as possible , and sum it all up . however , there are two main issues left to consider , cache optimization and parallelization .
let us start by looking at the cache usage . in a monte carlo application
, we want to estimate the two - point correlation function for , say , 1000 different maps .
the naive approach is simply to loop over the maps , applying equation ( [ eq : comp_twopt ] ) repeatedly .
however , this has two negative effects .
first of all , the cpu has to move the two - point table information from the main memory into the cache 1000 times .
secondly , the number of hits per memory page in the map array may be less than optimal , since consecutive rows may contain widely separated pixel numbers .
on the other hand , both the fact that the columns of the two - point tables are sorted , and the nested healpix pixelization , improve the situation .
if we have considerable amounts of physical memory available , we can do better by analyzing all maps simultaneously . in this case
we store all 1000 maps row - wise in a two - dimensional array , so that each column contains the map values of one single pixel . with this large super - array
we can loop over the maps in the innermost loop , rather than in the outermost .
thus we are guaranteed almost perfect cache usage . also , we only need to look up the two - point table once , further reducing the memory traffic .
the question is then , do we have enough memory available to store all 1000 maps ? for a map of , say , three million pixels , the answer is clearly `` no '' , since that would require about 12 gb of memory , assuming single precision floating point numbers .
on the other hand , for maps with less than 50000 pixels we only need 400 mb , and that is not a large amount for current computers .
even with our algorithms , estimation of @xmath0-point correlation functions is cpu intensive , and maps of several million pixels are out of reach for today s computers .
thus , cpu time is the limiting factor , rather than available memory . for this reason we often choose to analyze high - resolution maps region by region ( by the introduction of masks ) , where each region contains no more than , e.g. , 50000 pixels to @xmath75 , and then computing each correlation function in @xmath76 bins on a set of complementary hemispheres .
the smaller scales were studied by partitioning the sky into @xmath77 disks , on each of which the three - point function was estimated . ] . in these cases
we can discard all the unused map information , and retain the included pixels only .
explicitly , we now introduce a new pixel ordering scheme , counting relative to the mask rather than to the resolution parameter @xmath20 .
the maps are converted from the standard healpix numbering to the mask - based ordering by extracting all included pixels , while conserving their relative order .
thus , the first included pixel is assigned the new pixel number 1 , the second included pixel is assigned the new pixel number 2 etc . by storing these conversion rules in an external table , we may easily go back and forth between mask - based and map - based pixel ordering . in order to optimize the cache usage for high - resolution maps , we convert all maps and two - point tables to mask - based pixel ordering , and then proceed as before .
the final issue to consider from an efficiency point of view is parallelization .
there are two different cases , the geometric computations and the actual correlation function calculations .
the latter part is quite trivial , considering that we already in the algorithmic design have divided the problem into completely separate parts usually we want to compute many bins of each correlation function , and then we let each processor work on its own geometric configuration .
this obviously results in perfect speed - up ( i.e. , doubling the number of processors halves the total cpu time ) .
parallelization of the construction of the two - point tables is not quite as simple , at least not for the very first step , in which we generate a full @xmath29 table .
however , one possibility is to let each processor work on separate columns , and then merge the partial results into one complete array in the end .
once this task is performed , the problem is once again divided into separate parts , and the expansion steps may be trivially parallelized , like the @xmath0-point calculations .
if disk space is a limiting factor , one may be discouraged by the substantial size of the two - point tables . in such cases
one may wish to implement file compression .
one common algorithm for this purpose is the huffman ( 1952 ) coding method .
the idea behind this algorithm is to assign short bit strings to frequently occurring symbols ( i.e. , integers in our setting ) , and long bit strings to rarely occurring symbols .
the average number of bits per symbol is thus minimized .
applied to our two - point tables , it appears at first sight as though little is gained ; since any pixel number has to occur somewhere in each column , each number is more or less equally frequent in each two - point table .
however , one property of our two - point tables still makes huffman coding most efficient , namely the fact that each column is sorted according to increasing pixel number .
therefore , if we subtract the values in two consecutive rows , we get a small positive number , and the total histogram of the table is strongly skewed toward low values .
this new , transformed array may be compressed very efficiently .
a simple way of taking advantage of this idea is to do the following : first we transform the original array by subtracting consecutive rows , and store the resulting array in an external file .
then we compress that file using any available compression utility ( which usually relies on a combination of huffman and run - length coding ) , e.g. , ` gzip ` .
this method routinely reduces the required disk space by 80 - 90% for standard combinations of @xmath20 and @xmath39 . in the current implementation
there are special routines for reading and writing two - point tables .
compression may therefore be enabled simply by adding a few appropriate lines in each of those routines ( subtracting consecutive rows , making a system call to ` gunzip`/`gzip ` etc . ) , leaving the rest of the source code unchanged .
we have run several tests to determine what resources are required for application of the above algorithms .
the parameters of these tests mimic those used in the small - scale wmap analysis of @xcite , in that we let our region of interest be a disk on the healpix sky , pixelized at @xmath78 , and with a bin width of @xmath79 .
the only difference is that we let the radius of the disk vary between @xmath80 and @xmath81 ( in steps of @xmath82 ) in order to see how the algorithms scale with respect to number of pixels .
these disks correspond to pixel numbers between 5940 and 53464 , and we should therefore get a good picture of how the algorithms behave in real - world situations . the computer used in the following tests is a compaq ev6 alpha server .
we start by generating a set of two - point tables for each disk , and measure the cpu time and disk space spent in this process .
next we estimate the two- , three- and four - point correlation functions from 100 simulated maps , and find the number of pixel multiplets ( pairs , triplets and quadruples ) , the cpu time per realization and initialization cost as a function of @xmath22 .
the results from the two - point table generation tests are shown in figure [ fig : struct_gen ] . in panel
( a ) we see that the time spent on this process does not follow the expected @xmath37 relationship .
in fact , the cpu time scales as @xmath83 , while the input / output time ( i.e. , disk activity ) does nt even follow a power law .
the explanation is the same in both cases : each of these quantities follows a relation of the general form , @xmath84 .
algorithmic overhead and initialization costs are typically linear in @xmath22 , and these are likely to dominate for a small number of pixels . in other words , @xmath85 .
what we see in figure [ fig : struct_gen_cputime ] is the manifestation of this . for the relatively modest number of pixels we consider
, we do not reach the asymptotic @xmath43 region , and therefore the effective scaling coefficient in this region is smaller than 2 .
this result may seem somewhat counter - intuitive ; normally a low scaling coefficient is interpreted in a positive direction .
that is not the case here .
rather , it may suggest that the current implementation is sub - optimal , and that there is room for improvements .
however , it should also be noted that a truly @xmath86 implementation can probably only be devised if one is able to store the full @xmath87 distance matrix in the physical ram simultaneously , and in that case our methods are superfluous . the important conclusion , however , is that the current algorithms are in fact sufficiently efficient to handle data sets with at least a few hundred thousand pixels , considering that it only takes about an hour to generate a set of two - point tables for 50000 pixels . in figure
[ fig : struct_gen_diskspace ] the corresponding disk space usage is shown .
here we do see a virtually perfect @xmath86 scaling , a result which should come as no surprise .
the linear overhead for storing the two - point tables is basically just a few extra rows ( for storing the pixel centers and padding extra space with -1 s ) , and this is normally negligible compared to the total number of `` neighboring '' pixels .
note also that disk space is not a critical factor in these computations ; a data set of 50000 pixels requires less than 20 gb of space , a trivial amount of disk space for current computers .
if we want to include as many pixels as 200000 , we would need 160 gb , and then perhaps compression might be needed , as described in the previous section .
then the disk space requirements would be less than 30 gb , and once again well within the limits of current computers .
we may therefore conclude that the first step in the two - point table strategy , the construction of the two - point tables , do not pose a serious problem neither in terms of wall - clock time nor disk space , when analyzing data sets which contain fewer than a few hundred thousand pixels .
finally , we check that the amount of time spent on actually computing the correlation functions is not prohibitively large .
this experiment is carried out by measuring both the initialization and the addition / multiplication time for each of the two- , three- and four - point correlation functions , for one particular configuration on each of the disks described above .
the configurations were arbitrarily chosen to be the @xmath88 two - point , equilateral three - point , and rhombic four - point configurations .
the results are shown in figure [ fig : comp_resources ] . in panel
( a ) we see the numbers of pairs , triplets and quadruples contributing to the bin as a function of @xmath22 .
note that these numbers increase almost linearly with the number of pixels in the data set , with best - fit power - law indexes of 1.15 , 1.23 and 1.34 , respectively .
this is an intuitive result ; doubling the area of interest also doubles the number of polygons that fit into it . the deviation from a perfect linear relationship
is caused by boundary effects .
given this linear relationship , we also expect the cpu time required for multiplying and adding the pixel multiplets together to increase linearly with @xmath22 . in figure
[ fig : corr_init ] we see that this is the case .
it is important to note that the cpu time for computing any single bin of any @xmath0-point correlation function increases _ linearly _ with the number of pixels in the data set .
it is well known that the evaluation of a general @xmath0-point correlation function scales as @xmath89 . however
, this relation only applies to the computation of the full correlation function , i.e. , when including all @xmath90 pixel multiplets with all possible multiples of the bin size as lengths of the edges . in most applications ,
this is found unfeasible , and one uses only a small subset of all available geometric configurations of the @xmath0 vertices . in those cases the computational scalings are less severe ; the cost for computing one single bin increases only proportionally with the number of pixels in the data set , and that time is not radically different for the two- , three- , and four - point correlation functions . only the number of different available configurations change from correlation function to correlation function , not the cost for computing one given separation bin . in figure [ fig :
corr_init ] and [ fig : corr_comp ] we see that for a data set with 50000 pixels it takes about 90 seconds to find all equilateral triangles with @xmath88 edges , and 200 seconds to find all rhombi .
in general it takes about twice as long to find all quadruples as finding all triangles , an obviously logical result quadrilaterals are constructed by merging two triangles .
however , it is worth noting that this extra cost can be avoided when computing quadrilaterals constructed from isosceles triangles , since one then does not have to worry about the orientation of the triangles .
as mentioned in the introduction , our methods are designed for monte carlo studies , and in figure [ fig : corr_comp ] the cpu time for analyzing 1000 realizations simultaneously is shown , excluding initialization .
we see that for a 50000 pixel data set it takes about 30 seconds to estimate one two - point function bin , 150 seconds for one three - point configuration and 210 seconds for one four - point configuration . for the higher - order correlation functions
these numbers are about the same as the initialization costs , and we therefore conclude that it is crucial to organize our programs so that the initialization costs are paid once only
. a typical example of an @xmath0-point correlation function analysis could be the following : we analyze 5000 monte carlo realizations on a disk containing about 50000 pixels , with 100 configurations ( distance bins ) in the two - point correlation function , 500 different configurations in the three - point correlation function , and 1000 different configurations in the four - point correlation function .
we assume that the time for computing each of these configurations is constant and equal to the @xmath88 configurations . while this obviously is not strictly true ( more pixel multiplets are associated with larger configurations ) , this particular size is a good average of what is used in real - world cmb analysis .
first we have to generate the two - point tables for the disk , a process which is performed only once .
the total wall - clock time for this is about 3 hours , and we need about 20 gb of disk space to store the tables .
next , about 1 gb of ram is required for storing the maps , and an additional few hundred mb for generating the triangles and quadrilaterals .
this amount of disk space and ram is not a major obstacle for current workstations or super - computers .
next , the estimation of the two - point correlation functions takes about @xmath91 , while the three - point configurations require @xmath92 . finally , the four - point configurations need @xmath93 .
none of these numbers are prohibitively large , and we can therefore conclude that the proposed algorithms are able to handle real - world data sets using reasonable computational resources .
we have described a set of algorithms for estimating @xmath0-point correlation functions from a pixelized map . the fundamental idea is first to locate all pixel multiplets ( pairs , triplets , quadruples etc . ) corresponding to one given geometric configuration , store these in a table , and then finally use that information to efficiently add and multiply the pixel values together .
the pixel multiplets are found by searching through a set of two - point tables , motivated by ruler and compass construction of triangles .
the construction of the two - point tables only take a small amount of cpu time ( typically a few hours ) , and since this operation is performed only once , it is completely negligible compared to the following operations
. the storage requirements of these tables are well within the limits of current computers .
we have found that the initialization cost for each configuration roughly equals the computation time of 1000 realizations , so for a monte carlo simulation with , say , 5000 realizations , the major amount of time is spent on adding pixel values together .
thus , the geometric computations are removed from the problem . in other words ,
our methods are probably as fast as any fully @xmath1 algorithm can ever be , when applied to a monte carlo analysis . however , the main advantage of our method over a brute - force analysis is the fact that we estimate the correlation functions configuration - by - configuration no cpu time is spent on uninteresting configurations .
this effectively reduces the scaling of an experiment in which only a fixed number of configurations is desired ( such as only equilateral triangles , rhombic quadrilaterals , or configurations covering scales smaller than , say , @xmath94 ) to @xmath95 .
thus , even large data sets may be subjected to an @xmath0-point correlation function analysis .
another consequence of this configuration - wise division is optimal parallelization . by letting each processor work on separate configurations , optimal speed - up
is obtained , at the cost of some extra ram requirements .
hke and pbl thanks the research council of norway for economic support , including a phd studentship for hke .
this work has also received support from the research council of norway ( programme for supercomputing ) through a grant of computing time .
the authors acknowledge use of the healpix ( grski et al .
1999 ) software and analysis package for deriving the results in this paper .
barriga , j. , & gaztaaga , e. 2002 , , 333 , 443 bennett , c. l. et al .
2003 , , 148 , 1 bernardeau .
f. , van waerbeke , l. , & mellier , y. 2003 , , 397 , 405 connolly , a. j. , et al .
2002 , , 579 , 42 croft , r. a. c. , dalton , g. b. , & efstathiou , g. 1999 , , 305 , 547 dalton , g. b. , croft , r. a. c. , efstathiou , g. , sutherland , w. j. , maddox , s. j. , & davis , m. 1994 , , 271 , l47 davis , m. , efstathiou , g. , frenk , c. s. , & white , s. d. m. 1985 , , 292 , 371 eriksen , h. k. , banday , a. j. , & grski , k. m. 2002 , , 395 , 409 eriksen , h. k. , hansen , f. k. , banday , a. j. , grski , k. m. , & lilje , p. b. 2004 , apj , in press ( astro - ph/0307507 ) frieman , j. a. , & gaztaaga , e. 1999 , , 521 , l83 gangui , a. , lucchin , f. , matarrese , s. , & mollerach , s. 1994 , , 430 , 447 grski , k. m. , hivon , e. , & wandelt , b. d. , 1999 , in evolution of large - scale structure : from recombination to garching , ed .
j. banday , r. k. sheth , & l. n. da costa ( garching , germany : european southern observatory ) , 37 huffman , d. a. 1952 , proc .
ire , 40(9 ) , 1098 maddox , s. j. , efstathiou , g. , & sutherland , w. j. 1996 , , 283 , 1227 moore , a. w. et al .
2001 , in mining the sky , ed .
a. j. banday , s. zaroubi , & m. bartelmann ( berlin ,
heidelberg : springer - verlag ) , 71 peebles , p. j. e. , 1973 , , 185 , 413 szapudi , i. , prunet , s. , & colombi , s. 2001a , , 561 , 11 szapudi , i. , prunet , s. , pogosyan , d. , szalay , a. s. , & bond , j. r .. 2001b , , 548 , l115 takada , m. , & jain , b. 2003 , , 344 , 857 totsuji , h. , & kihara , t. 1969 , , 21 , 221 | we develop , implement and test a set of algorithms for estimating @xmath0-point correlation functions from pixelized sky maps .
these algorithms are slow , in the sense that they do not break the @xmath1 barrier , and yet , they are fast enough for efficient analysis of data sets up to several hundred thousand pixels .
the typical application of these methods is monte carlo analysis using several thousand realizations , and therefore we organize our programs so that the initialization cost is paid only once .
the effective cost is then reduced to a few additions per pixel multiplet ( pair , triplet etc . ) .
further , the algorithms waste no cpu time on computing undesired geometric configurations , and , finally , the computations are naturally divided into independent parts , allowing for trivial ( i.e. , optimal ) parallelization . |
of a micro - tpc , a time projection chamber ( tpc ) , which can detect three - dimensional fine tracks of charged particles , has been wanted for years .
we developed the @xmath0-pic , a two - dimensional fine position detector , and a fast readout electronics in order to realize such an `` electric cloud chamber '' . in this paper , the performance of the micro - tpc , together with one example of its applications , is described .
the micro pixel chamber , or the @xmath0-pic , is a pixel type gaseous two - dimensional imaging detector which takes over the outstanding properties of the microstrip gas chamber ( msgc@xcite ) such as the good position resolution and operating capacities under high flux irradiation .
the @xmath0-pic is manufactured by the print circuit board ( pcb ) technology in contrast to an older concept pixel type detector , the microdot chamber@xcite , which is made in the mos technology . with the pcb technology ,
large area detectors can be made cheaply , which is an important feature for developing various kinds of applications . as a consequence of the geometrical properties , discharge problems are less disturbing with @xmath0-pic @xcite , so that stable operations at high gas gain can be realized .
-pic . ] the schematic structure of the @xmath0-pic is shown in fig .
[ upicstruct ] .
the @xmath0-pic is a double - sided pcb with a 100 @xmath0m - thick polyimide substrate .
256 anode strips are formed on one side of the @xmath0-pic while 256 cathode strips are orthogonally placed on the other side .
both anode and cathode strips are placed with a pitch of 400 @xmath0 m .
cathode strips have holes of 200 @xmath0 m diameter and anode electrodes of 50@xmath170 @xmath0 m diameter are formed on the anode strips at the center of each cathode hole .
the signals from anode strips and cathode strips are of the same size , in contrast to the msgcs , whose pulse heights from back strips are 20@xmath130@xmath5 of those from anode strips@xcite .
we developed the @xmath0-pic of 10 @xmath2 10 @xmath6 detection area with 256 @xmath2 256 anode electrodes .
the @xmath0-pic is mounted on the mother board of 30 @xmath2 30 @xmath6 area by the bonding technique .
fe radioactive source.,width=288,height=240 ] we irradiated a non - collimated @xmath7 radioactive source and measured the output charge from @xmath8 pixels so that we calculated the gas gain .
the gas mixture of argon 80@xmath5 and ethane 20@xmath5 at 1 atm was flowed for the measurement .
the gas gain of the @xmath0-pic is plotted as a function of the anode voltage in fig .
[ gain_hv ] .
we achieved a maximum gas gain of 1.5 @xmath9 and the gas gain for the stable operation of 5000 without any other multipliers . in fig .
[ fe55_spec ] , the spectrum of the x - rays from the @xmath7 radioactive source is shown .
the energy resolution is @xmath10 ( fwhm ) for the 5.9 kev x - rays .
this result is worse than our previous results @xmath11 obtained with the trial piece which has 3 @xmath2 3 @xmath6 detection area@xcite .
the non - uniformity of the anode electrodes in the large detection area makes the resolution worse .
new technologies for forming the uniform electrodes are now sought for so as to achieve higher gas gain and better energy resolution .
long term stability is a necessary feature in order to realize the various applications of the @xmath0-pic .
we operated the @xmath0-pic for 120 hours with the gain of more than 3000 without serious discharge .
gas gain was measured every one hour by the same method described above .
the gain variation is shown in fig .
[ gain_time ] as a function of the elapsed time since the anode voltage supply .
slight increase of the gain was observed in the first 20 hours , which we think is due to the polarization effect of the substrate .
the gain was stable within 10@xmath5 after 20 hours .
we took the x - ray image of the test chart with an x - ray generator ( kevex x - ray cu028 , tungsten target ) .
the acceleration voltage of the x - ray generator was 18 kv .
low energy x - rays were filtered with 50 @xmath0 m thick copper and 1 mm thick aluminum and the energy of the x - rays was distributed between 14 kev and 18 kev with the intensity peak at 17 kev .
we used the gas mixture of xenon 70@xmath5 and ethane 30@xmath5 at 1 atm .
we chose xenon gas in order to achieve clear images .
the practical range of the 20 kev electron in argon gas is 3 mm while that in the xenon gas is 1 mm . the detection depth was set to be 2 mm .
the test chart is 5 @xmath2 5 @xmath6 area and slits of various widths are scribed in the thin lead layer on the plastic plate . in fig .
[ test_image ] , obtained x - ray image is shown .
the slits of 1 mm width are seen clearly .
we measured the spatial resolution from the edge image of the test chart .
acceleration voltage of the x - ray generator was set at 10 kv for this measurement in order to suppress the range of the electrons in the xenon gas .
the active area was set to be @xmath12 with one of the edges at the center of the active area .
the test chart image was projected along one of the edges for 1.3 cm .
the projected image was fitted with the function @xmath13 , where @xmath14 is the error function and @xmath15 and @xmath16 are the fitting parameters .
the calculated spatial resolution(@xmath16 ) was 160 @xmath0 m .
the histogram of the projected image of the test chart and the best fit function is shown in fig .
[ knife ] .
a pipe - line readout electronics system for the @xmath0-pic was developed as well as the detector itself@xcite .
the block diagram of the readout electronics is shown in fig . [ daq_block ] .
signals from the @xmath0-pic are amplified and discriminated by the preamplifier cards attached to the rear of the mother board .
the preamplifier cards output both analog and digital signals .
the digital signals are encoded by the position encoding module which consists of five field programmable gate arrays ( fpgas ) and recorded by the memory module which works at the clock of 20 mhz .
analog signals from the 16 cathode strips are summed and digitized by the 100 mhz flash adc ( fadc : repic rpv-160 ) to determine the event energy . in order to check the performance of the @xmath0-pic and the readout system for high rate signals
, we irradiated the @xmath0-pic with the x - rays from the x - ray generator .
the acceleration voltage was 20 kv and the intensity was controlled by way of varying the tube current between 0.01 ma and 0.6 ma .
the data acquisition rate is plotted against the x - ray intensity in fig .
[ highratetest ] .
we achieved the maximum data acquisition rate of 7.7 mhz .
this rate is twice higher than our previous readout system@xcite .
the saturation is thought to be due to the software for the fpgas and optimizations are now under way . -pic and the readout system.,width=288 ]
with the @xmath0-pic and the readout electronics system described above , we developed a time projection chamber ( tpc ) which has fine time and spatial resolutions@xcite .
a gas enclosure with a drift length of 8 cm was attached to the @xmath0-pic .
schematic drawings of the tpc gas enclosure is shown in fig .
[ tpc_structure ] .
a negative high voltage of 3.3 kv is supplied to the 1 mm thick aluminum window which is connected to the aluminum plane of 0.3 mm thick through 22 m@xmath17 .
fifteen field cage electrodes with 4 mm widths and 1 mm spacings are connected to the aluminum plane in series through 10 m@xmath17 resistors .
the electric field of 0.4 kv / cm is produced uniformly in the inner region .
we measured the drift velocity of electrons and the gain dependence on the drift length in the micro - tpc using the cosmic - ray muons .
we placed two plastic scintillators on both sides of the micro - tpc perpendicular to the direction of the electric field .
we acquired the micro - tpc events in coincidence with both of the two scintillators .
[ fadc_muon ] shows the shapes of signals from cathode electrodes for a cosmic - ray muon .
the signals are not identical because of the fluctuation of the gas gain due to the shape of the anode electrodes .
the drift velocity was measured at various electric fields .
the result shown in fig . [ driftv ] is consistent with the previous measurement @xcite .
no gain difference was observed throughout the drift length .
we detected the tracks of the electrons from a @xmath18 radioactive source which emits electrons with the maximum kinetic energy of 0.546 mev and 2.282 mev .
the micro - tpc was triggered by a plastic scintillator of 6 @xmath2 6 @xmath2 2 mm@xmath19 placed between the radioactive source and the micro - tpc .
the @xmath0-pic was operated at the gas gain of 7000 .
one of the typical three - dimensional tracks is shown in the upper panel of fig .
[ 3d_beta ] , while the projections of the several tracks on the @xmath0-pic plane are shown in the lower .
we only detected the points where large energy deposition took place because of the insufficient gas gain , therefore the electron tracks were detected as the dispersed points . since the principle of the micro - tpc is confirmed , dense tracks like ones taken by a cloud chamber will be taken when we ve achieved a sufficient gain .
-pic plane ( lower).,title="fig:",width=336 ] -pic plane ( lower).,title="fig:",width=288 ] one of the most interesting applications of the micro - tpc is a mev gamma - ray imaging detector which makes an event - by - event reconstruction possible@xcite .
typical mev gamma - ray imaging detectors of these days with double or multiple compton method@xcite measure the energy deposited to the electron ( @xmath20 ) , the energy of the scattered gamma - ray ( @xmath21 ) and the direction of the scattered gamma - ray ( @xmath22 ) .
the scattering angle of the gamma - ray ( @xmath23 ) is calculated by the following equations . @xmath24 here , @xmath25 is the mass of the electron .
the event cone is determined from @xmath22 and @xmath23 .
the principle of the gamma - ray reconstruction with double comptom method is shown in fig . [ eventcone ] .
our new concept for the mev gamma - ray imaging is to measure the direction of the compton scattered electrons ( @xmath26 ) with the micro - tpc in addition to @xmath20 , @xmath21 , and @xmath22 .
since the micro - tpc can detect the track of the electrons as we have shown in the previous section , the hybrid detector of the micro - tpc and the enclosing scintillators can realize this new concept . with this new concept detector , we can determine another angle @xmath27 so that the incident gamma - ray is reconstructed for each event . the principle is shown in fig . [ mev_prin ] .
we have a redundant measured value @xmath28 which can be used for the background rejection .
since no collimators are necessary , a large field of view can be realized .
we developed a prototype of the mev gamma - ray imaging detector with the micro - tpc of 10 @xmath2 10 @xmath2 8 @xmath29 volume and an nai ( tl ) scintillator of 4@xmath24@xmath21 size read by 25 single anode pmts of 3/4 diameters ( hamamatsu r116 ) .
the picture of the prototype is shown in fig .
[ prototype_photo ] .
we tested the prototype with a radioactive source of @xmath30ba which mainly radiates 356 kev gamma - rays .
the setup is shown in fig .
[ reconstructed ] .
we set the origin of the coordinates at the center of the @xmath0-pic detection plane and x , y and z axes are taken along the cathode , anode and drift direction of the micro - tpc , respectively .
the radioactive source was set at ( -4.0,-4.3,-4.8 ) without any collimators and the center of the nai ( tl ) scintillator was set at ( -5.5,-5.5,22.9 ) .
electron tracks are detected by the micro - tpc , while the energy and the position of the gamma - rays are taken by the nai ( tl ) scintillator .
a gate of two-@xmath0s was opened after the nai ( tl ) trigger for the micro - tpc signals and coincident events were recorded .
we operated the @xmath0-pic with a gas gain of 5000 . for this concept test
, we did not use the information on the electron energy but assumed @xmath31 as we knew the source energy .
this test has much meaning since the mev gamma - ray imaging for the radioactive sources of a know energy is useful for the medical use .
one of the well - reconstructed events is shown in fig . [
reconstructed ] .
the source is shown by the closed triangle .
the electron track is shown in the solid line and its projections are also drawn by the thin solid lines .
the measured energy by the nai ( tl ) scintillator was 260 kev and the calculated position of the gamma - ray in the nai ( tl ) scintillator is shown by the closed square .
we reconstructed the direction of the incident photon from @xmath32 , @xmath21 , @xmath22 and @xmath26 .
the reconstructed point in the z=-4.8 plane is ( -2.2 , -5.7 , -4.8 ) and is shown by the filled circle .
the reconstruction was realized by the angular accuracy of @xmath33 for this event .
the gamma - ray is well - reconstructed in this event so that we can claim that this new concept is a very promising method for the mev gamma - ray imaging .
we are going to use the energy information of the electrons , so that we can realize the mev gamma - ray imaging for the gamma - rays of unknown energies , or we can use the energy for the background rejection .
statistical studies are also under way and the angular resolution and the detection efficiency will appear in another report .
we improved the @xmath0-pic with the detection area of @xmath34 and obtained the maximum gas gain of more than @xmath4 without any other multipliers .
we detected x - rays of the rate as high as 7.7 mhz with the @xmath0-pic and the pipe - line readout electronics .
we developed a micro - tpc with the @xmath0-pic readout and took three - dimensional electron tracks .
we developed a prototype of the mev gamma - ray imaging detector with the micro - tpc and an nai ( tl ) scintillator so that we showed that this is a promising detector for the mev gamma - ray imaging .
this work is supported by a grant - in - aid in scientific research of the japan ministry of education , culture , science , sports and technology , and `` ground research announcement for space utilization '' promoted by japan space forum .
t. tanimori , a. ochi , s. minami and t. nagae , development of an imaging microstrip gas chamber with a 5 @xmath2 5 @xmath6area based on multi - chip module technology , _ nucl .
instrum . methods _ ,
a381 , no 2 - 3 , pp 280 - 288 , november 1996 .
a. ochi , t. nagayoshi , s. koishi , t. tanimori , t. nagae and m. nakamura , a new design of the gaseous imaging detector : micro pixel chamber , _ nucl . instrum . methods _
a471 , no 1 - 2 pp 264 - 267 ,
september 2001 .
h. kubo , k. miuchi , t. nagayoshi , a. ochi , r. orito , a. takada , t. tanimori , m. ueno , development of the time projection chamber with micro pixel electrodes , _ nucl .
instrum . methods _ , to be published .
t. tanimori , y. nishi , a. ochi and y. nishi , imaging gaseous detector based on micro - processing technology .
instrum . methods _ ,
a436 , no . 1 - 2 , pp .
188 - 195 , october 1999 .
u. becker , r. dinner , e. fortunato , j. kirchner , k. rosera and y. uchida , consistent measurements comparing the drift features of noble gas mixtures , _ nucl .
instrum . methods _ ,
a421 , no . 1 - 2 , pp .
54 - 59 , january 1999 . | micro - tpc , a time projection chamber ( tpc ) with micro pixel chamber ( @xmath0-pic ) readout was developed for the detection of the three - dimensional fine ( @xmath1 sub - millimeter ) tracks of charged particles .
we developed a two - dimensional position sensitive gaseous detector , or the @xmath0-pic , with the detection area of 10 @xmath2 10 @xmath3 and 65536 anode electrodes of 400 @xmath0 m pitch .
we achieved the gas gain of more than @xmath4 without any other multipliers . with the pipe - line readout system specially developed for the @xmath0-pic , we detected x - rays at the maximum rate of 7.7 mhz .
we developed a micro - tpc with the @xmath0-pic and three - dimensional tracks of electrons were detected with the micro - tpc .
we also developed a prototype of the mev gamma - ray imaging detector which is a hybrid of the micro - tpc and nai ( tl ) scintillator so that we showed that this is a promising method for the mev gamma - ray imaging .
gaseous detector ; time projection chamber ; micro - pattern detector ; gamma - ray imaging |
full - duplex ( fd ) communications is a promising technique which offers to double the spectral efficiency of radio links @xcite .
this impressive capability can response part of the explosive demands for high data - rate services . due to this and despite multiple drawbacks of fd implementation , this technique has been considered as a candidate for next generation 5 g wireless networks @xcite .
in addition , fd can eliminate hidden terminal problem in ad - hoc networks , congestion , and large end - to - end delays @xcite .
therefore , fd communication systems have attracted many attentions in research area recently @xcite-@xcite .
half - duplex ( hd ) and fd are two ways for connecting terminals in a wireless network .
traditionally , hd terminals transmit or receive either at different times or over different frequency bands .
therefore , for a simple connection , two time or frequency slots are consumed . on the other hand and
when the terminals support fd connection , transmission and reception can be accomplished simultaneously in a same time slot and frequency band .
therefore , resource consumption is reduced to the half of traditional hd links @xcite-@xcite .
clearly , in a dense environment with a large number of users , employment of fd terminals can release a high amount of resources and may lead to the significant increase in data rate of active users .
self - interference ( si ) from a transmitter to its own receiver is the evident result of using fd terminals . hence , management and reduction of si are vital for the practical implementation of fd systems @xcite-@xcite .
fortunately , recent researches show efficient techniques for si reduction which were not possible until a few years ago @xcite .
so , new research like our present paper can be considered for the practical purpose completely .
another challenging technology for next generation 5 g wireless networks is relaying which can provide lower transmit powers , higher throughput , and extensive coverage @xcite .
however , relaying suffers from low spectral efficiency .
for example , consider one - way transmission from a source to a destination together with @xmath0 parallel relays
. for complete transmission from the source and relays to the destination , @xmath1 different time slots or frequency bands are required .
this may be compared with direct transmission between the source and destination which uses only one time slot and frequency band . here , fd devices ( fd relays and fd destination ) can be organized to reduce this weakness @xcite .
therefore , fd relaying systems are of most interest among the other fd systems .
to alleviate the complexity of fd relays , some quasi fd schemes like successive relaying , two - way relaying or buffer aided relaying are investigated @xcite , @xcite-@xcite .
these techniques try to mimic fd relaying via two hd relays . in successive relaying
, two relays listen to the source and retransmit their received signal to the destination successively . in other words
, in each time slot , one of the relays receives a new data frame from the source while the other relay forwards the previous data frame to the destination .
then , the role is swapped in the next time slot . in this way
, a new data frame can be sent to the destination in each time slot as if fd relay was employed .
in this situation , while transmission and reception of two relays occur in a same time slot and frequency band , the transmitted signal from one relay can disturb the received signal by the other relay simultaneously .
therefore , inter - relay interference ( iri ) may degrade the whole performance .
so , it has to be managed well in successive relaying @xcite-@xcite .
due to high performance and simplicity of successive relaying , this paper concentrates on the power allocation of this scheme .
power allocation is widely take into account in different relay structures such as amplify and forward ( af ) @xcite-@xcite or decode and forward ( df ) @xcite-@xcite relaying and under various conditions such as hd @xcite-@xcite or fd @xcite-@xcite transmission mode .
in all these papers , the transmitted power at the source and/or relays is optimized to achieve certain target under power constraints .
several optimization targets have been adopted , such as outage probability minimization , capacity maximization , and signal to noise ratio ( snr ) maximization .
more precisely , successive relaying is also attracted attention and has been investigated from different aspects @xcite-@xcite .
for example in @xcite-@xcite , the iri suppression are studied and different schemes based on differential cancellation and network coding are proposed for two - path successive relays . in @xcite-@xcite ,
the authors have examined the throughput rate of a successive relay network which is suffered by the iri . then in @xcite , a spectrum sharing scheme for overlaid wireless networks based on the df successive relaying technique is proposed .
however , power allocation and effective capacity in successive relaying in addition to iri considerations is not investigated previously . here , we study the power allocation of af successive relays for effective capacity maximization .
the effective capacity is more general than the ergodic capacity which is defined as a maximum constant arrival rate with the delay quality - of - service ( qos ) guarantee @xcite-@xcite .
then , the allocated power is derived under short - term and long - term iri constraints and compared with different allocation techniques with different qos requirements .
iri is not always harmful , and , strong iri is more desirable in some states @xcite-@xcite .
therefore , in the following , we assume both weak and strong iri in our optimization problem separately . this paper is organized as follows : first , the system model is introduced in section [ sec:2 ] , where , successive relaying and interference management in this scheme are discussed .
in addition , channel model and the effective capacity are explained here . in section [ sec:3 ]
, we plan our optimization problem to maximize the effective capacity under different power constraints and then , the optimal allocated power is obtained in this section .
finally , the simulation results are presented in section [ sec:4 ] , and section [ sec:5 ] concludes the paper .
successive relaying behavior is shown in fig .
[ fig : successive_relaying ] , where two af relay nodes @xmath2 and @xmath3 , cooperate the source node @xmath4 , amplify its received signal and retransmit it to the destination @xmath5 successively .
the direct source to destination link is also ignored . for more details , we assume each data frame has @xmath6 symbols @xmath7 , which has to be transmitted from the source to the destination .
the specific steps for each time slot are described as follows @xcite : * in the @xmath8 time slot , @xmath4 transmits @xmath9 , @xmath2 listens to @xmath9 from @xmath4 and @xmath3 is silent ( see fig .
[ fig : successive_relaying](a ) ) . * in the @xmath10 time slot , @xmath4 transmits @xmath11 , @xmath3 listens to @xmath11 from @xmath4 while being interfered by @xmath9 from @xmath2 and @xmath2 amplifies and forwards @xmath9 to @xmath5 ( see fig . [
fig : successive_relaying](b ) ) . * in the @xmath12 time slot , @xmath4 transmits @xmath13 , @xmath2 listens to @xmath13 from @xmath4 while being interfered by @xmath11 from @xmath3 and @xmath3 amplifies and forwards @xmath11 to @xmath5 ( see fig .
[ fig : successive_relaying](c ) ) . * in the @xmath14 time slot ( @xmath15 is even ) , @xmath4 transmits @xmath16 , @xmath3 listens to @xmath16 from @xmath4 while being interfered by @xmath17 from @xmath2 and @xmath2 amplifies and forwards @xmath17 to @xmath5 ( see fig . [
fig : successive_relaying](d ) ) .
* in the @xmath18 time slot ( @xmath19 is odd ) , @xmath4 transmits @xmath20 , @xmath2 listens to @xmath20 from @xmath4 while being interfered by @xmath16 from @xmath3 and @xmath3 amplifies and forwards @xmath16 to @xmath5 ( see fig .
[ fig : successive_relaying](e ) ) . * finally , in the @xmath21 time slot ( @xmath22 is odd ) , @xmath3 amplifies and forwards @xmath23 to @xmath5 ( see fig . [
fig : successive_relaying](f ) ) .
now , we can observe that @xmath6 symbols are transmitted from the source and received completely in @xmath22 time slots by the destination .
so , the multiplexing ratio becomes @xmath24 which approaches @xmath25 for a large number of symbols @xmath6 .
this leads to the high spectral efficiency as can be obtained in fd relaying systems .
except for the first and last time slots , relays provide interference for each other .
note that , for achieving high performance of relaying systems , this iri has to be managed carefully . for proper relaying system performance , management of iri between two relays
is essential .
hence , applying interference cancellation techniques such as interference alignment ( ia ) @xcite-@xcite or successive interference cancellation ( sic ) @xcite is strongly recommended .
however for the sake of simplicity , we assume that the relays using sic for iri suppression and we explain the interference cancellation as follows : ( a ) when the iri is weak , the relay does not attempt to remove the interference .
so , the power of interference is added to the power of noise and the overall snr is reduced slightly .
( b ) on the other hand , when the iri is strong , the relay first detects the interferer signal .
then , subtracts the estimated interference from the received signal .
now , detection of the main signal without interference is possible . in this case
, there will be no snr reduction due to interference .
hence , the strong iri is not damaging in this case .
most of the time , interference is detrimental and interferer signal and symbols are discarded .
however , this subject is not always the best solution . in our successive relaying scenario ,
the interference signal can be detected and used for further improvements .
for example , according to description of [ subsec:21 ] , @xmath2 can extract just @xmath26 and @xmath3 can extract just @xmath27 when they transmitted from the source .
hence , these two relays do not have an access to all symbols completely .
but , by utilization of the iri , @xmath2 and @xmath3 can obtain @xmath27 and @xmath26 respectively .
well , using the complete set of symbols in the relays , may lead to higher performance .
therefore , the strong iri can be helpful in this scheme .
for this reason , the strong iri will be considered as one of the constraints in the optimization problem of section [ sec:3 ] . here , we assume uncorrelated quasi - static flat fading channel , where , @xmath28 and @xmath29 denote @xmath30 and @xmath31 channel coefficients and @xmath32 and @xmath33 show @xmath34 and @xmath35 channel coefficients and @xmath36 determines @xmath37 channel coefficient respectively .
the distances of @xmath30 and @xmath31 as well as @xmath34 and @xmath35 are nearly the same .
also , the average transmitted power by the source and each relay are assumed @xmath38 and @xmath39 and the power of additive white gaussian noise ( awgn ) is the same at the relays and destination .
therefore , the average snr at the @xmath30 or @xmath31 link is @xmath40 and at the @xmath34 or @xmath35 link is expressed as @xmath41 and at the @xmath37 link is written as @xmath42 .
now , assuming iri cancellation with the sic technique , the overall instantaneous snr at the destination is written as @xcite @xmath43 where at the even time slots , @xmath44 and @xmath45 and at the odd time slots @xmath46 and @xmath47 respectively . in this way
, the snr at the link between the relays can be expressed as @xmath48 . in the rayleigh fading channels , where @xmath49 , the probability density function ( pdf ) and the cumulative distribution function ( cdf ) of the snr are expressed as @xcite @xmath50\ ] ] @xmath51 where @xmath52 and @xmath53 are the first and zero - order modified bessel function of the first kind respectively ( * ? ? ? * eq .
8.407 ) .
the effective capacity is a similar concept to the capacity which investigates the effect of transmitter and receiver buffer ( queue ) delay on the capacity @xcite-@xcite . in this regard ,
the effective capacity is defined as the maximum arrival rate that a wireless channel can support , in order to guarantee the qos requirements such as the statistical delay constraint @xcite-@xcite .
consequently , the effective capacity is more realistic and can analyse the performance of the wireless channels against delay sensitive traffics more precisely . for a dynamic queuing system with stationary and ergodic arrival and service processes , under sufficient conditions , the queue length process @xmath54 converges in distribution to a random variable @xmath55 such that @xcite @xmath56 for a large @xmath57 , where @xmath58 represents the non - empty buffer probability , @xmath57 determines a certain threshold and
@xmath59 indicates the amount of required qos .
in addition , for the delay of a packet in the buffer as the main qos metric , we have a similar probability function as @xcite @xmath60 where @xmath61 indicates the tolerated delay , @xmath62 is a delay - bound , and @xmath63 is jointly determined by both arrival and service processes . the statistical delay constraint in
represents the qos which has to be guaranteed for the delay sensitive traffic sources .
it is apparent that the qos exponent @xmath59 has an important role here .
larger @xmath59 corresponds to more strict qos constraint , while smaller @xmath59 implies looser qos requirements .
effective capacity provides the maximum constant arrival rate that can be supported by the time - varying wireless channel under the statistical delay constraint .
since the average arrival rate is equal to the average departure rate when the buffer is in a steady - state @xcite , the effective capacity can be viewed as the maximum throughput in the presence of such a constraint . with all these explanations ,
the effective capacity is defined as @xcite-@xcite @xmath64 where @xmath65 is the time - varying rate of the channel and @xmath66 denotes the statistical expectation . for a specific application with a given statistical delay requirement
, the qos exponent @xmath59 can be determined from .
then , the maximum constant arrival rate of the sources that a wireless channel can support in order to guarantee the given qos is concluded from .
the power allocation scheme and the effective capacity of af successive relaying are discussed in the next section .
in order to maximize the effective capacity , the allocation power scheme at the relay is extracted here . as we explained before ,
high performance of successive relaying is obtained when the iri is considered and managed well .
so , we will assume weak or strong and short - term or long - term iri constraints in our optimization problem .
note that , weak constraint means the iri is lower than the specific threshold @xmath67 and subsequently , strong constraint means the iri is higher than the specific threshold @xmath67 .
besides , short - term constraint means the instantaneous iri and long - term constraint means the average iri respectively . with power allocation strategy , we assume that the relays can adapt their transmitted power to maximize the effective capacity . in this regard ,
the relays transmit @xmath68 instead of @xmath39 , where @xmath69 depicts the time - varying allocated power coefficient and we have @xmath70 for constant average transmitted power .
now , with this policy , the instantaneous snr at the destination is changed to @xmath71 and the instantaneous capacity is written as @xmath72 where @xmath73 specify the total required bandwidth and @xmath74 denotes one symbol duration or one transmission time slot in fig .
[ fig : successive_relaying ] .
since @xmath6 symbols are transmitted from the source to the destination in @xmath22 time slots , the instantaneous capacity is written as . by substituting into
, the effective capacity of af successive relays is obtained as @xmath75 where @xmath76 . as we explained before
, the iri has a significant influence in the performance of successive relays and this interference has to be managed properly . when the weak iri is desirable , we can adjust the transmitted power of the relays to obtain the short - term or long - term interference below some specific threshold @xmath67 .
consequently , we use @xmath77 as the short - term constraint or @xmath78 for the long - term constraint as well . as the same way ,
when we need the strong iri , we can adapt the transmitted power from the relays to obtain the short - term or long - term interference above the threshold @xmath67 .
therefore , we would have @xmath79 as the short - term constraint or @xmath80 for the long - term constraint , respectively .
now , we can summarize our optimization problem as @xmath81 @xmath82 @xmath83 [ eq:13 ] @xmath84 where and show the positive and constant average transmitted power constraints and depicts the short - term and long - term iri constraints respectively ( shows the weak and short - term constraint , shows the weak and long - term constraint , shows the strong and short - term constraint and finally shows the strong and long - term constraint ) . using the lagrangian optimization method @xcite , the problem can be solved and @xmath85 is derived .
however , applying the lagrange method in the optimization problem of , a high non - linear equation of @xmath85 is obtained and extraction of a closed - form solution for @xmath85 is not possible @xcite .
therefore , we continue with a novel approximation and replace @xmath86 at the denominator of by its average @xmath70 .
then , we can rewrite as @xmath87 and simplify it to @xmath88 where @xmath89 is introduced in . the constraints of the optimization remain unchanged .
now , we can obtain a closed - form solution for the allocated power in the following subsections .
now , we can use , and for the weak and long - term iri constraints . here
, the optimal power allocation is determined as @xmath90 where @xmath91 is a cut - off snr threshold which is calculated from @xmath92 against , for the strong and long - term iri constraints , we use , and and obtain @xmath93 for the @xmath94 .
note that , when @xmath95 , the power allocation with all constraints , and is impossible .
the cut - off snr threshold @xmath91 is calculatde from @xmath70 ( the proof of has been derived in appendix [ appx:1 ] and the proof of is similar to appendix [ appx:1 ] ) .
finally , we can replace or into to obtain the effective capacity of af successive relays with the weak or strong long - term iri constraints as follow ( for the proof , please see appendix [ appx:2 ] ) @xmath96\right.\nonumber\\ & -&\frac{\sqrt{\pi}\gamma_{\textrm{t}}}{\bar{\gamma}}\left[g_{23}^{30}\left(\frac{4\gamma_{\textrm{t}}}{\bar{\gamma}}\left|\begin{array}{l}0,3/2\\-1,2,0\end{array } \right.\right ) + g_{23}^{30}\left(\frac{4\gamma_{\textrm{t}}}{\bar{\gamma}}\left|\begin{array}{l}0,3/2\\-1,1,1\end{array } \right.\right)\right]\nonumber\\ & + & \left.\frac{\sqrt{\pi}}{4}\left(\frac{4\gamma_{\textrm{t}}}{\bar{\gamma}}\right)^{\frac{1 + 2\tilde{\theta}}{1+\tilde{\theta}}}\left[g_{23}^{30}\left(\frac{4\gamma_{\textrm{t}}}{\bar{\gamma}}\left|\begin{array}{l}0,\frac{1}{2}+\frac{1}{1+\tilde{\theta}}\\-1,1+\frac{1}{1+\tilde{\theta}},-1+\frac{1}{1+\tilde{\theta}}\end{array } \right.\right)\right.\right.\nonumber\\ & + & \left.\left.g_{23}^{30}\left(\frac{4\gamma_{\textrm{t}}}{\bar{\gamma}}\left|\begin{array}{l}0,\frac{1}{2}+\frac{1}{1+\tilde{\theta}}\\-1,\frac{1}{1+\tilde{\theta}},\frac{1}{1+\tilde{\theta}}\end{array } \right.\right)\right]\right\}\end{aligned}\ ] ] where @xmath97 represents gauss hypergeometric function ( * ? ? ?
9.10 ) , @xmath98 denotes the gamma function and @xmath99 is the meijer s g function defined in ( * ? ? ?
9.301 ) which is easily evaluated using the popular numerical softwares .
when the weak and short - term iri is required , the constraints of , and can be used and the optimal allocated power is obtained as @xmath100 on the other hand , with the strong and short - term iri assumption , the constraints of , and can be used to calculate the optimal allocated power as @xmath101 here , @xmath91 is a cut - off snr threshold which is determined from the average transmitted power , @xmath70 , and @xmath102 is . ]
@xmath103 finally , or can be replaced into and the effective capacity of af successive relays with the weak or strong short - term iri constraints is obtained . here , finding a closed - form solution for the effective capacity is not straightforward .
for the simulation we assume @xmath104 , @xmath105 and @xmath106 and for more simple drawing , the normalized effective capacity @xmath107 is plotted in the following figures .
the effective capacity of af successive relays with the weak and long - term iri constraints ( see eq . ) is plotted in fig .
[ fig:2 ] when @xmath108 with @xmath109 , @xmath110 and @xmath111 and also in fig .
[ fig:3 ] when we have @xmath112 and @xmath109 , @xmath113 and @xmath114 respectively .
we can see the tight agreement between theory and simulation results in these figures which verifies the derived closed - form solution of the effective capacity in section [ subsec:31 ] .
in addition , as we expected , the effective capacity increases when the threshold @xmath67 increases .
note that , when @xmath67 increases , the amount of tolerable interference between relays increases as well and therefore , the effective capacity of af successive relay is increased .
with @xmath109 , @xmath110 and @xmath111 here . ] with @xmath109 , @xmath113 and @xmath114 here . ] in a similar way , the effective capacity of af successive relays with the strong and long - term iri constraints ( see eq . ) is plotted in fig .
[ fig:4 ] .
note that , the power allocation is not possible when @xmath115 , because the strong constrain is not realized in this situation .
full agreement between theory and simulation is also clear in fig .
[ fig:4 ] . with @xmath116 and @xmath112 with @xmath117 here . ] after that , the effective capacity with the weak but short - term iri constraints ( see eq . ) is plotted in fig .
[ fig:5 ] for @xmath108 and fig .
[ fig:6 ] for @xmath112 for different values of @xmath67 .
we can observe that the effective capacity increases when the amount of acceptable iri is increased . without strict constraint on the interference ,
the relay can transmit more power temporally to improve the total effective capacity or requested qos .
now , we can compare the obtained results in fig .
[ fig:2 ] and [ fig:3 ] with the results of fig .
[ fig:5 ] and [ fig:6 ] accurately .
generally , the effective capacity with the long - term constraint outperforms the results with the short - term constraint .
however , despite the obtained results , short - term constraint is necessary in some sensitive applications and can not be ignored ever .
in addition , when @xmath59 increases and high qos is required , the effective capacity of successive af relay with short - term iri constraint drops rapidly . therefore , we can conclude that the performance with the high qos and short - term constraint is not suitable at all , and long - term constraint for management of interference is recommended for this situation properly . in a similar way , the effective capacity with the strong and short - term iri constraints ( see eq . ) , can be drawn and compared with fig .
[ fig:4 ] .
however to avoid duplication , this figure is not plotted here . and @xmath118 , @xmath119 , @xmath113 and @xmath120 .
] and @xmath120 , @xmath121 , @xmath122 and @xmath123 . ] for more details , in fig .
[ fig:7 ] , we have drawn the effective capacity of af successive relays with the optimal power allocation , constant power allocation @xmath124 and truncated channel inversion with @xmath125 where @xmath91 is obtained from @xmath126 . in constant power allocation technique , there is no constraint on the iri .
so , the performance at low value of @xmath59 with loose qos , tends to the performance with the optimal power allocation .
however , at high value of @xmath59 , the performance with the optimal power allocation , outperforms the other techniques of transmission .
in addition , when truncated channel inversion technique is used , the iri can remain constant but the effective capacity performance is very poor .
therefore , the optimal power allocation ( similar to ) at the relay is strongly recommended specially when high qos is required . and @xmath120 .
] finally , the effective capacity of successive af relying scheme is compared with the traditional hd relaying in fig .
[ fig:8 ] versus qos exponent @xmath59 . for the successive relays we assume the weak and long - term constraint and optimal power allocation with @xmath109 or @xmath113 or @xmath120 . on the other hand , in the hd mode
, there is no iri and therefore , the optimal power allocation coefficient @xmath86 can be obtained from when @xmath67 tends to very small value ( i.e. @xmath127 ) .
as we expected , we can see that successive relaying outperforms the traditional hd relaying especially when we manage the iri with high value of threshold @xmath67 ( i.e. @xmath120 ) .
when the threshold @xmath67 decreases and more strict constraint on the iri is required , the effective capacity of successive relays decreases and approaches the effective capacity of hd relaying .
however at low value of @xmath67 when the iri is not tolerable , using hd relaying is more efficient than the successive relays . .
in this paper , power allocation at the successive relay is studied . successive relaying
can be considered as a simple form of full - duplex relay implementation . in the successive relaying technique , two half - duplex relays with successive retransmission , act like a full - duplex relay .
however , inter - relay interference management is crucial for proper operation of this technique . here , the power allocation in af successive relays for the effective capacity maximization was the goal .
so , first , the power allocation coefficient with the long - term inter - relay interference constraint is calculated and a closed - form solution for the effective capacity is derived .
then , the power allocation with the short - term constraint is extracted .
finally , the effective capacity with the different type of power allocation is compared and we observed that in high qos and low latency requirement , interference management with long - term constraint have a better performance and effective capacity results .
first , we can write the optimization problem in a standard form as @xmath128 then , using lagrangian method , the cost function is written as @xmath129 where @xmath130 , @xmath131 and @xmath132 are the nonnegative lagrange multipliers corresponding to our constraints . taking the partials with respect to @xmath86
, we will have @xcite @xmath133 now , we can find @xmath85 , @xmath134 , @xmath135 and @xmath136 such that @xmath137 and @xmath138 @xmath139 and @xmath140 correspondingly . from , since @xmath141 is not acceptable , we can conclude that @xmath142 . consider and , we can break the analysis into four different cases as follows . 1 . if @xmath143 and @xmath144 , then converts to @xmath145 .
since we assume @xmath146 , @xmath147 , @xmath148 and @xmath149 , this case is not a feasible solution .
if @xmath150 and @xmath151 , then converts to @xmath152 . here , we can assume @xmath153 and @xmath147 which leads to @xmath154 note that , @xmath135 can be calculated from @xmath155 or equivalently @xmath156 .
since we assumed @xmath157 , then we have @xmath158 .
therefore , this case is valid when @xmath159 .
if @xmath160 and @xmath161 , then converts to @xmath162 . here , we can assume @xmath163 and @xmath147 which leads to @xmath164 note that , @xmath134 can be calculated from @xmath165 .
since we assumed @xmath166 , then we have @xmath167 .
therefore , this case is valid when @xmath168 .
if @xmath161 and @xmath169 , then converts to @xmath170 . here , we can assume @xmath171 and @xmath147 which leads to @xmath172 note that , @xmath173 can be calculated from both @xmath165 or @xmath155 with the same value .
therefore , this case is valid when @xmath174 . by substituting or into
, we will have @xmath177 where @xmath178 is presented in . for solving @xmath179 ,
it can be changed to two infinite integrals as @xmath180\nonumber\\ & -&\frac{\sqrt{\pi}\gamma_{\textrm{t}}}{\bar{\gamma}}\left[g_{23}^{30}\left(\frac{4\gamma_{\textrm{t}}}{\bar{\gamma}}\left|\begin{array}{l}0,3/2\\-1,2,0\end{array } \right.\right ) + g_{23}^{30}\left(\frac{4\gamma_{\textrm{t}}}{\bar{\gamma}}\left|\begin{array}{l}0,3/2\\-1,1,1\end{array } \right.\right)\right]\end{aligned}\ ] ] where the closed - form solution of @xmath181 and @xmath182 is possible using ( * ? ? ?
6.621 - 3 ) and ( * ? ?
* eq . 6.625 - 7 ) , respectively . once again
* eq . 6.625 - 7 ) can be used for finding @xmath183 .
therefore , we have @xmath184\end{aligned}\ ] ] where @xmath97 represents gauss hypergeometric function ( * ? ? ?
9.10 ) , @xmath98 denotes the gamma function and @xmath99 is the meijer s g function defined in ( * ? ? ?
now , connecting the obtained results for @xmath179 and @xmath183 , we have @xmath185\right.\nonumber\\ & -&\frac{\sqrt{\pi}\gamma_{\textrm{t}}}{\bar{\gamma}}\left[g_{23}^{30}\left(\frac{4\gamma_{\textrm{t}}}{\bar{\gamma}}\left|\begin{array}{l}0,3/2\\-1,2,0\end{array } \right.\right ) + g_{23}^{30}\left(\frac{4\gamma_{\textrm{t}}}{\bar{\gamma}}\left|\begin{array}{l}0,3/2\\-1,1,1\end{array } \right.\right)\right]\nonumber\\ & + & \left.\frac{\sqrt{\pi}}{4}\left(\frac{4\gamma_{\textrm{t}}}{\bar{\gamma}}\right)^{\frac{1 + 2\tilde{\theta}}{1+\tilde{\theta}}}\left[g_{23}^{30}\left(\frac{4\gamma_{\textrm{t}}}{\bar{\gamma}}\left|\begin{array}{l}0,\frac{1}{2}+\frac{1}{1+\tilde{\theta}}\\-1,1+\frac{1}{1+\tilde{\theta}},-1+\frac{1}{1+\tilde{\theta}}\end{array } \right.\right)\right.\right.\nonumber\\ & + & \left.\left.g_{23}^{30}\left(\frac{4\gamma_{\textrm{t}}}{\bar{\gamma}}\left|\begin{array}{l}0,\frac{1}{2}+\frac{1}{1+\tilde{\theta}}\\-1,\frac{1}{1+\tilde{\theta}},\frac{1}{1+\tilde{\theta}}\end{array } \right.\right)\right]\right\}.\end{aligned}\ ] ] 1 m. pashazadeh and f. s. tabataba , impact of loop - back interference and channel estimation errors on full - duplex relay networks , " _ springer wireless netw .
_ , doi : 10.1007/s11276 - 016 - 1205 - 3 , feb . 2016
. z. zhang , x. chai , k. long , a. v. vasilakos and l. hanzo , full duplex techniques for 5 g networks : self - interference cancellation , protocol design , and relay selection , " _ ieee commun . mag .
5 , pp . 2 - 10 ,
g. liu , f. r. yu , h. ji and v. c. m. leung , in - band full - duplex relaying : a survey , research issues and challenges , " _ ieee commun .
surveys tuts .
2 , pp . 500 - 524 , @xmath10 quarter 2015 .
a. sabharwal , p. schniter , d. guo , d. w. bliss , s. rangarajan and r. wichman , in - band full - duplex wireless : challenges and opportunities , " _ ieee j. sel .
areas commun .
1637 - 1652 , sept . 2014 .
d. l. prez , x. chu , a. v. vasilakos and h. claussen , on distributed and coordinated resource allocation for interference mitigation in self - organizing lte networks , " _ ieee / acm trans .
1145 - 1158 , aug .
d. l. prez , x. chu , a. v. vasilakos and h. claussen , power minimization based resource allocation for interference mitigation in ofdma femtocell networks , " _ ieee j. sel .
areas commun .
333 - 344 , aug . 2014 .
m. duarte , full - duplex wireless : design , implementation and characterization , " ph.d .
dissertation , rice university , houston , tx , usa , apr . 2012 .
t. riihonen , s. werner and r. wichman , mitigation of loopback self - interference in full - duplex mimo relays , " _ ieee trans .
signal process _ , vol .
5983 - 5993 , dec . 2011 .
l. j. rodrguez , n. h. tran , and t. le - ngoc , optimal power allocation and capacity of full - duplex af relaying under residual self - interference , " _ ieee wireless commun .
233 - 236 , apr . 2014 .
z. ding , i.krikidis , b. rong , j. thompson , c. wang and s. yang , on combating the half - duplex constraint in modern cooperative networks : protocols and techniques , " _ ieee wireless commun . _ , vol . 19 , no . 6 , pp .
20 - 27 , dec . 2012 .
s. zhang , q. f. zhou , c. kai and w. zhang , full diversity physical - layer network coding in two - way relay channels with multiple antennas , " _ ieee trans .
wireless commun .
13 , no . 8 , pp .
4273 - 4282 , aug .
y. fan , c. wang , j. thompson and h. v. poor , recovering multiplexing loss through successive relaying using repetition coding , " _ ieee trans .
wireless commun .
_ , vol . 6 , no . 12 , pp . 4484 - 4493 , dec . 2007 .
c. wang , y. fan , j. thompson and h. v. poor , a comprehensive study of repetition - coded protocols in multi - user multi - relay networks , " _ ieee trans . wireless commun .
_ , vol . 8 , no . 8 , pp . 4329 - 4339 , aug . 2009 .
y. zhao , r. adve and t. j. lim , improving amplify - and - forward relay networks : optimal power allocation versus selection , " _ ieee trans .
wireless commun .
_ , vol . 6 , no . 8 , pp
. 3114 - 3123 , aug .
l. j. rodrguez , n. h. tran , a. helmy and t. le - ngoc , optimal power adaptation for cooperative af relaying with channel side information , " _ ieee trans . veh .
62 , no . 7 , pp . 3164 - 3174 , sept .
h. kim , s. lim , h. wang and d. hong , optimal power allocation and outage analysis for cognitive full duplex relay systems , " _ ieee trans .
wireless commun .
3754 - 3795 , oct .
m. j. emadi , a. g. davoodi and m. r. aref , analytical power allocation for a full - duplex decodeand - forward relay channel , " _ iet commun .
_ , vol . 7 , no . 13 , pp
. 1338 - 1347 , sept .
b. yu , l. yang , x. cheng and r. cao , power and location optimization for full - duplex decode - and - forward relaying , " _ ieee trans .
4743 - 4753 , dec . 2015 .
g. farhadi and n. c. beaulieu , power - optimized amplify - and - forward multi - hop relaying systems , " _ ieee trans .
wireless commun .
_ , vol . 8 , no .
9 , pp . 4634 - 4643 , sept .
m. mohammadi , p. sadeghi and m. ardebilipour , node and symbol power allocation in time - varying amplify - and - forward dual - hop relay channels , " _ ieee trans .
432 - 439 , jan .
y. hua , an overview of beamforming and power allocation for mimo relays , " in _ milcom _ , san jose , ca , usa , nov .
2010 , pp .
375 - 380 .
t. riihonen , s. werner and r. wichman , hybrid full - duplex / half - duplex relaying with transmit power adaptation , " _ ieee trans .
wireless commun .
9 , pp . 3074 - 3085 , sept .
l. li , l. wang and l. hanzo , differential interference suppression aided three - stage concatenated successive relaying , " _ ieee trans .
60 , no . 8 , pp . 2146 - 2155 , aug .
h. lu , p. hong and k. xue , generalized interrelay interference cancelation for two - path successive relaying systems , " _ ieee trans
4113 - 4118 , oct .
y. ji , c. han , a. wang and h. shi , partial inter - relay interference cancellation in two path successive relay network , " _ ieee commun .
451 - 454 , mar .
r. zhang , on achievable rates of two - path successive relaying , " _ ieee trans .
2914 - 2917 , oct .
2009 . s. gupta , r. zhang and l. hanzo , throughput maximization for a buffer - aided successive relaying network employing energy harvesting , " _ ieee trans .
_ , vol . 65 , no . 8 , pp . 6758 - 6765 , aug
2016 . c. zhai , w. zhang and p. c. ching , cooperative spectrum sharing based on two - path successive relaying , " _ ieee trans .
61 , no . 6 , pp . 2260 - 2270 , jun .
l. li , h. v. poor and l. hanzo non - coherent successive relaying and cooperation : principles , designs and applications , " _ ieee commun . surveys tuts .
1708 - 1737 , @xmath12 quarter 2015 .
m. lari , a. mohammadi , a. abdipour and i. lee , characterization of effective capacity in af relay systems , " _ ieice electron . express _
, vol . 9 , no . 7 , pp . 679 - 684 , apr .
m. lari , a. mohammadi , a. abdipour and i. lee , characterization of effective capacity in antenna selection mimo systems , " _ j. commun .
476 - 485 , oct .
m. lari , effective capacity of receive antenna selection mimo - ostbc systems in co - channel interference , " _
j. wireless netw .
_ , doi:10.1007/s11276 - 016 - 1219-x , feb
m. lari , a. mohammadi , a. abdipour and i. lee , effective capacity in receive antenna selection and spatially correlated mimo - ostbc systems , " in _
telecommun . ,
( ist12 ) _ , tehran , iran , nov .
2012 , pp .
117 - 122 .
n. zhao , f. r. yu , h. sun and m. li , adaptive power allocation schemes for spectrum sharing in interference - alignment - based cognitive radio networks , " _ ieee trans . veh .
5 , pp . 3700 - 3714 , may .
x. li , n. zhao , y. sun and f. r. yu , interference alignment based on antenna selection with imperfect channel state information in cognitive radio networks , " _ ieee trans .
65 , no . 7 , pp . 5497 - 5511 , jul . 2016 .
p. l. yeoh , m. elkashlan and i. b. collings , selection relaying with transmit beamforming : a comparison of fixed and variable gain relaying , " _ ieee trans .
59 , no . 6 , pp .
1720 - 1730 , jun . 2011 .
i. s. gradshteyn and i. m. ryzhik , _ table of integrals , series , and products _ , @xmath186 ed . , academic press , 2007 .
d. wu and r. negi , effective capacity : a wireless link model for support of quality of service , " _ ieee trans .
wireless commun .
_ , vol . 2 no .
630 - 643 , jul . 2003 .
b. soret , c. a. torres and j. t. entrambasaguas ,
capacity with explicit delay guarantees for generic source over correlated rayleigh channel , " _ ieee trans .
wireless commun .
_ , vol . 9 , no . 6 , pp
1901 - 1911 , jun . 2010 .
x. kang , y. c. liang and a. nallanathan , optimal power allocation for fading channels in cognitive radio networks under transmit and interference power constraints , " in _ 2008 ieee int .
( icc08 ) _ , beijing , china , may .
2008 , pp .
3568 - 3572 . | in the relay based telecommunications with @xmath0 relays between the source and destination , @xmath1 time or frequency slots are required for a single frame transmission . however , without the relays , only one time or frequency slot is used for a single frame transmission . therefore , despite the benefits of relaying systems , this type of communications is not efficient from the spectral efficiency viewpoint .
one solution to reduce this issue might be the full - duplex ( fd ) relays .
an old technique which is reconsidered recently to improve the spectral efficiency of telecommunication systems .
however , fd relays have a certain complexity , so , some similar techniques such as successive relays with nearly the same performance but less complexity is taken into account now .
in successive relaying systems , two relays between the source and destination are employed which receive the transmitted frames from the source and relay it to the destination successively .
this structure generally acts like an fd relays . in this paper , the effective capacity performance of an amplify and forward ( af ) successive relaying systems with power allocation strategy at the relays
are studied perfectly . however , while the inter - rely interference ( iri ) between two successive relays has to be managed well , the power allocation and the effective capacity is derived under different assumptions about the iri . in this way
, we assume weak or strong , short or long - term constraints on the iri .
then we extract the optimal transmitted power at the relay to maximize the effective capacity under these constraints .
example.eps gsave newpath 20 20 moveto 20 220 lineto 220 220 lineto 220 20 lineto closepath 2 setlinewidth gsave .4 setgray fill grestore stroke grestore |
spin and orbital mechanisms , which are present simultaneously in various physical phenomena , result in two competitive contributions in the observable effects .
textbook examples therefore are pauli paramagnetism and landau diamagnetism that yield two comparable contributions to the magnetic susceptibility of an electron gas .
@xcite another bright manifestation is the fine structure of exciton lines in a magnetic field .
@xcite in all these cases , the electron systems are affected by a magnetic field in two different ways : via zeeman splitting of spin sublevels and due to cyclotron twisting of electron trajectories .
an interplay of spin and orbital mechanisms is also expected for linear and circular magneto - gyrotropic photogalvanic effects ( mpge ) .
the spin - related mechanisms of both mpges were widely discussed in the past and are driven by the spin - dependent relaxation of a non - equilibrium electron gas in gyrotropic two - dimensional electron systems ( 2des ) ( for reviews see refs . ) .
the microscopic mechanisms of the spin - driven mpges are based on spin - orbit coupling in 2des with structure and bulk inversion asymmetry ( sia and bia ) . in the case of the linear mpge
the absorption of linearly polarized or unpolarized radiation leads to electron gas heating . due to a spin - dependent energy relaxation of these heated electrons ,
two equal and oppositely directed electron flows in the spin - up and spin - down subbands result , what represents a pure spin current .
@xcite the zeeman splitting of the subbands induced by the in - plane magnetic field leads to the conversion of a spin flow into a measurable spin - polarized electric current .
the circular mpge yields a photocurrent , proportional to the radiation helicity .
it is caused by the spin - galvanic effect ( sge ) , @xcite in which the spin - flip relaxation of a spin - polarized non - equilibrium electron gas results in an electric current .
@xcite here , an in - plane magnetic field rotates the optically induced spin polarization into the plane of the 2des due to the larmor precession .
this provides an in - plane spin component , which is necessary for the sge in ( 001)-grown qws .
while the spin - based origin of the mpge was intensively discussed , most recently it was pointed out that orbital effects , caused by a magnetic - field - induced scattering asymmetry , may substantially contribute to the mpge @xcite and thus complicate the analysis of the spin currents .
the latter is of importance for the exploration of spin generation and spin - related transport in 2des , which are major and still growing fields in solid - state research .
@xcite here we report on experiments , which allow us to distinguish unambiguously between the spin - dependent and orbital origin of the mpge by investigating the qualitative difference in their behavior upon a variation of the @xmath0 factor .
we use the fact that the electric current resulting from the spin - related roots is proportional to the zeeman band spin splitting in contrast to the one due to orbital mechanisms . to explore this difference
, we utilize the well known fact that in gaas based quantum wells the zeeman splitting changes its sign at a certain qw width .
@xcite this inversion is mostly caused by the opposite signs of the @xmath0 factor in the gaas with respect to the algaas - barrier and the fact that for narrow qws the electron wave function deeply penetrates into the barrier .
therefore , we studied the mpge in a set of structures with qw widths varying from 4 to 30 nm .
the paper is organized as follows . in sec .
[ technique ] we describe details of the sample preparation and present an overview of the experimental techniques used to study the mpge . in sec .
[ kerr ] we give an account on combined photoluminescence ( pl ) and pl excitation ( ple ) measurements as well as time - resolved kerr rotation ( trkr ) experiments performed on the sample series in order to determine the electron confinement energies and the electron @xmath0 factor , respectively .
sections [ lmpge ] and [ cmpge ] deal with the experimental study and analysis of the linear and circular mpge , respectively .
finally , we summarize the study .
the experiments were carried out on ( 001)-oriented si-@xmath1-doped @xmath2-type gaas@xmath3al@xmath4ga@xmath5as structures grown by molecular - beam epitaxy at typical temperatures in excess of 600@xmath6c .
a set of structures with the same doping profile but different qw widths @xmath7 , between 4 nm and 30 nm , was grown .
figure [ fig1 ] sketches the conduction band edge of the multiple qw structures together with the corresponding @xmath1-doping positions , and table [ samples ] gives the sample parameters .
the doping layers are asymmetrically shifted by two different spacer layer thicknesses , @xmath8 and @xmath9 , with respect to the qw .
in addition , in all samples , except a , the sum @xmath10 was kept constant .
the impurities coulomb field yields an asymmetric potential profile inside the qws causing sia .
the investigated samples were square shaped ( @xmath11 ) and its edges oriented along the @xmath12 $ ] and @xmath13 $ ] crystallographic directions . for photocurrent measurements , ohmic contacts were alloyed on the sample corners allowing to probe the photocurrent along the @xmath14 $ ] and @xmath15 $ ] crystal axes .
the mpge is measured at room temperature by exciting the samples with linearly or circularly polarized terahertz ( thz ) laser radiation under normal incidence . as radiation source we used an optically pumped nh@xmath16 molecular laser @xcite operating at a wavelength @xmath17 of @xmath18 m .
the wavelength of @xmath18 m ( photon energy of 4.2 mev ) was chosen to cause free carrier absorption only .
the laser provided single pulses with a pulse duration of about 100 ns and a peak power @xmath19 of about @xmath20 kw . the photocurrent @xmath21 is measured via the voltage drop across a 50 @xmath22 load resistor with a digital oscilloscope . to study the linear mpge we applied linearly polarized radiation with the radiation field @xmath23 , and an external in - plane magnetic field of @xmath24 of @xmath25 t.
the resulting current is measured along the @xmath26-direction , normally to the magnetic field , @xmath24 . to excite the circular mpge we used
elliptically polarized radiation obtained by a crystal quartz quarter - wave plate .
the helicity of the radiation was varied by the rotation of the plate according to @xmath27 , where @xmath28 is the angle between the initial plane of polarization and the optical axis of the @xmath29-plate . in this case
the photocurrent @xmath30 is measured in the direction parallel to @xmath31 .
the experimental geometries with @xmath32 for linear and @xmath33 for the circular mpge are chosen to probe the currents solely caused by the structure inversion asymmetry .
@xcite this allows us to reduce the influence of the qw width on the degree of asymmetry .
for any other configuration we would also obtain a bia - induced mpge , which complicates the data analysis , because bia itself strongly depends on @xmath7 .
to study land factors and spin dynamics we used the time - resolved kerr rotation and photoluminescence techniques , applying a pulsed ti - sapphire laser system .
.sample parameters .
carrier density @xmath34 ( per qw - layer ) , mobility @xmath35 , momentum scattering time @xmath36 calculated from the mobility , measured @xmath0 factor and energy difference @xmath37 between the transitions ( e2hh2)-(e1hh1 ) are given for @xmath38 k. also listed : carrier density @xmath34(rt ) ( per qw - layer ) and mobility @xmath35(rt ) measured at room temperature . [ cols="^,^,^,^,^,^,^,^,^,^,^,^,^ " , ] [ samples ] it generates pulses with a spectral linewidth of about 3 - 4 mev and a duration of about 600 fs . the central wavelength of the laser is tuned to excite electron - hole pairs at the fermi energy of the two - dimensional electron gas in the samples .
the pulse train from the laser system is split into the pump beam , which is circularly polarized , and the weaker , linearly polarized probe beam .
both beams are focused onto the sample surface at near - normal incidence by an achromat , and the kerr rotation angle of the reflected probe beam is measured using an optical bridge detector .
the typical power density used in the experiments was about 40 w/@xmath39 for the pump beam , and 4 w/@xmath39 for the probe beam . a lock - in scheme
is used to increase the sensitivity of the experiment .
the samples were mounted in the @xmath40he insert of an optical cryostat and cooled by @xmath40he gas to a nominal temperature of 4.5 k. magnetic fields of up to 10 tesla were applied in the sample plane during the trkr measurements . for photoluminescence measurements ,
the ti - sapphire laser operated in cw mode and was tuned to higher energies to excite the samples nonresonantly .
the excitation density for pl measurements was about 4 w/@xmath39 .
the pl emitted from the samples was collected using a grating spectrometer and a charge - coupled device ( ccd ) sensor .
all pl measurements were performed without applied magnetic fields at a nominal sample temperature of 4.5 k. for ple measurements , a tunable cw ti - sapphire laser system was used .
the pl intensity of the low - energetic flank of the sample was recorded using the spectrometer and the ccd sensor as a function of the laser energy of the ti - sapphire laser .
factor changes its sign .
@xcite ( c ) energy difference between ( e1-hh1 ) and ( e2-hh2 ) transitions as a function of the qw width.,scaledwidth=45.0% ] to determine the electron confinement energies in the qws , we performed combined pl and ple measurements on the sample series . figure [ pl](a ) shows pl and ple spectra measured on sample e ( 15 nm wide qw ) .
the pl spectrum shows a single peak corresponding to the ( e1-hh1 ) transition in the qw . in the ple spectrum ,
the onset of absorption at the fermi energy of the two - dimensional electron system in the qw is clearly visible slightly above this pl peak .
the pronounced maximum in the ple spectrum , indicated by the arrow , corresponds to the ( e2-hh2 ) transition . due to the increasing confinement in the narrower qws , the pl peak position shifts to higher energies , as fig .
[ pl](b ) demonstrates . the energy difference between ( e1-hh1 ) and ( e2-hh2 ) transitions as a function of the qw width , shown in fig .
[ pl](c ) , and listed in table [ samples ] , is extracted from the pl and ple measurements .
it also increases as the qw width is reduced . for qws , which are thinner than 10 nm ,
the ( e2-hh2 ) transition lies outside of the tuning range of the ti - sapphire laser .
trkr measurements are used to determine the electron @xmath0 factor in our sample series .
it is well - established @xcite that in gaas / algaas qws the electron @xmath0 factor depends on the qw width , and even changes its sign from negative to positive values as a function of the qw thickness .
two effects contribute to this dependence : the conduction band of gaas is nonparabolic , and the @xmath0 factor of an electron depends on its energy relative to the conduction band edge .
@xcite additionally , in narrow qws , the electron wave function has a sizeable amplitude within the algaas barriers , leading to an effective admixture of the ( positive ) @xmath0 factor in the barrier material to that ( negative ) of the electron confined in the gaas qw . in our trkr measurements , we clearly observe this width dependence [ see fig .
[ kerr_gfactor](a ) ] .
the trkr traces for different samples , measured at the same magnetic field of 4 t , show the damped larmor precession of the optically oriented electron spin polarization about the applied magnetic field .
the precession frequencies are markedly different : 4 , 8 and 30 nm ) with an applied in - plane magnetic field of 4 tesla . ( b )
electron @xmath0 factor as a function of the qw width extracted from the trkr data .
the sign of the @xmath0 factor , which can not be directly extracted from the trkr traces , has been inferred from the pl transition energies.,scaledwidth=45.0% ] for the widest sample , the precession frequency is high , it decreases for narrower qws , then increases again for the most narrow sample .
the larmor precession frequencies were determined for all samples by measuring trkr traces in magnetic fields between 1 and 10 tesla , and the amplitude of the @xmath0 factor was calculated using a linear fit to the dispersion .
figure [ kerr_gfactor](b ) gives the qw width dependence of the @xmath0 factors in our samples . from trkr measurements , the sign of the @xmath0 factor is not immediately apparent , although it can be determined using slight modifications of a typical trkr setup and rigorous data analysis .
@xcite here , we infer the sign of the @xmath0 factor from the pl transition energies observed for our samples . in their studies on the @xmath0 factor s dependence on the qw width , yugova et al .
@xcite observed the zero crossing of the @xmath0 factor for a fundamental transition energy ( e1-hh1 ) of 1600 mev ( indicated by the dotted line in fig .
[ pl](b ) ) . in sample c , we observe the pl peak at 1572 mev , well below the value for zero crossing , and in the thinner sample b the pl peak is at 1616 mev , well above the value for zero crossing .
hence , we assign negative @xmath0 factor values to the samples with nominal qw widths between 30 nm and 8.2 nm , and positive @xmath0 factors to the two thinnest samples .
the trkr measurements also allow us to study the dependence of the spin dephasing time ( sdt ) , @xmath41 , on the qw width .
figure [ trkr_sdt_omega](a ) shows a series of trkr traces measured on different samples without applied magnetic fields . no simple correlation between the qw width and the sdt is apparent , as evidenced by the sdt data given in fig .
[ trkr_sdt_omega](b ) ( note the logarithmic scale for the sdt ) . in all samples , the dominant spin dephasing mechanism at low temperatures is the dyakonov - perel mechanism @xcite , and in the motional narrowing regime , the sdt is given by @xmath42 where @xmath36 is the momentum relaxation time , and @xmath22 is the precession frequency due to the effective spin - orbit fields . as the qw width is reduced , the magnitude of the dresselhaus field increases due to the momentum quantization along the growth axis , and the magnitude of the rashba field was also shown to increase monotonously with decreasing qw width in the thickness range investigated here .
@xcite the rising amplitude of both spin - orbit fields should lead to a more rapid dephasing in thinner qws according to eq .
( [ dp_motional ] ) .
we can explain the non monotonous dependence of the sdt on the qw width by taking into account the different momentum relaxation times in our samples .
the momentum relaxation times @xmath36 , calculated from the mobility data measured at 4.2 k , are listed in table [ samples ] . using these values , we calculate @xmath43 using eq .
( [ dp_motional ] ) .
the results , shown in fig .
[ trkr_sdt_omega](c ) , demonstrate a near - monotonous increase of @xmath43 with decreasing qw width , as expected from the qw width dependence of the effective spin - orbit fields . 4 , 6 and 15 nm .
( b ) zero - field sdt as a function of the qw width ( log .
( c ) @xmath43 as a function of the qw width.,scaledwidth=45.0% ]
the linear mpge excited by linearly polarized radiation has been detected in all samples . under thz irradiation
we observed a photocurrent signal @xmath30 , which is linearly increasing with rising magnetic field strength @xmath24 and changes its sign upon the inversion of the magnetic field direction from @xmath44 to @xmath45 , demonstrating the typical mpge behavior .
obtained at room temperature , @xmath46 t and photon energy @xmath47 mev and corresponding @xmath0 factors ( triangles , by trkr ) .
the inset shows the experimental geometry for the lmpge.,scaledwidth=50.0% ] as our experiments here are focused on the linear mpge , we eliminate any possible background in our experiments by @xmath48/2.\ ] ] figure [ fig2 ] shows the photocurrent @xmath49 as a function of @xmath7 obtained for magnetic fields of @xmath50 t. for comparison the effective land factor @xmath0 , extracted from the time - resolved kerr rotation , is also plotted ( details of these measurements are already discussed in sec . [ kerr ] ) . as an important result fig .
[ fig2 ] demonstrates that the sia - induced photocurrent @xmath30 , similarly to the @xmath0 factor , changes its sign upon the variation of @xmath7 .
@xcite however , there is a difference in the zero points : while the @xmath0 factor equals to zero at @xmath51 nm , the current vanishes for @xmath52 nm .
it will be shown below that this shift between the inversion points as well as the mpge s sign inversion can be well described by the interplay of spin and orbital mechanisms in the current formation . to explain
qualitatively our results we describe the basic physics of these mechanisms .
we start with the spin - related mechanism .
the generation of a spin - polarized current due to the mpge may be discussed in the frame of a recently proposed model for the spin - dependent asymmetric energy relaxation of a non - equilibrium electron gas heated by , e.g. , thz or microwave radiation .
@xcite free electrons are excited to higher energy states by absorbing radiation and then relax into an equilibrium state by emitting phonons .
figure [ fig5a ] sketches the hot electron energy relaxation processes in the two spin subbands ( @xmath53 that are split due to the zeeman effect in the presence of an external magnetic field .
the electron relaxation from higher to lower energies is shown by bent arrows . in gyrotropic media , like gaas low - dimensional semiconductors ,
the spin - orbit interaction makes the scattering probability spin - dependent : @xcite @xmath54_z \right\},\ ] ] where @xmath55 is the symmetric part of the scattering probability , which determines the mobility ( we consider sia only ) . here
@xmath56 and @xmath57 are the initial and scattered electron wave vectors and @xmath58 is the vector composed of the pauli matrices .
thus , the electron transitions to positive and negative @xmath59-states occur with different probabilities causing an imbalance in the distribution of carriers in both subbands between positive and negative @xmath60-states .
this is shown by different thicknesses of the bent arrows in fig .
[ fig5a ] . at zero magnetic ,
field the asymmetry of the electron - phonon interaction would lead to a spin current , but not an electric current : this is due to the fact that the oppositely directed electron fluxes , @xmath61 , are of equal strength and therefore , compensate each other .
the presence of an in - plane magnetic field @xmath24 , however , leads to the zeeman splitting of the @xmath62 subbands . as a consequence , the electron densities @xmath63 and @xmath64 in both spin subbands become different , and
the fluxes @xmath65 do not longer compensate each other and therefore a net electric current results .
obviously , this current , classified as linear mpge , is spin - polarized and its value is proportional to the zeeman spin splitting induced by the magnetic field .
its qw width dependence is described by the product @xmath66 . in order to estimate the parameter @xmath67 caused by sia
, we take into account that the remote impurities create an electric field @xmath68 along the growth direction , leading to the asymmetry of the qw .
therefore , the eigenstates in this structure are superpositions of the states of the rectangular qw . as a result ,
the envelope wave functions in the first and second subbands of size quantization @xmath69 are given by @xmath70 here @xmath71 is the growth direction , @xmath72 are the functions of size quantization in the ground and the first excited subbands of the rectangular qw of width @xmath7 , @xmath73 is the energy separation between these subbands , and @xmath74 is the coordinate matrix element calculated between these states .
this leads to @xmath75 , and to the following dependence of this contribution to the linear mpge current : @xmath76 here we disregard a weak dependence of the scattering probability on the qw width . in a set of samples with similar structure inversion asymmetry
, like in our samples , where sia is controlled by the asymmetric doping , the mpge should vanish for the sample with zero @xmath0 factor and change its sign upon a qw width variation .
thus , the spin - related mechanism of the mpge describes well the appearance of the sign inversion .
however , it can not explain the fact that the mpge s sign inversion takes place for an about 4 nm broader qw than the inversion of the @xmath0 factor .
@xcite the explanation of this fact requires an additional spin - independent orbital contribution to the mpge , recently addressed by tarasenko .
@xcite a process resulting in the orbital linear mpge is illustrated in fig .
[ fig5b ] .
here , we consider the in - plane motion of carriers subjected to the high - frequency electric field ( e.g. , thz radiation ) applied along the @xmath26 direction and an in - plane magnetic field along the @xmath77 direction .
for linearly polarized radiation , the electric field @xmath78 leads to a back and forth motion of electrons with velocity @xmath79 as illustrated in fig .
[ fig5b ] for two different times . without external field
the electron s wave function is localized closer to one interface due to the asymmetrical doping , and its momentum relaxation rate @xmath80 has a contribution controlled by scattering by this interface .
the orbital mpge originates from the combined action of electric and magnetic fields , described in the following .
if @xmath78 is applied , at a certain time @xmath81 , the electron is accelerated by the in - plane @xmath82 electric field along the negative @xmath26 direction . at the same time , the electron with velocity @xmath83 is subjected to the magnetic field @xmath84 .
this results in the lorentz force @xmath85 which shifts the wave function to the center of the quantum well and thus decreases the scattering probability ( see fig .
[ fig5b ] ) .
half a period later , at @xmath86 , the electron velocity gets reversed so that the direction of the lorentz force reverses as well .
now , the electron wave function is pushed closer to the interface , which increases the scattering rate and decreases the momentum relaxation .
the resulting imbalance of the relaxation times for the motion along positive and negative @xmath26 direction causes a net electric current proportional to the magnetic field strength .
obviously this mechanism yields a photocurrent , which does not change its sign with a variation of the qw width .
microscopically , the orbital contribution to the linear mpge current @xmath87 is caused by the effect of a magnetic field on the scattering .
the in - plane magnetic field results in a @xmath88-dependent mixing of the eigenstates @xmath89 and @xmath90 , and the envelope in the ground subband becomes @xmath91-dependent @xcite @xmath92 this leads to the following contribution in the scattering probability @xmath93_z \right\},\ ] ] where the parameter @xmath94 is caused by sia .
the dependence @xmath95 determines the dependence of @xmath96 on the quantum well width yielding @xmath97 the quadratic dependence on the parameter @xmath98 appears because , both , the electric field of the impurities and the magnetic field , lead to a mixing of the ground and first excited levels , eqs . and . here , like for eq .
, we also assume that additional factors originating from the scattering asymmetry depend weakly on the qw width . on the phenomenological level
both mechanisms are described by the same equations and the total current is given by the sum of both contributions @xmath99 the phenomenological similarity hinders the decomposition of both terms , because the spin contribution @xmath100 and the orbital one @xmath87 behave identically under a variation of the radiation s polarization state and the orientation of the magnetic field relative to the crystallographic axes .
our above consideration shows , however , that the behavior of the photocurrent upon a variation of the qw width allows us to distinguish between these two basically different mechanisms .
the observed sign inversion clearly demonstrates that the total current almost reflects the behavior of the @xmath0 factor and , therefore , is in most samples dominated by the spin mechanism .
the small shift of the mpge s inversion point to a larger @xmath7 compared to the one of the @xmath0 factor , however , demonstrates that the orbital mpge also contributes and its magnitude is smaller but still comparable to that of the spin mpge . for a qw with @xmath7 of about 6.5 nm
the @xmath0 factor is equal to zero and the mpge is solely caused by the orbital mechanism .
the dependence of @xmath96 on the quantum well width , calculated after eq .
is plotted in fig .
[ fig : theory ] .
figure [ fig : theory ] shows that this current rises with increasing @xmath7 due to the increase of the mixing parameter @xmath98 with @xmath7 but does not change its sign .
this behavior is explained as follows : in narrow qws the electric and magnetic fields can not efficiently mix size - quantized states because of their large energy separation @xmath73 . in wider
qws the confinement is weaker , and the photocurrent increases .
this strong dependence is valid for not too wide qws , where only the ground subband is occupied . in very wide qws
, the photocurrent dependence on the qw width has a maximum and then the photocurrent tends to zero since in bulk gaas photogalvanic effects are absent .
figure [ fig : theory ] also shows the interplay of both contributions . to obtain this curve , we have normalized both contributions in such a way that the total current vanishes for the qw width where the photocurrent s sign inversion has been detected .
one can see that the total photocurrent vanishes for a larger qw width than the @xmath0 factor .
this is due to the orbital contribution .
the sign inversion upon a variation of the qw width is also obtained for the sia - induced circular mpge .
figure [ fig3 ] shows the helicity dependence of the photocurrent @xmath30 , which clearly demonstrates the fingerprint of the circular mpge : the sign inversion upon switching the radiation s helicity @xmath101 from @xmath102 to @xmath103 at @xmath104 ( @xmath105 ) and @xmath106 ( @xmath107 ) , respectively .
we note , that the whole polarization dependence stems from the interplay of the circular and linear mpge . in agreement with the phenomenological theory @xcite the signal
can be well fitted by @xmath108 .
t and a photon energy of @xmath47 mev .
the ellipses on top illustrate the polarization states for various angles of the @xmath29 plate , @xmath28 .
the inset depicts the spin - galvanic effect.,scaledwidth=45.0% ] here the first term , given by the parameter @xmath109 , is due to the circular mpge and the following two , proportional to the coefficients @xmath110 and @xmath111 , stem from the linear mpge discussed above and vanish for circularly polarized radiation .
figure [ fig4 ] shows the quantum well width dependence of the circular photocurrent . in order to extract the circular mpge from the total current and to ensure that no spurious currents contribute to the signal we used the fact that the circular mpge reverses its sign upon switching the radiation s helicity from @xmath102 ( @xmath112 ) to @xmath103 ( @xmath113 ) and obtain the circular photocurrent after @xmath114 / 2 , \label{equcmpge}\ ] ] where @xmath115 and @xmath116 are photocurrents measured at @xmath112- and @xmath113-polarized excitation . because the circular mpge also reverses its sign upon the inversion of the magnetic field direction , the current after eq .
( [ equcmpge ] ) is further treated similarly to the linear mpge according to eq .
( [ mpge ] ) . for the circular mpge
, we also observed that the photocurrent reverses its sign at a certain qw width .
@xcite similarly to the behavior of the linear mpge , the sign inversion of the photocurrent in fig .
[ fig4 ] does not coincide with that of the land factor and takes place at @xmath117 nm .
we show below that , similarly to the linear mpge discussed earlier , this fact indicates the interplay of the spin and orbital mechanisms in the circular mpge .
the spin mechanism of the circular mpge is based on the spin - galvanic effect .
@xcite this effect results in the generation of an electric current due to a non - equilibrium spin polarization and is caused by its asymmetric spin relaxation .
the spin - galvanic current and the average non - equilibrium spin @xmath118 are related by a second rank pseudotensor , with components proportional to the parameters of the spin - orbit splitting , as follows @xmath119 for ( 001)-grown zinc - blende structure based qws of c@xmath120-symmetry , the spin - galvanic effect is allowed only for a spin polarized electron gas with spins oriented in the qw plane . in the coordinate system parallel to the cubic axes , eq .
( [ equ22 ] ) reduces to @xmath121 and @xmath122 , with @xmath123 and @xmath124 proportional to the rashba and the dresselhaus constants , respectively .
@xcite the spin - galvanic effect generally does not require optical excitation .
in fact , the non - equilibrium spin can be achieved by both optical and non - optical methods , e.g. , by electrical spin injection
. obtained at room temperature , @xmath125 t and a photon energy of @xmath47 mev and corresponding @xmath0 factors ( triangles , by trkr ) . the inset shows the experimental geometry for the cmpge.,scaledwidth=50.0% ] the spin - dependent contribution @xmath126 stems from the asymmetry in the spin relaxation , arising due to the rashba spin - orbit splitting of the electronic ground subband and being linear in the in - plane wave vector @xmath88 : @xmath127 here the sia - induced @xmath88-linear spin - orbit splitting @xmath128 is measured by trkr technique , sec . [ kerr ] .
an optical method , which provides a non - equal population of an uniform distribution in both spin subbands was proposed in ref . .
this method , sketched in the inset of fig .
[ fig3 ] , is based on the optical excitation with circularly polarized light yielding a steady - state spin orientation @xmath129 in the growth direction .
the in - plane magnetic field then rotates , due to the larmor precession , the optically oriented spins into the plane of the 2des . if @xmath130 is oriented along the @xmath26 axis we obtain a non - equilibrium spin polarization @xmath131 which reads after time averaging @xcite @xmath132 where @xmath133 and @xmath134 are the longitudinal and transverse electron spin relaxation times , the larmor frequency is given by @xmath135 , @xmath136 is the bohr magneton , and @xmath137 is the steady state electron spin polarization in the absence of a magnetic field . according to eq .
( [ equ22 ] ) the in - plane spin polarization @xmath131 causes a net electric current in the direction along the magnetic field . for the investigated qws at room temperature @xmath138 and ,
therefore , the photocurrent is proportional to @xmath139 .
the qw width dependence of @xmath126 is given by @xmath140 equation shows that this contribution is proportional to the @xmath0 factor and , consequently , reflects its behavior upon a qw width variation .
@xcite despite the qw width dependence of @xmath126 does not coincide with @xmath141 , this contribution to the photocurrent vanishes in qws with @xmath142 .
the fact that in gaas / algaas qws the current , similarly to the @xmath0 factor , changes its sign upon a variation of @xmath7 qualitatively explains the mpge s sign inversion but can not clarify the difference in the two zero points . to explain this shift we should again address the possible contribution of the orbital mpge .
an orbital mechanism yielding a helicity driven mpge current was recently suggested in ref . and is given by @xmath143 the second rank pseudotensor @xmath144 has the same space symmetry properties as the pseudotensor @xmath145 , describing the spin - galvanic effect .
however , the tensor @xmath144 is invariant under time inversion .
microscopically , the orbital contribution to the circular mpge appears similarly to the one to the linear mpge current described above .
the current is caused by the action of the lorentz force on the orbital motion of the 2d electrons in the radiation field .
under irradiation with circularly polarized light electrons perform a cyclic motion and , due to sia , the presence of an in - plane magnetic field forces them to flow predominantly along the direction of @xmath130 .
reversing the radiation s helicity , changes the current direction .
the microscopic theory of this effect is given in ref . .
the resulting current @xmath146 , is caused by the @xmath130-dependent corrections to the scattering probability , eq .
, as well as the corresponding contribution to the linear mpge .
the photocurrents due to the circular and linear mpge are linked to each other according to @xmath147 at @xmath148 ( ref . ) .
using eq . , we can , consequently , estimate the qw width dependence of @xmath149 as @xmath150 this equation shows that the orbital contribution to the circular mpge has a constant - sign dependence on the qw width .
thus , the observed sign inversion of the circular mpge and the shift between its zero crossing and that of the @xmath0 factor agrees with the picture of the interplay of comparable spin and orbital contributions to this phenomenon
. both mechanisms of the circular mpge contribute to the total current @xmath151 and are described phenomenologically by similar equations .
the observed sign inversion proves that , in correspondence with the results of the linear mpge , the dominant contribution comes from the spin - galvanic effect . while the spin - galvanic effect dominates the current for most of the investigated samples , the existence of the orbital circular mpge is also clearly demonstrated .
in particular , for qws with a width of about 6.5 nm the spin - galvanic effect vanishes and the current is caused solely by the orbital circular mpge .
to summarize , our experiments clearly demonstrate that , both , linear and circular mpge in gaas / algaas qw structures result from spin and orbital contributions .
our experiments show that for most quantum well widths the mpge is mainly driven by spin - related mechanisms , which result in a photocurrent proportional to the @xmath0 factor . for structures with a vanishingly small @xmath0 factor , however , the mpge caused by orbital mechanisms is clearly observed .
our work demonstrates that a variation of the electron @xmath0 factor by different means , like varying the qw width , as it is done here , doping with magnetic impurities or using narrow band materials , where @xmath0 and the spin - orbit interaction are enhanced , can be used for the separation of these qualitatively different mechanisms .
moreover , the orbital mpge can be studied independently in materials with a vanishingly small spin - orbit interaction , like si - based metal - oxide - semiconductor low dimensional structures .
we thank s.a .
tarasenko and e.l .
ivchenko for fruitful discussions and support , as well as m. schmalzbauer .
this work is supported by the dfg via programs spp 1285 , sfb 689 , by the linkage grant of ib of bmbf at dlr , russian ministry of education and sciences , rfbr and `` dynasty '' foundation icfpm . v. v. belkov and s. d. ganichev , _ zero - bias spin separation _ , in _ handbook of spintronic semiconductors _ , edited by w. m. chen and i. a. buyanova ( pan stanford publishing , singapore 2010 ) . s. d. ganichev , v. v. belkov , s. a. tarasenko , s. n. danilov , s. giglberger , c. hoffmann , e. l. ivchenko , d. weiss , w. wegscheider , c. gerl , d. schuh , j. stahl , j. de boek , g. borghs and w. prettl , nature physics * 2 * , 609 ( 2006 ) . v. v. belkov , p. olbrich , s. a. tarasenko , d. schuh , w. wegscheider , t. korn , c. schller , d. weiss , w. prettl , and s. d. ganichev , phys .
lett . * 100 * , 176806 ( 2008 ) .
s. d. ganichev , petra schneider , v. v. belkov , e. l. ivchenko , s. a. tarasenko , w. wegscheider , d. weiss , d. schuh , b. n. murdin , p. j. phillips , c. r. pidgeon , d. g. clarke , m. merrick , p. murzyn , e. v. beregulin , and w. prettl , phys .
b. * 68 * , 081302 ( 2003 ) .
s. giglberger , l. e. golub , v. v. belkov , s. n. danilov , d. schuh , ch .
gerl , f. rohlfing , j. stahl , w. wegscheider , d. weiss , w. prettl , and s. d. ganichev , phys .
b * 75 * 035327 ( 2007 ) .
v. lechner , l. e. golub , p. olbrich , s. stachel , d. schuh , w. wegscheider , v. v. belkov , and s. d. ganichev , appl .
. lett . * 94 * , 242109 ( 2009 ) . c. l. yang , j. dai , w. k. ge , and x. cui , appl .
* 96 * , 152109 ( 2010 ) .
m. i. dyakonov and v. i. perel , sov .
- jetp * 33 * , 1053 ( 1971 ) .
e. a. de andrada e silva , g. c.la rocca , and f. bassani , phys .
b * 55 * , 16293 ( 1997 ) . note , that @xmath152 of the sample with @xmath153 nm is twice as large as in the other samples , therefore sia and consequently the sia - induced current is enhanced compared to the samples with larger @xmath7 . c. drexler , v. v. belkov , b. ashkinadze , p. olbrich , c. zoth , v. lechner , ya .
v. terentev , d. r. yakovlev , g. karczewski , t. wojtowicz , d. schuh , w. wegscheider , and s. d. ganichev , appl .
97 * , 182107 ( 2010 ) .
note , that comparing our mpge data obtained at room temperature with the @xmath154-factor behavior measured at low temperatures , we do not consider the weak temperature dependence @xmath0-factor , @xcite which may result in a small shift of its zero crossing .
v. v. belkov , s. d. ganichev , e. l. ivchenko , s. a. tarasenko , w. weber , s. giglberger , m. olteanu , h .-
tranitz , s. n. danilov , petra schneider , w. wegscheider , d. weiss , and w. prettl , j. phys .
c : condens .
matter * 17 * , 3405 ( 2005 ) . here and
latter the calculations are limited to rather narrow quantum wells , for which only one subband is occupied . in a wider qws
other subbands should be taken into account which is out of scope of this paper .
we note , that recently an additional possible root of the spin - galvanic effect has been addressed in ref . . instead of sge
due to band spin splitting considered in ref .
, it is based on the spin - dependent scattering given by eq . .
however , this mechanism also yields a contribution proportional to the zeeman splitting and does not qualitatively change the discussion of the mpge variation upon the qw width change . | we report on the study of the linear and circular magneto - gyrotropic photogalvanic effect ( mpge ) in gaas / algaas quantum well structures . using the fact that in such structures the land - factor @xmath0 depends on the quantum well ( qw ) width and has different signs for narrow and wide qws , we succeeded to separate spin and orbital contributions to both mpges .
our experiments show that , for most quantum well widths , the mpges are mainly driven by spin - related mechanisms , which results in a photocurrent proportional to the @xmath0 factor . in structures with a vanishingly small @xmath0 factor , however , linear and circular mpge are also detected , proving the existence of orbital mechanisms . |
typically in quintessence theories with an asymptotically vanishing effective cosmological constant , the energy contrast in dark energy @xmath3 rises from near zero for redshifts @xmath4 to near one for @xmath5 , mimicking a true cosmological constant . at late times
the quintessence field may begin to oscillate about its minimum , behaving like nonrelativistic matter , or the quintessence field may evolve toward infinity in both cases with vanishing vacuum energy . in such theories , there is a period between roughly 3.5 gyr and 20 gyr after the big bang when @xmath0 .
however if the universe continues to expand forever , or even if positive curvature begins to dominate at late times ( after the quintessence field has evolved to its minimum ) and the universe enters a contracting stage , this period when the energy densities of dark energy and matter are comparable is a small or vanishing fraction of the total lifetime of the universe .
this is called the cosmic coincidence problem .
however in axion quintessence ( as in other unstable de sitter quintessence models ) , the cosmological era with @xmath6 may represent a significant fraction of the universe s lifetime if the minimum of the axion potential is negative ( unstable [ de sitter ] axion quintessence ) , thus resolving @xcite the cosmic coincidence problem .
( negative @xmath7 or @xmath8 and the fate of the universe are discussed in refs .
@xcite and @xcite plus references therein . )
vilenkin @xcite summarizes the currently predominant view that the dark energy is a cosmological constant , and points out the major weaknesses of most quintessence models : ( i ) most quintessence models assume that @xmath9 as the quintessence field @xmath10 evolves toward its minimum ( which may be at @xmath11 ) , conflicting with the expectation that @xmath8 ought to evolve to a nonzero vacuum energy density ; ( ii ) most quintessence models do not solve the coincidence problem that @xmath12 ( the subscript `` 0 '' will denote present values ) ; and ( iii ) since the present dark energy average equation of state parameter @xmath13 , perhaps @xmath14 and quintessence models are irrelevant . the unstable axion quintessence potential @xmath15 , where @xmath16 and the planck mass @xmath17 gev , addresses all of these issues , since the facts that the minimum of the potential is at @xmath18 , @xmath19 , and @xmath20 but @xmath21 are interrelated aspects of the model , and occur for an appreciable range of initial values for @xmath10 .
for @xmath22 , the initial value of the scalar field need only satisfy @xmath23 to produce a universe like ours @xcite ( due to symmetry , we can restrict our attention to @xmath24 ) .
thus there is a significant 23% range of the possible initial values @xmath25 which will produce a universe like ours . for @xmath26 . ] for these initial values , the contracting universe enters a late time era of kination ( during which the scalar field kinetic energy dominates over all other forms of energy ) , the negative hubble parameter acting like a negative friction term in the klein - gordon equation , and the axion field makes many transits of but never remains in its vacuum state . in section 2 ,
the basic cosmological equations are presented for the evolution of the scalar field , the scale factor , and the hubble parameter , and cast in the form of a scaled , dimensionless system of first - order equations in the conformal time , appropriate for a contracting ( or expanding ) universe .
these equations allow the simulations presented in section 3 ( see figs . [ fig - v][fig - ke - v ] ) to proceed much closer to the singularity at the end of the collapsing universe than any other simulations presented in the literature , and provide the basis for a more detailed analysis of the last stages of the collapsing universe than has appeared before .
in the quintessence / cold dark matter ( qcdm ) model , the total energy density @xmath27 , where @xmath28 , @xmath29 , and @xmath30 are the energy densities in ( nonrelativistic ) matter , radiation , and the axion quintessence scalar field @xmath10 , respectively . ratios of energy densities to the critical energy density @xmath31 for a flat universe will be denoted by @xmath32 , @xmath33 , and @xmath34 , while ratios of present energy densities @xmath35 , @xmath36 , and @xmath37 to the present critical energy density @xmath38 will be denoted by @xmath39 , @xmath40 , and @xmath41 , respectively .
@xmath3 will denote @xmath42 using wmap5 @xcite central values , we will set @xmath1 = 0.72 , @xmath43 , @xmath44 , and @xmath45 ev , with the present time @xmath46 gyr after the big bang for @xmath47cdm .
the homogeneous scalar field obeys the klein - gordon equation @xmath48 the evolution of the universe is described by the friedmann equations for the hubble parameter @xmath49 and the scale factor @xmath50 @xmath51 @xmath52 where the energy density @xmath53 and the pressure @xmath54 , with @xmath55 = 0 , @xmath56 , and @xmath57 the curvature signature @xmath58 , 0 , @xmath59 for a closed , flat , or open geometry .
. shows that @xmath60 for an accelerating universe .
the conservation of energy equation for matter , radiation , and the scalar field is @xmath61 equation gives the evolution of @xmath28 and @xmath29 , and with eq . the klein - gordon equation for the weakly coupled scalar field .
the time rate of change of the hubble parameter is given by @xmath62 only two of eqs .
, , , and are independent .
we will assume a flat universe after inflation and henceforth set @xmath63 = 0 .
the logarithmic time variable ( number of e - folds of the scale factor ) is defined as @xmath64 .
note that for de sitter space @xmath65 , where @xmath66 , and that @xmath67 is a natural time variable for the era of @xmath47-matter domination ( see e.g. ref .
we will make the simple approximations @xmath68 the equation of state parameter for the scalar field @xmath10 is @xmath69 . since @xmath70 is a natural time variable for the era of @xmath47-matter domination , we define the recent average of @xmath71 as @xmath72 we will take the upper limit of integration to correspond to @xmath73 = 1.75 .
the sne ia observations @xcite bound the recent average @xmath74 ( 95% cl ) . for numerical simulations ,
the cosmological equations should be put into a scaled , dimensionless form .
equations and can be cast @xcite in the form of a system of two first - order equations in @xmath70 plus a scaled version of @xmath75 : @xmath76 @xmath77 @xmath78 @xmath79 where @xmath16 , @xmath80 , @xmath81 , @xmath82 , @xmath83 , @xmath84 , @xmath85 , @xmath86 .
this scaling results in a set of equations that is numerically more robust , especially near and before the time of big - bang nucleosynthesis ( bbn)see ref .
@xcite , especially fig . 1 . for a contracting phase ( in which @xmath75 goes through zero ) , a different set of equations and a different scaling should be used . here
we will use the conformal time variable @xmath87 where @xmath88 corresponds to the big bang .
equations and can be cast in the form of a system of three first - order equations in @xmath89 : @xmath90 @xmath91 @xmath92 @xmath93 where @xmath16 , @xmath94 , @xmath95 , @xmath96 , @xmath97 , @xmath98 , @xmath99 , @xmath100 . this scaling results in a set of numerically more robust equations , especially near the turn - around time @xmath101 between expanding and contracting phases of the universe .
note that the conformal time @xmath89 is related to the logarithmic time @xmath70 by @xmath102
the original axion quintessence potential @xmath103 was based on @xmath104 supergravity @xcite , with @xmath105 . as @xmath106
, the universe evolves to minkowski space .
the unstable de sitter axion potential @xmath107 is based on m / string theory reduced to an effective @xmath104 supergravity theory @xcite , with @xmath108 at the maximum of @xmath109 .
both axion quintessence models are derivable ( up to a constant ) from string theory as axion monodromy @xcite .
the quintessence axion is a pseudo nambu - goldstone boson : at the perturbative level the theory is shift symmetric under @xmath110 with @xmath111 .
the shift symmetry is broken before or during inflation by nonperturbative instanton effects to a discrete symmetry @xmath112 , generating a potential @xmath113 . in these theories ,
quantum corrections to the classical axion potential are suppressed . for quintessence ( or for natural inflation @xcite ) , @xmath114 ; we will take @xmath115 .
@xmath116 and @xmath117 are the most interesting unstable axion and original axion cases , respectively . for the unstable axion quintessence computations ( with an expanding and contracting universe )
, we will use eqs . and with initial conditions @xmath25 and @xmath118 specified at matter - radiation equality @xmath119 , which corresponds to @xmath120 the constant @xmath121 in the potential is adjusted so that @xmath41 = 0.72 .
this involves the usual single fine tuning .
note that the same anthropic arguments that limit the magnitude of a present - day cosmological constant also limit @xmath122 , so the tuning of @xmath121 is no worse than the tuning of a cosmological constant .
results for the unstable axion potential are presented in table [ cos ] for various @xmath25 and for @xmath123 in figs .
[ fig - omega][fig - ke - v ] .
( as @xmath124 , classically @xmath125 , but quantum effects destabilize @xmath126 so that the maximum @xmath127 @xcite . ) for the values in the table , as @xmath25 increases , @xmath128 also increases and @xmath10 starts to move earlier , leading to a decrease in @xmath129 , @xmath130 , @xmath101 , and @xmath131 , and correspondingly to an increase in @xmath132 away from @xmath59 . note that for @xmath133 , the coincidence time ratio approaches 50% .
the qcdm universe mimics the @xmath47cdm model ( see fig .
[ fig - omega ] ; for clarity , only the beginnings of the contracting stage are shown in this figure ) until about @xmath134 , after which the qcdm universe begins to decelerate and ultimately to rapidly contract to a big crunch ( figs .
[ fig - ah ] and [ fig - a ] ) . during the contracting stage , @xmath135 acts as a _ negative _ friction in @xmath136 amplifying the axion field and its kinetic energy to bring about a late stage kination era during which the scalar field kinetic energy dominates over all other forms of energy .
the quintessence axion is an ultra - light scalar field with @xmath137 , so @xmath10 `` sits and waits '' during the early evolution of the universe , and only starts to move when @xmath138 ( figs .
[ fig - phi ] and [ fig - psi ] ) . in this way
it is easy to satisfy the bbn ( @xmath139@xmath140 ) , cosmic microwave background ( cmb ) ( @xmath141@xmath142 ) , and large scale structure ( lss ) ( @xmath143@xmath144 ) bounds on @xmath145 , as in fig .
[ fig - omega ] .
an ultra - light scalar field also reflects the observational evidence that the universe has only recently become dominated by dark energy . in fig .
[ fig - ah ] , the hubble parameter goes through zero at the turn - around time between an expanding and contracting universe . at the beginning of the contacting stage
, @xmath10 has yet to reach the minimum of the potential energy ( see fig .
[ fig - v ] ) , and thus the negative hubble parameter amplifies the kinetic energy of the scalar field , bringing about an era of kination with @xmath146 , as seen in figs .
[ fig - phi ] , [ fig - psi ] , [ fig - w ] , and [ fig - w - o ] .
also note that figs .
[ fig - ah ] , [ fig - phi ] , and [ fig - psi ] indicate that @xmath75 , @xmath10 , and @xmath147 are approaching a singularity near @xmath131 . in fact , in an era of kination during contraction during which @xmath148 , @xmath149 and @xmath150 , while @xmath151 ( see fig .
[ fig - a ] ) .
figure [ fig - v ] shows that as @xmath10 increases without bound , the potential energy @xmath152 , since a periodic function of @xmath10 , oscillates more and more rapidly .
depending on the strength of the coupling between the quintessence axion and other particles , @xmath10 may decay and populate the universe with additional radiation and matter .
( the monotonically increasing field @xmath10 in a periodic potential can be interpreted as an oscillating field . ) at the very least , there should be gravitational production of particles by @xmath10 during contraction .
figures [ fig - w ] and [ fig - w - o ] follow the contracting stage further , illustrating that although @xmath153 twice just after @xmath101 , the universe ultimately enters a stage of kination in which @xmath146 near @xmath131 .
note that @xmath154 until well after @xmath130 .
after matter - quintessence equality at @xmath155 , the scalar field energy density always dominates over the matter and radiation energy densities .
there is a period from @xmath156 until @xmath157 when the scalar field potential energy is comparable to its kinetic energy , and then the kinetic energy ( which scales as @xmath158 ) predominates during the rapid contraction to a big crunch ( see fig . [ fig - ke - v ] ) .
as the universe contracts , density inhomogeneities are amplified and presumably black holes are formed , similarly to the contracting stage of the ekpyrotic universe @xcite with @xmath159 . depending on the strength of the coupling ( which we have neglected here ) between the quintessence axion and other particles , @xmath10 may decay and produce radiation and matter .
there should at least be gravitational production of particles by @xmath10 during contraction . as the universe reheats during contraction , broken symmetries are restored .
it is possible that inflating patches may be generated , spawning new universes from the old .
these issues are currently under investigation . in summary
, the unstable axion quintessence potential resolves the issues concerning quintessence raised by vilenkin : the minimum of the potential is not at zero , but at a negative value @xmath160 , @xmath19 with only a single fine tuning ( in the anthropic range ) , and @xmath132 naturally satisfies @xmath161 , for an appreciable 23% range of possible initial values for the quintessence field . and for a universe like ours , the coincidence time when the energy densities of dark energy and matter are comparable varies ( as long as @xmath162 is not too small say , @xmath163 0.05 ) from 25%50% of the lifetime of the universe .
my thanks are due to lawrence krauss for valuable comments .
kallosh r , linde a d , prokushkin s , and shmakova m , _ supergravity , dark energy and the fate of the universe , _ 2002 phys .
d * 66 * 123503 [ arxiv : hep - th/0208156 ] .
krauss l m and turner m s , _ geometry and destiny , _ 1999 gen .
grav . * 31 * 1453 [ arxiv : astro - ph/9904020 ] .
vilenkin a , _ perspectives in cosmology , _ 2010 j. phys . conf .
ser . * 203 * 012001 [ arxiv:0908.0721 [ astro-ph.co ] ] .
gardner c l , _ quintessence and the transition to an accelerating universe , _ 2005 nucl .
b * 707 * 278 [ arxiv : astro - ph/0407604 ] .
komatsu e _ et al .
_ [ wmap collaboration ] , _ five - year wilkinson microwave anisotropy probe ( wmap ) observations : cosmological interpretation , _ 2009 astrophys
. j. suppl .
* 180 * 330 [ arxiv:0803.0547 [ astro - ph ] ] .
gardner c l , _ cosmological variation of the fine structure constant from an ultra - light scalar field : the effects of mass , _ 2003 phys .
d * 68 * 043513 [ arxiv : astro - ph/0305080 ] .
riess a g _ et al .
_ [ supernova search team collaboration ] , _ type ia supernova discoveries at z>1 from the hubble space telescope : evidence for past deceleration and constraints on dark energy evolution , _ 2004 astrophys .
j. * 607 * 665 [ arxiv : astro - ph/0402512 ] .
frieman j a , hill c t , stebbins a , and waga i , _ cosmology with ultralight pseudo nambu - goldstone bosons , _ 1995 phys .
lett . * 75 * 2077 [ arxiv : astro - ph/9505060 ] .
waga i and frieman j a , _ new constraints from high redshift supernovae and lensing statistics upon scalar field cosmologies , _ 2000 phys .
d * 62 * 043521 [ arxiv : astro - ph/0001354 ] .
choi k , _ string or m theory axion as a quintessence , _ 2000 phys .
d * 62 * 043509 [ arxiv : hep - ph/9902292 ] .
silverstein e and westphal a , _ monodromy in the cmb : gravity waves and string inflation , _ 2008 phys .
d * 78 * 106003 [ arxiv:0803.3085 [ hep - th ] ] .
freese k , savage c , and kinney w h , _
natural inflation : the status after wmap 3-year data , _ 2008 int .
j. mod .
d * 16 * 2573 [ arxiv:0802.0227 [ hep - ph ] ]
. lehners j l , steinhardt p j , and turok n , _ the return of the phoenix universe , _ 2009 int .
j. mod .
d * 18 * 2231 [ arxiv:0910.0834 [ hep - th ] ] .
khoury j , ovrut b a , steinhardt p j , and turok n , _ density perturbations in the ekpyrotic scenario , _ 2002 phys .
d * 66 * 046005 [ arxiv : hep - th/0109050 ] . | in axion quintessence , the cosmological era with an energy contrast in dark energy @xmath0 may represent a significant fraction of the universe s lifetime if the minimum of the axion potential is negative ( unstable axion quintessence ) , thus resolving the cosmic coincidence problem , as pointed out by kallosh , linde , prokushkin , and shmakova .
further details of the evolution of the quintessence field , the scale factor of the universe , and the hubble parameter are presented here , focussing on models with @xmath1 = 0.72 and recent dark energy average equation of state parameter @xmath2 . for these parameter values ,
the contracting universe enters a late time era of kination , the negative hubble parameter acting like a negative friction term in the klein - gordon equation , and the axion field makes many transits of but never remains in its vacuum state .
robust , scaled cosmological equations are derived for simulating the evolution of the scalar field , the scale factor , and the hubble parameter during a contracting phase of the universe .
these equations allow the simulations presented here to proceed much closer to the singularity at the end of the collapsing universe than any previous simulations . |
_ quantity has a quality all its own _ + thomas a. callaghan jr .
local features and their descriptors play pivotal roles in many computer vision systems . as such , research on improving these methods has been immense . over the years ,
this effort yielded progressively more accurate representations .
these improvements were often demonstrated on standard benchmarks designed to measure the accuracy of descriptor matching in the presence of various image transformations and other confounding factors .
it remains unclear , however , if the improved accuracy reported on these benchmarks reflects better , more useful representations when used in real world computer vision systems .
take , for example , recent attempts to use deep learning for image feature representation ( e.g. , @xcite ) .
there is no question that given sufficient training data and computational resources deep learning methods can achieve astonishing accuracy .
hence , using them to obtain local descriptors can result in better representations and by so doing impact a wide range of computer vision systems . but using deep learning for feature description and matching does not come without a price : most of these methods are computationally expensive and even with graphical processing units ( gpu ) , are relatively slow . even after extraction , their dimensions and floating point values
makes them slow to match .
finally , they require substantial training data which can be difficult to provide .
these limitations should be contrasted with evidence that accuracy , though important , is not the only property worth considering when choosing descriptors .
for example , simultaneous localization and mapping ( slam ) methods were shown to obtain better 3d reconstructions for the same computational effort by increasing the amount of feature points per keyframe @xcite .
this suggests that computationally cheaper features are more desirable for these systems .
consequently , state of the art slam techniques @xcite use orb @xcite rather than more accurate but computationally expensive representations such as sift @xcite : doing so allows for a greater number of features to be extracted without compromising reconstruction accuracy .
in fact , even classification systems appear to benefit from having more features over higher feature accuracy , as reported by @xcite
. one side effect to the success of deep learning is that the hardware enabling it
gpu processors is now becoming standard on systems running computer vision applications , including even consumer cellphone devices . beyond deep learning
, these gpus can also be used to accelerate extraction and matching of older , so - called _ engineered _ descriptors .
these representations may not reach the same benchmark performances as deep learning techniques , but their extraction on the gpu offers a potential trade off between accuracy and run time .
in particular , faster descriptor extraction and matching allows for more descriptors to be used and consequently better overall system performances .
gpu accelerated features were considered in the past . we , however , focus on a particular binary descriptor : the _ learned arrangement of three patches _ ( latch ) @xcite .
it was recently shown to offer a compromise between the high accuracy , low speeds of floating point representations such as sift @xcite , and the low accuracy , high speeds of binary descriptors ( e.g. , orb @xcite , brief @xcite ) . beyond these properties
, its design also happens to neatly fit gpu processing .
our contributions are : * ( 1 ) * we describe clatch , a cuda port for latch , enabling descriptor extraction and matching directly on the gpu . *
( 2 ) * we embed clatch in the openmvg library @xcite , along with a fast , gpu based hamming distance , brute force descriptor matcher . finally ,
* ( 3 ) * we compare sfm 3d reconstructions on scenes from @xcite using sift , recent deep learning based representations and our clatch .
these show that clatch reconstructions are comparable or even better than those obtained with other representations , yet clatch requires a fraction of the run time of its alternatives .
importantly , to promote reproducibility , the code used in this paper is publicly available from the project webpage : www.openu.ac.il/home/hassner/projects/latch .
due to their key role in many computer vision systems , local feature descriptors are extensively studied .
a comprehensive survey is therefore outside the scope of this paper .
below we provide only a cursory overview of this topic .
* floating point representations .
* for nearly two decades now , sift @xcite is very likely the most widely used local image descriptor .
it and the representations that followed ( e.g. , surf @xcite ) represent the region around an image pixel using a vector of typically 128 floating point values .
this vector is often a histogram of measurements extracted from the image , most commonly various functions of the local intensity gradients . *
binary descriptors . * despite the success of the older floating point representations , a prevailing problem was their extraction time and dimensionality ( which , in turn , affected their storage and matching time ) . in response ,
binary descriptors were proposed as low dimensional , efficient alternative representations .
these typically assign descriptor values by quick , pixel intensity comparisons .
one of the first binary descriptors was the binary robust independent elementary features ( brief ) @xcite , soon followed by the oriented fast and rotated brief ( orb ) descriptor @xcite which added rotation invariance , the binary robust invariant scalable keypoints ( brisk ) @xcite which used a more effective pixel sampling pattern , and the fast retina keypoint descriptor ( freak ) @xcite which sampled intensities using a pattern similar to the one found in human retinas .
the accelerated - kaze ( a - kaze ) was suggested in @xcite .
it builds on the earlier local difference binary ( ldb ) descriptor @xcite by computing the binary descriptor values from mean image intensities over a range of patch sizes .
the binary online learned descriptor ( bold ) @xcite improve accuracy yet retain high processing speeds .
finally and very recently , the latch binary descriptors were proposed in @xcite .
we defer discussion of latch to sec .
[ sec : latch ] .
hybrid binary / floating - point methods were also suggested .
one example is lda - hash @xcite which extracts sifts , projects them to a discriminative space and applies a threshold to obtain binary descriptors .
dbrief @xcite instead uses patch intensities directly , binboost @xcite learns a set of hash functions corresponding to each bit in the final descriptor and pr - proj @xcite uses learning and dimensionality reduction to produce compact binary representations .
the computational effort required to extract these descriptors is similar to ( if not greater than ) floating point descriptors .
the representations , however , are short binary vectors and so matching and storing them is relatively efficient
. * computing local descriptors on the gpu .
* of course , we are not the first to propose porting local feature extraction to the gpu . to our knowledge , nearly all these efforts used the gpu to aid in extracting _ floating point descriptors _ , including gpu - sift ( see , e.g. , @xcite ) and gpu - surf @xcite . these methods all used gpus in portions of the extraction process .
for example , @xcite used the gpu only to compute convolutions , all other stages performed on the cpu .
in addition , and more importantly , the gain in performance reported by these methods are modest and do not approach the speeds of contemporary binary descriptors , let alone our clatch .
interestingly , the only available gpu _ binary descriptor _ is cuda orb , implemented by opencv @xcite .
as we later discuss , due to the nature of gpu processing , the run time advantage of orb over the more accurate latch descriptor when computed on the cpu , vanishes on the gpu .
* deep features . * following the remarkable success of deep learning in computer vision , it is no surprise that these methods are also being applied to feature point description .
convolutional neural networks ( cnn ) were used in a number of previous attempts to learn local descriptor representations @xcite . in most cases ,
a siamese deep network is trained with hinge loss @xcite .
the training set used in these cases consists of positive and negative labeled patch pairs .
metric learning is then used to improve matching .
finally , @xcite proposed an efficient cnn design , bringing processing speeds down substantially .
as we later show , their run time is still slower than our proposed approach .
* the latch feature descriptor .
* latch was recently introduced in @xcite and is available as part of the opencv library since ver . 3.0 @xcite .
its design was inspired by the observation that _ pure _ binary descriptors such as brief and orb produce their values by comparing pairs of pixel intensities , a process which can be sensitive to local noise . to address this , these methods used various smoothing techniques before pixel values were compared .
smoothing , however , has the adverse effect of losing important high frequency image information . rather than smoothing the image and then comparing single pixel values ,
latch computes its binary values by comparing pixel _
patches_. the latch descriptor for image pixel @xmath0 is computed by selecting @xmath1 patch triplets , one for each latch bit . for triplet @xmath2 ,
three pixels are selected in the region around @xmath3 : an _ anchor _
pixel @xmath4 and two _ companion _ pixels @xmath5 and @xmath6 .
the @xmath7 pixel patches , @xmath8 , and @xmath9 centered on each of these three pixels are extracted .
finally , bit @xmath2 in the latch descriptor for @xmath3 is set by comparing the frobenious norm of the anchor to its two neighbors , as follows : @xmath10 the triplets latch uses are fixed but are not arbitrary : triplets are selected during training using the data set from @xcite , which contains same / not - same labeled image windows .
triplets were chosen by considering how well their bits correctly predicted the same / not - same labels over the entire training set . to prevent choosing correlated triplets , following @xcite
, triplets are skipped if their predictions are correlated with those of previously chosen triplets .
in their work @xcite , latch contained 512 bits ( selected triplets ) each one representing triplets of @xmath11 patches .
at matching time , its computational requirements were obviously equal to those of any other 512 bit binary descriptor . due to the use of patches and multiple frobenious norms , extracting latch was slower than pure binary descriptors of the same size .
experiments reported in @xcite , however , showed that the increase in extraction time was small .
this was balanced by improved accuracy which bested existing binary descriptors , sometimes rivaling even larger floating point representations . *
the gpu architecture and non - blocking programs . *
though the specific architectural designs of gpu processors changes from generation to generation , all have several multiprocessors .
a cpu can launch _ non - blocking _ ( parallel ) gpu programs on these multiprocessors , referred to as _
kernels_. that is , while a kernel is being executed on the gpu , the cpu is free to pursue other tasks and similarly , memory transfers to and from the gpu can take place without blocking either cpu or gpu .
this property is extremely important when designing computer vision systems using the gpu : it implies that if the gpu extracts descriptors independently of the cpu , _ the cpu is free to perform higher level processing_. related to the sfm application considered here are optimizations for recovering transformations @xcite and/or multiple view stereo for scene structure @xcite . some previous attempts to port descriptors to the gpu used it only for parts of the descriptor extraction process , using the cpu for others and requiring multiple memory transfers between processors @xcite .
this at least partially explains why these attempts showed only modest run time improvements over their original , cpu implementations . as a design goal
, we therefore limit the use of the cpu and any communications between it and the gpu when extracting and comparing our descriptors .
* why latch ? * latch was selected for following reasons . * * memory access vs. computation . *
the emphasis in gpu processing on raw arithmetic power results in memory access patterns often being the determining factor in performance rather than the actual computation .
latch requires more processing than pure binary representations ( e.g. , @xcite ) and therefore requires more cpu time to compute than they do . the memory transfer requirements of latch , however , are very similar to these other descriptors and hence it stands to gain more on the gpu . * * limited conditional branching . *
as mentioned above , gpus are optimized for processes which have few , if any , conditional branching ; under these circumstances , modern gpus are capable of up to 10 tera - flops .
most pure binary descriptors are therefore well suited for gpu processing , whereas porting more complex representations to the gpu is less trivial .
* * binary string comparisons .
* latch is a binary representation . like other binary representations
, it can be matched using fast hamming distance comparisons .
these can further be performed extremely fast on the gpu . finally , as demonstrated in the tests reported by @xcite
, latch outperforms other binary descriptors making it ideally suited for our purposes .
we have ported the latch representation to cuda 8 , building on the original latch opencv c++ implementation . in all our evaluations ,
clatch representations were extracted from @xmath12 pixel windows , using mini - patches of @xmath13 pixels giving a 64-byte feature vector .
to minimize cpu processing , differently from @xcite , we use the features from accelerated segment test ( fast ) @xcite feature detector .
fast is already available on the gpu as part of the opencv @xcite library . given a detected oriented keypoint , @xmath14 we extract latch from a @xmath15 intensities window around this point .
this process is described next .
* parallelizing latch on the gpu .
* gpu kernels include several identical , concurrently - executing , non - interacting _ blocks_. each one consists of groups ( _ warps _ ) of 32 threads . in our implementation ,
a clatch kernel sequentially computes 16 descriptors per block .
while the region of interest for one interest point is being processed , the next one is prefetched to pipeline the processing .
a single descriptor is extracted by multiple warps in each block .
each warp independently computes sixteen patch triplets , @xmath16 $ ] , four at a time , without any explicit synchronization during the main computation .
all told , two blocks of 32 warps , each one containing 32 threads ( total of 2048 threads ) are processed at a time per multiprocessor .
this coarse granularity was chosen to maximize performance across a variety of gpu architecture generations .
* memory optimization . *
given the fast orientation for an image region , the rotated @xmath15 pixel rectangle is loaded into shared memory as an upright square of single - precision floats .
we use texture memory accesses to efficiently load and process these values .
our implementation eliminates bank conflicts , with warp divergences or branches kept to a minimum .
thus , processing proceeds without ` i`f statements or communications between different warps .
this is achieved by strided access patterns of patches and careful padding of shared memory , and is critical to clatch s high performance .
specifically , patch comparisons are performed as follows .
a warp simultaneously processes four triplets . each thread ( in a warp of 32 threads ) performs two squared - distance comparisons per triplets in the f - norm of eq .
[ eq : latch ] . then
, fast warp shuffle operations are used to quickly sum the result from all pixel pairs in a novel , optimal manner . the original latch implementation used @xmath11 pixel patches .
we use @xmath17 patches instead as this implies 64 values which can be handled concurrently with no extra computation costs . to further optimize this process ,
instruction level parallelism was exploited by manual loop unrolling and carefully arranging operations to prevent stalls due to data dependency . * weighing pixels in latch patches . *
each pixel within a patch can optionally be given a unique weight at no overhead .
this is due to the gpu s emphasis on cheap fused - multiply - add operations .
we use this property to simulate the original latch patch size of @xmath11 by setting the relevant weights to zero , obtaining the exact same representation as the original latch . another potential use for this feature ,
not tested here , is applying gaussian weights to patch pixels thereby better emphasizing similarity at the patch center vs. its outer pixels .
latch ( and consequently clatch ) were shown to be slightly less accurate than some of the more computationally heavy , floating point descriptors .
it is not clear , however , how these differences in accuracy affect the overall accuracy and speed of an entire , descriptor - intensive computer vision system . to this end
, we test clatch vs. other descriptors on the challenging task of 3d sfm reconstruction .
our goal is to see how the final reconstruction and the time required to compute it are affected by the choice of descriptor .
we use the openmvg , multiple view geometry library @xcite , modifying it to include self - contained cuda streams and a gpu based , brute force hamming matcher .
these are detailed next .
* cuda integration . *
the kernel launching mechanism employed by cuda on its default stream disables concurrent launches of feature detection kernels .
we therefore modified the clatch descriptor and matching code to exclusively operate off of self - contained streams .
doing so allowed the gpu to concurrently execute feature detection and description kernels across multiple images at once , as well as perform feature matching .
* descriptor matching on the gpu . *
our tests compare the use of our binary descriptors with existing floating point representations . in all cases
, we used the gpu to compute the descriptor distances . because clatch is a binary representation , hamming distance is used to compute similarity of clatch descriptors .
to this end , we developed our own gpu based hamming distance brute force matcher and integrated it into openmvg . to provide a fair comparison , distances between floating point representations
were computed using the standard opencv gpu based l2 distance routine .
each block of our hamming - based brute force matching kernel processes half a probe descriptor per thread , though each descriptor is distributed throughout a half warp so that each thread holds parts of 16 probe descriptors .
gallery descriptors are alternatively prefetched into and processed from two shared memory buffers without intermediate synchronization . as the hamming distance between each pair of probe and gallery descriptors
is computed , partial results are distributed through a half warp .
this calls for a simultaneous reduction of several independent variables , which minimizes the number of additions and warp shuffles to be performed .
first , each thread halves the number of variables it must reduce by packing two variables into the lower and upper 16 bits of a 32 bit integer .
then , pairs of threads simultaneously exchange their packed variables in a warp shuffle , and sum the result with their original variable .
this results in pairs of threads with variables holding identical values .
the threads again pair off in the same manner , but exchange and sum a different variable .
the same pairs of threads now have two variables with identical values , so the second of each pair of threads overwrites the first packed variable with the second , before each thread discards the second packed variable .
this results in every thread in a warp having a unique value in the same variable , which allows efficient participation in subsequent warp shuffles until the reduction is complete .
this novel method requires only 16 additions to compare 16 descriptor pairs , while the standard warp reduction pattern would require 80 .
* descriptor extraction run time comparison . * the clatch descriptor is identical to latch and so their accuracy on different benchmarks are the same .
we therefore refer to the original paper for a comparison on standard benchmarks of latch and other representations @xcite . by using the gpu ,
clatch is much faster to extract .
this is demonstrated in table [ tab : runningtimes ] , which provides a comparison of the run times reported for extracting many popular existing feature point descriptors compared to clatch .
we report also the processor used to extract these representations and a price estimate for the processor in case of gpu based methods .
evident from the table is that even on affordable gpu hardware , extraction run times are orders of magnitude faster than standard representations on the cpu and even other gpu representations ( the only exceptions are the far less accurate cuda orb and the floating point representation cuda surf ) .
pn - net @xcite in particular , is designed to be a very fast deep learning based descriptor method , yet even with more expensive gpu hardware , it is more than an order of magnitude slower to extract than clatch .
more importantly , all floating point representations , including cuda surf and pn - net , require more time to match their bigger , real valued representations .
max width=0.55 l @ c c descriptor & extraction @xmath18 & gpu + sift @xcite & 3290 & - + surf @xcite & 2110 & - + cuda surf @xcite@xmath19 & 0.9 & gtx 970 m
( $ 280 usd ) + lda - hash @xcite & 5030 & - + lda - dif @xcite & 4740 & - + dbrief @xcite & 8750 & - + binboost @xcite & 3290 & - + brief @xcite & 234 & - + orb @xcite & 486 & - + cuda orb @xcite@xmath20 & 0.5 & gtx 970 m ( $ 280 usd ) + brisk @xcite & 59 & - + freak @xcite & 72 & - + a - kaze @xcite & 69 & - + latch @xcite & 616 & - + deepsiam @xcite@xmath21 & 6580 & titan ( $ 650 usd ) + matchnet @xcite@xmath22 & 575 & titan x ( $ 1,000 usd ) + cnn3 @xcite@xmath23 & 760 & titan black ( $ 1,100 usd ) + pn - net @xcite@xmath23 & 10 & titan x ( $ 1,000 usd ) + our clatch & 0.5 & gtx 970 m ( $ 280 usd ) + * sfm results .
* we use the incremental sfm pipeline implemented in openmvg , with its default values unchanged .
we compared the following descriptors in our tests : sift @xcite , often the standard in these applications , the deep learning based features , deepsiam and deepsiam2stream from @xcite , the fast deep feature representations , pn - net from @xcite and our own clatch .
max width= lcccccccc & avignon & bouteville & burgos & cognac garden & st .
jacques & mirebeau & murato & poitiers + number of images & 11 & 26 & 9 & 12 & 20 & 22 & 43 & 33 + + sift @xcite & 0.475 & 0.405 & 0.495 & 0.438 & 0.498 & 0.478 & 0.533 & 0.690 + deepsiam @xcite & 0.533 & 0.489 & 0.422 & 0.566 & 0.535 & 0.489 & 0.533 & 0.547 + deepsiam2stream @xcite & 0.505 & 0.457 & 0.419 & 0.536 & 0.522 & 0.459 & 0.496 & 0.529 + pn - net @xcite & 0.538 & 0.462 & 0.493 & 0.554 & 0.536 & 0.482 & 0.531 & 0.533 + our clatch & 0.556 & 0.414 & 0.466 & 0.478 & 0.409 & 0.466 & 0.494 & 0.454 + + sift @xcite & 174.30 & 454.13 & 143.728 & 155.61 & 296.431 & 401.64 & 958.778 & 1206.10 + deepsiam @xcite & 172.49 & 596.07 & 130.72 & 146.49 & 347.576 & 416.01 & 943.841 & 812.39 + deepsiam2stream @xcite & 269.39 & 922.95 & 226.123 & 301.67 & 628.03 & 739.629 & 1750.535 & 1379.80 + pn - net @xcite & 49.56 & 210.60 & 50.12 & 56.02 & 122.29 & 167.51 & 372.28 & 311.18 + our clatch & 18.91 & 69.07 & 15.907 & 19.089 & 27.877 & 47.534 & 86.377 & 61.868 + all descriptors used the cuda fast feature detector with the exception of sift which , for technical reasons , used its default dog based detector . following incremental sfm , point cloud densification @xcite , mesh reconstruction @xcite and mesh refinement @xcite were applied to produce the final reconstructions visualized in fig .
[ fig : reconstruct ] .
tests were performed on publicly available sets of high resolution photogrammetry images from @xcite , which include @xmath24 ( or @xmath25 ) pixels in each image .
table [ tab : sfmresults ] summarizes these results , providing the final scene reprojection rmse and the total time for descriptor extraction , matching and sfm reconstruction .
all these tests were run on our gtx 1080 gpu .
reconstruction run time is dominated by the brute force , nearest neighbor matcher .
hence , the gaps in run times between the different methods are smaller than those in table [ tab : runningtimes ] .
nevertheless , reconstructions with clatch required a fraction of the time for the runner up ( pn - net ) and far less than the others .
reprojection rmse , is low for all methods and is typically around half a pixel .
although these errors fluctuate between the different methods and scenes , these differences are often below 0.1 pixels .
considering the high resolutions of the input images , these differences are negligible . finally , fig .
[ fig : reconstruct ] additionally provides qualitative results , showing rendered views of our reconstructions .
evident from the figure is that despite large differences in run time , qualitatively , the reconstructions appear very similar .
in descriptor matching intensive application , such as sfm , accuracy per descriptor is sometimes balanced by the speed required to extract and match the descriptors . taking advantage of this
, we present clatch , a cuda port for the latch binary descriptor .
although clatch descriptor accuracy in standard benchmarks may fall slightly behind other representations , particularly recent deep learning based methods , they are far faster to extract and match .
clatch thereby provides a fast and accurate alternative means for 3d reconstruction . from a technical point , an outcome of this work is openly available code for extremely fast feature extraction and matching and a pipeline for sfm allowing convenient interchange of feature descriptors , including deep methods
this implementation can be improved in many ways .
for one thing , our use of the fast detector @xcite does not provide scale invariance .
clatch , however , can easily be extracted at multiple scales , potentially improving its accuracy .
use of clatch in other applications where descriptors are extracted and matched in large quantities , is also a priority .
one particularly appealing example is dense pixel matching @xcite where the clatch may be an alternative to methods such as patchmatch @xcite , providing similar run times without compromising spatial smoothness .
han , x. , leung , t. , jia , y. , sukthankar , r. , berg , a.c .
: matchnet : unifying feature and metric learning for patch - based matching . in : proc .
comput . vision pattern recognition .
32793286 ( 2015 ) simo - serra , e. , trulls , e. , ferraz , l. , kokkinos , i. , fua , p. , moreno - noguer , f. : discriminative learning of deep convolutional feature point descriptors . in : proc .
comput . vision .
118126 ( 2015 ) yang , x. , cheng , k.t .
: ldb : an ultra - fast feature for scalable augmented reality on mobile devices . in : mixed and augmented reality ( ismar ) , 2012 ieee international symposium on .
4957 . ieee ( 2012 ) | accuracy , descriptor size , and the time required for extraction and matching are all important factors when selecting local image descriptors . to optimize over all these requirements
, this paper presents a cuda port for the recent learned arrangement of three patches ( latch ) binary descriptors to the gpu platform .
the design of latch makes it well suited for gpu processing . owing to its small size and binary nature
, the gpu can further be used to efficiently match latch features . taken together , this leads to breakneck descriptor extraction and matching speeds .
we evaluate the trade off between these speeds and the quality of results in a feature matching intensive application . to this end
, we use our proposed cuda latch ( clatch ) to recover structure from motion ( sfm ) , comparing 3d reconstructions and speed using different representations .
our results show that clatch provides high quality 3d reconstructions at fractions of the time required by other representations , with little , if any , loss of reconstruction quality . |
after the discovery of a signal with a mass of about 125 gev in the higgs searches at the lhc @xcite , the prime goal is now to identify the underlying nature of the new state and to determine the mechanism of electroweak symmetry breaking . while the properties of the observed state are compatible with the ones predicted for the higgs boson of the standard model ( sm ) at the current level of precision , also a wide range of alternative interpretations could be possible , corresponding to very different underlying physics .
in particular , in models with an extended higgs sector the observed state would be accompanied by several other higgs bosons , in contrast to the minimal formulation of the sm where a single @xmath0-doublet is responsible for electroweak - symmetry breaking .
supersymmetry ( susy ) @xcite is commonly regarded as the most appealing extension of the sm , since it provides a solution for stabilising the huge hierarchy between the planck scale and the weak scale @xcite and offers further attractive features such as unification of the gauge couplings and a natural candidate for cold dark matter in the universe .
a crucial prediction of supersymmetric extensions of the sm are their extended higgs sectors : the holomorphicity of the superpotential ( as well as the cancellation of gauge anomalies ) implies that at least two @xmath0 doublets with opposite hypercharge have to be present , so as to generate mass terms for both up- and down - quarks ( in a type ii 2-higgs - doublet - model fashion ) .
the minimal supersymmetric extension of the sm ( mssm ) @xcite is based on the minimal higgs sector of this kind comprising two higgs doublets , whereas the higgs sector of the next - to - minimal supersymmetric extension of the sm ( nmssm ) @xcite contains an additional ( complex ) gauge - singlet .
it has long been recognised that the nmssm provides an elegant solution @xcite to the `` @xmath1-problem '' @xcite of the mssm . in the context of the relatively high mass value of about 125 gev of the observed state ,
this model has received particular attention lately since the mass of the light doublet - like state receives an additional contribution at lowest order as compared to the mssm , which dominates at low values of @xmath2 ( the ratio of the vacuum expectation values , v.e.v.s , of the two higgs doublets ) . in this case
significantly smaller higher - order corrections are required to obtain a higgs - boson mass in the appropriate range @xcite as compared to the mssm , where the lowest - order prediction for the mass of the light cp - even higgs boson is bounded from above by the mass of the z boson , @xmath3 .
furthermore , also the singlet doublet mixing can give rise to an uplift of the mass of the doublet state , provided that the cp - even singlet state is lighter than the doublet state .
it has been argued in this context that the relaxed requirement on the size of the higher - order corrections as compared to the mssm makes it possible to obtain a higgs - mass prediction of about 125 gev in a `` more natural '' way @xcite . in the following
we will focus on the nmssm as a theoretically well - motivated alternative to the sm with a potentially rich phenomenology in the higgs sector . for simplicity
, we will restrict to the cp - conserving case , for which the spectrum of physical higgs states of the nmssm consists of three cp - even , two cp - odd and a pair of charged higgs states ( while we do not explicitly consider cp - violating effects giving rise to a mixture between the five neutral states , it should be noted that cases where a cp - even and a cp - odd state are nearly mass - degenerate essentially mimick a scenario where a single state is an admixture of cp - even and cp - odd components ) .
furthermore , while several versions of the nmssm can be formulated , depending on the form of the singlet and singlet - doublet interaction terms in the superpotential , we will focus on the @xmath4-conserving version only , where the solution to the `` @xmath1-problem '' is more immediate ( on the other hand , this simple model could lead to a domain wall problem @xcite but we will not address this question here ) .
the higgs sector of this version of the nmssm is characterised by six parameters ( at tree - level ) , in contrast to the two parameters of the mssm .
while we shall borrow most of our notations from @xcite , we recall the higgs terms entering the superpotential of the model : @xmath5 where @xmath6 denotes the singlet ( super)field , @xmath7 and @xmath8 the doublets , while @xmath9 stands for the @xmath10 product . when confronting the predictions of an extended higgs sector with the observed signal and the limits from the higgs searches at lep , the tevatron and the lhc , the most obvious interpretation of the signal at about 125 gev is to associate it with the lightest cp - even higgs boson of the considered model .
the case where all other higgs ( as well as all new physics ) states of a supersymmetric extension of the sm ( and the same is true for various other extended higgs sectors ) are significantly heavier corresponds to the `` decoupling region '' of the model under consideration , where the couplings of the light higgs boson to gauge bosons and sm fermions are very close to the ones of the sm . revealing deviations of those couplings from their sm counterparts in such a case will require high - precision measurements , where in many cases the expected deviations do not exceed the level of a few per cent .
an additional source of possible deviations from the sm could be decays of the sm - like state into new - physics particles .
such a decay could in particular occur into a pair of dark matter particles , if the mass of the latter is less than half of the mass of the higgs state , i.e. below about 60 gev .
this would give rise to an invisible decay mode of the observed state , providing a strong motivation for searches of decays of the observed signal into invisible final states .
besides the interpretation of the observed state as the lightest cp - even higgs boson of an extended higgs sector , it is also possible , at least in principle , to identify the observed signal with the second - lightest state of an extended higgs sector .
this interpretation would have the immediate consequence that there should be an ( or more generally at least one ) additional higgs state _ below _
the one observed at about 125 gev .
the phenomenology of such a scenario is very different from the case of the decoupling limit discussed above , because of the presence of at least one more light state in the spectrum . within the mssm
this interpretation is in principle possible @xcite , but gives rise to a rather exotic higgs sector where in fact all additional higgs bosons are light , i.e.in the vicinity of the state at about 125 gev or below .
it is remarkable that a global fit within the mssm within this interpretation has resulted in an acceptable fit probability @xcite , but lately this interpretation , which implies in particular a light charged higgs boson below the mass of the top quark ( see in particular ref .
@xcite ) , has come under increased pressure from the limits obtained in the charged higgs searches by atlas @xcite and , more recently , cms @xcite .
the nmssm provides a well - suited and theoretically well motivated framework for investigating to what extent interpretations that go beyond the obvious case of a single light state in the decoupling limit are compatible with the latest experimental results both with respect to the properties of the observed state and to the limits obtained from higgs searches ( as well as other existing constraints ) .
it is the purpose of the present paper to perform such an analysis .
it is obvious that compatibility with the observed signal requires much more than just a higgs state ( or possibly more than one ) in the spectrum with a mass of about 125 gev . in order to properly take into account the latest experimental results from higgs search limits and from measurements of the properties of the observed state , we make use of the public tools higgsbounds @xcite and higgssignals @xcite , which incorporate a comprehensive set of results from atlas @xcite , cms @xcite and the tevatron @xcite .
we do not explicitly impose limits from the direct searches for susy particles at the lhc .
while some of the scenarios discussed in this paper could be affected by constraints from susy particle searches , we have checked that the qualitative features of the higgs phenomenology of those scenarios are maintained also for somewhat heavier susy particle spectra .
note however that most of the susy spectra that we employ ( especially for coloured particles ) are beyond the mass - range tested in the run - i of the lhc .
as mentioned above , already since the very early hints of a signal at about 125 gev the higgs sector of the nmssm has found a lot of attention in this context . besides the mass prediction in comparison with the mssm case @xcite ,
the possibility of modified rates has been discussed , particularly in the diphoton channel @xcite .
furthermore , the case of universal ( or semi - universal ) susy - breaking conditions at the gut scale @xcite , gauge - mediation @xcite and other related scenarios @xcite have been considered in this context .
the cp - violating version of the nmssm also received some attention @xcite .
@xcite analysed the fine - tuning in a @xmath4-violating version of the nmssm and variants .
other groups confronted the presence of a higgs state at this mass with direct searches for susy particles at the lhc or dark - matter constraints @xcite .
scenarios with a light singlet - like state around @xmath11 gev have found considerable interest @xcite .
another possibility involves a singlet and a doublet that are almost mass - degenerate at about 125 gev and may mix with each other , see ref .
@xcite ( and the last reference of @xcite ) .
several studies also suggested to exploit pair production processes at the lhc in order to distinguish the sm from the nmssm and/or to look for a light singlet in this fashion @xcite .
scenarios with a very light cp - odd ( or cp - even ) higgs boson were addressed with several search proposals in direct production , unconventional light charged - higgs decays , or cascade decays from sm - like / light singlet states ; large higgs - to - higgs decays were also considered from the point of view of the sm - compatible nature of the observed state @xcite .
recent studies of the properties of a light pseudoscalar in the nmssm @xcite have emphasized the relevance of indirect production modes for the investigation of this scenario at the lhc . in a different direction ,
the authors of @xcite focussed on nmssm higgs scenarios with a low - scale doublet sector .
furthermore , @xcite , and more recently @xcite , studied the discovery prospects of nmssm higgs states in the lhc run at @xmath12 tev . in our analysis
we go beyond the previous work in several respects : while many of the afore - mentioned analyses discussed scenarios which are compatible with existing limits , our inclusion of a fitting tool allows us to highlight the quality of the various scenarios in view of the available data .
furthermore , we aim at a comprehensive discussion from the point of view of the nmssm higgs phenomenology , hence do not confine to a specific scenario ( within our assumptions on the model , perturbativity constraints and choices of simplicity with regards to the susy spectrum ) .
we also focus on higgs physics and thus try , without spoiling the physical content , to avoid emphasis on questions of secondary importance with respect to this topic ( e.g. the details of the supersymmetric spectrum ) .
finally , much experimental data has become available in the last few years , narrowing the possibilities in the higgs sector , and most of the recent developments are included within the tools on which our discussion is based .
the paper is organised as follows : in sect .
2 we describe the framework that we use for the analyses in this paper , in particular the statistical approach used in our fits , the treatment of external constraints and the tools that we apply . as a first step of our analysis , in sect . 3
we briefly consider the sm case and the corresponding decoupling limits of the mssm and the nmssm .
the sm result is used for comparison with @xmath13 analyses in different nmssm scenarios , which we perform in sects .
[ lsing][ldoub ] . in sect .
[ fp ] , we focus on specific points of the nmssm parameter space and discuss in more details the higgs phenomenology and the consequences for future searches , should the corresponding spectrum be realised in nature . in sect . [ gs ] a more global scan is carried out , and the features observed in the global scan are discussed in view of the results obtained for the specific nmssm scenarios that we have considered before .
[ conc ] contains our conclusions .
the nmssm parameter space is explored with the help of the spectrum generator ` nmssmtools_4.4.0 ` @xcite , computing the higgs masses up to leading two - loop double - log order ( we will be using the default mode only ) , in an effective potential approach .
this code considers a certain number of phenomenological limits , several among which are kept within our analysis .
the first class of such tests are consistency requirements and are ( necessarily ) included as hard cuts : * stability of the ewsb - vacuum : positivity of the scalar squared - masses , absence of deeper minimum ; * absence of landau poles below the gut scale : while this requirement is sometimes omitted in order to probe effects associated with large values of the parameter @xmath14 and under the assumption that new - physics or specific properties of the non - perturbative regime would smoothen the theoretical difficulty of the landau poles , we choose to keep this theoretical limit ; * requirement for higgs soft squared - masses at the tev scale : the potential - minimization procedure in nmssmtools trades these masses for the higgs v.e.v.s , so that the naturalness requirement that soft masses intervene at the tev scale must be checked explicitly ; * requirement for a neutralino lsp ( the impact of which , however , is of secondary importance in our discussion ) .
* @xmath15 decay ( @xmath16 ) * mass lower limits on squarks ( @xmath17 gev , @xmath18 gev ) , gluino ( @xmath19 gev ) , sleptons ( @xmath20 gev ) , charginos ( @xmath21 gev ) ; * limits on @xmath22 , @xmath23 , @xmath24 .
we remind the reader that lhc limits on susy searches are not considered in our analysis
. however , the susy spectra that we employ are typically beyond the mass - range of the searches in the run - i . in this context
, the inclusion of lep limits as mentioned above has only a minor impact . * limits from the bottomonium sector : non - observation of a signal in @xmath25 , excessive contribution to the @xmath26 mixing @xcite . only light cp - odd higgs below @xmath27 gev
are concerned by these constraints and the limits are kept as a @xmath28 c.l .
cut . * limits from @xmath29-factories ( under a strong minimal flavour violation hypothesis , i.e. neglecting all possible tree - level flavour - changing neutral currents ) : @xmath30 , @xmath31 , @xmath32 , @xmath33 , @xmath34 @xcite . instead of treating the limits as a hard cut ,
we combine them in a @xmath13 function relying on the central value and error bars computed in nmssmtools : @xmath35 the corresponding experimental central values @xmath36 and standard deviation @xmath37 are summarized in table [ expblim ] .
the theoretical error estimate is the result of an involved calculation : errors relative to sm - like contributions are taken from the corresponding sm estimate ; the uncertainty on new - physics contributions is estimated to @xmath38 ( if only leading - order effects are included ) / @xmath39 ( if next - to - leading @xmath40 corrections are present ) of the total corresponding contributions and are added linearly to the sm error ; additional error sources are mostly ckm matrix elements ( taken from tree - level measurements exclusively ) and hadronic parameters ( decay constants , taken from lattice calculations ) ; to obtain the final theoretical uncertainty range , both higher - order and parametric uncertainties are varied within these previously - discussed limits .
* @xmath41 @xcite : similarly to @xmath29-observables , we add a contribution to the @xmath13 with experimental - sm input shown in table [ expblim ] , where the errors have been added in quadrature . the theoretical uncertainty associated to new - physics contributions and higher orders
is calculated as the sum of a fixed error @xmath42 , a @xmath43 error estimate on 1-loop contributions ( which do not involve coloured particles ) and a @xmath38 error estimate on 2-loop effects ( involving coloured particles ) .
additionally , given that a candidate for the interpretation of the signal observed at the lhc seems necessary , we require that the spectrum produces one cp - even higgs state in the mass - range @xmath44 $ ] gev .
* dark matter searches : relic - density ( via micromegas ) , xenon 100 @xcite .
the reasons for not taking such limits into account come from the observation that they are strongly dependent on the susy spectrum , while we want to focus on the higgs sector : confining to collider constraints allows us to handle simple supersymmetric spectra , which play a secondary part in our analysis , while these would likely have to be finely adjusted if one were to include , e.g. , the relic - density bounds .
we note also that the dark matter phenomenology may involve mechanisms ( e.g. light gravitino lsp ) which may alter the conclusions in the dark sector , while all such considerations are not the focus of our discussion . *
lep higgs searches : @xmath45 , @xmath46 , @xmath47 , @xmath48 ( z width ) ; * tevatron limits on @xmath49 , @xmath50 ; * lhc higgs limits : @xmath49 , @xmath51 $ ] gev , effective @xmath52 , @xmath53 , @xmath54 couplings excessive @xcite . 99 g. aad _ et al .
_ [ atlas collaboration ] , phys .
b * 716 * ( 2012 ) 1 [ arxiv:1207.7214 [ hep - ex ] ] .
s. chatrchyan _ et al . _
[ cms collaboration ] , phys .
b * 716 * ( 2012 ) 30 [ arxiv:1207.7235 [ hep - ex ] ] .
h. p. nilles , `` supersymmetry , supergravity and particle physics , '' phys .
( 1984 ) 1 .
e. witten , _ nucl .
phys . _ * b188 * ( 1981 ) 513 .
+ s. dimopoulos and h. georgi , _ nucl .
phys . _ * b193 * ( 1981 ) 150 .
+ e. witten , _ phys .
_ * b105 * ( 1981 ) 267 .
+ r. k. kaul and p. majumdar , _ nucl .
phys . _ * b199 * ( 1982 ) 36
. + n. sakai , _
z. phys .
_ * c11 * ( 1981 ) 153 .
martin , in * kane , g.l .
( ed . ) : perspectives on supersymmetry ii * 1 - 153 [ hep - ph/9709356 ] .
+ d. j. h. chung , l. l. everett , g. l. kane , s. f. king , j. d. lykken and l. -t .
wang , phys .
* 407 * ( 2005 ) 1 [ hep - ph/0312378 ] .
u. ellwanger , c. hugonie and a. m. teixeira , phys .
* 496 * ( 2010 ) 1 [ arxiv:0910.1785 [ hep - ph ] ] . m. maniatis , int .
j. mod .
a * 25 * ( 2010 ) 3505 [ arxiv:0906.0777 [ hep - ph ] ] .
fayet , _ nucl .
phys . _ * b90 * ( 1975 ) 104 . + p.
fayet , _ phys .
lett _ * b69 * ( 1977 ) 489 .
+ p. fayet , s. ferrara , _ phys .
rept._*32 * ( 1977 ) 249 .
j. e. kim and h. p. nilles , phys .
b * 138 * ( 1984 ) 150 .
u. ellwanger and c. hugonie , mod .
a * 22 * ( 2007 ) 1581 [ hep - ph/0612133 ] .
l. j. hall , d. pinner and j. t. ruderman , jhep * 1204 * ( 2012 ) 131 [ arxiv:1112.2703 [ hep - ph ] ] .
+ a. arvanitaki and g. villadoro , jhep * 1202 * ( 2012 ) 144 [ arxiv:1112.4835 [ hep - ph ] ] .
+ s. f. king , m. muhlleitner and r. nevzorov , nucl .
b * 860 * ( 2012 ) 207 [ arxiv:1201.2671 [ hep - ph ] ]
. + z. kang , j. li and t. li , jhep * 1211 * ( 2012 ) 024 [ arxiv:1201.5305 [ hep - ph ] ] .
+ j. cao , z. heng , j. m. yang and j. zhu , jhep * 1210 * ( 2012 ) 079 [ arxiv:1207.3698 [ hep - ph ] ] .
+ k. agashe , y. cui and r. franceschini , jhep * 1302 * ( 2013 ) 031 [ arxiv:1209.2115 [ hep - ph ] ] .
+ r. barbieri , d. buttazzo , k. kannike , f. sala and a. tesi , phys .
d * 87 * ( 2013 ) 115018 [ arxiv:1304.3670 [ hep - ph ] ] .
+ r. barbieri , d. buttazzo , k. kannike , f. sala and a. tesi , phys .
d * 88 * ( 2013 ) 055011 doi:10.1103/physrevd.88.055011 [ arxiv:1307.4937 [ hep - ph ] ] .
s. a. abel , s. sarkar and p. l. white , nucl .
b * 454 * ( 1995 ) 663 [ hep - ph/9506359 ] .
s. heinemeyer , o. stal and g. weiglein , phys .
b * 710 * ( 2012 ) 201 [ arxiv:1112.3026 [ hep - ph ] ] .
+ a. bottino , n. fornengo and s. scopel , phys .
d * 85 * ( 2012 ) 095013 [ arxiv:1112.5666 [ hep - ph ] ]
. + m. drees , phys .
d * 86 * ( 2012 ) 115018 [ arxiv:1210.6507 [ hep - ph ] ] .
+ k. hagiwara , j. s. lee and j. nakamura , jhep * 1210 * ( 2012 ) 002 [ arxiv:1207.0802 [ hep - ph ] ] .
+ a. arbey , m. battaglia , a. djouadi and f. mahmoudi , jhep * 1209 * ( 2012 ) 107 [ arxiv:1207.1348 [ hep - ph ] ] .
p. bechtle , s. heinemeyer , o. stal , t. stefaniak , g. weiglein and l. zeune , eur .
j. c * 73 * ( 2013 ) 2354 [ arxiv:1211.1955 [ hep - ph ] ] .
m. carena , s. heinemeyer , o. stl , c. e. m. wagner and g. weiglein , eur .
j. c * 73 * ( 2013 ) 2552 [ arxiv:1302.7033 [ hep - ph ] ] . the atlas collaboration , atlas - conf-2013 - 090 , atlas - com - conf-2013 - 107 .
p. bechtle , o. brein , s. heinemeyer , g. weiglein and k. e. williams , comput .
commun . * 181 * ( 2010 ) 138 [ arxiv:0811.4169 [ hep - ph ] ] .
+ p. bechtle , o. brein , s. heinemeyer , g. weiglein and k. e. williams , comput .
commun .
* 182 * ( 2011 ) 2605 [ arxiv:1102.1898 [ hep - ph ] ] .
+ p. bechtle , o. brein , s. heinemeyer , o. stal , t. stefaniak , g. weiglein and k. williams , pos charged * 2012 * ( 2012 ) 024 [ arxiv:1301.2345 [ hep - ph ] ] .
+ p. bechtle , o. brein , s. heinemeyer , o. stl , t. stefaniak , g. weiglein and k. e. williams , eur .
j. c * 74 * ( 2014 ) 2693 [ arxiv:1311.0055 [ hep - ph ] ] . + ` http://higgsbounds.hepforge.org/ ` t. aaltonen
_ et al . _ [ cdf and d0 collaborations ] , phys .
* 109 * ( 2012 ) 071804 [ arxiv:1207.6436 [ hep - ex ] ]
. t. aaltonen _ et al . _
[ cdf and d0 collaborations ] , phys .
d * 88 * ( 2013 ) 5 , 052014 [ arxiv:1303.6346 [ hep - ex ] ] .
u. ellwanger , jhep * 1203 * ( 2012 ) 044 [ arxiv:1112.3548 [ hep - ph ] ] .
+ j. -j .
cao , z. -x .
heng , j. m. yang , y. -m .
zhang and j. -y .
zhu , jhep * 1203 * ( 2012 ) 086 [ arxiv:1202.5821 [ hep - ph ] ] .
+ r. benbrik , m. gomez bock , s. heinemeyer , o. stal , g. weiglein and l. zeune , eur .
j. c * 72 * ( 2012 ) 2171 [ arxiv:1207.1096 [ hep - ph ] ] .
+ z. heng , adv .
high energy phys .
* 2012 * ( 2012 ) 312719 [ arxiv:1210.3751 [ hep - ph ] ] .
+ k. choi , s. h. i m , k. s. jeong and m. yamaguchi , jhep * 1302 * ( 2013 ) 090 [ arxiv:1211.0875 [ hep - ph ] ] .
+ s. f. king , m. mhlleitner , r. nevzorov and k. walz , nucl .
b * 870 * ( 2013 ) 323 [ arxiv:1211.5074 [ hep - ph ] ]
. + j. w. fan _ et al .
_ , chin .
c * 38 * ( 2014 ) 073101 [ arxiv:1309.6394 [ hep - ph ] ] .
j. f. gunion , y. jiang and s. kraml , phys .
b * 710 * ( 2012 ) 454 [ arxiv:1201.0982 [ hep - ph ] ] .
+ u. ellwanger and c. hugonie , adv .
high energy phys .
* 2012 * ( 2012 ) 625389 [ arxiv:1203.5048 [ hep - ph ] ] .
+ k. kowalska _ et al . _ [ bayesfits group collaboration ] , phys .
d * 87 * ( 2013 ) 11 , 115010 [ arxiv:1211.1693 [ hep - ph ] ] .
+ c. beskidt , w. de boer and d. i. kazakov , phys . lett .
b * 726 * ( 2013 ) 758 [ arxiv:1308.1333 [ hep - ph ] ] .
+ c. beskidt , w. de boer and d. i. kazakov , arxiv:1402.4650 [ hep - ph ] .
+ u. ellwanger and c. hugonie , arxiv:1405.6647 [ hep - ph ] .
l. randall and m. reece , jhep * 1308 * ( 2013 ) 088 [ arxiv:1206.6540 [ hep - ph ] ] .
+ b. kyae and j. -c .
park , phys .
d * 87 * ( 2013 ) 075021 [ arxiv:1207.3126 [ hep - ph ] ] .
k. s. jeong , y. shoji and m. yamaguchi , jhep * 1209 * ( 2012 ) 007 [ arxiv:1205.2486 [ hep - ph ] ] .
+ k. j. bae , k. choi , e. j. chun , s. h. i m , c. b. park and c. s. shin , jhep * 1211 * ( 2012 ) 118 [ arxiv:1208.2555 [ hep - ph ] ] .
+ i. gogoladze , b. he and q. shafi , phys .
b * 718 * ( 2013 ) 1008 [ arxiv:1209.5984 [ hep - ph ] ] .
+ j. cao , d. li , l. shang , p. wu and y. zhang , arxiv:1409.8431 [ hep - ph ] .
s. moretti , s. munir and p. poulose , phys .
d * 89 * ( 2014 ) 1 , 015022 [ arxiv:1305.0166 [ hep - ph ] ] . + s. moretti and s. munir , arxiv:1505.00545 [ hep - ph ] .
t. gherghetta , b. von harling , a. d. medina and m. a. schmidt , jhep * 02 * ( 2013 ) 032 [ arxiv:1212.5243 [ hep - ph ] ] .
+ a. kaminska , g. g. ross and k. schmidt - hoberg , jhep * 1311 * ( 2013 ) 209 [ arxiv:1308.4168 [ hep - ph ] ] .
+ a. kaminska , g. g. ross , k. schmidt - hoberg and f. staub , jhep * 1406 * ( 2014 ) 153 [ arxiv:1401.1816 [ hep - ph ] ] .
d. a. vasquez , g. belanger , c. boehm , j. da silva , p. richardson and c. wymant , phys .
d * 86 * ( 2012 ) 035023 [ arxiv:1203.3446 [ hep - ph ] ] .
+ t. cheng , j. li , t. li and q. -s .
yan , arxiv:1304.3182 [ hep - ph ] .
+ t. cheng and t. li , phys .
d * 88 * ( 2013 ) 015031 [ arxiv:1305.3214 [ hep - ph ] ] .
g. belanger , u. ellwanger , j. f. gunion , y. jiang , s. kraml and j. h. schwarz , jhep * 1301 * ( 2013 ) 069 [ arxiv:1210.1976 [ hep - ph ] ] .
+ m. badziak , m. olechowski and s. pokorski , jhep * 1306 * ( 2013 ) 043 [ arxiv:1304.5437 [ hep - ph ] ] .
+ m. badziak , m. olechowski and s. pokorski , arxiv:1310.4518 [ hep - ph ] .
j. f. gunion , y. jiang and s. kraml , phys .
d * 86 * ( 2012 ) 071702 [ arxiv:1207.1545 [ hep - ph ] ] .
+ s. munir , l. roszkowski and s. trojanowski , phys .
d * 88 * ( 2013 ) 5 , 055017 [ arxiv:1305.0591 [ hep - ph ] ] .
z. kang , j. li , t. li , d. liu and j. shu , phys .
d * 88 * ( 2013 ) 1 , 015006 [ arxiv:1301.0453 [ hep - ph ] ] .
+ j. cao , z. heng , l. shang , p. wan and j. m. yang , jhep * 1304 * ( 2013 ) 134 [ arxiv:1301.6437 [ hep - ph ] ] .
+ d. t. nhung , m. muhlleitner , j. streicher and k. walz , jhep * 1311 * , 181 ( 2013 ) [ arxiv:1306.3926 [ hep - ph ] ] .
+ u. ellwanger , jhep * 1308 * ( 2013 ) 077 [ arxiv:1306.5541 [ hep - ph ] ] .
+ c. han , x. ji , l. wu , p. wu and j. m. yang , jhep * 1404 * ( 2014 ) 003 [ arxiv:1307.3790 [ hep - ph ] ] .
+ l. wu , j. m. yang , c. p. yuan and m. zhang , phys .
b * 747 * ( 2015 ) 378 [ arxiv:1504.06932 [ hep - ph ] ] .
m. m. almarashi and s. moretti , arxiv:1205.1683 [ hep - ph ] .
+ j. rathsman and t. rossler , adv .
high energy phys .
* 2012 * ( 2012 ) 853706 [ arxiv:1206.1470 [ hep - ph ] ] .
+ d. g. cerdeno , p. ghosh and c. b. park , jhep * 1306 * ( 2013 ) 031 [ arxiv:1301.1325 [ hep - ph ] ] .
+ j. cao , f. ding , c. han , j. m. yang and j. zhu , jhep * 1311 * ( 2013 ) 018 [ arxiv:1309.4939 [ hep - ph ] ] .
n. e. bomark , s. moretti , s. munir and l. roszkowski , arxiv:1409.8393 [ hep - ph ] .
+ n. e. bomark , s. moretti and l. roszkowski , arxiv:1503.04228 [ hep - ph ] . n. d. christensen , t. han , z. liu and s. su , jhep * 1308 * ( 2013 ) 019 [ arxiv:1303.2113 [ hep - ph ] ] .
s. f. king , m. muhlleitner , r. nevzorov and k. walz , arxiv:1408.1120 [ hep - ph ] .
d. buttazzo , f. sala and a. tesi , arxiv:1505.05488 [ hep - ph ] .
+ m. guchait and j. kumar , arxiv:1509.02452 [ hep - ph ] .
u. ellwanger , j. f. gunion and c. hugonie , jhep * 0502 * ( 2005 ) 066 [ hep - ph/0406215 ] .
+ u. ellwanger and c. hugonie , comput .
commun .
* 175 * ( 2006 ) 290 [ hep - ph/0508022 ] .
+ u. ellwanger , j. f. gunion , c. hugonie , _ jhep _ * 0502 * ( 2005 ) 066 , arxiv : hep - ph/0406215 + u. ellwanger , c. hugonie , _ comput .
* 175 * ( 2006 ) 290 , arxiv : hep - ph/0508022 + ` http://www.th.u-psud.fr/nmhdecay/nmssmtools.html ` f. domingo , u. ellwanger , e. fullana , c. hugonie and m. -a .
sanchis - lozano , jhep * 0901 * ( 2009 ) 061 [ arxiv:0810.4736 [ hep - ph ] ] .
+ f. domingo , u. ellwanger and m. -a .
sanchis - lozano , phys .
* 103 * ( 2009 ) 111802 [ arxiv:0907.0348 [ hep - ph ] ] .
+ f. domingo , jhep * 1104 * ( 2011 ) 016 [ arxiv:1010.4701 [ hep - ph ] ] .
f. domingo and u. ellwanger , jhep * 0712 * ( 2007 ) 090 [ arxiv:0710.3714 [ hep - ph ] ] .
f. domingo and u. ellwanger , jhep * 0807 * ( 2008 ) 079 [ arxiv:0806.0733 [ hep - ph ] ] . g. belanger , f. boudjema , c. hugonie , a. pukhov and a. semenov , jcap * 0509 * ( 2005 ) 001 [ hep - ph/0505142 ] . + g. belanger , f. boudjema , a. pukhov and a. semenov , comput .
commun .
* 185 * ( 2014 ) 960 [ arxiv:1305.0237 [ hep - ph ] ] .
g. belanger , b. dumont , u. ellwanger , j. f. gunion and s. kraml , phys .
d * 88 * ( 2013 ) 075008 [ arxiv:1306.2941 [ hep - ph ] ] .
r. barate _
[ lep working group for higgs boson searches and aleph and delphi and l3 and opal collaborations ] , phys .
b * 565 * ( 2003 ) 61 [ hep - ex/0306033 ] .
r. aaij _ et al . _
[ lhcb collaboration ] , jhep * 1305 * ( 2013 ) 132 [ arxiv:1304.2591 [ hep - ex ] ] .
g. aad _ et al .
_ [ atlas collaboration ] , arxiv:1407.6583 [ hep - ex ] .
u. ellwanger , phys .
b * 698 * ( 2011 ) 293 [ arxiv:1012.1201 [ hep - ph ] ] .
r. v. harlander , s. liebler and h. mantler , comput .
commun .
* 184 * ( 2013 ) 1605 [ arxiv:1212.3249 [ hep - ph ] ] .
+ s. liebler , eur .
j. c * 75 * ( 2015 ) 5 , 210 [ arxiv:1502.07972 [ hep - ph ] ] .
b. a. dobrescu and k. t. matchev , jhep * 0009 * ( 2000 ) 031 [ hep - ph/0008192 ] . o. stal and g. weiglein , jhep * 1201 * ( 2012 ) 071 [ arxiv:1108.0595 [ hep - ph ] ]
. g. abbiendi _ et al . _
[ opal collaboration ] , eur .
j. c * 27 * ( 2003 ) 311 [ hep - ex/0206022 ] .
g. belanger , b. dumont , u. ellwanger , j. f. gunion and s. kraml , phys .
b * 723 * ( 2013 ) 340 [ arxiv:1302.5694 [ hep - ph ] ] .
e. fuchs , s. thewes and g. weiglein , eur .
j. c * 75 * ( 2015 ) 254 [ arxiv:1411.4652 [ hep - ph ] ] .
f. domingo and t. lenz , jhep * 1107 * ( 2011 ) 101 [ arxiv:1101.4758 [ hep - ph ] ] . | while the properties of the signal that was discovered in the higgs searches at the lhc are consistent so far with the higgs boson of the standard model ( sm ) , it is crucial to investigate to what extent other interpretations that may correspond to very different underlying physics are compatible with the current results .
we use the next - to - minimal supersymmetric standard model ( nmssm ) as a well - motivated theoretical framework with a sufficiently rich higgs phenomenology to address this question , making use of the public tools ` higgsbounds ` and ` higgssignals ` in order to take into account comprehensive experimental information on both the observed signal and on the existing limits from higgs searches at lep , the tevatron and the lhc .
we find that besides the decoupling limit resulting in a single light state with sm - like properties , several other configurations involving states lighter or quasi - degenerate with the one at about 125 gev turn out to give a competitive fit to the higgs data and other existing constraints .
we discuss the phenomenology and possible future experimental tests of those scenarios , and compare the features of specific scenarios chosen as examples with those arising from a more global fit . |
one goal of the inverse scattering problem is to find the shape of bulk or flaw in a medium using the scattered field data measured at the boundary .
this is an old but it has several of applications such as detection of cracks in material engineering , ultrasound imaging in medical sciences , and scanning anti - personnel mines hidden in the ground in military services . related researches may be found in @xcite and the references therein . in many studies , various shape reconstruction algorithms
have been developed , most of which are focused on the minimization of the least - square functional by using a newton - type iteration strategy , e.g. , level - set method @xcite .
the merits of the iteration strategy is that it does not require a very large number of boundary measurements for complete shape reconstruction ; instead , in order to guarantee successful reconstruction , it demands very high computational expenditures , optimal regularization terms related to the problems at hand , calculation of a complex frchet derivative at each iteration , and _ a priori _ information of the unknown target to be reconstructed .
nevertheless , even if these conditions are satisfied , the iteration strategy must begin with an initial guess that is close to the true one in order to avoid undesirable situations such as non - convergence or falling into the local minima .
therefore , development of a mathematical theory and an algorithm for generating a good initial guess is an important research topic .
conversely , certain non - iterative shape reconstruction algorithms have been proposed ; for example , multiple signal classification ( music ) algorithm @xcite , linear sampling method @xcite , kirchhoff migration @xcite , and fourier inversion based one @xcite .
in contrast with the iterative strategy , these algorithms require a large amount of incident field data and a significant number of boundary measurements .
however , if these conditions are satisfied , non - iterative shape reconstruction algorithms can prove to be effective , fast , and easy to extend to multiple targets .
topological derivative concept is a non - iterative strategy .
this concept was originally developed for the shape optimization problem , but its application to rapid shape reconstruction has only recently been proven .
related works can be found in @xcite and references therein .
one of the advantages of topological derivative concept is that it does not require a large amount of many incident field data ; however , a reduction in the amount of this data has been reported to result in poor resolution of the reconstructed shape .
the purpose of this literature is to establish an effective and fast reconstruction algorithm for detecting both the location and shape of perfectly conducting crack(s ) from boundary - measured scattered field data by using only a small amount of incident field data .
based on the structure of topological derivative imaging functional operated at a fixed - frequency @xcite and multi - frequency imaging schemes @xcite , we introduce a multi - frequency topological derivative concept .
note that recent work ( * ? ? ?
* formula ( 1.33 ) ) has briefly introduced a multi - frequency topological derivative but through a more detailed analysis , we investigate the structure of the proposed multi - frequency topological derivative and its benefits over the traditional topological derivative .
the remainder of this paper is organized as follows .
section [ sec2 ] briefly introduces the traditional topological derivative by creating a linear , narrow crack .
section [ sec3 ] describes the design of a multi - frequency topological derivative imaging function and analyzes it to investigate its properties .
section [ sec4 ] presents various numerical experiments to assess the performance of the proposed imaging function under various situations .
section [ sec5 ] contains a short conclusion and some remarks on future work .
assume that a perfectly conducting crack @xmath0 is completely hidden in a homogeneous domain @xmath1 with a smooth boundary @xmath2 .
we denote @xmath3 as a two - dimensional vector that lies on crack @xmath0 . throughout this paper , we consider the so - called transverse magnetic ( tm ) polarization case , letting @xmath4 be the ( single - component ) electric field that satisfies the following boundary value problem : @xmath5 where @xmath6 represents the unit outward normal to @xmath7 and @xmath8 , @xmath9 , denotes the two - dimensional vector on the unit circle @xmath10 . throughout this paper , we let @xmath11 denote a strictly positive wavenumber and presume @xmath12 to not be an eigenvalue of ( [ helmholtz ] ) .
similarly , let @xmath13 be the solution of equation ( [ helmholtz ] ) without @xmath0 .
then , the problem we consider here is the computation of the topological derivative of the residual function depending on the solution @xmath4 : @xmath14 where @xmath15 . in order to compute the topological derivative ,
let us create a small linear crack @xmath16 of length @xmath17 centered at @xmath18 such that @xmath19 and denote @xmath20 as that domain . then , because of the change in the topology of @xmath21 , we can consider the topological derivative @xmath22 based on the residual function @xmath23 with respect to point @xmath24 as follows : @xmath25 where @xmath26 as @xmath27 . from relationship ( [ topderivative ] ) , we have the following asymptotic expansion : @xmath28 then , the topological derivative @xmath22 with @xmath29 different incident waves at a given wavenumber @xmath11 is as follows ( see @xcite for its derivation ) .
[ thtop ] let @xmath30 denote the real part of @xmath31 .
then , the topological derivative corresponding to ( [ energy ] ) is given by @xmath32,\ ] ] where @xmath33 satisfies the following adjoint problem : @xmath34
in this section , we introduce a multi - frequency topological derivative imaging function and explore its properties . based on recent research @xcite
, a sufficiently large number of incident directions are required to obtain a good result .
this means that if @xmath29 is small , an image with poor resolution will be reconstructed .
moreover , when boundary measurement data are affected by a large amount of random noise , ( [ topderivative2 ] ) will fail to reconstruct a good image .
inspired by multi - frequency imaging algorithms @xcite , we consider the following multi - frequency topological derivative for several wavenumbers @xmath35 : @xmath36 where @xmath37 is given by ( [ topderivative2 ] ) for @xmath38 , @xmath39 . in this section , we investigate the structure of ( [ multifrequencytopologicalderivative ] ) . for this purpose , we recall a useful result ( see @xcite ) as follows .
[ lemmkp ] let @xmath40 imply that there exists a constant @xmath41 such that @xmath42 . then , for @xmath43 and @xmath44 , ( [ topderivative2 ] ) satisfies @xmath45.\ ] ] here @xmath46 denotes the two - dimensional time - harmonic fundamental solution ( or green s function ) of helmholtz equation @xmath47 where @xmath48 , @xmath49 , and @xmath50 denote the hankel function of the first kind , bessel function , and neumann function of order zero , respectively .
consequently , ( [ multifrequencytopologicalderivative ] ) is analyzed for the following three cases of interest .
[ theoremsymmetric ] under the configuration of a small number of symmetric incident directions , the structure of ( [ multifrequencytopologicalderivative ] ) is @xmath51 for a sufficiently large @xmath52 and @xmath53 .
assume that the symmetric incident directions are of the form : @xmath54 where @xmath29 is an even number , say , @xmath55 .
based on this form , @xmath56 satisfies @xmath57 for @xmath58 .
in contrast with lemma [ lemmkp ] , we assume that @xmath29 is not sufficiently large .
let @xmath59 denote the imaginary part of @xmath31 then since @xmath60&=\re\bigg[\exp(ik\vec{\theta}_{l+n}\cdot(\vec{x}_c-\vec{x}_s))\bigg]\\ \im\bigg[\exp(ik\vec{\theta}_l\cdot(\vec{x}_c-\vec{x}_s))\bigg]&=-\im\bigg[\exp(ik\vec{\theta}_{l+n}\cdot(\vec{x}_c-\vec{x}_s))\bigg],\end{aligned}\ ] ] ( [ topderivative3 ] ) becomes @xmath61\\ & = \re\left[\frac{1}{2k_f}\sum_{m=1}^{n}\cos(k_f\vec{\theta}_m\cdot(\vec{x}_c-\vec{x}_s))\{j_0(k_f|\vec{x}_c-\vec{x}_s|)+iy_0(k_f|\vec{x}_c-\vec{x}_s|)\}\right]\\ & = \frac{1}{2k_f}\sum_{m=1}^{n}\cos(k_f\vec{\theta}_m\cdot(\vec{x}_c-\vec{x}_s))j_0(k_f|\vec{x}_c-\vec{x}_s|).\end{aligned}\ ] ] therefore , @xmath62 let us recall that for @xmath63 ( see ( * ? ? ? * formula 11.4.37 ( p. 487 ) ) ) @xmath64 let @xmath65 , @xmath66 , and @xmath67 . then since @xmath68\geq0,\ ] ] by letting @xmath53 , ( [ multifrequencytopologicalderivative ] ) becomes @xmath69 [ theoremnonsymmetric ] under the configuration of a small number of non - symmetric incident directions , the structure of ( [ multifrequencytopologicalderivative ] ) becomes @xmath70 for a sufficiently large @xmath52 and @xmath53 . in constrast with theorem [ theoremsymmetric ] , we assume that the incident directions are non - symmetric . then , ( [ topderivative3 ] ) can be written as @xmath71\\ = & \re\left[\frac{1}{4k_f}\sum_{m=1}^{m}\exp(ik_f\vec{\theta}_m\cdot(\vec{x}_c-\vec{x}_s ) ) \{j_0(k_f|\vec{x}_c-\vec{x}_s|)+iy_0(k_f|\vec{x}_c-\vec{x}_s|)\}\right]\\ = & \frac{1}{4k_f}\sum_{m=1}^{m}\bigg[\cos(k_f\vec{\theta}_m\cdot(\vec{x}_c-\vec{x}_s))j_0(k_f|\vec{x}_c-\vec{x}_s|)\\ & -\sin(k_f\vec{\theta}_m\cdot(\vec{x}_c-\vec{x}_s))y_0(k_f|\vec{x}_c-\vec{x}_s|)\bigg],\end{aligned}\ ] ] and therefore @xmath72.\end{aligned}\ ] ] let us recall an integral formula in ( * ? ? ?
* formula 11.4.40 ( p. 487 ) ) , @xmath73 then , if @xmath53 , we can evaluate @xmath74 hence , by combining ( [ inequalityab ] ) , ( [ structuremultitopderivative1 ] ) , and ( [ approximationsinexp ] ) , we can obtain @xmath75 [ theoremmanydirection ] under the configuration of a sufficiently large number of incident directions , the structure of ( [ multifrequencytopologicalderivative ] ) becomes @xmath76 for an adequately large @xmath52 . here , @xmath77 is defined as @xmath78 and @xmath79 denotes the dirac delta function .
suppose that @xmath29 is large enough , i.e. , @xmath80 , then by ( * ? ? ?
* lemma 4.1 ) , ( [ topderivative3 ] ) can be approximated as @xmath81\\ & \approx\re\left[\frac{1}{4k}\bigg(j_0(k|\vec{x}_c-\vec{x}_s|)+iy_0(k|\vec{x}_c-\vec{x}_s|)\bigg ) \int_{\mathcal{s}^1}\exp(ik\vec{\theta}\cdot(\vec{x}_c-\vec{x}_s))d\vec{\theta}\right]\\ & = \frac{\pi}{2k}j_0(k|\vec{x}_c-\vec{x}_s|)^2 .
\end{aligned}\ ] ] assume that @xmath52 is large enough and that @xmath82 . then applying an integral formula of the bessel function @xcite @xmath83 and a change of variable @xmath84 yields @xmath85 where @xmath86 is given in ( [ functionlambda ] ) .
note that the term @xmath87 can be negligible since @xmath88 and @xmath89 , refer to @xcite . finally , by applying @xmath53 in ( [ structuremultitopderivative3 ] )
, we can immediately obtain the following result : @xmath90 [ remark1]based on structures ( [ structuremultitopderivative1 ] ) , ( [ structuremultitopderivative2 ] ) , and ( [ structuremultitopderivative3 ] ) , we can summarize some properties of ( [ multifrequencytopologicalderivative ] ) as follows : 1 . [ p1remark ]
formulas ( [ structuremultitopderivative1 ] ) and ( [ structuremultitopderivative2 ] ) reach their maximum value at @xmath43 that satisfies @xmath91 therefore , at @xmath92 , the map of ( [ multifrequencytopologicalderivative ] ) will yield @xmath93 that the location of @xmath0 can be accurately determined
. however , at point @xmath43 such that @xmath94 the map of ( [ multifrequencytopologicalderivative ] ) will show a large magnitude such that unexpected replicas @xmath95 will appear ( see figure [ remarkvector ] ) .
[ p2remark ] if @xmath43 satisfies @xmath96 then both the numerator and denominator of ( [ structuremultitopderivative2 ] ) become @xmath97 . hence , at this point , ( [ structuremultitopderivative2 ] ) will yield a large value at not only @xmath3 but also other locations in @xmath21 .
therefore , the configuration of symmetric incident directions will yield a better imaging result ( see figure [ gammanonsymmetric1 ] ) .
[ p3remark ] if the number of incident directions is increased to as many as possible , i.e. , @xmath80 , it is clear that a more accurate location of @xmath44 can be obtained . in this case
, an odd number of @xmath29 will not affect the imaging performance significantly .
[ p4remark ] application of multiple frequencies enhances the imaging performance .
illustration of remark [ remark1 ] . at the location @xmath24
such that @xmath98 ( black colored vector ) and @xmath99 ( green colored vector ) , true shape of @xmath0 and ghost replicas @xmath95 will appear , respectively.,scaledwidth=55.0% ]
in this section , we exhibit some numerical examples to demonstrate the effectiveness of the proposed algorithm . throughout this section ,
the homogeneous domain @xmath21 is chosen as the interior region of the two - dimensional unit disk centered at the origin , i.e. , @xmath100\right\}}.\ ] ] the adopted wavenumber has the form @xmath101 for @xmath39 ; here , @xmath102 is the given wavelength that is equi - distributed between @xmath103 and @xmath104 .
based on theorem [ theoremsymmetric ] , the incident direction @xmath8 is selected as @xmath54 for an even number @xmath29 . motivated by @xcite , we choose two single cracks @xmath105 and @xmath106 as @xmath107\right\}}\\ \mathcal{c}_2&={\left\{\left(1.5\sin\frac{(3t+4)\pi}{8}-1 , 0.8\sin\frac{(3t+4)\pi}{4}\right):t\in[-1,1]\right\}},\end{aligned}\ ] ] and inspired by @xcite , we choose multiple cracks @xmath108 as @xmath109\}\cup\{(t+0.2,t^3+t^2 - 0.5):s\in[-0.5,0.5]\}.\ ] ] note that in order to show the robustness of the proposed algorithm , a white gaussian noise with @xmath110 db signal - to - noise ratio ( snr ) is added to the unperturbed data .
first , let us consider the imaging result of @xmath105 .
the left - hand side of figure [ gamma1 ] shows the map of @xmath111 for @xmath112 different directions . in this result , because of its small magnitude at @xmath113 , @xmath105 appears divided .
moreover , owing to the large magnitude at certain points ( for example , @xmath114 ) , we can not recognize its shape at this stage .
the right - hand side of figure [ gamma1 ] is the map of @xmath115 .
the figure shows that there is undivided or whole @xmath105 ; this means that we obtained a more accurate result than the previous one .
however , because certain points have a large magnitude , we can not determine the true shape of @xmath105 .
map of @xmath116 with @xmath117 ( left ) and @xmath118 ( right ) for @xmath105.,title="fig:",scaledwidth=49.0% ] map of @xmath116 with @xmath117 ( left ) and @xmath118 ( right ) for @xmath105.,title="fig:",scaledwidth=49.0% ] figure [ gamma2 ] shows the map of @xmath116 for crack @xmath106 .
in contrast with the result of @xmath105 , it is difficult to identify the shape of the crack because of the great number of points with large magnitudes .
fortunately , similar to the results in figure [ gamma1 ] , map of @xmath115 yields a better image than that obtained by the map of @xmath111 .
map of @xmath116 with @xmath117 ( left ) and @xmath118 ( right ) for @xmath106.,title="fig:",scaledwidth=49.0% ] map of @xmath116 with @xmath117 ( left ) and @xmath118 ( right ) for @xmath106.,title="fig:",scaledwidth=49.0% ] the application of multiple cracks @xmath108 is shown in figure [ gammam ] .
this figure shows the maps of @xmath119 and @xmath116 . throughout these results
, we can observe that the map of @xmath22 yields a poor result even if the number of incident directions @xmath29 is large ; conversely , the multi - frequency imaging function @xmath116 yields an accurate result even if @xmath29 is small .
maps of @xmath119 with @xmath120 ( top , left ) and @xmath121 ( top , right ) , and maps of @xmath116 with @xmath117 ( bottom , left ) and @xmath118 ( bottom , right ) for @xmath108.,title="fig:",scaledwidth=49.0% ] maps of @xmath119 with @xmath120 ( top , left ) and @xmath121 ( top , right ) , and maps of @xmath116 with @xmath117 ( bottom , left ) and @xmath118 ( bottom , right ) for @xmath108.,title="fig:",scaledwidth=49.0% ] + maps of @xmath119 with @xmath120 ( top , left ) and @xmath121 ( top , right ) , and maps of @xmath116 with @xmath117 ( bottom , left ) and @xmath118 ( bottom , right ) for @xmath108.,title="fig:",scaledwidth=49.0% ] maps of @xmath119 with @xmath120 ( top , left ) and @xmath121 ( top , right ) , and maps of @xmath116 with @xmath117 ( bottom , left ) and @xmath118 ( bottom , right ) for @xmath108.,title="fig:",scaledwidth=49.0% ] in figure [ gammanonsymmetric1 ] , maps of @xmath115 with @xmath122 number of incident directions are displayed for cracks @xmath105 and @xmath108 in order to verify the [ p2remark ] of remark [ remark1 ] . by comparing figures [ gamma1 ] and [ gammam ] , it is difficult to discern the shape of true crack(s ) because of the appearance of unforseen replicas with large magnitudes .
note that if @xmath29 is an odd number but is sufficiently large enough , the map of @xmath116 yield a good result , refer to figure [ gammanonsymmetric2 ] .
map of @xmath116 with @xmath123 for @xmath105 ( left ) and @xmath108 ( right).,title="fig:",scaledwidth=49.0% ] map of @xmath116 with @xmath123 for @xmath105 ( left ) and @xmath108 ( right).,title="fig:",scaledwidth=49.0% ] map of @xmath116 with @xmath124 for @xmath105 ( left ) and @xmath108 ( right).,title="fig:",scaledwidth=49.0% ] map of @xmath116 with @xmath124 for @xmath105 ( left ) and @xmath108 ( right).,title="fig:",scaledwidth=49.0% ] at this moment , we increase the number of applied frequencies @xmath52 and incident directions @xmath29 .
figures [ gamma12 - 6416 ] , [ gamma12 - 6464 ] , and [ gammam-64 ] depict the map of @xmath116 when @xmath125 and @xmath112 or @xmath121 .
as we expected in [ p3remark ] of remark [ remark1 ] , the results yielded by @xmath116 improve with increasing number of @xmath52 and/or @xmath29 .
map of @xmath116 with @xmath126 for @xmath105 ( left ) and @xmath106 ( right).,title="fig:",scaledwidth=49.0% ] map of @xmath116 with @xmath126 for @xmath105 ( left ) and @xmath106 ( right).,title="fig:",scaledwidth=49.0% ] map of @xmath116 with @xmath127 for @xmath105 ( left ) and @xmath106 ( right).,title="fig:",scaledwidth=49.0% ] map of @xmath116 with @xmath127 for @xmath105 ( left ) and @xmath106 ( right).,title="fig:",scaledwidth=49.0% ] map of @xmath116 with @xmath126 ( left ) and @xmath127 ( right ) for @xmath108.,title="fig:",scaledwidth=49.0% ] map of @xmath116 with @xmath126 ( left ) and @xmath127 ( right ) for @xmath108.,title="fig:",scaledwidth=49.0% ] although multi - frequency topological derivative imaging yields very good imaging results for not only single but also multiple cracks , it still has the following limitations . 1 .
let us consider the imaging of a crack of a large curvature near one of the crack tips . to illustrate this , we choose a crack from @xcite : @xmath128 + from the results in figure [ gammaa ] , the imaging results yielded by @xmath116 are satisfactory .
however , at the point of the large curvature , the map of @xmath116 reconstructs a slightly different shape of the crack .
we apply the proposed algorithm for imaging an oscillation crack represented as follows : @xmath129 + figure [ gammab ] shows the map of @xmath116 .
the obtained result by the proposed algorithm does not improve even when the number of @xmath52 and @xmath29 are increased .
the proposed algorithm requires further optiization .
map of @xmath116 for @xmath130 with @xmath118 ( top , left ) , @xmath126 ( top , right ) , @xmath127 ( bottom , left ) and true shape ( bottom , right).,title="fig:",scaledwidth=49.0% ] map of @xmath116 for @xmath130 with @xmath118 ( top , left ) , @xmath126 ( top , right ) , @xmath127 ( bottom , left ) and true shape ( bottom , right).,title="fig:",scaledwidth=49.0% ] + map of @xmath116 for @xmath130 with @xmath118 ( top , left ) , @xmath126 ( top , right ) , @xmath127 ( bottom , left ) and true shape ( bottom , right).,title="fig:",scaledwidth=49.0% ] map of @xmath116 for @xmath130 with @xmath118 ( top , left ) , @xmath126 ( top , right ) , @xmath127 ( bottom , left ) and true shape ( bottom , right).,title="fig:",scaledwidth=49.0% ] same as figure [ gammaa ] except the crack is @xmath131.,title="fig:",scaledwidth=49.0% ] same as figure [ gammaa ] except the crack is @xmath131.,title="fig:",scaledwidth=49.0% ] + same as figure [ gammaa ] except the crack is @xmath131.,title="fig:",scaledwidth=49.0% ] same as figure [ gammaa ] except the crack is @xmath131.,title="fig:",scaledwidth=49.0% ]
we improved the traditional topological derivative concept to determine location and reconstruct the shape of arbitrary - shaped curve - like perfectly conducting cracks .
for this purpose , we designed a reconstruction algorithm based on the multi - frequency topological derivative , analyzed its structure , and investigated its properties .
we confirmed that the multi - frequency topological derivative contains singularity on the cracks , and therefore it reaches its maximum value at the location of the cracks .
moreover , we demonstrated that why ghost replicas appeared in the neighborhood of cracks and why the configuration of symmetric incident directions yields better results than the non - symmetric configuration .
numerical simulations under various situations depicted both the benefits and limitations of the proposed algorithm . because the algorithm requires only a one - step iteration procedure , it is very fast in terms of shape identification but does not yield the complete shape of cracks .
however , by adopting the obtained result as a starting point of a newton - type based algorithm @xcite , it is expected that complete shape reconstruction can be successfully accomplished .
this paper deals with the reconstruction of perfectly conducting cracks with dirichlet boundary condition .
accordingly , the extension of this research to the crack with neumann boundary condition will be an interesting work .
we believe that applying higher - order terms in the asymptotic expansion formula @xcite give a higher order topological derivative @xcite .
the calculation and analysis of higher - order topological derivative will be a valuable research topic . according to the statistical hypothesis testing @xcite
, multi - frequencies are expected to enhance the imaging performance .
the corresponding work will include a careful stability and resolution analysis of the multi - frequency topological derivative .
00 m. abramowitz and i. a. stegun , handbook of mathematical functions , with formulas , graphs , and mathematical tables , ( 1996 ) , dover , new york .
d. lvarez , o. dorn , n. irishina , and m. moscoso , crack reconstruction using a level - set strategy , j. comput .
phys . , 228 ( 2009 ) , 57105721 . h. ammari , an introduction to mathematics of emerging biomedical imaging , mathematics and applications series , 62 ( 2008 ) , springer - verlag , berlin . h. ammari , mathematical modeling in biomedical imaging ii : optical , ultrasound , and opto - acoustic tomographies , lecture notes in mathematics , 2035 ( 2011 ) , springer - verlag , berlin . h. ammari , j. garnier , v. jugnon , and h. kang , stability and resolution analysis for a topological derivative based imaging functional , siam j. control . optim . ,
50 ( 2012 ) , 4876 .
h. ammari , j. garnier , h. kang , w .- k .
park , and k. slna , imaging schemes for perfectly conducting cracks , siam j. appl .
, 71 ( 2011 ) , 6891 . h. ammari , e. iakovleva and d. lesselier , a music algorithm for locating small inclusions buried in a half - space from the scattering amplitude at a fixed frequency , multiscale model .
simul . 3 ( 2005 ) , 597628 . h. ammari , e. iakovleva and d. lesselier , two numerical methods for recovering small electromagnetic inclusions from the scattering amplitude at a fixed frequency , siam j. sci .
comput . , 27 ( 2005 ) , 130158 . h. ammari and h. kang , reconstruction of small inhomogeneities from boundary measurements , lecture notes in mathematics , 1846 ( 2004 ) , springer - verlag , berlin . h. ammari , h. kang , h. lee and w .- k . park , asymptotic imaging of perfectly conducting cracks , siam j. sci .
comput . , 32 ( 2010 ) , 894922 .
h. ammari , s. moskow and m. s. vogelius , boundary integral formulae for the reconstruction of electric and electromagnetic inhomogeneities of small volume , 9 ( 2003 ) , 4966 .
d. auroux and m. masmoudi , image processing by topological asymptotic analysis , esaim proc . , 26 ( 2009 ) , 2444 . m. bonnet , fast identification of cracks using higher - order topological sensitivity for 2-d potential problems , eng
, 35 ( 2011 ) , 223235 .
a. carpio and
rapun , solving inhomogeneous inverse problems by topological derivative methods , inverse probl . , 24 ( 2008 ) , 045014 . m. cheney , the linear sampling method and the music algorithm , inverse probl . , 17 ( 2001 ) , 591595 .
m. donelli , a rescue radar system for the detection of victims trapped under rubble based on the independent component analysis algorithm , prog .
m , 19 ( 2011 ) , 173181 .
r. douvenot , m. lambert and d. lesselier , adaptive metamodels for crack characterization in eddy - current testing , ieee trans .
magn . , 47 ( 2011 ) , 746755 .
o. dorn and d. lesselier , level set methods for inverse scattering , inverse probl .
22 ( 2006 ) , r67r131 .
r. griesmaier , multi - frequency orthogonality sampling for inverse obstacle scattering problems , inverse problems 27 ( 2011 ) 085005 .
y. d. jo , y. m. kwon , j. y. huh and w .- k .
park , structure analysis of single- and multi - frequency imaging functions in inverse scattering problems , submitted , available at http://arxiv.org/abs/1208.0641 .
a. kirsch and s. ritter , a linear sampling method for inverse scattering from an open arc , inverse probl . 16
( 2000 ) , 89105 .
r. kress , inverse scattering from an open arc , math .
methods appl .
, 18 ( 2003 ) , 267293 .
r. kress and p. serranho , a hybrid method for two - dimensional crack reconstruction , inverse probl .
, 21 ( 2005 ) , 773784 .
k . ma , p .- s .
kim and w .- k .
park , analysis of topological derivative function for a fast electromagnetic imaging of perfectly conducing cracks , prog .
res . , 122 ( 2012 ) , 311325 . n. nemitz and m. bonnet , topological sensitivity and fmm - accelerated bem applied to 3d acoustic inverse scattering , eng
. anal . bound .
, 32 ( 2008 ) , 957970 .
k . park , a study on multi - frequency electromagnetic imaging , available at http://arxiv.org/abs/1208.2063 .
w .- k . park , multi - frequency topological derivative for approximate shape acquisition of curve - like thin electromagnetic inhomogeneities , available at http://arxiv.org/abs/1207.0582 .
w .- k . park , non - iterative imaging of thin electromagnetic inclusions from multi - frequency response matrix , prog .
res . , 106 ( 2010 ) , 225241 .
park , on the imaging of thin dielectric inclusions buried within a half - space , inverse probl .
, 26 ( 2010 ) , 074008 .
w .- k . park , on the imaging of thin dielectric inclusions via topological derivative concept , prog .
res . , 110 ( 2010 ) , 237252 . w
park , topological derivative strategy for one - step iteration imaging of arbitrary shaped thin , curve - like electromagnetic inclusions , j. comput .
phys . , 231 ( 2012 ) , 14261439 .
w .- k . park and d. lesselier , electromagnetic music - type imaging of perfectly conducting , arc - like cracks at single frequency , j. comput .
phys . , 228 ( 2012 ) , 80938111 .
w .- k . park and d. lesselier , fast electromagnetic imaging of thin inclusions in half - space affected by random scatterers , waves random complex media , 22 ( 2012 ) , 323 .
w .- k . park and d. lesselier , music - type imaging of a thin penetrable inclusion from its far - field multi - static response matrix , inverse probl . , 25 ( 2009 ) , 075002 .
park and d. lesselier , reconstruction of thin electromagnetic inclusions by a level set method , inverse probl .
, 25 ( 2009 ) , 085010 .
w. rosenheinrich , tables of some indefinite integrals of bessel functions , 2011 , available at http://www.fh-jena.de/~rsh/forschung/stoer/besint.pdf b. scholz , towards virtual electrical breast biopsy : space frequency music for trans - admittance data , ieee .
trans . med .
imaging , 21 ( 2002 ) , 588595 .
j. sokoowski and a. zochowski , on the topological derivative in shape optimization , siam j. control optim . , 37 ( 1999 ) , 12511272 . | this paper concerns a fast , one - step iterative technique of imaging extended perfectly conducting cracks with dirichlet boundary condition . in order to reconstruct the shape of cracks from scattered field data measured at the boundary
, we introduce a topological derivative - based electromagnetic imaging function operated at several nonzero frequencies .
the properties of the imaging function are carefully analyzed for the configurations of both symmetric and non - symmetric incident field directions .
this analysis explains why the application of incident fields with symmetric direction operated at multiple frequencies guarantees a successful reconstruction .
various numerical simulations with noise - corrupted data are conducted to assess the performance , effectiveness , robustness , and limitations of the proposed technique . perfectly conducting cracks , topological derivative , multiple frequencies , numerical experiments |
author acknowledges hospitality of the institut fr halbleiter physik at the universitt linz , where the work was partly performed .
author appreciates valuable discussions with g. bauer , g. brunthaler , m.v .
entin , v. kravtsov , i.m .
suslov , e.i .
rashba , and v. volkov .
the work was supported by the russian foundation for basic research ( grant 97 - 02 - 17387 ) , by the programs on `` physics of solid - state nanostructures '' and `` statistical physics '' and by grant from nwo the netherlands .
kravchenko , w.e .
mason , g.e .
bowker , j.e .
furneaux , v.m .
pudalov , and m.diorio , phys .
b * 51 * , 7038 ( 1995 ) .
kravchenko , d. simonian , m.p .
sarachik , w. mason , and j.e .
furneaux , phys .
lett . * 77 * , 4938 ( 1996 ) . | we present a model that explains two phenomena , recently observed in high - mobility si - mos structures : ( 1 ) the strong enhancement of metallic conduction at low temperatures , @xmath0 k , and ( 2 ) the occurrence of the metal - insulator transition in 2d electron system .
both effects are prescribed to the spin - orbit interaction anomalously enhanced by the broken inversion symmetry of the confining potential well .
* introduction . * recently , in experiments with high mobility si - mos structures , a strong drop in resistivity @xmath1 has been found @xcite as temperature decreases below @xmath2 k. this effect is evidently in disagreement with the conventional interpretation of the one - parameter scaling theory ( opst ) @xcite , according to which all states in 2d system at zero magnetic field should be localized in the limit of @xmath3 .
the subsequent scaling analysis of the temperature and electric field dependencies of the conductivity @xcite has revealed a critical behavior , typical for the metal - insulator transition .
finally , convincing evidence for the existence of the extended states in 2d system at zero field has been obtained in experiments _ in magnetic field _ , in studies of the quantum hall effect to insulator transitions @xcite . the extended states which in high magnetic field are located in the centre of the corresponding landau bands , at decreasing field were found to remain in a finite energy range , giving rise to a mobility edge .
the experimental results thus suggest the existence of a true metallic state and of the metal - insulator ( m - i ) transition in 2 dimensions .
these results are in apparent contradiction with the conventional opst , and the origin of the metallic state remains puzzling . in this work , both experimental findings are explained as a consequence of the spin - orbit interaction enhanced by the broken inversion symmetry .
the suggested model provides a good agreement with the experimental data on the temperature dependence of the resistivity @xmath1 .
* analysis of the experimental results . *
fig .
1 shows a set of the curves @xmath4 vs @xmath5 typical for the high - mobility samples @xcite , at different electron densities @xmath6 . at @xmath7
k , the resistivity , @xmath4 , increases slowly as temperature decreases , the latter is characteristic for the weakly localized regime . at lower temperature
, @xmath4 drops sharply for all curves belonging to the `` metallic '' range of densities , @xmath8 .
the resistance drop is observed at densities in the range from @xmath9 to @xmath10 . the critical density , @xmath9 is sample dependent and is equal to @xmath11 @xmath12 for the sample shown in fig . 1 .
the drop in @xmath13 diminishes with decreasing sample mobility , and is almost replaced by a conventional rise in @xmath4 at @xmath3 in the sample with 8 times lower mobility , @xmath14 @xmath15/vs .
the latter behavior is consistent with that reported in earlier studies on low - mobility samples @xcite .
* empirical fit of the data . * the @xmath13- curves in fig . 1 may be fitted well by an empirical dependence which summarizes the scattering probabilities of two processes : @xmath16 the first term is independent of temperature , while the second one describes a scattering through an energy gap , @xmath17 .
the curves shown in fig .
1 by thick continuous lines were obtained using two fitting parameters for each density , @xmath18 and @xmath19 .
* possible microscopic mechanisms . * as seen from fig . 1 , the characteristic temperature , @xmath19 , is of the order of @xmath20 k in high mobility samples . searching for a proper microscopic mechanism , we find two small energy gaps intrinsic to si - mos structures at @xmath21 : the valley splitting @xmath22 k and the zero - field spin - gap @xmath23 k @xcite .
it seems attractive to link the resistance drop to the transitions between the two electron valleys , located close to the x - points in the brillouin zone @xcite .
however , the intervalley `` um - klapp '' scattering would hardly occur , since it requires a combination of the reciprocal lattice vectors of the very high order . on the other hand ,
the phonon - induced intervalley transition would require participation of high energy phonons , @xmath24k , and is therefore unlikely at low temperatures .
electron tunneling , as the intervalley transition mechanism , would not lead to a strong temperature dependence of scattering .
* spin - orbit splitting and interaction effects . * in the one - electron approximation , the spin - orbit interaction is described by the hamiltonian @xcite : @xmath25 { \bf \sigma},\ ] ] where @xmath26 and @xmath27 are the momentum and spin operators , correspondingly . for the 2d electron system in si
, the contribution of the bulk crystal potential in @xmath28 is small ( @xmath29 ) , and the lack of inversion symmetry of the triangular confining potential @xmath30 plays the major role .
this lifts the spin - degeneracy at zero magnetic field , and leads to the appearance of a linear term in the energy spectrum of 2d electrons @xcite : @xmath31 the corresponding spin - gap @xmath32 can be viewed as the difference in energy for the electron states with spin directed in the plane but to the left and right side with respect to @xmath33 , or , equivalently , along and opposite to the effective magnetic field @xmath34 $ ] in the frame related to electrons moving in 2d plane with fermi velocity .
we suggest that the empirically determined energy gap , @xmath35 originates from @xmath36 , and is equal to @xmath37 ( where @xmath38 ) , whereas the temperature independent contribution to the resistivity , @xmath39 , is related to the spin - independent scattering .
finite quantum relaxation time @xmath40 and the corresponding level broadening @xmath41 should reduce the effective gap : @xmath42 * comparison of the model with the experimental results . *
the spin splitting @xmath43 k for the same si - mos structures was determined from the magnitude of the quantum oscillations of the chemical potential @xcite , extrapolated to zero field .
it is clear from eqs .
( 3 ) and ( 4 ) , that this value corresponds to @xmath44 . with this result
we obtain @xmath45kcm and the total energy spectrum becomes known . in order to estimate
the quantum level broadening , @xmath46 , the temperature dependence of the diagonal resistance was measured in the quantum hall effect regime with fermi energy adjusted to the zeeman energy gap at @xmath47 in the field of @xmath48 t. as a result , we have obtained an estimate , @xmath49 k for @xmath50@xmath12 ( here we presumed the gaussian level broadening and @xmath51 to be the full width ) . comparing the model effective energy gap , @xmath52 with the empirical value , @xmath53 k obtained in the fit at @xmath54 ( the 5th curve from the bottom in fig . 1 )
, we eventually find @xmath55 @xcite .
the empirical energy gap , @xmath56 , decreases to zero at @xmath57 , as seen in fig . 1 . in the above model ,
eq .
( 5 ) , this occurs because ( i ) the level broadening increases , @xmath58 and ( ii ) the spin splitting , @xmath59 , diminishes @xmath60 .
the empirical fitting parameter @xmath51 shown in fig .
2 vs electron density , rises indeed as density decreases .
the total fit in fig .
1 is in surprisingly good agreement with the experimental data , despite the very simplified character of the above model .
as @xmath6 decreases and approaches @xmath9 , the resistance drop starts at lower temperatures .
the weak decrease in @xmath13 noticeable at @xmath61 k is presumably due to the weak localization corrections @xmath62 @xcite .
this effect was ignored in the above model and the @xmath13 points corresponding to the negative @xmath63 were not fitted ; these points are connected by dashed lines in fig .
1 . fig .
2 shows also two relaxation times : @xmath64 calculated from the mobility in the @xmath65 limit , and @xmath66 .
it is noteworthy that both @xmath67 and @xmath40 diminish almost linearly as density decreases ( but at @xmath8 ) and independently of each other .
the momentum relaxation time @xmath64 provides the necessary resistivity value , @xmath68 at the critical density @xcite , whereas @xmath40 provides the effective spin - gap equal to 0 at @xmath69 . as a result
, @xmath64 _ decays faster and becomes smaller than _ @xmath40 with decreasing @xmath6 . at the critical density
, @xmath70 is of the order of 0.1 m@xmath71/vs and @xmath72k , that corresponds to @xmath73s , and @xmath74s .
the conclusion on interception of the two curves , @xmath75 and @xmath76 is model independent , whereas the numerical values of @xmath77 and the interception point depend on the model chosen for level broadening .
for instance , the interception occurs at 11.5 or @xmath78 @xmath12 for gaussian or lorenzian broadening correspondingly .
* metal - insulator transition at @xmath79 : spin - orbit interaction and symmetry effects . *
it is not only the low - temperature resistance drop but the total scaling behavior strongly depends on the symmetry .
the corresponding universality classes of the symmetry for random systems were established by dyson @xcite . in the presence of the spin - orbit ( so ) interaction ,
the orthogonal symmetry of the system is replaced by the symplectic symmetry . correspondingly , the level - repulsion exponent in the random matrix statistics @xcite changes from @xmath80 to @xmath81 .
it appears therefore , that states are less easily localized in systems with large @xmath82 .
the effect of the spin - orbit interaction on weak localization was studied both theoretically and experimentally @xcite . for the strong localization regime ,
there have been suggestions @xcite that m - i transition can occur in the presence of a strong so interaction .
the scaling function in 2d was found to behaves asymptotically like @xmath83 in the high conductance limit @xmath84 , with @xmath85 in the orthogonal and @xmath86 in the symplectic case @xcite .
the @xmath87-function in the symplectic case may thus become positive at sufficiently large @xmath88 . as disorder increases and conduction @xmath89 decreases
, all states will be localized even in the symplectic case
. the critical level of disorder and the critical conduction @xmath90 correspond to the point at which @xmath91 .
the behavior of the symplectic @xmath92 function in 2d is qualitatively consistent with the experimental data presented in fig . 1 , in the vicinity of @xmath9 .
as temperature , or broadening increases , the energy relaxation time @xmath93 appears as a cut - off parameter and @xmath94 may become larger than the inverse spin relaxation time @xmath95
. then the system would again behave as in the orthogonal symmetry case .
the natural measure of the so interaction strength is the spin - orbit gap @xmath96 given by eqs .
( 4 ) and ( 5 ) , whereas as an estimate for disorder we adopted @xmath97 .
thus , one may expect the m - i transition would manifest in those samples where @xmath98 , which is also consistent with occurrence of the transition in the samples with peak mobility larger than 5000 @xmath15/vs @xcite .
* discussion .
* the above model explains why the resistance drop is seen only in low - disordered samples with large @xmath40-values @xcite .
the effect is dependent also on the symmetry of the potential well .
this provides a key for testing the driving mechanism . as for other systems , the zero field spin - gap in gaas / al(ga)as is smaller by a factor of @xmath99 , due to the smaller @xmath100-factor value and much smaller @xmath101 @xcite .
thus , even in ideal samples with zero broadening , the resistance drop may occur at temperatures @xmath102 lower than those for the si - mos structures . in accord with this
, no signatures of the resistance drop were revealed in recent measurements on gaas / al(ga)as heterojunctions at temperatures down to 20mk @xcite .
recently , there have been suggestions on other possible collective mechanisms , such as coulomb interaction @xcite , spin - triplet pairing @xcite , and non - fermi - liquid behavior @xcite .
however , the corresponding models are not developed yet to provide a comparison with the experimental data .
* summary .
* it seems likely that the recently observed metal - insulator transition in high - mobility si - mos structures is the first experimental manifestation of the spin - orbit interaction induced transition in 2d .
the enhancement of the metallic conduction in these samples at low temperatures fits the same framework .
strong so interaction energy relative to the level broadening , and broken inversion symmetry are favorable for the 2d metallic state .
the coulomb interaction in this model provides the small level broadening at density down to @xmath9 . in the recent experiments @xcite ,
the 2d metallic phase was found to be easily destroyed by the in - plane magnetic field ; this is a strong evidence for the spin - related origin of the 2d metal . |
large scale structures like galaxies and clusters of galaxies are believed to have formed by gravitational amplification of small perturbations . for an overview and original references , see , e.g. , @xcite .
initial density perturbations were present at all scales that have been observed @xcite . understanding evolution of density perturbations for such initial conditions is essential for the study of formation of galaxies and large scale structures . the equations that describe the evolution of density perturbations in an expanding universe have been known for a long time and these are easy to solve when the amplitude of perturbations is small .
these equations describe the evolution of density contrast defined as @xmath2 . here
@xmath3 is the density at point @xmath4 and time @xmath5 , and @xmath6 is the average density in the universe at time @xmath5 .
these are densities of non - relativistic matter , the component that clusters at all scales and is believed to drive the formation of large scale structures in the universe .
once density contrast at relevant scales becomes large , i.e. , @xmath7 , the perturbation becomes non - linear and coupling with perturbations at other scales can not be ignored .
the equation for evolution of density perturbations can not be solved for generic perturbations in this regime .
n - body simulations are often used to study the evolution in this regime .
alternative approaches can be used if one requires only a limited amount of information and in such a case either quasi - linear approximation schemes or scaling relations @xcite suffice . in cosmological n - body simulations ,
we simulate a representative region of the universe .
this is a large but finite volume and periodic boundary conditions are often used .
almost always , the simulation volume is taken to be a cube .
effect of perturbations at scales smaller than the mass resolution of the simulation , and of perturbations at scales larger than the box is ignored .
indeed , even perturbations at scales comparable to the box are under sampled .
it has been shown that perturbations at small scales do not influence collapse of perturbations at much larger scales in a significant manner if we study the evolution of the correlation function or power spectrum at large scales due to gravitational clustering in an expanding universe .
this is certainly true if the scales of interest are in the non - linear regime @xcite .
therefore we may assume that ignoring perturbations at scales much smaller than the scales of interest does not affect results of n - body simulations .
however , there may be other effects that are not completely understood at the quantitative level @xcite even though these have been seen only in somewhat artificial situations .
perturbations at scales larger than the simulation volume can affect the results of n - body simulations .
use of periodic boundary conditions implies that the average density in the simulation box is same as the average density in the universe , in other words we ignore perturbations at the scale of the simulation volume ( and at larger scales ) .
therefore the size of the simulation volume should be chosen so that the amplitude of fluctuations at the box scale ( and at larger scales ) is ignorable .
if the amplitude of perturbations at larger scales is not ignorable then clearly the simulation will not be a faithful representation of the model being studied .
it is not obvious as to when fluctuations at larger scales can be considered ignorable , indeed the answer to this question depends on the physical quantity of interest , the model being studied and the specific length / mass scale of interest as well . the effect of a finite box size has been studied using n - body simulations and the conclusions in this regard may be summarised as follows . *
if the amplitude of density perturbations around the box scale is small ( @xmath8 ) but not much smaller than unity , simulations underestimate the correlation function though the number density of small mass haloes does not change by much @xcite .
in other words , the formation of small haloes is not disturbed but their distribution is affected by non - inclusion of long wave modes . * in the same situation , the number density of massive haloes drops significantly @xcite .
* effects of a finite box size modify values of physical quantities like the correlation function even at scales much smaller than the simulation volume @xcite .
* the void spectrum is also affected by finite size of the simulation volume if perturbations at large scales are not ignorable @xcite . *
it has been shown that properties of a given halo can change significantly as the contribution of perturbations at large scales is removed to the initial conditions but the distribution of most internal properties remain unchanged @xcite . in some cases ,
one may be able to devise a method to `` correct '' for the effects of a finite box - size , but such methods can not be generalised to all statistical measures or physical quantities .
the effects of perturbations at scales larger than the box size can be added using map ( mode adding procedure ) after a simulation has been run @xcite .
this method makes use of the fact that if the box size is chosen to be large enough then the contribution of larger scales can be incorporated by adding displacements due to the larger scales independently of the evolution of the system in an n - body simulation .
the motivation for development of such a tool is to enhance the range of scales over which results of an n - body simulation can be used by improving the description at scales comparable to the box size .
such an approach ignores the coupling of large scale modes with small scale modes and this again brings up the issue of what is a large enough scale for a given model such that the effects of mode coupling can be ignored .
large scales contribute to displacements and velocities , and variations in density due to these scales modify the rate of growth for small scales perturbations @xcite .
effects of a finite box size modify values of physical quantities even at scales much smaller than the simulation volume @xcite ( br05 , hereafter ) . in br05 , we suggested use of the fraction of mass in collapsed haloes as an indicator of the effect of a finite box size .
we found that if the simulation volume is not large enough , the fraction of mass in collapsed haloes is underestimated .
as the collapsed fraction is less sensitive to box - size as compared to measures of clustering , several other statistical indicators of clustering to deviate significantly from expected values in such simulations .
a workaround for this problem was suggested in the form of an ensemble of simulations to take the effect of convergence due to long wave modes into account @xcite , the effects of shear due to long wave modes are ignored here .
it has also been shown that the distribution of most internal properties of haloes , e.g. , concentration , triaxiality and angular momentum do not change considerably with the box size even though properties of a given halo may change by a significant amount @xcite .
there is a clear need to develop a formalism for estimating the effect of perturbations at large scales on a variety of physical quantities .
without such a formalism , we can not decide in an objective manner whether a simulation box size is sufficiently large or not . in this paper
we generalise the approach suggested in br05 and write down an explicit correction term for a number of statistical indicators of clustering .
this approach allows us to study generic properties of the expected correction term in any given case , apart of course from allowing us to evaluate the magnitude of the correction as compared to the expected value of the given statistical indicator .
we apply this technique to mass functions in this paper .
initial conditions for n - body simulations are often taken to be a realisation of a gaussian random field with a given power spectrum , for details see , e.g. , .
the power spectrum is sampled at discrete points in the @xmath9 space between the scales corresponding to the box size ( fundamental mode ) and the grid size ( nyquist frequency / mode ) . here
@xmath9 is the wave vector .
sampling of the power spectrum in initial conditions of n - body simulations is dense towards the nyquist mode , but is sparse for modes near the fundamental mode .
power spectra for density , potential and the velocity field are related to each other in the linear regime .
the power spectra can be used to compute the second moment ; either two point functions or _
rms _ fluctuations . in view of the sampling of the power spectrum in initial conditions ,
the second moment can be expressed as a sum over power spectrum at these points , weighted by an appropriate window function . in the peak picture ,
most quantities of interest can be related to the two point correlation function @xcite , therefore a method for estimating box - size correction to the second moment can be used as a base for computing correction for other physical quantities .
( see eqn.([eqn : c1 ] ) ) for power law models with index @xmath22 normalised such that @xmath23 .
the curves here are for @xmath24 ( dot - dashed curve ) , @xmath25 ( dashed curve ) and @xmath26 ( solid curve ) .
@xmath27 is plotted as a function of @xmath28 and we find that the correction term increases as @xmath29
. see text for more details.,width=307 ] in the approach outlined above , the value of @xmath30 at a given scale is expressed as a combination of the expected value @xmath31 and the correction due to the finite box size @xmath32 . here
@xmath31 is independent of the box size and depends only on the power spectrum and the scale of interest .
it is clear than @xmath33 and also @xmath34 .
it can also be shown that for hierarchical models , @xmath35 , i.e. , @xmath36 increases or saturates to a constant value as we approach small @xmath13 .
if the scale of interest is much smaller than the box - size @xmath37 then , @xmath38 ^ 2 \nonumber \\ &
\simeq & \int\limits_0^{2\pi/\lbx } \frac{dk}{k } \frac{k^3 p(k)}{2 \pi^2 } \nonumber \\ & & - \frac{r^2}{5 } \int\limits_0^{2\pi/\lbx } \frac{dk}{k } \frac{k^5 p(k)}{2 \pi^2 } \nonumber \\ & & + \frac{3 r^4}{175 } \int\limits_0^{2\pi/\lbx } \frac{dk}{k } \frac{k^7 p(k)}{2 \pi^2 } + \mathcal{o}(r^6 ) \\ & = & c_1 - c_2 r^2 + c_3 r^4 + \mathcal{o}(r^6 ) \label{eqn : sig1_series}\end{aligned}\ ] ] the small parameter in the expansion is @xmath39 .
this expansion is useful if @xmath40 goes to zero as we approach @xmath41 .
it is interesting to note that the first term is scale independent .
the numerical values of @xmath42 can be used to estimate the scale below which @xmath43 can be approximated by a constant .
later terms are scale dependent and the noteworthy feature is that modes closer to @xmath44 contribute more significantly to the integral for most models .
it is noteworthy that the first term , @xmath27 , has the same value for all choices of window functions that approach unity at small @xmath16 . by virtue of this fact , @xmath27 is also the correction for the two point correlation function @xmath45 at sufficiently small scales . on the @xmath46 plane for the @xmath47cdm model ( see text for details ) .
lines mark @xmath48 , @xmath49 , @xmath50 , @xmath51 and @xmath52 , from top to bottom . the lower panel shows the scale of non - linearity @xmath53 as a function of redshift for the @xmath47cdm model.,title="fig:",width=307 ] on the @xmath46 plane for the @xmath47cdm model ( see text for details )
. lines mark @xmath48 , @xmath49 , @xmath50 , @xmath51 and @xmath52 , from top to bottom .
the lower panel shows the scale of non - linearity @xmath53 as a function of redshift for the @xmath47cdm model.,title="fig:",width=307 ] at large scales @xmath20 and @xmath36 have a similar magnitude and the _ rms _ fluctuations in the simulation become negligible compared to the expected values in the model .
as we approach small @xmath13 the correction term @xmath36 is constant and for most models it becomes insignificant in comparison with @xmath20 . in models where @xmath20 increases very slowly at small scales or saturates to a constant value
, the correction term @xmath32 can be significant at all scales .
this can be seen from the expression for @xmath27 for power law models ( @xmath54 ) @xmath55 clearly , this term becomes more and more significant as @xmath56 .
figure 1 illustrates this point , here @xmath27 is shown as a function of @xmath28 .
we fix @xmath57 by choosing a scale of non - linearity @xmath53 such that @xmath58 .
curves are plotted for three values of @xmath59 : @xmath24 ( dot - dashed curve ) , @xmath25 ( dashed curve ) and @xmath26 ( solid curve ) . as @xmath60 is unity at the scale of non - linearity and @xmath27
is the first order correction , clearly we require @xmath61 for the error due to box size to be small and hence ignorable .
if we fix @xmath62 then we can simulate @xmath63 with @xmath24 but for more negative indices we require a larger separation between the box size and the scale of non - linearity .
we can just about manage @xmath64 with @xmath25 with the same threshold on error , and with @xmath26 we can go up to @xmath65 .
as n - body simulations are most useful for studying non - linear evolution , even the largest simulations possible today are left with a small range of scales over which @xmath66 for @xmath67 .
this shows the pitfalls of simulating models with @xmath68 over the entire range of scales .
figure 2 ( top panel ) shows lines of constant @xmath69 in the @xmath70 plane for the @xmath47cdm .
we chose @xmath71 , @xmath72 , @xmath73 , @xmath74 and @xmath75 .
we ignored the effects of baryons on the power spectrum . from top to bottom , the lines correspond to @xmath48 , @xmath49 , @xmath50 , @xmath51 and @xmath52 .
it is noteworthy that a box - size smaller than @xmath52mpc is precluded if we insist on @xmath76 , irrespective of the scale of interest .
this implies that we can not expect to simulate scales smaller than about @xmath52 kpc in the @xmath47cdm model without considerable improvement in the dynamic range of cosmological n - body simulations .
as we are using linearly evolved quantities for our argument , the comments on box size are valid irrespective of the redshift up to which the simulation is run
. the contours do not change if we use @xmath32 instead of @xmath27 .
the lower panel of the same figure shows the scale of non - linearity for the @xmath47cdm model as a function of redshift .
this formalism can be used to estimate corrections for other estimators of clustering . for reference
, we have given expressions equivalent to eqn.([eqn : sig1_series ] ) for the correction to @xmath77 and @xmath78 in table 1 .
.this table lists corrections due to a finite box - size to indicators of clustering in the limit @xmath79 .
these expressions are equivalent to eqn.([eqn : sig1_series ] ) and constants @xmath42 are the same as in that equation . [ cols="^,^",options="header " , ] we present here some preliminary results of a comparison of our results with n - body simulations .
we do not try to fit either the press - schechter or the sheth - tormen mass functions to simulations here , instead we use a simulation with a larger @xmath37 as reference and compare the number density of haloes as a function of mass with another simulation run using a smaller @xmath37 .
more detailed results obtained from n - body simulations will be presented in a later publication .
we simulated the @xmath80 power law model in an einstein - de sitter universe with the normalisation @xmath81mpc at the final epoch .
we chose one grid length of the simulation to equal @xmath82mpc .
the simulation was run with two values of the box - size : @xmath83mpc and @xmath84mpc , with the latter being the reference .
the simulations were run using the treepm method @xcite .
the parallel treepm code was used for the @xmath85 simulation @xcite .
figure 5 shows the number density of haloes @xmath86 in the mass range @xmath87 for these two simulations .
note that following our definitions @xmath88 , where @xmath89 is the multiplicity function .
the solid line shows the number density of haloes in the reference simulation .
one can see the approximately power law variation at small @xmath90 and a rapid fall off at large @xmath90 .
number density of haloes in the simulation with @xmath83mpc is shown by the dashed line . as expected from the above discussion
, the deviation from power law starts at smaller masses as the number density of very massive haloes is under - estimated as compared to the reference simulation . at smaller @xmath90 , we find about @xmath91 more haloes in this simulation as compared to the reference .
it is noteworthy that the number density of low mass haloes remains above that in the reference simulation at all masses below the rapid drop around @xmath0 m@xmath1 .
both the features follow the predictions in the preceeding discussion , indeed we have shown that these features are independent of the specific analytical form for the mass function . here
we have also shown that the same behaviour is reproduced in n - body simulations .
a more detailed comparison is beyond the scope of this paper and the results will be presented in a later publication .
in the preceeding sections , we have described a method to estimate errors in the descriptors of clustering in the linear regime .
we have also shown that the key results of the analytical study are borne out by n - body simulations .
we have shown that the error is typically small if the scale of interest is sufficiently smaller than the box size .
an implicit requirement is that the scale of non - linearity too should be much smaller than the box size ; if this restriction is overlooked then we not only ignore power in modes larger than the simulation box but also the significant effects of mode coupling from scales in the mildly non - linear regime .
therefore , we require @xmath92 .
we propose using @xmath93 as an indicator of the significance of the finite box size , any descriptor of second moment can be used but @xmath94 has the virtue of being positive definite at all scales .
our proposal is that @xmath95 , for the finite box - size corrections to be ignorable .
all the @xmath94s are evolved linearly here .
conversely , the ratio @xmath96 at the scale of interest is indicative of the magnitude of correction due to the finite box - size . for a given relative magnitude of the correction term ( @xmath96 )
, @xmath59 is required to be larger for spectra with more large scale power .
indeed , the required @xmath59 approaches infinity as the slope of the power spectrum approaches @xmath97 . as a result of finite box - size corrections ,
the amplitude of density perturbations is not a power law and the range of scales over which it can be approximated by one becomes smaller as we approach @xmath29 . in the linear regime ,
the radial pair velocity is related directly with @xmath78 @xcite . as @xmath78 is not a pure power law in simulations due to box - size corrections
, we expect that the pair velocities must also deviate from expected values .
this , in turn leads to deviations from scale invariant growth of density perturbations .
this explains the difficulty in getting scale invariant evolution for models like @xmath80 in n - body simulations @xcite .
for realistic models like the @xmath47cdm , the correction term is significant only if the scales of interest are below a few kpc and becomes larger as we move to smaller scales ( see figure 2 ) .
indeed , at these small scales we may require @xmath98 or even greater in order to manage @xmath99 .
of course , a bigger simulation volume is required if we demand better accuracy . on the other hand , if we are interested in scales larger than @xmath0 kpc , present day simulations are sufficient for keeping @xmath100 .
we have shown that at sufficiently small scales the correction due to a finite box size can be written as a series of progressively smaller terms . the first correction term ( @xmath27 )
is shown to be positive definite .
we have also shown that the first correction term is the same for two point correlation function and @xmath30 , indeed it is the same for all descriptors of the second moment for which the effective window function goes to unity at small @xmath16 . as an application of our method
, we discussed the correction to mass function and multiplicity function using the press - schechter as well as the sheth - tormen approach .
we have given the explicit form of the correction term due to finite box size in each case .
we have also given approximate expressions for the correction term and have shown that the approximation is very good in case of mass function .
the mass function is always under estimated in simulations due to finite box - size corrections .
multiplicity function , and hence also the number density of haloes of a given mass are underestimated at @xmath101 . at smaller mass scales , however , the multiplicity function is over estimated and we find more haloes in a simulation than expected in the model .
the mass scale at which the cross over from under estimate to over estimate occurs is given by eqn.(10 ) for press - schechter and eqn.(14 ) for sheth - tormen mass function .
the over estimate at small scales is related to the under estimate of mass in haloes at large scales .
if the full power spectrum had been taken into account , the smaller haloes would have merged to form more massive haloes . in absence of large scale modes ,
the formation of massive haloes is slowed down and a larger number of low mass haloes survive .
a detailed analysis of the effect of finite box size correction on merger rates for haloes will be presented in a forthcoming paper .
of significant interest is the impact on rates of major mergers @xcite as these have implications for observations . we find that the over estimate in multiplicity function is large whenever the ratio @xmath102 is large . to illustrate this correlation , we have plotted the multiplicity function @xmath103 for the @xmath47cdm model in figure 6 .
this has been plotted for redshift @xmath104 and @xmath105 and the mass range has been chosen such that very large box size is required to keep @xmath106 smaller than @xmath50 .
we have also plotted @xmath107 here , with @xmath108h@xmath109kpc ( dotted curve ) , @xmath110h@xmath109kpc ( dot - dashed curve ) and @xmath111h@xmath109kpc ( dashed curve ) .
these correspond to @xmath112 , @xmath51 and @xmath113 , respectively .
the top row is for the press - schechter mass function and the lower row is for the sheth - tormen mass function .
an identical @xmath114 range has been used to highlight the differences between the two models for mass function as well .
it is noteworthy that the relative error is similar in both the cases even though the multiplicity function itself is different . at @xmath115 ,
the multiplicity function is under estimated by a large amount for @xmath108h@xmath109kpc , even though @xmath116 and if we are interested in scales around @xmath82 pc then @xmath117 .
the situation at small masses is better for the other two simulation volumes considered here .
for @xmath118 , the scale of non - linearity is @xmath119h@xmath109kpc , very close to @xmath108h@xmath109kpc and hence we do not expect believable results for this box - size . indeed , the two panels on the right demonstrate the large errors and the absurdly incorrect shape of the multiplicity function .
the difference in @xmath89 and @xmath103 at @xmath120 is about @xmath121 for @xmath122 , in this case @xmath123h@xmath109kpc and @xmath124 .
the error in the multiplicity function is slightly larger than @xmath91 for @xmath125h@xmath109kpc even though @xmath126 and @xmath127 .
the multiplicity function plotted here is the global function , and the conditional mass functions should be used in order to estimate errors in simulations where a high peak is studied at better resolution .
similar results are obtained for other mass functions that have been suggested as a better fit to simulation data @xcite .
the above discussion demonstrates the perils of using simulations where @xmath128 is close to unity .
one may argue that models for mass function have not been tested in this regime where the local slope of the power spectrum is very close to @xmath97 , but the fact that error in amplitude of density perturbations itself is large should be reason enough to worry about reliability of results .
further , the agreement in the magnitude of errors for the several approaches to mass functions also gives us some confidence in results .
majority of simulations are not affected by such serious errors , as the slope of power spectrum approaches @xmath97 only at very small scales ( large wave numbers ) .
however , high resolution simulations of earliest structure formation in the @xmath47cdm model need to have a very large dynamic range before the results can be believed within @xmath91 of the quoted value . indeed , our work may have some relevance to the ongoing discussion about the earth mass haloes @xcite .
conclusions of this work may be summarised as follows .
* we have presented a formalism that can be used to estimate the deviations of cosmological n - body simulations from the models being simulated due to the use of a finite box size .
these deviations / errors are independent of the specific method used for doing simulations . * for a given model , the deviations can be expressed as a function of the scale @xmath13 of interest and @xmath37 , the box size of simulations .
* we have applied the formalism to study deviations in _ rms _ fluctuations in mass in the initial conditions . *
we find that the errors are small except for models where the slope of the power spectrum is close to @xmath97 at scales of interest . *
the errors in case of the @xmath47cdm model are significant if the scale of interest is smaller than a kpc even if simulations as large as the millennium simulation @xcite are used .
* we have studied errors in mass function in the press - schechter model , as well as other models . *
the main error due to a finite box size is that the number of high mass haloes is under estimated . *
the number of low mass haloes is over predicted in simulations if the box size effects are important .
this happens as low mass haloes do not merge to form the ( missing ) high mass haloes .
* we have verified these trends using n - body simulations .
we note that it is extremely important to understand the sources of errors in n - body simulations and the magnitude of errors in quantities of physical interest .
n - body simulations are used to make predictions for a number of observational projects and also serve as a test bed for methods . in this era of `` precision cosmology '' , it will be tragic if simulations prove to be a weak link
. we would like to note that our results apply equally to all methods of doing cosmological n - body simulations , save those where techniques like map are used to include the effects of scales larger than the simulation volume . the method for estimating errors due to a finite box - size described in this paper can be used for several physical quantities . in this paper
we have used the method to study errors in clustering properties and mass functions .
we are studying the effect of finite box size on velocity fields and related quantities , the results will be presented in a later publication .
numerical experiments for this study were carried out at cluster computing facility in the harish - chandra research institute ( http://cluster.mri.ernet.in ) . this research has made use of nasa s astrophysics data system .
we thank the anonymous referee for useful comments . | n - body simulations are a very important tool in the study of formation of large scale structures .
much of the progress in understanding the physics of galaxy formation and comparison with observations would not have been possible without n - body simulations . given the importance of this tool ,
it is essential to understand its limitations as ignoring these can easily lead to interesting but unreliable results . in this paper
we study the limitations due to the finite size of the simulation volume .
we explicitly construct the correction term arising due to a finite box size and study its generic features for clustering of matter and also on mass functions .
we show that the correction to mass function is maximum near the scale of non - linearity , as a corollary we show that the correction to the number density of haloes of a given mass changes sign at this scale ; the number of haloes at small masses is over estimated in simulations .
this over estimate results from a delay in mergers that lead to formation of more massive haloes . the same technique can be used to study corrections to other physical quantities .
the corrections are typically small if the scale of non - linearity is much smaller than the box - size
. however , there are some cases of physical interest in which the relative correction term is of order unity even though a simulation box much larger than the scale of non - linearity is used . within the context of the concordance model
, our analysis suggests that it is very difficult for present day simulations to resolve mass scales smaller than @xmath0 m@xmath1 accurately and the level of difficulty increases as we go to even smaller masses , though this constraint does not apply to multi - scale simulations .
methods : n - body simulations , numerical
gravitation
cosmology : theory , dark matter , large scale structure of the universe |
in the recent progress @xcite , we have developed a transport model to describe the hadron production in relativistic heavy ion collisions . in the model , a decoupling hypersurface of the fluid which consists of a group of splitting qgp droplets @xcite
is used to emit hadrons .
the model satisfies most of light hadrons spectra , but disagrees with the distributions of @xmath0 @xcite , @xmath1 @xcite and some heavy resonances @xcite . as another remarkable phenomenon
, there seems to be significant enhancements at very low transverse momentum in some of the spectra for @xmath3 @xcite , @xmath4 and @xmath0 , to change the slopes . on the contrary ,
the model predicts a decrease when @xmath2 drops .
the idea of the hypersurface hadronization locks the emitted hadrons in the local rest frames of the hypersurface .
this is right when the fluid is dense enough in the early stage of the evolution . in the later stage , when the inelastic free path becomes as long as @xmath5 , the emission fixed in the hypersurface might be less strict . considering the spectra of light hadrons are well satisfied ,
as other hydrodynamic methods with decoupling hypersurface did , there must be some reasons to push the inside light hadrons to participate in the collective motions .
one of the possible candidate is the elastic collision @xcite .
although elastic collisions do not change the hadron abundances themselves , the effect of preventing hadrons from entering the more dense regions and being absorbed , will enhance their abundances or momenta .
thus , detailed simulations on hadron transport are applied to some of the heavy hadrons , such as @xmath0 and @xmath1 in the central events in rhic au - au 200 @xmath6 .
the advantages for @xmath0 and @xmath1 are based on the condition that almost no hadrons could decay to them and their widths are large enough to neglect the affections of decays in the expanding medium . as a simplification , only selected hadrons for the measurements are simulated . the medium evolution and the feedbacks to the medium are not simulated in microscopic frames .
as the cross sections of some heavy or multi - strange hadrons may be much smaller than those of light hadrons @xcite , it will provide us a possible solution to explain the various collective behaviors between different hadrons and @xmath2 regions .
@xmath7,\nonumber\\ { \partial v\over\partial t}&=&-{1\over 3-v^2}\bigg[{3(1-v^2)^2\over 4}{\partial\xi\over\partial r}+2v{\partial v\over\partial r}\nonumber\\ & & \ \ \ \ \ \ \ \
-(n-1){v^2(1-v^2)\over r}\bigg],\nonumber\\ { \partial n_\gamma\over \partial t } & = & -\left[v{\partial n_\gamma\over\partial r}+n_\gamma{\partial v\over \partial r}+(n-1)n_\gamma{v\over r}\right]\end{aligned}\ ] ] the decoupling of the fluid is a gradual progress in principle . as the feedback of the hadron simulation to the fluid is unjustified , the medium is separated to an ideal hydrodynamic part and a free expanding part with evaporation by a cut @xmath8 .
the @xmath9 hydrodynamic evolution with an invariant boost for the central events in rhic au - au 200 @xmath6 is determined by the method in ref @xcite .
the decoupling surface here is similar to the results of surface evaporation @xcite except in the very early stage which is considered to be less important . for a group of decoupled droplets , there are no collective motions and local rest frames indeed . approximately , the whole evolution is defined , @xmath10 where the former term in the integral stands for the free flights of the droplets and the latter term is the evaporation . then , @xmath11 where @xmath12 is the velocity and radius distribution of the number density of the droplets . as we do not know the distribution of the droplets @xmath13 and can not simulate all of the droplets either , the expansion is simplified as , @xmath14 where @xmath15 and @xmath16 is the symmetric dimension . considering that the droplet evolution by emitting and absorbing hadrons is not clear ,
the evaporation term is omitted in our calculations .
the classical transport equation for the evolution of hadron phase space distribution @xmath17 can be written as @xmath18 where the second term on the left hand side is the free streaming part , and on the right hand side the lose and gain terms @xmath19 and @xmath20 indicate hadron absorption and production in the medium , and @xmath21 stands for the elastic term . in the simulations , hadrons generated in qgp or decayed from heavier hadrons ,
will pass through the medium and decay at the same time .
thus , the gain term includes two parts , @xmath22 .
the decay contribution is easy to obtain , but the term @xmath23 is complicated .
the reason is that the quantitative temperature dependencies for some parameters are unknown @xcite , especially in a rapidly expanding system .
therefore , the production is simplified as ideal qgp with massless quarks and gluons plus the contributions of bound states which will never change via @xmath24 , as well as the coupling constants .
the simplification is based on such an idea that in large @xmath24 regions , where the inelastic free path is short and the produced hadrons may travel more distance before they escape from the medium , the affections may be less important for the simulations .
thus , the approximation will be sufficient enough if the productions are well satisfied near the critical temperature @xmath25 .
then , @xmath26 where @xmath27 for mesons and @xmath28 for baryons due to the quark - diquark frame @xcite .
@xmath29 is the ratio of the occupation volume of all the qgp droplets to the whole volume of the system , @xmath30 is the production rate at @xmath25 and @xmath31 where @xmath32 is contributions of bound states in qgp .
the loss rate includes the decay rate and the medium absorption , @xmath33 , where @xmath34 and @xmath35 with @xmath36 at the critical temperature @xmath25 . as the expression belongs in the local rest frame of the medium , a lorentz transformation on time step
is required when the simulation is applied in the selected frame .
details on @xmath30 and @xmath37 are discussed in ref @xcite .
it is noted , ( [ eq : pr ] ) and ( [ eq : abs ] ) are not independent , even the droplets are assumed under random distribution . when @xmath38 , the location of production and absorption is correlated .
if the qgp droplets are large enough , a factor @xmath39 should be inserted to the independent productions ( [ eq : pr ] ) to remove the absorption by the droplet where the hadron is produced by qgp combination in principle .
@xmath39 can be fitted as , @xmath40 where @xmath41 as discussed in above sections , the evolutions of the droplets are unjustified so far .
thus we set @xmath42 and @xmath43 to minimize the parameter dependency .
hadrons produced in time , position and momentum bins ( @xmath44 ) are tracked when they pass through the medium .
they are recorded separately , as @xmath45 where @xmath46 is the direct and reconstructed hadrons escaped from the medium by cuts and @xmath47 is the reconstructed hadrons which may decay in the medium , @xmath48 with recorded decay products @xmath49 and @xmath50 there is a small difference between our calculation and the experimental manipulations . in the experiments ,
usually one channel is measured for the reconstruction to provide the total abundance by dividing the corresponding branch ratio .
distribution of @xmath51 comparing with the data from star collaboration @xcite and the phenix collaboration @xcite .
( a ) comparison between hypersurface hadronization with transport emit distributions @xcite and inelastic transport simulation .
( b ) demonstration of the push and drag effects by elastic collisions .
( c ) demonstration of drag effects at different mean elastic cross sections .
subscripts for the mean elastic cross sections are removed for clarity . ]
distribution of @xmath52 comparing with the data from star collaboration @xcite .
, scaledwidth=40.0% ] inelastic simulations are applied with elastic cross sections omitted ( @xmath53 ) . for the consistency with @xcite ,
no parameter is changed .
there is no new parameter either .
the results for @xmath0 and @xmath1 are shown in fig .
[ fig : phi]a and fig .
[ fig : omega ] .
the simulated results look much better than the original estimations from hypersurface emissions @xcite .
the productions in very low @xmath2 region are enhanced and the collective flows are reduced .
the perfect @xmath1 spectrum shows the elastic collisions for @xmath1 could be neglected as they may not participate in the collective motions .
as the enhancement in very low @xmath2 region for @xmath0 is still smaller than expected , elastic collisions are required in our considerations .
the elastic process can be divided to two effects approximately .
the acceleration pushed by the pressure gradient and the deceleration dragged by the relative motion .
different from the inelastic cross sections , which is nearly stable via energy scales and parton species , the elastic cross sections vary from @xmath54 to @xmath55 @xcite , which makes it difficult to obtain a strict calculation with limited information of parton - hadron or hadron - hadron elastic differential cross sections , especially in the condition that only a few selected hadrons are simulated .
heavy hadrons are discussed here to present a rough estimation on elastic process .
as they are heavy , the changes of their momenta after each collision and the thermalization are supposed to be small .
thus , the push and drag effects can be estimated by integrating all of the collisions or the probabilities of them .
the push effect can be described in the local rest frame of the medium , @xmath56 where @xmath57 is the mean hadron volume of elastic collisions . as
detailed differential cross sections are not clear , it is estimated by a rigid ball approximation , @xmath58 on the other hand , the drag effect for a flat plate moving in normal direction is presented as , @xmath59 where @xmath60 is the momentum shift after each collision .
the integral could be simplified by assuming the medium particles to be massless and fixing some of the momentum symbols to a kind of average @xmath61 .
thus , @xmath62 where , @xmath63[e_i(1-v_i)+\langle{p_j}\rangle]},\ ] ] with @xmath64 for medium particles near @xmath65 .
@xmath66 is a correction function for the `` shape '' of the hadrons
. it should be noted , as the elastic cross sections vary in a large range for different energy scales , different averages of elastic cross sections @xmath67 and @xmath68 will be reached by equations ( [ eq : push ] ) and ( [ eq : drag ] ) .
the relation between them is not clear .
as eq ( [ eq : push ] ) affects slow hadrons more intensively and eq ( [ eq : drag ] ) affects fast hadrons more significantly , @xmath67 might be slightly larger than @xmath68 . to make the value @xmath68 similar to @xmath67
, we estimate @xmath69 as a rough approximation .
the elastic corrections are only applied in the medium before it decouples .
equations ( [ eq : push ] ) and ( [ eq : drag ] ) are inserted to the simulation separately .
the results are troublesome when the push and drag effects are inserted together .
they disturb each other in their invalid regions and the problem is not resolved so far . as shown in fig .
[ fig : phi]b , @xmath0 production at very low @xmath2 is enhanced by the drag effect of preventing hadrons from moving inside .
the enhancement @xcite could be well satisfied by fitting @xmath70 .
in higher @xmath2 regions , push effect will effectively accelerate the hadrons , like the hydrodynamics does , to participate in the collective motions until the medium decouples , when @xmath71 .
they do nt work well in opposite regions .
the effects of different @xmath68 are shown in fig .
[ fig : phi]c .
the result for @xmath1 implies the elastic cross section for @xmath1 is no more than @xmath72 .
distribution of @xmath73 comparing with the data from star collaboration @xcite .
@xmath74 for mesons is @xmath75 of the values listed .
, scaledwidth=40.0% ] as the momentum of each surviving daughter hadron may be shifted by elastic collisions , it is hard in current simulations to determine the efficiency of the reconstruction and the changed momenta of the reconstructed hadrons , which are different from those when they decay , until all of the related hadrons including backgrounds are simulated to commit a real reconstruction .
further more the elastic cross sections may be much different for hadrons and their decay products .
thus , the spectra for some strong decay hadrons , can not be simulated strictly .
the spectrum of @xmath76 is not satisfied at very low @xmath2 so far , as shown in fig .
[ fig : sig1385 ] .
@xmath77 has the same problem but looks more serious .
other possibilities may be the lack of information for some of heavier baryons which may decay to @xmath76 or @xmath77 , and inelastic hadron re - scattering .
distribution of @xmath78 comparing with the data from phenix collaboration @xcite .
@xmath74 for baryons is @xmath79 of the values listed .
, scaledwidth=40.0% ] although the equations ( [ eq : push ] ) and ( [ eq : drag ] ) are not suitable for light hadrons , we still applied them to test the push effect for pions . when @xmath80 , the inclusive pion distribution @xcite can be reproduced , as shown in fig .
[ fig : pion ] . for direct pions without any decay contributions , @xmath81 .
local thermalization as a boltzmann distribution was tried for strongly interacting regions for all the simulations and no significant difference or better results were found .
beyond the approximation of hypersurface hadronization developed from the transport model @xcite , we have applied transport simulations for heavy hadrons to reproduce the significant enhancement in very low @xmath2 region in some of transverse momentum spectra .
although the method is neither strict nor complete , and the values of mean elastic cross sections are just qualitative , the reason for the enhancement and therefore the different slopes in the spectra may be successfully explained by the push and drag effects .
the simulation shows that the elastic cross sections of @xmath0 and @xmath1 are much smaller than those of other hadrons .
the simulation can also explain the reason why the hypersurface hadronization works well for light hadrons by the push effect .
j. adams et al . , the star collaboration , phys . lett . * b612 * , 181 ( 2005 ) . s. s. adler et al , the phenix collaboration , phys . rev . * c72 * , 014903 ( 2005 ) .
j. adams et al . , the star collaboration , nucl . phys . * a757 * , 102 - 183 ( 2005 ) ; j. adams et al .
, the star collaboration , phys . rev . lett . * 98 * , 062301 ( 2007 ) .
j. adams et al . ,
the star collaboration , phys .
lett . * 97 * , 132301 ( 2006 ) ; s. salur , ph.d .
thesis , yale university , 2006 .
h. miao , c. s. gao and p. f. zhuang , commun .
. phys . * 46 * , 1040 - 1046 ( 2006 ) .
h. miao , z. b. ma and c. s. gao , j. phys .
* g29 * , 2187 - 2192 ( 2003 ) ; h. miao and c. s. gao , j. phys . * g31 * , 179 - 184 ( 2005 ) . w. m. yao et al . , ( particle data group ) , j. phys . *
g33 * , 1 ( 2006 ) . | based on the hadron transport frames , detailed simulations are carried out to discuss @xmath0 , @xmath1 productions and the significant enhancements in the very low @xmath2 region for some of the soft spectra in rhic .
elastic interactions are introduced in the simulations .
the elastic cross sections vary from different hadrons and energy scales , which can qualitatively explain the different collective motions of various hadrons . |
stellar populations are fundamental tracers of the formation and evolution history of their parent galaxy .
the accurate determination of their basic parameters ( as age , metallicity and distance ) are then crucial , but this is restricted to a number of relatively nearby stellar systems , for which single stars are spatially resolved and whose magnitudes can be accurately determined . at least the brightest portion of the red giant branch ( rgb ) is needed in order to have hits on metallicity and distance , while the detection of the main sequence turn - off ( ms - to ) is required to get the age . in this respect ,
detailed investigations of simple stellar populations ( ssp , _
i.e. _ coeval and chemically homogeneous stellar aggregates ) offer a unique opportunity to empirically calibrate suitable photometric indices and evolutionary features in terms of the overall metallicity of the system .
stellar clusters represent the best approximation of ssp known in the universe , hence they are ideal tools for this purpose . an empirical method to get simultaneously metallicity and reddening from the
morphology and location of the rgb in the @xmath0 color - magnitude diagram ( cmd ) has been presented a decade ago @xcite .
more recently , this method has been extended to the @xmath1 plane by @xcite and further improved by @xcite and @xcite by adopting the @xcite and the global ( @xmath2 $ ] ) metallicity scale , respectively .
the method presented here adopts the most recent calibrations as obtained from a large ir photometric database of gcs , collected by our group with different instruments at the eso telescopes ( la silla , chile ) and at the tng ( la palma , spain ) over the last decade in the framework of a long - term project devoted to study the photometric properties of the rgb @xcite .
in particular , we have recently presented a complete set of photometric indices ( colors , magnitudes and slopes ) describing the location and the morphology of the rgb and their calibrations in terms of the global cluster metallicity @xcite , together with the empirical calibrations of the ir luminosity of the two major rgb evolutionary features , namely the rgb bump and tip @xcite .
it is worth mentioning that , the calibration relations used in this study rely on a database , whose properties have been derived in a fully self consistent way .
in fact , the adopted estimates of the cluster reddening , metallicity and distance are based _
i ) _ on a homogeneous photometric system ; _ ii ) _ on the @xcite distance scale which relies on the most recent and largest database of galactic gcs ; and _ iii ) _ on a uniform and high resolution metallicity scale @xcite .
the calibration of suitable relations to derive metallicity , reddening and distance in the near ir plane is crucial in the study of extragalactic bulges , which can be characterized by high metallicities and can be affected by severe reddening .
the current generation of ground - based ir instrumentation with high resolution and wide field coverage , and the future availability of the james webb space telescope will allow us to resolve the brightest giants in galaxies up to several mpc away , and to derive their overall metallicity , reddening and distance modulus with great accuracy .
the paper is organized as follows .
2 discusses the adopted metallicity scale , while 3 and 4 present the equation sets for both the _ disk - like _ and _ bulge - like enrichment scenarios _ , respectively .
5 describes the computational routine which provides photometric estimates of metallicity , reddening and distance , while in 6 we test the described method on two template gcs in the galactic bulge and in the large magellanic cloud , namely ngc 6539 and ngc 1978 respectively , in order to demonstrate its reliability .
as widely discussed in @xcite a correct parameterization of the rgb characteristics as function of the metal content of the population does require the knowledge of the so - called _ global _ metallicity , which takes into account the iron as well as the @xmath3element abundances .
this is an important point since the location of the rgb strongly depends on the low ionization potential [ fe+mg+si / h ] abundance mixture @xcite rather than on [ fe / h ] abundance alone .
in fact , low
ionization potential elements are the main contributors to free electrons which generate the @xmath4 ion , the major component responsible for the continuum opacity in the rgb temperature range ( 30006000 k , * ? ? ?
* ) . in halo / disk field
stars the average [ @xmath3/fe ] abundance ratio shows a general enhancement of 0.30.5 dex with respect to the solar value up to [ fe / h]@xmath51 ( see e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* and references therein ) and a linear decreasing trend towards solar [ @xmath3/fe ] with further increasing metallicity .
a [ @xmath3/fe ] enhancement is also found in the metal poor halo gcs ( see e.g. * ? ? ?
* ; * ? ? ?
* and references therein ) .
however , until a few years ago , only a few measurements were available in the high metallicity regime , to properly define the [ @xmath3/fe ] trend in gcs ( see e.g. * ? ? ?
* ; * ? ? ?
the actual position of the knee ( i.e. the metallicity at which [ @xmath3/fe ] begins to decrease ) depends on the type ia sn timescales and it is also a function of the star formation rate , while the amount of @xmath6 enhancement depends on the initial mass function of the progenitors of the type ii sne ( see * ? ? ?
measurements of metal rich field and cluster giants towards the galactic bulge are a recent accomplishment of high resolution optical and ir spectroscopy .
bulge stars are indeed ideal targets to study the behavior of the abundance patterns in the high metallicity domain , but foreground extinction is so great as to largely preclude optical studies of any kind , particularly at high spectral resolution .
the most accurate abundance determinations obtained so far and based on high resolution optical spectroscopy refer to a sample of k giants in the baade s window @xcite and a few giants in two gcs , namely ngc 6553 and ngc 6528 @xcite .
recently , high resolution ir spectroscopic measurements of m giants in the baade s window @xcite as well as of bulge gcs @xcite dramatically extended the sample of measured bulge stars .
all these studies point toward a @xmath3-enhancement by a factor of 2 - 3 up to solar metallicity .
in the computation of the _ global _
metallicity for a sample of 61 ggcs , @xcite used a constant @xmath7=0.28 $ ] for @xmath8<-1 $ ] and linearly decreasing to zero at @xmath9<0 $ ] . in the following
we compute two independent sets of global metallicities and photometric relations , according to two different scenarios ( see fig .
[ scenario ] ) : _ scenario ( 1 ) - _ _ disk - like _ enhancement ( accordingly to @xcite ) @xmath7=0.3 $ ] for @xmath8<-1 $ ] and @xmath7 $ ] linearly decreasing to zero for @xmath9\le0 $ ] . _ scenario ( 2 ) - _ _ bulge - like _
enhancement ( accordingly to carney s ( 1996 ) suggestions and the recent results on the bulge field and cluster populations ) @xmath7=0.30 $ ] constant over the entire range of metallicity ( @xmath10\le0 $ ] ) .
the set of equations discussed in @xcite[hereafter paper i ] and @xcite[hereafter paper ii ] have been computed accordingly to _ scenario ( 1)_. here we present similar equations for _ scenario ( 2 ) _ , so the reader can choose the equation set which turns out to be the most suitable to the describe the stellar system . in both scenarios the contribution of the @xmath3element enhancement has been taken into account by simply re scaling standard models to the _ global _ metallicity [ m / h ] according to the following relation @xcite : @xmath11 = [ fe / h ] + log_{10 } ( 0.638 f_{\alpha}+0.362)\ ] ] where @xmath12 is the enhanced factor of the @xmath3elements . in this section
we present and discuss the equation set adopted to construct the computational routine which provides metallicity , reddening and distance of a stellar population , by using the rgb observables in the @xmath13 ir plane .
note that the validity of the relations has been extended up to [ m / h]@xmath14 + 0.4 dex , with respect to the results presented in paper i , including the recent results on the metallicity of ngc 6791 @xcite .
hence the relations presented here have been obtained in the metallicity range : -2.16@xmath15[fe / h]@xmath15 + 0.4 dex .
the overall metallicity of a stellar population can be linked to the slope of the rgb ( @xmath16 ) as defined in the near ir - cmd , by @xcite and @xcite , accordingly to the following relation : @xmath17 = -2.53 -20.83 ( slope^{jk}_{rgb})\ ] ] where @xmath18 is the rgb slope measured following the prescriptions discussed in paper i in the @xmath13-cmd .
the rgb - tip luminosity is a quantity well predicted by theoretical models and it can be easily measured in populous stellar systems like galaxies .
the luminosity of the rgb tip for stellar populations older than @xmath19 gyr is nearly independent from the population age .
moreover , it turns out to be particularly bright in the near - ir ( @xmath20 ) hence , it is a very promising standard candle up to large distances .
the main limitation of this method is the clear - cut determination of the rgb - tip luminosity that could be seriously affected by low - number statistics .
in fact , the rapid star evolution near the end of the rgb , makes the brightest portion of the rgb intrinsically poorly populated .
however , this problem is marginal in very populous stellar systems .
the dependence of this feature from the overall metallicity of the stellar population has been empirically determined by using a sample of galactic gcs ( see paper ii ) and it turns out to be in good agreement with theoretical expectations . here
we report the relation obtained in paper ii : @xmath21\ ] ] as well known ( see * ? ? ?
* ; * ? ? ?
* ) for clusters older than 2 gyr the level of the helium - burning red clump ( rc ) is mainly influenced by metallicity and it shows little dependence on the age .
hence for relatively old metal rich populations ( @xmath22>-1 $ ] ) it is also possible to use the level of the rc as an additional distance indicator .
we estimate the mean level of the rc by using a metal - rich ( -0.9@xmath23[fe / h]@xmath23 - 0.3 ) cluster subsample ( namely , 47 tuc , ngc 6342 , ngc 6380 , ngc 6441 , ngc 6440 , ngc 6553 , ngc 6528 , and ngc 6637 ) selected from the global database presented in paperi .
the value turns out to be : @xmath24 note that most selected clusters can be considered coeval within 10 - 20% accordingly to @xcite who measured the relative ages of 47 tuc , ngc 6342 , ngc 6637 and @xcite who measured the relative ages of 47 tuc , ngc 6528 , ngc 6553 .
the adopted uncertainty of @xmath25 in equation ( 3 ) , takes into account the metallicity dependence of the rc position in the metallicity range spanned by our metal rich gc subsample ( @xmath26\sim 0.6 dex$ ] ) .
in fact , accordingly to table 1 by @xcite the @xmath27 level is expected to vary @xmath28 at @xmath29 over a such metallicty range .
moreover in the age range 2 - 10gyr , at fixed metallicity ( @xmath2=-0.68 $ ] ) the @xmath27 varies 0.03 mag per gyr . hence the 0.2 mag uncertainty adopted as conservative estimate of the @xmath27 level
is expected to be reasonable over the considered range of age and metallicity .
the absolute magnitude of the rgb at fixed color is another observable that can be used to define useful relations with the overall metallicity .
here we report the calibrations discussed in paper i : @xmath30\ ] ] finally , the entire set of equations describing the rgb location in @xmath31 color at different level of magnitudes ( @xmath32 ) are listed .
@xmath33 = -4.37 + 3.84(j - k)^{-5.5}_0\ ] ] @xmath34 = -4.51 + 4.24(j - k)^{-5}_0\ ] ] @xmath35 = -4.87 + 5.20(j - k)^{-4}_0\ ] ] @xmath36 = -5.36 + 6.48(j - k)^{-3}_0\ ] ] an analogous set of 16 equations can be obtained in the _ bulge - like _ enrichment scenario .
here we list the complete equation set based on the database presented in paper i and ii by adopting a constant @xmath37 $ ] over the entire range of metallicity .
@xmath38 = -2.63 -22.50 ( slope^{jk}_{rgb})\ ] ] @xmath39\ ] ] @xmath40 @xmath41\ ] ] @xmath42 = -4.59 + 4.13(j - k)^{-5.5}_0\ ] ] @xmath43 = -4.74 + 4.55(j - k)^{-5}_0\ ] ] @xmath44 = -5.12 + 5.58(j - k)^{-4}_0\ ] ] @xmath45 = -5.64 + 6.94(j - k)^{-3}_0\ ] ]
the computational routine requires as input parameters : * the observed rgb mean ridge line in the @xmath46 cmd ; * the observed rgb slope ( @xmath18 ) ; * the observed rgb - tip ( @xmath47 ) and/or the rc level ( @xmath48 ) .
[ diagram ] shows the flow - diagram of the computational routine . in the following , references to eqs . 1 - 8 either equations in sect . [ diskeq ] or in sect
. [ bulge ] , depending on the adopted enrichment scenario .
since the rgb slope is independent of the cluster distance and reddening , we can use eq .
( 1 ) to get a first guess of the population metallicity . by using this first guess metallicity , it is possible to derive the expected level of the rgb tip from eq .
hence , the comparison with the observed value ( @xmath47 ) will directly yield the first estimate of the distance modulus @xmath49 .
for the ( closest ) metal - rich stellar systems for which the rc is observed , an independent estimate of the modulus can be also derived from the comparison with the mean rc levels reported in eq .
( 3 ) . by using the first guess metallicity and eq .
( 4 ) it is also possible to obtain the expected absolute magnitude ( @xmath50 ) at fixed color ( @xmath51 ) .
now , the mean ridge line of the program population can be corrected by using the distance modulus obtained above : the observed color ( @xmath52 ) in correspondence of the derived magnitude ( @xmath50 ) will yield the first guess estimate of the reddening : @xmath53 where the extinction coefficients from @xcite have been adopted .
the color of the input rgb mean ridge line can be now de - reddened and transformed into the absolute plane @xmath54 . once in the absolute plane ,
the color at fixed magnitude levels can be measured and inserted in eqs .
( 5 - 8 ) to get a new mean value of metallicity .
the latter can be inserted in eq .
( 2 ) , and the overall procedure iterated until the values of the three output quantities , namely metallicity , reddening and distance , converge within suitable ranges , which can be specified as input tolerances
. the formal error on each derived quantity can be estimate by using a simple monte carlo simulation . in doing this
we randomly extracted a large number of values for each of the three main observables ( namely @xmath18 , @xmath55 and @xmath48 ) from a gaussian distribution peaked on the observed value and with @xmath56 equal to the uncertainty of each observable .
these values are used ( following the scheme plotted in fig . [ diagram ] ) to derive metallicity , reddening and distance .
the standard deviation of each set of values is the error associated to that specific quantity .
however , beside the formal errors obtained from this procedure ( typically @xmath57 mag ; @xmath58 mag ; @xmath59=\pm0.10 - 0.15 $ ] dex ) we estimate that conservative uncertainties for the derived quantities are : @xmath60 mag ; @xmath61 mag ; @xmath59=\pm0.2 $ ] dex .
to test the reliability of the proposed technique we have chosen two metal - rich clusters for which accurate determination of the metallicity have been recently obtained : the large magellanic cloud cluster ngc 1978 , and the galactic bulge cluster ngc 6539 .
the cmds of ngc 1978 and ngc 6539 shown in fig .
[ 1978 ] and [ 6539 ] are from @xcite and from @xcite , respectively .
the rgb mean ridge lines computed following the prescriptions described in paper i are also overplotted as solid line onto the cmds of fig . [ 1978 ] and [ 6539 ] . in the case of ngc 1978 , the cmd shown in fig .
[ 1978 ] has been used to measure the rgb slope ( @xmath18=0.104 ) and to estimate the observed rgb - tip ( k@xmath62=11.89 ) .
it is worth noticing that this cluster is suspected to be a relatively young cluster ( @xmath63 ) , hence we do not use its rc level in deriving the distance modulus level has been used for this cluster . ] since for extreme young clusters the @xmath64 level is a sensible function of the age of the population ( see e.g. fig . 3 and table 1 by * ? ? ?
* ) . by running the computational routine with the input parameters quoted above
, a good convergence is reached after few iterations . in the _ bulge - like scenario _ ,
the computational routine leaded a reddening estimate of e(b - v)[email protected] mag , an intrinsic distance modulus of ( m - m)@[email protected] mag and a metallicity of [ m / h][email protected] dex .
in the _ disk - like scenario _ the following solutions have been obtained : e(b - v)[email protected] mag , ( m - m)@[email protected] mag , and [ m / h][email protected] dex .
these values are fully consistent with the results found by @xcite and @xcite for the extinction ( e(b - v)=0.100 , and e(b - v)=0.075 , respectively ) , by @xcite and @xcite for the distance ( ( m - m)@xmath66=18.48 and ( m - m)@xmath66=18.50 , respectively ) , and by @xcite and @xcite for the metallicity ( [ fe / h]=0.38 dex and [ fe / h]=0.42 dex , based on high - resolution optical spectra and on the calcium triplet , respectively ) .
adopting the same strategy followed in the case of ngc 1978 , we have analyzed the cmd of ngc 6539 @xcite , finding a slope@xmath67=0.091 and k@xmath62=8.47 mag . from the derived k luminosity function we also estimated the rc level to be k@xmath68=13.60 mag . by running the computational routine within the _ bulge - like scenario _ , we found a reddening of e(b - v)[email protected] mag , an intrinsic distance modulus of @xmath69 mag , and a global metallicity [ m / h][email protected] dex .
the derived metallicity is fully consistent within 0.2 dex with the recent estimate , based on high - resolution ( r@xmath70 ) ir spectroscopy , by @xcite , who found [ fe / h]=0.76 dex .
the obtained reddening value is also in excellent agreement with the estimate listed in the @xcite compilation ( e(b - v)=1.00 ) , and with the @xcite extinction map ( e(b - v)=1.09 ) .
our distance value turns out to be 0.1 mag larger than the @xcite estimate0.15 - 0.20 mag longer that the one by @xcite . ] .
this research is supported by the agenzia spaziale italiana ( asi ) and the ministero dellistruzione , delluniversit e della ricerca .
we thank the anonymous referee for his / her helpful comments and suggestions which significantly contributed to improve the paper presentation .
alves , d. r. 2004 , nar , 48 , 659 barbuy , b. , ortolani , s. & bica , e. 1997 , s , 122 , 483 boesgaard , a. m. , king , j. r. , deliyannis , c. p. , & vogt , s. s. 1999 , , 117 , 492 carney , b. w. 1996 , , 108 , 877 carretta , e. & gratton , r. g. 1997 , a&as , 121 , 95 carretta , e. , & bragaglia , a. 1998 , 329 , 937 carretta , e. , gratton , r. g. , & sneden , c. 2000 , , 356 , 238 carretta , e. , cohen , j. , gratton , r.g . , & behr , b. 2001 , , 122 , 1469 de angeli , f. et al .
2005 , , 130 , 116 fahlman , g. g. , douglas , k. a , thompson , i. b. , 1995 , , 110 , 2189 ferraro , f. r. , fusi pecci , f. , guarnieri , m. d. , moneti , a. , origlia , l. , & testa , v. 1994 , , 266 , 829 ferraro , f. r. , fusi pecci , f. , montegriffo , p. , origlia , l. , & testa , v. 1994 , , 298,461 ferraro , f.r . ,
messineo , m. , fusi pecci , f. , de palo , m.a . ,
straniero , o. , chieffi , a. & limongi , m. , 1999 , , 118 , 1738 ferraro f.r .
, montegriffo p. , origlia l. , fusi pecci f. , 2000 , , 119 , 1282 ferraro f.r . , mucciarelli , a. , carretta , e. , origlia , l. 2006 , apj , _ submitted _ fulbright , j.p .
, rich , r.m .
, & mcwilliam , a. 2004 , aas , 205 , 7706 lee , j. , carney , b. w. , balachandran , s. c. 2004 , , 128 , 2388 gratton , r. g. , sneden , c. , carretta , e. , & bragaglia , a. 2000 , , 354 , 169 gratton , r. , sneden , c. , & carretta , e. 2004 , , 42 , 385 grocholski , a. j. & sarajedini , a. 2002 , , 123,1603 harris , w. e. 1996 , , 112 , 1487 , for the 2003 updated version see http://physwww.mcmaster.ca/%7eharris/mwgc.dat kraft , r. p. 1994
, , 106 , 553 kuchinski , l. e. , frogel , j. a. , terndrup , d. m. , & persson , s. e. 1995 , , 109 , 1131 kuchinski , l. e. , & frogel , j. a. 1995 , , 110 , 2844 mcwilliam , a. , & rich , r.m .
1994 , , 91 , 749 mcwilliam , a. 1997 , , 35 , 503 melndez , j. , barbuy , b. , bica , e. , zoccali , m. , ortolani , s. , renzini , a. , & hill , v. 2003 , , 411 , 417 montegriffo , p. , ferraro , f. r. , fusi pecci , f. , & origlia , l. 1995 , , 276 , 739 mucciarelli , a. , origlia , l. , ferraro , f. r. , maraston , c. , testa , v. , 2006 , , in press , astro - ph/0604139 olszewski , e. w. , schommer , r. a. , suntzeff , n. b. , harris , h. c. 1991 , , 101 , 515 origlia , l. , rich , r. m. , & castro , s. 2002 , , 123 , 1559 origlia , l. , & rich , r. m. 2004 , , 127 , 3422 origlia , l. , valenti , e. & rich , r. m. 2005 , , 356 , 1276 origlia , l. , valenti , e. , rich , r. m. & ferraro , f. r 2005 , , 363 , 897 origlia , l. , valenti , e. , rich , r. m. & ferraro , f. r 2006 , , in press , astro - ph/0604030 ortolani , s. , renzini , a. , gilmozzi , r. , marconi , g. , barbuy , b. , bica , e. , rich , r. m. 1995 , natur , 337 , 701 ortolani , s. , barbuy , b. , bica , e. , renzini , a. , zoccali , m. , rich , r. m. , cassisi , s. 2001 , , 376 , 878 persson , s. e. , aaronson , m. , cohen , j. g. , frogel , j. a. , matthews , k. 1983 , , 266 , 105 renzini , a. 1977 , in advanced stages in stellar evolution , ed .
p. bouvier & a. maeder ( geneva : geneva obs . ) , 151 rich , r.m . , & origlia , l. 2005 , , 634 , 1293 rich , r.m . , &
mcwilliam , a. 2000 , spie , 4005 , 150 salaris , m. , & cassisi , s. 1996 , , 305 , 858 salaris , m. , chieffi , a. , straniero , o. , 1993 , , 414 , 580 salaris , m. & girardi , l. 2002 , , 337 , 332 sarajedini , a. 1994 , , 107 , 618 sarajedini , a. & layden , a. 1997 , , 113 , 264 savage , b.d . ,
mathis , j.s .
, 1979 , , 17 , 73 schlegel , d. j. , finkbeiner , d. p. , davis , m. 1998 , , 500 , 525 sneden , c. , kraft , r.p . ,
guhathakurta , p. , peterson , r.c.,fulbright , j.p .
2004 , , 127 , 2162 sollima , a. , ferraro , f. r. , origlia , l. , pancino , e. , & bellazzini , m. 2004 , , 420 , 173 straniero , o. , & chieffi , a. 1991 , , 76 , 525 valenti , e. , ferraro , f. r. , perina , s. , & origlia , l. 2004 , , 419 , 139 valenti , e. , ferraro , f. r. , & origlia , l. 2004 , , 351 , 1204 , _ paper i _ valenti , e. , ferraro , f. r. , & origlia , l. 2004 , , 404 , 157 , _ paper ii _ van den bergh , s. 1998 , , 110 , 1377 zocali , m. , barbuy , b. , hill , v. , ortolani , s. , renzini , a. , bica , e. , momany , y. , pasquini , l. , minniti , d. , rich , r. m. 2004 , , 423 , 507 | we present an empirical method to derive photometric metallicity , reddening and distance to old stellar populations by using a few major features of the red giant branch ( rgb ) in near ir color magnitude diagrams .
we combine the observed rgb features with a set of equations linking the global metallicity [ m / h ] to suitable rgb parameters ( colors , magnitudes and slope ) , as calibrated from a homogeneous sample of galactic globular clusters with different metallicities .
this technique can be applied to efficiently derive the main population parameters of old stellar systems , in the view of using ground - based adaptive optics and space facilities to probe the stellar content of remote galaxies . |
theoretical description of hadronic amplitudes of the quark currents is one of the key problems of particle physics as such amplitudes provide a bridge between qcd formulated in the language of quarks and gluons and observable phenomena which deal with hadrons . in particular , the knowledge of such amplitudes
is necessary for extraction of the parameters of the quark mixing matrix in the standard model from the experiments on weak heavy hadron decays .
the difficulty in such calculations lie in the fact that hadron formation occurs at large distances where perturbative qcd methods are not applicable and nonperturbative consideration is necessary . various theoretical frameworks more or less directly related to qcd have been applied to the description of meson transition form factors : among them constituent quark models @xcite , lattice qcd simulations @xcite , qcd sum rules @xcite , analytical constraints @xcite .
lattice qcd simulations is the most direct qcd based nonperturbative approach and thus should in principle provide most reliable results .
however it faces serious technical problems with placing heavy particles on the lattice and performing calculations with inclusion of quark loops .
so far direct calculations with @xmath0-quark are not possible and extrapolation in the quark mass is used which considerably reduces the predictive power of the method .
various versions of qcd sum rules to meson decays give very uncertain predictions strongly dependent on the technical subtleties of the particular version .
a recent analysis @xcite disregards the three point qcd sum rules in favor of the light
cone sun rules which however involve more phenomenological inputs .
constituent quark models have proved to be a fruitful phenomenological method for the description of heavy meson transition form factors .
constituent quark models employ the fact that heavy meson consists of a heavy quark and light degrees of freedom with have quantum numbers of the quark state and assume that these light states can be approximately described by an effective constituent quark .
the application of the various versions of the constituent quark picture to the decay processes has a long history .
the first models were based on a semirelativistic @xcite or nonrelativistic @xcite considerations and did nt treat the quark spins thoroughly .
further developments demonstrated the relativistic description of the quark spins to be important for consistent applications to meson decays .
the exact solution to a complicated dynamical problem of treating the spins of the interacting particles is not known , but a simplified self consistent relativistic consideration of the quark spins can be performed within the light cone quark model ( lcqm ) @xcite .
the difficulty with the application of this approach to the decay processes lies in the fact that the applicability of the model is restricted by the condition @xmath1 , while the physical region for hadron decays is @xmath2 , @xmath3 being the initial and final hadron masses , respectively .
the problem is connected with the contribution of the so
called pair creation subprocesses which can not be taken into account thoroughly in the model , except for a trivial case of a pointlike interaction . at spacelike momentum
transfers the contribution of these subprocesses can be killed by choosing an appropriate reference frame , whereas at timelike momentum transfers such a frame does not exist and thus pair creation contributes together with the partonic contribution .
an estimate of the form factors at timelike momentum transfers within the lcqm has been done in @xcite where the transition form factors have been calculated in the spacelike region and then numerically extrapolated to the timelike region assuming some particular @xmath4behavior .
this works well if the accessible momentum transfer interval is not large .
however , in the heavy to light meson decays this interval is of order of the heavy meson mass squared and the extrapolation procedure yields large uncertainties .
the approach of refs @xcite calculates the partonic contribution to the form factors at timelike momentum transfers and thus avoids extrapolations .
however , the nonpartonic contribution is omitted .
such treatment can be justified if the latter is small that is not obvious @xmath5 .
thus , one can see that for a reliable description of the decay processes within the lcqm it might be reasonable to find another formulation of the model appropriate also at timelike momentum transfers .
such a formulation has been proposed in @xcite .
the approach of @xcite is based on the dispersion formulation of the lcqm .
namely , the transition form factors obtained within the light
cone formalism at @xmath6 @xcite , are represented as dispersion integrals over initial and final hadron masses through their light cone wave functions .
the spectral densities of these representations can be calculated from the feynman graphs .
the transition form factors at @xmath7 are derived by performing the analytical continuation in @xmath4 from the region @xmath1 . as a result , for a decay caused by the weak transition of the quark @xmath8 , form factors in the region @xmath9 are expressed through the light cone wave functions of the initial and final hadrons and can be calculated in the decay region avoiding dangerous extrapolations .
at the same time , such a dispersion formulation of the lcqm allows also a bit different view on the problem of constructing the form factors within the quark model : actually , we calculate the double spectral densities from the feynman graphs and thus obtain unsubtracted spectral representations .
on the other hand we know that in general subtractions might be added to such representations and we need some additional arguments to decide on the necessity and the structure of the subtraction terms .
let us notice that once we are working within an approach not directly deduced from qcd it is important to preserve essential features of the underlying fundamental theory in the model .
thus matching the results obtained within the quark model to rigorous qcd results might be helpful for bringing more realistic features to the model .
we use such matching to qcd results for determining the subtraction terms in the double spectral representations of the form factors .
there are few qcd predictions on meson form factors in the decay region .
fortunately , there is an important case in which qcd provides rigorous and model independent results on transition form factors : namely , meson decays induced by transitions between heavy quarks . in this case the heavy
quark symmetry simplifies all the form factors in the leading @xmath10 order to the universal isgur
wise function @xcite .
a systematic expansion in inverse powers of the heavy quark mass can be constructed within the heavy quark effective theory @xcite based on qcd with the heavy quarks .
although the form factors appearing in different @xmath10 orders can not be determined within the hqet , the latter restricts the structure of the expansion and the number of these universal form factors in each @xmath10 order and provides constraints on some of them at zero recoil .
the idea of matching the quark model form factors to the structure of the heavy
quark expansion of hqet has been employed in the isgw2 model @xcite .
however the isgw2 model does not calculate the form factors within a relativistic dynamical approach on the basis of considering the meson structure but rather gives prescription for constructing the form factors .
the approach of ref .
@xcite also employs matching to the heavy
quark expansion and obtains interesting results within the constutuent quark picture assuming a strong peaking of the momentum distribution of the constituent quarks inside a meson and avoiding details of the dynamics of the process .
this model imposes constraints on the form factors but does not allow one to calculate them .
the idea implemented in this work is to perform the heavy
quark expansion in the dispersion formulation of the quark model and to match this expansion to the hqet .
this matching is used for determining the structure of the subtraction terms in the spectral representations .
we consider the form factors of pseudoscalar meson transitions into pseudoscalar and vector final mesons through the vector , axial vector , and currents and analyze the expansion in the leading and next to leading order .
further constraints on the subtraction terms are obtained by considering the case of the heavy to light quark transition and analyzing the relations between the form factors of the tensor , vector and axial vector currents .
comparing the relations obtained within the dispersion quark model to the general relations found by isgur and wise @xcite allows us to fix the subtraction terms .
our main results are the following : + 1 . applying the approach of @xcite to heavy meson transition
throught the tensor current , we derive the decomposition of the form factors @xmath11 , @xmath12 , @xmath13 , and @xmath14 ( the definitions are given in the next section ) in the next to leading order in @xmath10 .
in particular , we have found that the @xmath10 corrections to the form factor @xmath11 at zero recoil vanish just as for the form factors @xmath15 and @xmath16 ( the luke theorem ) .
although the heavy meson transitions through the tensor current do not correspond to a real experimental situation we need the expansion for such form factors as a bench mark for comparison of the quark model .
we present an improved dispersion formulation of the lcqm : the transition form factors of pseudoscalar ( @xmath17 ) mesons into pseudoscalar ( @xmath17 ) and vector ( @xmath18 ) mesons induced by the vector , axial
vector and tensor currents are given by double spectral representations through the wave functions of the initial and final mesons .
the unsubtracted double spectral densities are calculated from the feynman graphs and the subtraction terms in the spectral representations are determined by matching the @xmath10 expansion in the quark model to the hqet .
we perform the heavy
quark expansion of the form factors of the dispersion quark model up to nlo and calculate the isgur wise function and the nlo universal form factors assuming a strong peaking of the soft wave function near the @xmath19 threshold with a width of order @xmath20 .
we obtain in the lcqm the relations [ nloffs ] _ 2()=0;_3()=0;_3()>0 , _
3(1)=z /3 , where @xmath21 is an average kinetic energy of the light quark in the heavy meson rest frame .
we observe that matching the @xmath10 expansion of the quark model to hqet in lo and nlo requires subtractions in some of the form factors describing the @xmath22 transition and restricts the form of the subtraction terms .
we consider the heavy to light meson transitions in which case a small parameter @xmath23 emerges and analyze the form factors in the leading @xmath23 order .
requiring the fulfillment of the isgur wise relatins for the heavy to light transitions @xcite further constrains the subtraction terms providing explicit spectral representations with subtractions for the form factors of interest . in the next section
we remind the definitions of the form factors and list the hqet results for the amplitudes of the @xmath24 and @xmath22 transitions through the vector , axial
vector and tensor currents up to nlo . in section 3
we present the quark
model results for the meson decay form factors at arbitrary masses : namely , we calculate the unsubtracted spectral densities and explicit form of the subtraction terms which form is motivated in the subsequent section . in section 4
the @xmath10 expansions of the wave functions and the form factors in quark model for the heavy to light transitions are considered . in section 5 the heavy
to light case is discussed .
section 6 illustrates the main results with numerical estimates and evaluate the isgur wise function for various quark model parameters .
conclusion summarizes the results .
the amplitudes of meson decays induced by the quark transition @xmath25 through the vector @xmath26 , axial vector @xmath27 , and tensor @xmath28 currents have the following structure @xcite @xmath29 , \nonumber \\
< p(m_2,p_2)|t_{\mu\nu}(0)|p(m_1,p_1)>&=&-2i\,s(q^2)\,(p_{1\mu}p_{2\nu}-p_{1\nu}p_{2\mu } ) , \nonumber \\
< v(m_2,p_2,\epsilon)|t_{\mu\nu}(0)|p(m_1,p_1)>&=&i\epsilon^{*\alpha}\ , [ \,g_{1}(q^2)\epsilon_{\mu\nu\alpha\beta}p^{1\beta}+ g_{2}(q^2)\epsilon_{\mu\nu\alpha\beta}p^{2\beta}+ g_0(q^2)p_{1\alpha}\epsilon_{\mu\nu\beta\gamma}p_1^{\beta}p_2^{\gamma}\ , ] , \end{aligned}\ ] ] with @xmath30 , @xmath31 .
we use the notations : @xmath32 , @xmath33 $ ] , @xmath34 , and @xmath35 . the relativistic
invariant form factors contain the dynamical information on the process and should be calculated within a nonperturbative approach for any particular initial and final mesons . for analysing the transition in the case when both the parent and the daughter quarks inducing the meson transition are heavy , i.e. @xmath36 it is convenient to introduce a new dimensionless variable @xmath37 and velocity dependent form factors connected with 4velocities and not 4momenta as in ( [ ffdef ] ) in the following way @xmath38 , \nonumber \\ < v(m_2,p_2,\epsilon)|v_\mu(0)|p(m_1,p_1)>&=&\sqrt{m_1m_2}\;h_g(\oh)\epsilon_{\mu\nu\alpha\beta } \epsilon^{*\nu}\,v_1^{\alpha}\,v_2^{\beta } , \nonumber \\
< v(m_2,p_2,\epsilon)|a_\mu(0)|p(m_1,p_1)>&= & i\epsilon^{*\alpha}\,\sqrt{m_1m_2}\ , [ h_f(\oh)(1+\oh)g_{\mu\alpha}-h_{a_1}(\oh)v_{1\alpha}v_{1\mu } -h_{a_2}(\oh)v_{1\alpha}v_{2\mu}\ , ] , \nonumber \\
< p(m_2,p_2)|t_{\mu\nu}(0)|p(m_1,p_1)>&=&-2i\,\sqrt{m_1m_2}\;h_s(\oh)\,(v_{1\mu}v_{2\nu}-v_{1\nu}v_{2\mu } ) , \nonumber \\
< v(m_2,p_2,\epsilon)|t_{\mu\nu}(0)|p(m_1,p_1)>&=&i\epsilon^{*\alpha}\sqrt{m_1m_2}\ , [ \,h_{g_+}(\oh)\epsilon_{\mu\nu\alpha\beta}(v_1+v_2)^{\beta}+ h_{g_-}(\oh)\epsilon_{\mu\nu\alpha\beta}(v_1-v_2)^{\beta}\nonumber\\ & & + h_{g_0}(\oh)v_{1\alpha}\epsilon_{\mu\nu\beta\gamma}v_1^{\beta}v_2^{\gamma}\ , ] . \end{aligned}\ ] ] these form factors are related to the form factors introduced by the relations ( [ ffdef ] ) as follows @xmath39,\quad g_1= \frac{m_2}{\sqrt{m_1m_2}}[h_{g_+}+h_{g_-}],\quad \nonumber\\ f_2&=&\frac{m_1}{\sqrt{m_1m_2}}[h_{f_+}-h_{f_-}],\quad g_2= \frac{m_1}{\sqrt{m_1m_2}}[h_{g_+}-h_{g_-}],\quad \nonumber\\ g&=&\frac{1}{2\sqrt{m_1m_2}}h_g,\quad s=\frac{1}{2\sqrt{m_1m_2}}h_s,\quad \nonumber\\ f&=&\sqrt{m_1m_2}(1+\oh)h_f,\quad a_2=-\frac{1}{\sqrt{m_1m_2}}h_{a_2},\quad a_1=-\frac{1}{\sqrt{m_1m_2}}\frac{m_2}{m_1}h_{a_1},\quad g_0=\frac{1}{\sqrt{m_1m_2}}\frac{1}{m_1}h_{g_0}\end{aligned}\ ] ] the form factors @xmath40 are convenient quantities as in the leading @xmath10 order all of them are expressed through a single universal function of the dimensionless variable @xmath41 the isgur wise function . a consistent heavy
quark expansion of the form factors , i.e. expansion in inverse powers of the heavy
quark mass , can be constructed within the heavy quark effective theory based on qcd with heavy quarks .
the general structure of the @xmath10 expansion of the heavy quark form factors in qcd for the meson transition @xmath42 induced by heavy quark transition @xmath43 have the form ( omitting corrections @xmath44 : @xmath45 in the leading @xmath10 order ( lo ) all the form factors are represented through the single universal isgur
wise function @xmath46 , whereas in the next - to - leading order ( nlo ) the 4 new form factors @xmath47 , and @xmath48 appear .
the universal form factors are functions of a single variable @xmath41 .
the form factor @xmath48 originates from the expansion of the transition quark current , and the form factors @xmath49 are connected with the nontrivial relationship between the mesonic states in the full and the effective theory .
the universal form factors satisfy the conditions @xmath50 whereas no constraints on @xmath48 and @xmath51 are imposed by the heavy quark symmetry . as found by luke @xcite ,
the @xmath10 corrections to the form factors @xmath16 and @xmath15 vanish due to kinematical or dynamical reasons .
one can see that the same is true for the form factor @xmath11 : namely , @xmath52 the parameter @xmath53 in ( [ hqexp ] ) comes from the @xmath10 expansion of the mass of a meson consisting of the heavy quark and light degrees of freedom @xmath54 in our notations for heavy quarks and mesons , this gives @xmath55 for the parent and daughter particles , respectively .
it is straightforward to derive the following useful relations @xmath56\nonumber\\ \frac{m_1-m_2}{\sqrt{m_1m_2}}&=&\frac{m_2-m_1}{\sqrt{m_1m_2 } } \left[{1-\left ( \frac1{m_1}+\frac1{m_2}\right)\bl2+\ldots } \right]\nonumber\\ \sqrt{m_1m_2}&=&\sqrt{m_1m_2 } \left[{1+\left ( \frac1{m_1}+\frac1{m_2}\right)\bl2+\ldots } \right],\end{aligned}\ ] ] where the dots denote higher order terms . using the relations ( [ ff2hqff ] ) and ( [ m2 m ] )
, we obtain for the form factors ( [ ffdef ] ) the following expansions @xmath57,\\ \label{f2exp } f_2&=&\frac{m_2}{\sqrt{m_1m_2}}\left[{\xi+\frac1{m_1}\left(\rho_1-\xi_3\right ) + \frac1{m_2}\left(\rho_1+\xi_3\right)}\right],\\ \label{sexp } s&= & \frac{1}{2\sqrt{m_1m_2}}\left[{\xi+\frac1{m_1}\left(\rho_1-\xi_3\right ) + \frac1{m_2}\left(\rho_1-\xi_3\right)}\right],\\ \label{g1exp } g_1&=&-\frac{m_1}{\sqrt{m_1m_2}}\left[{\xi+\frac1{m_1}\rho_2+\frac1{m_2}\left(\rho_1-\xi_3\right)}\right],\\ \label{g2exp } g_2&=&-\frac{m_2}{\sqrt{m_1m_2}}\left[{\xi+\frac1{m_1}\rho_2+\frac1{m_2}\left(\rho_1+\xi_3\right)}\right],\\ \label{gexp } g&= & \frac{1}{2\sqrt{m_1m_2}}\left[{\xi+\frac1{m_1}\rho_2 + \frac1{m_2}\left(\rho_1-\xi_3\right)}\right],\\ \label{a1exp } a_1&=&-\frac{1}{\sqrt{m_1m_2}}\frac1{m_2}\frac1{\oh+1}\left[{-\bar\lambda\xi+2(\oh+1)\chi_2-\xi_3}\right],\\ \label{a2exp } a_2&=&-\frac{1}{\sqrt{m_1m_2}}\left[{\xi+\frac1{m_1}\left(\rho_1-\xi_3\right ) + \frac1{m_2}\left(\rho_1-\xi_3\right ) -\frac{\bar\lambda}{m_1}\frac1{\oh+1}\xi-\frac{2\chi_2}{m_1}+\frac1{m_1}\frac{\oh}{\oh+1}}\xi_3\right],\\ \label{fexp } f&=&{\sqrt{m_1m_2}}(\oh+1)\left[{\xi+\frac{\bar\lambda\xi\oh}{\oh+1}\left(\frac1m_1+\frac1m_2\right)+\frac1m_1\rho_2 + \frac1m_2\left(\rho_1-\frac{\oh-1}{\oh+1}\xi_3\right ) } \right]\\ \label{g0exp } g_0&=&\frac{1}{(m_1m_2)^{3/2}}\left[\frac{\bla\xi+\xi_3}{\oh+1}+2\chi_2\right ] \end{aligned}\ ] ] for the following analysis it is worth noting that the behavior of the combination @xmath58 and @xmath59 in lo and nlo coincide , namely @xmath60 it is also convenient to introduce the form factor @xmath61 such that @xmath62 in what follows we need the expansions of the following linear combinations of the form factors @xmath61 and @xmath63 @xmath64,\\ \label{aplus } a_1m_2+a'_2m_1&=&-\frac1{\sqrt{m_1m_2}}\left [ -\frac{2\bla\xi}{\oh+1}+\xi_3 \frac{\oh-1}{\oh+1 } \right].\end{aligned}\ ] ]
the results presented in the previous section are strict consequences of qcd in the heavy quark limit which however can not provide more information on the universal form factors @xmath65 and @xmath66 .
they must be calculated within a nonperturbative dynamical approach .
we study the form factors within the dispersion formulation of the quark model @xcite which has proved to be a reasonable framework for describing meson decays .
we start with @xmath6 and represent the form factors as double spectral representations in the invariant masses of the initial and final @xmath19 pairs .
the form factors at @xmath7 are derived by performing the analytical continuation .
the transition of the initial meson @xmath67 with the mass @xmath68 to the final meson @xmath69 with the mass @xmath70 induced by the quark transition @xmath43 through the current @xmath71 is described by the diagram of fig.[fig : trianglegraph ] . for constructing the double spectral representation we must consider a double cut graph where all intermediate particles go on mass shell but the initial and final mesons have the off
shell momenta @xmath72 and and @xmath73 such that @xmath74 and @xmath75 with @xmath76 kept fixed . for the transition @xmath77 @xmath78 , and @xmath79 .
the constituent quark structure of the initial and final mesons is given in terms of the vertices @xmath80 and @xmath81 , respectively .
the initial pseudoscalar meson vertex has the spinorial structure @xmath82 ; the final meson vertex has the structure @xmath83 for a pseudoscalar state and the structure @xmath84\,g_2/\sqrt{n_c}$ ] , @xmath85 , @xmath86 for an @xmath87wave vector meson .
the double spectral densities @xmath88 of the form factors are obtained by calculating the relevant traces and isolating the lorentz structures depending on @xmath72 and @xmath73 .
the invariant factors of such lorentz structures provide the double spectral densities @xmath88 corresponding to taking into account contributions of the two particle singularities in the feynman graph .
let us point out that this procedure allows one to obtain unsubtracted double spectral densities , whereas subtraction terms should be determined independently . in this paper
we determine the subtraction terms from matching the @xmath10 expanded form factors of the quark model to the heavy
quark expansion in qcd . at @xmath6
the spectral representations of the form factors have the form @xcite @xmath89 where the wave function @xmath90 and @xmath91 and @xmath92 is the triangle function .
the unsubtracted double spectral densities @xmath93 of the form factors read @xcite : @xmath94 , \\
\label{f1 } \tilde f_1&=&2m_1\tilde s+4\alpha_2[s_2-(m_1-m_3)^2]-2m_3\tilde s , \\
\label{f2 } \tilde f_2&=&2m_2\tilde s+4\alpha_1[s_1-(m_2-m_3)^2]-2m_3\tilde s , \\
\label{g } \tilde g&=&a\tilde s-4b\beta , \\
\label{g1 } \tilde g_{1}&=&a\tilde f_1 - 8\beta+8b(m_1+m_3)\beta , \\
\label{g2 } \tilde g_{2}&=&a\tilde f_2 + 8b(m_2-m_3)\beta , \\
\label{a2 } \tilde a_{2d}&=&-2\tilde s+ 4bc_2\alpha_1+\alpha_{12}c_0 , \\
\label{a1 } \tilde a_{1d}&=&-4a\,(2m_3+bc_1)\alpha_1+\alpha_{11}c_0 , \\
\label{f } \tilde f_d&=&-4a[m_1m_2m_3+\frac{m_2}2(s_2-m_1 ^ 2-m_3 ^ 2 ) + \frac{m_1}2(s_1-m_2 ^ 2-m_3 ^ 2)-\frac{m_3}2(s_3-m_1 ^ 2-m_2 ^ 2)]+c_0\beta , \\
\label{g0 } \tilde g_{0d}&=&-8a\alpha_{12}-8b\,[-m_3\alpha_{1}+(m_3-m_2)\alpha_{11}+(m_3+m_1)\alpha_{12}],\end{aligned}\ ] ] where @xmath95/{\lambda(s_1,s_2,s_3 ) } , \\
\label{alpha2 } \alpha_2&= & \left[(s_1+s_2-s_3)(s_1-m_2 ^ 2+m_3 ^ 2)-2s_1(s_2-m_1 ^ 2+m_3 ^ 2)\right]/{\lambda(s_1,s_2,s_3 ) } , \\
\label{beta } \beta&=&\frac14\left[2m_3 ^ 2-\alpha_1(s_1-m_2 ^ 2+m_3 ^ 2)-\alpha_2(s_2-m_1 ^ 2+m_3 ^ 2)\right ] , \\ \label{alpha11 }
\alpha_{11}&=&\alpha_1 ^ 2 + 4\beta { s_2}/{\lambda(s_1,s_2,s_3 ) } , \quad \alpha_{12}=\alpha_1\alpha_2 - 2\beta(s_1+s_2-s_3)/\lambda(s_1,s_2,s_3 ) , \\
\label{cc } c_0&=&-8a(m_2-m_3)+4bc_3,\quad c_1=s_2-(m_1+m_3)^2 , \\
c_2&=&s_1-(m_2-m_3)^2 , \quad c_3=s_3-(m_1+m_2)^2-c_1-c_2 . \nonumber\end{aligned}\ ] ] we label with a subscript d the double spectral densities of the form factors which require subtractions .
we fix this subtraction procedure by requiring the @xmath10 expansion of the form factors to have a proper form in accordance with qcd in leading and next to leading orders for the case of a meson transition caused by heavy to heavy quark transition . as we shall see this yields the double spectral densities which include properly defined subtraction terms @xmath96\tilde g , \\
\label{a1+sub } \tilde a_1&=&\tilde a_{1d } + \frac{1}{(\oq+1)m_2}\left(\frac{m_1 ^ 2-s_1}{\sqrt{s_1}}+\frac{m_2 ^ 2-s_2}{\sqrt{s_2}}\right ) \frac{\tilde g}2,\\ \label{a2+sub } \tilde a_2&=&\tilde a_{2d } + \frac{1}{(\oq+1)m_1}\left(\frac{m_1 ^ 2-s_1}{\sqrt{s_1}}+\frac{m_2 ^ 2-s_2}{\sqrt{s_2}}\right ) \frac{\tilde g}2,\\ \label{g0+sub } \tilde g_0&=&\tilde g_{0d } + \frac{1}{(\oq+1)m_1m_2}\left(\frac{m_1 ^ 2-s_1}{\sqrt{s_1}}+\frac{m_2 ^ 2-s_2}{\sqrt{s_2}}\right ) \frac{\tilde g}2.\end{aligned}\ ] ] as the analytical continuation to the timelike region is performed , in addition to the normal contribution which is just the expression ( [ ff ] ) taken at @xmath7 , anomalous contribution emerges @xcite .
the origin of the anomalous contribution is connected with the motion of a zero of the triangle function @xmath97 from the unphysical sheet at @xmath6 onto the physical sheet at @xmath7 through the normal cut between the points @xmath98 and @xmath99 . to be more specific
, we can write ( s_1,s_2,s_3)=(s_1-s_1^l)(s_1-s_1^r ) , where @xmath100 and @xmath101 . at @xmath6 both singularities @xmath102 and @xmath103 are located on the unphysical sheet and do not contribute to the spectral representation of the form factor . as @xmath4 becomes positive , @xmath103 moves around @xmath104 and appears on the physical sheet .
the pinching of the points @xmath103 and @xmath104 occurs at @xmath105 such that @xmath106 .
the corresponding value @xmath107 reads @xmath108 the spectral representation of the form factor at @xmath7 takes the form @xcite [ final ] f(q^2)= _ ( m_1+m_3)^2^_2(s_2 ) _
s_1 ^ -^s_1^+ _ 1(s_1 ) + ( q^2)_s_2 ^ 0^_2(s_2 ) _ s_1^r^s_1 ^ - ds_1_1(s_1 ) .
one should also take into account that if the double spectral density @xmath109 is singular at the point @xmath103 then a properly defined spectral representation contains also another kind of subtraction terms : these terms appear as the contribution of a small circle around @xmath103 which is the lower integration limit in the anomalous part .
these subtractions have quite different nature and different form than subtractions in the spectral representations at @xmath6 .
the corresponding expression can be found in @xcite .
the normal contribution dominates the form factor at small timelike and vanishes as @xmath110 while the anomalous contribution is negligible at small @xmath4 and steeply rises as @xmath111 . for pseudoscalar and vector mesons with the mass @xmath112 built up of the constituent quarks @xmath113 and @xmath114 ,
the function @xmath115 is normalized as follows @xcite @xmath116=1.\ ] ] this equation is the normalization of the elastic charge form factor at @xmath117 .
the meson wave function can be written in the form @xmath118 where @xmath119 is the ground
state @xmath87wave radial wave function . in the next sections we analyze the form factors given by the dispersion representation ( [ final ] ) with the spectral densities ( [ s][g2 ] ) and ( [ f+sub][g0+sub ] ) and
demonstrate them to have the following properties in the case of a heavy parent meson : for the transition induced by the heavy to heavy quark transition they satisfy the lo and nlo relations @xcite of the @xmath10 expansion in accordance with qcd provided the functions @xmath120 are localized near the @xmath19 threshold with the width of order @xmath20 .
for the meson decay induced by the heavy to light quark transition they satisfy the lo relations between the form factors of the vector , axial vector and tensor currents @xcite .
in this section we consider the form factors of the dispersion quark model in the case when both @xmath121 and @xmath122 are large .
we calculate the universal form factors and demonstrate that requiring the structure of the @xmath10 expansion in the quark model to be consistent with the structure of such expansion in qcd allows us to determine the subtraction terms .
first , we need to specify the properties of the soft wave function of a heavy meson .
a basic property of such soft wave function @xmath123 is a strong peaking near the @xmath19 threshold . for elaborating the @xmath10 expansion ,
it is convenient to formulate such peaking in terms of the variable @xmath124 such that @xmath125 ( hereafter we denote the mass of the light quark as @xmath126 ) .
the region above the @xmath19 threshold which contributes to the spectral representation corresponds to @xmath127 .
a localization of the soft wave function in terms of @xmath124 means that the wave function is nonzero as @xmath128 . in the heavy meson case @xmath129 .
let us notice that for a heavy meson the localization in terms of @xmath124 is equivalent to the localization in terms of the relative momentum in the meson rest frame @xmath130 the normalization condition ( [ norma ] ) which is a consequence of the vector current conservation in the full theory provides an ( infinite ) chain of relations in the effective theory .
namely , expanding the soft wave function in @xmath10 as follows @xmath131,\ ] ] we come to the normalization condition in the form @xmath132=1.\ ] ] this exact relation is equivalent to an infinite chain of equations in different @xmath10 orders .
lowest order relations take the form @xmath133 in the description of the transition processes the dispersion formulation of the quark model has the following feature : since the underlying process is the quark transition , the relevant kinematical variable for the description of the dynamics of the process is the quark recoil @xmath134 which is defined as follows q^2=(m_2-m_1)^2 - 2m_1m_2(-1 ) .
the relationship between @xmath41 and @xmath134 is given by the condition that the spectral representation for the form factor is written at fixed value of @xmath4 and hence q^2=(m_1-m_2)^2 - 2m_1m_2(-1)=(m_2-m_1)^2 - 2m_1m_2(-1 ) . in the case of heavy particle transitions
these quantities are related to each other as = + ( + ) ( -1)+o(1/m_q^2 ) .
we shall obtain the representations of the form factors as functions of the variable @xmath134 .
the variables @xmath41 and @xmath134 are different by the terms of order @xmath10 at nonzero recoil . on the other hand ,
the quark and meson zero recoil points coincide with @xmath135 accuracy .
this means that in the analyses of the @xmath10 expansion at nonzero recoil the difference between the @xmath41 and @xmath134 might be ignored in the isgur wise function , but gives nontrivial contribution to the nlo form factors . at the same time , at zero recoil the difference between @xmath41 and @xmath134 might be neglected both in the leading and next to leading orders .
namely , the quark model provides the expansion of the form factor in the following form h&=&h_0()+h^(1)_1()+ h^(2)_1()+
= + & = & h_0()+h_0()(+)(-1 ) + h^(1)_1()+|h^(2)_1()+ .
as we shall see later , among the nlo form factors only @xmath136 are affected by the the difference between @xmath41 and @xmath134 whereas @xmath137 are not .
we are going now to demonstrate that the anomalous contribution comes into the game only in close vicinity of the zero recoil point whereas beyond this region is negligible .
let us study the behavior of the anomalous contribution in the region -1m_q^-(2 + ) . introducing the variables @xmath138 and
@xmath139 such that @xmath140 and @xmath141 we find that the magnitude of the anomalous contribution is controlled by the value of @xmath142 such that @xmath143 which is the lower boundary of the @xmath139 integration . if @xmath142 becomes large , i.e. of the order @xmath144 , the anomalouis contribution is suppressed by the fall down of the soft wave function .
this suppression is at least stronger than @xmath135 .
this means that the anomalous contribution is nonvanishing only if z_2 ^ 0()=-m_1+o(m_3 ) . in the region
@xmath145 , one finds @xmath146 .
hence the anomalous contribution comes actually into the game only in the @xmath147 vicinity of the zero recoil point but otherwise might be neglected . on the other hand , at the quark zero recoil point @xmath148 , the normal contribution vanishes and the form factor is given by the anomalous contribution . we shall calculate the form factors in the region @xmath149 where only the normal contribution should be taken into account in leading and subleading orders . to perform the lo analysis of the form factors
let us start with the integration measure . with @xmath10 accuracy it can be represented in the form 116 ^ 2 14 ^ 2dz_2(z_2)(^2<1 ) , and the expression for @xmath138 reads [ z1 ] z_1=z_2+m_3(-1)++o(1/m_q ) .
let us point out that the lo integration measure is symmetric in @xmath138 and @xmath139 .
next , we need expanding the spectral densities ( [ s][g0 ] ) . to this end
we must take into account that under the integral sign @xmath138 and @xmath139 are localized in the region @xmath150 due to the soft wave functions @xmath151 . in lo the kinematical coefficients ( [ alpha1][cc ] ) in the region
@xmath149 simplify to ( s_1,s_2,q^2)&=&4m_1 ^ 2m_2 ^ 2(^2 - 1 ) , + _
1&=&1m_2(+1 ) , + _
2&=&1m_1(+1 ) , + & = & 12 , + _ 11&=&_1 ^ 2 + , _ 12=_1_2- , + c_0&=&-4m_2(-1),b=12m_1 , c_1=2m_1z_2,c_2=2m_2(z_1 + 2m_3 ) .
one finds the lo behavior of the form factor densities ( [ s][a2 ] ) is determined by the term proportional to @xmath152 .
the latter reads in the lo s2 ( m_3 + ) .
the lo expression for @xmath88 takes the form f_d=(+1)(m_3 + ) .
the spectral densities @xmath153 and @xmath154 vanish in the leading order .
hence the lo relations ( [ f1exp][g0exp ] ) are fulfilled with the isgur wise ( iw ) function @xmath155 in ( [ xi ] ) we used the equality of @xmath134 and @xmath41 with @xmath10 accuracy .
the normalization condition of the lo wave functions ( [ normaphi ] ) yields @xmath156 . for the slope of the iw function at zero recoil , @xmath157 ,
one finds @xmath158 let us point out that the subtraction terms in the spectral densities do not contribute in the lo relations . as we shall see later , they are important in the nlo analysis .
first , let us concentrate on the nlo relations ( [ f1exp][gexp ] ) .
it is convenient to analyze the linear combinations of the form factors which do not contain the lo contribution .
these combinations are @xmath165,\\ \label{2exp } g_1+f_1&=&\frac{1}{\sqrt{m_1m_2}}\left[-\rho_2+\rho_1+\xi_3\right],\\ \label{3exp } g_2+f_2&=&-\frac{m_2}{m_1}\frac{1}{\sqrt{m_1m_2}}\left[\rho_2-\rho_1+\xi_3\right],\\ \label{4exp } f_1 - 2m_1s&=&\frac{2\xi_3}{\sqrt{m_1m_2}}. \end{aligned}\ ] ] the spectral densities of the form factor combinations in the l.h.s . of eqs .
( [ 1exp][3exp ] ) read @xmath166 comparison with the eqs .
( [ 1exp][3exp ] ) yield the relation _
1()=_2 ( ) , for the form factor @xmath48 we come to the representation @xmath167 , \end{aligned}\ ] ] with @xmath138 given by the expression ( [ z1 ] ) .
in ( [ xi3a ] ) we have neglected the @xmath168 difference between @xmath41 and @xmath134 . on the other hand ,
the equation ( [ 4exp ] ) yields the representation for the form factor @xmath48 in a different form @xmath169 .
\end{aligned}\ ] ] one can check that for the soft wave functions providing convergency of the integrals and nonsingular at @xmath170 the representations ( [ xi3a ] ) and ( [ xi3b ] ) are equivalent . at zero recoil one finds _ 3(1)=dz_2(z_2 + 2m_3)^3/2 ^ 2_0(z_2)3 , where @xmath21 is an average kinetic energy of the light constituent quark inside an ( infinitely ) heavy meson in its rest frame .
it is worth noting that the function @xmath48 is positive for all @xmath41 .
the universal form factor @xmath171 can be found from the expansion of @xmath172 with a nlo accuracy . in this case
the nlo terms in the @xmath10 expansions of the integration measure , the wave function , and the spectral density @xmath152 contribute .
namely , we can write for the form factor @xmath173 the expression h_s()&=&_0(z_1 ) ( 1+_1(z_1 ) ) _ 0(z_2 ) ( 1+_1(z_2 ) ) + & = & ( ) + ( 1m_1 + 1m_2)^(1)()+ =()+ ( 1m_1 + 1m_2)[^(1)()+()(-1)]+ , and hence @xmath174 . we do not present explicit expression for @xmath175 .
however , let us consider @xmath175 at zero recoil .
the analysis of the anomalous contribution at @xmath148 gives h_s(=1)=1 + ( 1m_1 + 1m_2 ) ( 2-)+ using the relations @xmath176 and @xmath177 we obtain @xmath178 just as required by the hqet .
first , let us demonstrate that the dispersion representation of the form factor @xmath59 requires subtraction . to this end
consider the anomalous contribution at @xmath148 .
for the form factor @xmath179 constructed from the spectral density @xmath180 through the unsubtracted double dispersion representation we find the form factor constructed from the spectral density @xmath181 through unsubtracted double dispersion representation . ] f_d(1)=2this value contradicts the luke theorem which requires the @xmath10 corrections to vanish at zero recoil .
let us demonstrate that the form factor @xmath59 constructed from the spectral density with the included subtraction term [ tildeffull ] f = f_d+(m_1 ^ 2-s_1+m_2 ^ 2-s_2)g = f_d+(2p_1p_2 - 2p_1 p_2)g , satisfies the nlo relation ( [ fexp ] ) . here
@xmath182 .
we have noticed above that to lo and nlo the expansion of the form factor @xmath59 and the combination @xmath58 coincide ( [ fmodified ] ) . hence checking the nlo relations for the form factor
@xmath59 is equivalent to checking with the nlo accuracy the relation [ fnew ] f_df_2p_1p_2 g -m_1 g_2 .
the spectral density of the r.h.s . of eq .
( [ fnew ] ) can be written as [ fsym ] 2p_1p_2 g -m_1 g_2 2m_1m_2(+1)s+ [ 2m_1(z_2+m_3)+2m_2(z_1+m_3)]s-8(m_1+m_2)-4m_2(-1 ) .
for checking the expression ( [ fnew ] ) in nlo we need the expansion of the spectral density @xmath152 in lo and nlo which has the structure 12s = m_1_2 + m_2_1+m_3(1-_1-_2)m_3 + + + + , with s_1=12 z_2(z_2 + 2m_3)-z_1m_3-(z_2 + 2m_3)- , and @xmath183 is obtained from @xmath184 by replacing @xmath138 and @xmath139 .
the spectral density @xmath180 reads f_d=4m_1m_2(+1)(m_3 + ) + 2m_2z_2(z_2 + 2m_3)+2m_1z_1(z_1 + 2m_3)-4m_2(-1 ) .
notice that for checking the nlo relation ( [ fnew ] ) between the form factors we do not need explicit expression for the integration measure in the nlo : in the lo the spectral densities are equal and hence the nlo contributions from the integration measure into both sides of the eq ( [ fnew ] ) are equal too .
finally , the eq .
( [ fnew ] ) is satisfied if the following relation is valid @xmath185=0,\end{aligned}\ ] ] with @xmath138 given by ( [ z1 ] ) .
one can check this relation to be true for any function @xmath186 regular at @xmath170 .
hence , the form factor @xmath59 calculated with the subtracted double spectral density ( [ tildeffull ] ) satisfies the hqet relations in lo and nlo at all @xmath41 . strictly speaking
, the nlo analysis does not allow us to uniquely specify the subtraction term : namely , any spectral density of the form f = f_d + ( m_1 ^ 2-s_1+m_2 ^ 2-s_2 ) _ f has a proper nlo behavior in accordance with ( [ fexp ] ) provided the spectral density @xmath187 behaves in the lo as @xmath188 as we demonstrate in the next section , the analysis of the heavy to light lo relations requires the identification @xmath189 let us notice that the form factor obtained within the light
cone approach @xcite can be represented as dispersion representation with the spectral density @xcite f_lc = f_d+ m_2(- ) 2 . at zero
recoil one finds f_lc(1)=2that contradicts the luke theorem .
the unsubtracted spectral densities of the form factors @xmath63 and @xmath191 in the lo read a_1d&=&4 , + a_2d&=&4 .
the quantity @xmath192 is more convenient than @xmath193 for calculations as the lo term of the heavy quark expansion of @xmath192 is zero .
the unsubtracted spectral representation for @xmath194 in combination with eq ( [ aminus ] ) gives _
2()=_3()-14 dz_2 _
0(z_2)_-1 ^ 12 _ 0(z_1 ) . substituting the representation ( [ xi3a ] ) for @xmath48 we find _ 2()=0 .
let us now consider the linear combination @xmath195 .
as a first step , show that the unsubtracted dispersion representation is not compatible with hqet .
to this end calculate the unsubtracted form factors @xmath196 and @xmath197 at zero recoil : a_1d(1)&=&11m_2 , + a_2d(1)&=&11m_1 , and hence a_1d(1)m_2+a_2d(1)m_1=1[z + m_3 ] in contradiction with the hqet result eq .
( [ aplus ] ) .
this fact suggests a necessity of subtraction in the quantity @xmath198 .
let us write the spectral density with subtraction in the form a_1m_2+a_2m_1=a_1dm_2+a_2dm_1 + ( + ) 2 with @xmath199 then corresponding representation for the form factor reads & a_1()m_2+a_2()m_1= 14 dz_2 _
0(z_2)_-1 ^ 12_0(z_1 ) + & + . according to the hqet eq ( [ aplus ] ) this relation should be equal to the term proportional @xmath46 yields @xmath200 . one can check that this value also makes other parts of both expressions equal .
hence we arrive at the subtracted spectral density a_1m_2+a_2m_1=a_1dm_2+a_2dm_1 + ( + ) _ a .
the resulting spectral densities of the form factors @xmath63 and @xmath193 with the built in subtraction terms take the form a_1&=&a_1d + ( + ) 2 , + a_2&=&a_2d + ( + ) 2 . the unsubtracted spectral density @xmath154 has the form g_0d4 .
at zero recoil one finds g_0d(1)=. on the other hand , taking into account our earlier finding @xmath202 , the hqet result ( [ g0exp ] ) reads g_0d(1)=. our experience obtained in considering @xmath203 hints that the subtraction procedure adds a term proportional at zero recoil to @xmath204 , and hence we expect subtraction to work properly also in the case of @xmath201 . as a matter of fact ,
the spectral density [ tildegsub ] g_0=g_0d+ ( + ) 2 satisfies the hq expansion ( [ g0exp ] ) provided the function @xmath205 behaves in the lo as @xmath206 concluding this section let us summarize our main results : we have calculated the universal form factors and demonstrated the dispersion representations with relevant subtractions in the case of heavy to heavy transitions to reproduce the structure of the heavy quark expansion in qcd in the leading and next to leading orders .
however , we have not been able to fix uniquely these subtractions .
as we shall see in the next section , the heavy to light transitions provide further restrictions on the form of the subtraction terms .
in this section we discuss meson decays induced by the heavy to light quark transitions in which case @xmath207 is large , while @xmath208 is kept finite .
as found by isgur and wise @xcite , in the region @xmath209 the form factors of the tensor current can be expressed through the form factors of the vector and axial vector currents in the leading @xmath210 order as follows @xmath211 , \\
\label{hlg1 } g_{1}(q^2)&=&\frac{1}{m_1}[-f(q^2)+2p_1p_2\cdot g(q^2 ) ] .
\end{aligned}\ ] ] we address the two issues : ( i ) perform the leading order @xmath210 expansion of the form factors and show the fulfillment of the relations ( [ hls][hlg1 ] ) and ( ii ) discuss the scaling behavior of the form factors of the transition of different heavy mesons into a fixed final light state . in the case of heavy
light transitions one observes the appearance of the two scales : the light scale @xmath212 , and the heavy scale @xmath208 , and we may expand the form factors in inverse powers of the small parameter @xmath213 .
the kinematical coefficients in the leading @xmath213 order in the region @xmath149 simplify to [ hlcoefficients ] ( s_1,s_2,q^2)&&4m_2 ^ 2,=o(1 ) , + _ 1 & & = o ( ) , + _ 2 & & = o(1 ) , + _ 12 & & _ 1_2-=o ( ) , + _ 11 & & _ 1 ^ 2+=o ( ) . for the double spectral densities @xmath214 these expressions yield the following lo relations [ hlds ] s&=&f_2 , + [ hldg2 ] g_2&=&2m_2 g , + [ hldg0 ] g_0d&= & ( 2g+a_2d ) , + [ hldg1 ] g_1&=&[-f_d+2p_1p_2 g ] , with @xmath215 .
notice that these relations hold also for @xmath216 .
first two equations directly give the lo equality of the corresponding form factors .
the relation ( [ hldg0 ] ) between the unsubtracted spectral densities is more interesting and requires the lo identity of the subtraction terms _ a_g_0 .
the choice @xmath217 is acceptable although there are no firm backgrounds to justify this very choice .
the most informative is the relation ( [ hldg1 ] ) : this relation not only suggests a necessity of subtraction in the form factor @xmath59 but also determines the subtraction term .
when considering the heavy to heavy transitions we have found that the subtracted spectral density of the form ( [ tildeffull ] ) f = f_d+(m_1
^ 2-s_1+m_2 ^ 2-s_2)_f matches the hqet expansion if @xmath218 in lo .
( [ hldg1 ] ) prescribes @xmath219 .
let us point out that in the case of the heavy to light transition there is no parametrical suppression of the anomalous contribution to the form factors compared with the normal one as it was in heavy to heavy transitions and both parts contribute on equal footings .
we shall present all the results for the normal part but the same holds also for the anomalous part . in the region
@xmath220 one finds for the normal contribution @xmath221 where @xmath222 using the relations ( [ hlcoefficients ] ) one finds that the spectral densities scale at large @xmath121 as @xmath223 namely , @xmath224 hence , the form factors have the scaling behaviour of the form @xmath225 the variables @xmath41 and @xmath134 are connected with each other as follows @xmath226 and hence the ratio @xmath227 is universal for the transition of any heavy meson into a fixed light state .
this behaviour reproduces the results of @xcite up to the logarithmic corrections which arise from the anomalous scaling of the quark currents in qcd
. to summarize , matching the lo @xmath210 relations between the form factors in the quark model to the corresponding relations of @xcite allowed us to determine the subtraction term in the form factor @xmath59 and equated to each other subtraction terms in the form factors @xmath190 and @xmath201 .
in this section we apply the derived results to the model dependent estimates of the universal form factors .
first , let us study the dependence of the iw function on the parameters of the quark model .
we adopt the exponential parametrization of the radial wave functions in the form @xmath228 . for obtaining the lo wave function @xmath229 we must take into account the relationship between the parameters @xmath124 and @xmath230 which is found from the following equation = + = m_q+m_3+z .
this yields the relation k^2=z(z+2m_3)+o(1/m_q ) . using the eq .
( [ 5vertex ] ) we come to the following form of the lo wave function _ 0(z)(- ) , where @xmath231 is the lo harmonic oscillator parameter ( m_q)=_0+o(1/m_q ) . for calculations we need to specify @xmath231 . the exact value of @xmath231 is not known but extrapolating the known values of @xmath232 , @xmath233 , and @xmath234 we find @xmath235 both in the wsb @xcite and the isgw2 @xcite models . in the model of ref .
@xcite from the analysis of the axial vector decay constants of pseudoscalar mesons the following approximate dependence has been proposed ( m_q)2.5 , which gives @xmath236 .
table [ table : parameters ] presents the relevant numerical parameters .
the results of calculating the iw function through the eq .
( [ xi ] ) are shown in fig .
[ fig : iw ] .
table [ table : parameters ] demonstrates the values @xmath237 calculated with the eq .
( [ rho2 ] ) and the parameters of the quadratic fit of the form ( ) = 1-_i^2(-1)+_i(-1)^2 , obtained by interpolating the results of calculations in the range @xmath238 .
one can observe the value of @xmath239 to be considerably larger than the parameter @xmath240 obtained by the interpolation procedure .
.[table : parameters ] parameters of the quadratic fit to the iw function and the nlo form factor @xmath48 in various quark model versions . [ cols="^,^,^,^,^,^,^,^,^",options="header " , ] the form factor @xmath16 can be seen to agree with the results of a combined fit to the aleph , argus , cleo , delphi , and opal data @xcite ^2_f=1.070.20 .
one observes sizeable difference between the form factors @xmath16 and @xmath15 and the isgur wise function both in the value at zero recoil and the slope parameter .
this indicates sizeable higher order corrections . at zero recoil
the @xmath10 corrections to @xmath16 vanish and writing @xcite [ deltam2 ] h_f(1)=1+_1/m^2 we find @xmath241 that agrees well with the value @xmath242 @xcite . at the nonzero
recoil the higher order corrections yield a more flat @xmath41dependence of the form factor @xmath16 compared with the iw function such that the slope of the iw function in our model turns out to be considerably larger compared with the slope of @xmath16 : an approximate relation @xmath243 is found .
this is different from the sr result @xmath244 @xcite .
let us point out that @xmath245 turns out to be considerably larger than the parameter @xmath240 of the quadratic fit obtained by the interpolation over the range @xmath238 just as in the case of the isgur wise function .
we have performed a detailed analysis of the dispersion formulation of the quark model for meson decays based on representing the transition form factors as double spectral integrals in the invariant masses of the initial and final @xmath19 pairs through the soft wave functions of the initial and final mesons .
the unsubtracted double spectral densities of the form factors at spacelike momentum transfers are calculated from the feynman graphs , and subtractions in the spectral representations are determined by considering the two following cases : ( i ) by performing the heavy
quark expansion of the form factors in the case of heavy to heavy transitions and matching them to the hqet and ( ii ) by considering the relations between the form factors of the vector , axial vector and tensor currents in the case of the heavy to light transition and matching them to the general relations of ref .
this procedure gives the spectral representations with appropriate subtractions at spacelike momentum transfers .
the form factors at timelike momentum transfers are obtained by performing the analytical continuation in @xmath4 .
the analytical continuation yields the appearance of the anomalous contribution to the form factors at @xmath7 . as a result
, we have come to an explicit model of the form factors describing the meson decays induced by the transition between finite
mass quarks which develops the correct structure of the heavy quark expansion .
let us emphasize that the obtained representations allow us to calculate the form factors directly at the timelike region .
notice that for deriving the heavy
quark expansion we impose only a strong localization of the meson momentum distribution with a width of order of the confinement scale .
no other constraints on the soft wave functions or numerical parameters of the model have emerged .
our main results are : + 1 .
we apply the method of ref .
@xcite to meson amplitudes of the tensor current and find the expansions of the relevant form factors in lo and nlo in hqet .
the @xmath10 corrections in the form factor @xmath11 are found to vanish at zero recoil just as in @xmath15 and @xmath16 .
the @xmath10 expansion of the soft wave function is constructed .
the wave function normalization condition which is a consequence of the vector current conservation in the full theory yields an infinite chain of normalization conditions for the wave function components emerging in various @xmath10 orders . the lo normalization condition which is connected with the current conservation in the effective theory provides the correct normalization of the isgur wise function at zero recoil .
+ 3 . in the lo
all the transition form factors are represented through the isgur wise function in accordance with hqet .
the isgur wise function is calculated through the lo component of the heavy meson wave function .
the subtraction terms in spectral representations of the form factors do not contribute in lo .
the anomalous contribution which emerges when the analytical continuation to the timelike region is performed comes into the game only in the @xmath135 vicinity of the of the zero recoil point but otherwise is suppressed by at least the second inverse power of the heavy
quark mass .
the nlo analysis of the @xmath24 transitions shows that the unsubtracted dispersion representations develop @xmath10 structure in accordance with hqet and allows calculating @xmath48 through the lo wave function .
+ 5 . in the case of the @xmath22 transition the unsubtracted representations for the form factors @xmath164 , @xmath162 , @xmath163 , and
@xmath246 have the nlo behavior compatible with hqet and yield the relation @xmath171 ( or @xmath247 ) and @xmath202 . at the same time , the unsubtracted form factors @xmath59 , @xmath201 , and @xmath248 do not agree with hqet in nlo .
the matching procedure allows restricting the form of the subtraction terms .
+ 6 . analyzing the heavy to light transitions and requiring the fulfillment of the isgur wise relations between the form factors of the tensor and vector and axial vector currents @xcite
further constrains the subtraction terms .
we observe a discrepancy between the predictions of the various versions of the quark model on the nlo universal form factors : namely , the analysys of the wsb model @xcite performed in @xcite resulted in @xmath249 and @xmath250 , but @xmath251 ; the quasipotential quark model @xcite predicts all the nlo form factors to be nonzero ; a consistent relativistic treatment of only @xmath19 intermediate states in feynman graphs performed in this paper gives @xmath171 , @xmath252=0 , and @xmath253 .
the origin of this discrepancy between the quark models is not fully understood and should be considered in more detail .
the numerical results of the dispersion quark model using several parameter sets ( i.e. constituent quark masses and wave functions ) demonstrate a moderate dependence of the isgur wise functions on the light
quark mass and the shape of the wave function .
namely , the light
quark mass @xmath254 and the wave function width @xmath255 yield @xmath256 that is a bit smaller than the sr result @xmath257 @xcite .
the form factor @xmath16 is found to have the behavior in agreement with predictions of other models and experimental results .
we observe sizeable difference between the form factor @xmath16 and the isgur wise function both in the value at zero recoil and the slope parameter .
the slope of the iw function in our model turns out to be considerably larger compared with the slope of @xmath16 : an approximate relation @xmath243 is found .
this is opposite to the sr result @xmath244 @xcite .
we would like to notice that the slope parameter is very sensitive to the interpolation procedure : namely , the value @xmath245 turns out to be considerably larger than the result of the quadratic interpolation over the range @xmath238 . the dispersion quark model for the transition form factors can be further improved by performing the heavy
quark expansion in higher orders and matching to the hqet order by order .
it should be taken into account that the quark model effectively describes the whole @xmath10 series , but the short distance corrections are not contained in the model .
the inclusion of such short distance corrections into consideration is necessary for the application of the model to the analysis of the experimental results .
i am grateful to i.narodetskii , s.simula , b.stech and k.termartirosyan for discussions and to m.neubert for helpful communication .
the work was supported by the rfbr under grants 950204808a and 960218121a .
50 m.wirbel , b.stech , and m.bauer , z.phys .
c29 , 637 ( 1985 ) . t.altomari and l.wolfenstein , phys.rev .
d37,681 ( 1988 ) .
j.g.krner and g.a .
schuler , z.phys .
c38 , 511 ( 1988 ) .
n.isgur , d.scora , b.grinstein , and m.wise , phys.rev .
d39 , 799 ( 1989 ) . d.scora and
n.isgur , phys.rev .
d52 ( 1995 ) 2783 .
w.jaus , phys.rev .
d41 ( 1990 ) 3394 ; _ ibid _ , d53 ( 1996 ) 1349 ; w.jaus and d.wyler , _ ibid _ , d41 ( 1990 ) 3405 .
b.stech , phys.lett .
b354 ( 1995 ) 447 ; hep - ph/9608297 .
r.n.faustov , v.o.galkin , yu.a.mishurov , z.phys .
c66 , 119 ( 1995 ) ; phys.rev .
d53 , 6302 ( 1996 ) .
g.zoller , s.hainzl , c.mnz , and m.beyer , z.phys .
c68 , 103 ( 1995 ) .
yaouanc , l.oliver , o.pene , j.c.raynal , phys.lett .
b365 , 319 ( 1996 ) .
i.grach , i.narodetskii , and s.simula , phys.lett .
b385 , 317 ( 1996 ) .
h.y.cheng , c.y.cheung , c.w.hwang , phys.rev.d55 , 1559 ( 1996 ) .
n.b.demchuk , p.yu.kulikov , i.m.narodetskii , p.j.odonnell , hep - ph/9701388 .
d.melikhov , phys.rev .
d53 ( 1996 ) 2460 .
d.melikhov , phys .
b380 ( 1996 ) 363 ; _ ibid _ , b394 ( 1997 ) 385 .
a.abada et al .
( elc collaboration ) , nucl.phys .
b416 , 675 ( 1994 ) .
c.r.allton et al .
( ape collaboration ) , phys.lett .
b345 , 513 ( 1995 ) .
j.m.flynn et al.(ukqcd collaboration ) , nucl.phys .
b461 , 327 ( 1996 ) .
p.ball , v.m.braun , and h.dosch , phys.rev .
d44 , 3567 ( 1991 ) .
p.ball , phys.rev .
d48 , 3190 ( 1993 ) .
s.narison , phys.lett .
b345 , 166 ( 1995 ) .
p.colangelo , f.de fazio , p.santorelli , and e.scrimieri , phys.rev .
d53 , 3672 ( 1996 ) .
v.m.belyaev , a.khodjamirian , r.rckl , z.phys .
c60 , 349 ( 1993 ) .
p.ball and v.m.braun , phys.rev .
d55 , 5561 ( 1997 ) .
t.m.aliev , a. ozpineci , m. savci , hep - ph/9612480 .
c.g.boyd , b.grinstein , and r.f.lebed , nucl.phys .
b461 , 493 ( 1996 ) .
d.becirevic , phys.rev .
d54 , 6842 ( 1996 ) .
n.isgur and m.b.wise , phys.lett .
b232 ( 1989 ) 113 ; _ ibid _ , b237 ( 1990 ) 527 .
m.luke , phys.lett .
b252 ( 1990 ) 447 .
m.neubert and v.rieckert , nucl.phys .
b382 ( 1992 ) 97 .
m.neubert , phys.rep .
245 , 259 ( 1994 ) and references therein . n.isgur and m.b.wise , phys.rev .
d42 , 2388 ( 1990 ) .
s.simula , phys.lett .
b373 , 193 ( 1996 ) .
m.neubert , phys.rev .
d47 , 4063 ( 1993 ) .
m.p.jimack , in _ heavy quark physics _ , proceedings of the iii german russian workshop , dubna , 1996 , edited by m.a.ivanov and v.e.lyubovitskij , p.80 ( dubna , 1996 ) . | meson transition amplitudes of the vector , axial
vector , and tensor quark currents are analyzed within dispersion formulation of the constituent quark model .
the form factors in the decay region are given by relativistic double spectral representations through the wave functions of the inital and final mesons .
we perform heavy quark expansion of the quark model mesonic matrix elements with a next to leading order accuracy and demonstrate that matching this expansion to the heavy
quark expansion in qcd requires subtractions in the double spectral representations for the form factors and allows fixing the subtraction terms .
the isgur wise function and next to leading order universal form factors are calculated .
= -0.5 cm = 0.3 cm = 24.5 cm 22 |
physicists have long realized the importance of the earth s magnetic field and that this field originates in the earth s fluid core .
the geodynamo is the process by which the rotating , convecting , electrically conducting molten iron in the earth s fluid core maintains the geomagnetic field against ohmic decay .
the convective processes in the core that produce the velocity fields required for dynamo action are a combination of thermal and compositional convection .
the full dynamo problem requires the examination of the full three dimensional partial differential equations governing convective , incompressible magnetohydrodyamics ( mhd ) . in the past decades
computer models have been used to simulate the actual geodynamo . however , current computers and numerical methods require the imposition of diffusivities that are several orders of magnitude larger than those which are realistic .
it is therefore reasonable to attempt to gain some insight into the geodynamo by considering a reduction of the full mhd equations to a system that is more tractable , but one that retains many of the essential features of the problem .
the magnetogeostrophic equation proposed by moffatt and loper ( @xcite-@xcite ) is one such model .
the physical postulates of this model are the following : slow cooling of the earth leads to slow solidification of the liquid metal core onto the solid inner core and releases latent heat of solidification that drives compositional convection in the fluid core .
the arguments for the appropriate ranges of the characteristic length , velocity , and perturbation density are based on these physical postulates .
we first present the full coupled three - dimensional mhd equations for the evolution of the velocity vector @xmath0 , the magnetic field vector @xmath10 and the buoyancy ( or temperature ) field @xmath1 in the boussinesq approximation and written in the frame of reference rotating with angular velocity @xmath11 .
the physical forces governing this system are coriolis forces , lorentz forces , and gravity . following the notation of moffatt and loper @xcite we write the equations in terms of dimensionless variables .
the orders of magnitude of the resulting nondimensional parameters are motivated by the physical postulates of the moffatt and loper model : @xmath12&=-\nabla p+e_2\cdot\nabla b+r_m b\cdot\nabla b+n^2\theta e_3+\varepsilon_\nu \delta u,\label{1.1 - 1}\\ r_m[\partial_t b+u\cdot\nabla b - b\cdot\nabla u]&=e_2\cdot\nabla u+\delta b,\label{1.1 - 2}\\ \partial_t\theta+u\cdot\nabla\theta&=\varepsilon_\kappa\delta\theta+s,\label{1.1 - 3}\\ \nabla\cdot u&=0,\nabla\cdot b=0\label{1.1 - 4}.\end{aligned}\ ] ] here @xmath13 is a given smooth function that represents the forcing of the system .
the mathematical statement of the geodynamo problem asks whether there are initial data for the mhd system for which the evolution of the perturbation of the magnetic field @xmath14 grows for sufficiently long time .
this can be interpreted as a question of the existence of instabilities of - .
the notation in - is the following .
the cartesian unit vectors are denoted by @xmath15 and @xmath16 . for simplicity
, we have assumed that the axis of rotation and the gravity @xmath17 are aligned in the direction of @xmath16 .
we have assumed that the magnetic field @xmath10 consists of the sum of an underlying constant field @xmath18 and a perturbation @xmath14 .
our choice of @xmath18 as the underlying magnetic field is consistent with the models where the magnetic field is believed to be predominantly toroidal ( see @xcite ) .
the dimensionless parameters in - are @xmath19 ( the inverse elsasser number ) , @xmath20 ( the rossby number ) , @xmath21 ( the magnetic reynolds number ) , @xmath22 ( the inverse square of the hartman number ) and @xmath23 ( the inverse peclet number ) . the definitions of these numbers in terms of the relevant dimensionless quantities are given in @xcite where the authors argue that in the region of the fluid core that they are modeling the parameters have the following orders of magnitude : @xmath19 is order unity while @xmath20 is order @xmath24 and @xmath21 is relatively small .
the moffatt and loper model neglects the terms multiplied by @xmath20 and @xmath21 in comparison with the remaining terms .
essentially this means that the evolution equations for the velocity and magnetic field take a simplified `` quasi - static '' form and are linear in the perturbation vector fields @xmath0 and @xmath14 .
the diffusion parameters @xmath22 , which is proportional to the viscosity @xmath25 , and @xmath23 , which is proportional to the thermal diffusivity @xmath26 , are plausibly extremely small .
however their roles are to multiply the laplacian which is the highest spatial derivatives in the system and these terms are retained in the model . in the moffatt - loper model the dominant balance of the leading order terms in equations - gives the following reduced system @xmath27 the linear system - determines the differential operators that relate the vector fields @xmath0 and @xmath14 with the scalar buoyancy ( or temperature ) @xmath1 .
these operators encode the vestiges of the physics in the problem , namely coriolis force , lorentz force , and gravity .
vector manipulations of - give the expression @xmath28
^ 2+n^4(e_3\cdot\nabla)^2\delta\right\}u =- n^2[\varepsilon_\nu\delta^2-(e_2\cdot\nabla)^2]&\nabla\times(e_3\times\nabla\theta)\notag\\ \qquad&+n^4(e_3\cdot\nabla)\delta(e_3\times\nabla\theta).\end{aligned}\ ] ] the sole remaining nonlinearity in the system comes from the coupling of with the equation for the time evolution of @xmath1 .
we call this nonlinear active scalar equation the _ magnetogeostrophic _ ( mg ) equation .
the physically relevant boundary for a model of the earth s fluid core is a spherical annulus .
however for the purposes of studying the mathematical properties of the mg equations we consider the more tractable case of the active scalar equation in the domain @xmath29 ^ 3\times(0,\infty)$ ] ( i.e. with periodic boundary conditions ) . without loss of generality
we may assume that @xmath30 for all @xmath31 , since the mean of @xmath32 is conserved by the flows .
we study the active scalar equation @xmath33,\theta(x,0)=\theta_0(x ) \end{array}\right.\end{aligned}\ ] ] via an examination of the fourier multiplier operator @xmath34 obtained from , which relates @xmath35 and @xmath32 .
more precisely , @xmath36=(\widehat{m_j}\hat\theta)^\vee\end{aligned}\ ] ] for @xmath37 and the explicit expression for the components of @xmath38 as functions of the fourier variable @xmath39 with @xmath40 are @xmath41d(k)^{-1},\label{1.1 - 10}\\ \widehat m_2(k)&=[-n^4k_1k_3|k|^2-n^2k_2k_3(k_2 ^ 2+\varepsilon_\nu|k|^4)]d(k)^{-1}\label{1.1 - 11},\\ \widehat m_3(k)&=[n^2(k_1 ^ 2+k_2 ^ 2)(k_2 ^ 2+\varepsilon_\nu|k|^4)]d(k)^{-1},\label{1.1 - 12}\end{aligned}\ ] ] where @xmath42 on the set @xmath43 , we let @xmath44 since for self - consistency of the model we assume that @xmath32 and @xmath35 both have zero vertical mean .
we point out that @xmath45 and hence the velocity field @xmath35 given by is indeed divergence - free . in a series of recent papers (
@xcite , @xcite , @xcite , @xcite ) properties were proved for the _ inviscid _ mg equation ( i.e. the system - when the viscosity @xmath22 is set to zero ) . in this case
the fourier multiplier symbols @xmath46 , given by - with @xmath47 are not bounded in all regions of fourier space . more specifically in `` curved '' regions where @xmath48 , @xmath49 the symbols are unbounded as @xmath50 with @xmath51 for some positive constant @xmath52 .
thus the relation between the velocity field @xmath35 and the scalar @xmath32 is given by a _
singular _ operator of order 1 .
the implications of this fact for the inviscid mg equation are summarized in the survey paper friedlander - rusin - vicol @xcite . in particular , it is proved that when the thermal diffusivity @xmath23 is set to zero , the inviscid mg equation is _ ill - posed _ in the sense of hadamard in sobolev spaces .
in contrast when @xmath7 , the inviscid but thermally dissipative mg equation is globally well - posed .
in this present article we study the _ viscous _ mg equation ( i.e. @xmath53 ) .
the situation is then dramatically different because the operator @xmath34 whose symbols are given by - now is bounded on @xmath54 and @xmath55 .
thus the constitutive law for the _ viscous _ mg model produces two orders of smoothing .
we remark that the mg equations fall into a hierarchy of active scalar equations arising in fluid dynamics in terms of the nature of the operator that produces the drift velocity from the scalar field : 1 . for the inviscid mg equation the operator is singular of order 1
. 2 . for the surface quasi - geostrophic equation
the operator is the riesz transform which is singular of order zero .
3 . for the 2d euler equation in vorticity form the operator is smoothing of degree one .
4 . for the viscous mg equation the operator is smoothing of degree two . in this sense
the viscous mg equation , even without thermal diffusion , is `` better behaved '' than the 2d euler equation . in the following sections of this article we will prove strong properties of the system - that are a consequence of this fact . the paper is organized as follows . in section 2 , we state the main results of this article which are summarized in various theorems . in section 3 , we introduce some notations and recall some useful embeddings which can be found in the literature . in section 4 , we study the case when @xmath2 . we prove that initial data @xmath56 implies the existence of unique , global weak solutions ; while for @xmath57 we obtain the single exponential growth in time on @xmath58 . in section 5 , we consider the thermally dissipative case ( @xmath7 ) and prove that the solution @xmath59 exists for all time .
in particular , we show that the solution is instantaneously @xmath8-smoothed out and in the class @xmath60 for all positive time with @xmath61 and @xmath62 . in section 6
, we address the convergence of @xmath63 as @xmath23 approaches to zero . under the assumption that @xmath56 , the sequence of solutions @xmath64 to - with @xmath7 converges to @xmath1 weakly in @xmath65 for all @xmath9 , where @xmath1 is the solution to - when @xmath2 .
moreover , if @xmath66 , then for all @xmath67 , the term @xmath68 tends to zero as @xmath69 . in section 7 , we illustrate some dynamical features of the _ forced _ mg equation by demonstrating that the moffalt and loper model can sustain very rapid exponentially growing instabilities .
we examine perturbations of a background temperature gradient and we construct unstable eigenvalues in the spectrum of the linearized equation .
we use a suitable modification of the method of continued fractions first introduced in the context of the navier - stokes equations by meshalkin and sinai @xcite .
the anisotropy of the mg symbols @xmath70 given by ( 1.10)-(1.13 ) permits eigenvalues that are very large when the parameters @xmath22 and @xmath23 are extremely small .
we now give a precise formulation of our results .
the following theorems will be proved in section 4 to section 6 .
assume that @xmath3 has mean zero on @xmath71 and @xmath2 .
there exists unique global weak solution to - such that @xmath72 in particular , @xmath73 weakly in @xmath74 as @xmath75 . here
@xmath76 stands for _ `` bounded continuous functions''_. assume that @xmath4 has mean zero on @xmath71 with @xmath5 and @xmath2 .
there exists a unique solution to - such that @xmath77 for all @xmath78 with @xmath73 weakly in @xmath74 as @xmath75 . in particular , for @xmath79 , we have the following single exponential growth in time on @xmath80 : @xmath81 where @xmath82 are constants which depend only on the spatial dimension .
assume that @xmath7 in - .
let @xmath3 be given and has mean zero on @xmath71 .
then there exists a unique global - in - time mild solution @xmath63 to - such that @xmath83 for all @xmath84 satisfying @xmath61 and @xmath85 .
in particular , @xmath86 in @xmath74 as @xmath75 and @xmath87 as @xmath88 . if we assume @xmath89 for @xmath90 , then by a similar argument given in the proof of theorem 2.3 , one can show that there exists a unique global - in - time mild solution @xmath63 to - with @xmath91 for all @xmath84 satisfying @xmath92 and @xmath85 , and @xmath93 as @xmath88 .
let @xmath56 be given and has mean zero on @xmath71 .
let @xmath63 be the solution to - when @xmath7 with initial data @xmath94 as obtained in theorem 2.3 .
there exists a sequence @xmath95 with @xmath96 such that , @xmath97 where @xmath32 is the solution to - when @xmath2 with initial data @xmath94 as obtained in theorem 2.1
. moreover , if we further assume that @xmath98 , then for any @xmath67 , we have @xmath99
we introduce the following notations .
we say @xmath100 is a weak solution to - for @xmath2 if they solve the system in the weak sense , that means for all @xmath101 , we have @xmath102 @xmath103 and @xmath104 are the usual inhomogeneous sobolev space and homogeneous sobolev space with norm @xmath105 and @xmath106 respectively .
we also define @xmath107 to be the log - lipschitz norm given by @xmath108 we recall the following facts from the literature ( see for example azzam - bedrossian @xcite , bahouri - chemin - danchin @xcite and ziemer @xcite ) : there exists a constant @xmath109 such that @xmath110 and for @xmath111 and @xmath112 , there are constants @xmath113 such that @xmath114 for simplicity , we write @xmath115 , @xmath116 , etc . unless otherwise specified .
we study the non - diffusive case when @xmath2 in - .
we first prove the global - in - time well - posedness of - in the lebesgue space @xmath65 without any smallness conditions . here
@xmath65 is the _
critical _ lebesgue space with respect to the natural scaling of the system - in the sense that if @xmath1 is a solution , then @xmath117 is also a solution with corresponding drift given by @xmath118 $ ] for @xmath119 .
we let @xmath100 be a local smooth solution to - with smooth initial data defined on @xmath120 $ ] for some @xmath67 . under the assumption that @xmath56
, we show the following two lemmas about some _ a priori _ estimates on @xmath100 . for any @xmath111
, there exists @xmath121 such that for all @xmath122 , @xmath123 and @xmath124 the assertion follows immediately from the first equation in together with the divergence - free property on @xmath35 . to prove
, we observe that , by the definition of the operator @xmath34 given in - , for all @xmath125 and all @xmath37 , we have @xmath126 is a fixed constant . in other words
, @xmath34 is a smoothing operator of degree 2 .
hence with the help of the fourier multiplier theorem ( see stein @xcite ) , given @xmath111 , there exists some constant @xmath121 such that @xmath127 therefore follows from .
there exists a constant @xmath109 such that , for any @xmath122 , we have @xmath128 and @xmath129 using and , we can choose some @xmath109 such that @xmath130 on the other hand , with the help of from lemma 4.1 , we can bound @xmath131 in terms of @xmath132 . hence follows .
can proved similarly by using and and we omit the details . in view of lemma 4.2 ,
one can obtain bounds on both @xmath133 and @xmath134 in terms of @xmath135 without any further assumption on @xmath136 .
a uniform - in - time bound on the log - lipschitzian norm of @xmath35 is essential to assure the existence and uniqueness of the flow map @xmath137 ( which will be given in the proof of theorem 2.1 , see below ) and hence the existence and uniqueness of the solution . in the case for 2d euler equation ,
initial conditions on the vorticity @xmath138 of the type @xmath139 ( or @xmath140 ) are required in getting a uniform - in - time bound of @xmath133 . by utilizing the 2-order smoothing effect of @xmath34
, we obtain enough regularity on @xmath35 which gives the desired bound on @xmath133 in terms of @xmath135 .
we are ready to give the proof of theorem 2.1 .
the main idea is to apply lemma 4.2 which gives uniform bounds on @xmath133 and @xmath134 in terms of @xmath135 only . once these bounds are established , the existence and uniqueness follow from the similar argument given by bernicot - keraani @xcite for 2d euler equation .
the proof is divided into two parts .
* existence : * consider the standard mollifier @xmath141 , and we set @xmath142 for @xmath143 and @xmath144 . for the rest of this section , @xmath109 denotes a generic constant which is independent of @xmath145 unless otherwise stated . by standard argument
, we can obtain a sequence of global smooth solution @xmath146 to with @xmath147 .
define @xmath148 to be the flow map given by @xmath149 one can show ( for example in @xcite ) that @xmath150 where the norm @xmath151 is given by @xmath152 with @xmath153 so using and , we have @xmath154 , where @xmath155 are some continuous functions which depends on @xmath156 . and also , for @xmath157 , using , @xmath158 in view of - , the family @xmath159 is bounded and equicontinuous on every compact set in @xmath160 . arzela - ascoli theorem then impies the existence of a limiting trajectory @xmath137 . by similar analysis on @xmath161 ( where @xmath162 is the inverse flow map of @xmath163 ) , we can obtain a limit @xmath164 to @xmath162 and that @xmath165 .
so @xmath137 is a lebesgue measure preserving homeomorphism .
we define @xmath166 and @xmath167 $ ] .
it follows that @xmath168 which implies @xmath169 uniformly , using the fact that @xmath170 so the above allows us to pass the limit in the integral equation on @xmath171 and prove that @xmath100 is a weak solution to .
the continuity of @xmath172 and the preservation of lebesgue measure imply that @xmath173 is continuous with values in @xmath65 , in particular implies @xmath174 .
finally , the assertion that @xmath73 weakly in @xmath74 as @xmath75 follows by a similar argument as given by diperna - lions @xcite and kato - ponce @xcite .
* uniqueness : * we only give a sketch of the proof .
let @xmath67 and suppose that @xmath175 and @xmath176 solve - with @xmath177 .
following the similar argument given in @xcite , there exists a constant @xmath178 such that for all @xmath179 and @xmath180 , we have @xmath181 where @xmath182 s are the usual nonhomogeneous dyadic blocks and @xmath183 .
define @xmath184:\mathcal{c}\int_0^t(\|u^1\|_{\overline{l.l.}}+\|u^2\|_{\overline{l.l.}})(\tau)d\tau\le\frac{1}{2}\right\},\end{aligned}\ ] ] then by - , @xmath185 is well - defined . for each @xmath186 $ ] , we let @xmath187 . using theorem 3.28 in @xcite
, we have for all @xmath188 , @xmath189}\|(\theta^1-\theta^2)\|_{b^{-\delta_t}_{\infty,\infty}}.\end{aligned}\ ] ] summing over @xmath190 , we conclude that @xmath191 on @xmath192 $ ] . by repeating the argument a finite number of times
we obtain the uniqueness on the whole interval @xmath120 $ ] .
we finish the proof of theorem 2.1 .
theorem 2.2 can then be proved by a similar argument as that of theorem 2.1 for @xmath193 with @xmath5 . by controlling the term @xmath194 ,
we further obtain the single exponential growth in time on @xmath58 when @xmath57 .
the details are given as follows .
we only need _ a priori _ bounds on @xmath32 . for the rest of the proof
, @xmath109 denotes a generic constant which is independent of time .
given @xmath5 , we apply fourier transform on the first equation of , multiply it by @xmath195 , rearrange the terms and take the inverse fourier transform to obtain @xmath196,\end{aligned}\ ] ] which implies @xmath197 it suffices to estimate @xmath198 .
we divide it into two cases .
* case 1 : * @xmath199 .
define @xmath200 .
then @xmath201 and we have the following embedding ( see for example nezzaa - palatuccia - valdinocia @xcite ) that @xmath202 and hence @xmath203 on the other hand , by the similar argument as given in the proof of lemma 3.2 , we get @xmath204 therefore , using , , and , we conclude @xmath205 * case 2 : * @xmath206 . using the embeddings @xmath207 and @xmath208 , we follow the similar argument as given in case 1 to get @xmath209 we substitute the above estimates on @xmath210 into and obtain the _ a priori _ required bounds on @xmath32 . finally , we prove the single exponential growth in time on @xmath58 under the assumption that @xmath57 .
we consider the sequence of global smooth solution @xmath146 to as given in the proof of theorem 2.1 with mollified initial data @xmath211 .
we then differentiate the first equation of with respect to @xmath212 , integrate over space - time and use gronwall s inequality to get @xmath213 where @xmath214 are fixed constant which are independent of @xmath215 . on the other hand , we have another fixed constant @xmath216 which depends only on the spatial dimension and is independent of @xmath215 such that @xmath217 hence we conclude from that @xmath218 by taking @xmath219 , follows immediately .
we finish the proof of theorem 2.2 . our result for the single exponential growth in time on @xmath58 is better than the well - known double exponential growth in time of the vorticity gradient for the 2d euler equation ( see , for example , majda - bertozzi @xcite , denisov @xcite for more discussion ) .
we further consider the non - diffusive system - with `` damping '' , namely @xmath220,\theta(x,0)=\theta_0(x ) \end{array}\right.\end{aligned}\ ] ] where @xmath221 is the _ damping constant _ and @xmath34 is the operator as defined in - . using a similar argument to the one given in the proof of theorem 2.2 , we have the following estimate on @xmath222 provided that @xmath57 : @xmath223 $ for all $ t>0$,}\end{aligned}\ ] ] where @xmath224 are the constants defined in the proof of theorem 2.2 which depend only on the spatial dimension .
hence the solution @xmath32 to remains _ bounded _ in @xmath225 provided that @xmath226 .
next we study the thermally diffusive case when @xmath7 in - .
we also mention a recent work obtained by ferreira - lima @xcite which proved a global well - posedness result for a family of dissipative active scalar equations ( via a different method ) with a smallness condition on the weak norm of a fourier - besov - morrey space that allowed to consider some types of large initial data in @xmath227 and sobolev spaces .
we first state the following lemmas for which the proofs can be found in carrillo - ferreira @xcite and lewis @xcite .
let @xmath228 be the convolution operator with kernel given in fourier variables by @xmath229 .
then for @xmath230 , @xmath231 and @xmath232 , there is a constant @xmath109 such that @xmath233 for all @xmath234 .
let @xmath235 be a banach space with norm @xmath236 and @xmath237 be a continuous bilinear map , that means there exists @xmath238 such that @xmath239 for all @xmath240 .
given @xmath241 and @xmath242 such that @xmath243 , there exists a solution @xmath244 for the equation @xmath245 such that @xmath246 .
the solution @xmath212 is unique in the closed ball @xmath247 .
moreover , the solution depends continuously on @xmath248 in the following sense : if @xmath249 , @xmath250 and @xmath251 , then @xmath252 we will prove the following theorem about the local - in - time existence of solutions to - with initial data @xmath56 when @xmath7 .
it is crucial in proving theorem 2.3 .
the argument is similar to the one given by ferreira - lima @xcite .
let @xmath56 .
for any @xmath7 , there exists @xmath67 such that - has a unique mild solution @xmath63 in the class @xmath253 for all @xmath84 satisfying @xmath61 and @xmath85 .
we convert the system - into the integral equation : @xmath254 where @xmath228 is the convolution operator as defined in lemma 5.1 and @xmath255 is the bilinear form @xmath256d\tau.\end{aligned}\ ] ] we will prove that there exists a constant @xmath238 such that for any @xmath257 $ ] , @xmath258 in view of the operator @xmath34 with @xmath167 $ ] as given by - , for any @xmath259 and @xmath260 , there exists @xmath261 such that given @xmath262 , we have the following estimates on @xmath263\|_{w^{{s+2},p}}$ ] : @xmath264\|_{w^{{s+2},p}}\le k_1\|f\|_{w^{s , p}},\end{aligned}\ ] ] using , and , there exists @xmath238 such that , given @xmath265 , we have @xmath266\|_{l^p}d\tau\\ & \le k\int_0^t(t-\tau)^{-\frac{1}{2}}\|\nabla\cdot(u(\phi(\cdot,\tau))\psi(\cdot,\tau))\|_{\dot{w}^{-1,p}}d\tau\\ & \le k\int_0^t(t-\tau)^{-\frac{1}{2}}\|u(\phi(\cdot,\tau))\psi(\cdot,\tau)\|_{l^p}d\tau\\ & \le k\int_0^t(t-\tau)^{-\frac{1}{2}}\|u(\phi(\cdot,\tau))\|_{l^\infty}\|\psi(\cdot,\tau)\|_{l^p}d\tau\\ & \le k\int_0^t(t-\tau)^{-\frac{1}{2}}\|u(\phi(\cdot,\tau))\|_{w^{2,p}}\|\psi(\cdot,\tau)\|_{l^p}d\tau\\ & \le k\int_0^t(t-\tau)^{-\frac{1}{2}}\|\phi(\cdot,\tau)\|_{l^p}\|\psi(\cdot,\tau)\|_{l^p}d\tau\\ & \le k\left[\int_0^t(t-\tau)^{-\frac{1}{2}}\tau^{-1+\frac{3}{p}}d\tau\right]\\ & \qquad\times\left(\sup_{0<\tau < t}\tau^{\frac{1}{2}-\frac{3}{2p}}\|\phi(\cdot,\tau)\|_{l^p}\right)\left(\sup_{0<\tau < t}\tau^{\frac{1}{2}-\frac{3}{2p}}\|\psi(\cdot,\tau)\|_{l^p}\right)\\ & \le k t^{-\frac{1}{2}+\frac{3}{2p}}t^\frac{3}{2p}\left(\int_0 ^
1(1-z)^{-\frac{1}{2}}z^{-1+\frac{3}{p}}dz\right)\\ & \qquad\times\left(\sup_{0<\tau < t}\tau^{\frac{1}{2}-\frac{3}{2p}}\|\phi(\cdot,\tau)\|_{l^p}\right)\left(\sup_{0<\tau < t}\tau^{\frac{1}{2}-\frac{3}{2p}}\|\psi(\cdot,\tau)\|_{l^p}\right).\\\end{aligned}\ ] ] by the assumption that @xmath267 and @xmath85 , it implies @xmath268 and so follows . for any @xmath257 $ ] , we now define @xmath269 and @xmath270 .
it is clear that @xmath271 is a banach space , and by , we have @xmath272 given @xmath3 , using there exists @xmath273 such that , @xmath274 so it implies for every @xmath275 , @xmath276 since @xmath277 , we conclude that for @xmath3 , @xmath276 hence for any @xmath278 , there exists @xmath279 such that @xmath280 we can apply lemma 4.2 to obtain a unique solution @xmath63 to such that @xmath281 next , we consider the picard sequence defined by : @xmath282 we notice that the solution @xmath63 as given by can be obtained as the limit in @xmath271 of @xmath283 .
moreover , @xmath284 we now prove that for any @xmath143 , the sequences @xmath283 and @xmath285 are uniformly bounded respectively in @xmath286 and @xmath287 .
we first show that @xmath285 is uniformly bounded in @xmath287 . using
, there exists a constant @xmath288 such that @xmath289 for any @xmath290 , using the definition of @xmath291 , we have @xmath292 using , , and , there exists a constant @xmath293 such that @xmath294 by the assumptions on @xmath295 and @xmath296 , we have @xmath297 , @xmath298 and @xmath299 .
hence @xmath300 replacing @xmath301 if necessary , choosing @xmath302 and reducing @xmath303 , we conclude from that @xmath304 next we prove that @xmath283 is uniformly bounded in @xmath286 . again using ,
there exists a constant @xmath305 such that @xmath306 and by the definition of @xmath291 , for @xmath290 , @xmath307 using , , and , there exists a constant @xmath308 such that the term @xmath309 can be estimated as follows .
@xmath310 since @xmath267 and @xmath311 , replacing @xmath312 if necessary , we obtain @xmath313 by choosing @xmath314 small enough and reducing @xmath303 if necessary , an induction argument on shows that @xmath283 is uniformly bounded in @xmath286 .
since the sequence @xmath283 is uniformly bounded in @xmath286 , we can see that there exists a subsequence of @xmath283 which converges towards some @xmath315 weak- * in @xmath286 and consequently in @xmath316 .
on the other hand , we know that @xmath317 in @xmath271 , which implies convergence in @xmath316 as well
. therefore , @xmath318 the time - continuity of @xmath63 follows by using the fact that @xmath63 belongs to @xmath271 and it solves ( see kato @xcite-@xcite ) .
hence we obtain . by similar methods
, we can prove that @xmath319 as well , and we finish the proof of theorem 5.3 .
we point out that , the indexes @xmath295 , @xmath296 as appeared in theorem 5.3 are independent of each other .
in other words , there is no further restriction on @xmath84 except @xmath61 and @xmath85 .
we give the proof of theorem 2.3 which involves showing the global - in - time existence of @xmath63 and the time decay of @xmath320 for all @xmath321 $ ] and @xmath85 when @xmath9 . in view of theorem 5.3 , we first extend the local solution @xmath63 satisfying - .
standard parabolic theory ( see for example @xcite ) shows that @xmath322 for all @xmath201 , @xmath323 and multi - index @xmath324 , where @xmath67 is the existence time obtained in theorem 5.3 .
hence , we have @xmath325 and @xmath326 for all @xmath122 .
now we prove .
the following argument is reminiscent of the one given in @xcite .
we notice that for @xmath327 , the existence time @xmath303 ( as in theorem 5.3 ) can be chosen as @xmath328 where @xmath241 , and @xmath329 are given as in the proof of theorem 5.3 .
let @xmath56 . by theorem 5.3
, there exists constants @xmath330 and a unique solution @xmath63 to - defined on @xmath331 such that @xmath332 next we define @xmath333 we claim that @xmath334 .
we argue by contradiction .
suppose that @xmath335 , and we let @xmath336 which will be chosen later . by theorem 5.3
, we have @xmath337 and @xmath338 . by choosing @xmath339 as initial data , given @xmath340
, there exists @xmath341 and a unique solution @xmath342 to - defined on @xmath343 such that @xmath344 @xmath345^{\frac{1}{2}-\frac{3}{2p}}\|\tilde\theta^{\varepsilon_\kappa}(t)\|_{l^p}&\le 2d_1,\\ \sup_{t-\delta < t < t-\delta+t_1}\|\tilde\theta^{\varepsilon_\kappa}(t)\|_{l^p}&\le\|\theta^{\varepsilon_\kappa}(t-\delta)\|_{l^3}.\end{aligned}\ ] ] by the uniqueness of solution , we have @xmath346 for all @xmath347 .
we now choose @xmath348 then @xmath349 and there exists a solution @xmath350 to - such that @xmath351 for all @xmath352 , which contradicts the maximality of @xmath303 .
hence we must have @xmath334 and we finish the proof of . finally we consider .
we first claim that there exists a constant @xmath353 independent of @xmath354 such that @xmath355 for all @xmath9 .
the proof for is the same as the one given in @xcite and we include here for completeness .
because @xmath63 exists for all time , it follows from and that @xmath356 in particular , @xmath357 satisfies in the classically sense for all @xmath9 . in view of @xmath358
, we integrate by parts to obtain , for any @xmath359 , @xmath360 for all @xmath9 . using the gagliardo - nirenberg inequality for @xmath361 ,
there exists a constant @xmath362 such that @xmath363 and hence we obtain the following inequality @xmath364 by considering the sequence @xmath365 for @xmath366 , we can solve inductively to get @xmath367 where @xmath368 are defined by @xmath369 and we have used to get @xmath370 for all @xmath9 . therefore , @xmath371 taking @xmath372 , we obtain a constant @xmath373 such that , for @xmath9 , @xmath374 which implies by choosing @xmath375 . by interpolation , for @xmath201 , we can further obtain a constant @xmath376 such that @xmath377 to estimate @xmath378 , using , , , and the integral equation , there exists constants @xmath379 such that for all @xmath9 , @xmath380 hence follows and we have @xmath87 as @xmath88 .
we finish the proof of theorem 2.3 .
let @xmath63 be the solution to - when @xmath7 with initial data @xmath56 as obtained in theorem 2.3 .
we now consider the convergence of @xmath63 as @xmath69 .
we first prove the following lemma which gives some uniform bounds on @xmath63 independent of time and @xmath23 .
given @xmath56 , there exist constants @xmath381 independent of @xmath354 and @xmath23 such that , for any @xmath78 and @xmath7 , @xmath382 @xmath383 furthermore , for any @xmath359 , there exists @xmath384 independent of @xmath354 and @xmath23 such that @xmath385 to show , we choose @xmath386 in to obtain @xmath387 and follows .
- can be proved by similar arguments given in the proof of lemma 4.1 - 4.2 , and we omit the details here . the convergence of @xmath388 can be proved by a standard argument involving weak limits ( for example , see hoff @xcite ) , while the convergence of @xmath389 can be shown by controlling the term @xmath390 when @xmath98 . without loss of generality , we assume that @xmath391 and @xmath392 .
let @xmath393 be a countable dense subset of @xmath394 . by ,
we have that @xmath388 is bounded in @xmath65 , uniformly in @xmath23 . using a diagonalization process , we can obtain a sequence @xmath395 with @xmath96 for which @xmath396 converges weakly in @xmath65 , say to @xmath397 for each @xmath398 .
given @xmath399 , using - , we have for all @xmath7 , @xmath400 for some constant @xmath401 independent of time and @xmath23 .
so it follows that @xmath402 converges weakly to an element @xmath403 for every @xmath404 . on the other hand , using the uniform bound , we have @xmath405 it shows that every subsequence of @xmath406 has a further subsequence ( still call it @xmath406 for simplicity ) which converges weakly in @xmath65 , necessarily to @xmath407 , for every @xmath404 .
hence @xmath408 .
the assertion that @xmath409 solves - with @xmath2 in the weak sense can be verified in a very similar way as given by friedlander and vicol in @xcite for the _ inviscid _ mg equation ( i.e. @xmath47 ) ( see appendix a in @xcite ) . by the uniqueness of the solution @xmath32 to - with @xmath2 as obtained in theorem 2.1 , we conclude @xmath410 and therefore @xmath411 weakly in @xmath65 for every @xmath404 .
finally , it remains to prove .
we assume further that @xmath98 . in view of
, we can differentiate @xmath412 with respect to @xmath212 and integrate to obtain @xmath413 for all @xmath122 . using gronwall s inequality , there exists a constant @xmath414 independent of @xmath415 such that @xmath416 it suffices to estimate @xmath417 . using , and
, there exists a constant @xmath418 independent of @xmath419 such that for all @xmath420 , @xmath421 therefore we have @xmath422 integrating over @xmath354 , we conclude from that @xmath423\end{aligned}\ ] ] and follows immediately by taking @xmath69 .
we finish the proof of theorem 2.5 .
in this section we demonstrate the existence of unstable eigenvalues for the forced mg equation @xmath424 where the divergence free velocity @xmath35 is obtained from @xmath32 via @xmath425 with the fourier multiplier symbols given by ( 1.10)-(1.13 ) .
we consider perturbations of a particular steady state , namely @xmath426 where the amplitude @xmath427 is an arbitrary constant .
we consider the linear evolution of the perturbation temperature @xmath428 and velocity @xmath429 and make the assumption that , for some fixed integers @xmath430 and @xmath431 has the form @xmath432 which from ( 1.10)-(1.13 ) implies that the perturbation velocity @xmath433}{n^4n^2 ( k^2_1 + k^2_2 + n^2 ) + [ k^2_2 + \varepsilon_\nu ( k^2_1 + k^2 + n^2)^2]^2 } \sin ( nx_3).\end{aligned}\ ] ] substituting ( 7.2)-(7.4 ) into the linearized version of ( 7.1 ) gives @xmath434}{n^4n^2 ( k^2_1 + k^2_2 + n^2 ) + [ k^2_2 + \varepsilon_\nu ( k^2_1 + k^2_2 + n^2)^2]^2 } \sin ( nx_3)= 0 .
\label{7.5}\end{aligned}\ ] ] we now follow the construction of an unstable eigenvalue @xmath435 and associated @xmath8 smooth eigenfunction satisfying ( 7.5 ) which is given in @xcite for the inviscid mg equation . a suitable modification of the method of continued fractions utilized for the navier - stokes equations by meshalkin and sinai @xcite produces a characteristic equation which gives a positive _ lower _ bound on a real root @xmath436 .
this procedure applied to the viscous mg equation demonstrates the existence of an eigenvalue @xmath436 such that @xmath437}{4n^4m^2(k^2_1+k^2_2 + 4m^2 ) + [ k^2_2 + \varepsilon_\nu ( k^2_1+k^2 + 4m^2)^2]^2 } \notag\\ & - \varepsilon_\kappa ( k^2_1+k^2_2 + 4m^2).\end{aligned}\ ] ] we note that the characteristic equation also produces an upper bound @xmath438}{n^4m^2(k^2_1+k^2_2+m^2 ) + [ k^2_2 + \varepsilon_\nu ( k^2_1+k^2_2+m^2)^2]^2 } \notag\\ & - \varepsilon_\kappa ( k^2_1+k^2_2+m^2).\end{aligned}\ ] ] from ( 7.6 ) it follows that we may choose the amplitude @xmath427 of the steady temperature distribution sufficiently large in terms of @xmath439 and @xmath19 so that there exists at least one pair of wave numbers @xmath440 that produce a positive eigenvalue @xmath436 .
we now discuss the implications of the lower bound ( 7.6 ) in terms of the limiting behavior of the small parameters @xmath22 and @xmath23 and the existence eigenvalues whose magnitudes grow as @xmath22 or @xmath23 tend to to zero . * case ( i ) : * both @xmath441 and @xmath442 . in this case
the estimate ( 7.6 ) gives @xmath443 choosing @xmath444 such that @xmath445 we obtain that for any integer @xmath446 there exists an eigenvalue @xmath436 bounded from below as @xmath447 and the upper bound ( 7.7 ) gives @xmath448 where @xmath52 is a constant depending only on @xmath449 and @xmath19 .
thus in the absence of viscous or thermal diffusion we can construct eigenvalues with arbitrarily large real part .
this property was used in @xcite to prove that the cauchy problem for the inviscid , nondiffusive mg equation is ill - posed in the sense of hadamard .
* case ( ii ) : * @xmath441 and @xmath450 . in this case
( 7.6 ) gives @xmath451 hence there are at most finitely many values @xmath440 for which the lower bound is positive .
the maximum value of this lower bound as a function of @xmath430 and @xmath452 occurs when @xmath453 ( 7.7 ) and ( 7.9 ) give @xmath454 where @xmath455 is a positive constant depending only on @xmath456 and @xmath19 . thus the lower bound ( 7.11 ) is positive for sufficiently large amplitudes @xmath427 and grows like @xmath457 for @xmath450 .
hence the limit of vanishing thermal diffusivity produces very large , but finite , unstable eigenvalues .
* case ( iii ) : * @xmath442 and @xmath458 .
the lower bound ( 7.6 ) gives @xmath459}{4n^4m^2 ( k^2_1 + k^2_2 + 4m^2 ) + [ k^2_2 + \varepsilon_\nu ( k^2_1 + k^2_2 + 4m^2)^2]^2}.\end{aligned}\ ] ] hence in this case there are unstable eigenvalues for all values of @xmath427 and for all finite values of @xmath430 and @xmath452 .
the maximum values of this lower bound as a function of @xmath430 and @xmath452 occurs when @xmath460 ( 7.7 ) and ( 7.12 ) gives @xmath461 where @xmath462 is a positive constant depending only on @xmath456 and @xmath19 .
thus the limit of vanishing viscosity also produces very large , but finite , unstable eigenvalues . * case ( iv ) : * @xmath450 and @xmath458 . as we observed the bound ( 7.6 )
implies that provided @xmath427 is sufficiently large there always exists unstable eigenvalues for some small values of @xmath430 and @xmath452 . to examine the behavior of the lower bound in terms of the limiting behavior of the two small parameters we write @xmath463 .
the maximum value of the lower bound occurs when @xmath464 where ( 7.6 ) gives @xmath465\end{aligned}\ ] ] and ( 7.7 ) gives @xmath466\end{aligned}\ ] ] where @xmath467 and @xmath468 are positive constants depending only on @xmath456 and @xmath19 .
thus when @xmath469 we again conclude that for sufficiently large @xmath427 there exist eigenvalues that grow like @xmath470 as @xmath23 approaches zero .
in contrast , when @xmath471 the bound ( 7.14 ) is negative for @xmath450 and this construction does not produce large unstable eigenvalues , in other words when @xmath471 the viscosity prevents rapid instabilities .
in particular , this observation implies that there will be no large " unstable eigenvalues in the context of section 6 where @xmath472 and @xmath473 .
we note that in each case ( ii ) , ( iii ) and ( iv ) @xmath474 the maximum lower bound on @xmath436 is attained on the parabolic curves @xmath475 and this bound grows with @xmath476 or @xmath457 .
the limit @xmath477 and the limit @xmath478 are consistent with the inviscid , non - diffusive case ( i ) which permits _ arbitrarily _ large eigenvalues associated with the curves @xmath475 as @xmath479 .
we also note that when @xmath480 instability for the linearized equation implies nonlinear lyapunov instability for the active scalar equation ( 7.1 ) .
this result was proved for the inviscid mg equation in @xcite using a modification of the bootstrap techinques applied to the navier - stokes equations in @xcite .
the proof in @xcite directly carries over to the viscous mg equation where @xmath472 . to summarize ,
the classical geodynamo problem of magnetic field generation via the flow of an electrically conducting fluid is closely connected with the existence of unstable eigenvalues ( see , for example , @xcite ) .
we have shown that the moffatt and loper geodynamo model permits such instabilities . the active scalar equation ( 7.1 )
linearized about a suitable background temperature field supports unstable eigenvalues .
if either one of @xmath22 or @xmath23 is zero and the other is very small and positive there exist perturbations that grow very rapidly on certain parabolic curves in wave number space .
such fast instabilities also occur when both @xmath22 and @xmath23 are positive with @xmath481 .
the existence of very strong instabilities is a consequence of certain features of the structure of the fourier multiplier symbol @xmath482 that are produced by the specific constitutive law in the model relating the temperature with the velocity and the magnetic field ( see equations ( 1.5)-(1.7 ) ) .
the crucial features are the anisotropy of the symbol and the fact that it is even with respect to the wave numbers @xmath483 and @xmath484 . in the case where both @xmath23 and @xmath22 are zero a consequence of the structure of the symbol
is the existence of _ arbitrarily _ large eigenvalues .
this implies that the mg equation without the control of any diffusion is hadamard ill - posed , in contrast with the diffusive mg equation which is globally well posed .
s. friedlander , w. rusin and v. vicol , _ the magneto - geostrophic equations : a survey _ , proceedings of the st .
petersburg mathematical society , volume xv : advances in mathematical analysis of partial differential equations .
( 2014 ) d. apushkinskaya , and a.i .
nazarov , eds .
. 5378 .
moffatt , _ magnetostrophic turbulence and the geodynamo _ , in y. kaneda , editor , iutam symposium on computational physics and new perspectives in turbulence , nagoya , japan ,
september , 11?14 , 2006 , volume 4 of iutam bookser . , pages 339?346 .
springer , dordrecht , 2008 . | we study the three dimensional active scalar equation called the magneto - geostropic equation which was proposed by moffatt and loper as a model for the geodynamo processes in the earth s fluid core . when the viscosity of the fluid is positive , the constitutive law that relates the drift velocity @xmath0 and the scalar temperature @xmath1 produces two orders of smoothing .
we study the implications of this property .
for example , we prove that in the case of the non - diffusive ( @xmath2 ) active scalar equation , initial data @xmath3 implies the existence of unique , global weak solutions .
if @xmath4 with @xmath5 , then the solution @xmath6 for all time . in the case of positive diffusivity ( @xmath7 ) , even for singular initial data @xmath3 , the global solution is instantaneously @xmath8-smoothed and satisfies the drift - diffusion equation classically for all @xmath9 .
we demonstrate , via a particular example , that the viscous magneto - geostrophic equation permits exponentially growing dynamo type " instabilities . |
frustrated quantum magnets have been studied intensively to unveil possible novel phenomena which can be caused by the interplay of frustration and quantum effects .
spin-1/2 two dimensional ( 2d ) antiferromagnet on the kagome lattice is a central example of such frustrated quantum magnets . on the theoretical front ,
various studies have suggested several candidate ground states for the magnetically disordered phase of the spin-1/2 antiferromagnetic heisenberg model : a variety of spin liquids @xcite and valence bond solids @xcite .
experimentally , a lot of effort has been made to realize the material system of such geometrically frustrated antiferromagnet .
the herbertsmithite @xcite zncu@xmath1(oh)@xmath5cl@xmath0 has the most ideal structure with uniform exchange couplings and shows no magnetic ordering down to much lower temperatures than the curie - weiss temperature . on the other hand , there may be several additional factors that would affect the nature of the ground state .
for example , the dzyaloshinskii - moriya interaction and impurity spins may be present in the material .
@xcite because of these complications , it has been difficult to pin down the true ground state of the system .
these considerations also suggest that it is important to understand the effect of various perturbations to the ideal kagome lattice structure and magnetic anisotropies to identify the true ground state of the system .
to @xmath6 in @xmath7 is clockwise as indicated by the arrows in the lower right triangles . for the vbs state with the 12-site unit cell ,
the valence bonds are formed in the interaction links with the strongest couplings , @xmath8 and @xmath9 ( thick black ) . [
fig : deformed_kagome_lattice ] ] rb@xmath0cu@xmath1snf@xmath2 is another compound that was synthesized as an attempt to materialize the kagome lattice antiferromagnet . unlike the herbertsmithite , however , it has a deformed kagome lattice structure ( see fig .
[ fig : deformed_kagome_lattice ] ) .
magnetic susceptibility and magnetization measurements show that it is a magnetically - disordered 2d spin - gapped system.@xcite as a model to describe the system , a spin-1/2 antiferrmagnetic heisenberg model was considered with four different exchange couplings due to its deformed structure , labeled by @xmath8 , @xmath10 , @xmath11 , and @xmath12 .
the exchange coupling constants are estimated by fitting the thermodynamic data : @xmath8 , @xmath10 , and @xmath11 are quite similar in magnitude ( @xmath13200 k ) , while @xmath12 is about half of the others ( @xmath14 ) .
the size of the spin gap is estimated to be @xmath15=21 k. the recent neutron scattering experiment in conjunction with a dimer series expansion study has shown that the ground state of rb@xmath0cu@xmath1snf@xmath2 is indeed the valence bond solid ( vbs ) with a 12-site unit cell.@xcite the valence bonds are formed in the strongest interaction links with the coupling constant @xmath8 ( black line segments in fig .
[ fig : deformed_kagome_lattice ] ) .
further , it turned out that the dzyaloshinskii - moriya ( dm ) interactions are not negligible in rb@xmath0cu@xmath1snf@xmath2 as seen in the triplon dispersions . if there are only ( isotropic ) heisenberg interactions , the system has so(3 ) global spin rotation symmetry so that its triplon dispersions should be triply - degenerate . when there are ( anisotropic ) dm interactions , the degeneracy of triplon dispersions is lifted .
it was observed in neutron scattering experiment that the triplon dispersions of rb@xmath0cu@xmath1snf@xmath2 are only doubly - degenerate .
this fact indicates that there exist significant dm interactions in rb@xmath0cu@xmath1snf@xmath2 .
the valence bond solid state with a 12-site unit cell was studied by some of us for the spin-1/2 antiferromagnetic heisenberg model on the deformed kagome lattice.@xcite in that study , the weakest interaction was controlled with @xmath16 and @xmath17 .
the vbs state with a 12-site unit cell turned out to have lower energy than the vbs state with a 36-site unit cell , which was suggested for uniform kagome lattice antiferromagnet ( @xmath18 case ) , when @xmath19 is less than 0.97 within the bond operator mean - field theory .
concerning rb@xmath0cu@xmath1snf@xmath2 , the triplon dispersions were predicted as well .
the predicted triplon energy bands are triply - degenerate due to the so(3 ) symmetry of the heisenberg hamiltonian and spin gap is found at the k point .
however , the spin gap ( the minimum triplon gap ) was observed at @xmath4 point in the recent neutron scattering experiment.@xcite this presents another evidence for the importance of the dm interaction in rb@xmath0cu@xmath1snf@xmath2 . in this paper
, we theoretically investigate the effect of the dzyaloshinskii - moriya interaction in the deformed kagome lattice antiferromanget via various methods .
the model hamiltonian for the system consists of the heigenberg ( @xmath20 ) and dzyaloshinskii - moriya ( @xmath21 ) interactions between the nearest neighbors ; @xmath22 for the heisenberg interactions , we consider four different exchange couplings , @xmath14 , depending on the interaction link as shown in fig .
[ fig : deformed_kagome_lattice ] . as for the dm interactions , we use the result from series expansion study.@xcite according to the series expansion
, the dm vectors @xmath23 have small in - plane components compared to the out - of - plane component , and the overall behavior of excitation spectrum is determined by the out - of - plane component whereas the in - plane ones make minor corrections . due to this reason ,
the dm vectors are assumed to have only the out - of - plane component in our study ( @xmath24 ) . like the heisenberg interactions
, we consider four different dm interactions , the magnitude of which is labeled by @xmath25 .
@xmath26 is assumed to be proportional to @xmath27 , _
i.e. _ @xmath28 .
the strength of the dm interactions is controlled by the parameter @xmath29 in this paper .
the direction of @xmath23 is determined by the moriya s rule@xcite to be along the positive z - direction if the orientation from the site @xmath30 to @xmath6 in @xmath7 is clockwise in each triangle of the kagome lattice . in the lower right part of fig .
[ fig : deformed_kagome_lattice ] , the orientations from @xmath30 to @xmath6 for the positive dm vector are denoted with arrows in two possible types of triangles in the kagome lattice .
it must be noted that the model hamiltonian has so(2 ) global spin rotation symmetry in the xy - plane because every dm vector is along the z - direction .
therefore , the generator of spin rotation in the xy - plane , @xmath31 , is a good quantum number of the hamiltonian . to analyze this hamiltonian , we take three different approaches : bond operator mean - field theory , exact diagonalization , and strong coupling expansion .
first , the bond operator mean - field theory is used to investigate the effect of the dm interaction on the triplon dispersions . within the bond operator theory
, we consider the vbs state with the 12-site unit cell in fig .
[ fig : dimer_configuration ] as the ground state and obtain the triplon dispersions .
these disperions are shown to be consistent with the results of the neutron scattering experiment .
the dispersions are twofold degenerate , which results from the so(2 ) symmetry of the hamiltonian . by increasing the strength of the dm interactions ( @xmath29 ) from zero , we observe that the position of the spin gap ( the minimum triplon gap ) in the first brillouin zone shifts from the k to @xmath4 point at @xmath32 . beyond @xmath32 ,
the size of the spin gap decreases gradually , but the position of spin gap does not change from the k point .
it is interesting to note that the spin gap is linearly decreasing as a function of @xmath29 after @xmath32 . by performing linear fitting
, we estimate the critical value @xmath33 , at which the spin gap becomes zero .
the value of @xmath29 in rb@xmath0cu@xmath1snf@xmath2 was estimated to be 0.18 in the series expansion study . at @xmath34 ,
the spin gap in the bond operator theory is 0.464 @xmath8 .
this value is about four times larger than the neutron scattering result , 0.126 @xmath8 .
but , it is not surprising because the bond operator mean - filed theory tends to overestimate the spin gap .
secondly , we conduct the exact diagonalization study for finite size clusters in order to calculate the spin gap more accurately . in the 24-site cluster
, the spin gap is found to be 0.229 @xmath8 , which is not far from the experimental value . as a matter of fact
, this value is the best among the results of various approaches in our study . via the exact diagonalization
, we also study magnetization process because magnetically - disordered spin - gapped systems often show interesting behaviors in magnetization process .
exact diagonalization is restricted to small size clusters due to the large unit cell of the deformed kagome lattice . to overcome this limitation ,
we also employ strong coupling expansion approach .
strong coupling expansion is a degenerate perturbation theory which provides an effective hamiltonian near the magnetic fields strong enough to close the spin gap .
this approach allows the larger system sizes than the clusters used in the exact diagonalization study because the full hilbert space is projected into the subspace of the states relevant to the magnetization process . from the exact diagonalization and strong coupling expansion studies ,
we obtain qualitatively the same magnetization curves .
the magnetization curves increase continuously from zero to saturated magnetization as the magnetic field is increased , which is consistent with the magnetization measurement .
the outline of the rest of the paper is as follows . in sec .
[ sec : bomft ] , we develop the bond operator mean - field theory for the deformed kagome lattice antiferromanget . exact diagonalization study is presented in sec .
[ sec : ed ] . in sec .
[ sec : sce ] , we perform the strong coupling expansion .
lastly , results are compared with the recent experimental results and summarized in sec .
[ sec : discussion ] .
in this section , we construct the bond operator mean - field theory for the vbs state with the 12-site unit cell . the dimer configuration for the vbs state
is depicted in fig .
[ fig : dimer_configuration ] , where dimers are denoted with thick blue line segments and six dimers in the 12-site unit cell are labeled by @xmath35 . to introduce the basis for the bond operator representation , we consider the following hamiltonian of a pair of spins , @xmath37 and @xmath36 , which corresponds to an isolated spin dimer with the coupling constants @xmath8 and @xmath9 in fig .
[ fig : dimer_configuration ] ; @xmath38 where @xmath39 and @xmath40 @xmath41 .
the energy eigenstates are given by it must be noted that @xmath45 is not the pure spin singlet state and @xmath46 is not one of the pure spin triplet states unless @xmath47 ( or @xmath48 ) .
@xmath49 , @xmath50 , @xmath51 , and @xmath52 are operators to create a bond particle in the associated state and satisfy the bosonic commutation rules .
the subscript of @xmath53-operator , @xmath54 , implies the quantum number of @xmath55 , which originates from the so(2 ) spin rotation symmetry of the hamiltonian in eq .
( [ h_dimer ] ) .
in other words , @xmath54 indicates the magnetic quantum number of the @xmath53-bosons . in the above representation , we set @xmath57 .
there are two reasons why we choose the states ( [ eq : bot_basis ] ) as the basis for the bond operator representation against the conventional one.@xcite first , if we use the conventional basis , the dm interactions with the coupling constant @xmath9 do not appear in the quadratic part of the hamiltonian . by taking the above basis
, those dm interactions can be included from the beginning .
secondly , the ground state can not be a pure spin singlet state due to the broken spin rotation symmetry in the presence of the dm interactions . in the conventional representation , or when
@xmath58 , @xmath53-bosons are sometimes called the triplons because they represents the spin triplet states .
although all of the @xmath53-bosons in our representation are not in the pure spin triplet states , we will call them the triplons throughout the paper . in order to write the hamiltonian ( [ eq : hamiltonian ] ) in terms of the bond operators , it is helpful to arrange the hamiltonian into the following form ; @xmath59 where @xmath60 includes the interactions with coupling constants @xmath27 and @xmath26 .
that is , where @xmath65 , and the set @xmath66 is given by @xmath67 in eq .
( [ eq : h_n ] ) and ( [ eq : i_3 ] ) , @xmath68 implies the lattice vector , @xmath69 and @xmath70 are dimer indices within the unit cell , @xmath71 and @xmath72 are indices to distinguish spins in a given dimer , and @xmath73 where @xmath74 is the lattice spacing . in our convention ,
spin @xmath72 is located at the vertices of the hexagon denoted with circle in fig .
[ fig : dimer_configuration ] . rewriting
the hamiltonian ( [ eq : h_n ] ) in terms of the bond operators is straightforward .
in addition , the following hardcore constraint is required at each dimer to restrict the hilbert space into the physical one ; @xmath75 moreover , the operators @xmath76 and @xmath49 are replaced by a number @xmath77 to describe the vbs order .
@xmath77 denotes the condensate density of the @xmath76-bosons .
therefore , we consider the following hamiltonian in the bond operator theory ; @xmath78 , \label{eq : bot_hamiltonian}\end{aligned}\ ] ] where @xmath68 is the lattice vector , @xmath69 dimer index within the unit cell , and @xmath79 the lagrange multiplier for the hardcore constraint . in principle , the condensate density @xmath77 and lagrange multiplier @xmath79 can depend on @xmath68 and @xmath69 .
however , based on the symmetry of the underlying lattice , they are taken to be uniform on the lattice . arranging the hamiltonian with respect to the degree of @xmath53-boson operators , @xmath80 where @xmath81 is the number of unit cells , @xmath82,\end{aligned}\ ] ] and @xmath83 and @xmath84 consist of quadratic and quartics terms of @xmath53-operators , respectively .
cubic terms are ignored in the above hamiltonian because magnetically disordered vbs states are of interest here .
the quartic part of the hamiltonian is decoupled in such a way that the resulting mean - field hamiltonian conserves @xmath90 .
readers interested in the details are referred to appendix [ appendix : mean - field ] .
the corresponding mean - field parameters are defined in ( [ eq : mean_field_paramters ] ) .
the resultant mean - field hamiltonian can be arranged in the following form .
@xmath91 where @xmath92^{t } , \nonumber\\\end{aligned}\ ] ] @xmath93 is a 12@xmath9412 hermitian matrix , and @xmath95 is the mean - field contribution to the energy and defined in ( [ eq : epsilon_pq ] ) . at this point
, it is important to note that the mean - field hamiltonian describes the triplon excitations and it is block diagonalized into three triplon modes of @xmath96 . as mentioned earlier
, @xmath97 is the magnetic quantum number of the triplon , which reflects the so(2 ) global spin rotation symmetry of the original and mean - field hamiltonian .
this fact can also be seen from the following expression of the magnetization ; @xmath98 the mean - field hamiltonian is diagonalized via the bogoliubov transformation@xcite which transforms @xmath53-bosons to bosonic quasiparticles ( @xmath99 ) ; @xmath100 where @xmath101^{t } , \nonumber\\\end{aligned}\ ] ] and @xmath102 is the transformation matrix .
the transformation matrix satisfies the following relation due to the bosonic statistics of the quasiparticles .
@xmath103 where @xmath104 is the diagonal matrix with the first ( second ) six diagonal elements being @xmath105 ( @xmath106 ) .
because of this bosonic constraint , @xmath107 is diagonalized instead of @xmath108 , so the eigenvalue problem has the following form ; @xmath109 where @xmath110 is the diagonal matrix containing eigenvalues @xmath111 in the first and second half of its diagonal .
it must be noticed that @xmath112 contains eigenvectors of the eigenvalue problem in its columns and the eigenvectors are normalized according to eq .
( [ eq : bosonic_constraint ] ) . diagonalizing the mean - field hamiltonian with
the above transformation leads to the following ground state energy per site . @xmath113\end{aligned}\ ] ] the energy is a function of the condensate density @xmath77 , lagrange multiplier @xmath79 , and mean - field parameters in ( [ eq : mean_field_paramters ] ) .
the ground state is determined by solving eq .
( [ eq : sc_eq_s_mu ] ) and ( [ eq : mean_field_paramters ] ) self - consistently .
@xmath114 in this subsection , the results of the bond operator mean - field theory are presented .
we use the coupling constants estimated by the series expansion study@xcite for rb@xmath0cu@xmath1snf@xmath2 ; @xmath115 where @xmath116 . for the dm interactions , increasing @xmath29 from 0 to 0.18
, we investigate the effect of the dm interactions on the ground state energy and the triplon dispersions . ) and ( b ) the spin gap ( @xmath117 ) versus the strength of the dm interaction ( @xmath29 ) . in the plots , the color represents the position of the spin gap ; light blue for the spin gap at the k point and red for the spin gap at the @xmath4 point .
the shift of the spin gap position from the k to @xmath4 point happens at @xmath32 . in figure ( b )
, the dashed line indicates the linear fitting for the spin gaps occurring at the @xmath4 point . according to the linear fitting , the spin gap is closed and a magnetic transition occurs at @xmath33 .
[ fig : sc_sol],title="fig : " ] ) and ( b ) the spin gap ( @xmath117 ) versus the strength of the dm interaction ( @xmath29 ) . in the plots , the color represents the position of the spin gap ; light blue for the spin gap at the k point and red for the spin gap at the @xmath4 point .
the shift of the spin gap position from the k to @xmath4 point happens at @xmath32 . in figure ( b )
, the dashed line indicates the linear fitting for the spin gaps occurring at the @xmath4 point . according to the linear fitting , the spin gap is closed and a magnetic transition occurs at @xmath33 .
[ fig : sc_sol],title="fig : " ] the ground state energy is plotted for various values of @xmath29 in fig .
[ fig : sc_sol ] ( a ) .
as shown in the figure , the ground state energy decreases monotonically once we include the dm interactions .
the energy at @xmath118 is slightly off from the general tendency .
this is due to the fact that we employed different mean - field decouplings for the cases of @xmath118 and @xmath119 because each case has different symmetry ; so(3 ) spin rotation symmetry when @xmath118 and so(2 ) symmetry when @xmath119 .
how the mean - field decoupling is done in each case is explained in the appendix [ appendix : mean - field ] . from the plot of the ground state energy , it turns out that the dm interaction tends to stabilize the valence bond state with the 12-site unit cell in the deformed kagome lattice antiferromagnet .
the ground state energy per site at @xmath34 is @xmath120 . .
the ground state energy per site ( @xmath121 ) , the spin gap ( @xmath117 ) , the critical magnetic field ( @xmath122 ) , and the saturation magnetic field ( @xmath123 ) when @xmath34 . [
tab : energy_spingap ] [ cols="<,^,^,^",options="header " , ] the above parameters are defined for @xmath125 listed in table [ tab : r_a_b ] and assumed to be independent of the lattice vector @xmath68 . with these mean - field parameters , the quartic part of the hamiltonian is decoupled into the following form .
@xmath126 where @xmath127 + h.c .
\nonumber\\ & + & \sum_{(m , n ) } b_{mn } \left [ -\left| p_{mn}^{++ } \right|^2 -\left| p_{mn}^{-- } \right|^2 + \left| q_{mn}^{+- } \right|^2 + \left| q_{mn}^{-+ } \right|^2 \right ] , \nonumber\\ \label{eq : epsilon_pq}\end{aligned}\ ] ] and a few comments on the mean - field decoupling are in order .
first , it must be noted that there is another possibility in decoupling the last terms with @xmath129 in ( [ eq : h_quartic ] ) .
it is to decouple them by using the direct channel ; @xmath130 \left [ t_{+,j}^{\dagger } t_{+,j } - t_{-,j}^{\dagger } t_{-,j } \right ] \nonumber\\ & \rightarrow & \left [ \langle t_{+,i}^{\dagger } t_{+,i } \rangle - \langle t_{-,i}^{\dagger } t_{-,i } \rangle \right ] \left [ t_{+,j}^{\dagger } t_{+,j } - t_{-,j}^{\dagger } t_{-,j } \right ] \nonumber\\ & + & \left [ t_{+,i}^{\dagger } t_{+,i } - t_{-,i}^{\dagger } t_{-,i } \right ] \left [ \langle t_{+,j}^{\dagger } t_{+,j } \rangle - \langle t_{-,j}^{\dagger } t_{-,j } \rangle \right ] \nonumber\\ & - & \left [ \langle t_{+,i}^{\dagger } t_{+,i } \rangle - \langle t_{-,i}^{\dagger } t_{-,i } \rangle \right ] \left [ \langle t_{+,j}^{\dagger } t_{+,j } \rangle - \langle t_{-,j}^{\dagger } t_{-,j } \rangle \right ] , \nonumber\\\end{aligned}\ ] ] where @xmath30 and @xmath6 mean @xmath131 and @xmath132 , respectively .
however , the direct channel mean - fields do not make any effect because they cancel each other so that their contribution becomes zero .
the cancellation originates from the time reversal symmetry of the original hamiltonian ( [ eq : h_j_h_dm ] ) . due to the symmetry ,
@xmath133 and the direct channel mean - fields have no effect .
secondly , the above mean - field decoupling is intended for the hamiltonian with nonzero dm interactions ( @xmath119 ) . for the case without the dm interactions ( @xmath118 ) , reader is referred to ref .
32 .
k. morita , m. yano , t. ono , h. tanaka , k. fujii , h. uekusa , y. narumi , and k. kindo , j. phys .
. jpn . * 77 * , 043707 ( 2008 ) ; t. ono , k. morita , m. yano , h. tanaka , k. fujii , h. uekusa , y. narumi , and k. kindo , phys .
b * 79 * , 174407 ( 2009 ) . | motivated by the recent neutron scattering experiment on rb@xmath0cu@xmath1snf@xmath2 [ nat .
phys .
@xmath3 , 865 ( 2010 ) ] , we investigate the effect of dzyaloshinskii - moriya interactions in a theoretical model for the magnetic structure of this material . considering the valence bond solid ground state , which has a 12-site unit cell , we develop the bond operator mean - field theory .
it is shown that the dzyaloshinskii - moriya interactions significantly modify the triplon dispersions around the @xmath4 point and cause a shift of the spin gap ( the minimum triplon gap ) position from the k to @xmath4 point in the first brillouin zone .
the spin gap is also evaluated in exact diagonalization studies on a 24-site cluster .
we discuss a magnetic transition induced by the dzyaloshinskii - moriya interactions in the bond operator framework .
moreover , the magnetization process under external magnetic fields is studied within the exact diagonalization and strong coupling expansion approaches .
we find that the results of all above approaches are consistent with the experimental findings . |
fluctuations of currents of matter and energy is an important subject of nonequilibrium statistical mechanics .
a prototypical model problem , which attracted much attention , involves a diffusive lattice gas driven by two reservoirs of particles kept at different densities @xcite .
this simple setting has a direct relevance to experiment in at least two different contexts : statistics of electric current in mesoscopic conductors @xcite and statistics of wave transmission through disordered media @xcite .
here we consider a different but closely related problem : transport of diffusing molecules into the living cell through receptors distributed on its surface @xcite .
the surrounding gas serves as a finite - density reservoir , whereas the cell receptors can be modeled as reservoirs kept at zero gas density . in their pioneering 1977 paper berg and purcell @xcite evaluated the expected steady - state current of non - interacting diffusing particles into a single receptor .
their motivation was to assess physical limitations of the cell s ability to sense changes in the environmental concentration @xcite .
building on their work , here we aim at ( i ) evaluating the current _ fluctuations _ at each receptor during a given time , and ( ii ) accounting for inter - particle _ interactions _ in the surrounding gas .
these extensions are presently possible due to recent advances in the fluctuating hydrodynamics @xcite and the macroscopic fluctuation theory ( mft ) @xcite of diffusive lattice gases .
these formalisms have already been successfully used in the simplest two - reservoir setting @xcite , and in several non - stationary settings which involved a single current .
as we will show here , the presence of multiple absorbing patches , leading to multiple currents , allows one to ask new questions , and brings qualitatively new effects .
one new question concerns the joint probability distribution of , and correlations between , the absorption currents at different receptors .
one qualitatively new effect is vorticity which , as we will see , is typically present in the most probable particle flux field conditioned on a specified joint absorption statistics .
we will model the surrounding gas of interacting particles as a diffusive lattice gas @xcite .
the large - scale long - time behavior of such gases can be described by fluctuating hydrodynamics @xcite .
the average particle density @xmath0 of a lattice gas obeys a diffusion equation @xmath1 , \label{diff}\ ] ] whereas macroscopic fluctuations are described by the conservative langevin equation @xcite @xmath2 where @xmath3 is a zero - mean gaussian noise , delta - correlated in space and in time . as one
can see from eq .
( [ lang ] ) , a diffusive lattice gas is completely specified by two transport coefficients : the diffusivity @xmath4 and the mobility @xmath5 . for lattice gases @xmath6 .
the simplest example of a diffusive lattice gas is a gas of non - interacting random walkers ( rws ) . for the rws one has @xmath7 , and @xmath8 @xcite .
an example of _ interacting _ lattice gas is the simple symmetric exclusion process ( ssep ) . here at each time
step a particle can jump , with equal probability , to any _ empty _ neighboring site .
the average behaviors of the rws and the ssep turn out to be identical : they share the same density - independent diffusivity @xmath9 .
the inter - particle interactions of the ssep are manifested at the level of fluctuations , as the ssep s mobility @xmath10 is a non - linear function of @xmath11 @xcite .
the lattice gases are not the only systems describable by the langevin equation ( [ lang ] ) .
important additional examples describe transport of noninteracting electrons in mesoscopic materials at zero temperature @xcite and wave transmission in disordered media @xcite .
now we formulate the model which we will study in this paper . consider a lattice gas of initially uniform density @xmath12 which fills the whole space outside of a simple - connected three - dimensional domain ( the cell ) of the characteristic linear size @xmath13 .
the domain boundary @xmath14 includes absorbing patches ( the cell receptors ) @xmath15 .
whenever a particle hits any of these , it is immediately absorbed .
whenever a particle hits the rest of the surface , @xmath16 ( the cell wall ) , it is reflected , see fig .
[ hure ] . .
a gas of particles ( black dots ) with a constant density surrounds a domain outlined by the thick line .
the reflecting part @xmath16 of the domain boundary is shown in gray , the absorbing patches @xmath17 , @xmath18 and @xmath19 are shown in white.,width=170 ] at times @xmath20 much longer than the characteristic diffusion time @xmath21 ,
the system reaches a non - equilibrium steady state . in the steady state
the average gas density @xmath22 and the average absorption current @xmath23 into each absorbing patch are independent of time . here
@xmath24 is the average number of particles absorbed by the @xmath25-th patch during the time interval @xmath26 . in this work
we will determine the joint probability distribution , @xmath27 , of observing small fluctuations , @xmath28 , of the absorption currents during the time @xmath20 . as we show here ,
this multi - variate probability distribution is gaussian and given by the expression @xmath29^{1/2}}}\exp\left(-\frac{t}{2}\sum_{i , j=1}^{s}\delta n_ic_{ij}^{-1}\delta n_j\right),\ ] ] where @xmath30 is an @xmath31 positive - definite symmetric matrix which depends on @xmath32 and on the geometry of the problem , but is independent of time .
we obtain a general condition when the _ sign _ of the currents cross - correlation @xmath33 is independent of the geometry and completely specified by the transport coefficients @xmath34 and @xmath35 of the gas .
we also determine the optimal ( most probable ) spatial profiles of the fluctuating gas density and flux fields , conditioned on the specified absorption current into each patch .
our calculations employ the mft in conjunction with the additivity principle , proposed by bodineau and derrida @xcite in the context of statistics of current in a one - dimensional lattice gas driven by the boundaries . by virtue of the additivity principle
the optimal absorption current into the @xmath25-th patch is independent of time and equal to @xmath36 . in order to calculate the variance of the probability distribution ( [ mainresult ] )
, we use the approach of ref .
@xcite and linearize the mft equations around the deterministic solution .
the additivity principle and linearization enable us to solve the problem in quite a general form and for an arbitrary diffusive lattice gas .
we show that the optimal flux field , conditioned on the specified absorption current into each patch , exhibits a vortex structure .
this feature is unique to multi - reservoir systems sustaining multiple currents , and it appears even when the surrounding gas is modeled as non - interacting rws .
the vorticity is absent in systems with a single current @xcite .
remarkably , it is also absent when we condition the process on the total absorption current into _ all _ patches combined .
although the original motivation for this work came from a living cell sensing via multiple receptors , many of our results can be generalized to other diffusive systems which are driven by multiple reservoirs and therefore sustain multiple currents .
the reservoirs can be disjoint ( rather than placed on a common reflecting boundary ) , and the system can be finite .
still , for concreteness we focus here on the setting shown in fig .
[ hure ] . here is how the remainder of the paper is organized . in sec
[ gena ] we present the mft formulation of the problem .
the deterministic limit is discussed in sec .
the joint distribution of fluctuating absorption currents is obtained in sec . [ solving ] and
analyzed in sec .
[ covv ] . in sec .
[ compa ] we briefly discuss the shot - noise - driven fluctuations of current in multi - terminal diffusive conductors , previously studied in ref . @xcite . in sec .
[ tot ] we determine the statistics of total absorption current into all patches combined , and relate it to previous findings for single - current settings . in sec .
[ flu2 ] we study optimal fluctuations of the density and of the flux fields and report a vortex structure of the optimal flux field .
section [ flu2 ] also discusses the particular case of the density and flux fields conditioned on the total absorption current .
section [ conc ] contains a discussion of our main results .
some of the technical derivations are relegated to appendices .
the starting point of the mft formulation is the langevin equation ( [ lang ] ) with the boundary conditions @xmath37 at the absorbing patches , and @xmath38 at the reflecting surface .
the fluctuating flux field @xmath39 is defined in eq .
( [ lang ] ) . here and in the following @xmath40 denotes a local unit vector normal to the domain boundary and directed into the domain . at @xmath41
the gas has a uniform density @xmath32 , @xmath42 the boundary condition at infinity is , therefore , @xmath43 in appendix [ mform ] we derive the mft equations for the joint absorption statistics .
the derivation , by now pretty standard @xcite , yields the governing equations [ eqs .
( [ d11 ] ) and ( [ d12 ] ) below ] and problem - specific boundary conditions .
we repeat the derivation here in order to establish the previously unknown boundary conditions on the absorbing patches .
the derivation starts from a path - integral formulation for eq .
( [ lang ] ) with specified numbers of absorbed particles @xmath44 by time @xmath20 .
the derivation exploits a large parameter the typical number of particles in relevant regions of space to perform a saddle - point evaluation of the path integral .
the ensuing minimization procedure yields the euler - lagrange equations which can be cast into a hamiltonian form for the optimal density history @xmath0 ( where @xmath11 plays the role of a coordinate " ) and the conjugate momentum density @xmath45 : @xmath46 , \label{d11 } \\
\partial_t p & = & -\frac{\delta h}{\delta \rho}=- d(\rho ) \nabla^2 p-\frac{1}{2 } \,\sigma^{\prime}(\rho ) ( \nabla p)^2 . \label{d12}\end{aligned}\ ] ] the hamiltonian @xmath47 is given in eq .
( [ hamilton1 ] ) , and the prime denotes the derivative with respect to the single argument .
the optimal flux field is given by : @xmath48 the boundary conditions in time are @xmath49 where here and in the following @xmath50 is outside of the domain .
the boundary conditions in space are the following .
far from the domain the gas is unperturbed , so @xmath51 on the domain boundary we have @xmath52 where @xmath53 are _ a priori _ unknown lagrange multipliers which are ultimately set by the @xmath54 constraints of having the specified numbers @xmath44 of particles absorbed by the patches by time @xmath20 .
the boundary conditions ( [ bcgen12 ] ) for @xmath55 generalize their simple analogs in single - current settings @xcite .
having solved the coupled nonlinear partial differential equations ( [ d11 ] ) and ( [ d12 ] ) with the boundary conditions in space and time , one determines the optimal history of the system conditioned on @xmath44 , @xmath56 . with the solutions at hand
, one can calculate the action @xmath57 which yields @xmath58 up to a pre - exponential factor : @xmath59 here and in the following the volume integral @xmath60 is performed over all space outside of the domain . for typical fluctuations the pre - exponential factor in @xmath61 , see eq .
( [ mainresult ] ) , is determined from normalization to unity .
the choice @xmath62 sets @xmath55 to vanish at all times and describes the deterministic solution , where all @xmath44 are equal to their expected values @xmath24 . in this case
( [ d11 ] ) reduces to the deterministic diffusion equation ( [ diff ] ) for the average density . at long times , @xmath63 ,
the solution approaches a stationary one , @xmath22 , obeying the time - independent equation @xmath64=0 \label{mfq}\end{aligned}\ ] ] and the boundary conditions in space . as a result , a steady - state particle flux field , and steady - state currents into the absorbing patches , set in . essentially , this is the problem which berg and purcell @xcite solved for gases with constant diffusivity @xmath9 .
they did it using an illuminating electrostatic analogy which can be easily generalized to a density - dependent diffusivity .
let us introduce the following function of the steady - state density @xmath65 : @xmath66=\int_{\bar{\rho}(\mathbf{x})}^{\rho_0}d(w)dw.\ ] ] then eq . (
[ mfq ] ) for @xmath22 becomes the laplace s equation for @xmath67 : @xmath68 the boundary conditions for @xmath22 transform to the following boundary conditions for @xmath67 : @xmath69 as in ref .
@xcite , @xmath67 can be interpreted as the electrostatic potential outside an insulating domain with boundary @xmath14 over which @xmath54 conducting patches @xmath70 , held at voltage @xmath71 , are distributed .
having solved for this potential , one obtains the complete solution of the deterministic problem : @xmath72 is obtained by inverting the relation for @xmath73 in eq .
( [ phi ] ) . for gases with constant diffusivity @xmath9 , such as the rws and ssep , eq .
( [ phi ] ) defines a linear relation @xmath74 which we will use in the following .
the deterministic steady - state flux field @xmath75 is minus the _ electric field _ of this system : @xmath76 in their turn , the average steady - state currents @xmath23 are ( up to a factor of @xmath77 ) the _ electric charges _ accumulated on the conducting patches : @xmath78 these charges are linearly related to the voltage @xmath79 via an @xmath31 symmetric _ capacitance matrix _ ( which we will denote by @xmath80 ) , determined solely by the problem s geometry .
the @xmath81 element of @xmath80 is the charge on the patch @xmath25 induced by the unit voltage applied to the patch @xmath82 , the rest of the patches being grounded @xcite .
@xmath80 can be expressed via a set of @xmath54 characteristic electrostatic potentials @xmath83 .
each of them appears when the corresponding conducting patch @xmath84 is held at unit voltage , the rest of the conducting patches are grounded , and the neumann boundary condition is specified at the reflecting part of the boundary @xmath16 .
the capacitance matrix is given by @xmath85 in its turn , @xmath86 from eq .
( [ phi ] ) can be expressed as @xmath87 using eqs .
( [ current])-([phii ] ) , we obtain @xmath88 to highlight the symmetry of the capacitance matrix @xmath89 , let us rewrite eq .
( [ a ] ) as @xmath90 where we used the boundary conditions for @xmath91 , the gauss theorem , and the fact that @xmath91 are harmonic functions in the bulk .
we will call the integrand of the last expression in eq .
( [ a2 ] ) , @xmath92 , _ the capacitance density _ of the system .
finally , the total absorption current into all patches in this interpretation is equal to the total charge on all the patches when they are held at voltage @xmath79 : @xmath93 where @xmath94 is the total capacitance of the system , that is the total charge accumulated on all the patches when they are kept at unit voltage .
as already mentioned , we will solve the mft problem under two simplifying assumptions .
the first is the additivity principle @xcite : being interested in the long - time limit , @xmath63 , we look for _ stationary _ solutions of eqs .
( [ d11 ] ) and ( [ d12 ] ) which satisfy the boundary conditions ( [ inftyq])-([bcgen13 ] ) @xcite .
the second assumption involves linearization of the stationary mft equations around the steady - state deterministic solution @xmath95 and @xmath96 .
this corresponds to typical , small fluctuations and suffices for the evaluation of the variance of the joint probability distribution @xcite .
note that , for the typical fluctuations , the additivity principle appears to be a safe assumption for all diffusive lattice gases @xcite .
let us denote small deviations of @xmath11 and @xmath55 from their average values as @xmath97 and @xmath98 .
they determine time - independent current deviations @xmath99 .
the linearized steady - state mft equations read : @xmath100+\sigma(\bar{\rho})\nabla p_1,\label{qstatglin}\\ \nabla^2p_1&=&0 \label{pstatglin}.\end{aligned}\ ] ] the boundary conditions for @xmath101 and @xmath102 follow from eqs .
( [ inftyq])-([bcgen13 ] ) : @xmath103 as one can see , the laplace s equation ( [ pstatglin ] ) for @xmath102 is decoupled from eq .
( [ qstatglin ] ) .
subject to the boundary conditions ( [ dbcgen12])-([dbcgen14 ] ) , it has a unique solution which can be expressed in terms of the auxiliary potentials @xmath83 , introduced in the previous section : @xmath104 this solution suffices for determining the action ( [ actionmain ] ) in terms of @xmath53-s .
indeed , in the leading order eq .
( [ actionmain ] ) yields @xmath105 plugging eq .
( [ p1 ] ) into eq .
( [ actionmainlin ] ) , we obtain the action in terms of a bilinear form in @xmath53-s : @xmath106 where @xmath107 is a vector with components @xmath53 , and @xmath30 is a @xmath31 symmetric matrix , @xmath108 given in terms of the volume integral of the product of the capacitance density of the system @xmath92 and the effective local noise magnitude @xmath109 $ ] .
as we will see shortly , this is the _
covariance matrix_. it is fully determined by the deterministic solution , and it plays a crucial role in our results . what is left is to express the @xmath53-s in eq .
( [ actionmainlin1 ] ) in terms of the current deviations @xmath110 .
this requires solving eq .
( [ qstatglin ] ) which , once @xmath102 is known , is a poisson s equation for @xmath101 : @xmath111=\nabla\cdot\left[\sigma(\bar{\rho})\nabla p_1\right]=\sum_{i=1}^s\lambda_i\nabla\cdot\left[\sigma(\bar{\rho})\nabla\phi_i\right].\label{pois}\ ] ] in appendix [ qrws ] we present an explicit solution of this equation for the rws . for a general lattice gas
an explicit solution is unavailable .
still , we were able to derive an explicit relation for @xmath110 vs. @xmath53 by using a green s function identity , see appendix [ lamcur ] .
this relation is given by the same matrix @xmath30 defined in eq .
( [ mat ] ) : @xmath112 where @xmath113 is the vector with components @xmath110 .
as shown in appendix [ mat1 ] , the symmetric matrix @xmath30 is positive definite and therefore invertible , enabling one to solve eq .
( [ sigmalin ] ) for @xmath114 . plugging this inverse relation in the bilinear form ( [ actionmainlin1 ] ) and using eq .
( [ actionmainlin ] ) , we obtain @xmath115 this probability distribution describes gaussian fluctuations .
normalizing it to unity , we arrive at the announced result ( [ mainresult ] ) . as it is clear from eq .
( [ lin ] ) , the joint statistics of absorption is encoded in the covariance matrix @xmath116 .
the diagonal elements of @xmath30 describe the variance of the current into the patch @xmath25 : @xmath117 where the brackets denote averaging with respect to the gaussian distribution ( [ mainresult ] ) .
the off - diagonal elements of @xmath30 describe cross - correlations between the currents into different patches : @xmath118 a necessary condition for nonzero cross - correlations is inter - particle interactions .
this can be seen from an alternative expression for @xmath30 which we derived in appendix [ offdiag1 ] : @xmath119^{\prime},\label{offdiag}\ ] ] where @xmath120 is the kronecker delta . for the non - interacting rws one has @xmath121 , so the integrand vanishes , and one is left with @xmath122 , no correlations . from eq .
( [ vari ] ) one can see that , for the rws , the variance of the number of absorbed particles in each patch , @xmath123 , is equal to the mean value @xmath24 , regardless of the system s geometry .
in fact , for the rws the steady - state mft problem can be solved exactly , without linearization .
we performed these calculations and found the complete joint distribution of the number of absorbed particles .
as to be expected , this distribution is equal to a product of independent poisson distributions with the mean @xmath24 .
a simple _ interacting _ lattice gas with zero cross - correlations of absorption by different patches is the zero - range process , where a particle can hop to a neighboring site with a rate which increases with the number of particles on the departure site but is independent of the number of particle on the target site @xcite .
for the zero - range process one has , in the hydrodynamic limit , @xmath124 @xcite , and the integrand in eq .
( [ offdiag ] ) vanishes .
going back to general @xmath34 and @xmath35 , we notice that the expression @xmath125 under the integral in eq .
( [ offdiag ] ) is everywhere positive @xcite .
therefore , if @xmath126^{\prime}<0 , \quad \mbox{or } \quad d(\bar{\rho})\sigma^{\prime\prime}(\bar{\rho})<d^{\prime}(\bar{\rho})\sigma^{\prime}(\bar{\rho } ) , \label{condition}\ ] ] for any value of @xmath127 $ ] , then the currents into different patches @xmath128 are all anti - correlated , @xmath129 , regardless of the system s geometry . in particular , this is true for the ssep , where @xmath130^{\prime}=-4 $ ] . if the inequality in eq .
( [ condition ] ) is reversed , the absorption currents are all positively correlated .
this happens for the kipnis - marchioro - presutti ( kmp ) model @xcite which describes an ensemble of agents on a lattice which randomly redistribute energy among neighbors .
using the kmp model one can study fluctuations of _ energy _
absorption by absorbing patches located on the domain boundary . for the kmp model @xmath131 and @xmath132 , and so @xmath130^{\prime}=4 $ ] .
remarkably , the same inequality ( [ condition ] ) guarantees the validity of the additivity principle for arbitrary currents @xcite , and also determines the sign of the two - point _ density _
correlation function @xcite , in single - current systems .
sukhorukov and loss @xcite studied shot - noise - driven fluctuations of current in multi - terminal diffusive conductors .
although theirs and our geometries are different , their expression ( 3.16 ) for the zero - frequency mode of the power spectrum of correlations of the current has a mathematical structure which resembles that of our covariance matrix ( [ mat ] ) . in their case
it is a volume integral over the conductance density " an analog of our capacitance density times the local noise magnitude .
if we closely examine the governing equations of both systems , this resemblance should not come as a surprise .
sukhorukov and loss started from a boltzmann - langevin description of the distribution function of electrons ( where noise comes from electron scattering on static impurities ) .
then , applying a diffusion approximation , they derived a time - independent linear langevin equation for the electrical potential @xmath133 inside the conductor : eqs .
( 2.12 ) and ( 2.13 ) of ref .
for an isotropic medium their equation can be written as @xmath134=0,\ ] ] where @xmath135 is a gaussian noise , delta - correlated in @xmath50 and @xmath136 .
the transport coefficients @xmath137 and @xmath138 depend on @xmath50 but are independent of the fluctuating potential @xmath133 and of time . equation ( [ multilan ] ) , therefore , is very different from the nonlinear time - dependent langevin equation ( [ lang ] ) which was the starting point of our analysis .
however , in order to make a progress within the mft formalism , we employed the additivity principle and linearization .
in fact , we could have derived the same final results from the following langevin equation which describes small quasi - stationary density fluctuations @xmath139 around the steady - state average density profile @xmath22 : @xmath140\nabla\rho_1+d^{\prime}\left[\bar{\rho}(\mathbf{x})\right ] \rho_1\nabla\bar{\rho}+\sqrt{\sigma\left[\bar{\rho}(\mathbf{x})\right]}\,\boldsymbol{\eta}\right\}=0.\ ] ] the mathematical difference between the linear langevin equations ( [ multilan ] ) and ( [ linlan ] ) comes from the fact that the transport coefficients @xmath34 and @xmath35 in eq .
( [ linlan ] ) are @xmath11-dependent . in particular , the @xmath11-dependence of @xmath141 causes an additional contribution to the fluctuating flux field , @xmath142 , which is absent in eq .
( [ multilan ] ) . of course , this difference reflects different physical problems which ref .
@xcite and this work address . notwithstanding these differences
, there is a close similarity between the two problems .
an extensively discussed regime in ref .
@xcite is that of zero - temperature elastic scattering , manifested in a particular form of the spatial mobility profile @xmath143 in eq .
( [ multilan ] ) .
an important finding of ref .
@xcite for this case is the strictly negative sign of the cross - correlations of the current , regardless of the system s geometry .
remarkably , the profile @xmath143 for such a system turns out to be equivalent , up to irrelevant factors , to the mobility coefficient @xmath144 $ ] for the ssep . as we saw in the previous section ,
the cross - correlations of the absorption currents for the ssep are indeed strictly negative . here
we deal with arbitrary @xmath34 and @xmath35 , where the currents cross - correlations can behave differently .
in particular , the sign of the cross - correlations can be strictly positive , as it happens for the kmp model .
in addition , the mft formalism , which we employ here , makes it possible to determine the optimal fluctuating profiles , as we report in sec .
[ flu2 ] . given the joint absorption statistics , eq .
( [ mainresult ] ) or ( [ lin ] ) , what is the probability density that the _ total _ current into all the patches combined , @xmath145 , is equal to a prescribed value @xmath146 ? to answer this question , we can minimize the action ( [ lin ] ) under the constraint @xmath147 or , equivalently , @xmath148 .
the minimization , performed in appendix [ optimum ] , yields the following optimal values of the individual currents @xmath149 : @xmath150 the corresponding lagrange multipliers @xmath53-s [ see eq .
( [ bcgen12 ] ) ] turn out to be all equal : @xmath151 where @xmath152 note that @xmath153 , see eq .
( [ v ] ) .
now we can compute the action by using either eqs .
( [ lin ] ) and ( [ optn ] ) or eqs .
( [ actionmainlin1 ] ) and ( [ optlam ] ) .
an explicit result can be obtained with the help of the relation @xmath154 derived in appendix [ optimum ] . in this way
we arrive at a gaussian distribution of the total absorption current @xmath146 : @xmath155 the first two distribution cumulants of the number of absorbed particles @xmath156 , following from eq .
( [ sn ] ) , are @xmath157 where we have used eq .
( [ current2 ] ) . remarkably , these cumulants are equal to the total capacitance @xmath158 multiplied by the cumulants of the integrated current , obtained for a _ one - dimensional _ lattice gas driven by reservoirs at @xmath159 and @xmath160 , see eq .
( 3 ) of ref .
@xcite ) .
relations of the same type were established , for _ all _ cumulants of the current , in ref .
@xcite which dealt with the ssep driven by two reservoirs in a finite system in any spatial dimension .
there too the cumulants of the current are equal to the corresponding cumulants of the one - dimensional system multiplied by a constant geometrical factor @xcite . as was noticed in ref .
@xcite , this geometric factor is the electric capacitance .
our setting involves an infinite system with a single reservoir at infinity .
still , as we have just shown , the first two cumulants of the _ total _ absorption current have the same structure as those in the finite two - reservoir settings .
( for a spherically symmetric absorber this property was established earlier , and for the complete absroption statistics , in ref .
@xcite . ) extending the results of refs . @xcite and @xcite ,
the relations ( [ com ] ) show that the problem s geometry affects the current distribution only through the electric capacitance @xmath158 , or equivalently through the value of the average current @xmath161 .
here we determine the optimal profiles of the gas density and flux fields conditioned on specified absorption currents into each patch .
let us start with the gas of non - interacting rws . here
the calculations are straightforward , see appendix [ qrws ] . the resulting stationary optimal density profile , @xmath162 can be written as @xmath163,\label{qrw}\ ] ] where , using eqs .
( [ phirw ] ) and ( [ phii ] ) , we have : @xmath164.\label{rorw}\ ] ] the optimal flux field is ( see appendix [ qrws ] ) : @xmath165 let us calculate the vorticity @xmath166 . taking the curl of both sides of eq .
( [ jrw ] ) we obtain , after cancelation of two terms , @xmath167 this quantity is , in general , non - zero . that is , the optimal fluctuating flux field is , in general , _ not _ a potential vector field , in contrast to the average flux field ( [ j ] ) .
equation ( [ vorrw ] ) can be generalized to an arbitrary lattice gas . for small fluctuations
the flux field is equal to @xmath168+\sigma(\bar{\rho})\nabla p_1,\ ] ] see eq .
( [ qstatglin ] ) .
the first two terms in this right - hand - side are potential vector fields , so a non - zero contribution to @xmath169 can come only from the last term . substituting @xmath102 from eq .
( [ p1 ] ) and using eqs .
( [ phi ] ) and @xmath170 , we obtain @xmath171 the explicit result ( [ vorrw ] ) for the rws is a particular limit of this relation . as it is clear from eq .
( [ vortex ] ) , for a fluctuating system ( that is , when @xmath172 for some @xmath25 ) the vorticity vanishes if and only if the two vector fields , @xmath173 and @xmath174 , entering eq .
( [ vortex ] ) , are parallel .
this happens in a single - current system , that is if there is only one absorbing patch .
( this also happens in the extensively studied finite two - reservoir system , also sustaining a single current . )
the vorticity also vanishes if all @xmath53 are equal to each other . as shown in the previous sec .
[ tot ] , this special case appears when the process is conditioned on the _ total _ absorption current into all patches combined , @xmath175 .
we will return to this special case in sec .
[ flutot ] .
let us examine the large-@xmath176 asymptotic of the optimal vorticity field .
the fields @xmath177 , comprising the cross product in eq .
( [ vortex ] ) , can be expanded in multipoles , see e.g. ref .
@xcite , chapter 4 .
the two monopole terms , each decaying as @xmath178 , do not contribute to the cross product , as they are directed in the radial direction . the leading contribution to the vorticity comes from the cross product of a monopole field of one potential and a dipole field of the other .
this product decays as @xmath179 regardless of the problem s geometry or specific lattice gas . to illustrate our results on the optimal density and flux fields
, we consider a gas of rws in contact with a sphere of radius @xmath180 placed in the origin .
the whole sphere is absorbing , and we are interested in the absorption statistics of each of the two hemispheres : the northern one , @xmath181 $ ] , which we denote by @xmath182 and the southern one , @xmath183 $ ] , denoted by @xmath184 . here
@xmath185 is the polar angle of the spherical coordinates .
the problem possesses cylindrical symmetry .
the effective potential , defined in eq .
( [ phi ] ) , is @xmath186 , where @xmath187 is the radial coordinate
. the average absorption currents ( [ current ] ) to each of the two hemispheres are identical and equal to @xmath188 .
the set of effective potentials @xmath91 can be found explicitly via expansion in spherical harmonics : @xmath189 where the plus and minus signs refer to the northern and southern hemispheres , respectively .
the function @xmath190 is given by @xmath191 @xmath192 are the legendre polynomials , and @xmath193 is the gamma function . plugging expressions ( [ hemi ] ) and ( [ hemif ] ) in eq .
( [ vorrw ] ) we obtain , after some algebra , the vorticity : @xmath194 where @xmath195 is the unit vector in the azimuthal direction @xmath196 in spherical coordinates . using eq .
( [ hemif ] ) , we can obtain the large-@xmath187 behavior of the vorticity : @xmath197 the @xmath198 dependence is expected from the general argument given above .
the vorticity exhibits a delta - function singularity at the equator of the sphere : @xmath199 we can also find the optimal flux and optimal density fields in this example by plugging expressions ( [ hemi ] ) and ( [ hemif ] ) in eqs .
( [ qrw])-([jrw ] ) .
figure [ qrwfig ] shows a vertical cross - section of the optimal density field @xmath200 for @xmath201 and @xmath202 .
as intuitively expected , the optimal density is higher / lower than the mean density close to the northern / southern hemisphere , respectively . the optimal flux field ( [ jrw ] ) and the vorticity field ( [ vortex2 ] ) in this case are shown in figs . [ jrwfig ] and [ rotjrwfig ] , respectively .
evident in fig .
[ rotjrwfig ] is the singularity of the vorticity at @xmath203 , displayed by eq .
( [ sing ] ) .
the flux field ( [ jrw ] ) exhibits a milder singularity at @xmath203 , where the radial component of the flux is discontinuous as a function of @xmath185 at @xmath204 . these singularities are due to the discontinuous boundary conditions for the potentials @xmath91 s , and they are generic when considering several patches ( or reservoirs ) in direct contact with each other .
the density field ( [ qrw ] ) is continuous at @xmath205 .
the optimal density and flux fields , conditioned on the total absorption current deviation @xmath206 , are given by the solution of the linearized mft equations ( [ qstatglin ] ) and ( [ pstatglin ] ) alongside with eqs . ( [ optn ] ) and ( [ optlam ] ) .
in this case the equations can be explicitly solved for any lattice gas , see appendix [ totqp ] , and we obtain @xmath207 , \label{qq}\\ \!\!\!\!\!\!\!\!\!\!p_1&=&\frac{\delta n}{\bar{n}}\frac{i_1(\rho_0)}{i_2(\rho_0)}\tilde{i}_1(\bar{\rho},\rho_0),\label{pp}\end{aligned}\ ] ] where @xmath208\equiv\int_{\bar{\rho}(\mathbf{x})}^{\rho_0}d(w)\sigma(w)^{k-1}dw.\end{aligned}\ ] ] note that @xmath209 $ ] , see eq .
( [ phi ] ) .
( [ qq ] ) and ( [ pp ] ) show , both @xmath210 , and @xmath211 depend on @xmath22 alone .
remarkably , these @xmath22-dependences are described by ( the small - fluctuation limit of ) the optimal profiles @xmath212 and @xmath213 of the _ one - dimensional _ problem for the reservoir density values @xmath159 and @xmath160 , found in ref .
@xcite : @xmath214\right],\\ & & p(\mathbf{x})=p^{(1)}\left[\phi\left[\bar{\rho}\left(\mathbf{x}\right)\right]\right],\end{aligned}\ ] ] where @xmath73 is defined in eq .
( [ phi ] ) .
as one can see from eq .
( [ q1d ] ) , the average density @xmath22 plays the role of the natural spatial coordinate " of the problem .
( the same property was uncovered and exploited in a different geometry in ref .
@xcite . ) with respect to this spatial coordinate the problem is effectively one - dimensional , hence the relation to the current fluctuations of the one - dimensional system , described at the end of sec .
[ tot ] . finally , using eqs .
( [ qq ] ) and ( [ pp ] ) , we obtain the optimal fluctuating flux field ( [ qstatglin ] ) , conditioned on the total absorption current : @xmath215 where @xmath75 is the average steady state flux field ( [ j ] ) . as eq .
( [ totflux ] ) shows , in order to generate a total absorption current which is larger by a factor @xmath216 than the average one , the system has to increase the optimal flux by the same factor everywhere in space ( see also refs .
it is not surprising , therefore , that the optimal flux field , conditioned on the total absorption current , is vortex - free .
in this work we employed the macroscopic fluctuation theory ( mft ) to determine the statistics of particle absorption by several patches located on the surface of a domain immersed in a gas composed of interacting diffusing particles .
essentially , this work extends a number of previous results @xcite , obtained for different single - current settings , to a multiple - current setting . among central results of this work
are eqs .
( [ mat ] ) and ( [ offdiag ] ) for the covariance matrix of the absorption currents into different patches , and eq .
( [ condition ] ) which establishes , independently of the system s geometry , whether the absorption currents correlate or anti - correlate .
the same condition ( [ condition ] ) has recently gained much attention in a different context as a sufficient condition for the validity of the additivity hypothesis for arbitrary currents @xcite .
this coincidence hints at a possible relation between correlations and dynamical phase transitions via which the additivity property breaks down .
one particular example of diffusive transport , extensively studied in the past , describes electronic transport in mesoscopic wires @xcite . in this case
( [ lang ] ) , with @xmath141 and @xmath217 corresponding to the ssep , serves as a suitable mathematical description @xcite . as follows from eq .
( [ condition ] ) , the cross - correlations in this system are strictly negative . this result has been known for some time : it was obtained both by the scattering matrix method @xcite and by the effective langevin description @xcite .
it has been verified in experiment and gained much attention , see _
e.g. _ @xcite .
it is often referred to as the electronic hanbury brown and twiss effect @xcite , and is intimately related to the fermi statistics of charge carriers .
the mft framework which we developed here enables one to address a much broader class of diffusive systems by considering more general diffusivity and mobility .
as we have seen , this may lead to qualitatively different fluctuational behaviors , such as different signs of the correlation terms .
an important example of positive correlations in the ( energy ) absorption appears in the context of diffusive wave propagation in disordered media @xcite .
indeed , mesoscopic wave transport can often be described by eq .
( [ lang ] ) which relates the wave energy flux to the local wave intensity @xmath218 : an equivalent of particle density @xmath11 .
remarkably , this lattice gas " can be described in terms of the kmp model .
indeed , the corresponding diffusion coefficient is constant here , but the mobility behaves as @xmath219 @xcite . in this case
( [ condition ] ) predicts strictly positive cross - correlations in the photon absorption .
it would be very interesting to test this prediction in experiment .
the experimental setting can be similar to that of ref .
@xcite , where a waveguide filled with a disordered material was used to study the light - intensity correlations between two distant points along the waveguide .
another important advantage of the macroscopic fluctuation theory is that it predicts , for all settings , the optimal ( most probable ) spatial profiles of the density and of the flux , conditioned on a given multiple - current statistics .
as we have shown here , the corresponding optimal flux field generically exhibits vorticity .
the vorticity disappears if the process is conditioned on the _ total _ absorption current into all the patches combined .
it would be interesting to see whether the vorticity still disappears for the total absorption current if one goes beyond typical , small fluctuations which we addressed here .
we found that , _ under the assumption _ that the optimal flux field is vortex - free , the simple solution ( [ q1d ] ) [ which solves the time - independent _ nonlinear _ mft equations exactly ] , is unique . as a result
all the cumulants [ and not only the first two as in eq .
( [ com ] ) ] , are equal to the corresponding one - dimensional cumulants times the capacitance of the system . the irrotational character of the flux field is , however , a strong assumption .
its breaking , at a critical value of the total absorption current , may have a character of phase transition .
finally , the mft formalism can be readily extended to other geometries sustaining multiple currents .
it can also be extended , with some modifications , to the case when different reservoirs are kept at different densities .
we are grateful to eric akkermans , yaron bromberg , giovanni jona - lasinio and naftali smith for useful discussions of different parts of this work .
we acknowledge financial support from the israel science foundation ( grant no .
807/16 ) and the united states - israel binational science foundation ( bsf ) ( grant no .
2012145 ) .
there are several methods for derivation of the mft equations @xcite . here
we use the martin - siggia - rose formalism @xcite .
we start from eq .
( [ lang ] ) and represent the probability of observing a joint density and flux history @xmath220 , constrained by the conservation law @xmath221 , as a path integral : @xmath222 ^ 2}{2\sigma(\rho)}\right\}.\nonumber\end{aligned}\ ] ] using an integral representation for the @xmath223-function with the help of an auxiliary field @xmath45 , we can rewrite eq .
( [ path1 ] ) as a path integral over three unconstrained fields @xcite : @xmath224 \right\ } , \\\nonumber & & \mathcal { l}=\int_0^tdt\int d\mathbf{x}\left\{\frac{\left[\mathbf{j}+d(\rho)\nabla \rho\right]^2}{2\sigma(\rho)}+p\left(\partial_{t}{\rho}+\nabla \cdot \mathbf{j}\right)\right\}.\nonumber\\\end{aligned}\ ] ] we wish to evaluate the path integral over only those histories which led to @xmath44 particles being absorbed , by time @xmath20 , by the @xmath25-th patch , @xmath225 .
assuming that all characteristic length scales are macroscopic and include large numbers of particles , we can evaluate this path integral via a saddle - point approximation .
the dominant contribution comes from the optimal fluctuation : the most probable history @xmath226 leading to a specified number of absorbed particles @xmath44 .
the problem , therefore , reduces to finding the minimum of @xmath227 under the @xmath54 constraints : @xmath228 the latter can be incorporated via @xmath54 lagrange multipliers @xmath53 , @xmath225 : @xmath229\\ + \sum_i\lambda_in_i-\sum_i\lambda_i\int_0^tdt\oint_{\omega_i}\mathbf{j}\cdot \hat{n}ds\biggr\},\end{gathered}\ ] ] where the _ a priori _ unknown lagrange multipliers are ultimately set by the @xmath54 constraints ( [ cons ] ) .
taking the first variation of @xmath57 we obtain : @xmath230 ^ 2\\ -d(\rho)\nabla\cdot\left[\frac{\mathbf{j}+d(\rho)\nabla \rho}{\sigma(\rho)}\right]\biggr\}\\ + \int_0^tdt\oint_{\omega_r}\delta \rho \frac{d^2(\rho)}{\sigma(\rho)}\nabla \rho\cdot\hat{n } ds+\int d\mathbf{x}\delta \rho\ , p|_{t = t}\\ + \int_0^tdt\int d\mathbf{x}\boldsymbol{\delta j}\cdot\left[\frac{\mathbf{j}+d(\rho)\nabla \rho-\sigma(\rho)\nabla p}{\sigma(\rho)}\right]\\ + \sum_i\int_0^tdt\oint_{\omega_i}\left(p-\lambda_i\right)\boldsymbol{\delta j}\cdot\hat{n } ds,\\ + \int_0^tdt\int d\mathbf{x}\delta p\left(\partial_{t}{\rho}-\nabla\cdot \mathbf{j}\right).\end{gathered}\ ] ] here we have used the conditions @xmath231 following from the boundary conditions ( [ blang])-([blang3 ] ) .
we have also used gauss s theorem to transform volume integrals to surface integrals .
there are no contributions from the surface integral at infinity . setting @xmath232 to zero and using the fact that @xmath233 , @xmath234 , and @xmath235 can be arbitrary
, we see that the the optimal flux field is given by @xmath236 and the optimal profiles @xmath11 and @xmath55 satisfy the mft equations ( [ d11 ] ) and ( [ d12 ] ) in the bulk and the boundary conditions ( [ bt0])-([bcgen13 ] ) . the mft equations ( [ d11 ] ) and ( [ d12 ] ) are hamiltonian , where @xmath11 and @xmath55 play the role of the conjugate coordinate " and momentum " density fields .
the hamiltonian is @xmath237=\int d\mathbf{x}\,\mathcal{h},\end{aligned}\ ] ] where the hamiltonian density is given by @xcite @xmath238 finally , the action ( [ mform2 ] ) can be simplified to ( [ actionmain ] ) by using eqs .
( [ d11 ] ) , ( [ cons ] ) and ( [ jj ] ) .
the formal solution to eq .
( [ pois ] ) is given by green s function @xmath239 which satisfies the equations @xmath240 @xmath239 can be interpreted as the electrostatic potential at point @xmath50 , induced by a point charge @xmath241 at point @xmath242 outside the domain , when all the patches @xmath70 are conducting are grounded . with @xmath239 , the solution to eq .
( [ pois ] ) can be written as @xmath243(\mathbf{x})=\int d\mathbf{x'}g(\mathbf{x},\mathbf{x'})\nabla\cdot\left[\sigma(\bar{\rho})\nabla p_1\right](\mathbf{x'}).\label{qg}\ ] ] now let us evaluate the flux deviation @xmath244+\sigma(\bar{\rho})\nabla p_1,\ ] ] whose surface integrals over the patches give the corresponding current deviations : @xmath245+\sigma(\bar{\rho})\nabla p_1\right\}\cdot\hat{n}ds=\\\nonumber & & -\int d\mathbf{x'}\nabla\cdot\left[\sigma(\bar{\rho})\nabla p_1\right](\mathbf{x'})\oint_{\omega_i}\!\!\!\!\!\nabla_{\mathbf{x}}g(\mathbf{x},\mathbf{x'})\cdot\hat{n}ds\\ & & + \oint_{\omega_i}\sigma(\bar{\rho})\nabla p_1\cdot\hat{n}ds.\end{aligned}\ ] ] using a green s function identity @xmath246 @xcite , we obtain @xmath247\phi_i + \oint_{\omega_i}\sigma(\bar{\rho})\nabla p_1\cdot\hat{n}ds.\\\end{aligned}\ ] ] as the last step , we use gauss s theorem to transform the surface integral in eq .
( [ delta1 ] ) to a volume integral with the help of the potential @xmath91 defined in sec .
[ m ] : @xmath248\phi_i + \nabla \cdot \left[\sigma(\bar{\rho})\nabla p_1\phi_i\right]\big\}\\\nonumber & & = \int d\mathbf{x}\,\sigma(\bar{\rho})\nabla p_1\cdot\nabla\phi_i=\sum_j\lambda_j\int d\mathbf{x}\,\sigma(\bar{\rho})\nabla\phi_j\cdot\nabla\phi_i,\\\end{aligned}\ ] ] where in the last equality we have substituted @xmath102 from eq .
( [ p1 ] ) .
this linear relation can be written in a matrix form as in eq .
( [ sigmalin ] ) .
to prove this statement one needs to show that , for every nontrivial @xmath54-dimensional vector field @xmath249 , @xmath250 .
we have @xmath251 ^ 2,\label{bform}\end{aligned}\ ] ] which is clearly non - negative .
furthermore , since the fields @xmath177 are linearly independent , it is positive definite .
we start with rewriting ( [ mat ] ) : @xmath252,\end{aligned}\ ] ] where we have used the fact that @xmath91-s are harmonic functions . using gauss s theorem
, we can transform the last integral into a surface integral .
there are no contribution from the surface integral at infinity since both @xmath253 , and @xmath254 decay as @xmath255 as @xmath256 .
therefore , the last integral in eq .
( [ 2 ] ) becomes @xmath257\\ & & = \oint_{\omega } \left[\frac{\sigma(\bar{\rho})}{2}\nabla\left(\phi_i\phi_j\right)-\phi_i\phi_j\frac{\nabla\sigma(\bar{\rho})}{2}\right]\cdot\hat{n}ds \label{bb}.\end{aligned}\ ] ] as @xmath258 the surface integral is taken over the absorbing patches @xmath70 alone .
moreover , since @xmath259=\sigma(0)=0,\ ] ] the surface integral of the first term vanishes . finally , since @xmath260 we obtain @xmath257\\ & & = -\delta_{i , j}\oint_{\omega_i}\frac{\nabla\sigma(\bar{\rho})}{2 } \cdot\hat{n}ds\\\nonumber & & = \delta_{i , j}\oint_{\omega_i}\frac{\sigma^{\prime}(\bar{\rho})}{2d(\bar{\rho})}\left[-d(\bar{\rho})\nabla\bar{\rho } \right]\cdot\hat{n}ds=\delta_{i , j}\frac{\sigma^{\prime}(0)}{2d(0)}\bar{n}_i,\end{aligned}\ ] ] where in the last equality we have used the fact that , on the absorbing patches , @xmath261 , and identified the average currents ( [ current ] ) expressed through the average flux field ( [ j ] ) .
what is left to arrive at eq .
( [ offdiag ] ) is to rewrite the integrand of the second integral in eq .
( [ 2 ] ) as @xmath262=\frac{1}{2}\phi_i\phi_j\frac{\left(\nabla\phi\right)^2}{d(\bar{\rho})}\left[\frac{\sigma^{\prime}(\bar{\rho})}{d(\bar{\rho})}\right]^{\prime},\end{aligned}\ ] ] where we have used eq . ( [ phi ] ) and the fact that @xmath67 is harmonic function .
we seek the minimum of the bilinear form @xmath263 subject to the constraint @xmath264 introducing an additional lagrange multiplier @xmath265 , we minimize the function @xmath266\ ] ] with respect to all @xmath110 .
the minimization yields @xmath267 comparing this expression with eq .
( [ sigmalin ] ) , we see that @xmath268 now we can find the currents and all @xmath53-s in terms of the total absorption current @xmath146 .
we plug the identity @xmath269 in eq .
( [ mat ] ) and obtain @xmath270 now we introduce a single - variable function @xmath271 as the inverse function to @xmath272 defined in eq .
( [ phi ] ) . using @xmath271 , we can rewrite the integrand as @xmath273\nabla\phi\cdot\nabla\phi_i=\nabla\cdot\left\{\nabla\phi_i\int_0^{\phi}\sigma\left[\bar{\rho}(w)\right]dw\right\},\ ] ] where we have also used the fact that @xmath91 is a harmonic function . applying gauss s theorem for this form of the integrand in eq .
( [ sum ] ) we have ( the surface integral at infinity vanishes due to our choice of the lower bound of the integral on @xmath217 , and that @xmath177 decays as @xmath274 ) : @xmath275dw\right\}\nabla\phi_i \cdot \hat{n}ds\nonumber \\\nonumber & & = \frac{1}{v(\rho_0)}\sum_j\oint_{\omega_j } \left\{\int_0^{\phi}\sigma\left[\bar{\rho}(w)\right]dw\right\}\nabla\phi_i \cdot \hat{n}ds\\\nonumber & & = \frac{\int_0^{v(\rho_0)}\sigma\left[\bar{\rho}(w)\right]dw}{v(\rho_0)}\sum_j\oint_{\omega_j } \nabla \phi_i\cdot\hat{n}ds\\\nonumber & & = \frac{\int_0^{v(\rho_0)}\sigma\left[\bar{\rho}(w)\right]dw}{v^2(\rho_0)}\oint_{\omega_i } v(\rho_0)\sum_j\nabla \phi_j\cdot\hat{n}ds\\ & & = \frac{\int_0^{v(\rho_0)}\sigma\left[\bar{\rho}(w)\right]dw}{v^2(\rho_0)}\bar{n}_i= \frac{\int_0^{\rho_0}d(w)\sigma(w)dw}{v^2(\rho_0)}\bar{n}_i , \label{sumsigma1 } \end{aligned}\ ] ] where in the second and third equalities we used the boundary conditions : @xmath276 respectively . in the forth equality we employed the identity @xmath277 which reflects the symmetry of the capacitance matrix ( [ a ] ) . as a result , @xmath278 where @xmath279 is defined in ( [ alpha ] ) .
it is left to impose the constraint @xmath280 in order to set the value of @xmath265 .
this yields @xmath281 equations ( [ 111 ] ) and ( [ 222 ] ) yield eqs .
( [ optn ] ) and ( [ optlam ] ) .
for the non - interacting rws one can solve eq .
( [ pois ] ) explicitly . since
both @xmath91 and @xmath282 are harmonic functions [ see eq .
( [ phirw ] ) ] , we can rewrite the r.h.s . of eq .
( [ pois ] ) as @xmath283=2d_0\nabla\cdot\left(\bar{\rho}_{\text{rws}}\nabla\phi_i\right)=d_0\nabla^2\left(\bar{\rho}_{\text{rws}}\phi_i\right).\ ] ] therefore , eq . ( [ pois ] ) can be rewritten as a laplace s equation : @xmath284=0.\label{poisrw}\ ] ] by virtue of the boundary conditions for @xmath72 , @xmath91 and @xmath101 , the function under the laplacian vanishes on all the patches and at infinity , whereas its normal derivative vanishes on the reflecting part of the boundary @xmath16 . as a result , this function vanishes everywhere , and we obtain @xmath285 now we use eq .
( [ sigmalin ] ) alongside with @xmath286 ( see sec . [ covv ] ) and obtain : @xmath287 plugging this relation into eq .
( [ q1rw ] ) we arrive at eq .
( [ qrw ] ) . also , using eq .
( [ q1rw ] ) , we can evaluate the flux deviation @xmath288+\sigma(\bar{\rho})\nabla p_1.\ ] ] substituting here @xmath102 from eq .
( [ p1 ] ) and using ( [ lamrw ] ) , we arrive at eq .
( [ jrw ] ) .
we start from @xmath102 from eq .
( [ p1 ] ) .
all @xmath53 are given by eq .
( [ optlam ] ) . then , using eqs .
( [ phii ] ) and ( [ alpha ] ) , we obtain : @xmath289 where @xmath290 is defined by eq .
( [ itild ] ) .
now we turn to eq .
( [ pois ] ) for @xmath101 .
using @xmath291 we obtain @xmath292.\end{aligned}\ ] ] as a result , eq .
( [ pois ] ) for @xmath101 becomes laplace s equation , @xmath293=0,\ ] ] with inhomogeneous boundary conditions on the absorbing patches [ as can be deduced from the boundary conditions ( [ dbcgen1])-([dbcgen14 ] ) for @xmath101 and the definition ( [ itild ] ) of @xmath294 .
the solution gives @xmath101 as a function of the average density field @xmath22 in eq .
( [ qq ] ) . in analogy with other problems of this type @xcite , we expect that the full time - dependent solution of the mft problem exhibits narrow boundary layers in time near @xmath41 and @xmath295 which accommodate the boundary conditions ( [ bt0 ] ) and ( [ bt1 ] ) .
these boundary layers give only a subleading contribution to the action ( [ actionmain ] ) which we do not attempt to calculate .
each of the potentals @xmath91 decays monotonically from @xmath296 at the @xmath25-th patch to @xmath297 at infinity ( and on the other patches ) in compliance with the maximum principle for the laplace s equation . | we study fluctuations of particle absorption by a three - dimensional domain with multiple absorbing patches .
the domain is in contact with a gas of interacting diffusing particles .
this problem is motivated by living cell sensing via multiple receptors distributed over the cell surface .
employing the macroscopic fluctuation theory , we calculate the covariance matrix of the particle absorption by different patches , extending previous works which addressed fluctuations of a single current .
we find a condition when the sign of correlations between different patches is fully determined by the transport coefficients of the gas and is independent of the problem s geometry .
we show that the fluctuating particle flux field typically develops vorticity .
we establish a simple connection between the statistics of particle absorption by all the patches combined and the statistics of current in a non - equilibrium steady state in one dimension .
we also discuss connections between the absorption statistics and ( i ) statistics of electric currents in multi - terminal diffusive conductors and ( ii ) statistics of wave transmission through disordered media with multiple absorbers . |
the mechanisms driving the shutdown of star - forming galaxies ( sfgs ) , quenching , remains one of the least understood puzzles in galaxy formation to date . in the low - redshift universe ,
galaxies show a bimodal color distribution , accompanied by a bimodality of morphologies ( see * ? ? ?
* ; * ? ? ?
spiral galaxies have low bulge - to - disk ratios and are commonly the site of active star formation leading to blue colors .
passive galaxies are observed to be mostly spheroid - dominated .
the color bimodality has also been observed at higher redshifts ( e.g. , * ? ? ?
* ; * ? ? ?
these observations have been interpreted as evolutionary paths , in which one or several quenching processes cause the sfg to become red and passive on a short timescale ( see * ? ? ?
* ; * ? ? ?
studies of the shape of the mass function of passive galaxies over cosmic time using large surveys such as sdss , nmbs , zcosmos , ultravista and zfourge showed that the probability for a galaxy being quenched increases with its mass ( peng et al .
2010 ; brammer et al .
2011 ; ilbert et al . 2013 ; muzzin et al . 2013
; woo et al .
2013 ; tomczak et al . 2014 ) .
further evidence for quenching also comes from studies employing abundance matching techniques , where the cumulative abundance of galaxies is matched to that of haloes using the results of cosmological dark matter simulations alongside with observational constraints on the stellar mass function .
those infer low baryon fractions within dark matter haloes , hinting at an efficient quenching mechanism associated with significant gas mass loss for both low and high mass galaxies @xcite .
several quenching mechanisms have been proposed , which act to either remove the gas from the galaxy or prevent the existing / inflowing gas within the galaxy to form stars . for the high - mass regime of galaxies ,
powerful agn feedback , which may be induced by galaxy mergers ( e.g. , * ? ? ? * ) or internal evolutionary processes triggered by ( violent ) gravitational disk instabilities ( e.g. , * ? ? ?
* ) , may drive energetic outflows expelling gas out of the galaxy and heating the halo .
in addition , a so - called radio - mode feedback may suppress the cooling of gas onto the galaxy over a longer timespan ( croton et al .
2006 ) .
moreover , morphological quenching , proposed by @xcite may switch off or reduce the efficiency of star formation in galaxies through the existence of a dominant bulge which stabilizes the gas disk against gravitational instabilities ( see saintonge et al .
2012 and crocker et al .
2012 for observational hints in the local universe , and genzel et al .
2014 for galaxies at high redshift ) .
another process proposed to be responsible for the shut - down of galaxies is halo mass quenching ( e.g. , birnboim & dekel 2003 ; kere et al .
2005 ) . in this scenario ,
dark matter haloes exceeding a critical mass of @xmath5 are able to stop the flow of incoming cold gas onto their central galaxies via virial shock heating , which leads to a decrease of star formation and/or suppresses star formation over longer timescales . however , at @xmath6 ( with @xmath7 ) , gas is predicted to penetrate to the central galaxy through cold streams even in massive haloes ( e.g. , dekel et al .
2009 ) .
the above quenching mechanisms could plausibly leave an imprint on the structure of galaxies , as they may lead to high central concentrations due to internal processes and/or major mergers .
this idea triggered a plethora of studies investigating the relation between star formation activity ( or absence thereof ) and galaxy structure .
local galaxy surveys demonstrated that quiescence is strongly linked with structural and morphological parameters such as the index , stellar mass density and the central velocity dispersion @xcite .
this suggests that a high stellar mass surface density in the center of a galaxy is connected to its quenching .
recently , high - resolution imaging facilities onboard the hubble space telescope ( hst ) , including critically the nir wide - field camera 3 ( wfc3 ) , have allowed us to trace the relation between structure and stellar populations further back in cosmic time .
early studies on the basis of high - resolution hst - imaging found a correlation between color and bulge fraction for large samples of galaxies at @xmath8 ( bell et al .
2004 ; weiner et al . 2005
; koo et al . 2005 ) .
particularly the cosmic assembly near - infrared deep extragalactic legacy survey ( candels , * ? ? ?
* ; * ? ? ?
* ) has played a pivotal role due to its unique combination of multi - wavelength high - resolution imaging , large sample size , and depth .
a common theme of several of the early candels studies is that a correlation between galaxy structure and stellar populations ( a ` hubble sequence ' ) has already been in place since at least @xmath9 ( wuyts et al . 2011 ; bell et al .
2012 ; wang et al .
2012 ; lee et al .
2013 ; see also cheung et al .
. additionally , bruce et al .
( 2012 ) show that the majority of star - forming galaxies have disk - dominated light profiles and passive galaxies have bulge - dominated profiles directly from decomposed bulge+disk light - profile fitting of @xmath10 galaxies at @xmath11 .
however , it should be noted that they also find a significant fraction ( @xmath12 ) of passive disk - dominated galaxies and star - forming bulge - dominated systems in this high - mass regime .
these results have set on a firmer footing the pioneering studies by @xcite and @xcite , that used ground - based data and significantly smaller samples , respectively . at all epochs since at least @xmath9 , sfgs follow a relatively tight relation of @xmath13 dex scatter between their ongoing star formation rate ( sfr ) and assembled stellar mass , with a zero point that increases with lookback time .
this relation has been dubbed the main sequence ( ms ) of star - forming galaxies ( noeske et al .
2007 ; elbaz et al . 2007 ; daddi et al . 2007 ; rodighiero et al . 2011 ) .
typical sfgs that lie on the ms are best approximated by exponential disk profiles , while the quiescent galaxy population below the ms is better described with higher indices @xcite . in common to all of the above studies
, however , is that they were carried out on monochromatic ( albeit mostly rest - optical ) imaging , furthermore often exploiting only a subset of the entire candels data set .
if a galaxy component ( e.g. , disk ) forms significantly more stars than another component ( e.g. , bulge ) , their relative weight in light can differ significantly from their contributions in physically more relevant units of stellar mass .
deep panchromatic high - resolution imaging datasets enable us to go beyond measurements of light profiles , by reconstructing the distribution of stellar mass ( see , e.g. , zibetti et al .
2009 for a detailed application to nearby galaxies ) .
while pioneering work on pre - existing high - z datasets using the nicmos camera onboard hst was limited to small sample sizes ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , @xcite made use of resolved 7-band stellar - population synthesis modeling on a pixel - by - pixel basis of over 600 massive ( @xmath14 ) sfgs at @xmath15 and found that galaxies are overall smoother and more centrally concentrated in mass than they appear in light
. caused by a combination of internal extinction and star formation history variations , the presence of mass - to - light ratio variations within galaxies emphasizes the importance of making measurements on mass maps when characterizing the stellar structure of high- redshift galaxies .
guo et al .
( 2011 ) furthermore demonstrated that also quiescent galaxies at high redshift feature color ( and hence mass - to - light ratio ) gradients . in order to further shed light on the connection between galaxy structure and quenching , we reconstruct stellar mass maps of a large mass - selected sample ( @xmath16 ) of galaxies at @xmath15 and subject those to a detailed structural analysis . for this purpose
, we exploit the available multi - wavelength data in all 5 candels/3d - hst fields with accurate redshifts from the 3d - hst survey @xcite .
we do nt restrict the structural measurements to one - component ( ) fits , but rather also explore two - component ( bulge + disk ) decompositions .
the latter allows us to carry out a more direct comparison to semi - analytic models ( sams ; specifically those by somerville et al .
2008 , 2012 , further developed by porter et al .
2014 ) , whose prescriptions are formulated in units of disk and bulge components .
the paper is structured as follows . in section [ obs_sample.sec ] , we give an overview of the observations and sample .
the construction of stellar mass maps through resolved stellar population modeling , and the methodology to derive structural parameters , is described in section [ method.sec ] .
we present our observational results in section [ structure.sec ] , followed by a comparison with sams in section [ sam_results.sec ] . finally , we summarize and discuss the implications of our findings in section [ discussion.sec ] . throughout this paper
, we quote magnitudes in the ab system , assume a chabrier ( 2003 ) initial mass function ( imf ) , and adopt the cosmological parameters @xmath17 .
the core dataset used for this work is the deep space - based hst imaging from the candels multi - cycle treasury program @xcite , complemented with redshift information from the 3d - hst grism survey ( van dokkum et al .
2011 ; brammer et al .
2012 ) . in order to determine galaxy - integrated masses and sfrs , we make use of additional multi - wavelength data in the candels/3d - hst fields including space - based photometry from _
spitzer_/irac , _ spitzer_/mips and _ herschel_/pacs and an array of ground - based facilities ( see skelton et al .
2014 for a detailed description of the @xmath2-selected multi - wavelength catalogs produced by the 3d - hst team ) .
the galaxy - integrated properties are derived following identical procedures as @xcite .
these include stellar masses , based on @xmath18- to - @xmath19 broad - band sed modeling using population synthesis models from @xcite and sfrs derived from a ` ladder of sfr indicators ' .
the latter method uses detected emission in either uv + pacs for pacs - detected galaxies ( lutz et al . 2011 ; magnelli et al .
2013 ) or uv + mips 24@xmath20 for mips - detected galaxies to compute the sum of the obscured and unobscured sfr . for galaxies lacking an ir detection , the sfr is adopted from the best - fit sed model .
we verified that consistent results are obtained when splitting star - forming and quiescent galaxies based on the above measures of star formation activity or when adopting the rest - uvj diagram ( wuyts et al .
2007 ; williams et al .
2009 ) the hst candels observations used for this study comprise high - resolution imaging in five distinct fields : goods - south , goods - north , cosmos , uds , and egs , covering a total area of 625 arcmin@xmath21 .
typical limiting depths in 160 are 27.0 mag for candels / wide and 27.7 mag for candels / deep ( the central halves of the goods fields ) , respectively .
additional data used for this work include pre - existing acs imaging in the goods , egs , and cosmos fields ( giavalisco et al .
2004 ; davis et al . 2007 ;
koekemoer et al .
2007 ) , plus wfc3 imaging in ers @xcite .
the available passbands are @xmath22,@xmath23,@xmath24,@xmath25,@xmath26,@xmath27 for the goods fields and @xmath23,@xmath28,@xmath26,@xmath27 for the remaining fields . additionally , @xmath29 imaging was available for ers as part of goods - south and @xmath30 for the regions with candels / deep coverage .
all imaging used in our analysis was drizzled to a @xmath31 pixel scale . for details on the observations and data reduction
, we refer the reader to @xcite and @xcite .
the main focus of this paper is on the structural shape of stellar mass distributions in a mass - selected sample of galaxies in the redshift range @xmath32 . for this purpose
, we only apply a mass cut @xmath14 to select our galaxies , well above the mass completeness limit of the @xmath27-selected catalogs for all five fields .
our sample consequently spans a wide range of sfrs , from normal sfgs on the main sequence to quiescent galaxies ( qgs ) below , that already formed the bulk of their stars . for selection
, we use the sed modeled galaxy - integrated mass estimates , but note that they match well the masses obtained by summing the resolved stellar mass distributions ( see also * ? ? ?
* ) . in order to determine redshifts for the galaxies in our sample ,
we use ground - based spectroscopic redshift information whenever available .
otherwise , redshifts are fitted to the combination of 3d - hst grism data and broad - band photometry .
the total sample comprises 6764 galaxies , of which 3839 and 2925 lie within the redshift range @xmath33 and @xmath34 , respectively . in the following ,
these two redshift ranges are referred to as @xmath35 and @xmath36 , respectively .
a detailed description of the resolved sed modeling can be found in @xcite . here , we review only the key steps involved and additional processing steps .
first , all images in the available wavelength bands are brought to the wfc3 160 resolution ( @xmath37 ) by using the iraf psfmatch algorithm .
next , a voronoi - binning scheme from @xcite is applied in order to ensure a minimum s / n level of 10 for each bin in the corresponding @xmath2-band image .
the binned multi - wavelength images are then fit with stellar population synthesis models from @xcite , assuming a chabrier ( 2003 ) imf , a uniform solar metallicity , and a calzetti et al .
( 2000 ) reddening law with visual extinctions in the range @xmath38 .
the adopted sfhs are exponentially declining , allowing for e - folding timescales down to 300 myr .
this choice is simplistic and may not be representative of the real sfhs ( e.g. , for sfgs see maraston et al .
however , we are here interested in the stellar mass , which , among sed - derived properties , is most robust against variations ( and uncertainties ) in model assumptions ( e.g. , papovich et al . 2001 ; shapley et al . 2005 ; maraston et al .
2010 ) . in order to conduct further structural analysis using the ( binned ) output of this sed modeling technique , we constructed pixelized ( instead of voronoi binned ) images of the stellar mass surface density distribution in the following way .
first , we construct a @xmath39 map which , within the galaxy s @xmath27 segmentation map , has uniform values for pixels belonging to the same voronoi bins .
pixels outside the segmentation map ( i.e. , containing sky noise and possibly faint wings of the galaxy extending below the signal - to - noise threshold ) are assigned the average @xmath39 of the nearest three voronoi bins .
the resulting expanded @xmath39 map is then combined with the 160 image to construct a final mass map at full @xmath27 resolution .
this ensures a smooth transition of the galaxy s mass profile from the brighter central parts to the faintest regions . in order to conduct a structural analysis of our galaxy sample ,
we employ the galfit ( peng et al .
2010 ) morphology fitting code to fit 2-dimensional parametric models to the stellar mass distribution . for comparison purposes , we model the 2d @xmath27 surface brightness distributions as well , following identical procedures .
the models we use comprise on the one hand single models to parametrize the galaxy s shape with the effective radius ( @xmath40 ) and index @xmath41 .
@xmath40 is the galactocentric distance containing half of the total light / mass , and @xmath41 is a measure of the cuspiness of the overall light / mass profile .
we allow @xmath41 to vary within the range @xmath42 . while the index @xmath41 is often used and indeed serves as an approximate measure of the contribution of the bulge , it is important to note that @xmath41 does not translate one to one to the bulge - to - total ratio @xmath3 ( see appendix a ; also andredakis et al .
1995 ; de jong et al . 2004
; cibinel et al . 2013 ; bruce et al .
2012 , bruce et al .
2014 in prep . ) . for a given bulge + disk composite profile ,
the best - fit index can be increased both by increasing its @xmath3 , and by leaving @xmath3 constant while growing the extent of the disk relative to that of the bulge ( i.e. , lowering @xmath43 ) .
for this reason , and to facilitate a more direct comparison to ( semi - analytic ) models that are parametrized in terms of bulge and disk units , we furthermore decompose the galaxies by fitting 2-component ( bulge + disk ) models .
in addition , we also performed one - component pure disk and pure bulge fits by forcing @xmath44 and @xmath45 , respectively .
for the bulge - to - disk decomposition , we adopt a procedure similar to that implemented in bruce et al .
( 2012 ) for the bulge+disk decomposition of @xmath27 light profiles , where we fixed the indices to @xmath44 for the disk and to @xmath45 for the bulge .
for the bulge component to address the impact on the derived @xmath3 values .
we find a median difference of @xmath46 for quiescent galaxies and @xmath47 for sfgs , with the errors marking the 1@xmath48 scatter .
any systematic trends are small , and our results are therefore robust against the precise value of @xmath49 adopted . ]
the centers are left free , but we restrict the relative distance between the bulge and disk centers to be less than 2 pixels . all other parameters defining the two components ( @xmath50 , the axial ratio @xmath51 , and the total magnitude / mass of both components ) are allowed to vary independently . in 2-component modeling , the higher number of degrees of freedom increases the odds of the fit being trapped in a local minimum . in order to mitigate this risk , we ran galfit using a grid of initial starting values .
our grid was constructed by using a range of size ratios between bulge and disk ( @xmath43 ) ranging from 0.1 to 1 , in steps of 0.1 . for each initial guess of @xmath43 , the corresponding initial
guess on @xmath3 was then set such that the index matching this initial configuration matches the one measured in the single - component fit ( see appendix a ) .
likewise , the initial magnitudes and absolute values of the initial size guesses for bulge and disk were set such that the total magnitude and half - light / mass radius of the composite profile matches the respective values determined from fitting . in cases
where galfit yields solutions with implausibly small bulge sizes ( @xmath52 pixel , corresponding to @xmath53 of the resolution , or @xmath54 50 pc at @xmath55 ) and flags the outcome as potentially not converged and unphysical , we excluded the respective run from the grid . also , solutions yielding a disk smaller in size than the bulge ( i.e. , @xmath56 ) were not included . galaxies . ] after performing the fits for each point of the grid with initial guesses , we assigned a final @xmath3 ratio for each object as the solution of the fit with the lowest @xmath57 . here , the pure disk and pure bulge fits were also included .
their solutions generally show higher @xmath57 than the 2-component fits , but are occasionally preferred over those in a @xmath57 sense ( oftentimes , these are bulgeless systems with @xmath58 ) .
the two - component decompositions are statistically preferred over the single models ( as based on both their @xmath57 values and the akaike information criterion , aic , where @xmath59 is the number of free parameters in the fit and @xmath60 is the number of data points used in the fit . ] ) for @xmath61 of the total sample . those systems
for which the single model yields a lower value of @xmath57 and aic typically feature shallow profiles with @xmath58 . in both single- and two - component fitting , we use an automated scheme which pre - determines neighboring sources that need to be masked or fitted simultaneously , and passes initial guesses of fitting parameters to galfit .
those include estimates on size , total magnitude and center .
the initial guess on size is based on the distance from the center to the radius at which the curve of growth reaches 50% of the galaxy s total flux .
the mass - weighted center of the galaxy derived within its segmentation map is adopted as initial estimate for the center .
galfit takes into account the convolution of the model with the point - spread - function ( psf ) . both for fitting light and mass
, we use a psf which is a combination of stacked stars and a tiny - tim @xcite model psf . for a more detailed description of the used psf , see @xcite .
we emphasize that throughout this paper our working definition adopted for bulge and disk components is based on the above bulge+disk modeling of stellar mass or light maps , as empirical constraints on whether or not stars assigned to a bulge / disk component are dynamically hot / cold are currently lacking . as the sins survey of @xmath55 galaxy kinematics demonstrated that high - redshift sfgs are dynamically ( frster schreiber et al .
2009 ) and morphologically ( frster schreiber et al . 2011 ) distinct from local spiral galaxies , we caution that bulge fractions as derived by our decompositions may have a somewhat different meaning than they would have in the local universe .
in addition , we estimated a typical measurement uncertainty on @xmath3 by setting up an array of model galaxies with a range of mass , size , @xmath3 ratio , and @xmath43 , which is similar to our data .
the grid of model galaxies consisted of 5 , 3 , 11 , and 10 grid points in mass , @xmath40 , @xmath3 and @xmath43 respectively .
these were then inserted in multiple empty sky regions of the candels uds field to mimic the typical background noise .
we next ran galfit on all 8250 mock galaxies using our 2-component fitting scheme .
the measurement error on @xmath3 of each galaxy in our sample given its mass , radius , and profile shape is then finally assigned as the scatter among the recovered @xmath3 ratios of the corresponding model galaxy . typical measurement errors in @xmath3 for star - forming and quiescent galaxies are on average @xmath62 and @xmath63 at @xmath64 and @xmath65 and @xmath66 at @xmath67 , respectively .
the distribution of errors peaks below the median ( @xmath68 ) , and shows a tail towards higher @xmath3 errors .
two alternative methods to estimate the uncertainty in @xmath3 , namely the formal random uncertainty reported by galfit and re - fitting the observed galaxies after applying additional background and poisson noise , generally lead to lower estimated uncertainties ( by a factor of @xmath69 in the case of galfit ) . in the remainder of the paper
, we therefore adopt the most conservative error estimates inferred from our analysis of the inserted mock galaxies .
exploiting the bulge - disk decompositions of the stellar mass maps derived for our sample of massive galaxies , we first evaluate the average mass budget of disks and bulges .
let us consider picking a random star out of our sample of massive galaxies above @xmath0 . at @xmath70 ,
the probability that this star belongs to a bulge component is @xmath71 . increasing the mass limit to @xmath72 or 11 yields a higher probability for the star to be associated to the bulge , of @xmath73 and @xmath74 , respectively .
perhaps somewhat surprisingly , the fraction of stars residing in a bulge component rises only slightly to @xmath75 , to @xmath76 , @xmath77 , and @xmath78 for galaxies more massive than @xmath79 , 10.5 , and 11 respectively .
the formal uncertainties to the above stated probabilities including sample variance and typical measurement errors on @xmath3 are limited to a few percent .
the total error budget is likely dominated by systematics , for example related to the assumptions made in stellar population modeling ( see section 3.1 ) .
we note , however , that only @xmath39 uncertainties with a differential impact on bulges and disks will affect the above numbers .
even if the @xmath39 ratio of bulges were systematically under- or overestimated by 0.2 ( 0.3 ) dex with respect to those of disks , the change in the above numbers would be limited to @xmath80 . as bulges ,
unlike stellar disks , can be considered sinks in the continuous assembly of a galaxy s stellar component , the rising mass density of stars in bulges ( by a factor of @xmath81 from the higher to the lower redshift bin ) therefore seems to be compensated largely by the continuing assembly of new disks . splitting our sample in finer redshift bins ,
we do find the fraction of stellar mass in bulges to increase more significantly , by a factor @xmath82 over the entire 6 gyr timespan sampled by our study .
overall , the bulge mass fractions are higher than what would be inferred from fits to the @xmath2-band surface brightness profiles , as the median mass - to - light ratio of disk components is 0.2 dex lower than of bulge components .
the above numbers address the evolving mass budget of disks and bulges for a mass - limited sample including both star - forming and quiescent galaxies . in the remainder of the paper
, we will delve into more depth by breaking down our sample by star formation activity . in recent years
, several hst - based studies have investigated the structural differences between star - forming and quiescent galaxies at high redshift . in common to all of these analyses ,
star - forming systems are found to have significantly larger rest - optical sizes than their quiescent counterparts at the same mass and redshift ( e.g. , toft et al . 2009 ; van der wel et al .
in addition , their surface brightness profile shapes tend to be shallow ( @xmath83 ) , while quiescent galaxies feature cuspier light profiles ( e.g. , wuyts et al . 2011 ; bell et al .
2012 ; cheung et al .
2012 , although see also bruce et al .
2012 for decomposed light profiles at high stellar masses and redshifts ) .
at the same time , the same multi - wavelength high - resolution lookback surveys have also established that substantial mass - to - light ratio variations in the rest - optical can occur , not only between but also within galaxies ( see , e.g. , wuyts et al .
2012 ; guo et al .
2012 ; boada et al . in prep ) .
typically , as new stars tend to form from gas settled in a disk configuration , such @xmath39 ratio variations are anticipated to give rise to a composite light profile in which the disk component has a relatively larger weight ( per unit mass ) than the bulge .
it is therefore important to address to which degree the above structural distinction between the two classes of galaxies is intrinsic to their distribution of stellar mass , or , conversely , can be attributed to stellar population effects .
the answer to this question is of immediate relevance to our understanding of quenching , as in principle the latter scenario could imply that compact quiescent systems can evolve from the star - forming main sequence by simple fading , without invoking an associated morphological transition .
kriek et al .
( 2009 ) investigate this scenario for a spectroscopically confirmed sample of massive @xmath36 galaxies , finding that 3 out of 6 massive star - forming systems have dense cores , and thus may passively evolve into compact galaxies due to fading of the outer star - forming regions .
szomoru et al .
( 2010 ) , on the other hand , exploit the exquisite depth of the hubble ultra deep field to probe the surface brightness profile of a massive compact quiescent galaxy at @xmath84 , ruling out the existence of a faint extended envelope or disk around the observed galaxy .
another argument against fading comes from cheung et al .
( 2012 ) , who derived the stellar masses of bulges for both star - forming and quiescent galaxies at @xmath85 .
they found that bulges of sfgs are half as massive as those of similar - mass quiescent galaxies , implying they can not simply fade onto the red sequence without structural evolution . using the stellar mass maps reconstructed for our mass - selected sample of 6764 galaxies at @xmath15 with @xmath86 , we are now able to draw statistically significant conclusions on the structural distinction between high - z galaxies prior to and after quenching . in figure
[ struc.fig ] , we compare the shape of the stellar mass distributions ( i.e. , corrected for spatial @xmath39 variations ) of star - forming and quiescent galaxies , and study their dependence on the total galaxy stellar mass .
we consider profile parameters based on single - component ( i.e. , ) fits as well as two - component ( bulge + disk ) decompositions , and show the results for two separate redshift intervals : @xmath87 and @xmath88 . in both cases
, we identified a galaxy as quiescent if its specific sfr ( ssfr ) satisfied @xmath89 , where @xmath90 is the hubble time at the redshift of the galaxy , and as star - forming otherwise .
we tested that a definition of quiescence based on the location of a galaxy in the uvj diagnostic diagram ( wuyts et al .
2007 ; williams et al . 2009 ) yields effectively identical results .
figure [ struc.fig ] immediately highlights that the distinct structural appearance of star - forming and quiescent galaxies is intrinsic to its internal distribution of stellar mass , and not just driven by stellar population or obscuration effects .
in fact , a comparison to the equivalent plots based on @xmath2-band surface brightness profiles rather than mass maps ( see appendix b ) indicates that stellar population effects ( when measuring at rest - optical wavelengths ) only induce a modest , albeit non - negligible shift . at all masses ,
quiescent galaxies feature cuspier stellar mass distributions ( i.e. , higher @xmath41 ) than star - forming systems .
their typical best - fit index is furthermore an increasing function of galaxy mass . interestingly ,
also among sfgs the profile shape is not independent of stellar mass .
an increase in @xmath41 is apparent above @xmath4 , both at @xmath35 and at @xmath36 .
a similar trend of increasing cuspiness at the tip of the ms was noted by wuyts et al .
( 2011 , figure 1 ; see also nelson et al . in prep ) .
next , it is worthwhile reflecting on what it is that we measure when fitting profiles .
appendix a illustrates that , when considering galaxies as superpositions of bulge and disk components , a given best - fit index does not necessarily correspond one - to - one to a unique @xmath3 value , even though it is often interpreted as such . given a bulge+disk system with associated best - fit @xmath41
, one can increase its @xmath41 either by boosting @xmath3 , or , alternatively , by growing the extent of the disk with respect to that of the bulge without any change to @xmath3 .
turning to the right - hand panel of figure [ struc.fig ] , we now explore the @xmath3 ratio as a function of galaxy mass , for sfgs and quiescent galaxies separately .
again , we find a clear anti - correlation between star formation activity and bulge prominence .
focussing on the star - forming population , the median @xmath3 is limited to below 30% for intermediate mass sfgs ( @xmath91 ) , while typical bulge mass fractions rise to 40 - 50% above @xmath4 .
we note that there is a significant scatter in the distribution of individual @xmath3 values around the median for both quiescent and star - forming galaxies .
we investigated the variation in median trends when varying the binning intervals , finding negligible changes at lower masses , while the median @xmath3 of the most massive ( @xmath92 ) sfg bin changes by @xmath93 , depending on the applied binning intervals is such that the most massive bin still contains more than 10 galaxies , allowing a robust estimation of the median . ] .
we note that measurements on the @xmath2-band yield bulge mass fractions among sfgs that are lower by on average @xmath94 , as can be understood from a disk component composed of a younger , lower @xmath39 stellar population than the bulge . from the two - component fits ,
we infer a typical @xmath43 size ratio of @xmath95 , albeit with significant scatter ( see figure [ n_bd.fig ] ) .
the median size ratio shows little dependence on star formation activity or mass , over the range probed by our sample .
given the enhanced @xmath3 values in quiescent galaxies , and the fact that bulges have smaller half - mass radii than disks , one may wonder if the difference in total size between sfgs and quiescent galaxies can be accounted for completely by a redistribution of stellar material from the disk to the bulge , without changing the extent of each of the components individually .
our analysis confirms that the change in @xmath3 of sfgs prior or during quenching is to a large extent responsible for the size difference between the quiescent and star - forming population .
however , some fraction of the shrinking size is still attributed to the individual components being smaller . in figure [ radii.fig ]
, we show the total sizes as well as the sizes of the individual components for star - forming and quiescent galaxies , as measured on the mass maps . while the total sizes of sfgs and qgs are noticeably different ( by a factor @xmath96 at @xmath97 ) , the difference in size of bulge and disk components between sfgs and qgs respectively is smaller , typically by a factor @xmath82 . with the morphological parameters of the mass maps for our entire galaxy sample in hand
, we now proceed to relate those with galaxy - integrated star formation properties .
the three panels of figure [ fquench_all.fig ] show , from left to right , the fraction @xmath98 of quenched galaxies as a function of total stellar mass , bulge mass , and disk mass , respectively . here
, we again define galaxies as quenched / quiescent when @xmath99 , and as star - forming otherwise .
the uncertainties in @xmath98 are derived via a bootstrapping method and represent the 68% confidence levels .
they include both sample variance and the typical measurement errors on @xmath3 , which are derived as described in section 3.2 .
for the bootstrapping , we computed @xmath98 for 1000 samples , which are randomly drawn from the original sample , with replacement . for each bootstrap iteration
, we displace the @xmath3 values for each galaxy by the typical measurement error in @xmath3 , given the galaxy s magnitude , size and measured profile shape .
the top left panel of figure [ fquench_all.fig ] illustrates that the fraction of quenched galaxies increases with increasing mass , from @xmath100 at around @xmath101 to @xmath102 at @xmath103 .
a second conclusion is that the fraction of quenched galaxies is overall higher for the lower redshift bin , by a factor of @xmath104 on average .
both of these results are well established in the literature .
the rising mass function of quiescent galaxies over cosmic time has most recently been quantified on a firm statistical footing by muzzin et al .
( 2013 ) and ilbert et al .
( 2013 ) , both of which exploit the wide - area ultravista survey .
what candels lacks in number statistics compared to ultravista , it adds in depth and high resolution .
exploiting these key strengths , we now turn to the dependence of the quenched fraction on galaxy subcomponents : the mass of their bulge ( middle panel ) and disk ( right - hand panel ) .
clearly , the dependence of @xmath98 on the bulge mass is much stronger than on the disk mass , which does not show any significant correlation with @xmath98 above @xmath105 . towards lower disk masses ,
@xmath98 increases rapidly , but we point out that this trend is entirely driven by the ( total ) stellar mass limit of our sample ( @xmath106 ; i.e. , the galaxies occupying the lowest @xmath107 bins are necessarily heavily bulge - dominated systems , that tend to form relatively few stars ) .
if lower mass galaxies were to be included , less massive , disk - dominated sfgs would likely outnumber these massive spheroids with small residual disks in the low @xmath107 bins , producing a flat relation of @xmath98 with @xmath107 over the full range probed . above respective masses of @xmath108 , @xmath98 increases more rapidly with bulge mass than with total stellar mass in both redshift ranges ( @xmath109 per dex of @xmath110 compared to @xmath111 per dex @xmath112 , or @xmath100 per dex of @xmath107 ) .
with the bulge - to - disk decompositions in hand , we next split the galaxy sample in bins of @xmath3 , and explore second parameter dependencies .
considering first the dependence of @xmath98 on the total stellar mass , it is apparent that , at a given total mass , @xmath98 is increasing significantly with increasing @xmath3 ratio .
the middle panels of figure [ fquench_all.fig ] illustrate that , when considering the dependence of @xmath98 on bulge mass , the different @xmath3 bins align along a much tighter locus .
in contrast , a large spread is seen as a function of disk mass ( right - hand panels of figure [ fquench_all.fig ] ) . in order to quantify these trends , we compute the spearman s rank correlation coefficient ( @xmath113 ) for the relations of @xmath98 with @xmath114 , @xmath115 , and @xmath116 for respective masses @xmath86 .
we find that @xmath117 is indeed significantly higher for the relation @xmath98 vs. @xmath116 ( @xmath118 ) than for both @xmath98 vs. @xmath119 ( @xmath120 ) and @xmath98 vs. @xmath115 ( @xmath121 ) , as measured for @xmath35 .
consistent results are found for @xmath36 .
we investigated the impact of defining quiescence based on a @xmath122 color - color criterion instead of a ssfr cut . when applying a uvj - based selection of quiescent galaxies , we find an overall good agreement with the trends presented in figure [ fquench_all.fig ] .
quantitatively , small changes occur , with @xmath98 increasing by @xmath123% for the entire @xmath15 sample integrated over all masses .
the good agreement is not surprising , since the precise threshold in ssfr used to select quiescent galaxies ( @xmath99 ) was chosen to yield maximum overlap with the @xmath122 selection criterion . taken together , this demonstrates that the build - up of a bulge seems to play a critical role in the quenching process of galaxies , whereas the disk does not .
the amount of stars in the disk component of a galaxy has little to no predictive power regarding its star - forming or quenched state , unless also @xmath3 ( and hence the bulge mass ) is known .
we find a qualitatively similar behavior at @xmath36 as at @xmath35 , but note that the cosmic evolution in the quiescent fraction can not solely be attributed to continuing bulge growth over time , as galaxies in the same @xmath112 and @xmath3 bin at @xmath35 are more likely to be quenched than those at @xmath36 .
appendix b illustrates how the equivalent diagrams composed from fits to the @xmath2-band surface brightness rather than the stellar mass distribution exhibit a larger spread from low to high @xmath3 bins .
this generic behavior can be understood from a physical picture where the disk component has a relatively larger weight in light than in mass .
our work is in agreement with , and takes the next step beyond previous reports that the inner stellar mass density is better related to the star formation history than the total stellar mass @xcite , as inferred from rest - optical imaging of smaller samples of high - redshift galaxies ( see kauffmann et al .
2003 and fang et al .
2013 for a local universe reference , and cheung et al .
2012 for intermediate redshifts @xmath124 ) importantly , the same behavior explored here over the redshift range @xmath15 extends in a strikingly similar fashion all the way to the present day , as demonstrated by bluck et al .
( 2014 ) who exploit the large number statistics of sdss .
semi - analytic models ( sams ) have a rich history of trying to reproduce galaxy scaling relations and abundances , with the goal of guiding our interpretation of the observational results . here
, we focus specifically on the sam developed by somerville et al .
( 2008 ) and further updated by somerville et al .
( 2012 ) and porter et al .
( 2014 ) , which is rooted in the bolshoi cosmological dark matter simulation ( klypin et al .
as is generic to all sams , the model relies on simplified analytic prescriptions for the dynamical and astrophysical processes down from entire galaxy scales , rather than on kiloparsec to parsec scales ( the resolution below which state - of - the - art cosmological and zoom - in hydro - simulations resort to subgrid physics , respectively ) .
this limitation , however , yields the enhanced flexibility of a relatively inexpensive runtime , allowing the straightforward generation of statistically significant model galaxy populations , and the tuning of parameters to observational constraints such as mass functions and scaling relations ( only empirical constraints from the nearby universe were used in tuning the parameters of the model considered here ) .
the fact that sams conceptually are formulated in units of bulge and disk components furthermore makes them suitable for a direct and meaningful comparison to the diagnostics explored in this paper .
a detailed description of the prescriptions for cooling , star formation , feedback , and structural growth is provided by somerville et al .
( 2008 , 2012 ) , with extensions and applications to the candels data set presented by porter et al .
the input to and output from the model is further contrasted to that of other sams by lu et al . ( 2013 ) .
for a detailed discussion of the physical recipes of this sam and the resulting output in the context of a larger set of sams , we refer the reader to lu et al .
for the sake of the comparison presented here , we emphasize that none of the relations investigated in this paper formed part of the set of observational constraints to which the model parameters were tuned .
model parameters of the sam were tuned to ( approximately ) match the global stellar mass function , the stellar mass function of early- and late - type galaxies , the gas fraction as a function of stellar mass for disks , and the mass - metallicity relation for stars . also of particular relevance
is the fact that , in the model , stars form either in the disk following a kennicutt - schmidt law ( kennicutt 1998 ) , where the disk scalelength is set by a similar methodology as mo , mao & white ( 1998 ) , or , in the event of a merger , during a starburst .
bulge formation as well as feeding of the central supermassive black hole can happen through two channels : mergers or disk instabilities ( see porter et al .
2014 ) . the starburst , black hole accretion and morphological transformation induced by mergers depends on the mass ratio and gas fraction , as calibrated using a large suite of binary merger simulations ( somerville et al .
2008 ; hopkins et al .
2009 ; somerville et al .
star formation is moderated through heating of gas by supernovae as well as through agn feedback .
in addition to the quasar mode , during which agn can drive powerful outflows , black holes also grow more gradually over longer timespans through the so - called radio mode ( i.e. , suppression of cooling via radio jets ) . no explicit connection between the bulge mass and quenching ( as may for example be expected from the toomre q stability criterion in a gravitational quenching scenario , see section 6.2 ) was built into the model . using galaxies from the somerville et al .
galaxies with @xmath125 are assigned as star - forming , the others as quiescent .
uncertainties are derived via bootstrapping to reflect the sample variance .
[ sam_somerville1.fig],scaledwidth=49.0% ] in figure [ sam_somerville1.fig ] , we show the model equivalent of figure [ fquench_all.fig ] , describing how the fraction of galaxies that are quenched depends on the total stellar mass , the mass of stars in the bulge , and the mass of stars in the disk component respectively .
as for the observations , we define the threshold for a galaxy to be quenched with a ruler moving with redshift : @xmath126 . this definition , rather than an application of the uvj diagnostic , remains closer to the direct output of the sam , avoiding a translation to mock spectral energy distributions , which would introduce additional assumptions and uncertainties . at first glance ,
the sam features several of the characteristic trends noted earlier for the candels galaxies : @xmath98 rises towards later cosmic times , increases with the total stellar mass , more steeply so with bulge mass , and shows no appreciable correlation with disk mass above @xmath127 . in the interval @xmath128 , a sharp drop with increasing @xmath107 is noted , as also seen in the observations .
given the @xmath129 threshold of our sample selection , the latter objects are necessarily heavily bulge - dominated . without imposing such a mass limit ,
less massive , disk - dominated sfgs would outnumber massive early - type galaxies with small residual disks in the lower @xmath107 bins . despite the qualitative success , quantitative differences in the quenched fractions of model galaxies with respect to those observed are clearly present .
the discrepancy is most severe at @xmath36 , where modeled @xmath98 values are , on an average over the whole displayed mass range , of order a factor @xmath130 short of observed , hinting at an underestimated quenching rate and/or inefficiency to prevent quiescent systems from rejuvenating . when splitting the sam galaxies at each @xmath112 , @xmath110 , and @xmath107 in bins of @xmath3 ( middle and bottom panels of figure [ sam_somerville1.fig ] ) , we reproduce a similar behavior as found for the real universe in section [ f_quench.sec ] .
namely , the total stellar mass acts as a poorer predictor of the quenched state of a galaxy . this situation can be remedied if in addition to @xmath112 also @xmath3 ( and hence @xmath110 ) is known .
quantitatively , the correlation between @xmath98 and @xmath110 is measured to be the strongest ( @xmath131 ) , whereas @xmath112 and @xmath107 only show weak correlation with @xmath98 ( @xmath132 and @xmath133 , respectively ) .
the quoted values of @xmath117 are measured for respective masses of @xmath134 and at @xmath35 .
the values for @xmath113 at @xmath36 are similar , with the correlation between @xmath98 and @xmath119 as well as between @xmath98 and @xmath115 being somewhat stronger . at @xmath87 ,
less than 20% of all massive galaxies with @xmath135 are classified as quiescent .
conversely , the majority of galaxies in the upper @xmath3 bin ( with @xmath136 ) have low ssfr .
these inferences are in common between the sam and the observations . also in agreement , is the fact that @xmath110 serves as a better predictor of @xmath98 than the total stellar mass , with different @xmath3 bins being more ( albeit not perfectly ) aligned along a single locus in the @xmath98 versus @xmath110 diagram . at @xmath88
, the model predictions are skewed towards too low @xmath98 values , as noted earlier . however , in relative terms the same generic behavior as a function of bulge prominence is notable .
we note that most of the trends in figure [ sam_somerville1.fig ] are driven by physical prescriptions in the sam affecting central galaxies rather than satellites , as centrals account for 80% ( 90% ) of the model galaxy population above @xmath137 ( 11 ) .
those massive galaxies classified as satellites are further subjected to additional environmental quenching processes , resulting in a higher @xmath138 for this particular subpopulation . as a function of halo and black hole mass using galaxies from the somerville et al .
sam . in the middle and bottom rows ,
the galaxy sample has been split in bins of @xmath3 .
galaxies with @xmath125 are assigned as star - forming , the others as quiescent .
uncertainties are derived via bootstrapping to reflect the sample variance .
[ sam_somerville2.fig],scaledwidth=45.0% ] given the qualitative agreement between model and observations , we can now pose the question how , in the context of the somerville sam , the relation between structure and stellar populations could be interpreted . to this end
, we consider the dependence of @xmath98 on two physical properties of galaxies in the sam that are not observationally accessible for our candels sample : the halo mass @xmath139 and the mass of the central supermassive black hole @xmath140 .
the top panels of figure [ sam_somerville2.fig ] illustrate that the probability of a galaxy being quenched increases towards high @xmath139 and high @xmath140 . in detail , however , the dependencies on the two look different .
the increase in @xmath98 with @xmath139 is gradual over nearly two orders of magnitude .
in contrast , a much sharper upturn of @xmath98 emerges above @xmath141 .
this behavior is especially notable at @xmath35 , but a rise above the same threshold is present in the @xmath36 population as well .
breaking the model galaxy population down by its structural properties , we find a wide spread in @xmath98 for different @xmath3 at a given @xmath139 .
naively , a different , less scattered behavior would be expected if halo mass quenching were the sole and dominant mechanism ( birnboim et al .
2007 ; dekel & birnboim 2008 ; dekel et al .
2009 ) . as a function of @xmath140 , on the other hand , a similar upturn in @xmath98 is present for all @xmath3 bins equally above @xmath141 .
the increased scatter in the relation between @xmath98 and @xmath110 compared to the tight correlation of @xmath98 and @xmath140 can be explained by the sam s @xmath140 - @xmath110 relation . for a given @xmath3 bin , the scatter in the @xmath140 - @xmath110 relation
is significant compared to the dynamic range in the bulge masses plotted . at a given bulge mass ,
the scatter stems in part from an anti - correlation between the black hole mass and the level of star formation activity ( ssfr ) .
our analysis illustrates that in the somerville model , which includes feedback from both supernovae and agn , the central supermassive black hole acts as the primary agent of quenching in massive galaxies , and its accumulated mass ( i.e. , the integral over past accretion activity ) is tightly related to the probability of finding a galaxy in a quenched state .
since the physical processes giving rise to bulge and black hole growth are the same in the sam ( mergers and disk instabilities ) , the stronger relation of @xmath98 with @xmath110 than with @xmath112 , present in both observations and model predictions , is not surprising .
we stress that in the somerville model , no direct causal link between the presence of a bulge and quenching is implemented .
the bulge is simply the accessible observable that correlates most tightly with the actual agent of quenching in this particular model : the supermassive black hole .
observationally , there is increasing evidence for agn - driven outflows of massive @xmath142 galaxies ( frster schreiber et al . 2013
; nesvadba et al .
2011 ; harrison et al .
2012 ; cano - daz et al .
2012 ) . given the shortcomings of the sam in a quantitative sense , notably its underprediction of the quiescent population at @xmath36 , additional or other quenching processes may be at play in the real universe .
one such process , that is causally linked to the presence of a bulge , could be morphological quenching ( martig et al . 2009 ; genzel et al .
in such a scenario , the high central stellar density provided by a bulge , stabilizes the gas disk and prevents it from forming stars .
while plausibly only a temporary measure ( as no gas is expelled , nor stopped from accreting through this mechanism ) , it could potentially contribute to suppressing star formation in @xmath36 galaxies more efficiently and/or preventing them from returning to the star - forming branch in the sam . in the local universe ,
star formation efficiencies of bulge - dominated systems are reduced by factors of @xmath143 compared to disk - dominated galaxies ( saintonge et al .
2012 ; martig et al .
in order to study the morphological differences between sfgs and quiescent galaxies , and to draw conclusions on the possible structural changes of star - forming galaxies as they move along the ms , we first examined the mass dependence of the profile shape of sfgs and quiescent galaxies as traced by the index and @xmath3 ratio .
we have shown that quiescent galaxies are structurally distinct from the star - forming population as seen by overall higher indices and @xmath3 ratios at a given stellar mass .
sfgs show rising trends of their median index and @xmath3 ratio with increasing stellar mass , with the latter rising up to @xmath144 above @xmath4 .
these findings give insights about the link between the structural evolution of sfgs and the quenching process as they move along the ms . analyzing the schechter functional forms of the sfg and qg mass function as a function of redshift , @xcite
conclude that the quenching rate of galaxies climbing the ms rises proportionally to the sfr ( and given the near - linear ms slope therefore also proportionally to the stellar mass , hence their terminology mass quenching ) .
this corresponds to a survival probability on the ms that drops exponentially with mass , implying that , while nearly all low - mass sfgs are destined to continue growing along the ms , toward the high - mass end the ms becomes progressively more dominated by near - to - be - dead sfgs .
in fact , the sub - unity slope of the ms , and possible flattening at the high - mass end ( whitaker et al . 2012 ) , may well be interpreted in this context : the typical sfg above @xmath4 is already undergoing some level of quenching , thereby deviating from the projected path along a sfr - mass relation of slope unity , that could be expected from cosmological accretion rates in the absence of quenching . tying in our observational results on galaxy structure ,
the deviation toward high median @xmath41 and @xmath3 at the massive end reflects the typical structure of soon - to - be - dead star - formers that account for the bulk of sfgs above @xmath4 .
the fact that they are structurally distinct implies that the morphological transition happens first , to be followed later by the departure from the ms .
bulge growth precedes quiescence .
such a morphological change prior to quenching is in line with qualitative predictions based on a toy model by dekel & burkert ( 2013 ) . in the latter study , about half of the star - forming disk galaxies at @xmath55
are predicted to evolve into compact star - forming blue nuggets due to violent disk instabilities before they are quenched into compact quiescent galaxies ( red nuggets ) .
an observed population of blue nuggets has been proposed by barro et al .
( 2013a , b ) to represent an evolutionary link , originating from extended disk galaxies , and evolving into compact quiescent systems .
this does not refute that galaxies also undergo further structural evolution after they are quenched .
at least part of the size growth ( cassata et al .
2013 ; van der wel et al . 2014 ) and evolution toward rounder axial ratios
( chang et al . 2013 ) has been attributed to ( minor and/or major ) dry mergers , and it is conceivable that similar processes contribute to the observed trend of increasing @xmath3 toward the massive end for the quiescent population .
we have demonstrated that the bulge mass of a system is well correlated with its quenched state and has a stronger predictive power of quiescence than the total stellar mass .
the observed trends of @xmath98 with total stellar mass , bulge mass and disk mass as viewed among galaxies in different @xmath3 bins are in good qualitative agreement with predictions from the somerville et al .
sam . in the context of this model ,
the growth of the central supermassive black hole , which is the primary quenching agent for massive galaxies in this sam , is tightly coupled with the growth of bulges through both merging and disk instabilities .
if a black hole - bulge scaling relation is in place during the peak of cosmic star formation as it is in the present - day universe ( hring & rix 2004 ) , our observational results together with the model comparison could therefore hint at the bulge not being the causal link to quenching , but rather the most accessible observational proxy for the agn acting as the quenching agent . in detail , however , there are quantitative differences between the sam and our observations , most severely in the highest redshift bin ( @xmath88 ) , where the observed quenched fraction exceeds the value predicted by the sam by a factor of @xmath145 .
the latter difference could hint at a need for more frequent , efficient , or lasting quenching , a possible mechanism we speculate about below .
we also note that the same behavior is not necessarily a generic feature to all sams ( see appendix c ) .
it is tempting to draw connections between the emerging bulges in massive ms galaxies out to @xmath9 revealed by our analysis , and recent observational results based on deep ao - assisted integral field data sets and grism spectroscopy over the same redshift range .
frster schreiber et al . (
2013 ) found a high prevalence of powerful nuclear outflows in @xmath146 galaxies driven by agn , which appear to be absent in galaxies at lower masses .
along with star formation driven winds in the outer parts of the galaxies , such outflows could efficiently remove gas out of galaxies and , in this way , contribute to the quenching process .
meanwhile , the 3d - hst and candels legacy programs have yielded evidence for nuclear depressions in the h@xmath147 equivalent width in @xmath35 sfgs ( nelson et al .
2012 , 2013 ; wuyts et al .
2013 ) . at the highest stellar surface mass densities
, star formation no longer appears to proceed in lockstep with the assembled stellar mass .
likewise , genzel et al .
( 2014 ) report on ring - shaped h@xmath147 distributions in @xmath36 sfgs , surrounding a more quiescent center where the dynamically inferred toomre q parameter significantly exceeds unity , owing to the emergence of a stellar bulge . as such
, the toomre stability criterion is satisfied in the central galaxy regions , which consequently could prevent the gas reservoir , if present there , from fragmenting and forming stars .
while this result suggests that some causal connection between bulge growth and quenching may be at play , it should be noted ( as is done also by genzel et al .
2014 ) that gravitational quenching by itself does not expel the gas present , neither does it stop the accumulation of a larger gas reservoir by continuing cosmological accretion .
additional maintenance mode might be required for a long - term shut - down of further gas supply .
we analyzed the structural properties of a sample of 6764 massive ( @xmath148 ) galaxies in the redshift range @xmath32 , by exploiting the multi - wavelength candels hst imaging data set in all five candels/3d - hst fields .
we carried out single - component ( ) fits and two - component ( bulge + disk ) decompositions , on stellar mass maps reconstructed from a resolved panchromatic sed modeling technique ( wuyts et al .
2012 , 2013 ) , as well as on images of the @xmath2-band surface brightness distribution .
in addition , we compared our findings to predictions by the state - of - the - art semi - analytic model from somerville et al .
( 2008 , 2012 ; with extensions including disk instabilities presented by porter et al . 2014 ) .
our main results are the following : 1 . at fixed stellar mass ,
quiescent galaxies have overall higher indices and @xmath3 ratios than sfgs as measured on their mass maps , in line with previous findings using monochromatic observations .
we find an increase of indices among sfgs with increasing total stellar mass , with the median mass profiles increasing from ( @xmath149 ) at @xmath0 , to @xmath150 above @xmath4 .
two - component bulge - disk decompositions confirm that the same rising trend is present when considering the median @xmath3 ratio of sfgs , which is rising up to @xmath144 above @xmath4 .
the same characteristic behavior is seen at @xmath35 and @xmath36 .
2 . quantifying the same trends on the h - band light profiles rather than the mass maps , the indices and @xmath3 fractions are overall lower for sfgs , confirming previous non - parametric measurements for a subset of our sample ( wuyts et al .
the emergence of bulges above @xmath4 in sfgs appears to be also slightly less prominent when viewed in light , consistent with the steepest color gradients ( blue disks with red central bulges ) being found among massive sfgs .
3 . the likelihood of a galaxy being quenched , as traced by the fraction of quiescent galaxies , is better correlated with the bulge mass than the total stellar mass and further shows no appreciable correlation with the amount of stellar mass in the disk component .
the quenched fraction at redshift 1 is on average higher by a factor @xmath151 than at redshift 2 .
4 . at a given total stellar mass ,
the quenched fraction exhibits a strong positive correlation with @xmath3 , while different @xmath3 bins are confined to a significantly tighter locus in a diagram of @xmath98 versus @xmath110 .
these findings imply that the bulge mass of a system is the single observable parameter with the most predictive power regarding its quenched state , although a somewhat tighter constraint on the probability of quiescence can be obtained if in addition also the total stellar mass is known .
the same trend is seen over the full redshift range probed , with the distinction that quenched fractions are lower at higher lookback times .
we find a good qualitative agreement between the semi - analytic model by somerville et al .
sam and our observational findings . since bulge and black hole growth
are tightly coupled in the sam , the strong dependence of @xmath98 on bulge mass follows rather naturally in this model .
our observational results can _ in the context of this model _ therefore be interpreted as the bulge being the closest observable proxy to the underlying agent of quenching : the black hole .
quantitatively , the largest discrepancy between model and observations is found in the highest redshift bin ( @xmath88 ) , where the observed quenched fraction is larger by a factor of @xmath145 than predicted by the sam .
we note that the same behavior is not necessarily a generic feature to all sams .
[ [ section ] ] the authors acknowledge fruitful discussions with edmond cheung , david c. koo , yu lu , casey j. papovich , mohammadtaher safarzadeh , benjamin j. weiner and steven p. willner .
support for program number hst - go-12060 and hst - go-12177 was provided by nasa through a grant from the space telescope science institute , which is operated by the association of universities for research in astronomy , incorporated , under nasa contract nas5 - 26555 .
the combination of our one and two - component fits on the mass maps enables us to examine empirically how the index relates the amplitude of the bulge ( as parametrized by @xmath3 ) on the one hand , and the relative size - ratio to the bulge and disk component ( @xmath43 ) on the other hand .
figure [ n_bd.fig ] illustrates that , while @xmath3 shows a clear correlation with @xmath41 , there is no unique one - to - one translation between the two . instead , the best - fit @xmath41 to a composite bulge+disk system additionally depends on the size ratio of the two components . in other words ,
a galaxy s sersic index could be increased by placing more material in the bulge , but also by growing the disk at fixed bulge size .
the observed candels galaxies occupy a surface in this three - parameter space ( @xmath41 , @xmath3 , @xmath43 ) that is in good agreement with what would be anticipated from sersic fits to idealized , noise - free bulge ( @xmath152 ) plus disk ( @xmath153 ) profiles ( squares and curves in the left - and right - hand panels of figure [ n_bd.fig ] , respectively ) .
evidently , for systems with @xmath3 close to 0 or 1 , the size ratio of the two components is ill - constrained as one of them contains barely any mass .
the galaxies in our full @xmath15 sample span the full range of @xmath3 values , and are located predominantly around bulge - to - disk size ratios of @xmath154 .
here , we investigate how the results of our structural analysis change when conducting the measurements on the @xmath2-band light images rather than on the stellar mass maps . to this end
, we present in figure [ light_struc.fig ] the same figures as discussed in section [ structure.sec ] , now using the index and @xmath3 values as inferred from the candels @xmath2-band imaging .
likewise , we compute the bulge and disk mass as @xmath155 and @xmath156 , respectively . overall , our analysis reveals a qualitatively similar mass dependence of @xmath41 and @xmath3 , and distinction between sfgs and quiescent galaxies as inferred from the mass maps . in detail , however , modest changes in @xmath41 and @xmath3 are notable . while the median @xmath35 ( 2 ) sfg has @xmath157 ( 0.20 ) , the typical bulge fractions increase to above 20% , and reach up to @xmath158 at the massive end , once spatial @xmath39 variations are corrected for
. likewise , the corresponding indices measured on @xmath2-band imaging for sfgs below @xmath159 are consistent with exponential disk profiles ( see also wuyts et al .
2011 ) , but are slightly cuspier as quantified on mass maps .
this is in line with findings based on smaller subsets of candels data by wuyts et al .
( 2012 ) and guo et al .
( 2012 ) .
the central 50th percentile intervals marked by the red and blue polygons are somewhat less confined in the plots based on stellar mass maps compared to the @xmath2-band results .
we interpret this to be due to an additional source of random uncertainty introduced by the resolved stellar population modeling .
the resolved stellar population modeling itself was motivated by the need to reduce the systematic biases associated with spatial @xmath39 ratio variations .
focussing on figure [ light_struc.fig ] , the qualitative trends of @xmath98 with total stellar mass , bulge and disk mass are very similar .
however , the lower two rows of figure [ light_struc.fig ] compared to figure [ fquench_all.fig ] show that the bins of lowest and highest @xmath3 are more separated from each other in @xmath98 in light than in mass .
this observation too is in line with the disks of sfgs being dominating by a younger stellar population than that of the bulge , shifting sfgs to lower @xmath3 .
the middle panels of figure [ light_struc.fig ] show a larger scatter than the corresponding panels in figure [ fquench_all.fig ] .
in order to evaluate to which extent the characteristic trends of the somerville et al .
sam , as presented in section [ sam_results.sec ] , are generic to all semi - analytic models , we considered an independent semi - analytic model by guo et al .
the guo et al .
( 2013 ) model is rooted in the dark matter backbone of the millennium simulation ( springel et al .
2005 ) , and its output tables with galaxy properties for snapshots of different lookback times are publicly available on the virgo millennium database ( g. lemson & the virgo consortium 2006 ) .
we present the equivalent plots of figure [ sam_somerville1.fig ] and [ sam_somerville2.fig ] in figure [ sam_guo.fig ] . as in the somerville model ,
quenched fractions are rising with increasing mass , increasing bulge mass , and decreasing redshift , but broken down in bins of @xmath3 , significant differences are notable . specifically
, @xmath98 does not monotonically increase with increasing @xmath3 at a given stellar mass .
andredakis , y. c. , peletier , r. f. , & balcells , m. 1995 , , 275 , 874 barro , g. , faber , s. m. , prez - gonzlez , p. g. , et al . 2013a , , 765 , 104 barro , g. , faber , s. m. , perez - gonzalez , p. g. , et al .
2013b , apj , submitted ( arxiv:1311.5559 ) behroozi , p. s. , conroy , c. , & wechsler , r. h. 2010 , , 717 , 379 bell , e. f. , wolf , c. , meisenheimer , k. , et al .
2004 , , 608 , 752 bell , e. f. 2008 , , 682 , 355 bell , e. f. , van der wel , a. , papovich , c. , et al .
2012 , , 753 , 167 birnboim , y. , & dekel , a. 2003 , , 345 , 349 birnboim , y. , dekel , a. , & neistein , e. 2007 , , 380 , 339 bluck , a. f. l. , mendel , j. t. , ellison , s. l. , et al .
2014 , , submitted ( arxiv:1403.5269 ) bournaud , f. , dekel , a. , teyssier , r. , et al .
2011 , , 741 , l33 brammer , g. b. , whitaker , k. e. , van dokkum , p. g. , et al .
2009 , , 706 , l173 brammer , g. b. , whitaker , k. e. , van dokkum , p. g. , et al .
2011 , , 739 , 24 brammer , g. b. , van dokkum , p. g. , franx , m. , et al .
2012 , , 200 , 13 bruce , v. a. , dunlop , j. s. , cirasuolo , m. , et al .
2012 , , 427 , 1666 bruzual , g. , & charlot , s. 2003 , , 344 , 1000 calzetti , d. , armus , l. , bohlin , r. c. , et al .
2000 , , 533 , 682 cano - daz , m. , maiolino , r. , marconi , a. , et al .
2012 , , 537 , l8 cappellari , m. , & copin , y. 2003 , , 342 , 345 cassata , p. , giavalisco , m. , williams , c. c. , et al .
2013 , , 775 , 106 chabrier , g. 2003 , , 586 , l133 chang , y .- y . , van der wel , a. , rix , h .- w .
, et al .
2013 , , 773 , 149 cheung , e. , faber , s. m. , koo , d. c. , et al .
2012 , , 760 , 131 ciambur , b. c. , kauffmann , g. , & wuyts , s. 2013 , , 432 , 2488 cibinel , a. , carollo , c. m. , lilly , s. j. , et al . 2013 , , 776 , 72 crocker , a. , krips , m. , bureau , m. , et al .
2012 , , 421 , 1298 croton , d. j. , springel , v. , white , s. d. m. , et al .
2006 , , 365 , 11 daddi , e. , dickinson , m. , morrison , g. , et al .
2007 , , 670 , 156 davis , m. , guhathakurta , p. , konidaris , n. p. , et al .
2007 , , 660 , l1 dekel , a. , & birnboim , y. 2008 , , 383 , 119 dekel , a. , birnboim , y. , engel , g. , et al .
2009 , , 457 , 451 dekel , a. , & burkert , a. 2013 , , 2954 de jong , r. s. , simard , l. , davies , r. l. , et al .
2004 , , 355 , 1155 elbaz , d. , daddi , e. , le borgne , d. , et al . 2007 , , 468 , 33 elmegreen , b. g. , elmegreen , d. m. , fernandez , m. x. , & lemonias , j. j. 2009 , , 692 , 12 faber , s. m. , willmer , c. n. a. , wolf , c. , et al .
2007 , , 665 , 265 fang , j. j. , faber , s. m. , koo , d. c. , & dekel , a. 2013 , , 776 , 63 ferrarese , l. , van den bosch , f. c. , ford , h. c. , jaffe , w. , & oconnell , r. w. 1994 , , 108 , 1598 frster schreiber , n. m. , genzel , r. , bouch , n. , et al .
2009 , , 706 , 1364 frster schreiber , n. m. , shapley , a. e. , genzel , r. , et al .
2011 , , 739 , 45 frster schreiber , n. m. , genzel , r. , newman , s. f. , et al .
2013 , , submitted ( arxiv:1311.2596 ) franx , m. , van dokkum , p. g. , schreiber , n. m. f. , et al .
2008 , , 688 , 770 genzel , r. , frster schreiber , n. m. , lang , p. , et al .
2014 , , 785 , 75 giavalisco , m. , ferguson , h. c. , koekemoer , a. m. , et al .
2004 , , 600 , l93 grogin , n. a. , kocevski , d. d. , faber , s. m. , et al .
2011 , , 197 , 35 guo , y. , giavalisco , m. , cassata , p. , et al .
2011 , , 735 , 18 guo , y. , giavalisco , m. , ferguson , h. c. , cassata , p. , & koekemoer , a. m. 2012 , , 757 , 120 guo , q. , white , s. , angulo , r. e. , et al .
2013 , , 428 , 1351 hring , n. , & rix , h .- w .
2004 , , 604 , l89 harrison , c. m. , alexander , d. m. , swinbank , a. m. , et al .
2012 , , 426 , 1073 hopkins , p. f. , hernquist , l. , cox , t. j. , et al .
2006 , , 163 , 1 hopkins , p. f. , somerville , r. s. , cox , t. j. , et al .
2009 , , 397 , 802 ilbert , o. , mccracken , h. j. , le fvre , o. , et al .
2013 , , 556 , a55 jaffe , w. , ford , h. c. , oconnell , r. w. , van den bosch , f. c. , & ferrarese , l. 1994 , , 108 , 1567 kauffmann , g. , heckman , t. m. , white , s. d. m. , et al . 2003 , , 341 , 54 kauffmann , g. , heckman , t. m. , de lucia , g. , et al .
2006 , , 367 , 1394 kennicutt , r. c. , jr .
1998 , , 498 , 541 kere , d. , katz , n. , weinberg , d. h. , & dav , r. 2005 , , 363 , 2 klypin , a. a. , trujillo - gomez , s. , & primack , j. 2011 , , 740 , 102 koekemoer , a. m. , aussel , h. , calzetti , d. , et al .
2007 , , 172 , 196 koekemoer , a. m. , faber , s. m. , ferguson , h. c. , et al .
2011 , , 197 , 36 koo , d. c. , simard , l. , willmer , c. n. a. , et al .
2005 , , 157 , 175 kormendy , j. , & kennicutt , r. c. , jr . 2004 , , 42 , 603 kriek , m. , van dokkum , p. g. , franx , m. , illingworth , g. d. , & magee , d. k. 2009 , , 705 , l71 krist , j. 1995 , astronomical data analysis software and systems iv , 77 , 349 lanyon - foster , m. m. , conselice , c. j. , & merrifield , m. r. 2012 , , 424 , 1852 lee , b. , giavalisco , m. , williams , c. c. , et al .
2013 , , 774 , 47 lemson , g. , & virgo consortium , t. 2006 , arxiv : astro - ph/0608019 lu , y. , wechsler , r. h. , somerville , r. s. , et al .
2013 , apj , submitted ( arxiv:1312.3233 ) lutz , d. , poglitsch , a. , altieri , b. , et al .
2011 , , 532 , a90 magnelli , b. , popesso , p. , berta , s. , et al .
2013 , , 553 , a132 maraston , c. , pforr , j. , renzini , a. , et al .
2010 , , 407 , 830 martig , m. , bournaud , f. , teyssier , r. , & dekel , a. 2009 , , 707 , 250 martig , m. , crocker , a. f. , bournaud , f. , et al .
2013 , , 432 , 1914 mo , h. j. , mao , s. , & white , s. d. m. 1998 , , 295 , 319 moster , b. p. , somerville , r. s. , maulbetsch , c. , et al .
2010 , , 710 , 903 muzzin , a. , marchesini , d. , stefanon , m. , et al .
2013 , , 777 , 18 nelson , e. j. , van dokkum , p. g. , brammer , g. , et al .
2012 , , 747 , l28 nelson , e. j. , van dokkum , p. g. , momcheva , i. , et al .
2013 , , 763 , l16 nesvadba , n. p. h. , polletta , m. , lehnert , m. d. , et al .
2011 , , 415 , 2359 noeske , k. g. , weiner , b. j. , faber , s. m. , et al .
2007 , , 660 , l43 papovich , c. , dickinson , m. , & ferguson , h. c. 2001 , , 559 , 620 peng , c. y. , ho , l. c. , impey , c. d. , & rix , h .- w .
2010 , , 139 , 2097 peng , y .- j .
, lilly , s. j. , kova , k. , et al .
2010 , , 721 , 193 porter , l. a. , somerville , r. s. , primack , j. r. , johansson , p. h. 2014 , , sumitted rodighiero , g. , daddi , e. , baronchelli , i. , et al .
2011 , , 739 , l40 rosario , d. j. , santini , p. , lutz , d. , et al .
2013 , , 771 , 63 saintonge , a. , tacconi , l. j. , fabello , s. , et al .
2012 , , 758 , 73 schiminovich , d. , wyder , t. k. , martin , d. c. , et al .
2007 , , 173 , 315 shapley , a. e. , steidel , c. c. , erb , d. k. , et al .
2005 , , 626 , 698 skelton , r. e. , whitaker , k. e. , momcheva , i. g. , et al .
2014 , , submitted ( arxiv:1403.3689 ) somerville , r. s. , hopkins , p. f. , cox , t. j. , robertson , b. e. , & hernquist , l. 2008 , , 391 , 481 somerville , r. s. , gilmore , r. c. , primack , j. r. , & domnguez , a. 2012 , , 423 , 1992 springel , v. , white , s. d. m. , jenkins , a. , et al .
2005 , , 435 , 629 strateva , i. , ivezi , . ,
knapp , g. r. , et al .
2001 , , 122 , 1861 szomoru , d. , franx , m. , van dokkum , p. g. , et al .
2010 , , 714 , l244 szomoru , d. , franx , m. , van dokkum , p. g. , et al .
2013 , , 763 , 73 toft , s. , franx , m. , van dokkum , p. , et al .
2009 , , 705 , 255 tomczak , a. r. , quadri , r. f. , tran , k .- v .
h. , et al .
2014 , , 783 , 85 van den bosch , f. c. , ferrarese , l. , jaffe , w. , ford , h. c. , & oconnell , r. w. 1994 , , 108 , 1579 van der wel , a. , bell , e. f. , hussler , b. , et al .
2012 , , 203 , 24 van der wel , a. , franx , m. , van dokkum , p. g. , et al .
2014 , , submitted ( arxiv:1404.2844 ) van dokkum , p. g. , brammer , g. , fumagalli , m. , et al .
2011 , , 743 , l15 wang , t. , huang , j .- s . , faber , s. m. , et al .
2012 , , 752 , 134 weiner , b. j. , phillips , a. c. , faber , s. m. , et al .
2005 , , 620 , 595 whitaker , k. e. , labb , i. , van dokkum , p. g. , et al .
2011 , , 735 , 86 whitaker , k. e. , van dokkum , p. g. , brammer , g. , & franx , m. 2012 , , 754 , l29 williams , r. j. , quadri , r. f. , franx , m. , van dokkum , p. , & labb , i. 2009 , , 691 , 1879 windhorst , r. a. , cohen , s. h. , hathi , n. p. , et al .
2011 , , 193 , 27 woo , j. , dekel , a. , faber , s. m. , et al .
2013 , , 428 , 3306 wuyts , s. , labb , i. , franx , m. , et al .
2007 , , 655 , 51 wuyts , s. , frster schreiber , n. m. , van der wel , a. , et al .
2011 , , 742 , 96 wuyts , s. , frster schreiber , n. m. , genzel , r. , et al .
2012 , , 753 , 114 wuyts , s. , frster schreiber , n. m. , nelson , e. j. , et al .
2013 , , 779 , 135 zibetti , s. , charlot , s. , & rix , h .- w .
2009 , , 400 , 1181 | exploiting the deep high - resolution imaging of all 5 candels fields , and accurate redshift information provided by 3d - hst , we investigate the relation between structure and stellar populations for a mass - selected sample of 6764 galaxies above @xmath0 , spanning the redshift range @xmath1 .
for the first time , we fit 2-dimensional models comprising a single fit and two - component ( i.e. , bulge + disk ) decompositions not only to the @xmath2-band light distributions , but also to the stellar mass maps reconstructed from resolved stellar population modeling .
we confirm that the increased bulge prominence among quiescent galaxies , as reported previously based on rest - optical observations , remains in place when considering the distributions of stellar mass .
moreover , we observe an increase of the typical index and bulge - to - total ratio ( with median @xmath3 reaching 40 - 50% ) among star - forming galaxies above @xmath4 . given that quenching for these most massive systems is likely to be imminent
, our findings suggest that significant bulge growth precedes a departure from the star - forming main sequence .
we demonstrate that the bulge mass ( and ideally knowledge of the bulge and total mass ) is a more reliable predictor of the star - forming versus quiescent state of a galaxy than the total stellar mass .
the same trends are predicted by the state - of - the - art semi - analytic model by somerville et al . in the latter , bulges and black holes
grow hand in hand through merging and/or disk instabilities , and agn - feedback shuts off star formation .
further observations will be required to pin down star formation quenching mechanisms , but our results imply they must be internal to the galaxies and closely associated with bulge growth . |
the einstein vacuum equations determine a manifold @xmath1 with a lorentzian metric with vanishing ricci curvature : @xmath2 the set @xmath3 : standard minkowski metric @xmath4 on @xmath5 describes the minkowski space - time solution of the system ( [ eq.1 ] ) , which is stable under small perturbation according to the remarkable work @xcite of d. christodoulou and s. klainerman in 1993 . in this paper
, the authors showed that for @xmath6 given @xmath7-dimensional manifold @xmath8 with a riemannian metric @xmath9 and a symmetric two tensor @xmath10 , such that @xmath11 is close to the euclidean space , @xmath10 is close to @xmath12 and @xmath13 satisfy the constraint equations : @xmath14_j^i[k_0]_i^j+[k_0]_i^i[k_0]_j^j=0,\quad \nabla^j[k_0]_{ij}-\nabla_i[k_0]_j^j=0\ ] ] where @xmath15 is the scalar curvature of @xmath9 and @xmath16 is the covariant differentiation w.r.t .
@xmath9 , then we can find out a @xmath17-dimensional lorentzian manifold @xmath18 satisfying ( [ eq.1 ] ) and an embedding @xmath19 such that @xmath9 is the restriction of @xmath20 to @xmath8 and @xmath10 is the second fundamental form .
the einstein vacuum equations are invariant under diffeomorphism . in the work of y. choquet - bruhat , @xcite followed by @xcite ,
this allows her to choose a special _ harmonic gauge _ to prove the existence and uniqueness up to diffeomorphism of a maximal globally hyperbolic smooth space - time arising from any set of smooth initial data .
the harmonic gauge is also referred as _ wave coordinates _ , in which the einstein vacuum equations become a system of quasilinear wave equations on the components of the unknown metric @xmath21 : @xmath22 where @xmath23 is the geometric wave operator of @xmath20 and @xmath24 depends quadratically on @xmath25 and analytically on @xmath26 for @xmath26 small . by using christoffel symbols , @xmath27 , which reduces to @xmath28 in wave coordinates .
this is because wave coordinates are required to be solutions to the equations @xmath29 for @xmath30 .
the metric @xmath31 relative to wave coordinates @xmath32 satisfies the _ harmonic gauge condition _ : @xmath33 for @xmath30 , where @xmath34 under this condition the stability of minkowski space - time can reformulate as follows : given a pair of small symmetric two tensors @xmath35 on @xmath36 such that @xmath37 for @xmath30 , we want to find a lorentian metric @xmath38 of signature @xmath39 satisfying the reduced einstein equations ( [ eq.2 ] ) and such that @xmath40 .
[ thm.lr ] given asymptotically flat cauchy data @xmath35 in coordinates @xmath41 satisfying the harmonic gauge condition ( [ harmonic.1 ] ) and small , then the solution to the reduced einstein equations ( [ eq.2 ] ) provides a solution to the einstein vacuum equations ( [ eq.1 ] ) .
the reduced einstein vacuum equations satisfies the _ null condition _ when the spacial dimension @xmath0 , which ensures the global existence theorem for small cauchy data . for @xmath42
, it can be shown that the einstein vacuum equations in harmonic gauge do not satisfy the null condition . moreover , y. choquet - bruhat showed in @xcite that even without imposing a specific gauge the einstein equations violate the null condition .
however , the reduced einstein vacuum equations ( [ eq.2 ] ) satisfy the _ weak null condition_. in the work of @xcite and @xcite , h. lindblad and i. rodnianski reproved the global existence for einstein vacuum equations in harmonic gauge for general asymptotic flat initial data by combining this condition with the vector field method . in this paper
, we apply the radiation field theory due to f. g. friedlander to study the asymptotic behavior of solutions to einstein vacuum equations in harmonic gauge .
friedlander s radiation field was used by l. hrmander to study the asymptotic behavior of solutions to linear hyperbolic equation in the following coordinates : @xmath43 for @xmath44 large .
for instance , consider the cauchy problem in minkowski space - time as follows : @xmath45 where @xmath46 near @xmath47 , by writing @xmath48 , @xmath49 satisfies the following equation @xmath50 where @xmath51 . f. g. friedlander showed that @xmath49 is smooth up to @xmath47 .
then the radiation field is the image of the map @xmath52 which is an isometric isomorphism .
here @xmath53 is the free space translation representation of lax and phllips .
i generalize this idea by considering the conformal transformation of the reduced einstein vacuum equations on a suitable compactification of @xmath5 as follows .
$ ] , width=240 ] here @xmath54 is the compactification of null infinity and @xmath55 is cauchy surface . let @xmath56 be the total boundary defining function and @xmath57 be the defining functions for corresponding boundary hypersurfaces @xmath58 . with a conformal change @xmath59 and @xmath60 , the reduced einstein equations ( [ eq.2 ] ) are equivalent to @xmath61 where @xmath62 and @xmath63 .
see section 5 for more details .
denote by @xmath64 the space of elements in the weighted b - sobolev space @xmath65 with norm less then @xmath66 and satisfying the harmonic gauge conditions ( [ harmonic.1 ] ) , for some @xmath67 , @xmath68 and @xmath69 small .
then by energy estimate method we can show that for @xmath0 , if @xmath70 then @xmath60 is @xmath71 up to @xmath54 and hence the radiation field for @xmath72 is well defined which satisfies the corresponding harmonic gauge conditions : @xmath73 for @xmath30 , where @xmath74 . combining this with the linear radiation field theory and the implicit function theorem ,
we further show that @xmath75 also lies in some weighted b - sobolev space on @xmath54 .
more explicitly , denote by @xmath76 the elements in @xmath77\ ] ] with norm less than @xmath78 and satisfying ( [ harmonic.2 ] ) .
here we take the equivalence @xmath79 .
then @xmath80 for some @xmath81 . here
both @xmath64 and @xmath76 are banach manifolds .
see section [ sec.geosetting ] for the definition of weighted b - sobolev spaces and section [ sec.initialdata ] , [ sec.mollerop ] for more details about @xmath64 and @xmath76 .
the main theorem of this paper is the following : [ mainthm ] for @xmath0 , @xmath67 , @xmath68 and @xmath69 small , if @xmath82 , then the einstein equation ( [ eq.1 ] ) has a global solution @xmath72 such that @xmath83 is @xmath71 up to @xmath54 .
hence the radiation field is well defined and the nonlinear mller wave operator defines a continuous and open map @xmath84 for some constant @xmath85 .
the open property guarantees the solution for characteristic initial value problem for einstein vacuum equations . for @xmath0 ,
given characteristic data @xmath86 for some @xmath87 and @xmath69 small enough , there exists an asymptotically flat solution @xmath20 to the einstein vacuum equation ( [ eq.1 ] ) such that @xmath88 .
moreover , @xmath41 are wave coordinates w.r.t . @xmath20 globally . for @xmath42 ,
the cauchy data of interest for reduced einstein vacuum equations has an asymptotic leading term @xmath89 for some constant @xmath90 small , which has a long range effect at null infinity .
since @xmath89 provides a solution to the linearization of ( [ eq.2 ] ) in a neighborhood of null infinity in @xmath91 , we may expect that the essential change of the geometry of perturbed solution @xmath38 only comes from the asymptotic leading term , i.e. the constant @xmath92 . by a different change of coordinates to @xmath93 , which was suggested by friedlander to study the linear equation @xmath94 with such background metric
, i find a corresponding compactification of @xmath5 such that the radiation filed is well defined in a similar way .
this will be discussed in details in a separate paper .
the outline of this paper is as follows : in section [ sec.geosetting ] we define a compactification of @xmath5 and give basic geometric notations used through out the paper ; in section 3 , we study the solution spaces of constraint equations ( [ constraint.1 ] ) and of harmonic gauge conditions ( [ harmonic.1 ] ) separately ; in section 4 , we make a conformal transformation of the reduced einstein equations and study the metric and wave operator in the conformal setting ; in section 5 , we define a family of time - like functions , which will be used to define a special type of energy norm , called the weighted b - sobolev norm ; in section 6 , we do energy estimates for the equations ( [ eq.conf ] ) , show the regularity of einstein solutions near boundary and define the radiation field ; in section 7 , we modify the result obtained in section 6 and obtain the isomorphism property for the nonlinear mller wave operator . * acknowledgment * : i want to thank to my adviser richard melrose for inspirations and valuable discussion on this topic . without his help
, this paper would not exist .
i also would like to apologize to all who expressed interest in this work for the long delay between the first announcement of these results and the appearance of this monograph .
this section is a preparation for describing the problem and dealing with it in a manifold with corners arising from a suitable compactification of @xmath95 . to simplify the notation ,
we take @xmath96 and use the lower case english alphabet @xmath97 as indices taking value in @xmath98 and the greek alphabet @xmath99 as indices taking value in @xmath100 .
moreover , we use the capital english alphabet @xmath101 as multi indices . the change between superscript and subscript is taken over minkowski metric @xmath102 .
for example , @xmath103 .
the usual radial compactification of @xmath5 , denoted by @xmath104 , can be realized by embedding @xmath105 and taking @xmath106 , which is a closed half sphere of dimension @xmath17 .
then all null rays w.r.t .
minkowski metric converge to an embedded submanifold of @xmath104 on the bounday : @xmath107 the compactification space @xmath91 , which we use in this paper to study the asymptotic behavior of einstein vacuum solutions that are close to minkowski metric , is @xmath104 blown up @xmath108 . hence @xmath91 is a manifold with corners up to codimension @xmath109
. see figure [ fig.1 ] .
more explicitly , let @xmath110 be the distance function on @xmath104 w.r.t .
standard spherical metric and @xmath111 .
choose some @xmath69 small and define @xmath112 as follows : if @xmath113 then @xmath114 ; if @xmath115 then @xmath116 where @xmath117\longrightarrow [ \epsilon/2,\epsilon]$ ] is smooth and satisfies : @xmath118 , @xmath119 and @xmath120 for @xmath121 . then the blowup can be realized by taking @xmath122 obviously , @xmath123 embeds @xmath95 into @xmath124 with image @xmath125 , the interior of @xmath91 .
this compactification is closely related to penrose s diagram .
the three essentially different boundary hypersurfaces can be represented with the three regions at infinity : spatial infinity @xmath126 , null infinity @xmath54 and temporal infinity @xmath127 where the latter two both have forward and backward components .
in particular , we can identify @xmath128 with @xmath129 , where @xmath130 is the radial compactification of @xmath131 , and identify @xmath127 with closed ball of dimension @xmath7 .
denote the set of boundary hyper surfaces of @xmath91 by @xmath132 [ def.definingfun ] for @xmath133 , we say @xmath134 a defining function for @xmath135 iff @xmath136 two boundary defining functions @xmath137 are equivalent on a domain @xmath138 iff @xmath139 we say @xmath56 a total boundary defining function for @xmath91 iff @xmath140 throughout this paper , we use @xmath57 to denote the defining functions for @xmath141 and @xmath142 the total boundary defining function for @xmath143 .
we make a special choice of them in local coordinates in the following .
, width=240 ] choose a covering @xmath144 for @xmath91 such that @xmath145 is a bounded domain in the interior of @xmath91 and @xmath146 covers @xmath143 .
the corresponding local coordinates can be chosen from the following functions : @xmath147 more explicitly , * in the domain @xmath148 choose @xmath149 * in the domain @xmath150 , choose @xmath151 * in the domain @xmath152 , choose @xmath153 * in the domain @xmath154 for some @xmath155 , choose @xmath156 * in the domain @xmath157 , choose @xmath158 * in the domain @xmath159 , choose @xmath160 notice that , in the intersection of any two domains , the different choices of defining functions are equivalent . in many cases , for example when doing energy estimates , equivalent defining functions give equivalent statements .
so we change the choice of @xmath161 freely from one set of coordiantes to the other in the rest of this paper if they give equivalent results . however ,
when performing energy estimates , we may not been able to restrict the discussion in a single coordinate chart and hence have to consider some of them together .
see section 6 for more details .
we also denote by @xmath162 ^ 2 $ ] a double space of @xmath91 across the null infinity @xmath163 such that @xmath162 ^ 2 $ ] is a manifold with boundary and the smooth structure on @xmath91 extends to @xmath162 ^ 2 $ ] , i.e. @xmath164 ^ 2)|_{x}$ ] . here
^ 2 $ ] is not uniquely determined .
refer to @xcite for more details .
the boundary of @xmath162 ^ 2 $ ] consists of three components , @xmath165 ^ 2 $ ] , @xmath166 ^ 2 $ ] and @xmath167 ^ 2 $ ] , which are also double spaces of @xmath126 , @xmath168 and @xmath169 across their boundaries and hence are close manifolds .
moreover , @xmath165 ^ 2 $ ] can be identified with @xmath170 and @xmath171 $ ] can be identified with @xmath172 .
we introduce @xmath162 ^ 2 $ ] for the reason that the conformal minkowski metric , or with small perturbation , extends to a b - type lorentz metric on @xmath162 ^ 2 $ ] .
see section [ sec.conf ] for more details .
let @xmath173 ( resp .
@xmath174 ^ 2)$ ] ) be the space of smooth vector fields on @xmath91 which are tangent to @xmath143 ( resp .
@xmath175 ^ 2 $ ] ) .
the relation between @xmath174 ^ 2)|_x$ ] and @xmath173 is @xmath176 ^ 2)|_x = \mathscr{v}_b(x ) + c^{\infty}(x)v_1.\ ] ] where @xmath177 ^ 2)|_{x}$ ] transverse to @xmath54 , i.e. for any @xmath178 , @xmath179 . in particular , in domain @xmath180
, we can choose @xmath181 .
set @xmath182 here @xmath183 are the projection of @xmath184 on @xmath185 satisfying @xmath186 .
we denote by @xmath187 any vector field in @xmath188 and @xmath189 any element in @xmath190 for @xmath191 .
[ lem.0 ] the vector fields defined above satisfy the following properties : @xmath192 the first property follows directly by @xmath193 and @xmath194 . to prove ( ii ) and ( iii ) , we only need to check them in local coordinates near @xmath143 : * in @xmath195 with coordinates @xmath196 , @xmath197 .
hence ( ii ) and ( iii ) follow by @xmath198 * in @xmath199 with coordinates @xmath200 , @xmath201 .
hence ( ii ) and ( iii ) follow by @xmath202 * in @xmath203 with coordinates @xmath204 , @xmath205 .
hence ( ii ) and ( iii)follow by @xmath206 * in @xmath207 with coordinates @xmath208 , @xmath209 .
hence ( ii ) and ( iii ) follow by @xmath210 * in @xmath211 with coordinates @xmath212 , @xmath213 .
hence ( ii ) and ( iii ) follow by @xmath214 we finish the proof .
we define the b - type sobolev space and conormal symbol space on @xmath91 in the following .
these spaces are used to characterize the regularity and asymptotic properties near boundary for einstein vacuum solutions .
let @xmath215 be the riemannian metric on @xmath91 induced from euclidean metric on @xmath124 via ( [ eq.blowup ] ) and denote by @xmath216 the corresponding volume form .
[ def.1 ] define the b - sobolev space @xmath217 for any @xmath218 the completion of @xmath219 w.r.t .
the norm : @xmath220 define the conormal symbol space by @xmath221 for @xmath222 .
the relationship between the b - sobolev spaces and the conormal symbol spaces are stated in the following lemma .
[ lem.2.3 ] if @xmath223 , then @xmath224 moreover , there exists a constant @xmath225 for any @xmath226 such that @xmath227 which implies that @xmath228 definition [ def.1 ] and lemma [ lem.2.3 ] can be generalize to any manifold with corners , including closed @xmath229-submanifolds of @xmath91 with induced riemannian metric and boundary defining functions . see @xcite for definition of @xmath229-submainifold . in this paper
, we mainly consider the cauchy surface @xmath230 and null infinity @xmath231 , which are both @xmath229-submanifolds of @xmath91 .
we study the map from cauchy data to characteristic data for einstein vacuum equations , which lie in some weighted b - sobolev spaces or cornormal symbol space on @xmath8 and @xmath231 correspondingly .
in this section , we study the spaces of small asymptotically flat initial data for einstein vacuum equations which satisfy either the constraint equations ( [ constraint.1 ] ) or the harmonic gauge conditions ( [ harmonic.1 ] ) , and the relation between them . denote by @xmath38 a small perturbation of minkowski metric : @xmath232 .
we can view @xmath233 as functions of @xmath41 valued in @xmath234 symmetric matrices .
then ( [ constraint.1 ] ) and ( [ harmonic.1 ] ) can be written in terms of the components of @xmath235 , which turn out to be undetermined elliptic systems after suitable decomposition .
let us first derive the constraint equations ( [ constraint.1 ] ) in terms of components of @xmath35 .
denote by @xmath236 then the components of ricci curvature of @xmath20 can be expressed as @xmath237 where @xmath238 denote by @xmath239 then @xmath240 are quadratic forms of @xmath241 with coefficients analytically depending on @xmath242 .
the einstein vacuum equations ( [ eq.1 ] ) imply that @xmath243 for @xmath244 , which cancel the @xmath245 terms and impose @xmath246 constraint equations on @xmath247 at @xmath248 : @xmath249 for @xmath244 . here
( [ constraint.2 ] ) is the system of constraint equations ( [ constraint.1 ] ) in fixed coordinates .
we are interested in small solutions to ( [ constraint.2 ] ) in the following space : @xmath250 let @xmath251 [ prop.2 ] for @xmath6 , @xmath252 and @xmath69 small enough , @xmath253 is a frchet manifolds .
we prove the statement by studying the linearization of ( [ constraint.2 ] ) at @xmath254 : @xmath255 where @xmath256 , and applying implicit function theorem .
notice that by @xcite , for @xmath6 and @xmath257 , @xmath258 is an isomorphism and hence @xmath259 are both isomorphism
. see proposition [ app.1 ] in appendix .
assume @xmath35 is in space ( [ initialspace ] ) and take a linear transformation of as follows : for @xmath260 , @xmath261 , @xmath262 , @xmath263 then the first equation in ( [ constraint.3 ] ) becomes @xmath264 denote by @xmath265 and @xmath266 , @xmath267-forms on the euclidean space , i.e. @xmath268 then the second equation in ( [ constraint.3 ] ) can be expressed as @xmath269 since @xmath270 has trivial kernel on @xmath271 for all @xmath272 , the hodge decomposition theorem implies @xmath273 with @xmath274 and @xmath275 the subspace of closed and coclosed forms respectively .
more explicitly , let @xmath276 then @xmath277 and @xmath278 is a projection .
writing @xmath279 and @xmath280 with @xmath281 closed and @xmath282 coclosed , i.e. @xmath283 then ( [ eq.3 ] ) split into @xmath284 for the first equation in ( [ eq.5 ] ) , notice that we have the following commutative diagram : @xmath285 ^{d,\delta } \ar[d]_{\triangle^{-\frac{1}{2 } } } & \rho^{\lambda+2}h_b^{n-1}(\overline{\mathbb{r}^n}:\lambda(\mathbb{r}^n ) ) \ar[d]^{\triangle^{-\frac{1}{2 } } } \\
\rho^{\lambda}h_b^{n+1}(\overline{\mathbb{r}^n}:\lambda(\mathbb{r}^n ) ) \ar[r]_{d,\delta } & \rho^{\lambda+1}h_b^{n}(\overline{\mathbb{r}^n}:\lambda(\mathbb{r}^n ) ) } \ ] ] where @xmath286 .
since @xmath287 , @xmath288 hence @xmath289 is equivalent to @xmath290 substitute @xmath291 by @xmath292 , ( [ constraint.3 ] ) is equivalent to the elliptic system : @xmath293 in ( [ constraint.3 ] ) , the components @xmath294 are free . to solve @xmath295 from ( [ constraint.3 ] ) , it is equivalent to solve @xmath296 from ( [ constraint.4 ] )
. given arbitray @xmath297 and @xmath298 for @xmath299 and @xmath300 , we can solve ( [ constraint.4 ] ) uniquely and have @xmath301 and hence the solution space of ( [ constraint.3 ] ) , denoted by @xmath302 , is isomorphism to @xmath303 due to the ellipticity of ( [ constraint.4 ] ) and invertibility of @xmath270 and @xmath304 , when @xmath69 small enough , @xmath305 is a submanifold in space ( [ initialspace ] ) , which has tangent space @xmath302 at @xmath254 .
we first deriving the equations ( [ harmonic.1 ] ) in terms of components of @xmath35 .
first @xmath306 is equivalent to @xmath307 for @xmath244 .
secondly @xmath308 is equivalent to @xmath309 for @xmath244 .
when assuming @xmath310 , the reduced einstein vacuum equations ( [ eq.2 ] ) at @xmath248 is @xmath311 multiplying ( [ eq.7 ] ) by @xmath312 and substituting @xmath313 terms by the right hand side of ( [ eq.0 ] ) , we have @xmath314 for @xmath244 . [ lem.1 ] the harmonic gauge conditions ( [ constraint.6 ] ) and ( [ constraint.7 ] ) are equivalent to the constraint conditions ( [ constraint.2 ] ) combined with ( [ constraint.6 ] ) in the fixed coordinates , i.e. ( [ constraint.6 ] ) + ( [ constraint.7 ] ) = ( [ constraint.2 ] ) + ( [ constraint.6 ] ) .
we only need to show under the condition ( [ constraint.6 ] ) , ( [ constraint.7 ] ) is equivalent to ( [ constraint.2 ] ) .
first ( [ constraint.6 ] ) gives @xmath315 for @xmath244 .
plugging them into the first equation of ( [ constraint.7 ] ) , we have @xmath316 where @xmath317 is a quadratic form of @xmath318 , @xmath319 which gives the first equation of ( [ constraint.2 ] ) .
rewrite the second equation of ( [ constraint.7 ] ) as follows : @xmath320 plugging ( [ eq.00 ] ) into the first three terms in above equation , we get @xmath321 where @xmath322 is a quadratic form of @xmath318 , @xmath323 which gives the second equation in ( [ constraint.2 ] ) . since the substituting action is reversible , we finish proving the equivalence of ( [ constraint.6])+([constraint.7 ] ) and ( [ constraint.6])+([constraint.2 ] ) .
notice that ( [ constraint.6 ] ) is an elliptic system of equations of @xmath324 for @xmath30 , which are free terms when solving the linearization of ( [ constraint.2 ] ) .
hence by a similar decomposition as in the proof of proposition [ prop.2 ] , the linearization of ( [ constraint.6])+([constraint.7 ] ) form an elliptic system w.r.t .
@xmath325 when considering the solutions in space ( [ initialspace ] ) for @xmath6 and @xmath252 .
let @xmath326 then by the same proof of proposition [ prop.2 ] , we have [ prop.3 ] for @xmath6 , @xmath252 and @xmath69 small enough , @xmath327 is a frchet manifolds .
here @xmath327 has tangent space @xmath328 at @xmath254 which is the space of solutions to linearization of ( [ constraint.6])+([constraint.7 ] ) and is isomorphic to @xmath329 suppose @xmath330 is an einstein vacuum solution which is close to the minkowski space - time with cauchy data @xmath331 satisfying the constraint equations ( [ constraint.2 ] ) . here
@xmath332 in the sense of matrices .
let @xmath333 be a diffeomorphism preserving the cauchy surface : @xmath334 we want to find out all @xmath335 such that @xmath41 are wave coordinates w.r.t .
the pull back metric @xmath336 : @xmath337 denote by @xmath35 the corresponding cauchy data in coordinates ( t , x ) : @xmath338|_{t=0 } \\ & \ + [ \bar{g}_{\mu\beta}\partial_{\nu}\partial_tf^{\beta}+ \bar{g}_{\alpha\nu}\partial_{\mu}\partial_tf^{\alpha}+ \bar{g}_{\alpha\beta}(\partial_t\partial_{\mu}f^{\alpha}\partial_{\nu}f^{\beta}+ \partial_{\mu}f^{\alpha}\partial_t\partial_{\nu}f^{\beta})]_{t=0}. \end{aligned}\right.\ ] ] for simplicity , denote @xmath339 .
it shows that @xmath247 is determined by @xmath340 . to make @xmath20 satisfy the harmonic gauge condition ( [ harmonic.3 ] ) , it is equivalent to require @xmath341 satisfying the following equations : @xmath342 for @xmath343 , where @xmath344 are analytic functions of @xmath345 for @xmath346 small : @xmath347 notice that @xmath348 . in particular , @xmath349 here ( [ eq.8 ] ) is a system of quasilinear wave equations for @xmath341 .
given any @xmath350 and @xmath351 with enough regularity , we can solve ( [ eq.8 ] ) in a neighborhood of cauchy surface .
furthermore , if @xmath352 are also small enough , @xmath341 is determined globally .
since we assume @xmath353 , @xmath352 have @xmath354 degree of freedom .
however , we want to choose the gauge before we find out the global einstein solution @xmath355 for the cauchy problem . by theorem [ thm.lr ] , this can be done by requiring @xmath35 in the new coordinates @xmath41 to satisfy the harmonic gauge conditions ( [ harmonic.1 ] ) , or equivalently , ( [ constraint.6])+([constraint.7 ] ) .
first ( [ constraint.6 ] ) give @xmath246 linear equations of @xmath356 and allow us to write @xmath357 in terms of @xmath352 .
more explicitly @xmath358 where @xmath359 are analytic functions of @xmath360 . then
( [ constraint.7 ] ) is satisfied automatically by lemma [ lem.1 ] since the constraint equations ( [ constraint.1 ] ) , or equivalently ( [ constraint.2 ] ) , is invariant under deffeomorphism .
this shows again that the global diffeomorphism and the global harmonic coordinates set are determined by @xmath352 . from ( [ eq.10 ] )
, the set of @xmath361 gives a group action on the space of @xmath35 which satisfies the harmonic gauge conditions ( [ constraint.6])+([constraint.7 ] ) .
we investigate this group action in more details in the following . for @xmath362 ,
let @xmath363 be the nonabelian group generated by @xmath364 for some @xmath81 small with group operation @xmath365 and identity @xmath254 . here
@xmath366 denotes the composition of functions .
it is clearly that @xmath363 does not depend on @xmath81 small .
then @xmath363 has two subgroups generated by @xmath367 and @xmath368 respectively , which transversely intersect at @xmath254 .
denote @xmath369 [ lem.13 ] with above notation , @xmath370 denote by @xmath371 where @xmath372 is given in proposition [ prop.3 ] . by lemma [ lem.13 ] , we can study the action of @xmath373 and @xmath374 on @xmath327 separately .
first , @xmath373 forms an nonabelian group with operation @xmath375 by ( [ eq.10 ] ) , given @xmath376 and @xmath377 , then @xmath378 is defined as follow : @xmath379 here @xmath380 is defined in ( [ eq.50 ] ) .
obviously @xmath373 preserves the coordinates on the cauchy surface and hence preserves the restricting metric and second fundamental form on cauchy surface .
given initial data @xmath381 , let @xmath9 be the induced metric on cauchy surface and @xmath10 the corresponding second fundamental form .
this defines the map @xmath382 : @xmath383_{ij}= \delta_{ij}+h^0_{ij } , \quad [ k_0]_{ij}=-(1-h^0_{00}+g_0^{kl}h^0_{0l}h^0_{0k})^{\frac{1}{2 } } ( \partial_ih^0_{0j}+\partial_jh^0_{0i}-h^1_{ij } ) .
\end{aligned}\ ] ] in our setting , @xmath384_{ij}-\delta_{ij}\in \rho_0^{\lambda}h_b^{n+1}(\overline{\mathbb{r}^n}),\quad [ k_0]_{ij}\in \rho_0^{\lambda+1}h_b^{n}(\overline{\mathbb{r}^n}).\ ] ] let @xmath385 .
obviously , @xmath386 & \mathcal{g}_1^{n,\lambda}\cdot \widetilde{\mathcal{v}}_{\epsilon}^{n,\lambda } \ar[d]^{\pi } \\ & \widetilde{c}^{n,\lambda}_{\epsilon}}\ ] ] provides a fibration structure .
secondly , @xmath374 is obvious equivalent to the diffeomorphism group on cauchy surface with the corresponding regularity and decay at infinity .
the group operation is as follows : @xmath387 the action of @xmath374 on @xmath388 is just a pull - back .
denote by @xmath389 .
then for any @xmath390 , we have the commutativity @xmath391 which implies that @xmath386 & \widetilde{\mathcal{v}}^{n,\lambda } \ar[d]^{\pi } \\ & \widetilde{c}^{n,\lambda}}\ ] ] is also a fibration .
let @xmath392 and @xmath393 the quotient map .
denote by @xmath394 be the euclidean metric and @xmath395 be the symmetric 2-tensor bundle on @xmath396 .
summarizing above discussion , we have for @xmath6 and @xmath257 , the quotient space @xmath397 is an open neighborhood of @xmath398 in the solution space of constraint equation ( [ constraint.1 ] ) in @xmath399 with the fibration structure : @xmath400 & \widetilde{\mathcal{v}}^{n,\lambda } \ar[d]^{\pi } \\
\mathcal{g}_0^{n,\lambda}\ar@{-}[r ] & \widetilde{c}^{n,\lambda } \ar[d]^{\pi'}\\ & \widetilde{d}^{n,\lambda } . } \ ] ]
to study the asymptotic behavior of einstein vacuum solutions which are close to minkowski space - time , we will work on their conformal transformation on @xmath91 , the compactification of @xmath95 defined in section [ sec.geosetting ] , and perform energy estimates for a conformal transformation of the reduced einstein vacuum equation ( [ eq.2 ] ) with cauchy data satisfying the harmonic gauge conditions ( [ constraint.6])+([constraint.7 ] ) .
we clarify the conformal transformation of metrics and equations in this section by describing them in the local coordinates specified in section [ sec.geosetting ] .
let @xmath401 be the minkowski metric and @xmath38 a small perturbation of it : @xmath402 .
in the fixed coordinates @xmath403 , it is equivalent to say that @xmath72 is a function on @xmath95 valued in @xmath234 symmetric matrix space @xmath404 . in this sense , denote the inverse of @xmath20 by @xmath405 for any @xmath406 , denote by @xmath407 a real analytic function of @xmath72 with coefficients in @xmath408 such that @xmath409 for any @xmath410 , denote by @xmath411 an @xmath412-form in @xmath413 with coefficients @xmath407 , and @xmath414 an @xmath415-form in @xmath416 with coefficients @xmath411 .
for any @xmath417 @xmath418 with above notation @xmath419 for the purpose of studying the geometry of @xmath20 near null infinity of @xmath420 , we also denote by @xmath421 with this notations , the harmonic gauge condition says that some components decay faster than the others near null infinity .
[ lem.4 ] suppose @xmath41 are harmonic coordinates w.r.t .
we define a vector field @xmath422 for each @xmath423 as follows : @xmath424 then in each domain @xmath425 for @xmath423 , with the choice of @xmath56 and @xmath426 specified in section [ sec.geosetting ] , @xmath427 satisfies @xmath428=[d,\theta]=0,\ ] ] and the harmonic gauge condition implies @xmath429 for @xmath30 . here
@xmath430 .
we only need to prove the statement in @xmath180 . by writing out the harmonic gauge condition in local coordinates
, we have * in @xmath199 , @xmath431 , @xmath432 , @xmath433 , @xmath434 , @xmath435 * in @xmath203 , @xmath436 , @xmath437 , @xmath438 , @xmath439 * in @xmath207 , @xmath440 , @xmath441 , @xmath442 , @xmath443 , @xmath444 here @xmath445 .
we finish the proof .
recall that @xmath56 denotes a total boundary defining function of @xmath91 and @xmath161 denotes the defining function for boundary hypersurfaces @xmath446 for @xmath447 .
set @xmath448 similarly , denote @xmath449 under the assumption and notation in lemma [ lem.4 ] , for @xmath30 , @xmath450 [ prop.4 ] if @xmath451 such that for some @xmath452 and all @xmath453 @xmath454 for some @xmath69 small , then @xmath455 extends to a @xmath456 lorentzian b - metric of signature @xmath39 on @xmath457 with @xmath458 being its characteristic surfaces .
the statement is independent of choice of boundary defining functions .
hence we only need to write the metric @xmath459 in local coordinates near @xmath143 with particular choice of @xmath56 and @xmath161 as stated in section [ sec.geosetting ] . * in @xmath195 , @xmath460 , @xmath436 , @xmath461 , @xmath462 * in @xmath199 , @xmath431 , @xmath432 , @xmath433 , @xmath434 , @xmath463 * in @xmath203 , @xmath436 , @xmath437 , @xmath438 , @xmath464 * in @xmath207 , @xmath440 , @xmath441 , @xmath442 , @xmath443 , @xmath465 * in @xmath211 , @xmath466 , @xmath467 , @xmath468 , @xmath469 here in @xmath425 for @xmath470 we use the polar coordinates .
notice here @xmath471 is restricted on @xmath472 , i.e. @xmath473 and @xmath474 .
then the statement is obviously true if @xmath475 . for @xmath476 ,
let @xmath477 and change the coordinates near @xmath478 as follows : * in @xmath199 , set @xmath479 ; * in @xmath203 , set @xmath480 ; * in @xmath207 , set @xmath481
. in the new coordinates the metric components is uniformly bounded and @xmath482 is its characteristic surface .
since the coordinates changing preserves the boundary hypersurface @xmath231 , we prove the statement .
notice that the coordinates change extends to a @xmath71 homeomorphism of @xmath483 .
denote by @xmath484 and @xmath485 the laplace - beltrami operators w.r.t . @xmath20 and @xmath59 .
the relation between @xmath484 and @xmath485 is stated as follows . for any @xmath486 , @xmath487
the formula can be proved by direct computation as friedlander did in @xcite .
@xmath488 [ lem.2 ] suppose @xmath41 are harmonic coordinates w.r.t .
@xmath20 . then in each domain @xmath425 for @xmath489 , with the choice of @xmath490 specified in section [ sec.geosetting ] ,
we have @xmath491,\quad \\ \box_{\tilde{m}}=\ 2\partial_{\rho_1}d+ \tilde{\partial}^2+\tilde{\partial } , \quad \gamma= \gamma_0+\tilde{\rho}^{\frac{n-1}{2}}\theta_1(\tilde{h } ) , \end{gathered}\ ] ] where @xmath422 is defined in lemma [ lem.4 ] and @xmath492 . here
@xmath493 is a constant in each domain .
if @xmath41 are harmonic coordinates w.r.t .
@xmath20 , then @xmath494 , which implies that @xmath495 we write @xmath496 , @xmath497 , @xmath498 and @xmath499 in local coordinates in each domain @xmath425 for @xmath423 . * in @xmath195 , @xmath460 , @xmath436 , @xmath461 , @xmath500\partial^2_s + [ h^{ij}\theta_i\theta_j](\rho\partial_{\rho})^2 + h^{ij}\slashpar_i\slashpar_j \\ & + 2[-h^{0i}\theta_i+sh^{ij}\theta_i\theta_j]\partial_s(\rho\partial_{\rho } ) + 2[h^{0i}-sh^{ij}\theta_j]\partial_s\slashpar_i \\ & + 2[-h^{ij}\theta_j](\rho\partial_{\rho})\slashpar_i + [ -(n+1)h^{0i}\theta_i+sh^{ij}(-\delta_{ij}+(n+2)\theta_i\theta_j)]\partial_s \\ & + [ h^{ij}(-\delta_{ij}+(n+1)\theta_i\theta_j)]\rho\partial_{\rho } + [ -nh^{ij}\theta_j]\slashpar_i \\ = & \ \box_{\tilde{m}}+\rho^{\frac{n-1}{2}}\theta_1(\tilde{h } ) ( \tilde{\partial}^2+\tilde{\partial } ) , \\ \gamma = & \ \gamma_0+[h^{ij}(-\tfrac{n-1}{2}\delta_{ij}+\tfrac{(n-1)(n+3)}{4}\theta_i\theta_j ) ] , \quad \gamma_0=-\tfrac{(n-1)(n-3)}{4}. \end{aligned}\ ] ] * in @xmath199 , @xmath431 , @xmath432 , @xmath433 , @xmath501(b\partial_{b}-a\partial_{a})^2 + [ h^{ij}\theta_i\theta_j](a\partial_{a})^2 \\ & + 2a^{-1}[-h^{0i}\theta_i+h^{ij}\theta_i\theta_j ] ( a\partial_{a})(b\partial_{b}-a\partial_{a})+h^{ij}\slashpar_i\slashpar_j \\ & + 2a^{-1}[h^{0i}-h^{ij}\theta_j]\slashpar_i(b\partial_{b}-a\partial_{a } ) + 2[-h^{ij}\theta_j]\slashpar_i(a\partial_{a } ) \\ & + a^{-2}[h^{00}-2h^{i0}\theta_i+h^{ij}\theta_{i}\theta_{j } + ah^{ij}(-\delta_{ij}+n\theta_i\theta_j ) -(n-1)ah^{0i}\theta_i](b\partial_{b}-a\partial_{a } ) \\ & + [ h^{ij}(-\delta_{ij}+(n+1))\theta_i\theta_j]a\partial_{a } + [ -nh^{ij}\theta_j]\slashpar_i , \\ = & \ \box_{\tilde{m}}-a^{\frac{n-5}{2}}b^{\frac{n-1}{2}}\tilde{h}_{\rho\rho } [ ( b\partial_{b}-a\partial_{a})^2+(b\partial_{b}-a\partial_{a})]+ a^{\frac{n-3}{2 } } b^{\frac{n-1}{2 } } \theta_1(\tilde{h})(\tilde{\partial}^2+\tilde{\partial } ) , \\ \gamma = &
\ \gamma_0 + [ h^{ij}(-\tfrac{n-1}{2}\delta_{ij}+\tfrac{(n-1)(n+3)}{4}\theta_i\theta_j ) ] , \quad \gamma_0=-\tfrac{(n-1)(n-3)}{4}. \end{aligned}\ ] ] * in @xmath203 , @xmath436 , @xmath437 , @xmath461 , @xmath502\partial^2_{\tau } + [ h^{ij}\theta_i\theta_j](\rho\partial_{\rho})^2 + h^{ij}\slashpar_i\slashpar_j \\ & + 2[-h^{ij}\theta_j](\rho\partial_{\rho})\slashpar_i + 2\rho^{-1}[-h^{i0}\theta_i+h^{ij}\theta_i\theta_j]\partial_{\tau}(\rho\partial_{\rho } ) \\ & + 2\rho^{-1}[h^{i0}-h^{ij}\theta_{j}]\partial_{\tau}\slashpar_i + [ h^{ij}(-\delta_{ij}+(n+1)\theta_i\theta_j)]\rho\partial_{\rho } \\ & + \rho^{-1}[-(n-1)h^{0i}\theta_i+h^{ij}(-\delta_{ij}+n\theta_i\theta_j)]\partial_{\tau } + n[-h^{ij}\theta_j]\slashpar_i \\ = & \ \box_{\tilde{m}}-\rho^{\frac{n-5}{2}}\tilde{h}_{\rho\rho}\partial^2_{\tau } + \rho^{\frac{n-3}{2}}\theta_1(\tilde{h})(\tilde{\partial}^2+\tilde{\partial } ) , \\
\gamma = & \ \gamma_0+[h^{ij}(-\tfrac{n-1}{2}\delta_{ij}+\tfrac{(n-1)(n+3)}{4})\theta_i\theta_j ] ,
\quad \gamma_0=-\tfrac{(n-1)(n-3)}{4}. \end{aligned}\ ] ] * in @xmath207 , @xmath503 , @xmath441 , @xmath442 , @xmath504 ( \bar{b}\partial_{\bar{b}}-\bar{a}\partial_{\bar{a}})^2 + h^{00}(\bar{a}\partial_{\bar{a}})^2 \\ & + ( 1-\bar{a})^{-2}h^{ij}\slashpar_i\slashpar_j + 2\bar{a}^{-1}[h^{00}-h^{0i}\theta_i](\bar{a}\partial_{\bar{a } } ) ( \bar{b}\partial_{\bar{b}}-\bar{a}\partial_{\bar{a } } ) \\
& -2(1-\bar{a})^{-1}h^{0i}(\bar{a}\partial_{\bar{a}})\slashpar_i + 2(\bar{a}(1-\bar{a}))^{-1}[-h^{i0}+h^{ij}\theta_j](\bar{b}\partial_{\bar{b } } -\bar{a}\partial_{\bar{a}})\slashpar_i \\ & + ( 1-\bar{a})^{-2}[-h^{ij}\theta_j-(n-1)h^{0i}]\slashpar_i + nh^{00}\bar{a}\partial_{\bar{a } } + \bar{a}^{-2}[h^{00}-2h^{i0}\theta_i+h^{ij}\theta_i\theta_j \\ & + ( n-1)\bar{a}(h^{00}-h^{0i}\theta_i ) -\bar{a}(1-\bar{a})^{-1}h^{ij}(-\delta_{ij } + \theta_i\theta_j)](\bar{b}\partial_{\bar{b } } -\bar{a}\partial_{\bar{a } } ) \\ = & \ \box_{\tilde{m}}-\bar{a}^{\frac{n-5}{2}}\bar{b}^{\frac{n-1}{2}}\tilde{h}_{\rho\rho } [ ( \bar{b}\partial_{\bar{b } } -\bar{a}\partial_{\bar{a}})^2+(\bar{b}\partial_{\bar{b } } -\bar{a}\partial_{\bar{a}})]+ \bar{a}^{\frac{n-3}{2 } } \bar{b}^{\frac{n-1}{2 } } \theta_1(\tilde{h})(\tilde{\partial}^2+\tilde{\partial } ) , \\
\gamma=&\ \gamma_0+\tfrac{n^2 - 1}{4}h^{00},\quad \gamma_0=-\tfrac{n^2 - 1}{4}. \end{aligned}\ ] ] * in @xmath211 , @xmath466 , @xmath467 , @xmath468 , @xmath505 we finish the proof . next , we consider the commutator of @xmath496 and a basis of @xmath173 . by lemma [ lem.0 ] , we can choose a basis @xmath506 for @xmath507 for @xmath508 as follows : @xmath509 [ lem.4.7 ] suppose @xmath41 are harmonic coordinates w.r.t .
then in each domain @xmath510 for @xmath511 , with the choice of @xmath56 specified in section [ sec.geosetting ] , @xmath512=[\box_{\tilde{m}},z_{ij}]=0 , \quad [ \box_{\tilde{m}},z_{0i}]=c_i(\box_{\tilde{m}}+\gamma_0 ) + \tilde{\partial}+c'_i,\ ] ] for @xmath453 , where @xmath513 and @xmath514 .
in particular , in @xmath203 , @xmath515=0 , \quad [ \box_{\tilde{m}},\rho\partial_{\rho } ] = \box_{\tilde{m}}-(\rho\partial_{\rho})^2-\rho\partial_{\rho}-\triangle_{\theta } .\ ] ] for any @xmath516 , @xmath517 u = & \ ( \box_{\tilde{m}}+\gamma_0)z_{\mu\nu}u -z_{\mu\nu}(\box_{\tilde{m}}+\gamma_0)u \\ = & \ \tilde{\rho}^{-\frac{n+3}{2}}\box_m\tilde{\rho}^{\frac{n-1}{2}}z_{\mu\nu}u - z_{\mu\nu}\tilde{\rho}^{-\frac{n+3}{2}}\box_{m}\tilde{\rho}^{\frac{n-1}{2}}u \\ = & \ \tilde{\rho}^{-\frac{n+3}{2}}[\box_m , z_{\mu\nu}]\tilde{\rho}^{\frac{n-1}{2}}u + \tilde{\rho}^{-\frac{n+3}{2}}\box_m[\tilde{\rho}^{\frac{n-1}{2}},z_{\mu\nu}]u -[z_{\mu\nu},\tilde{\rho}^{-\frac{n+3}{2}}]\box_{m}\tilde{\rho}^{\frac{n-1}{2}}u \\ = & \
c_{\mu\nu}(\box_{\tilde{m}}+\gamma_0)u -\tfrac{n-1}{2}(\box_{\tilde{m}}+\gamma_0)\tilde{\rho}^{-1}z_{\mu\nu}(\tilde{\rho})u + \tfrac{n+3}{2}\tilde{\rho}^{-1}z_{\mu\nu}(\tilde{\rho})(\box_{\tilde{m}}+\gamma_0)u \\ = & \ ( c_{\mu\nu}+2\tilde{\rho}^{-1}z_{\mu\nu}(\tilde{\rho } ) ) ( \box_{\tilde{m}}+\gamma_0)u -\tfrac{n-1}{2}[\box_{\tilde{m } } , \tilde{\rho}^{-1}z_{\mu\nu}(\tilde{\rho})]u \end{aligned}\ ] ] in the following , we study the term @xmath518 and its commutator with @xmath496 .
recall that @xmath519=c_{\mu\nu}\box_m$ ] with @xmath520 and @xmath521 hence @xmath522=[\box_{\tilde{m}},z_{ij}]=0 $ ] .
notice that we choose @xmath461 in @xmath523 and @xmath468 in @xmath524 and @xmath525 hence @xmath526 for all @xmath453 and @xmath511 . since @xmath527\subset c^{\infty}(\overline{\omega_j}),\ ]
] it is obviously for @xmath528 or @xmath529 , @xmath530 \in \mathscr{v}_b(\omega_j)+c^{\infty}(\overline{\omega_j}).\ ] ] for @xmath531 , by lemma [ lem.2 ] , we only need to consider @xmath532 = 2\partial_{\rho_1}d ( \tilde{\rho}^{-1}z_{0i}(\tilde{\rho } ) ) + 2d ( \tilde{\rho}^{-1}z_{0i}(\tilde{\rho } ) ) \partial_{\rho_1}+ 2\partial_{\rho_1 } ( \tilde{\rho}^{-1}z_{0i}(\tilde{\rho } ) ) d\ ] ] where @xmath533 and @xmath534 . hence @xmath532
\in \mathscr{v}_b(\omega_j)+c^{\infty}(\overline{\omega_j}).\ ] ] we finish the proof . with the conformal transformation ( [ eq.11 ] ) , the reduced einstein vacuum equations ( [ eq.2 ] ) is equivalent to the following system : @xmath535 where @xmath536 is given in lemma [ lem.2 ] and @xmath537 using the explicit formula of @xmath538 given in @xcite , we have [ lem.4.8 ] for @xmath6 , @xmath539 in particular , in domains @xmath180 , @xmath540 where @xmath541 is defined in lemma [ lem.4 ] . by @xcite , @xmath542 , where @xmath543 @xmath544 satisfies the null condition and @xmath545 first notice that @xmath546\in\tilde{\rho}^{\frac{n+1}{2}}\mathscr{a}^{0,0,0}(x)$ ] .
hence by the conformal transformation ( [ eq.11 ] ) , @xmath547 in particular , in domain @xmath180 , @xmath548 we finish the proof .
suppose @xmath72 is a solution of the equation ( [ eq.12 ] ) .
then we can insert vector fields which are tangent to the boundary of @xmath91 and have the following lemma .
[ lem.6 ] let @xmath549 be any vector field in @xmath506 for @xmath511 and @xmath550 a multi - index .
then in each domain @xmath510 for @xmath511 , the equations ( [ eq.12 ] ) implies that @xmath551 where in @xmath552 , @xmath553 in @xmath554 , @xmath555 it is obviously true if @xmath556 by lemma [ lem.4.8 ] . for @xmath557 , @xmath558 \tilde{h}_{\mu\nu}+\tilde{d}^i\tilde{f}^0_{\mu\nu}(\tilde{h } ) + [ \box_{\tilde{m}},\tilde{d}^i]\tilde{h}_{\mu\nu}.\ ] ] it is obvious the first two terms on the right hand side are nonlinear and can be expressed as above . for the third term , from lemma [ lem.4.7 ] , we have @xmath559=c\box_{\tilde{m}}+\tilde{\partial}^2+\tilde{\partial}+c'\ ] ] for some @xmath560 .
then for @xmath561 @xmath562\tilde{h}_{\mu\nu } = & \ \sum_{|k|+|j|=|i|-1}c_{k , j}\tilde{d}^{k}[\box_{\tilde{m}},\tilde{d}]\tilde{d}^{j}\tilde{h}_{\mu\nu } \\ = & \ \sum_{|k|+|j|=|i|-1}c_{k , j}\tilde{d}^{k}\big(c(\box_{\tilde{g}}+\gamma)+c(\box_{\tilde{m}}-\box_{\tilde{g}}+\gamma_0-\gamma)+\tilde{\partial}^2+\tilde{\partial}+c'-c\gamma_0\big)\tilde{d}^{j}\tilde{h}_{\mu\nu } \end{aligned}\ ] ] where @xmath563 are constants . by induction on @xmath550 and lemma
[ lem.2 ] , @xmath564\tilde{h}_{\mu\nu}$ ] is equal to @xmath565 plus a nonlinear term which can be expressed as above .
we finish the proof .
time - like function is the most important concept associated to a lorentzian metric : we use it to define space - like hypersurfaces , positive quadratic forms as well as energy norms . in this section
, we consider the time - like functions w.r.t . @xmath566 and @xmath567 with @xmath83 small in the interior of @xmath91 .
suppose @xmath455 extends to a lorentz b - metric on @xmath162 ^ 2 $ ] of signature @xmath39 , @xmath568 ^ 2))$ ] and @xmath569 ^ 2)$ ] .
we say @xmath570 is _ time - like _ w.r.t .
@xmath455 at @xmath229 iff @xmath571 _ and null _ w.r.t .
@xmath455 at @xmath229 iff @xmath572 a hypersurface @xmath573 is called _ space - like _ ( resp .
_ null _ ) w.r.t .
@xmath455 iff @xmath574 has a defining function @xmath568 ^ 2))$ ] such that @xmath570 is time - like ( resp .
null ) on @xmath574 .
[ def.4 ] for a time like function @xmath570 w.r.t .
@xmath455 , define a vector field @xmath575 associate to @xmath570 and quadratic in @xmath25 by @xmath576 here @xmath16 is the covariant derivative w.r.t . @xmath455 and @xmath577 is some constant .
suppose @xmath578 are two time - like functions w.r.t .
a lorentzian metric @xmath455 at @xmath569 ^ 2)$ ] such that @xmath579 at @xmath229 , then @xmath580 is a strict positive quadratic form of @xmath581 at @xmath229 .
see chapter 6 in @xcite for proof in details .
moreover , we have the following formulas from the divergence theorem directly , which implies the energy estimates if choosing @xmath570 and @xmath582 properly .
[ lem.3 ] suppose that the domain @xmath583 has piecewise smooth boundary @xmath584 with defining function @xmath585 and @xmath570 is a @xmath586 time - like function w.r.t .
@xmath455 on @xmath587 .
assume @xmath588 .
then @xmath589 here @xmath590 is the volume form on @xmath584 such that @xmath591 and @xmath584 is oriented by @xmath592 .
notice that @xmath16 is the covariant derivative w.r.t . @xmath455 and @xmath593 by space like hypersurfaces and null hypersurfaces.,width=360 ] for the purpose of doing energy estimates
, we divide @xmath594 into four domains @xmath595 by space - like hypersurfaces and null hypersurfaces w.r.t .
@xmath596 , or equivalently w.r.t .
see figure [ fig.2 ] .
we list the time - like functions used to define those surfaces , which are also used to define the energy norms in next section , domain by domain . for @xmath470
, we use @xmath597 to denote time - like functions in @xmath598 such that @xmath599 extends to a null function on @xmath600 .
we use @xmath601 to denote the weight functions if necessary .
we also compute @xmath602 , @xmath603 , @xmath604 and @xmath605 for @xmath470 .
* assumption . * through out this section , we assume the following : * @xmath41 are harmonic coordinates w.r.t .
* @xmath607 , @xmath608 , @xmath609 , @xmath610 are small enough for some @xmath611 .
this means @xmath612 is bounded by @xmath613 when approaching the null infinity @xmath54 . *
@xmath614 satisfies @xmath615 and @xmath616 . in section 6 ,
we will prove given small initial data @xmath617 with enough regularity the einstein vacuum solutions will satisfy ( a1 ) and ( a2 ) . and with the choice of @xmath618 satisfy ( a3 ) , the time - like functions we choose in the following will always be time - like . in @xmath620 ,
define @xmath621 where @xmath622 is defined by @xmath623 hence @xmath624 . then in @xmath619 , @xmath625 for some constant @xmath85 .
hence @xmath626 is a regular time - like function w.r.t . @xmath596 all over @xmath619 and bounded .
we choose @xmath627 and @xmath628 in @xmath619 . here
@xmath629 is bounded in the interior of @xmath91 . near the boundary of @xmath619 ,
i.e. in @xmath630 , recall that we use the following coordinates and boundary defining functions : @xmath631 let @xmath577 be a constant and we have the following two lemmas .
[ lem.5.1 ] in @xmath630 , @xmath632 here @xmath16 is the covariant derivative w.r.t . @xmath596 . in @xmath630 ,
change variable to @xmath633 and hence @xmath634 and @xmath635 .
then in coordinates @xmath636 , @xmath637 it is a product type metric .
wlog , at any fixed point @xmath638 , let @xmath471 denote the normal coordinate w.r.t .
standard spherical metric @xmath639 on @xmath472 .
then at @xmath229 , we can write the metric components as follows : @xmath640 , \quad \tilde{m}^{-1}=\left[\begin{array}{ccc}-1+s^2 & -s & 0\\ -s & 1 & 0\\ 0&0 & \mathbbm{1}_{n-1}\end{array}\right ] . \end{gathered}\ ] ] for the quadratic form , notice that @xmath641 where @xmath642 then the formula for @xmath643 follows directly .
for @xmath644 , by lemma [ lem.3 ] we only need to compute @xmath645 and @xmath646 .
using to the formula of @xmath496 in the proof of lemma [ lem.2 ] , we have @xmath647 here @xmath648 . at @xmath229 ,
the non - zero connection components are @xmath649 hence @xmath650 and @xmath651 follows .
those two formulae do not change much in the following sense when we take the background metric to be @xmath455 .
[ lem.5.2 ] in @xmath619 , @xmath652 we only need to concern the region near boundary . in @xmath653 , let us do the similar computation as in the proof of lemma [ lem.5.1 ] by using the same coordinate @xmath196 .
first , by the proof of proposition [ prop.4 ] , write @xmath654 hence the first formula follows directly by @xmath655 according to the formula of @xmath485 in the proof of lemma [ lem.2 ] , @xmath656 which implies the second formula . here
@xmath648 . in @xmath199 ,
recall that the coordinates and boundary defining functions are as follows : @xmath658 define for @xmath155 @xmath659 here @xmath660 is a natural extension of @xmath661 into @xmath199 . then in @xmath657 , @xmath662 hence @xmath663 are time - like functions all over @xmath664 and @xmath665 is null on @xmath666 w.r.t .
@xmath596 . in @xmath657 , with @xmath577 a constant , we have the following two lemmas .
[ lem.5.3 ] in @xmath657 , @xmath667 in @xmath657 , change variable to @xmath668 and hence @xmath669 . then in coordinate @xmath200
, @xmath670 with @xmath471 the normal coordinates w.r.t .
standard spherical metric @xmath639 at @xmath671 , we can write the metric components at @xmath229 as follows : @xmath672 ,
\quad \tilde{m}^{-1}=\left[\begin{array}{ccc}-a(2-a ) & -1&0 \\ -1 & 0&0\\0&0&\mathbbm{1}_{n-1}\end{array}\right].\ ] ] then @xmath673 and @xmath674 follow from the following computation : @xmath675 for @xmath676 , we have @xmath677 and the non - zero connection components at @xmath229 are @xmath678 which implies that @xmath679 . \end{aligned}\ ] ] then @xmath676 follows .
[ lem.5.4 ] for @xmath0 and in @xmath657 , @xmath680 we can do the similar computation as in the proof of lemma [ lem.5.1 ] by writing out the metric components in local coordinates @xmath681 with @xmath682 and @xmath471 is normal w.r.t .
the standard spherical metric @xmath639 at @xmath683 : @xmath684_{1\times ( n-1 ) } \\ -a(\tilde{h}_{\rho\rho}-a\tilde{h}_{0\rho } ) & a^{2}(\tilde{h}_{\rho\rho } - 2a\tilde{h}_{0\rho } + a^2\tilde{h}_{00 } ) & a^{2}[\theta_1(\tilde{h})]_{1\times ( n-1 ) } \\ a[\theta_1(\tilde{h})]_{(n-1)\times 1 } & a^{2}[\theta_1(\tilde{h})]_{(n-1)\times 1 } & a^{2}[\theta_1(\tilde{h})]_{(n-1)\times ( n-1 ) } \end{array}\right ] . \\
\tilde{g}^{-1}=\tilde{m}^{-1}+ a^{\frac{n-5}{2}}b^{\frac{n-1}{2 } } \left[\begin{array}{ccc } -a^{2}(\tilde{h}_{\rho\rho } + a\theta_1(\tilde{h } ) ) & -a(\tilde{h}_{\rho\rho}+a\theta_1(\tilde{h } ) ) & a^{2}[\theta_1(\tilde{h})]_{1\times ( n-1 ) } \\ -a(\tilde{h}_{\rho\rho}+a\theta_1(\tilde{h } ) ) & -(\tilde{h}_{\rho\rho } + a^{\frac{n-1}{2 } } \theta_2(\tilde{h } ) ) & a[\theta_1(\tilde{h})]_{1\times ( n-1 ) } \\ a^{2}[\theta_1(\tilde{h})]_{(n-1)\times 1 } & a[\theta_1(\tilde{h})]_{(n-1)\times 1 } & a^{2}[\theta_1(\tilde{h})]_{(n-1)\times ( n-1 ) } \end{array}\right ] .
\end{gathered}\ ] ] hence @xmath685 then the estimates of @xmath686 and @xmath687 follow directly . for @xmath688 ,
first by the formula of @xmath485 in the proof of proposition [ prop.4 ] , we have @xmath689 hence @xmath690 and moreover , @xmath691 this implies that @xmath692 we prove the last formula . in @xmath694 ,
define for @xmath695 @xmath696 where @xmath697 are defined by @xmath698 hence @xmath699 .
then in @xmath693 @xmath700 we divide @xmath693 into two parts : @xmath701 where @xmath702 is compact and @xmath703 .
recall that in @xmath203 we use the coordinates @xmath704 hence in @xmath705 , @xmath706 here @xmath707 is time - like functions all over @xmath708 ; @xmath709 is time - like in the interior of @xmath710 and null on @xmath711 .
with @xmath577 a constant , we have the following two lemmas .
[ lem.5.5 ] in @xmath712 , @xmath713 in @xmath705 under coordinates @xmath204 with @xmath471 normal w.r.t .
standard spherical metric at @xmath714 , we can write the metric components of @xmath51 as follows : @xmath715 , \quad \tilde{m}^{-1}=\left[\begin{array}{ccc } 0 & 1 & 0 \\1&\rho^2&0\\0&0&\mathbbm{1}_{n-1}\end{array}\right].\ ] ] then @xmath716 and @xmath717 follow from @xmath718 for @xmath719 , we have @xmath720 and the non - zero connection components at @xmath229 are @xmath721 which implies that @xmath722 then @xmath719 follows . [ lem.5.6 ] for @xmath0 and in @xmath705 , @xmath723 we can do the similar computation as in the proof of lemma [ lem.5.1 ] by writing out the metric components in local coordinates @xmath204 with @xmath471 normal w.r.t .
the standard spherical metric at @xmath724 : @xmath725_{1\times(n-1)}\\ \rho^2\theta_1(\tilde{h } ) & \tilde{h}_{\rho\rho } & \rho[\theta_1(\tilde{h})]_{1\times ( n-1 ) } \\ \rho^3[\theta_1(\tilde{h})]_{(n-1)\times 1 } & \rho[\theta_1(\tilde{h})]_{(n-1)\times ( n-1 ) } & \rho^2[\theta_1(\tilde{h})]_{(n-1)\times ( n-1 ) } \end{array}\right ] .
\\ \tilde{g}^{-1}=\tilde{m}^{-1}+\rho^{\frac{n-5}{2 } } \left[\begin{array}{ccc } -\tilde{h}_{\rho\rho}+\rho^{\frac{n-1}{2}}\theta_2(\tilde{h } ) & \rho^2\theta_1(\tilde{h } ) & \rho[\theta_1(\tilde{h})]_{1\times(n-1)}\\ \rho^2\theta_1(\tilde{h } ) & \rho^4\theta_1(\tilde{h } ) & \rho^3[\theta_1(\tilde{h})]_{1\times ( n-1 ) } \\ \rho[\theta_1(\tilde{h})]_{(n-1)\times 1 } & \rho^3[\theta_1(\tilde{h})]_{(n-1)\times ( n-1 ) } & \rho^2[\theta_1(\tilde{h})]_{(n-1)\times ( n-1 ) } \end{array}\right ] .
\end{gathered}\ ] ] hence @xmath726 can be estimated by the following @xmath727 for @xmath728 , we first have @xmath729 hence @xmath730 moreover , @xmath731 hence @xmath732 then @xmath728 follows . in @xmath207 , recall that we have the coordinates and boundary defining functions @xmath734 and in @xmath211 , we have @xmath735 both sets of coordinates can be extended to @xmath736 with the transition map @xmath737 define @xmath738 and let @xmath739 then in @xmath740 , @xmath741<0 , \\ & \langle \nabla t'_4,\nabla t'_4\rangle_{\tilde{m } } = -2\bar{a}(\tfrac{2-\bar{a}}{2})^{2 - 2\alpha}\bar{b}^{2\alpha } ( \tfrac{2-\bar{a}}{2}\alpha-1+\bar{a})(1+\tfrac{\alpha-1}{2-\bar{a}}\bar{a } ) < 0 , \\ & \langle \nabla t'_4,\nabla t_4\rangle_{\tilde{m } } = -(\tfrac{2-\bar{a}}{2})^{1-\alpha}\bar{b}^{\alpha}\big ( \alpha\bar{a}+\tfrac{2(1-\bar{a})^2}{2-\bar{a}}(1- ( 2-\alpha)(\tfrac{\bar{a}(2-\bar{a})}{2})^{\delta'+\frac{1}{2 } } ) \big)<0 . \end{aligned}\ ] ] hence @xmath742 are time - like in @xmath740 and @xmath743 is null on @xmath744 .
with @xmath577 a constant , we have the following two lemmas . [ lem.5.7 ] in @xmath733 , @xmath745 where @xmath746 under coordinates @xmath747 with @xmath748 and
@xmath749 normal w.r.t .
the standard spherical metric @xmath639 at @xmath750 , the metric components are : @xmath751,\quad \tilde{m}^{-1}= \left[\begin{array}{ccc } \bar{a}(2-\bar{a})&1&0\\1 & 0 & 0\\ 0&0 & \frac{1}{(1-\bar{a})^{2}}\mathbbm{1}_{n-1 } \end{array}\right].\ ] ] then @xmath752 follows from @xmath753\big ) , \\
\langle \nabla t_4,\nabla t''_4\rangle_{\tilde{m } } = & \ -1-(1-\bar{a})^2(\tfrac{\bar{a}(2-\bar{a})}{2})^{\delta'-\frac{1}{2 } } , \\
\langle \nabla t_4,\nabla v\rangle_{\tilde{m } } = & \ -((\tfrac{\bar{a}(2-\bar{a})}{2})^{\delta'-\frac{1}{2}}+\tfrac{1}{2-\bar{a}})\partial_{\xi}v+ ( 1-\bar{a}-2(\tfrac{\bar{a}(2-\bar{a})}{2})^{\delta'+\frac{1}{2}})\partial_{\bar{a}}v , \\
\langle \nabla t'_4,\nabla v\rangle_{\tilde{m } } = & \ \bar{b}^{\alpha}(\tfrac{2-\bar{a}}{2})^{-\alpha}\big ( -(\tfrac{2-\bar{a}}{2}+\tfrac{\alpha-1}{2}\bar{a})\partial_{\xi}v + \tfrac{\alpha-2}{2}(1-\bar{a})(2-\bar{a})\bar{a}\partial_{\bar{a}}v\big ) , \\
\langle \nabla t''_4,\nabla v\rangle_{\tilde{m } } = & \ -\tfrac{1}{2-\bar{a}}\partial_{\xi}v+(1-\bar{a})\partial_{\bar{a}}v , \\
\langle \nabla v,\nabla v\rangle_{\tilde{m } } = & \ 2\partial_{\bar{a}}v\partial_{\xi}v+\bar{a}(2-\bar{a})\partial_{\bar{a}}v+\tfrac{|\slashpar_{\theta}v|^2}{(1-\bar{a})^2}. \end{aligned}\ ] ] for @xmath754 , we have @xmath755 and the non - zero connection components at @xmath229 are @xmath756 which implies @xmath754 . [ lem.5.8 ] for @xmath0 and @xmath757 $ ] , if @xmath758 then in @xmath733 , @xmath759 let @xmath760 .
we only need to show that @xmath761 where @xmath762 first , @xmath763 \\
\geq&\ \lambda \tfrac{(1-\bar{a})^2}{2-\bar{a}}[1-(\tfrac{\bar{a}(2-\bar{a})}{2})^{\delta'+\frac{1}{2 } } ] \geq 0 . \end{aligned}\ ] ] similarly , @xmath764 \\ \geq & \ \tfrac{\bar{a}}{2}[\alpha'-n+n\bar{a}(2-\bar{a})+(\alpha'-2)(1-\bar{a})^2 ] + \lambda \tfrac{(1-\bar{a})^2}{2-\bar{a}}[1-(\tfrac{\bar{a}(2-\bar{a})}{2})^{\delta'+\frac{1}{2 } } ] \\ \geq & \
\tfrac{\bar{a}(1-\bar{a})^2}{2}[2\alpha'-n-2+\tfrac{\lambda}{4 } ] \geq 0 . \end{aligned}\ ] ] hence we prove the first inequality in ( [ eq.5.4.1 ] ) . for the second inequality in ( [ eq.5.4.1 ] ) , there are two cases .
first , if @xmath765 , then @xmath766 . so we only need to show @xmath767 . in this case ,
notice that @xmath768 .
then @xmath769 here @xmath770 and hence @xmath771 \\
\geq&\ \tfrac{1}{(2-\bar{a})^2 } [ \tfrac{n+\alpha'}{2}-\tfrac{n}{\sqrt{2}}-(n-\alpha'-\tfrac{n-1-\alpha'}{\sqrt{2}})(1-\bar{a } ) ( \tfrac{\bar{a}(2-\bar{a})}{2})^{\frac{1}{2 } } ] \\
\tfrac{1}{4(2-\bar{a})^2 } [ ( 4 - 2\sqrt{2})n-\tfrac{3\sqrt{2}}{2}-\tfrac{5}{2 } ] \geq 0 \end{aligned}\ ] ] for all @xmath0 .
second , if @xmath772 , then the second inequality in ( [ eq.5.4.1 ] ) is equivalent to @xmath773 in this case , @xmath774 and @xmath775 \\ \geq & \bar{a}^2 [ \alpha'(1-\bar{a})(1-(\tfrac{\bar{a}(2-\bar{a})}{2})^{\delta'+\frac{1}{2 } } ) + \tfrac{2-\bar{a}}{2 } + ( \alpha'\bar{a}-(n-1)(2-\bar{a}))(\tfrac{1}{2}-(\tfrac{\bar{a}(2-\bar{a})}{2})^{\delta'+\frac{1}{2 } } ) ] \\ \geq & \
\bar{a}^2 ( \alpha'+2-n)(2-\bar{a})(\tfrac{1}{2}-(\tfrac{\bar{a}(2-\bar{a})}{2})^{\delta'+\frac{1}{2 } } ) \geq 0 .
\end{aligned}\ ] ] hence ( [ eq.5.4 ] ) follows by @xmath776 by direct computation , @xmath777 hence @xmath778 \bar{a}^2 ( \tfrac{1}{2}-(\tfrac{\bar{a}(2-\bar{a})}{2})^{\delta'+\frac{1}{2}})^2 \\ \geq & \ \alpha ' [ 2\alpha'-1 ] \bar{a}^2 ( \tfrac{1}{2}-(\tfrac{\bar{a}(2-\bar{a})}{2})^{\delta'+\frac{1}{2}})^2 \\ \geq&\ \bar{a}^2(2-\bar{a})^2d_3 ^ 2 .
\end{aligned}\ ] ] we finish the proof . [ lem.5.9 ] for @xmath0 and in @xmath733 @xmath779 the proof is similar as the proof of lemma [ lem.5.4 ] by substituting @xmath780 by @xmath781 if restricting to @xmath207 and similar as the proof of lemma [ lem.5.2 ] if restricting to @xmath211 .
in this section , we mainly prove the following theorem by energy estimates .
[ thm.est ] for @xmath0 if given cauchy data @xmath782 where @xmath783 is defined in ( [ initial.2 ] ) , then there exists a global einstein vacuum solution @xmath38 such that @xmath784 is @xmath71 up to @xmath785 and hence the radiation field is well defined : @xmath786 which provides the leading term of @xmath72 along the bicharacteristic curves .
moreover , @xmath787 for some @xmath788 and @xmath85 .
the theorem follows from proposition [ prop.est.1 ] , [ prop.est.2 ] , [ prop.est.3 ] , [ prop.est.4 ] and corollary [ cor.6.2 ] , [ cor.6.3 ] , [ cor.6.4 ] .
we do energy estimates for @xmath83 satisfying the conformal transformation of reduced einstein equations ( [ eq.12 ] ) by using the time - like functions defined in section @xmath789 .
the boundary defining functions are given in section @xmath109 in each domain @xmath790 for @xmath470 .
here @xmath791 are fixed all over this section and @xmath66 is chosen to be small enough .
the image space of @xmath792 will be refined in section [ sec.mollerop ] .
to solve the equations ( [ eq.12 ] ) in @xmath619 , it is in fact a local existence theorem .
we prove the local existence and uniqueness theorem in the conformal setting for @xmath793 $ ] .
the local well - posedness is a classical result .
see @xcite or @xcite for more details . here
we restate it in weighted b - sobolev spaces , which will lead to the asymptotical analysis of einstein vacuum solutions .
is bounded by @xmath8 , @xmath794 and @xmath126.,width=288 ] recall that @xmath619 is bounded by @xmath795 , @xmath796 and @xmath126 .
we use @xmath626 to slice the domain . for @xmath797 $ ] , denote @xmath798 the space - like hypersurface @xmath794 is a p - submanifold of @xmath91 with induced boundary defining function @xmath799 . let @xmath800 be the domain bounded by @xmath8 , @xmath794 and @xmath126 .
in particular , @xmath801 .
define the high energy norm on @xmath794 by @xmath802 .
\end{aligned}\ ] ] here @xmath803 and @xmath804 where @xmath805 is a smooth cutoff function such that @xmath806 if @xmath807 and @xmath808 if @xmath809 . here @xmath810 we choose @xmath811 to define @xmath812 . with above notation and given initial data ( [ cauchy ] ) , we have for some constant @xmath813 such that @xmath814 [ lem.6.1 ] there exists @xmath815 such that if @xmath816 then on @xmath794 @xmath817 where @xmath292 is independent of @xmath818 if it is small enough . near the boundary of @xmath794 , we can choose the coordinates @xmath819 and with this coordiantes , @xmath820 where @xmath821_{t_1}\times [ 0,1]_{\rho}\times \mathbb{s}^{n-1})$ ] is bounded above and below .
hence @xmath822 gives the b - sobolev norm @xmath823 .
the equivalence of @xmath822 and @xmath824 comes from the equivalence between @xmath825 and @xmath826 , as well as the equivalence between the volume forms @xmath827 and @xmath828 by lemma [ lem.5.2 ] . since @xmath829 for some @xmath830 , the second inequality comes from sobolev embedding theorem .
see lemma [ lem.2.3 ] .
we find out the existence of @xmath72 to the equations ( [ eq.2 ] ) ( resp . @xmath83 to the euqations ( [ eq.12 ] ) ) by considering it as the limit of a sequence @xmath831 ( resp .
@xmath832 ) , which are solutions to the linear equations ( resp . their conformal transformation ) for @xmath833 starting with @xmath834 ( resp .
@xmath835 ) : @xmath836 here @xmath837 and @xmath831 has cauchy data @xmath35 . notice that there is no guarantee that @xmath41 are wave coordinates w.r.t .
@xmath838 . however
, we do nt need this property here . in the conformal transformation of equations ,
@xmath839 we also study the equation for @xmath840 and @xmath841 to get the convergence of @xmath831 and @xmath832 : @xmath842 [ lem.6.2 ] there exists @xmath843 such that if @xmath844\ ] ] is true for @xmath845 then @xmath846.\ ] ] here @xmath292 is a constant independent of @xmath847 and @xmath848 for @xmath849 small enough . when @xmath850 , since @xmath835 , the condition in lemma [ lem.6.1 ] is automatically satisfied .
hence @xmath851 . here
@xmath852 is the constant in lemma [ lem.6.1 ] independent of @xmath853
. choose @xmath849 such that @xmath854 and we can apply lemma [ lem.6.1 ] until @xmath855 , i.e. on @xmath794 , @xmath856 for @xmath857 . to get the high energy estimate for @xmath858 , we first deduce the equation for @xmath859 for @xmath860 : @xmath861\tilde{h}^{l+1 } \\
\rho_0^{\frac{n-1}{2 } } \sum_{|\alpha|\leq k+1,i\leq k+1,|\alpha|+i\leq k+2,1\leq j}\theta_0(\tilde{h}^{l})(\tilde{\partial}^{\alpha_1}\tilde{h}^{l}, .. ,\tilde{\partial}^{\alpha_j}\tilde{h}^{l } , \tilde{\partial}^{i}\tilde{h}^{l+1 } ) \end{aligned}\ ] ] by lemma [ lem.3 ] , @xmath862 by lemma [ lem.5.1 ] and lemma [ lem.5.2 ] , we have @xmath863 which implies that @xmath864 .
\end{gathered}\ ] ] here @xmath85 is a constant independent of @xmath849 if it is small enough .
we finish the proof .
[ prop.est.1 ] there exists @xmath865 such that if @xmath866 then the equations ( [ eq.12 ] ) have a unique solution in @xmath619 such that @xmath867,\ ] ] where @xmath85 is a constant independent of @xmath868 if it is small enough .
since @xmath834 and @xmath242 is the solution of standard wave equation on minkowski space - time , we can apply lemma [ lem.6.2 ] and get @xmath869 . here
@xmath830 is the constant in lemma [ lem.6.2 ] independent of @xmath870 $ ] .
choose @xmath865 small such that @xmath871 and @xmath872.\ ] ] we do energy estimates for @xmath873 which satisfies equations ( [ eq.29 ] ) and hence @xmath874(\tilde{h}^{l+1}-\tilde{h}^l ) + \tilde{d}^i\big((\box_{\tilde{g}^{l-1}}-\box_{\tilde{g}^l})\tilde{h}^l + ( \gamma(\tilde{h}^{l-1})-\gamma(\tilde{h}^{l}))\tilde{h}^l\big ) .
\end{aligned}\ ] ] here the estimates for @xmath841 are deduced in the same way as in the proof of lemma [ lem.6.2 ] : @xmath875 where @xmath876 is a constant independent of @xmath849 .
this implies that @xmath877 for @xmath833 , let @xmath878 } m_1^n(\mathfrak{t};\tilde{h}^{l}-\tilde{h}^{l-1},\tilde{m}).\ ] ] then @xmath879 and for @xmath880 , @xmath881 choose @xmath849 and hence @xmath868 even smaller if necessary such that @xmath882 .
then for any @xmath870 $ ] , @xmath883 @xmath884 hence @xmath832 converges to @xmath83 with @xmath885 for all @xmath870 $ ] . here
@xmath83 provides a solution to equations ( [ eq.12 ] ) .
choose @xmath849 and hence @xmath868 even smaller if necessary such that lemma [ lem.6.2 ] holds for @xmath83 : @xmath886 we finish proving the existence .
for the uniqueness , suppose there is another solutions @xmath887 with the same cauchy data such that @xmath888 for some constant @xmath85 .
then choose @xmath868 even smaller if necessary such that @xmath889 .
notice that @xmath890 satisfy the following equations : @xmath891(\tilde{h}^{l+1}-\tilde{h}^l ) \\ & \
+ d^i\big((\box_{\tilde{g}'}-\box_{\tilde{g}})\tilde{h}'+(\gamma(\tilde{h}')-\gamma(\tilde{h}))\tilde{h}'\big ) . \end{aligned } \end{gathered}\ ] ] a similar estimates as above shows that @xmath892 for some constant @xmath893 .
therefore @xmath894 . when restricting to @xmath619 , all the proofs work for spacial dimension @xmath6 .
to solve the equations ( [ eq.12 ] ) up to @xmath657 with given cauchy data ( [ cauchy ] ) , we first notice that the solution we find in @xmath619 can be extended a bit over @xmath895 since @xmath619 is closed .
if the cauchy data is small enough , we show that this extension will never stop in @xmath657 until it arrives @xmath785 .
is bounded by @xmath895 , @xmath896 , @xmath794 and @xmath126.,width=288 ] recall that @xmath657 is bounded by @xmath796 , @xmath897 for @xmath155 and @xmath898 .
we use @xmath660 to slice the domain . for @xmath899
$ ] , denote by @xmath900 the space - like hypersurface w.r.t . @xmath596 with defining function
notice that @xmath901 can be viewed as a closed subset of a p - submanifold of @xmath91 , which has induced boundary defining function @xmath902 .
let @xmath903 be the domain bounded by @xmath895 , @xmath896 , @xmath126 and @xmath904 .
then @xmath905 .
define the high energy norm on @xmath901 and @xmath896 by @xmath906 here @xmath907 , @xmath908 and @xmath909 .
we choose @xmath910 in the definition of @xmath911 .
recall that we choose @xmath912 such that @xmath913 in the definition of @xmath914 . with above notation , proposition [ prop.est.1 ]
implies that for some @xmath915 , @xmath916 [ lem.6.3 ] for @xmath0 , there exists @xmath917 such that if @xmath918 then @xmath919 for some constant @xmath292 independent of @xmath920 if it is small enough .
moreover , there exists @xmath921 such that if @xmath922 , then @xmath923 for some constant @xmath852 independent of @xmath924 if it is small enough .
similar as the proof of lemma [ lem.6.1 ] , the equivalence between @xmath925 and @xmath926 comes from the equivalence between the quadratic forms @xmath927 and @xmath928 , as well as it between the volume forms @xmath827 , @xmath828 by lemma [ lem.5.3 ] and lemma [ lem.5.4 ] . here on @xmath901 , we can choose the coordinates @xmath929 , so @xmath930 hence @xmath925 gives the b - sobolev norm : for some @xmath876 @xmath931 the second and third inequalities come from sobolev embedding theorem .
see lemma [ lem.2.3 ] next we show that the harmonic gauge conditions make @xmath612 have a bit more decay than @xmath83 when approaching @xmath932 .
by lemma [ lem.4 ] , @xmath933 so for @xmath934 , @xmath935 and for @xmath936 , @xmath937 and the last inequality is proved similarly as above : for @xmath938 , @xmath939 \right|^2 b^{n-1 - 2\delta}\frac{db}{b}d\theta \\ \leq & \ c ' a\left(m_2^n(\mathfrak{t};\tilde{h},\tilde{m})\right)^2 ; \end{aligned}\ ] ] and for @xmath940 , @xmath941 we finish the proof .
[ prop.est.2 ] for @xmath0 , there exists @xmath942 such that if @xmath943 , then the equations ( [ eq.12 ] ) have a unique solution in @xmath657 such that @xmath944,\ ] ] where @xmath85 is a constant depending on @xmath945 but independent of @xmath946 if it is small enough . by the proof of existence theorem in @xmath619 , we see that first we can choose @xmath942 small such that the we can solve the equations ( [ eq.12 ] ) for @xmath947 with the conditions in lemma [ lem.6.3 ] holds
. then by lemma [ lem.3 ] , for @xmath948 @xmath949 here the divergence term can be estimated according to lemma [ lem.6 ] , [ lem.5.3 ] , [ lem.5.4 ] : @xmath950 ( m_2^n(\mathfrak{t};\tilde{h},\tilde{g}))^2 \end{aligned}\ ] ] where @xmath876 is a constant independent of @xmath946 if it is small enough .
notice the faster decaying of @xmath612 and @xmath951 in @xmath426 direction guarantees that the perturbation of @xmath952 all fall into the smaller term and the leading term @xmath953 is determined by the linear equations .
here @xmath954 since it is an increasing function of @xmath955 .
choose @xmath956 smaller if necessary such that @xmath957 then we have @xmath958 hence @xmath959 for some @xmath876 independent of @xmath946 if it is small enough . notice that @xmath960(c '' m^n_2(-\tfrac{1}{4};\tilde{h},\tilde{g } ) ) ^2 , \\
\longrightarrow l_2^n(\mathfrak{t};\tilde{h},\tilde{g } ) \leq c ' '' |\mathfrak{t}|^{\delta-\frac{1}{2 } } m^n_2(-\tfrac{1}{4};\tilde{h},\tilde{g}),\quad \forall \mathfrak{t}\in [ -\tfrac{1}{4},\mathfrak{t}_0 ) \end{gathered}\ ] ] where @xmath893 is a constant independent of @xmath961 too .
so we prove ( [ eq.6.2 ] ) for @xmath962 .
moreover , @xmath963 for some @xmath964 independent of @xmath946 if it is small enough . hence choosing @xmath946 even smaller if necessary such that above two inequalities guarantee that the conditions in lemma [ lem.6.3 ] holds up to @xmath965 , which allows us to extend the solution up to and hence a bit over @xmath965 with all the above estimates hold without changing of constants . therefore the extension will never stop until it arrives @xmath666 .
we finish the proof .
[ cor.6.2 ] for @xmath0 , suppose @xmath966 where @xmath946 is chosen such that proposition [ prop.est.2 ] holds
. then @xmath83 is @xmath71 up to @xmath967 and @xmath968 for some @xmath85 depending on @xmath969 but independent of @xmath946 if it is small enough . by lemma [ lem.6.3 ] and proposition [ prop.est.2 ] , for some constant @xmath830 @xmath970
hence @xmath83 is @xmath71 up to @xmath971 and the radiations field @xmath972 is well defined .
moreover , define @xmath973 then for some @xmath974 @xmath975 here the @xmath976 term comes from the derivative of the volume form @xmath977 . since @xmath978 is bounded , this implies that @xmath979.\ ] ] finally , @xmath980 gives the norm bound of @xmath981 . to solve the equations ( [ eq.12 ] ) in @xmath693
we first notice that with cauchy data small enough , the solution we find in @xmath982 can be extended to @xmath983 since it is a compact domain in the interior of @xmath91 .
we show that if the cauchy data is small enough then this extension can continue into @xmath710 and never stop until it arrives @xmath932 .
is bounded by @xmath896 , @xmath984 , @xmath985 and @xmath986.,width=288 ] recall that @xmath897 , @xmath987 and @xmath988 . for @xmath989 $ ] ,
denote by @xmath990 the space like hypersurfaces w.r.t . @xmath596 and @xmath991 the domain bounded by @xmath992 and @xmath993
. then @xmath994 .
define the high energy norm on @xmath896 , @xmath985 and @xmath993 by @xmath995 here @xmath996 , @xmath997 , @xmath998 and @xmath999 we choose @xmath1000 in the definition of @xmath1001 in @xmath705 .
notice that in @xmath657 , we use the quadratic form @xmath1002 on @xmath896 to define the norm @xmath1003 . here
the two quadratic forms @xmath1004 and @xmath1002 are equivalent on @xmath1005 , i.e. @xmath1006 for some constant @xmath85 .
on the other hand , @xmath660 and @xmath709 are equivalent on @xmath1005 , i.e. @xmath1007 for some constant @xmath1008 . hence the estimate of @xmath1009 in @xmath657
give the estimate of @xmath1010 . by proposition [ prop.est.1 ] ,
proposition [ prop.est.2 ] and classical theorems for wave equations in compact domain , we have @xmath1011 where @xmath1012 independent of @xmath66 if it is small enough . in particular we choose @xmath912 in the definition of @xmath707 such that @xmath913 and @xmath155 large in the definition of @xmath709 .
similar as lemma [ lem.6.3 ] , we have [ lem.6.4 ] assume @xmath0 .
there exists @xmath1013 such that if @xmath1014 , then on @xmath986 for @xmath1015 , @xmath1016 for some constant @xmath292 independent of @xmath1017 if it is small enough .
moreover , there exists @xmath1018 such that if @xmath1019 , then @xmath1020 the proof of equivalence between @xmath1021 and @xmath1022 is the same as before , which is directly from the equivalence of quadratic forms @xmath1023 , @xmath1024 and equivalence of volume forms @xmath1025 , @xmath1026 . and @xmath1027 gives the sobolev norms : @xmath1028 for some @xmath876 .
changing coordinates from @xmath1029 to @xmath1030 , we can prove the rest inequalities similar as in @xmath657 .
[ prop.est.3 ] for @xmath0 , there exists @xmath1031 such that if @xmath1032 , then the equations ( [ eq.12 ] ) have a unique solution up to @xmath985 such that @xmath1033,\ ] ] where @xmath85 is a constant depending on @xmath945 but independent of @xmath1034 if it is small enough . we first can choose @xmath1034 such that we can solve the equation in @xmath1035 and ( [ eq.30 ] ) holds .
choose @xmath1034 even smaller if necessary such that we can extend the solution up to @xmath948 for some @xmath1036 with lemma [ lem.6.4 ] satisfied and @xmath1037 then by lemma [ lem.3 ] , @xmath1038 here the divergence term can be estimated by lemma [ lem.5.5 ] , [ lem.5.6 ] as follows : @xmath1039 hence @xmath1040 here @xmath1041 and @xmath876 depends on @xmath1042 only if @xmath1034 small enough .
moreover , @xmath1043 where @xmath893 is independent of @xmath1034 if it is small enough . we have proved ( [ eq.6.3 ] ) for all @xmath1044 .
notice that by lemma [ lem.6.4 ] , @xmath1045 choose @xmath1034 even smaller if necessary such that the above two inequalities guarantee that lemma [ lem.6.4 ] holds .
then when we extend the solution up to and a bit over @xmath1046 , all the above estimates are still true with the same constants
. hence the extension does not stop until it arrives @xmath1047 .
[ cor.6.3 ] for @xmath0 suppose @xmath1032 with @xmath1034 chosen in proposition [ prop.est.3 ] .
then @xmath83 is @xmath1048 up to @xmath1049 and @xmath1050 for some constant @xmath85 depending on @xmath1051 but independent of @xmath1034 if it is small enough . by lemma [ lem.6.4 ] and proposition [ prop.est.3 ] , for some constant @xmath830 @xmath1052
hence @xmath83 is @xmath71 up to @xmath1053 and the radiations field @xmath1054 is well defined .
moreover , define @xmath1055 then for some @xmath1056 , @xmath1057 similar as the proof of corollary 2 , we can show that @xmath1058.\ ] ] finally , @xmath1059 gives the norm bound of @xmath1060 . to solve the equation ( [ eq.12 ] ) in @xmath733
, we first notice that we can extend the solution a bit over @xmath985 on a compact region .
we show that if @xmath69 small enough then this extension can be taken all over @xmath733 and hence we solve the equation globally .
is bounded by @xmath985 and @xmath1061 .
, width=288 ] recall that @xmath988 .
let @xmath1062 . for @xmath1063 $ ] , denote by @xmath1064 the space - like hypersurfaces w.r.t @xmath596 with defining function @xmath743 and @xmath1065 the domain bounded by @xmath985 and @xmath1066
. then @xmath1067 .
define the high energy norm on @xmath1066 and @xmath985 by @xmath1068 here @xmath1069 , @xmath997 and @xmath1070 .
we choose @xmath1071 in the definition of @xmath1072 .
notice that , @xmath1073 is defined in a different way from @xmath1074 and @xmath1075 . by the proof of lemma [ lem.4.7 ] @xmath1076 for some @xmath1077 .
this modified definition will simplify the energy estimates in proposition [ prop.est.4 ] .
( if we define @xmath1074 and @xmath1075 in above way , then we can erase @xmath1078 terms in the estimates of divergence terms when proving proposition [ prop.est.2 ] and [ prop.est.3 ] . )
notice that the two quadratic forms @xmath1079 and @xmath1080 are equivalent on @xmath985 , i.e. @xmath1081 for some @xmath85 .
moreover , @xmath709 and @xmath1082 are also equivalent on @xmath985 , i.e. @xmath1083 for some @xmath1008 .
hence the estimate of @xmath1084 in @xmath693 gives the estimate of @xmath1085 .
by proposition [ prop.est.1 ] , [ prop.est.2 ] , [ prop.est.3 ] , for all @xmath1086 $ ] @xmath1087 where @xmath1088 independent of @xmath66 if it is small enough .
in particular , in the definition of @xmath1082 we choose @xmath1089 such that lemma [ lem.5.8 ] holds for @xmath1090 , and in the definition of @xmath1091 we choose @xmath912 which is close to @xmath1051 such that @xmath1092 and @xmath1093 .
similar as lemma [ lem.6.4 ] , we have [ lem.6.5 ] assume @xmath0 .
there exists @xmath1094 such that if @xmath1095 , then on @xmath1066 for @xmath1096 , @xmath1097 for some constant @xmath292 independent of @xmath1098 if it is small enough
. moreover , there exists @xmath1099 such that if @xmath1100 then @xmath1101 the proof of equivalence between @xmath1102 and @xmath1103 is the same as before , which is directly from the equivalence of quadratic forms @xmath1104 , @xmath1105 and equivalence of volume forms @xmath1025 , @xmath1026 .
and @xmath1106 gives the sobolev norm : @xmath1107 for some @xmath876 .
the sobolev embedding inequalities here come from thinking of @xmath1066 as a union of two domains , @xmath1108 a bounded domain with compact metric and @xmath1109 a domain in manifold with boundary which carries b - metric .
changing coordinates from @xmath200 to @xmath1110 , we can prove the rest inequalities similar as in @xmath657 .
[ prop.est.4 ] for @xmath0 , there exists @xmath1111 such that if @xmath1112 , then the equations ( [ eq.12 ] ) have a unique solution all over @xmath733 such that @xmath1113,\ ] ] where @xmath85 is a constant depending on @xmath945 but independent of @xmath1114 if it is small enough .
we first can choose @xmath1114 such that we can solve the equation in @xmath1115 and ( [ eq.6.4.1 ] ) holds .
choose @xmath1114 even smaller if necessary such that we can extend the solution to @xmath1116 for some @xmath1117 with lemma [ lem.6.5 ] satisfied and @xmath1118 then by lemma [ lem.3 ] , if denoting @xmath1119 , then @xmath1120 here the divergence term can be estimated by lemma [ lem.5.7 ] , [ lem.5.8 ] , [ lem.5.9 ] as follows : @xmath1121 ( m_4^n(\mathfrak{t};\tilde{h},\tilde{g}))^2 \\ \leq&\ [ ( 1 - 2\delta)|\mathfrak{t}|^{-1}+2 c'|\mathfrak{t}|^{\delta-\delta'-1}|\ln|\mathfrak{t}|| ] ( m_4^n(\mathfrak{t};\tilde{h},\tilde{g}))^2 . \end{aligned}\ ] ] hence @xmath1122 ( m_3^n(\mathfrak{t};\tilde{h},\tilde{g}))^2 \\ \longrightarrow \quad \partial_{\mathfrak{t}}\big ( f(\mathfrak{t } ) \big ) \leq f(\mathfrak{t } ) \end{gathered}\ ] ] where @xmath1123 is positive and uniformly bounded for @xmath1086 $ ] .
this implies @xmath1124 here @xmath292 depends on @xmath1042 only if @xmath1114 small enough .
so we have proved ( [ eq.6.4 ] ) for @xmath1125 . by sobolev embedding theorem and lemma [ lem.6.5 ] , this gives @xmath1126 where @xmath1127 , and @xmath976 only depends on @xmath1128 if @xmath1114 is small enough .
choose @xmath1114 even smaller if necessary such that @xmath1129 implies the conditions in lemma [ lem.6.5 ] .
then we can extend the solution up to and a bit over @xmath1130 such that all the above inequalities still hold with the same constants .
hence the extension does not stop until it arrives at @xmath1131 , i.e. we solve the equation ( [ eq.12 ] ) globally .
[ cor.6.4 ] for @xmath0 suppose @xmath1132 with @xmath1114 chosen in proposition [ prop.est.4 ] .
then @xmath83 is @xmath1048 up to @xmath1133 and for any @xmath1134 where @xmath1135 , @xmath1136 for some constant @xmath85 depending on @xmath1137 only but independent of @xmath1114 if it is small enough . by lemma [ lem.6.5 ] and proposition [ prop.est.4 ] , in @xmath733 for some constant @xmath830 @xmath1138
hence @xmath83 is @xmath71 up to @xmath1139 and the radiations field @xmath1140 is well defined . moreover , define @xmath1141 then for some @xmath876 , @xmath1142 here @xmath1143 , which implies that @xmath1144,\ ] ] where @xmath292 depends on @xmath1137 .
finally , @xmath1145 gives the norm bound of @xmath1146 .
for @xmath0 , suppose @xmath70 with @xmath1147 and @xmath69 small such that theorem [ thm.est ] holds
. then the global solution @xmath72 obtained in theorem [ thm.est ] provides a true solution to einstein equation ( [ eq.1 ] ) since @xmath72 satisfies the harmonic gauge condition ( [ harmonic.3 ] ) globally by theorem [ thm.lr ] .
this imposes constraint conditions on the radiation field .
@xmath1148 this is equivalent to @xmath1149 when @xmath1150 has some decay as @xmath1151 . to refine the image space for the radiation field ,
first let us consider the cauchy data with only conormal singularity at @xmath1152 .
[ lem.7.1 ] if @xmath1153 , then @xmath1154 , where @xmath788 is given in theorem [ thm.est ] . for all @xmath1155
, we estimate @xmath1156 in the same way as we do for @xmath1157 .
the only difference is that in this case , we can replace the cubic term @xmath1158 in the estimates of divergence term by @xmath1159 as a result of interpolation .
hence the nonlinear estimating reduces to linear estimating : @xmath1160 for some @xmath830 . for some @xmath1161
, we have @xmath1162 then by the same proof as corollary [ cor.6.2 ] , [ cor.6.3 ] , [ cor.6.4 ] , we have @xmath1163 for some @xmath292 only depending on @xmath1164 .
since @xmath1165 is arbitrary large , @xmath83 only have conormal singularity at @xmath143 . by lemma [ lem.2.3 ] ,
we have @xmath1166 . hence @xmath1167 .
[ lem.7.2 ] if @xmath1153 , then @xmath1168 applying the linear wave operator @xmath1169 to the solution @xmath72 gives @xmath1170 hence @xmath1171 if @xmath1172 .
see @xcite for details . repeating @xmath1173 times until @xmath1174 and
finally we have @xmath1175 fix cauchy data @xmath1176 and let @xmath72 be the solution obtained in theorem [ thm.est ] . consider the cauchy problem for linear wave equation with background metric @xmath38 : @xmath1177 then @xmath1173 globally exists and the radiation field is well defined .
moreover , @xmath1178 is a solution if @xmath1179 .
denote by @xmath1180 and define the following map @xmath1181 for @xmath1182 .
[ lem.7.3 ] for @xmath1183 , @xmath1184 and @xmath1185 have the same boundary operator .
with @xmath1186 , by similar energy estimates in the proof of lemma [ lem.7.1 ] , we first have @xmath1187 . then applying @xmath1169 as in the proof of lemma [ lem.7.2 ] , we get @xmath1188 .
consider the equation @xmath1189 then @xmath1190 hence @xmath1191 . since @xmath1192 for @xmath0 , this term contributes zero to the boundary operator .
see @xcite for details . by lemma [ lem.7.3 ] and the mapping property of @xmath1185 given in @xcite ,
we show the following lemma .
[ lem.7.4 ] the linear mller wave operator @xmath1184 defines a continuous map for @xmath1193 @xmath1194 \end{aligned}\ ] ] here @xmath1195 for some constant @xmath85 independent of @xmath1176 .
moreover , @xmath1185 is an isomorphism .
denote by @xmath1196 be the collection of elements in @xmath77\ ] ] with norm less than @xmath66 and satisfying the harmonic gauge condition ( [ harmonic.char ] ) .
since @xmath452 , @xmath1196 is a small neighborhood of @xmath12 in the sobolev space @xmath1197 .
[ thm.op ] with the assumption in theorem [ thm.est ] and choosing @xmath69 even smaller if necessary , the nonlinear mller wave operator defines a continuous map and open map @xmath1198 for some @xmath85 .
first , @xmath792 is well defined and continuous .
when restricting to @xmath1199 , it is obvious that @xmath1200 where @xmath292 is given in lemma [ lem.7.4 ] . by density argument
, it holds for all @xmath82 .
secondly , the linearization of @xmath792 at @xmath254 is @xmath1201 , where @xmath1202 is the tangent space of @xmath64 .
consider the wave equation @xmath1203 and let @xmath1204 then the solution space of the linearization of harmonic gauge conditions ( [ constraint.6 ] ) and ( [ constraint.7 ] ) is @xmath1205 since @xmath1206 , we have the restriction map @xmath1207 is still an isomorphism . by implicit function theorem
, @xmath1208 is open if @xmath66 is small enough . *
proof of theorem [ mainthm ] * : it follows from theorem [ thm.est ] and theorem [ thm.op ] directly .
in this section , we prove some mapping property of @xmath1209 on @xmath396 for @xmath6 by the same method we used in @xcite , where we show the mapping properties of radon transform and radiation field map for standard wave equation .
define for @xmath1210 @xmath1211 then a similar proof as for the inverse radon transform in @xcite shows that : @xmath1212 where @xmath1213 and @xmath1214 is the radon transform @xmath1215 with @xmath1216 the induced lebesgue measure on hyperplane @xmath1217 and @xmath1218 are defined as follows : for @xmath1219 , @xmath1220 by the mapping property of @xmath270 , we only need to show the above map is continuous . for @xmath1225 , by a similar proof as in @xcite , @xmath1222 extends to a continuous map @xmath1226 which restricts to the boundary to define @xmath1227 here @xmath1228 is a meromorphic family of semicalssical pseudodifferential operator with simple poles at @xmath1229
. moreover , let @xmath1230 defined by the schwartz kernel @xmath1231 .
then @xmath1232 a(\tau ) & \ n=\mathrm{odd } \end{cases } \end{gathered}\ ] ] where @xmath1233 is the antipodal map and @xmath1234 hence the schwartz kernel of @xmath1228 can be split into three parts : @xmath1235 with @xmath1236 the semiclassical parameter as @xmath1237 . from standard estimates on the boundedness of semiclassical families
, @xmath1228 defines two continuous maps : @xmath1238 by mellin transform and taking @xmath1239 , we finish the proof .
mm y. choquet - bruhat , _
theorem dexistence pour certains systemes dequations aux derivees partielles nonlinearires . _
acta . math . * 88*(1982 ) , 141 - 225 .
y. choquet - bruhat , _ the null condition and asymptotic expansions for the einstein s equations _ , ann .
phys , ( leipzig ) * 9*(2000 ) , 258 - 266 .
y. choquet - bruhat , r.p .
geroch , _ global aspects of the cauchy problem in general relativity _ , cmp * 14*(1969 ) , 329 - 335 .
d. christodoulou , _ global solutions of nonlinear hyperbolic equations for small initial data _ , comm .
pure appl . math . * 39*(1986 ) , 267 - 282 .
d. christodoulou , s. klainerman , _ the global nonlinear stability of the minkowski space _ , princeton mathematical series , * 41*(1993 ) , princeton university press .
friedlander , _ radiation fields and hyperbolic scattering theory _ , math .
cambridge philos .
* 88 * ( 1980 ) , no.3 , 483 - 515 .
s. helgason , _ radon transform _ , birkhauser boston , 1999 .
l. hrmander , _ lectures on nonlinear hyperbolic differential equations _ , mathematiques and applications , vol.*26*(1997 ) , springer , berlin .
s. klainerman , _ the null condition and global existence to nonlinear wave equations_. lectures in applied mathematics * 23*(1986 ) , 293 - 329 .
h. lindblad , i. rodnianski , _ global existence for the einstein vacuum equations in wave coordinates _ , comm .
phys . * 256*(2005 ) , no.1 , 43 - 110 .
h. lindblad , i. rodnianski , _ the global stability of the minkowski space - time in harmonic gauge _ ,
arxiv : math/0411109 .
r. b. melrose , _ differential analysis on manifolds with corners _ , in preparation .
r. b. melrose , _ transformation of boundary problems _ , acta * 147*(1981 ) , 149 - 236 .
r. b. melrose , f. wang , _ radon transform and radiation field _
, in preparation . | in this paper , the radiation field is defined for solutions to einstein vacuum equations which are close to minkowski space - time with spacial dimension @xmath0 .
the regularity properties and asymptotic behavior of those einstein vacuum solutions are established at the same time . in particular , the map from cauchy intial data to the radiation field is proved to be an isomorphism when restricting to a small neighborhood of minkowski data in suitable weighted b - sobolev spaces . |
* derivation of the smoluchowski equation for fb model :* according to the liouville theorem ( @xmath143 ) the time evolution of probability density @xmath144 in terms of @xmath14 and @xmath13 satisfies @xmath145 insertion of two langevin equations ( eq.(1 ) in the main text ) for the fluctuating bottleneck model @xmath146 $ ] and @xmath147 into eq.[eqn : liouville ] leads to where @xmath149 and @xmath150 . using the vector notation as in the second line of eq.[eqn : liouville2 ] , one can formally solve for the probability density @xmath151 as @xmath152 averaging @xmath151 over noise after iterating @xmath151 into the noise related term in the integrand and exploiting the fluctuation - dissipation theorem , we obtain the smoluchowski equation for @xmath144 in the presence of a reaction sink , @xmath153 , @xmath154\overline{\varphi}(x , r , t ) , \label{eqn : smol}\ ] ] where @xmath155 and @xmath156 . integrating both sides of the equation over @xmath14 by defining @xmath157 leads to @xmath158 . by setting @xmath159 where @xmath160
, we get @xmath161\overline{c}(r , t ) , \label{eqn : step2}\ ] ] where @xmath162 with @xmath163 . in all likelihood , @xmath164 represents the product of diffusion coefficient @xmath165 associated with barrier crossing dynamics and the contribution of dynamics at the barrier top .
thus , under tension @xmath10 , one can set @xmath166 where @xmath167 and @xmath168 describes the correction due to geometrical information of the cross section of bottleneck @xcite .
therefore , under tension @xmath10 , eq.[eqn : step2 ] becomes eq.(2 ) in the main text .
+ * solution of the smoluchowski equation with time - dependent sink :* for the problem with a constant loading rate , the sink function of our smoluchowski equation becomes time - dependent , resulting in the following equation for the flux @xmath39 , with @xmath170 .
although a time - dependent sink term , in general , makes smoluchowski equations analytically intractable , the ansatz @xmath171 used in the ref .
@xcite allows us to solve the above problem exactly .
substitution of @xmath171 leads to two odes for @xmath172 and @xmath173 ( with @xmath174 denoting derivative with respect to @xmath41 ) , @xmath175 and @xmath176 with @xmath177 .
the equation for @xmath173 in eq.[eqn : mu ] is the riccati equation , @xmath178 with @xmath179 . in general
, the riccati equation can be reduced to a second order ode .
the variable is changed in two steps : ( i ) @xmath180 leads to @xmath181 where @xmath182 and @xmath183 .
( ii ) another substitution @xmath184 leads to @xmath185 , i.e. , @xmath186 introducing the variable @xmath187 with @xmath188 and @xmath189 one can modify the second - order ode in eq.[eqn : riccati ] into a more familiar modified bessel equation ,
@xmath190u=0 \label{eqn : mod_bessel}\ ] ] where @xmath191 . the solution of the above ode is the linear combination of @xmath192 for non - integer @xmath86 , and the linear combination of @xmath84 and @xmath88 when @xmath86 is integer .
thus , the solution of eq.[eqn : mod_bessel ] is @xmath193 for simplicity , we use the notation @xmath194 to represent either @xmath195 or @xmath88 , @xmath196 thus one obtains @xmath173 using @xmath197 .
@xmath198 note that @xmath199 .
the initial condition @xmath177 determines the constant @xmath200 in eq.[eqn : mu ] @xmath201^{-1}i_{\beta}(\rho_0)}{[\kappa(0)]^{-1}\mathcal{q}_{\beta}(\rho_0)-\mathcal{q}_{\beta}'(\rho_0)}.\ ] ] where @xmath202 .
thus , one obtains @xmath203
\label{eqn : mu_main}\end{aligned}\ ] ] where @xmath204^{-1}\left\{i_{\beta}(\rho_0)\mathcal{q}_{\beta}(\rho)-\mathcal{q}_{\beta}(\rho_0)i_{\beta}(\rho)\right\}$ ] .
recall that @xmath79 with @xmath205 , @xmath189 , and @xmath82 .
note that @xmath206 is satisfied .
integration of eq.[eqn : nu_org ] with @xmath41 using eq.[eqn : mu_main ] and change of variable @xmath207 results in the expression for @xmath172 : @xmath208 with @xmath173 ( eq.[eqn : mu_main ] ) and @xmath172 ( eq.[eqn : nu ] ) in hand , we can solve the @xmath5-dependent unbinding time distribution @xmath211 are obtained from the relation @xmath212 , @xmath213 \left[\frac{\mathcal{i}(\rho)}{\mathcal{i}(\rho_0)}\right]^{-1/2}\left[1+\kappa(t)\frac{\mathcal{i}'(\rho)}{\mathcal{i}(\rho)}\right]^{-3/2}. \label{eqn : p}\ ] ] * illustration using synthetic data :*
although @xmath217 in eq.[eqn : p ] is complicated , the familiar expression used in the dynamic force spectroscopy ( dfs ) for @xmath4 is restored when @xmath92 ( see below ) . in order to obtain insight into the behavior of @xmath217 we generated several synthetic distributions for varying @xmath5 values and loading rates .
we find that @xmath217 with varying @xmath218 shows the standard pattern of force distribution in dfs ( fig.[dfs_fig]-a , b ) @xcite .
the effect of varying @xmath5 on @xmath217 is shown in fig.[dfs_fig]-c , d. it is of particular interest that if @xmath219 then the most probable forces @xmath220 from @xmath217 are insensitive to the variation in @xmath5 even though the shapes of @xmath221 and @xmath222 are very different from each other ( fig.[dfs_fig]-c ) . however ,
when @xmath223 , @xmath224 changes with @xmath5 ( fig.[dfs_fig]-e ) and the shape of @xmath221 differs from @xmath225 qualitatively ( fig.[dfs_fig]-d ) .
+ * asymptotic behavior at @xmath226 :* to obtain the asymptotic behavior we will use the following uniform asymptotic expansion of the modified bessel function for large orders ( @xmath227 ) @xcite .
the asymptotic behavior at large negative orders can be obtained by using the relation @xmath231 @xmath232 using these asymptotics , we obtain the following relations at @xmath233 . @xmath234\nonumber\\ \lim_{\beta\rightarrow\infty}\mathcal{i}'(\rho)&\sim 2\left(\frac{\sin{\beta\pi}}{\beta\pi}\right)\frac{1}{\kappa(0)\kappa(t)}\left[e^{\beta(\eta-\eta_0)}\left(s(t)+1\right)-e^{-\beta(\eta-\eta_0)}\left(s(t)-1\right)\right ] \ ] ] where @xmath235 .
therefore @xmath236 \label{eqn : i'/i}\ ] ] and @xmath237.\ ] ] with @xmath238 and @xmath239 where @xmath240 , it is now easy to show @xmath241 \label{eqn : mu_asymp}\ ] ] and @xmath242^{-1/2}}. \label{eqn : nu_asymp}\ ] ] thus , substituting eq.[eqn : mu_asymp ] and [ eqn : nu_asymp ] into @xmath243 recovers the previous result for survival probability in zwanzig s fb model @xcite @xmath244^{-1/2}. \label{eqn : survival_zwanzig}\ ] ] for @xmath92 and @xmath107 , @xmath245 and @xmath246 , respectively . +
* survival probability ( @xmath247 ) and rupture force distribution ( @xmath248 ) for @xmath249 and @xmath107 :* for @xmath249 , taking @xmath250 on eq.[eqn : pde ] with pre - averaged rate constant @xmath251 and transforming @xmath41 into @xmath215 , we obtain @xmath252 , which leads to @xmath253}\ ] ] and the rupture force distribution ( @xmath254 ) @xmath255 the most probable force ( @xmath256 ) is obtained using @xmath257_{\tilde{f}=\tilde{f}^*}=0 $ ] , which is equivalent to @xmath258_{\tilde{f}=\tilde{f}^*}=[k(\tilde{f})]^2_{\tilde{f}=\tilde{f}^*}\theta$ ] .
using @xmath259 , one can easily show that @xmath260}. \label{eqn : f_lambda_infinity}\ ] ] this expression is equivalent to the standard dfs theory except for the presence of the @xmath261 term .
the fast variation of @xmath13-coordinate effectively modifies the reactivity @xmath262 into @xmath113 . for @xmath107 the bottleneck radius
is quenched to a single value , say , @xmath263 . in this case
the noise - averaged probability of the molecule found at the configuration of @xmath263 at force @xmath215 , @xmath264 , should be weighted by @xmath265 $ ] as @xmath266 to give the survival probability , @xmath267^{-1/2}.\ ] ] a similar procedure as in eqs .
[ eqn : pf_lambda_infinity ] and [ eqn : f_lambda_infinity ] leads to @xmath268 ^ 3\ ] ] and @xmath269 + * comparison between @xmath225 and @xmath221 one - dimensional models :* asymptotic behaviors of @xmath4 with two limiting @xmath5 values at large @xmath270 , @xmath225 and @xmath221 are obtained by using the bell model for @xmath271 .
comparison between @xmath225 and @xmath221 can be made by using the explicit form of @xmath259 .
@xmath272}\ ] ] and @xmath273^{-3/2}. \label{eqn : oned1}\ ] ] for @xmath274 , @xmath248 behaves as @xmath275 it is worth noting that depending on the @xmath5 value ( @xmath92 or 0 ) @xmath217 differs in its asymptotic behavior with respect to @xmath215 ( see eqs.[eqn : oned1 ] and [ eqn : oned2 ] ) .
the asymptotic behavior of the so - called microscopic model @xcite , whose force range is limited by the critical force ( @xmath276 , is reduced to that of gumbel distribution only if @xmath277 .
if @xmath278 then the unbinding force distribution decays precipitously to zero as @xmath279(@xmath280 : cubic potential ) and linearly ( @xmath281 : harmonic cusp potential ) ( @xmath249 corresponds to the bell model ) .
note that the model in @xcite corresponds to @xmath249 .
in contrast , for @xmath282 , @xmath283}\nonumber\\ \lim_{\tilde{f}\rightarrow 0}p_{\lambda\rightarrow 0}(\tilde{f } ) & \sim \frac{k_0\theta}{\tilde{\gamma}}\times\left[1+\left(1 - 3\frac{k_0\theta}{\tilde{\gamma}}\right)\tilde{f}+\mathcal{o}(\tilde{f}^2)\right].\end{aligned}\ ] ] the initial slope of @xmath248 is determined by the value of @xmath284 . | heterogeneity in biological molecules , resulting in molecule - to - molecule variations in their dynamics and function , is an emerging theme . to elucidate the consequences of heterogeneous behavior at the single molecule level
, we propose an exactly solvable model in which the unfolding rate due to mechanical force depends parametrically on an auxiliary variable representing an entropy barrier arising from fluctuations in internal dynamics . when the rate of fluctuations @xmath0 a measure of dynamical disorder @xmath0 is comparable to or smaller than the rate of force - induced unbinding , we show that there are two experimentally observable consequences : non - exponential survival probability at constant force , and a heavy - tailed rupture force distribution at constant loading rate . by fitting our analytical expressions to data from single molecule pulling experiments on proteins and dna ,
we quantify the extent of disorder .
we show that only by analyzing data over a wide range of forces and loading rates can the role of disorder due to internal dynamics be quantitatively assessed .
complex systems , characterized by processes that occur over a wide spectrum of time and length scales , often exhibit heterogeneous behavior .
spin glasses with quenched randomness @xcite and structural glasses in which randomness is self - generated @xcite are two classic examples where heterogeneity is indicated by the violation of the law of large numbers @xcite .
these systems exhibit sub - sample to sub - sample variations in measurable quantities . for biological systems
, it is increasingly becoming appreciated that there are cell - to - cell variations ( on length scales @xmath1 @xmath2 m ) @xcite as well as molecule - to - molecule variations ( @xmath1 nm scales ) .
manifestation of heterogeneous behavior on the @xmath2 m length scale is easier to fathom than on the molecular scale . in a pioneering study , evidence for disorder in enzymes
was presented using single molecule experiments @xcite .
two recent studies @xcite have further established that on the nm scale biological molecules display heterogeneity and broken ergodicity just as found in cells @xcite and glasses @xcite .
time traces generated using single molecule fluorescence energy transfer experiments on the holliday junction , a mobile junction of four dna strands involved in exchange of genetic information , showed that the conformational space is partitioned into disjoint basins of attraction @xcite .
interconversions between the basins do not occur on long time scales unless the system is annealed by first reducing the concentration of mg@xmath3 ions for a period of time , and then increasing mg@xmath3 .
more recently , it has been demonstrated that the speed of the dna unwinding motor recbcd varies from one molecule to another .
persistent heterogeneity in speed can be reset " by an annealing protocol , which involves depleting and then reintroducing mg@xmath3-atp to the enzyme @xcite .
both these experiments on unrelated systems show that there must be intrinsic disorder between chemically identical molecules
. if the dynamical variations from molecule - to - molecule in the holliday junction and recbcd helicase are due to disorder , then it should be possible to discern the consequences in single molecule pulling experiments , which probe the response of proteins and nucleic acids to mechanical force .
previously , such experiments have been particularly useful in directly measuring some features of the folding landscape @xcite of biological molecules that are difficult to access by other methods . here , we show two signatures of molecular disorder : deviations from exponential kinetics in force - induced unfolding of proteins @xcite , and the presence of fat tails in the distribution of rupture forces , @xmath4 , characterizing the unzipping of dna @xcite .
these features can not be explained using standard theories , which involve crossing a one - dimensional free energy barrier in the presence of force .
instead , we propose a generic mechanism , based on a model coupling molecular disorder and function . as an illustration ,
consider the unbinding ( or binding ) kinetics of a ligand from a receptor molecule , where the dynamics depends on the time - varying conformation of the receptor ( open or closed ) @xcite .
the ligand is more tightly bound in the closed than in the open conformation .
depending on the gating rate @xmath5 ( the frequency of transitions between the conformations ) the ligand is expected to exhibit very different unbinding kinetics .
if @xmath6 is the mean rate of unbinding , and @xmath7 or @xmath8 , the environment appears _
static _ to the ligand @xcite .
the ligand experiences either _ quenched disorder _ ( @xmath7 ) , unbinding via parallel paths over a spectrum of multiple barriers , or _
annealed disorder _
( @xmath8 ) , unbinding via a single path over a rapidly averaging barrier . if @xmath9 , the gating produces a fluctuating environment along the dynamic pathway of the ligand and affects the unbinding process in a non - trivial fashion .
this regime is often termed _ dynamic disorder _ @xcite .
the gating mechanism has been extensively studied in both experiments and theories in the context of oxygen binding to myoglobin @xcite .
the presence of dynamical disorder in the oxygen - myoglobin system results in a power law decay of unreacted oxygen and a fractional order dependence of binding rate constant on solvent viscosity @xcite . to account for the origin of this phenomenon , zwanzig proposed a fluctuating bottleneck ( fb ) model @xcite , which considers a rate process controlled by passage through a bottleneck whose cross - sectional area , responsible for the reactivity , undergoes stochastic fluctuations . while the frequency @xmath5 governing the internal dynamics , which is intrinsic to a molecule , can in principle
be varied to a certain extent by changing viscosity @xcite , the unbinding rate @xmath6 can be more easily altered by changing force @xmath10 or the corresponding most probable force @xmath11 under the constant loading rate condition in single - molecule pulling experiments , thus providing a way to infer dynamic disorder . here , we adopt zwanzig s fb concept as a general mechanism for probing the internal disorder in biological molecules , with explicit experimental consequences . by fitting our analytical expressions to single - molecule force data ,
we extract a measure of dynamic disoder in proteins and dna . to model the effect of mechanical force on the dynamics of crossing a free energy barrier in the presence of molecular gating , we modified zwanzig s fb model @xcite using an effective potential @xmath12 that depends parametrically on @xmath13 , the auxiliary variable characterizing the internal dynamics , and explicitly on the molecular extension @xmath14 conjugate to the applied force , @xmath10 @xcite .
the fb model is governed by two langevin equations of motion : @xmath15 where @xmath16 is the friction coefficient along @xmath14 .
the precise functional form of @xmath17 is arbitrary except it should have a local minimum corresponding to a bound ( folded ) state at @xmath18 , separated by a free energy barrier at @xmath19 from the unbound ( unfolded ) ensemble at large @xmath14 .
the variable @xmath13 is the dimensionless bottleneck radius , imposed with a reflecting boundary condition at @xmath20 to satisfy @xmath21 @xcite .
both the noise - related random force @xmath22 along @xmath14 and @xmath23 , the stochastic fluctuation of @xmath13 , satisfy the fluctuation - dissipation theorem : @xmath24 and @xmath25 , with @xmath26 being the thermal energy , and @xmath27 .
forced - unbinding occurs on first passage from @xmath28 to @xmath29 , with a rate @xmath30 that in general varies with both @xmath10 and @xmath13 . in traditional models of barrier crossing
, there is no coupling between reaction dynamics in @xmath14 and other degrees of freedom , so @xmath31 only depends on @xmath10 .
for example , in the bell approximation @xmath32 , where @xmath33 . in the fb model , the coupling to @xmath13
is incorporated by making the reaction sink proportional to the area of the bottleneck , @xmath34 .
the form of @xmath30 is physical for the applications here because the rate of unfolding of proteins or unzipping of dna should increase as the solvent accessible area ( @xmath35 ) increases . for simplicity ,
we assume the force - dependence is described by the bell approximation , @xmath36 , though the calculations below can be generalized to more complicated models where @xmath37 reflects movement of the transition state under force @xcite . the langevin equations in eq.[eqn :
fb_model ] can be translated into the following smoluchowski equation ( see supporting information ( si ) for details ) : @xmath38\overline{c}(r , t ) \label{eqn : step2}\ ] ] where @xmath39 is the mean probability of finding the system still bound ( @xmath40 ) with bottleneck value @xmath13 at time @xmath41 , and @xmath42 @xcite . depending on whether @xmath10 is constant or is a linearly varying quantity with time , i.e. , @xmath43 , our problem is classified into unbinding under force - clamp or force - ramp conditions , respectively . _
force - clamp _ : for a constant @xmath10 , eq .
[ eqn : step2 ] for @xmath39 is solved analytically with a reflecting boundary condition at @xmath20 , and an initial condition @xmath44 with @xmath21 .
the resulting survival probability @xmath45 is an extension of the result by zwanzig @xcite in the presence of force , @xmath10 @xcite : @xmath46^{-1/2 } \label{eqn : solution}\end{aligned}\ ] ] where @xmath47 and @xmath48 . in two asymptotic limits of @xmath5 ,
the expression for @xmath49 becomes simple .
( i ) for @xmath50 , we have @xmath51 and the survival probability decays exponentially , @xmath52 , with @xmath53 acting as an effective rate constant .
( ii ) for @xmath54 , we get @xmath55 and the survival probability exhibits a power - law decay , @xmath56 at short times @xmath57 , changing over into an exponential decay with rate @xmath53 at long times @xmath58 . in the limit of quenched disorder , as @xmath59 , the power - law decay extends to all times . using the bell force dependence for @xmath37 in eq .
[ eqn : solution ] , we find @xmath60 , where @xmath61 with @xmath62 .
two limiting conditions arise : ( 1 ) for @xmath63 , @xmath64 changes sign from negative to positive at @xmath65 .
therefore , as @xmath10 is increased , a crossover occurs from an exponential kinetics with @xmath66 ( @xmath67 ) to a power - law kinetics with @xmath55 ( @xmath68 ) .
( 2 ) for @xmath69 , @xmath70 and @xmath55 ; hence the power - law behavior persists at all @xmath10 .
_ force - ramp : _ in constant loading rate experiments , where the external force @xmath71 is ramped at a fixed rate @xmath72 , @xmath37 in eq .
[ eqn : step2 ] is time - dependent , i.e. , @xmath73=k_0e^{\tilde{\gamma } t}$ ] where @xmath74 .
thus , @xmath39 satisfies a smoluchowski equation with a time - dependent sink , @xmath75 , with an initial condition @xmath76 , and a reflecting boundary condition @xmath77 .
the survival probability at time @xmath41 in this case can be analytically computed ( see si for details of derivation ) , leading to the result @xmath78^{-1/2}\left[1+\kappa(t)\frac{\mathcal{i}'(\rho)}{\mathcal{i}(\rho)}\right]^{-1/2 } , \label{eqn : surv}\end{aligned}\ ] ] where @xmath79 , @xmath80 , @xmath81 , @xmath82 , and @xmath83^{-1}\left\{i_{\beta}(\rho_0)\mathcal{q}_{\beta}(\rho)-\mathcal{q}_{\beta}(\rho_0)i_{\beta}(\rho)\right\}$ ] . here
@xmath84 is a modified bessel function of the 1st kind , @xmath85 when @xmath86 is not an integer , and @xmath87 , when @xmath86 is an integer , where @xmath88 is a modified bessel function of the 2nd kind .
an analytical but rather complicated expression for the rupture force distribution @xmath89 is obtained from @xmath90 ( see si ) .
as will be discussed below , the distributions @xmath89 for finite @xmath5 exhibit fat tails at large @xmath10 , a consequence of disorder . by fitting the theoretical expression for @xmath91 ( eq .
s19 in the si ) to measured force distribution data , one should be able to quantify the extent of disorder in terms of @xmath5 . in the two limits of @xmath92 and @xmath93
, one can obtain explicit expressions for the most probable rupture force @xmath11 as a function of @xmath94 : for @xmath95 , @xmath96}$ ] , while for @xmath59 , @xmath97 ( see si ) .
the latter equation is valid when @xmath98 . for @xmath59 and @xmath99 , @xmath100
is peaked at @xmath101 .
the presence of finite probability at @xmath101 is due to unfolding through spontaneous transitions @xcite .
the difference in @xmath102 between the two asymptotic limits of @xmath5 is maximized when @xmath103 ( illustrated in fig .
s1e using synthetic data ) , but disappears for sufficiently high loading rates , @xmath104 . in conventional dynamic force spectroscopy ( dfs )
theory , @xmath11 is linear in @xmath105 .
however , a positive curvature often develops when the force response of the molecule is ductile and hence the transition state location moves towards the bound state @xcite .
we predict that if a system has disorder , negative curvature should be discernible at low loading rates in the @xmath11 versus @xmath105 plot especially for @xmath106 when disorder is quenched , i.e. , @xmath107 ( see fig .
s1e ) . besides fitting @xmath108 to force distribution data , a careful investigation of @xmath11 vs @xmath105 plots at small
@xmath72 could be useful to capture the fingerprints of dynamical disorder .
we use our theory , derived from a single model , to analyze two representative sets of force data , one from a force - clamp and one from a force - ramp experiment .
pn ( digitized from fig.1 in ref .
@xcite ) .
the line is the fit using eq .
[ eqn : solution ] .
the inset shows @xmath109 using a logarithmic scale .
( b ) ( top ) solid lines show our theoretical fits to the survival probability data ( colored circles obtained by digitizing the results in fig.2 in ref.@xcite ) at different values of the force @xmath110 pn .
the extracted parameters @xmath53 , @xmath5 , and their ratio @xmath111 are plotted against @xmath10 on the two panels at the bottom .
[ kuo_fit],width=288 ] pn / nm ) @xcite using three different models .
the fits using @xmath91 ( solid lines ) , based on our fb model , yield ( @xmath112 [ nm ] , @xmath113 [ s@xmath114 , @xmath5 [ s@xmath114)=(1.1 , 0.017 , @xmath115 ) for @xmath116 nm / s and ( 0.66 , 0.99 , 0.48 ) for @xmath117 nm / s .
the fits using @xmath118=\frac{k(\varepsilon)}{\gamma}\exp{\left[\frac{k_0}{\gamma\delta x^{\ddagger}}\left(1-\frac{k(\varepsilon)}{k_0}\varepsilon^{-1/2}\right)\right]}$ ] ( dashed lines ) with @xmath119 and @xmath120 @xcite yield @xmath121 [ nm ] , @xmath122 @xmath123 $ ] , @xmath124 [ pn@xmath125nm])=(0.12 , 0.12 , 31.9 ) for @xmath116 nm / s and ( 0.09,7.16,19.0 ) for @xmath117 nm / s .
the fits using @xmath126 ( dotted lines ) yield @xmath121 [ nm ] , @xmath122 @xmath123$])=(0.49 , 0.13 ) for @xmath116 nm / s and ( 0.32 , 8.19 ) for @xmath117 nm / s .
note that @xmath127 describes the heavy tails of the distributions better than @xmath128 or @xmath129 , implying that unbinding of these dna duplexes by force can not be accounted for using one - dimensional models .
[ analysis],width=268 ] _ polyubiquitin stretching : _ the first is polyubiqutin stretched at constant @xmath10 in afm , where the survival probabilities @xmath109 , obtained from dwell time analysis , exhibited non - exponential decay @xcite .
the experimental data were further interpreted using a gaussian distribution for free energy barriers and transition state locations separating the folded and unfolded states of ubiquitin @xcite .
we find that the physics of the measured non - exponential decay can be quanitatively explained using our theory ( eq . [ eqn : solution ] ) based on the fb model .
the almost exact fit of the theory to experiment for @xmath49 at @xmath130 pn in fig.[kuo_fit](a ) shows the presence of dynamical disorder . for polyproteins ( used in afm experiments )
there are two possible origins for disorder : ( i ) couplings between neighboring modules ; ( ii ) disorder inherent to each module .
because a simple statistical relationship such as binomial factorization associated with the kinetics of individual modules @xcite is expected to break down for the scenario ( i ) , experiments that change the number of modules can discriminate between the two scenarios .
regardless of the origin of heterogeneity , our theory can be used to extract the parameters characterizing disorder effects .
if the bottleneck represents molecular gating or breathing dynamics in a multimodular construct of polyubiquitin , @xmath5 should in principle be an increasing function of @xmath10 . as the applied tension increases , the amplitude of transverse fluctuations decreases , hence increasing the corresponding frequency . even in bell model
, the extracted parameters @xmath112 and @xmath122 should be interpreted as capturing the geometry of the landscape , which changes with @xmath10 @xcite . from the fits of @xmath131 at different values of force we extracted @xmath53 and @xmath5 at each @xmath10 ( bottom of fig.[kuo_fit](b ) ) .
the effective rate constant @xmath53 changes exponentially with @xmath10 , in accord with bell model . by using @xmath132
we obtain @xmath133 nm and @xmath134 s@xmath135 . the value of @xmath112 is in excellent agreement with experimental measurements @xcite .
interestingly , we also find a rough exponentially increasing trend in @xmath5 with @xmath10 .
increase of @xmath5 implies that the rate of change of accessible surface increases as @xmath10 increases , supporting our physical intuition about the influence of @xmath10 on the internal dynamics .
a nonexponential @xmath109 in the force - clamp condition corresponds to a heavy tailed @xmath4 in the force - ramp condition .
the parameters determined for polyubiquitin satisfy @xmath136 for all @xmath10 values ( fig.[kuo_fit](b ) ) .
therefore , a heavy tailed @xmath4 will manifest itself over the entire range of @xmath137 pn / s that gives rise to the most probable forces @xmath138 pn .
_ unbinding of dna duplexes : _ the effects of disorder manifest themselves dramatically in unzipping experiments on dna duplexes between 5-ggctcccttctaccactgac-3 and 3--5 , where underlined sequences are complimentary to each other @xcite .
the measured rupture force distributions have heavy tails at high @xmath10 , though the physical reasons for the tails were not discussed in the original paper .
fig .
[ analysis ] shows the unzipping force distribution at two pulling speeds , 8 nm / s and 1600 nm / s , and fits using our model , compared to two other models commonly employed in analyzing dfs experiments .
it is clear that our theory for @xmath108 using the fb model most accurately fits the force data .
two other models , based on the bell model @xcite and a cubic potential @xcite , fail to capture the tail part of the data because they incorporate no disorder , only unbinding through a one - dimensional free energy profile . for 8
nm / s , the gating frequencies are almost zero ( @xmath139 s@xmath135 ) .
however , @xmath5 increases by nearly four orders of magnitude to @xmath140 s@xmath135 at @xmath117 nm / s .
given that the bubble dynamics of a dna duplex occur with a characteristic time scale of @xmath141 @xmath2s @xcite , our extracted @xmath5 values are too large for breathing motion of base pairs to be a source of disorder .
it is more reasonable to surmise that each duplex is pulled from starkly different and very slowly interconverting conformations , similar to that found in holliday junctions and recbcd .
the origin of this disorder is likely to be in the heterogeneity of base pairings , although the experiment was intended to probe the unbinding dynamics from a homogeneous sample made of two dna strands with perfect complimentarity .
the large increase in @xmath5 as @xmath142 increases suggests that tension facilitates interconversion between states , which accords well with the expectation that force lowers barriers between distinct bound states .
applications of our theory reveal that by analyzing data from single - molecule pulling experiments over a range of forces and loading rates one can infer the role that dynamical disorder , intrinsic to the molecule , plays in unfolding or unbinding kinetics .
the observed non - exponential kinetics in survival probability or fat tails in the unfolding force distributions can not be captured by theories based on one - dimensional free energy profiles .
our work shows that by using an auxiliary coordinate in addition to extension we can quantitatively predict the consequences of disorder in the dynamics of biological molecules .
the theory provides a conceptual framework for analyzing future single molecule pulling experiments on complexes involving proteins , dna , and rna in which heterogeneity is sure to play a prominent role .
we thank shaon chakraborty for useful discussions .
c.h .
thanks korea institute for advanced study for providing computation resources .
d.t . acknowledges a grant from the national institutes of health ( gm 089685 ) . 31 natexlab#1#1bibnamefont # 1#1bibfnamefont # 1#1citenamefont # 1#1url # 1`#1`urlprefix[2]#2 [ 2][]#2 , , , _ _ ( , ) . ,
* * , ( ) . ,
* * , ( ) . , , , * * , ( ) . , * * , ( ) . , * * , ( ) .
, , , , , , , , , , , * * , ( ) . , , , * * , ( ) .
, , , , , * * , ( ) . , , , * * , ( ) .
, , , , , , * * , ( ) . , * * , ( ) . , * * , ( ) . , , , , , * * , ( ) . , , , * * , ( ) . , , , , , , , , ,
* * , ( ) . , , , , * * , ( ) . ,
* * , ( ) . ,
* * , ( ) .
, , , , , , , , , * * , ( ) . , , , , ,
* * , ( ) . ,
* * , ( ) . , , , * * , ( ) . , * * , ( ) . , * * , ( ) . , * * , ( ) . , , , , , * * , ( ) . , * * , ( ) . , , , * * , ( ) . , eds . , _ _ ( , ) . ,
* * , ( ) . |
the euv imaging spectrometer ( eis ) on hinode uses multilayer optics to produce stigmatic spectra in two 40 wide euv bands centered at 195 and 270 .
emission lines present in these bands originate in the solar transition region and corona , permitting detailed study of the temperature , density , and dynamical structure of these regions .
the hinode mission is described in @xcite .
an overall description of eis is available in @xcite .
@xcite provides more detailed information on the eis calibration , and @xcite provides details of the optics and mechanisms .
eis can image the sun using 1 and 2 slits and 40 and 266 slots . the latter result in monochromatic images of the sun in the stronger spectral lines . obtaining the full diagnostic capability of the instrument ,
however , is only possible when data are acquired with the 1 or 2 slits .
this means that images must be constructed by moving the eis mirror to raster an area of interest .
there is , of course , a scientific tradeoff to be made for constructing a spectroheliogram in this manner it can take considerable time .
for an exposure time of 30 s , a typical value for capturing both the strongly- and weakly - emitting portions of an active region , it can take an hour or more to image a substantial area of the disk .
when high time cadence is important , a better approach is to use eis in a sit - and - stare mode . in this case , the 1 or 2 slit is placed at one location on the sun and repeated exposures are obtained while the hinode spacecraft tracks the solar rotation so that the slit always samples the same solar plasma .
these kinds of observations can be enhanced by preceding and/or following them with eis spectroheliograms , and are , of course , further enhanced by using imaging data from the hinode x - ray telescope ( xrt ) to monitor the overall evolution of the area being examined with the sit - and - stare observation .
details of the capabilities of the xrt on hinode are available in @xcite . in this paper , we present some initial results from one eis sit - and - stare data set .
the focus here is on small fluctuations in an active region .
@xcite discuss additional eis dynamical observations during the gradual phase of an x - class flare .
the observations discussed in this contribution cover portions of noaa active region 10930 and were obtained on 2006 december 14 , when it was located at approximately s05w39 .
eis spectroheliograms covering portions of the active region were obtained beginning at 3:23 and 19:20 ut .
each spectroheliogram consisted of 256 30 s exposures with the 1 slit and a spatial step size between exposures of 1 , taking approximately 2 hr 23 m to complete . a 120-exposure eis sit - and - stare observation within the region
was begun at 17:00 ut .
that study used 60 s exposures . for both the spectroheliograms and the sit - and - stare observations ,
nine data windows were selected on the eis detectors .
this paper only presents results for the lines in six of those windows .
each window in the spectroheliograms was 24 spectral pixels wide and covered a 256 region of the sun in the n / s direction .
for the sit - and - stare data set , each window was 32 spectral pixels wide , but the slit covered a 512 region in the n / s direction .
each eis spectral pixel is 22.3 m wide .
the measured fwhm at 185 is 47 m @xcite .
table [ table : lines ] lists the emission lines included in this study and their temperatures of formation .
.emission lines observed with eis [ cols="<,<,<",options="header " , ] figure [ fig : xrt_full ] shows an xrt image of the active region taken at 19:19:52 ut with a box marking the region covered by the eis spectroheliogram that followed the sit - and - stare observation and a vertical line showing the position of the eis slit during the sit - and - stare observation .
the image was obtained with the xrt be_thin / open filter combination , which has peak sensitivity at a temperature of roughly 10 mk and is two orders of magnitude less sensitive at about 2 mk @xcite .
thus , the image is mostly dominated by emission at higher temperatures than those seen in most of the eis emission lines used in this study .
( 2.70in,2.70in)mariska_fig1.eps all the data were processed using the current version of the eis data preparation software .
this removed the detector bias and dark current , as well as hot pixels and cosmic ray hits , and then applied the eis absolute calibration , yielding intensities in ergs @xmath1 s@xmath0 sr@xmath0 @xmath0 .
the software replaces the hot pixels and those affected by cosmic ray hits with interpolated values from adjoining pixels , but the data at those locations were flagged as bad and the line fitting software gave those locations no weight in the subsequent line profile fitting process .
the data on the long - wavelength detector were also shifted in the y - direction to account for the roughly 17 pixel offset between the two detectors .
figure [ fig : spectroheliogram ] shows the appearance of the active region in the eis spectroheliograms taken immediately after the sit - and - stare study .
the vertical line on each panel shows the location of the eis slit during the sit - and - stare observation . as the images show , the portions of the active region imaged by eis had a complex structure .
examination of movies made from xrt images of the active region shows that considerable flare - like activity was taking place throughout the time of the eis observations .
data from the goes x - ray monitors also show significant flare - like fluctuations at the b to c level , and an x - flare took place later in the day .
the bright , compact loops in the xrt image are the result of a small flare that took place before the eis slit reached those positions . by the time the eis slit
reached those positions , the strong emission had decayed and the loop morphology is less evident .
( eis spectroheliograms are constructed by scanning the slit from west to east on the sun . )
for the sit - and - stare observation , the emission lines in the six wavelength bands shown in figure [ fig : spectroheliogram ] were all fitted with single gaussian line profiles .
thus , for each spectral line at each location along the eis slit , we have measured values of the total line intensity , the location of the line center , and the line width .
figure [ fig : sns - image ] shows a greyscale representation of the resulting total intensity in each spectral line as a function of time for all 512 positions along the slit .
a number of locations show obvious evidence for rapid time variations in the total intensity .
comparison with the eis spectroheliograms in figure [ fig : spectroheliogram ] shows that these locations appear to correspond generally to areas near the footpoints of loops .
examination of movies made from the xrt images and those from the euv imaging telescope ( eit ) on the solar and heliospheric observatory ( soho ) suggest that these are due to intermittent heating events in some of the loops crossed by the eis slit .
( 6.3in,4.5in)mariska_fig3.eps extracting doppler - shift information is a somewhat complex undertaking .
the eis slit is not perfectly vertical on the camera ccds , so a correction for the slit tilt must be made .
in addition , over the course of each orbit , eis exhibits a periodic variation in the measured line center position for all spectral lines observed in both detectors .
the peak - to - peak variation is about 2 spectral pixels a significant doppler signal at the wavelengths covered by eis .
the cause of this variation is still under investigation , but appears to be due to temperature changes within the instrument .
the slit tilt does not appear to change with time and results in a difference in the location of the rest wavelength for an emission line of approximately 0.37 spectral pixels from the bottom to the top of the 512 slit length used in the sit - and - stare observations . while this is small , it represents a velocity change of about 13 km s@xmath0 at the fe xii 195.12 emission line and
thus is important for detailed dynamical studies .
the slit tilt was removed from the sit - and - stare data set by shifting the fitted line center position in each spectral line by an amount determined using an analysis of doppler - shift data taken early in the mission . to a first approximation
, the orbital variation is the same at each location on the two eis detectors .
thus , it is possible to use the data in a high - signal - to - noise channel during the sit - and - stare observation to determine the orbital variations for all the wavelengths .
to do this we use the top 112 rows of fitted line center positions ( corrected for slit tilt ) in the fe xii 195.12 emission line data to determine an average orbital variation .
this orbital variation is then subtracted from the line center positions measured in all the wavelength windows .
this particular region was selected because examination of the xrt and eit data indicates that no significant dynamical activity is present at these locations .
note that eis has no absolute wavelength reference .
once the orbital trend has been removed using the fe xii 195.12 emission line data , the average value of the line center position in the top 112 rows in each of the six emission lines used in this study was used as the rest wavelength for that emission line .
figure [ fig : sns - doppler ] shows a greyscale representation of the fitted line center positions with the orbital variations removed as a function of time for all 512 positions along the slit for the six lines included in this study . in these images
darker shades represent smaller values for the doppler shift ( blueshifts ) . if one excludes data points that are obvious outliers because of poor gaussian fits , the range of doppler - shift velocities seen in the panels is on the order of 20 to 100 km s@xmath0 .
examination of the sit - and - stare images shown in figures [ fig : sns - image ] and [ fig : sns - doppler ] shows that all the significant features shown along the slit location in the the spectroheliograms shown in figure [ fig : spectroheliogram ] are present in the sit - and - stare data .
near the top of this @xmath2-range ( roughly @xmath3 to @xmath4 , there are nearly constant intensity bands in all the emission lines .
examination of the spectroheliograms and especially of a movie made from the eit 195 images shows that these are quiescent loops interconnecting portions of the active region .
the doppler - shift sit - and - stare data at these locations show little or no time - dependent fluctuation in any of the emission lines observed with eis .
in contrast , the @xmath2-range from @xmath5 to @xmath6 shows considerable time - dependent behavior in both the line intensity and the doppler shift .
while the intensity variations are visible in all the emission lines , there is a clear tendency for the doppler shifts to become more pronounced as the temperature of formation of the emission line increases .
figure [ fig : sns - expanded ] shows greyscale images of the fe xii and fe xv intensity and doppler - shift data along the portion of slit where considerable activity is taking place .
some features are clearly impulsive , such as the upflow at @xmath7 and 10 min and a second at @xmath8 and 15 min seen in the fe xii and fe xv lines .
more gradual changes in the doppler shifts are also evident .
for example , in both the fe xii and fe xv doppler images in the vicinity of @xmath8 and 60 min , the redshift velocities appear to be changing gradually over 20-min and longer timescales .
figure [ fig : sns-185 ] shows intensity and doppler - shift data in the fe xii and fe xv emission lines as a function of time at a solar @xmath2-position of approximately @xmath9 .
the graphs quantify the impression given in figure [ fig : sns - expanded]an impulsive event of some kind has taken place causing a relative blueshift followed by a relative redshift . in the doppler - shift plots and to a lesser degree in the intensity plots , the events appear to be superimposed on a more smoothly varying background .
( 6.3in,4.5in)mariska_fig4.eps ( 3.15in,3.15in)mariska_fig5.eps ( 3.15in,2.88in)mariska_fig6.eps the doppler shifts in both the fe xii and fe xv emission lines appear to show roughly one complete oscillation . in the intensity plots , the full oscillation is less obvious , possibly because the signal is superimposed on the more gradual trend present in the data .
measuring the time between the maximum blueshift and the maximum redshift in each line and assuming that it corresponds @xmath10 the oscillation period , results in oscillation periods of 8.2 min for the fe xii emission line and 10.2 min for the fe xv emission line .
given the one - minute cadence of the sit - and - stare observation , these numbers should be considered roughly equal .
measuring the time between the intensity peak and the following local minimum and assuming that it corresponds to @xmath10 the oscillation period results in oscillation periods of 14.4 and 16.4 min for the fe xii and fe xv emission lines , respectively .
thus , the intensity fluctuations do not have the same period as the doppler - shift fluctuations .
note also that the peak in the intensity oscillation appears to be later than the maximum blueshift .
the characteristics of the intensity and doppler - shift oscillations show in figure [ fig : sns-185 ] are very similar to those observed for example by @xcite and @xcite using the sumer spectrometer on soho .
@xcite reported average oscillation periods of @xmath11 min with a range from 8.6 to 23.3 min , comparable to the oscillation shown in the figure .
they reported a wide range of doppler shift amplitudes ranging from 12 to 353 km s@xmath0 , with an average value of @xmath12 km s@xmath0 .
we find amplitudes of roughly 6.5 and 25.8 km s@xmath0 for the fe xii and fe xv emission lines , respectively , within the range seen in the sumer observations .
note that for the most part the doppler shift velocities are quite small , and one might ask whether eis is capable of measuring such small velocity changes . figure [ fig : sns-185-profile ] shows sample fe xii and fe xv spectral data taken near the peak in the doppler - shift plots along with the best - fit gaussians .
the one-@xmath13 uncertainties on the fits are 0.44 and 0.63 km s@xmath0 for the fe xii and fe xv data , respectively .
this level of accuracy is achievable with eis whenever the signal level is high and a gaussian represents a good functional form for the line profile .
the sit - and - stare observations presented in this paper and other similar eis active region observations clearly show an abundance of dynamical phenomena in active regions , much of it periodic .
cotemporal xrt imaging data often show that these phenomena are the dynamical manifestation of disturbances that can be seen following field lines that interconnect portions of the active region .
thus , careful analysis of combined eis and xrt data sets should provide an improved picture of how active regions evolve .
ultimately , the addition of chromospheric images and vector magnetograms from the solar optical telescope on hinode should make it possible to begin to connect fully changes in the corona with the evolution of the underlying magnetic field .
the particular example shown in figure [ fig : sns-185 ] is similar to oscillatory phenomena observed in high - temperature emission lines with the sumer spectrometer on soho .
the intensity signal shown near @xmath9 in figure [ fig : sns - expanded ] does begin with a fairly low intensity , and one might speculate that higher temperature plasma in emitting at that time at that location . the data set we have analyzed for this study does not include the hotter line of fe xvii at 204.65 available in the eis bandpass , and , while we do have data for the ca xvii emission line at 192.82 , the line suffers from a blend with a cooler line of fe xi , which can only be removed with information from other fe xi lines .
thus , a full comparison with sumer results must await the analysis of additional data .
many of these events are thought to be standing slow - mode waves , the detailed study of which should lead to a better understanding of heating in coronal loops . because eis is a more sensitive instrument than sumer , shorter exposure times are possible , which should allow us to expand the range of periods that can be investigated .
moreover , the eis wavelength bands contain lines whose ratios are sensitive to coronal electron densities , leading to better constraints on the characteristics of the oscillations . extracting subtle dynamical information from eis data
does , however , require both carefully planned observations and an analysis that pays attention to all the possible instrumental effects . many of the instrumental effects are just beginning to be understood ,
so it may be some time before it is possible to fully realize the potential eis offers for more completely characterizing solar coronal dynamics .
hinode is a japanese mission developed and launched by isas / jaxa , with naoj as domestic partner and nasa and stfc ( uk ) as international partners .
it is operated by these agencies in cooperation with esa and nsc ( norway ) .
we are grateful to the entire hinode team for all their efforts in the design and operation of the mission .
jtm , hpw , iuu , and dhb acknowledge support from the nasa hinode program . | the euv imaging spectrometer ( eis ) on the hinode satellite is capable of measuring emission line center positions for gaussian line profiles to a fraction of a spectral pixel , resulting in relative solar doppler - shift measurements with an accuracy of less than a km s@xmath0 for strong lines .
we show an example of the application of that capability to an active region sit - and - stare observation in which the eis slit is placed at one location on the sun and many exposures are taken while the spacecraft tracking keeps the same solar location within the slit . for the active region
examined ( noaa 10930 ) , we find that significant intensity and doppler - shift fluctuations as a function of time are present at a number of locations .
these fluctuations appear to be similar to those observed in high - temperature emission lines with other space - borne spectroscopic instruments . with its increased sensitivity over earlier spectrometers and its ability to image many emission lines simultaneously
, eis should provide significant new constraints on doppler - shift oscillations in the corona . |
the advance of precision laser spectroscopy of atomic and molecular systems has , over the past decades , been closely connected to the development of experimental techniques such as tunable laser technology @xcite , saturation spectroscopy @xcite , two - photon doppler - free spectroscopy @xcite , cavity - locking techniques @xcite , and ultimately , the invention of the frequency comb laser @xcite , developments to which prof .
theodore hnsch has greatly contributed .
these inventions are being exploited to further investigate at ever - increasing precision the benchmark atomic system the hydrogen atom , evinced by the advance in spectroscopic accuracy of atomic hydrogen measurements by more than seven orders of magnitude since the invention of the laser @xcite .
the spectroscopy of the 1s-2s transition in atomic hydrogen , at @xmath10 relative accuracy @xcite , provides a stringent test of fundamental physical theories , in particular quantum electrodynamics ( qed ) .
currently , the theoretical comparison to precision measurements on atomic hydrogen are limited by uncertainties in the proton charge radius @xmath11 .
the finding that the @xmath11-value obtained from muonic hydrogen spectroscopy is in disagreement by some 7-@xmath12 @xcite is now commonly referred to as the _ proton - size puzzle_. molecular hydrogen , both the neutral and ionic varieties , are benchmark systems in molecular physics , in analogy to its atomic counterpart .
present developments in the _ ab initio _ theory of the two - electron neutral molecule and the one - electron ionic molecule , as well as the respective isotopologues , have advanced in accuracy approaching that of its atomic counterpart despite the increased complexity .
the most accurate level energies of the entire set of rotational and vibrational states in the ground electronic state of were calculated by komasa et al .
an important breakthrough in these theoretical studies was the inclusion of higher - order relativistic and qed contributions , along with a systematic assessment of the uncertainties in the calculation .
recently , further improved calculations of the adiabatic @xcite as well as non - adiabatic @xcite corrections have been performed , marking the steady progress in this field . in the same spirit as in atomic hydrogen spectroscopy ,
the high - resolution experimental investigations in molecular hydrogen are aimed towards confronting the most accurate _ ab initio _ molecular theory . for the and hd@xmath13 ions , extensive efforts by korobov and co
- workers over the years , have recently led to the theoretical determination of ground electronic state level energies at 0.1 ppb accuracies @xcite .
the latter accuracy enables the extraction of the proton - electron mass ratio , @xmath14 , when combined with the recent hd@xmath13 spectroscopy using a laser - cooled ion trap @xcite .
these are currently at lower precision than other methods but prospects exist that competitive values can be derived from molecular spectroscopy .
similarly karr et al .
@xcite recently discussed the possibility of determining @xmath15 using ( or hd@xmath13 ) transitions as an alternative to atomic hydrogen spectroscopy .
even for the neutral system of molecular hydrogen , the determination of @xmath11 from spectroscopy is projected to be achievable , from the ongoing efforts in both calculation @xcite and experiments @xcite .
molecular spectroscopy might thus be posed to contribute towards the resolution of the proton - size puzzle .
in contrast to atomic structure , the added molecular complexity due to the vibrational and rotational nuclear degrees of freedom could constitute an important feature , with a multitude of transitions ( in the ground electronic state ) that can be conscripted towards the confrontation of theory and experiments . from both experimental and theoretical perspectives ,
this multiplicity allows for consistency checks and assessment of systematic effects . in recent years
, we have tested the most accurate quantum chemical calculations using various transitions , for example , the dissociation limit @xmath16 or binding energy of the ground electronic state @xcite ; the rotational sequence in the vibrational ground state @xcite ; and the determination of the ground tone frequency @xmath17 @xcite .
the comparisons exhibit excellent agreement thus far , and have in turn been interpreted to provide constraints of new physics , such as fifth - forces @xcite or extra dimensions @xcite .
recently , we reported a precision measurement on highly - excited vibrational states in @xcite .
the experimental investigation of such highly - excited vibrational states probe the region where the calculations of komasa , pachucki and co - workers @xcite are the least accurate , specifically in the @xmath18 range .
the production of excited offers a unique possibility on populating the high - lying vibrational states , that would otherwise be practically inaccessible by thermodynamic means ( corresponding temperature of @xmath19 k for @xmath1 ) . here , we present measurements of level energies of @xmath1 rovibrational quantum states that extend the spectroscopy in ref .
@xcite and that implement improvements , leading to a narrowing of the resonances .
this is achieved by the use of a resonant ionization step to molecular rydberg states , thereby enhancing the detection efficiency significantly .
the enhancement allows for the use of a low intensity spectroscopy laser minimizing the effect of ac - stark induced broadening and shifting of lines .
the ac - stark effect is identified as the major source of systematic uncertainty in the measurements , and a detailed treatment of this phenomenon is also included in this contribution .
the production of excited from the photodissociation of hydrogen sulfide was first demonstrated by steadman and baer @xcite , who observed that the nascent molecules were populated at predominantly high vibrational quanta in the two - photon dissociation of at uv wavelengths . that study used a single powerful laser for dissociation , for subsequent spectroscopy , and to induce dissociative ionization for signal detection .
niu et al .
@xcite utilized up to three separate laser sources to address the production , probe , and detection steps in a better controlled fashion . the present study , targeting @xmath20 levels ,
is performed using the same experimental setup as in ref .
@xcite , depicted schematically in fig .
[ fig : setup ] .
the photolysis laser at @xmath21 nm , generated from the second harmonic of the output of a commercial pulsed dye laser ( pdl ) with rhodamine b dye , serves in the production of by photolysing .
the narrowband spectroscopy laser radiation at around @xmath22 nm is generated by frequency upconversion of the output of a continuous wave ( cw)-seeded pulsed dye amplifier system ( pda ) running on rhodamine 640 dye .
the ionization laser source at @xmath23 nm , from the frequency - doubled output of another pdl ( also with rhodamine 640 dye ) , is used to resonantly excite from the @xmath8 to autoionizing rydberg states to eventually form ions . this @xmath24 resonance - enhanced multiphoton ionization ( rempi ) scheme results in much improved sensitivities compared to our previous study @xcite .
production from h@xmath3s dissociation ; the _ probe _ laser from a narrowband dye amplifier ( pda ) for the @xmath25 spectroscopy transition ; and the _ ionization _ laser for resonant h@xmath26 ion production , subsequently detected as signal .
doppler - free two - photon excitation is facilitated by the sagnac interferometric alignment of the counter - propagating probe beams .
absolute frequency calibration is performed with respect to i@xmath3 hyperfine reference lines , aided by the relative frequency markers from the transmission fringes of a length - stabilized fabry - prot etalon . the frequency offset between the cw - seed and pda pulse output , induced by chirp effects in the dye amplifier ,
is measured and corrected for , post - measurement .
[ cw : continuous wave ; pdl : pulsed dye laser ; @xmath27 : frequency - doubling stage ; tof : time - of - flight region ; mcp : multichannel plates ; pmt : photomultiplier tube ] , scaledwidth=50.0% ] the molecular beam , produced by a pulsed solenoid valve in a source vacuum chamber , passes through a skimmer towards a differentially pumped interaction chamber , where it intersects the laser beams perpendicularly . the probe or spectroscopy laser beam
is split into two equidistant paths and subsequently steered in a counter - propagating orientation , making use of a sagnac interferometer alignment for near - perfect cancellation of the residual first - order doppler shifts @xcite .
moreover , the probe laser beams pass through respective lenses , of @xmath28-cm focal length , to focus and enhance the probe intensity at the interaction volume .
finally , the ionization beam is aligned in almost co - linear fashion with the other laser beams to ensure maximum spatial overlap . to avoid ac - stark shifts during the spectroscopic interrogation , induced by the photolysis laser
( @xmath29 mj typical pulse energy ; @xmath30-ns pulse duration ) , a 15-ns delay between the photolysis and probe pulses is established with a delay line . for a similar reason , the ionization pulse ( @xmath31 mj typical pulse energy ; @xmath30-ns pulse duration ) is also delayed by 30 ns with respect to the probe pulse .
the 1 mj ionization pulse energy is sufficient for saturating the ionization step .
the ions produced in the interaction volume are accelerated by ion lenses , further propagating through a field - free time - of - flight ( tof ) mass separation region before impinging on a multichannel plate ( mcp ) detection system .
scintillations in a phosphor screen behind the mcp are monitored by a photomultiplier tube ( pmt ) and a camera , culminating in the recording of the mass - resolved signals . in the non - resonant ionization step as in ref .
@xcite , predominantly ions were produced and were thus used as the signal channel for the @xmath25 excitation . in contrast , the resonant ionization scheme employed here predominantly produces ions .
in addition to the enhancement of sensitivity , the channel offers another important advantage as it is a background free channel , whereas the channel includes significant contributions from , as well as sh , dissociative ionization products . to avoid dc - stark effects on the transition frequencies ,
the acceleration voltages of the ion lens system are pulsed and time - delayed with respect to the probe laser excitation .
niu et al .
@xcite confirmed the observation of two - photon transitions in various @xmath25 ( @xmath32,10 - 12 ) bands , first identified by steadman and baer @xcite , but only for transitions to the outer @xmath33well of the @xmath8 electronic potential in h@xmath3 .
franck - condon factor ( fcf ) calculations , to assess the transition strengths of the photolysis - prepared levels of @xmath34 to levels in the combined inner ( @xmath35 ) and outer ( @xmath36 ) wells of the @xmath8 double well potential , were performed by fantz and wnderlich @xcite . while niu et al .
@xcite performed precision measurements probing the @xmath37 levels , presently @xmath38 levels are probed .
note that for the excited state two different numberings of vibrational levels exist : one counting the levels in the combined @xmath8 well , and the other counting the levels in the @xmath35 and @xmath36 wells separately
. then @xmath39 corresponds to @xmath40 , while @xmath41 corresponds to @xmath42 . in the following
, we will refer to the vibrational assignments following the @xmath36-well notation .
the fcf for the @xmath43 band , used in ref .
@xcite , amounts to 0.047 @xcite and that of the presently used @xmath44 band amounts to 0.17 , making the latter band s transitions three times stronger .
rempi study .
two - photon doppler - free spectroscopy is applied on the @xmath45 ( 0,11 ) band .
resonant excitation to the @xmath7 state by the detection laser follows the spectroscopic excitation , leading to subsequent autoionization yielding enhanced ion signal .
, scaledwidth=45.0% ] the non - resonant ionization step was the major limitation in @xcite , since this prohibits the spectroscopy to be carried out at sufficiently low probe laser intensities .
due to ac - stark effects , the lines were broadened to more than 1 ghz , while the expected instrumental linewidth is less than 200 mhz .
moreover , at higher probe intensities asymmetric line profiles are observed , reducing the accuracy of the line position determination and ultimately limiting the ac - stark extrapolation to the unperturbed line position . while signal improvement was observed when employing a detection laser in the range between 202 - 206 nm in ref .
@xcite , the enhancement was limited since no sharp resonances were found , indicating excitation to some continuum . for the present study ,
a thorough search for resonances from the @xmath36 state was undertaken .
the @xmath46 and @xmath47 rydberg series , with principal quantum number @xmath48 were identified as potential candidates based on the fcfs for the outer @xmath36-well .
the search was based on reported fcfs for the @xmath49 ( @xmath32,1 ) bands @xcite .
it was further assumed that the fcfs for the @xmath50 electronic systems are comparable to that of @xmath51 , since the potential energy curves for the @xmath52 rydberg states are similar as they all converge to the ionic potential .
note the particular characteristic of the @xmath48 @xmath53 rydberg states , that dissociate to a ground state atom and another with a principal quantum number h(@xmath54 ) , i.e. @xmath55 h(1s ) + h(4f ) ; @xmath56 h(1s ) + h(5f ) ; @xmath57 h(1s ) + h(6d ) @xcite .
the electron configuration changes as a function of the internuclear distance @xmath58 , e.g. the low vibrational levels of the @xmath59 follow a diabatic potential that extrapolates to the @xmath60 limit , and not the @xmath61 dissociation limit at @xmath62 .
this peculiarity is explained by @xmath63-uncoupling @xcite , as the molecule changes from hund s case b ( born - oppenheimer ) to hund s case d ( complete nonadiabatic mixing ) with increasing principal quantum number , corresponding to the independent nuclear motion of the residual ion core and the excited rydberg electron @xmath64 .
such transition in hund s cases occurs at @xmath65 for low rotational quantum numbers @xcite .
the @xmath52 rydberg states decay via three competing channels : by fluorescence to lower @xmath66 electronic configurations ; by predissociation , where the nascent h atom is further photoionized to yield ions ; and lastly , by autoionization to yield ions .
recent analysis of one - photon absorption measurements using xuv synchrotron radiation in the range of @xmath67 nm demonstrated that autoionization completely dominates over the other two competing channels in the case of @xmath68 @xcite . using the @xmath48 level energies of the rydberg levels reported in @xcite ,
the detection laser was scanned in the vicinity of the expected transition energies , where it turned out that transitions to @xmath69 resulted in sufficient signal enhancement .
the maximum fcf overlap for the @xmath49 system is @xmath70 for the ( 8,0 ) band , while the corresponding @xmath71 ( 5,0 ) band only has an fcf of @xmath72 @xcite .
although the ( 8,0 ) band with better fcf could be used , the ionization step is already saturated using the weaker ( 5,0 ) band .
autoionization resonances are shown in fig .
[ fig : rydberg ] and assigned to @xmath73 and @xmath74 lines in the @xmath75 ( 5,0 ) bands , whose widths are in good agreement with the synchrotron data @xcite .
the neighboring resonances of the @xmath73 line in fig .
[ fig : rydberg ] are not yet assigned but is not relevant to the @xmath45 investigation here .
appropriate @xmath76 transitions are used for the ionization of particular @xmath45 two - photon @xmath77 transitions . and @xmath74 resonances of the @xmath75 ( 5,0 ) band which are crucial in enhancing the signal strength of the @xmath45 spectroscopy .,scaledwidth=48.0% ] a representative high - resolution spectrum of the @xmath78@xmath79 ( 0,11 ) @xmath80 transition taken at low probe intensity is displayed in fig .
[ fig : hires_spectra ] .
simultaneous with the spectroscopy , the transmission fringes of the pda cw - seed radiation through a fabry - prot interferometer were also recorded to serve as relative frequency markers , with the free spectral range of fsr=148.96(1 ) mhz .
the etalon is temperature - stabilized and its length is actively locked to a frequency - stabilized hene laser .
the absolute frequency calibration is obtained from the i@xmath3 hyperfine - resolved saturation spectra using part of the cw - seed radiation . for the @xmath80 line in fig .
[ fig : hires_spectra ] , the i@xmath3 @xmath81 transition is used , where the line position of the hyperfine feature marked with an * is @xmath82 @xcite .
the accuracy of the frequency calibration for the narrow transitions is estimated to be 1 mhz in the fundamental or 4 mhz in the transition frequency , after accounting for a factor of 4 for the harmonic up - conversion and two - photon excitation . for sufficiently strong transitions probed at the lowest laser intensities , linewidths as narrow as 150 mhz were obtained .
this approaches the fourier - transform limited instrumental bandwidth of 110 mhz , for the 8-ns pulsewidths at the fundamental , approximated to be gaussian , that also includes a factor two to account for the frequency upconversion .
the narrow linewidth obtained demonstrates that despite of the photodissociation process imparting considerable kinetic energy on the produced , additional doppler - broadening is not observed .
although not unexpected due to the doppler - free experimental scheme implemented , this strengthens the claim that residual doppler shifts are negligible .
( 0,11 ) @xmath80 transition is shown at a probe laser intensity of 280 mw/@xmath83 and detection in the signal channel .
the transmission markers of a length - stabilized fabry - prot etalon ( fsr=148.96 mhz ) are used in the relative frequency calibration , while the hyperfine feature marked with * of the i@xmath3 @xmath81 transition serves as an absolute frequency reference ( * at 16482.83312 @xmath6 ) .
, scaledwidth=50.0% ] since the frequency calibration is performed using the cw - seed while the spectroscopy is performed using the pda output pulses , any cw pulse frequency offset need to be measured and corrected for @xcite .
a typical recording of the chirp - induced frequency offset for a fixed pda wavelength is shown in fig .
[ fig : chirp ] . while the measurements can be done online for each pulse , this comes at the expense of a slower data acquisition speed , and was only implemented for a few recordings in order to assess any systematic effects .
a flat profile of the cw - pulse offset when the wavelength was tuned over the measurement range accessed in this study justifies this offline correction . typical cw pulse frequency
offset values were measured to be @xmath84 mhz in the fundamental , which translates to @xmath85 mhz in the transition frequency .
the sources of uncertainties and the respective contributions are shown in table [ tab : uncertainty ] .
the contributions of each source are summed in quadrature in order to obtain the final uncertainty for each transition .
data sets from which separate ac - stark extrapolations to zero - power were performed on different days , and were verified to exhibit consistency within the statistical uncertainty of 0.0014 .
note that the estimates shown in table [ tab : uncertainty ] are only for the low probe intensity measurements used to obtain the highest resolutions .
the uncertainties of the present investigation constitute more than a factor of two improvement over our previous study in @xcite .
the dominant source of systematic uncertainty is the ac - stark shift and is discussed in more detail in the following section .
in the perturbative regime , the leading - order energy level shift @xmath86 of a state @xmath87 induced by a linearly - polarized optical field with an amplitude @xmath88 , and frequency @xmath89 can be described as @xmath90 where @xmath91 is the transition dipole moment matrix element between states @xmath66 and @xmath92 , with an energy @xmath93 for the latter @xcite . thus @xmath86 has a quadratic dependence on the field , or a linear dependence on intensity for this frequency - dependent ac - stark level shift . in a simple case , when there is one near - resonant coupling to state @xmath92 whose contribution dominates @xmath86 , the sign of the detuning with respect to transition frequency @xmath94 determines the direction of the light shifts with intensity .
when the probing radiation is blue - detuned , i.e. @xmath95 , the two levels @xmath87 and @xmath96 shift toward each other , while for red - detuning , @xmath97 , the levels repel each other . in the case when all accessible states are far off - resonant , both terms in eq .
( [ eq : deltae ] ) contribute for each state @xmath92 , and numerous @xmath92-states need to be included in the calculations to explain the magnitude and sign of @xmath86 .
the energy shifts of the upper ( @xmath98 ) and lower ( @xmath99 ) levels in turn translate into an ac - stark shift , @xmath100 where @xmath101 is the ac - stark coefficient .
the measured _ transition _ energy is @xmath102 , where @xmath103 is the unperturbed ( zero - field ) transition frequency .
the ac - stark coefficient @xmath101 depends on the coupling strengths of the @xmath104 and @xmath105 levels to the dipole - accessible @xmath96 states , as well as the magnitude and sign of the detuning .
we note that in a so - called _ magic wavelength _
configuration , the frequency @xmath89 is selected so that the level shifts of the upper and lower states cancel out , leading to ac - stark free transition frequencies @xcite .
.uncertainty contributions in units of @xmath106 . [ cols="<,^,^",options="header " , ] to extract the ground electronic @xmath107 level energies , from the @xmath25 transition energy measurements , we use the level energy values of the @xmath42 states determined by bailly et al .
the derived experimental level energies are listed in table [ tab : levels ] , where the uncertainty is limited by the present @xmath45 determination except for the @xmath108 .
the calculated values obtained by komasa et al .
@xcite are also listed in table [ tab : levels ] .
the experimental and theoretical values are in good agreement , except for @xmath109 that deviate by 1.5-@xmath12 .
the combined uncertainty of the difference is dominated by the theoretical uncertainty .
however , improvements in the calculations of the nonrelativistic energies , limited by the accuracy of fundamental constants @xmath14 and @xmath15 , have recently been reported @xcite , and improved calculations of qed corrections up to the @xmath110-order is anticipated . as has been pointed out previously in ref .
@xcite , the uncertainties in the calculations @xcite are five times worse for the @xmath111 in comparison to the @xmath112 level energies . along with the previous measurements on the @xmath113 levels in ref .
@xcite , the measurements presented here probe the highest - uncertainty region of the most advanced first - principle quantum chemical calculations .
transition energies of @xmath114 rovibrational states were determined at @xmath5 @xmath6 absolute accuracies .
enhanced detection efficiency was achieved by resonant excitation to autoionizing @xmath7 electronic rydberg states , permitting excitation with low probe laser intensity that led to much narrower transitions due to reduced ac - stark effects .
the asymmetric line broadening , induced by the ac - stark effect , at high probe intensities was found to be well - explained by taking into account the spatial and temporal intensity beam profile of the probe laser .
the extracted ac - stark coefficients for the different transitions @xmath45 , as well as previously determined @xmath115 transitions , are consistent with qualitative expectations . however , a quantitative explanation awaits detailed calculations of the ac - stark effect that account for molecular structure , i.e. including a proper treatment of relevant intermediate states . using the @xmath36 level energies obtained by bailly et al .
@xcite , the level energies of @xmath116 states are derived with accuracies better than 0.002 @xmath6 except for @xmath117 , limited by @xmath36 level energy accuracy .
the derived experimental values are in excellent agreement with , thereby confirming , the results obtained from the most advanced and accurate molecular theory calculations .
the experimental binding energies reported here are about thrice more accurate than the present theoretical values , and may provide further stimulus towards advancements in the already impressive state - of - the - art _ ab initio _ calculations .
this work is part of the research programme of the foundation for fundamental research on matter ( fom ) , which is part of the netherlands organisation for scientific research ( nwo ) .
p. w. received support from laserlab - europe within the ec s seventh framework programme ( grant no .
. w. u. received funding from the european research council ( erc ) under the european union s horizon 2020 research and innovation program ( grant no . | accurate @xmath0 transition energies in molecular hydrogen were determined for transitions originating from levels with highly - excited vibrational quantum number , @xmath1 , in the ground electronic state .
doppler - free two - photon spectroscopy was applied on vibrationally excited h@xmath2 , produced via the photodissociation of h@xmath3s , yielding transition frequencies with accuracies of @xmath4 mhz or @xmath5 @xmath6 .
an important improvement is the enhanced detection efficiency by resonant excitation to autoionizing @xmath7 electronic rydberg states , resulting in narrow transitions due to reduced ac - stark effects . using known @xmath8 level energies , the level energies of @xmath9 states
are derived with accuracies of typically 0.002 @xmath6 .
these experimental values are in excellent agreement with , and are more accurate than the results obtained from the most advanced _ ab initio _ molecular theory calculations including relativistic and qed contributions . |
quantum cryptography @xcite , being the first practical realization of quantum physics at the single quanta level , is the art of creating data channels physically secure against eavesdropping .
most of its successful practical realizations are based on the use of single photon information coding though optical fiber links thus being based on a weak laser pulses .
the security of the protocols is based on the state perturbation during the eavesdropping @xcite or the measurement correlations analysis with the bell inequalities check @xcite . since
key bits are encoded in the single photon states the appearance of the additional photons may seriously undermine the protocol security by leading to the successful listening - in because additional photons states can be imperceptibly measured by an eavesdropper . thus the security reasons require the minimizing of the multi - photon pulses appearance probability making most of the pulses empty , which limits the key rate and results in an additional error rate caused by the single - photon counting detectors , which are inclined to `` dark counts '' , clicking when the photon is missing @xcite .
this leads to some contradiction between the effectiveness and security of the existing quantum key distribution ( qkd ) schemes .
the contradiction can be overcame by using multi - photon pulses and establishing the channel security on the realistic basis of multi - photon statistics rather then the fermion pairs statistics which the first qkd protocols where starting with . besides avoiding empty pulses by making carrier beams more intensive another point for increasing the effectiveness of quantum communications and qkd in particular is the alphabet size , which can be extended beyond the usual two - letter one , which corresponds to a classical bit .
it was shown that the large - alphabet coding essentially increases the bit rate of a quantum channel with no loss in security even in the case of a single - photon polarization coding @xcite . in this work
we prove the effectiveness of the alphabet extension for the new multiphoton qkd protocol which was proposed recently @xcite on the basis of the special beam states called two mode coherently correlated ( tmcc ) .
these states have the strong correlation between the photon numbers in each of the two spatially parted modes .
this correlation leads to the fact that laser shot noise shows itself equally in the both of the modes thus enabling the use of a tmcc - source as a generator of some random bit sequence which is then shared between two legitimate parties who perform independent photon - number measurements on each of the modes and extract the key bits from the measurement results comparing them to the constant average photon number . the quantum channel capacity for the 4- and 8-letter extended alphabets is estimated .
the eavesdropping attacks on the extended - alphabet tmcc - channels are considered in terms of the introduced qber .
it is shown that the alphabet extension leads to the increase of the effectiveness of the qkd protocol providing channel capacity of up to 3 bits per pulse .
moreover the alphabet extension strengthen the protocol security against the intercept - resend attacks by making eavesdropper introduce larger qber on each successfully intercepted bit exceeding 70% qber per bit for the 8-letter alphabet .
the tmcc - protocol is based on the use of a special states of light , the two mode coherently correlated ones . such states were mathematically studied by agarwal @xcite .
they are defined as fully correlated and at once the eigenstates for the product of annihilation operators of the both modes .
the latter condition makes them the special case of the wider class of the two - mode correlated states , also referred to as twin beams , which are broadly investigated in the past time @xcite and are usually obtained in the process of parametric down - conversion ( pdc ) in nonlinear crystals .
the tmcc - states can be presented through series by fock states : @xmath0 here we use the designation @xmath1 , where @xmath2 and @xmath3stand for the states of the @xmath4 and @xmath5 mode accordingly , represented by their photon numbers .
the main feature of the states ( [ eq : tmcc ] ) is that only the terms with equal photon numbers in the both of the modes are present in the expansion .
this leads to the strong correlation between the observables concerned with each of the modes . at the same time
the tmcc states differ from the usual two - mode correlated coherent states , because the average for any of the linear in field observable ( e.g. the vector - potential ) is equal to zero for each of the tmcc modes @xcite .
thus each of the modes is not coherent to itself , but as the square in field observables ( e.g. correlation function , momentum , energy ) have non - zero average values , the tmcc states possess the second order coherence and so can be referred to as the coherently correlated states .
the separate intensity measurements on each of the beams give the results which are proportional to the average of the @xmath6 , which corresponds to the number of photons in the mode .
the probability distribution for different photon numbers detection is @xmath7 the photon statistics for each of the beams is sub - poisson as confirmed by the mandel parameter , which is negative even for the small intensities @xcite .
this fact can be useful for the tmcc state identification and the quantum channel eavesdropping detection .
the shot noise photon number fluctuations in the tmcc modes enable the use of the tmcc source as a generator of a random key encoded in the photon numbers value .
the strong correlation between the independent measurements of two modes makes it possible to securely distribute such random key between two remote legitimate users , alice and bob .
the security of the quantum channel is based on the fact that intermediate photon measurements will perturb the states which can be checked locally by measurement statistics @xcite .
the simple protocol based on the two mode coherently - correlated states was described @xcite as follows : alice and bob simultaneously start the independent photon number measurements each on the corresponding tmcc mode .
they compare the obtained photon number values for each next unit time to the average which is constant during the overall key transmission procedure .
if the obtained number is larger than the average , the next bit is considered to be equal to `` 1 '' .
if the photon number for the next time window is less than average , the corresponding bit is equal to `` 0 '' .
thus , the protocol uses the two - bit photon number alphabet coding by the multiphoton two mode coherently correlated states .
the state coherence is an important feature of the tmcc - protocol distinguishing it from the existing single - photon incoherent state protocols because coherence allows establishing the security for the multiphoton pulses transmission since the state perturbation leads to the decoherence
. the loss of coherence can be revealed by checking the state statistics which can be done even locally by estimating differences between the obtained and expected state density matrices @xcite .
the classical information channel is well known to be described by the shannon mutual information between the observables of some classical macroscopic systems .
the measurements on such systems return the probability distributions for the sets of the observables discrete values .
these value sets when used for encoding and decoding the information are called the alphabets which contain the discrete values of the observables as the letters .
the input - output mutual information depends on the observables shannon entropies : @xmath8 where @xmath9 is the input observable entropy and the @xmath10 is the mutual entropy of the input relative to the output which describes the information loss in the channel : @xmath11 since @xmath10 depends on the probability distribution of the values set @xmath12 it depends on the alphabet size and increases with the increase of the number of its letters .
this discourse clearly fits the quantum channels which differ from the classical ones by the fact that quantum observables , being the parameters of the quantum microscopic states , are used for information encoding and decoding . since quantum cryptography deals with the key sharing across the quantum channel it can also gain from the alphabet expansion . in 1999
bechmann - pasquinucci and tittel @xcite proposed the use of a larger alphabet for the bb84 single - qubit protocol @xcite , which is extended on the four - level quantum systems - the so - called quantum quarts ( qu - quarts ) .
alice still selects randomly between the two possible bases , but she is now preparing one of the four states thus making eight possible choices for the qu - quart based protocol instead of the four choices for the qubit - based one .
it was shown that such a development of the bb84 protocol increases the information flux and makes the qkd scheme more secure against realistic eavesdropping because eavesdropper introduces the much higher qber for the given amount of the acquired information . lately in 2003 sych , grishanin and zadkov @xcite made the further development of the single - qubit qkd scheme by proposing the use of the continuous alphabet for the key bits coding .
the idea was to identify key bits from the unselected qubit states which was shown to result in the increase of the protocol effectiveness , security and reliability at noisy channels .
thus it makes it clear that the protocol extension may be quite useful for the tmcc - based protocol especially since the multi - photon states in the fock presentation can be considered as the multi - dimensional systems .
the maximum shannon entropy of a photon - counting beam measurement , i.e. the highest possible information capacity for a channel built on such beams can obviously be achieved if each of the different photon - number events are identified as the different measurement outcomes and correspond to the different alphabet letters . in other words ,
the m - letter alphabet for the m - photon state will give the maximum information which can be encoded into and transmitted by such a state . for the tmcc - beam this maximal information will be @xmath13 the dependence of this highest possible information gain of a state measurement on the average photon number of the state is given at ( [ maxent ] ) .
one can easily see that depending on the beam intensity , the tmcc state can carry from 1 up to 4 bits of information .
thus it is possible to build the effective tmcc - based qkd protocols utilizing 4-letter alphabets for beams carrying about 3 - 5 photons in average which will raise the measurement information gain to 2 bits and the 8-letter alphabets for more intensive beams which will result in the effectiveness of up to 4 bits per measurement .
here we examine two possible alphabet extensions for the tmcc channels , containing 4 and 8 state `` letters '' .
the quaternary channel with two bits per measurement capacity can be established on the basis of the tmcc beams which are distinguished by the photon numbers between 4 possible states ( see figure [ bits ] ) : @xmath14}-1 & 00 $ ( letter 0)$\\ \{n = { \left[\left\langle n \right\rangle\right ] } & 01 $ ( letter 1)$\\ \{n = { \left[\left\langle n \right\rangle\right]}+1 & 10 $ ( letter 2)$\\ \{n \geq { \left[\left\langle n \right\rangle\right]}+2 & 11 $ ( letter 3)$ \end{array}\right.}\ ] ] knowing the photon numbers registration probability distribution for a tmcc - beam ( [ eq : nphotprob ] ) one can easily estimate the probability for each of the 4 quart states ( [ eq : quart ] ) realization : @xmath15}-1 } { \frac{|\lambda|^{2n}}{n!^2 } } ; \
p_1(\lambda ) = \frac{1}{i_0 ( 2|\lambda| ) } \frac{|\lambda|^{2{\left[\left\langle n \right\rangle\right]}}}{{\left[\left\langle n \right\rangle\right]}!^2};\ ] ] @xmath16}+1)}}{({\left[\left\langle n \right\rangle\right]}+1)!^2};\ p_3(\lambda ) = \frac{1}{i_0 ( 2|\lambda| ) } \sum_{n={\left[\left\langle n \right\rangle\right]}+2}^\infty { \frac{|\lambda|^{2n}}{n!^2 } } \ ] ] these probability values can be then used for estimating the shannon entropy ( [ eq : shentropy ] ) of a photon - number measurement on a tmcc - beam with four possible interpretation outcomes , which is the information gain for such a measurement .
the octuple tmcc - based channel with the capacity of three bits per measurement can be established in a similar way if the states are distinguished by 8 possible measurement outcomes which differ in photon numbers ( see figure [ bits ] ) : @xmath17}-3 & 000 $ ( letter 0)$ \\
\{n = { \left[\left\langle n \right\rangle\right]}-2 & 001 $ ( letter 1)$ \\
\{n = { \left[\left\langle n \right\rangle\right]}-1 & 010 $ ( letter 2)$ \\
\{n = { \left[\left\langle n \right\rangle\right ] } & 011 $ ( letter 3)$ \\
\{n = { \left[\left\langle n \right\rangle\right]}+1 & 100 $ ( letter 4)$ \\ \{n = { \left[\left\langle n \right\rangle\right]}+2 & 101 $ ( letter 5)$ \\
\{n = { \left[\left\langle n \right\rangle\right]}+3 & 110 $ ( letter 6)$ \\ \{n \geq { \left[\left\langle n \right\rangle\right]}+4 & 111 $ ( letter 7)$ \end{array}\right.}\ ] ] again , knowing the distribution ( [ eq : nphotprob ] ) we can obtain the probability for each octo - bit value occurrence and thus build the shannon information available from a measurement . the dependencies for the measurement information gain for 2- , 4- and 8-letter alphabets on the state average photon number are given on a graph at ( [ entropies ] ) .
the security of a quantum channel i.e. the impossibility to carry out a successful eavesdropping is obviously the most important property for any qkd scheme .
traditionally the security of a quantum channel is examined in the frames of the two approaches - the ideal and the realistic ones @xcite . in the ideal case
the eavesdropper eve is supposed to have the unlimited technological power with possibilities restricted only by the laws of quantum mechanics @xcite .
the realistic approach takes into account the technical possibilities of an eavesdropper compatible with today s and foreseeable technology at this the most successful realistic eavesdropping techniques on the secure quantum channels are the intercept - resend strategy also referred to as state cloning and the beam splitting @xcite .
the most general security criterion known from the classical cryptography already states that bob has to possess more information on the transferred key than eve ( if the bob s mode was eavesdropped ) . in this case
the privacy amplification post - transfer algorithms will be helpful in distilling the truly secure key , otherwise they will fail @xcite . the eavesdropping in quantum cryptography is usually detected by checking the qber ( quantum bit error rate ) which is the measure of the errors in the obtained key . since all of the errors are considered to be caused by an eavesdropper , knowing the qber one can estimate how much information on the key does the eavesdropper have and thus determine is the key distillation will be successful .
the tmcc - based scheme was examined against the realistic eavesdropping and shown to be secure against splitting and cloning attacks . in the case of the beam splitting
this is quite intuitive because the installation of a splitter at any of the modes removes the correlation between modes thus simply destroying the channel ( for more detailed examination of a beam splitting attack on a tmcc - channel see @xcite ) . in the case of a state cloning
it was shown that eavesdropper significantly changes the statistics of the re - emitted mode which can be detected by the local calculations of the mandel parameter or the distances between the received and the expected states density matrices @xcite . in this work
we estimate the security of the tmcc - based channel in the terms of the qber , which is introduced to the channel upon the eavesdropping .
we use this measure to compare the security of the tmcc qkd scheme for different alphabet sizes .
let s consider eve carrying out a cloning intercept - resend attack on a tmcc - based channel .
in order to do so she installs a photon - counting detector which measures one of the modes ( suppose the one which goes to bob ) and a tmcc laser source assigned for re - creating the state ( fig [ cloning ] ) . for each incoming pulse ( i.e. in each time slot ) eve measures the photon number @xmath18 and tries to re - emit the same photon number in the bob s direction .
we assume eve calculates the value of state parameter @xmath19 which corresponds to the photon number @xmath18 to be emitted and sets her source up to this value .
though due to the quantum fluctuations she is unable to emit exactly the same number and thus bob obtains the state which density matrix @xmath20 , differs from the original @xmath21 measured by eve .
this re - emitted state matrix is the mixture of the k - photon states @xmath22 with probabilities @xmath23 the probability for bob to incorrectly interpret the received state i.e. to obtain the wrong letter @xmath24 of the alphabet set @xmath25 of size @xmath26 ( so that @xmath27 ) is equal to @xmath28 where @xmath29 is the probability for bob to obtain the correct letter @xmath24 given alice obtained the same letter @xmath24 .
the qber introduced by eve for each intercepted bit is then @xmath30 the sum in @xmath31 can be presented and calculated through the probability distributions as @xmath32 where @xmath33 is the photon number corresponding to the first alphabet letter @xmath34 and @xmath35 is the photon number corresponding to the last letter @xmath36 , @xmath37 is the probability distribution ( @xmath38 ) and @xmath39 hence the qber can be calculated numerically .
the results are presented on the graphs at fig @xmath40 as the qber dependence on the average photon number ( i.e. on the original pulse intensity ) for different alphabet sizes .
one can easily see that qber introduced for each intercepted bit during the state cloning attack on a 2-letter ( 1-bit ) alphabet channel is about 20% , growing to about 50% in the case of 4-letter ( 2-bit ) case and exceeding 70% for the 8-letter ( 3-bit ) alphabet .
thus the tmcc - based qkd protocol turns out to be more secure at the larger alphabets .
if eve uses a usual single mode laser source , producing a coherent beam with the poisson statistics for the tmcc - state cloning , the calculations give qber values of about 30% for 2-letter , up to 60% for 4-letter and over 80% for the 8-letter alphabet thus proving that tmcc - state cloning based on a usual laser source is even less effective .
the quantum cryptography scheme with larger alphabets based on the use of the two - mode coherently correlated multiphoton beams is proposed .
the alphabet extension is shown to result in the increase of the effectiveness of the qkd scheme .
the protocol security against the realistic state cloning is examined in terms of the introduced qber .
it is shown that the tmcc - based qkd scheme becomes more secure for the larger alphabet sets i.e. for the more intense laser pulses .
usenko v. c. , usenko c. v. , proceedings of the international conference on quantum communication , measurement and computing ( glasgow , uk , 2004 ) , am .
inst . of phys .
cp734 , p. 319
/ quant - ph/0407175 ( 2004 ) laurat j. , coudreau t. , treps n. , maitre a. and fabre c. , conditional preparation of a quantum state in the continuous variable regime : generation of a sub - poissonian state from twin beams , phys .
91 , 213601 / quant - ph/0304111 ( 2003 ) | the large - alphabet quantum cryptography protocol based on the two - mode coherently correlated multi - photon beams is proposed . the alphabet extension for the protocol
is shown to result in the increase of the qkd effectiveness and security .
pacs numbers : 03.67.-a,03.67.dd,03.67.hk |
to model a gas cloud in hydrostatic equilibrium we make the simplified assumption that the gas pressure consists of the sum of thermal ( subscript `` th '' ) , turbulent wave ( `` w '' ) , and magnetic ( `` m '' ) pressure components , @xmath0 we compute the thermal pressure through a detailed thermal equilibrium calculation , but the wave and magnetic pressures are assumed to obey a polytropic equation of state , @xmath1 i.e. the pressure only depends on density , and @xmath2 , the polytropic exponent , is constant in a given object . given the thermal and non - thermal gas pressure , it is possible to construct hydrostatic equilibria for any surface pressure and central to surface density contrast .
we assume spherical symmetry for simplicity . whether a given hydrostatic equilibrium model cloud is gravitationally stable depends on its response to perturbations in pressure or density .
different equations of state apply for such perturbations .
for example , if the perturbation occurred on a time scale shorter than the cooling time , the scaling of the thermal pressure with density would be stiffer than if the perturbation would be allowed to reach thermal equilibrium . to analyze the stability we make the simple assumption that the perturbation obeys a polytropic equation of state , but with some different `` adiabatic index '' , @xmath3 : @xmath4 where @xmath5 are infinitesimal perturbations , and the subscript `` 0 '' refers to values before the perturbation .
a hydrostatic equilibrium cloud is stable against spontaneous contraction or expansion when at any point in the cloud , the pressure increases during a compression or decreases during an expansion , i.e. , @xmath6 where @xmath7 is the radius from the cloud center .
marginally stable equilibria have @xmath8 at some point in the cloud . observed column densities of prestellar cores ( _ symbols _ ) and upper limits for stable equilibria ( _ lines _ ) .
the lower line applies for pure thermal support , the one above for an equal thermal and magnetic pressure in the cloud center , the one above that for equal thermal and turbulent , the top one for equal thermal , magnetic and turbulent pressures in the cloud center .
_ dotted lines _ connect independent measurements of the same object by different authors . ] at any radial point in the cloud we solve the coupled thermal balance for the gas and dust .
we adopt the gas heating and cooling rates and thermal gas - dust coupling from goldsmith @xcite , and the dust heating and cooling rates for the solar neighborhood from zucconi et al . @xcite .
the heating is due to cosmic rays and the interstellar radiation field , for which we allow some shielding , @xmath9 , due to the material surrounding the core .
when analyzing the stability , we assume that the heating or cooling due to a perturbation is negligible .
the heating and cooling rates themselves however change , as they depend on the perturbed density and visual extinction .
we treat the kinetic pressure due to non - thermal motions , which we call turbulence , in the very simplifying framework of alfvn waves .
the pressure due to alfvn waves can be described by a polytropic exponent @xmath10 , and an adiabatic exponent @xmath11 for the case where no energy flow occurs @xcite .
if we allow for a flow of the turbulent energy between the core and its surroundings during perturbations , then the adiabatic exponent would be equal to the polytropic one , i.e. @xmath12 .
this requires that the heat flows are faster than the duration of the perturbation . the magnetic pressure due to the average magnetic flux threading the cloud depends on the distribution of field lines in the cloud , which is a result of how the cloud has formed while the gas was frozen to the field lines .
a somewhat more realistic case was adopted for the 2-dimensional hydrostatic equilibria computed by tomisaka et al.@xcite : here a cloud was assumed to initially have been spherical and at uniform density within a uniform magnetic field .
we examined these 2-d equilibria to determine the magnetic field
density scaling , finding @xmath13 . the magnetic adiabatic exponent adopted in our work , @xmath14 ,
is chosen such that the critical density contrasts of tomisaka et al.@xcite are reproduced .
we adopt these values for our idealizing spherical model to study the behavior of 2-d magnetized equilibria .
in our yet preliminary study we have looked at clouds with a particular ratio between the non - thermal and thermal pressure _ at the cloud center : _ we investigated clouds with @xmath15 or 1 , and with @xmath16 or 1 , which is meant to give a first impression on the effect of non - thermal pressure support on the cloud stability .
we adopt an extinction due to material surrounding the clouds of 5 magnitudes in the visual . as a function of cloud mass we computed the maximum central column density for a stable hydrostatic equilibrium .
these are plotted in figure [ fig:1 ] for clouds with and without non - thermal pressure .
a comparison with the column densities measured for all starless low - mass cores we found in the literature shows that some of the more massive cores have column densities which are larger than what could be accounted for by a stable cloud with pure thermal pressure support . with a magnetic or turbulent pressure at least comparable to the thermal pressure in the cloud center , higher central condensations comparable to those observed would be possible for stable clouds .
if this non - thermal pressure were due to turbulent motions , then the observed cores non - thermal velocity dispersions should be comparable to the sound speed , but observations typically show smaller non - thermal linewidths @xcite .
thus turbulence is unlikely to be the dominant support in such cores , and magnetic fields remain as the stabilizing force .
a second argument for the need of non - thermal support in the more massive low - mass cores derives from their maximum possible surface pressure .
a core of given mass in a stable hydrostatic equilibrium can only exist if the kinetic surface pressure due to thermal and turbulent particle motions is below some critical value .
significant non - thermal pressure is necessary to stabilize the more massive cores ( @xmath17 ) in such environments . to summarize ,
our preliminary analysis of the stability of dense , self - gravitating cores and our comparison with observed cores and their environments suggests the need for non - thermal pressure support , most likely provided by magnetic fields of less than @xmath18 .
p. caselli , p.j .
benson , p.c .
myers , m. tafalla : apj * 572 * , 238 ( 2002 ) d. galli , c.m .
walmsley , j. gonalves : a&a * 394 * , 275 ( 2002 ) c.f .
gammie , e.c .
ostriker : apj * 466 * , 814 ( 1996 ) p.f .
goldsmith : apj * 557 * , 736 ( 2001 ) c.f .
mckee , j.h .
holliman ii : apj * 522 * , 313 ( 1999 ) c.f .
mckee , e.g. zweibel : apj * 440 * , 686 ( 1995 ) k. tomisaka , s. ikeuchi , t. nakamura : apj * 326 * , 208 ( 1988 ) a. zucconi , c.m .
walmsley , d. galli : a&a * 376 * , 650 ( 2001 ) | following an approach initially outlined by mckee & holliman @xcite , we investigate the structure and stability of dense , starless molecular cloud cores . we model those as spherical clouds in hydrostatic equilibrium and supported against gravity by thermal , turbulent , and magnetic pressure .
we determine the gas pressure by solving for thermal equilibrium between heating and cooling , while the turbulent and magnetic pressures are assumed to obey polytropic equations of state . in comparing the models to observed cloud cores we find that the observed
peak column densities often exceed the limit for stable equilibria supported by thermal pressure alone , suggesting significant non - thermal pressure if the cores are to be stable .
non - thermal support is also needed to stabilize cores embedded in molecular clouds with high average pressures . since the observed molecular linewidths of cores suggest that the turbulent pressure is lower than the thermal pressure , magnetic field are likely a dominant pressure component in many such cores . |
heavy quark production processes provide a powerful insight into our understanding of quantum chromodynamics ( qcd ) . the study of heavy mesons properties is also a subject of interest for understanding of quark - gluon interaction dynamics .
generally , two mechanisms are investigated for the production of heavy mesons : recombination and fragmentation @xcite . in the first scheme ,
heavy mesons are formed from heavy - heavy or heavy - light quarks which are produced independently in hard subprocesses . in the second mechanism
, the fragmentation refers to the process of a parton which carries large transverse momentum and subsequently forms a jet containing the expected hadron @xcite . at sufficiently large transverse momentum of the heavy meson production , the direct leading - order production scheme ( recombination mechanism )
is normally suppressed while the fragmentation scheme becomes dominant , though it is formally of higher order in the strong coupling constant @xmath1 @xcite .
+ the fragmentation mechanism is described by the function @xmath2 which refers to the probability for a parton @xmath3 at the factorization scale @xmath4 to fragment into a hadron @xmath5 carrying away a fraction @xmath6 of its momentum @xcite .
the fragmentation functions ( ffs ) are key quantities to calculate hadroproduction cross sections and their specific importance is for their model - independent predictions of the cross sections at the large hadron collider ( lhc ) . in this respect , one needs to determine these functions with high accuracy as possible .
+ basically , there are two approaches to determine the ffs where one calculates these functions in the initial scale of fragmentation and then can evolve them to higher scales using the dokshitzer - gribov - lipatov - altarelli - parisi ( dglap ) renormalization group equations @xcite . in the first approach ( _ phenomenological _ approach ) the free parameters in the proposed forms of the ffs are extracted form experimental data analysis . since the hadronization mechanism is universal and independent of the perturbative processes which produce the partons one can exploit , for example , the existing data on @xmath7 events to fit the proposed models for the @xmath8 transition . in @xcite , considering a power model for the ffs we determined the @xmath9 ffs , both at lo and nlo , through a global fit to the single - inclusive @xmath10 annihilation data and the semi - inclusive deep inelastic scattering asymmetry data from hermes and compass .
this phenomenological approach is frequently used to obtain the nonperturbative ffs .
the ffs are related to the low - energy part of the hadroproduction processes but , fortunately , it was found that these functions for heavy hadron productions are analytically calculable by virtue of perturbative qcd ( pqcd ) with limited phenomenological parameters @xcite .
an alternative fragmentation model which does contain spin information has been proposed by suzuki @xcite . in this approach
, suzuki calculates the heavy ffs using the same feynman diagrams , as in the pqcd approach , for the parton level of the process and also by considering the wave function of heavy meson which contains the bound state nonperturbative dynamic of produced meson . in the suzuki s model ,
the analytical expression of ffs depends on the transverse momentum @xmath11 of the initial parton , while in the pqcd scheme one integrates over the invariant mass of the fragmenting quark .
the invariant mass is related to the transverse momentum @xmath11 of the meson relative to the fragmenting quark @xcite .
in fact , rather than integrating over @xmath12 , the suzuki s model chooses to evaluate the integrand at a typical value @xmath13 , see ( [ ayda2 ] ) . in @xcite , using the suzuki s model we calculated the initial scale fragmentation function for @xmath14-quark to split into s - wave @xmath15-mesons at leading - order of perturbative qcd . here
, we review the suzuki s formalism at lo and detail the extension of this formalism at nlo by considering the real and virtual gluon radiative corrections .
finally we will present , for the first time , our nlo analytical expression of the heavy quark ff in the suzuki s model and compare our result with the lo one @xcite . to show the importance of our calculations we will also compare both results with experimental data from belle @xcite and cleo @xcite .
we also compare our analytical results with a well - known phenomenological model .
as will be shown , the nlo corrections improve our theoretical results at lo and make good agreement with experimental data .
the theoretical approaches to calculate the heavy quark ffs depend on the fact that the ffs for hadrons containing heavy quarks can be calculated theoretically using the perturbative qcd ( pqcd ) @xcite .
the first theoretical effort to illustrate the production procedure of hadrons containing heavy quarks was established by bjorken @xcite . in a naive quark - parton model
, he deduced that the inclusive distribution of heavy mesons should peak almost at @xmath16 , where @xmath17 refers to the longitudinal momentum fraction of the hadron state .
the pqcd scheme was followed by suzuki @xcite , ji and amiri @xcite by considering more elaborate models . while in this approach suzuki computes the heavy ffs by applying a feynman diagram similar to that in fig .
[ lo ] , amiri and ji calculate their ffs in @xmath10 annihilation process in the same order of pqcd . in their models
the total amplitude for the fragmentation of a heavy quark into a heavy meson is obtained by perturbative calculations of quark - gluon non - abelian interaction up to the order of @xmath18 and use of a delta function to represent the s - wave heavy meson bound state .
in fact , they consider a heavy bound state as a nonrelativistic system and reduce its wave function to a delta function @xcite . the lowest order feynman diagram contributing to the fragmentation of a heavy quark q into a heavy meson @xmath19 . ]
production of a heavy meson @xmath19 at next - to - leading order ( nlo ) .
real gluon radiative contributions to @xmath20 are shown at nlo . the spins ( @xmath21 ) and four - momenta
are also labeled . ]
the suzuki s model includes most of the kinematical and dynamical properties of the splitting process and gives us a detailed insight about the fragmentation process . specifically ,
if we want to know about further details such as the spin property of fragmentation , this model can be instructive .
briefly , the suzuki s model is a dynamical model which is more predictive and describes spin - dependent effects and also includes the kinematical details of the fragmentation process .
it mixes a perturbative picture with nonperturbative dynamics of fragmentation and not only predicts the z - dependence of the ffs , but also their dependence on @xmath11 , the transverse momentum of the meson relative to the jet .
+ in @xcite , using the suzuki s model we derived an analytical expression for the heavy quark ff at lowest order ( @xmath18-order ) of pqcd by considering the typical feynman diagram shown in fig . [ lo ] , where a heavy quark q creates a bound state @xmath19 along with a light quark @xmath22 through a single gluon .
the result for the fragmentation function @xmath23 was dependent on the transverse momentum @xmath11 of the initial heavy quark . here , we present a compacted expression of our previous result as @xmath24\left\langle k_t^{2}\right\rangle+m_qm_q\big[m_q-(1-z)m_q\big]\bigg ) \nonumber\\ & & + \big[\left\langle k_t^{2}\right\rangle+m_q^2\big]\big[\left\langle k_t^{2}\right\rangle+m_q^2\big]\frac{m_qz^4}{m_q}\bigg\},\end{aligned}\ ] ] where , @xmath25 ^ 2\times \bigg[z^2 \left\langle k_t^{2}\right\rangle+(m(z-1)-zm_q)^2\bigg]^2,\end{aligned}\ ] ] and @xmath26 , but the coefficient b is related to the normalization condition @xmath27 @xcite . in computing eq .
( [ last ] ) , following ref .
@xcite we adopted the infinite momentum frame where the fragmentation parameter in the usual light - cone form , @xmath28 , is reduced to a more popular form @xmath29 which is more convenient when the masses of partons and outgoing meson are ignored . in reality , the scaling variable @xmath17 refers to the energy fraction of the fragmenting heavy quark which is taken away by the produced meson and takes the values as @xmath30 . in @xcite , authors studied the theoretical uncertainties due to the freedom in the choice of fragmentation parameter in the presence of heavy quark and meson masses .
in fact , hadron mass modifies the relations between partonic and hadronic variables and is responsible for the low-@xmath17 threshold , although this additional effect is not expected to be truly sizable numerically , its study is nevertheless necessary in order to fully exploit the enormous statistics of the lhc data . in the present work we derive an analytical form of the transverse momentum dependent heavy quark ff at next - to - leading order ( nlo ) with assumption of a delta function for the meson bound state , as in @xcite .
the underlying link between hadronic phenomena in qcd at large and small distances is the hadronic wave function .
in fact , the nonperturbative aspect of the hadroprodution processes is contained in the bound state of the meson which is described by the wave function .
following ref .
@xcite and according to the lepage - brodsky s approach @xcite we neglect the relative motion of the constituent quarks @xmath31 and @xmath32 therefore we assume , for simplicity , that the quark pair @xmath33 are emitted collinearly with each other and move along the @xmath34-axes .
in fact , in the suzuki s model a meson is replaced by collinear constituents with neglecting the fermi motion and the nonperturbative aspect of the hadroproduction is included in the wave function of the heavy meson bound state .
+ the feynman diagrams of the real gluon corrections are shown in fig .
considering these diagrams we set the relevant four - momenta as @xmath35 , & & \quad s_\mu=[s_0 , \vec{0 } , s_l ] , \nonumber\\ s_\mu^\prime = [ s_0^\prime , \vec{s^\prime}_t , s_l^\prime ] , & & \quad t_\mu=[t_0 , \vec{0 } , t_l],\\ t_\mu^\prime = [ t_0^\prime , \vec{t^\prime}_t , t_l^\prime ] , & & \quad \bar p_\mu=[\bar p_0 , \vec{0 } , \bar p_l],\nonumber\end{aligned}\ ] ] where @xmath36 refers to the four - momentum of the produced meson , so @xmath37 .
+ to proceed , we define the momentum fractions carried by the constituent quarks as : @xmath38 and @xmath39 so that @xmath40 . in the infinite momentum frame
these fractions are reduced to simpler forms as @xmath41 thus , @xmath42 and @xmath43 stand for the meson energy fractions carried by the constituent quarks . considering the definition of fragmentation parameter , @xmath44 , we also may write the parton energies in terms of the initial heavy quark energy @xmath45 as @xmath46 in the suzuki s model the fragmentation function for the production of a s - wave heavy meson @xmath47 in the fragmentation of a quark q may be put in the following form @xcite @xmath48 where , @xmath4 is the fragmentation scale , @xmath49 refers to the spin of the fragmenting quark and the summation is going over the spins and colors . in ( [ first ] ) , @xmath50 is the probability amplitude of the meson production which , at the large momentum transfer , is expressed in terms of the hard scattering amplitude @xmath51 and the process - independent distribution amplitude @xmath52 @xcite as @xmath53 t_h(\bar p , s^\prime , t^\prime , x_i ) \phi_m(x_i , q^2),\end{aligned}\ ] ] where @xmath54=dx_1dx_2\delta(1-x_1-x_2)$ ] .
this scheme is convenient to absorb the soft behavior of the bound state into the hard scattering amplitude @xcite .
the short - distance coefficient @xmath51 can be calculated perturbatively from quark - gluon subprocesses at lo or nlo approximations .
the long - distance distribution amplitude @xmath52 which contains the bound state nonperturbative dynamic of the outgoing meson , is the probability amplitude for a @xmath55-pair to evolve into a particular bound state .
the distribution amplitude @xmath52 is related to the valence wave function of meson @xmath56 @xcite . with the heavy quark mass , the relative motion of the constituent quarks inside
the heavy meson is effectively nonrelativistic and this allows one to estimate the nonrelativistic mesonic wave function as a delta function form .
therefore , the distribution amplitude for a s - wave heavy meson with neglecting the fermi motion reads @xcite @xmath57 where @xmath58 stands for the meson mass in the nonrelativistic limit , @xmath59 refers to the decay constant of meson which can be also related to the nonrelativistic mesonic s - wave function @xmath60 at the origin as @xmath61 . in the meson decay
constant , @xmath62 is the binding energy of the mesonic bound state . in @xcite
, we studied the effect of meson wave function on the heavy quark ff by considering a typical mesonic wave function which is different of the delta function and is the nonrelativistic limit of the solution of bethe - salpeter equation with the qcd kernel @xcite .
however , due to the lengthy and cumbersome expressions of the new ff we just presented the two - dimensional integrals which must be evaluated numerically .
+ here , for simplicity , we consider a delta function for the hadron bound state as well . with this approximation ( [ delta ] )
, we are assuming that the contribution of each constituent quark from the meson energy is proportional to its mass , i.e. @xmath63 and @xmath64 ( [ zare ] ) so that @xmath40 . using eqs .
( [ base ] ) and ( [ delta ] ) , one has @xmath65 considering the nlo feynman diagrams shown in fig .
[ nlo ] , where a produced meson is replaced by collinear constituent quarks , we make the nlo approximation for the fragmentation function of @xmath66-meson . in ( [ pepe ] ) , the qcd amplitude @xmath51 is , in essence , the partonic cross section to produce a quark pair @xmath33 with certain quantum number that in the old fashioned perturbation theory is expressed as @xmath67 where , @xmath68 is the color factor for the process @xmath69 . in ( [ forth ] ) , the amplitudes @xmath70 stand for each feynman diagrams shown in fig .
[ nlo ] , and include an appropriate combination of the quark propagators and the spinorial parts of the amplitude .
we set the amplitudes @xmath71 for fig .
[ nlo]a , @xmath72 for fig .
[ nlo]b , @xmath73 for fig .
[ nlo]c and the amplitude @xmath74 for fig .
these amplitudes read @xmath75u(p^\prime , r_1)\big\}\times\nonumber\\ & & \hspace{3cm}\big\{\bar{u}(t^\prime , r_4)\gamma_\nu v(t , r_3)\big\},\nonumber\end{aligned}\ ] ] where , the denominator of propagators are expressed as @xmath76 next , using the kinematics ( [ kinematic ] ) we put the dot products of the relevant four - vectors in the following form @xmath77 substituting ( [ base ] ) and ( [ forth ] ) in ( [ first ] ) and carrying out the necessary integrations , the fragmentation function @xmath78 reads @xmath79 where @xmath80 and the factor @xmath81 is the energy denominator .
+ to proceed one needs to determine the phase space integrations in ( [ hehe ] )
. then we start with the following integral @xmath82 where @xmath83 . considering the dot products of the four - vectors ( [ mohsen ] ) ,
it is simplified as @xmath84 for simplicity , we also assume that the emitted gluon and outgoing light quark @xmath85 move almost in the same direction .
this assumption is justified by the fact that the very high momentum of the initial heavy quark is predominantly carried in the forward direction . due to momentum conservation ,
the total transverse momentum of the emitted gluon and light quark will be identical to the transverse momentum of the initial heavy quark .
therefore , we have @xmath86 . according to the suzuki s model , and for simplicity
, we also replace the transverse momentum integrations by their average values as @xmath87 where @xmath88 is a free parameter which can be specified phenomenologically . substituting all in ( [ first ] ) , we obtain the fragmentation function for the @xmath89 as follows @xmath90 the next step will be to compute @xmath91 and sum ( or average ) over gluon and quark polarization states . at first , we calculate the contribution of the forth feynman diagram ( fig .
[ nlo]d ) to the radiative corrections by considering @xmath74 in ( [ fifth ] ) .
then one has @xmath92u(p^\prime , r_1)\big\}\times\nonumber\\ & & \hspace{2.25cm}\big\{\bar{u}(t^\prime , r_4)\gamma_\nu v(t , r_3)\big\}=0,\end{aligned}\ ] ] and for the remaining contributions we have @xmath93 in calculating the above terms we used the identities @xmath94 and @xmath95 for the polarization sums .
there is a similar trick for summing over gluon polarization vectors .
the correct prescription is to make the replacement : @xmath96 ( see sec.5.5 of ref .
@xcite ) . considering the dirac algebra and using the dot products of four - momenta
, these expressions can be simplified as @xmath97 ^ 2\bigg[z^2k_t^2+m_q^2(1 - 2z+3z^2)\bigg],\nonumber\\ \sum_{r_i}\gamma_1\cdot\gamma_3^\star&=&\frac{4m_q}{g_1g_2g_5g_6}\bigg\{\frac{m_q^5(z-1)}{z^3 } ( 5z^4 - 2z^3 + 6z^2 - 2z+1)-\frac{m_q^3k_t^2}{z(1-z)}(12z^4 - 19z^3 + 21z^2 - 9z+3)-\nonumber\\ & & \frac{zm_qk_t^4}{1-z}(9z^2 - 6z+3)+\frac{z^3(2z-1)k_t^6}{m_q(1-z)^2}\bigg\},\nonumber\\ \sum_{r_i}\gamma_2\cdot\gamma_3^\star&=&\frac{-8}{g_3g_4g_5g_6}\frac{m_q}{z^2(1-z)^2m_q}\bigg [ z^2k_t^2+m_q^2(1 - 2z+3z^2)\bigg]\bigg[z^2k_t^2+(1-z)^2m_q^2\bigg]^2.\end{aligned}\ ] ] virtual gluon contributions to the @xmath20 at nlo . ] at nlo approximation , in addition to the real gluon radiative corrections there are some feynman diagrams related to the virtual corrections .
this class of contributing diagrams interferes with the lo amplitude .
in fact , the nlo full amplitude is the sum of the amplitudes of the born term ( @xmath98 ) , virtual one - loop ( @xmath99 ) , and the real contributions ( @xmath100 ) , @xmath101 the qcd nlo contributions result from the square of the amplitudes : @xmath102 so its related fragmentation function is of order @xmath18 ( [ last ] ) , @xmath103 and @xmath104 ( [ ali ] ) so that the nlo fragmentation function is of order @xmath105 ( [ lasttt ] ) .
the feynman diagrams related to the virtual gluon radiative corrections are classified into two classes .
the first class of contributing diagrams includes the fermion loop diagram , the three - gluon vertex and a four - gluon vertex .
these are shown in fig .
[ virtual ] .
it is shown that these virtual contributions interlock in an essential way .
in general , feynman diagrams with @xmath106 loops typically contain correction terms proportional to @xmath107 , where @xmath108 is a renormalization scale .
fortunately , we can absorb these corrections into the lowest - order terms by using the renormalization group equations .
in other words , their effect is to modify the gluon propagator by replacing the fixed renormalized coupling with a running coupling constant . by solving the renormalization group equations , the one - loop corrections shown lead to evolve the qcd coupling constant at the energy q , as @xcite @xmath109 with @xmath110 , where @xmath111 refers to the active quark flavor numbers . in the above equation
, @xmath108 is a typical qcd scale which shows the border between the perturbative and nonperturbative regimes of qcd . in practice
the value of @xmath108 is ambiguous and is determined by experiments .
experimental measurements of the rate of @xmath112 reaction and others yield a value of @xmath113 mev @xcite so the qcd perturbation theory is valid only when @xmath31 is somewhat larger than this , say above @xmath114 gev , where @xmath115 .
+ besides these 1pi diagrams , there are also three @xmath116 diagrams ; one - loop diagrams with a propagator that connects back to its originating vertex . it is shown that these automatically vanish .
the second class of virtual corrections includes the gluon - quark loops on the incoming or the outgoing quark legs .
generally , these amplitudes need to be considered and include in order to maintain the infrared stability of the overall result . indeed ,
these virtual corrections consist of both infrared ( ir ) and ultraviolet ( uv ) singularities where the uv - divergences appear when the integration region of the internal momentum of the virtual gluon goes to infinity and the ir - divergences arise from the soft - gluon singularities .
all uv - singularities are canceled after summing all virtual contributions up , whereas the ir singularities are remaining .
the real gluon radiative corrections also include ir - divergences which arise from the soft- and collinear gluon emissions . according to the lee - nauenberg theorem , after summing all radiative corrections up
the ir - singularities cancel each other and the final result is free of all singularities .
more details can be found in our previous works @xcite , where we calculated the decay rate of top quarks at nlo by working at dimensional regularization scheme .
note that in the suzuki s model , to compute the contribution of the real corrections into the fragmentation function we do not integrate over the momentum of the emitted real gluon and instead , we replace the gluon momentum integration by its average value , see eq .
( [ ayda2 ] ) .
therefore , by this simplification on one side we shall not deal with the ir - singularities in the real gluon radiative corrections ( [ ali ] ) and on the other side the contribution of the virtual corrections can be ignored .
we checked that the contribution of the virtual gluon corrections into the qcd amplitude @xmath51 ( [ forth ] ) is small and then their corresponding ffs are tiny .
specifically , this point is confirmed in fig .
[ belle ] .
the behavior of @xmath117 as a function of @xmath17 when @xmath118 is normalized to unity . ]
the pqcd ff of @xmath119 at lo ( solid line ) and nlo ( dashed line ) approximations .
the fragmentation scale is @xmath120 and we set @xmath121 gev . the theoretical results are also compared with data from belle @xcite and cleo @xcite . ] comparison of pqcd ff with data from belle @xcite on @xmath122 production at the initial scale @xmath120 , considering lo ( solid line ) and nlo ( dashed line ) approximations . ]
we are now in a position to present our numerical results for the heavy quark fragmentation function .
our result at the leading - order approximation is shown in ( [ last ] ) and the result at nlo is obtained as follows @xmath123 where @xmath124 is obtained through the normalization condition @xmath125 @xcite , and the @xmath126 and @xmath127 are presented in ( [ last ] ) and ( [ lasttt ] ) , respectively .
generally , the fragmentation function @xmath128 depends on both the fragmentation parameter @xmath129 and the fragmentation scale @xmath130 .
this scale is normally arbitrary , but in a high energy process of electron - positron annihilation where a jet is produced with transverse momentum @xmath131 , large logarithms of @xmath132 in the partonic cross section of the process @xmath133 can be avoided by choosing @xmath130 on the order of @xmath131 .
also , the @xmath17 dependence of the ff is not yet calculable at each desired scale .
however , once they are computed at some initial fragmentation scale @xmath4 , their @xmath130 evolution is specified by the dglap evolution equations @xcite .
therefore , the function ( [ final ] ) should be regarded as a model for the heavy quark ff at the scale @xmath4 of order @xmath134 and the @xmath135 at larger scales can be obtained by solving dglap equations . here , as an example
, we consider the fragmentation of @xmath14-quark into @xmath136- and @xmath122-mesons with the constituent quark structures @xmath137 and @xmath138 using @xmath139 gev , @xmath140 mev , @xmath141 mev and @xmath142 gev @xcite . in this work
we also consider @xmath143 adjusted such that @xmath144 with @xmath145 gev .
+ it is worth mentioning here that the average transverse momentum is not a constant but a function of the fragmentation parameter @xmath17 .
our prediction for @xmath17 dependence of this quantity is shown in fig .
this plot justifies the fact that the choice of @xmath121 gev is an extreme value for this quantity and any lower value will produce the peak even at higher - z regions . in figs .
[ belle ] and [ dplus ] , the behavior of @xmath136 and @xmath122 ffs at the initial scale @xmath120 is shown for the lo ( solid line ) and nlo ( dashed line ) approximations .
for comparison , data from belle @xcite and cleo @xcite are also shown , see also fig . 3 from ref .
as is seen , there is reliable consistency between our analytic result at nlo and experimental data .
however , we may also think of other effects such as the fermi motion of constituent quarks , the meson relativistic wave function @xcite and the effects of meson mass and so on , which can make a better agreement with experimental data . concerning the data shown in figs . [ belle ] and [ dplus ] we shall discuss , in this section , later . besides the theoretical schemes there is another current approach to determine the ffs ; _ phenomenological _ approach .
this is based on data analyzing where the ffs are mainly determined by hadron production data of electron - positron annihilation , hadron - hadron and lepton - hadron scattering processes .
however , among all the @xmath10 annihilation provides a clean environment to determine the fragmentation densities , specifically one does not need to consider the parton distribution functions ( pdfs ) of initial hadrons . in this scheme the collin s factorization theorem of the qcd improved parton model @xcite is an important tool to study this process . according to this theorem , in the high energy scattering @xmath146 the cross section of hadron production
is described by the convolution of partonic hard scattering cross sections @xmath147 , which are calculable in perturbative qcd @xcite , and the realistic and nonperturbative ffs @xmath148
describing the transition of a parton into an outgoing hadron ( @xmath149 ) , i.e. @xmath150 where , the momentum fraction @xmath17 is defined as @xmath151 where @xmath152 is the energy of hadron and @xmath153 is the square of total @xmath10 center - of - mass energy ( @xmath154 ) . in ( [ monaa ] ) , @xmath155 is defined as @xmath156 and @xmath157 stands for the unobserved jets and @xmath158 is the total partonic cross section at nlo @xcite . in this scheme
, the ffs are parameterized in terms of a number of free parameters which are determined by an @xmath159 analysis of the @xmath10 annihilation data at the scale @xmath160 where @xmath161 is the squared center - of - mass energy .
these parameterizations should include some restrictions .
for example , they must be zero at @xmath162 and @xmath16 .
various phenomenological models like peterson model @xcite , lund model @xcite , cascade model @xcite and etc . , have been developed to describe the ffs .
+ in ref .
@xcite , authors computed the ffs of @xmath163 and @xmath164 mesons through a global fit to electron - positron data from the belle , cleo , aleph , and opal collaborations . according to the bowler model @xcite
, authors have parameterized the @xmath17 distributions of the charm quark ff at its starting scale @xmath120 , as @xmath165 with three free parameters .
their result for @xmath136-meson reads @xmath166 , @xmath167 and @xmath168 with the value of @xmath169 achieved . in fig .
[ bowler ] , using ( [ final ] ) the behavior of @xmath136 ff at the starting scale @xmath120 is compared with the bowler model , as a well - known phenomenological model .
since to obtain the constant @xmath124 ( [ final ] ) we have used the normalization condition then to compare our result with the bowler model , the fragmentation function in the bowler model should be divided by the @xmath119 branching fraction @xmath170 @xcite .
the branching fraction is defined as @xmath171 where the cut @xmath172 excludes the @xmath17 range in which the result is not valid . as fig .
[ bowler ] shows our result at nlo is in reliable consistency with the phenomenological model . in this comparison
we set @xmath173 gev as in @xcite . concerning the data shown in figs . [ belle ] and [ dplus ]
, it should be noted that according to the definition of ff presented in @xcite , the ff @xmath174 is related to the differential cross section for the inclusive meson production as @xmath175 with normalization condition @xmath176 .
+ in @xcite , authors have compared their results for the fragmentation functions of @xmath177- and @xmath62-quarks into @xmath0- and @xmath178-mesons with various experimental data . in @xcite ,
suzuki have compared the ff of @xmath179 with the inclusive antineutrino data in the process @xmath180 . however , the mentioned authors ( suzuki @xcite , amiri , ji @xcite and _ etc _ ) define the ffs as the full differential hadron - level cross section , which is what is measured ( [ aydaa2 ] ) , but in phenomenological schemes ( [ monaa ] ) , one usually writes a hadronic cross section as a convolution of coefficient functions and fragmentation functions .
in fact , it is just a matter by definition and notation . as long as one is consistent , both definitions are possible : according to the definition
, one can say that the experimental data are a differential cross section or a fragmentation function .
similar issues hold for structure functions and parton distribution functions . in practice , it is also possible to show that both definitions are consistent at lo and at higher - orders they are equal approximately . according to the second definition , the cross section for @xmath10 annihilation can be expressed as in ( [ monaa ] ) .
+ at lo , the wilson coefficient functions are expressed as @xcite @xmath181 then one has @xmath182 which is the definition introduced by amiri , ji , suzuki and _ etc _ ( [ aydaa2 ] ) . at nlo approximation ,
the wilson coefficients read @xcite @xmath183 so that in high energy @xmath10 annihilation ( the condition applied by amiri and _ etc _ ) , the qcd coupling constant is tiny then the definition is approximately valid .
besides the @xmath184 ffs themselves , also their first moment is of phenomenological interest and subject to experimental determination .
it corresponds to the average fraction of energy that the @xmath185-mesons receive from the @xmath177 quark , @xmath186 where the cut @xmath187 excludes the problematic @xmath17 range where the formalism is not valid . on the other hand , as may be seen from fig .
[ belle ] there are no experimental data at @xmath188
. our results for the average energy fraction is @xmath189 at lo and @xmath190 at nlo approximations .
these results can be compared with the values quoted by belle , cleo , aleph and opal which are listed in @xcite .
also , if one takes the bowler model ( [ bw ] ) with the values of free parameters presented above , the result would be @xmath191 .
there is good consistency between our result and the phenomenological results , however one must keep in mind that experimental results naturally include all orders and also contributions from gluon and light - quark fragmentation , while ours are evaluated at nlo .
the @xmath192 ff at the initial scale @xmath120 as a function of @xmath17 in the pqcd approach ( dashed line ) and bowler model ( solid line ) .
the result in the pqcd approach is obtained at nlo approximation . ]
the dominant mechanism to produce hadronic bound states with large transverse momentum is fragmentation , that is the splitting of a high - energy parton into a hadronic state and other partons .
it is tempting to use the heavy - quark limits of the perturbative qcd fragmentation functions as phenomenological models for the fragmentation of a heavy quark q into heavy - light mesons @xmath33 , where @xmath193 or @xmath62 and @xmath194 , or @xmath153 . in this work , using the suzuki s model we studied the perturbative qcd fragmentation functions for a heavy quark to fragment into s - wave heavy - light mesons in the heavy - quark limit at nlo . in this model , the nonperturbative aspect of the hadroproduction processes
is emerged in the bound state of the meson which is described by the wave function .
our result describes not only the @xmath6 dependence of the fragmentation probabilities , but also their dependence on the transverse momentum of the initial parton . as a numerical example
, we studied the initial ffs of c - quark to split into s - wave @xmath195-mesons to leading order in @xmath1 and next - to - leading order .
specifically , we compared the lo and nlo ffs for @xmath195-mesons with available @xmath112 annihilation data from belle @xcite and cleo @xcite and we found good agreement between the nlo result and experimental data .
our results are also compared with a well - known phenomenological model ( bowler model @xcite ) for the heavy - quark fragmentation and found reliable consistency .
the full agreement between our result and the experimental data can reach by considering some additional effects ignored in this work such as , the fermi motion of constituent quarks , the meson wave function effects , etc .
it should be noted that , the next - to - next - to - leading order ( nnlo ) effects for the @xmath196 ff will be of order @xmath197 and these effects would not be expected to be sizable numerically . to include the nnlo effects , apart form the virtual corrections
, one needs to consider many real gluon feynman diagrams including the gluon , light and heavy quark propagators .
normally , we expect that the nnlo corrections would increase the ff at @xmath17-large and decrease then at @xmath17-low to make a better fit with the data , as we had for the nlo effect , see figs .
[ belle ] and [ dplus ] . s. m. moosavi nejad , eur .
j. a * 52 * ( 2016 ) no.5 , 127 .
r. seuster _ et al . _ [ belle collaboration ] , phys .
d * 73 * ( 2006 ) 032002 .
m. artuso _ et al . _
[ cleo collaboration ] , phys .
d * 70 * ( 2004 ) 112001 .
s. m. moosavi nejad , eur .
j. c * 72 * ( 2012 ) 2224 ; s. m. moosavi nejad and m. balali , eur .
j. c * 76 * ( 2016 ) no.3 , 173 ; s. m. moosavi nejad , nucl . phys .
b * 905 * ( 2016 ) 217 ; s. m. moosavi nejad and m. balali , phys .
d * 90 * ( 2014 ) no.11 , 114017 ; s. m. m. nejad , phys .
d * 88 * ( 2013 ) no.9 , 094011 ; b. a. kniehl , g. kramer and s. m. moosavi nejad , nucl .
b * 862 * ( 2012 ) 720 ; s. m. moosavi nejad , phys .
d * 85 * ( 2012 ) 054010 . | it is well - known that the dominant mechanism to produce hadronic bound states with large transverse momentum is fragmentation .
this mechanism is described by the fragmentation functions ( ffs ) which are the universal and process - independent functions . here
, we review the perturbative ffs formalism as an appropriate tool for studying these hadronization processes and detail the extension of this formalism at next - to - leading order ( nlo ) . using the suzuki s model
, we calculate the perturbative qcd ff for a heavy quark to fragment into a s - wave heavy meson at nlo . as an example
, we study the lo and nlo ffs for a charm quark to split into the s - wave @xmath0-meson and compare our analytic results both with experimental data and well - known phenomenological models . |
the recent experimental realization of a single layer of carbon atoms arranged in a honey - comb lattice has prompted much excitement and activity in both the experimental and theoretical physics communities @xcite .
carriers in graphene ( both electrons and holes ) have a linear bare kinetic energy dispersion spectra around the @xmath1 and @xmath2 points ( the dirac points " ) of the brillouin zone . the ability of experimentalists to tune the chemical potential to lie above or below the dirac point energy ( by application of voltages to gates in close proximity to the graphene sheets ) allows the carriers to be changed from electrons to holes in the same sample .
this sets graphene apart from other two dimensional ( 2d ) carrier systems that have a parabolic dispersion relation , and typically have only one set of carriers , i.e. either electrons or holes .
another unique electronic property , the absence of back - scattering , has led to the speculation that carrier mobilities of 2d graphene monolayers ( certainly at room temperature , but also at low temperature ) could be made to be much higher than any other field - effect type device , suggesting great potential both for graphene to be the successor to si - mosfet ( metal - oxide - semicondcutor field effect transistor ) devices and for the discovery of new phenomena that normally accompanies any significant increase in carrier mobility @xcite .
it is therefore of considerable fundamental and technological interest to understand the electronic properties of graphene @xcite .
graphene samples that are currently being fabricated are far from pure , based on the relatively low electronic mobilities compared to epitaxially grown modulation - doped two - dimensional electron gases ( 2degs ) such as gaas - algaas quantum wells .
it is therefore important to understand the effects of disorder on the properties of graphene .
disorder manifests itself in the finite lifetimes of electronic eigenstates of the pure system . in the presence of scattering from an impurity potential that is not diagonal in the these eigenstates ,
the lifetime scattering rate of the eigenstate , @xmath3 , is non - zero and can be measured experimentally by fitting the line - shape of the low - field shubnikov - de haas ( sdh ) oscillations . @xcite the effect of disorder scattering on the sdh line shape is equivalent to increasing the sample temperature and one can therefore measure this dingle temperature ( @xmath4 ) and relate it to the single - particle lifetime through @xmath5 . to avoid potential confusion , we mention that the lifetime damping rate @xmath3 discussed in this paper is _ not _ equal to the _ transport _ scattering rate which governs the electrical conductivity .
the lifetime damping rate is the measure of the rate at which particles scatters out of an eigenstate , whereas the transport scattering rate is a measure of the rate of _ current decay _ due to scattering out of an eigenstate . in normal 2degs
, the transport scattering time can be much larger than the impurity induced lifetime @xmath6 , particularly in high mobility modulation doped 2d systems where the charged impurities are placed very far from the 2deg@xcite .
recently , the issue of transport scattering time versus impurity scattering lifetime in graphene has been discussed.@xcite the single - particle level broadening due to the impurity potential changes many of the physical properties of the system including the electronic density of states.@xcite the electronic density of states is an important property which directly affects many experimentally measurable quantities such as the electrical conductivity , thermoelectric effects , and differential conductivity in tunneling experiments between graphene and scanning - tunneling microscope tips or other electron gases .
changes in the density of states also modify the electron screening,@xcite which is an important factor in the determination of various properties of graphene .
it is therefore imperative to take into account the effects of disorder on the density of states , particularly since disorder is quite strong in currently available graphene samples . in the present work ,
we present calculations of the average density of states of disordered graphene .
this problem has been investigated using various models and techniques , both analytical and numerical.@xcite we take into account scattering effects from long - range and short range impurity potentials .
we consider both unscreened and screened short - ranged and screened charged impurities , using the born approximation .
in addition , for unscreened short - ranged ( usr ) impurities , we go beyond the born approximation and include self - consistent effects .
there is another class of disorder in graphene called off - diagonal " or
random gauge potential " disorder , in which the hopping matrix elements of the electrons in the underlying honeycomb lattice are random . in this paper
we do not consider in this type of disorder , which can result from height fluctuations ( ripples ) in the graphene sheet and lead to qualitatively different results from the ones presented in this paper.@xcite the rest of the paper is organized as follows : in sec .
[ sec : approx ] , we describe the approximation schemes that we use . secs .
[ sec : usr ] and [ sec : screened ] deal with unscreened short range impurities and screened short - range / charged impurities , respectively . in sec .
[ sec : comparison ] , we compare our results to those from other workers , and we conclude in sec . [
sec : conclusion ] .
the single - particle density of states for a translationally invariant 2deg is given by@xcite @xmath7 \label{eq : dos}\end{aligned}\ ] ] where @xmath8 is the degeneracy factor ( for graphene @xmath9 due to valley and spin degeneracies ) @xmath10 is the band index , @xmath11 is the retarded green s function , and the @xmath12-integration is over a single valley , which we assume to be a circle of radius @xmath13 .
@xmath11 expressed in terms of the retarded self - energy @xmath14 is @xmath15^{-1}.\label{eq : g}\ ] ] where @xmath16 is the bare band energy of the state @xmath17 and @xmath18 is an infinitesimally small positive number .
( in this paper , the green s functions and self - energies are all assumed to be retarded . )
( [ eq : dos ] ) and ( [ eq : g ] ) show that if @xmath19 and @xmath20 are known , the density of states can be obtained in principle from @xmath21+\eta}{(e - e_{\bm k,\lambda } - { \rm re}[\sigma_\lambda(\bm k , e)])^2+({\rm im}[\sigma_\lambda(\bm k , e)]-\eta)^2}. \label{eq : dos1}\ ] ] for pure graphene systems , @xmath22 ( excluding electron electron and electron phonon interactions , which are not considered here ) , and hence @xmath23 = -\pi \delta(e - e_{\bm k\lambda})$ ] .
close to the dirac points ( which we choose to the the zero of energy ) , the dispersion for graphene is ( we use @xmath24 throughout this paper ) @xmath25 where @xmath26 and @xmath27 for the conduction and valence bands , respectively , @xmath28 is the wavevector with respect to the dirac point , and @xmath29 is the fermi velocity of graphene .
performing the @xmath12-integration in eq .
( [ eq : dos ] ) for the pure graphene case gives @xmath30 where @xmath31 is the band energy cut - off .
the average density of states for a disordered 2deg can be obtained by averaging the green s function over impurity configurations .
the averaging procedure gives a non - zero @xmath32 which , in general , can not be evaluated exactly .
various approximation schemes for @xmath32 have therefore been developed , four of which are described below .
= 1.7 in _ born approximation _ in the born approximation , the self - energy is given by the feynman diagram shown in fig . 1(a)@xcite , and the expression for the self - energy is @xmath33 where @xmath34 is the impurity density
, @xmath35 is the fourier transform of the impurity potential , @xmath36 is the bare green s function and @xmath37 is square of the the overlap function between the part of the wavefunctions of @xmath38 and @xmath39 that are periodic with the lattice ( here @xmath40 are band indices ) . for graphene states near the dirac point , @xmath41 where @xmath42 is the angle between @xmath12 and @xmath43 .
_ self - consistent born approximation _ the feynman diagram for this self - energy is the same as fig .
1(a ) , except that the bare green s function is replaced by the full one . consequently
, the expression for @xmath44 is the same as in eq .
( [ eq : sigma_b ] ) , except with @xmath36 replaced by @xmath11 .
_ @xmath0-matrix approximation _ the @xmath0-matrix approximation is equivalent to the summation of feynman diagrams shown fig
the expression for @xmath45 is the same as in eq .
( [ eq : sigma_b ] ) , except with @xmath46 replaced by @xmath0 , the @xmath0-matrix for an individual impurity . _ self - consistent @xmath0-matrix approximation _ the feynman diagram for this approximation is the same as in the @xmath0-matrix approximation , except that the bare green s functions are replaced by full ones . the expression for @xmath47 is the same as in eq .
( [ eq : sigma_b ] ) , except with @xmath46 replaced by @xmath0 , and @xmath36 replaced by @xmath11
. when the potentials for the impurities are not all identical ( for example , in the case where there is a distribution of distances of charged impurities from the graphene sheet ) , one averages @xmath48 or @xmath49 over the impurities .
in the present context , short - ranged impurities are impurities which result from localized structural defects in the honeycomb lattice , which are roughly on the length scale of the lattice constant . in this case
, it is acceptable to approximate @xmath50 , a real constant , for intravalley scattering processes . in this paper
, we ignore the intervalley processes .
( however , we note that if the matrix element joining intervalley states is constant , inclusion of intervalley scattering in our calculations is not difficult . ) this simplification allows us to obtain some analytic expressions for the self - energies in the approximation schemes mentioned above . for usr disorder , the four approximation schemes we use
give self - energies that are independent of @xmath10 and @xmath51 . the self - energy for graphene with usr scatterers in the born approximation , using @xmath52 ,
( [ eq : dirac_dispersion ] ) and ( [ eq : f ] ) in eq .
( [ eq : sigma_b ] ) , is @xmath53 as a function of complex @xmath54 , @xmath55 is real and positive ( negative ) along the real axis from @xmath56 to @xmath57 ( @xmath58 to @xmath59 ) .
furthermore , it has a branch cut in on the real axis of @xmath54 in between @xmath58 and @xmath56 , so that for @xmath60 , @xmath61 fig .
2 shows @xmath62 . for real @xmath63 and @xmath64 , @xmath65.\ ] ] the born approximation damping rate for state @xmath66 is @xmath67 = \tilde\gamma_b\,v_f k$ ] .
[ see eq .
( [ eq : h_0e ] ) ] , where @xmath68 is the self - energy for usr impurities in the born approximation . ] the self - energy @xmath69 for unscreened short - ranged scatterers in the self - consistent born approximation is given by the self - consistent equation @xmath70 \ln\left[1 - \frac{e_c^2 } { [ \zeta-\sigma_{\rm sb}^{\rm ( usr)}(\zeta ) ] ^2}\right ] \nonumber \\ & = h_0\left [ \zeta-\sigma_{\rm sb}^{\rm ( usr)}(\zeta)\right ] , \label{eq : sigma_sb}\end{aligned}\ ] ] this shows that if @xmath71 , then @xmath72 ( except possibly around @xmath73 , which is usually not experimentally relevant ) .
in general the impurity averaged @xmath0-matrix for a potential @xmath74 is @xmath75.\label{eq : tmatrix}\ ] ] in this approximation , the self energy is @xmath76 if @xmath52 , a constant , the term in the square parentheses in eq .
( [ eq : tmatrix ] ) is @xmath77 .
the sum then gives @xmath78 as in self - consistent born approximation , the @xmath79 in eq .
( [ eq : sigma_t ] ) is replaced by the self - consistent @xmath80 , giving @xmath81 as in the self - consistent born approximation , this shows that if @xmath82 , then @xmath83 ( except possibly around @xmath73 ) .
when the self - energy is @xmath12- and @xmath10-independent , the density of states for graphene can be calculated analytically from eq .
( [ eq : dos1 ] ) .
the result is @xmath84\biggr\},\label{eq : dos2}\end{aligned}\ ] ] where @xmath85 $ ] and @xmath86 $ ] .
there has been considerable interest in the minimum dc electrical conductivity of disordered graphene as the fermi energy moves through the dirac point.@xcite there is still no consensus on whether the minimum conductivity is a universal value or not .
since the electrical conductivity is directly proportional to the density of states at the fermi energy , it is important to be able to determine the density of states at the dirac point of disordered graphene .
because the minimum conductivity is non - zero as the fermi energy passes through the dirac point , the density of states should be nonzero .
( [ eq : dos1 ] ) shows that if @xmath87 when @xmath88 , then the density of states at the dirac point @xmath89 .
( note that we have not included the term that is first order in the impurity potential in our self - energy . since this first - order term merely rigidly shifts the band by an amount @xmath90 , ignoring this term is equivalent to shifting the zero of the energy by @xmath91 , and hence the dirac point is still at @xmath92 . )
a non - zero density of states at the dirac point depends on a non - zero @xmath93 $ ] . since @xmath94
, it is clear from eqs .
( [ eq : sigma_b1 ] ) and ( [ eq : sigma_t ] ) that the born and @xmath0-matrix approximations give zero density of states at the dirac point . for the self - consistent born approximation , eq .
( [ eq : sigma_sb ] ) can be rewritten as @xmath95^{-1},\end{aligned}\ ] ] which shows that @xmath96 , and therefore the self - consistent born approximation also gives @xmath89 . in the case of the self - consistent @xmath0-matrix approximation , re - writing eq ( [ eq : sigma_st ] ) as @xmath97 setting @xmath92 , using eq .
( [ eq : h_0 ] ) and taking the imaginary parts of both sides of this equation gives @xmath98 \equiv\gamma(0 ) & = { \rm i m } \left[\frac{n_i u_0}{\displaystyle 1-i\frac{u_0\gamma(0)}{2 v_f^2}\ln\left(1+\frac{e_c^2}{\gamma(0)^2}\right)}\right ] .
\label{eq : gamma0}\end{aligned}\ ] ] in the weak scattering limit , when @xmath99 , the imaginary term in the denominator of eq .
( [ eq : gamma0 ] ) is much less than one ( we check for self - consistency later ) , and this gives @xmath100 = 2\tilde\gamma_b\gamma_w(0 ) \ln\left(\frac{e_c}{\gamma_{w}(0)}\right),\end{aligned}\ ] ] which implies that @xmath101 note that the result is non - analytic in @xmath102 . inserting eq .
( [ eq : gamma0result ] ) into the imaginary part of the denominator of eq .
( [ eq : gamma0 ] ) ( which we had assumed to be much smaller than 1 in magnitude ) gives the self - consistent criterion @xmath103 for the validity of eq .
( [ eq : gamma0result ] ) . substituting this into eq .
( [ eq : dos2 ] ) gives an average density of states at the dirac point for weak scattering of approximately @xmath104 similar results to eq .
( [ eq : dosweak ] ) have been reported@xcite in studies of disordered systems of fermions with linear dispersions using other methods .
we mention that our calculation of the graphene density of states at the dirac point should only be considered as demonstrative since electron - electron interaction effects are crucial @xcite at the dirac point , and the undoped graphene system is not a simple fermi liquid at the dirac point .
free carriers will move to screen a bare impurity potential @xmath105 , resulting in a screened interaction @xmath106 , where @xmath107 is the static dielectric function .
the @xmath107 results in an @xmath108-dependent effective electron - impurity potential @xmath46 , even in the case of short - ranged ( @xmath108-independent ) bare impurity potentials .
this makes the calculations much more involved than in the usr case .
therefore , in this paper , we limit our investigation of @xmath108-dependent screened potentials to the level of the born approximation . for the dielectric function @xmath107 , we use the random phase approximation ( rpa ) for dielectric function appropriate for graphene , given by @xmath109 where @xmath110 is the two - dimensional fourier transform of the coulomb potential ( @xmath111 is the dielectric constant of the surrounding material ) , and @xmath112 is the static irreducible rpa polarizability for graphene.@xcite we use @xmath113 for charged impurities , and @xmath114 , a constant , for short - range point defect scatterers .
we first look at the density of states at the fermi surface ; _ i.e. _ , at energy @xmath115 . to obtain this
, we calculate the single - particle lifetime damping rate @xmath3 in the born approximation , which is given by @xmath116 \nonumber \\ & = & \frac{n_i}{2\pi } \frac{k}{v_f}\int_0^{\pi}d\theta\ \frac { \langle\left |{v_{ei}(q)}\right |^2\rangle}{\varepsilon(q)^2 } ( 1+\cos\theta ) , \label{scat_s}\end{aligned}\ ] ] where @xmath117 .
then , assuming that @xmath118 $ ] is relatively constant for @xmath51 close to @xmath119 , we substitute @xmath120 for @xmath121 $ ] into eq .
( [ eq : dos1 ] ) , which gives eq .
( [ eq : dos2 ] ) with @xmath122 .
we assume that the charged or neutral impurities are distributed completely at random on the surface of the insulating substrate on which the graphene layer lies , the areal density for the charged and neutral impurities is @xmath123 and @xmath124 , respectively , and the density of carriers in the graphene layer is @xmath125 , where @xmath119 is the fermi wavevector relative to the dirac point .
( this relationship between @xmath126 and @xmath119 takes into account the spin and valley degeneracy @xmath127 and @xmath128 . )
we use the rpa screening function at @xmath129,@xcite to obtain the effective impurity potential .
the key dimensionless parameter that quantifies the screening strength is @xmath130 , which is corresponding to the interaction strength parameter of a normal 2d system ( _ i.e. _ the ratio of potential energy to kinetic energy ) .
the born approximation lifetime damping rates for screened charged - impurities @xmath131 and @xmath132-correlated neutral impurities @xmath133 at @xmath119 are @xmath134 [ eq : gamma ] in these equations , @xmath135 where @xmath136 & \mbox{for $ x < 1 $ } ; \\ 1 & \mbox{for $ x = 1 $ } ; \\
\frac{1}{\sqrt{x^2 - 1}}\cos^{-1}\frac{1}{x } & \mbox{for $ x > 1$}. \end{array } \right .
\label{fx}\end{aligned}\ ] ] in fig .
[ damping ] we show the calculated damping rates scaled by @xmath137 as a function the interaction parameter @xmath138 .
for @xmath139 we have @xmath140 and @xmath141 .
for @xmath142 we have @xmath143 and @xmath144 .
thus , for small ( large ) @xmath138 the damping rate due to the short - ranged impurity dominates over that due to the long - ranged charged impurity . on the other hand , since @xmath145 and @xmath146 [ eq . ( [ eq : gamma ] ) ] , in the low ( high ) carrier density limit the lifetime damping of single particle states at the fermi surface is dominated by charged impurity ( short - ranged impurity ) scattering .
the crossover takes place around a density @xmath147 as a function of @xmath138 .
graphene on a sio@xmath148 ( air ) substrate has an @xmath149 ( 2 ) .
solid lines indicate damping rates ( @xmath131 ) due to charged impurities with an impurity density @xmath150 for different electron densities @xmath151 , 10 , 50@xmath152 @xmath153 ( from top to bottom ) , respectively . dashed line indicate the damping rate ( @xmath154 ) due to short - ranged impurity with impurity density @xmath155 @xmath153 and potential strength @xmath1561 kev @xmath157 , which correspond to @xmath158 .
note @xmath159 is independent on the electron density . ]
using @xmath160 in eq .
( [ eq : dos2 ] ) gives ( assuming @xmath161 , @xmath162 ) @xmath163.\end{aligned}\ ] ] for @xmath164 , this gives @xmath165 \right\ } , \label{eq : dosdelta}\ ] ] and for @xmath166 @xmath167 + \frac{g_sg_v}{2\pi^2 } \frac{\gamma } { v_f^2}\ln\left(\frac{e_c}{\gamma}\right ) .
\label{eq : doscharge}\ ] ] we can apply eq .
( [ eq : dosdelta ] ) for short - ranged impurity scattering and for charged impurity scattering in high carrier density limits , and eq .
( [ eq : doscharge ] ) for charged impurity scattering in low density limits . taking the limit @xmath168 in eq .
( [ eq : doscharge ] ) , it appears that for the case of screened charged impurities , we obtain a finite density of states at the dirac point with the born approximation ( since @xmath169 ) . however , recall that in deriving eqs .
( [ eq : dosdelta ] ) and ( [ eq : doscharge ] ) , we have assumed that @xmath170 is constant with respect to @xmath51 , which is not necessarily the case at the dirac point . for screened coulomb scattering potential ( solid lines ) and for screened neutral short - ranged scattering potential ( dashed lines).,title="fig : " ] for screened coulomb scattering potential ( solid lines ) and for screened neutral short - ranged scattering potential ( dashed lines).,title="fig : " ] in general , the damping rate ( or the imaginary part of the self energy ) is a function of energy and wave vector rather than a constant . from eq .
( [ eq : sigma_b ] ) , we calculated the self - energy of disordered graphene .
[ self ] we show the self - energy of a conduction band electron ( @xmath171 ) for both screened coulomb scattering potential and screened neutral short - ranged potential . for coulomb scatterers we use the impurity density @xmath172 @xmath153 , and for neutral short - ranged scatterers the impurity density @xmath155 @xmath153 and potential strength @xmath1561 kev @xmath157 .
the self - energies in the valence band ( @xmath173 ) are related to the self - energy in the conduction band by @xmath174 and @xmath175 . as @xmath176 , @xmath177 , and @xmath178 for both scattering potentials . however , for large value of @xmath179 the asymptotic behaviors are different , that is , as @xmath180 @xmath181 for coulomb scattering potential and @xmath182 for short - ranged potential .
note that by using only the ( non - self - consistent ) born approximation in this section , we assume weak scattering and ignoring multiple - scattering events in calculated @xmath183 .
therefore , the results are unreliable in the strong disorder limit ( _ i.e. _ , when @xmath183 is modified significantly from its lowest - order form ) . , 1 , @xmath184 @xmath153 ( from bottom to top ) , and in ( b ) the short - ranged impurity density @xmath185 , 0.5 , @xmath186 @xmath153 ( from bottom to top ) and potential strength @xmath1871 kev @xmath157 .
, title="fig : " ] , 1 , @xmath184 @xmath153 ( from bottom to top ) , and in ( b ) the short - ranged impurity density @xmath185 , 0.5 , @xmath186 @xmath153 ( from bottom to top ) and potential strength @xmath1871 kev @xmath157 .
, title="fig : " ] in fig .
[ dos_f ] the density of states in the presence of impurity is shown for different impurity densities . in fig .
[ dos_f](a ) we show the density of states for coulomb impurity potential with impurity densities , @xmath188 , @xmath189 @xmath153 , @xmath184 @xmath153 ( from bottom to top ) , and in fig .
[ dos_f](b ) we show the density of states for neutral short - ranged impurity potential with densities , @xmath185 , @xmath190 @xmath153 , @xmath189 @xmath153 ( from bottom to top ) and potential strength @xmath1871 kev @xmath157 .
the calculated density of states is normalized by @xmath191 .
the density of states is enhanced near dirac point ( @xmath92 ) , but as @xmath192 it goes zero as @xmath193 for both types of screened impurity scattering .
based on the results of section iii , we expect that this result is an artifact of the born approximation , and that the density of states should in fact be non - zero .
the enhancement of the density of states can be explained as follows . in normal 2d system with finite disorder
the band edge @xmath194 of the pure conduction ( valence ) band is shifted to @xmath195 ( @xmath196 ) and a band tail forms below ( above ) the band edge of a pure system .
thus , the density of states in the presence of impurities is reduced for @xmath197 ( @xmath198 ) because the states have been shifted by the impurity potential into the band tail .
however , for graphene since the conduction band and the valence band meet at the dirac point the band tail ( or shift of band edge ) can not be formed , which gives rise to enhancement of density of states near dirac point . before concluding we point out that our perturbative calculation of the graphene density of states assumes that the system remains homogeneous in the presence of impurities .
it is , however , believed @xcite that graphene carriers develop strong density inhomogeneous ( i.e. electron - hole puddle ) at low enough carrier densities in the presence of charged impurities due to the breakdown of linear scattering .
in such an inhomogeneous low - density regime close to the dirac point , our homogeneous perturbative calculation does not apply .
in this section , we compare and contrast our model of disorder and results to other works in the field .
et al._@xcite studied the effect of disorder in graphene by considering the effect of vacancies on the honeycomb lattice . for a finite density of vacancies , they found that the density of states at the dirac point is zero for the full born approximation " ( equivalent to our @xmath0-matrix approximation ) and non - zero for the full self - consistent born approximation " ( equivalent to our self - consistent @xmath0-matrix approximation ) .
our results are consistent with theirs , even though the regimes that are studied are different .
vacancies correspond to the limit where the impurity potential @xmath199 , whereas this work is more concerned with the weak impurity - scattering limit . pereira _ at al._@xcite considered , among several different models of disorder , both vacancies and randomness in the on - site energy of the honeycomb lattice .
they numerically calculated the density of states for these models of disorder . for compensated vacancies ( same density of vacancies in both sub - lattices of the honeycomb structure )
they found that the density of states increased around the dirac point .
( ref . also studied the case of uncompensated vacancies , but that has no analogue in our model of disorder . ) for the case of random on - site impurity potential , they find that `` there is a marked increase in the dos ( density of states ) at @xmath200 ( the dirac point ) '' and the dos becomes finite at @xmath200 with increasing concentration " of impurities .
our self - consistent @xmath0-matrix approximation result is consistent with their numerical results , although it should be mentioned again that it strictly does not apply to the case of vacancies .
et al._@xcite numerically investigated the average density of states of graphene for the case of on - site disorder .
they found that for weak disorder , the density of states at the dirac point increased with both increasing density of impurities and strength of the disorder .
( in their work , they did not absorb the shift in the band due to the impurity potential in their definition of the energy , so the dirac point had a shift @xmath201 where @xmath202 is the concentration of impurities and @xmath203 is the on - site impurity energy . ) their numerical results for the minimum in the average density of states ( fig .
3(b ) in ref . ) seem to indicate a non - linear dependence of the value of the minimum as a function of the strength of the disorder potential , and is at least not inconsistent with eq .
( [ eq : dosweak ] ) .
the issue of the effect of screening of the impurity interactions on the density of states discussed in section iv , to the best of our knowledge has not yet been treated in the literature .
qualitatively , the effect of screened impurities away from the dirac point is the increase the density of states , and is consistent with numerical results for random on - site disorder.@xcite ( since our treatment of screened impurities is at the level of the born approximation , we do not obtain either a non - zero density of states at the dirac point nor resonances in the density of states.@xcite )
we have calculated the density of states for disordered graphene . in the case of unscreened short - ranged impurities
, we utilized the non - self - consistent and self - consistent born and @xmath0-matrix approximations to calculate the self - energy . among these ,
only the self - consistent @xmath0-matrix approximation gave a non - zero density of states at the dirac point , and the density of states is a non - analytic function of the impurity potential .
we investigated the density of states in the case of screened short - ranged and charged impurity potentials at the level of the born approximation .
we find that , unlike the case of parabolic band 2degs , in graphene near the band - edge ( _ i.e. _ , the dirac point ) the density of states is enhanced by impurities instead of being suppressed . at very low carrier densities , however , graphene develops strong carrier density inhomogeneity in the presence of charged impurities , an effect not captured by the homogeneous many - body theory in our description .
morozov , k.s .
novoselov , m.i .
katsnelson , f. schedin , d.c .
elias , j.a .
jaszczak , a.k .
geim , phys .
. lett . * 100 * , 016602 ( 2008 ) ; j. h. chen , c. jang , s. xiao , m. ishigami , m. s. fuhrer , nature nanotechnology * 3 * , 206(2008 ) ; e. h. hwang and s. das sarma , phys .
b * 77 * , 115449 ( 2008 ) .
victor m. periera , f. guinea , j. m. b. lopes dos santos , n. m. r. peres , and a. h. castro neto , phys .
. lett . * 96 * , 036801 ( 2006 ) ; victor m. periera , j. m. b. lopes dos santos , and a. h. castro neto , phys .
b * 77 * , 115109 ( 2008 )
. see _ e.g. _ , g. rickaysen , _
green s functions and condensed matter _
( academic press , new york , 1980 ) ; s. doniach and e. h. sondheimer , _ green s functions for solid state physicists _ ( imperial college , london , 1998 ) ; g. d. mahan , _ many particle physics _ , 3rd ed .
( plenum , new york , 2000 ) . | we calculate the average single particle density of states in graphene with disorder due to impurity potentials . for unscreened short - ranged impurities ,
we use the non - self - consistent and self - consistent born and @xmath0-matrix approximations to obtain the self - energy . among these
, only the self - consistent @xmath0-matrix approximation gives a non - zero density of states at the dirac point .
the density of states at the dirac point is non - analytic in the impurity potential . for screened short - ranged and charged long - range impurity potentials ,
the density of states near the dirac point typically increases in the presence of impurities , compared to that of the pure system . |
chirality in liquid crystals produces a fascinating variety of phases , such as the blue phases .
three blue phases ( bps ) designated bp i , bp ii , and bp iii have been identified , and their structures are now well understood .
bp i and bp ii exhibit long - range periodic order at the half micron scale and bragg scatter visible light . for these reasons ,
the optical study of blue phases is an active field of research , with a particular emphasis on spatial dispersion effects such as the optical activity . the first experiments on the optical activity in the pretransitional region of the isotropic phase above the isotropic - cholesteric transition were carried out by cheng and meyer @xcite .
using a general formulation due to de gennes , they calculated and confirmed experimentally that the pretransitional optical activity depends on the temperature as @xmath2 , where @xmath3 is the metastability temperature of the isotropic phase .
two years later , brazovskii and dmitriev developed the first complete theory of phase transitions in cholesteric liquid crystals , and they predicted the existence of the blue phases @xcite .
the optical activity in the pretransitional region was derived using this approach by dolganov et al .
@xcite . at this time , a detailed landau theory of the cholesteric blue phases was obtained by hornreich and shtrikman , who also provided an outstanding study of the light scattering in the blue phases including a treatment of the polarization of the light , based on the formalism of the mller matrices @xcite .
bensimon , domany , and shtrikman studied the optical activity in the pretransitional regime and in the blue phases @xcite , confirming and extending the work of dolganov . in their paper , the blue phase bp iii was considered as an amorphous polycrystalline structure distinct from the isotropic phase , and their treatment of the optical activity relied on the long wavelength approximation . on the experimental side ,
the wavelength dependence of the optical activity in the blue phases was measured by collings @xcite .
the structure of bp iii still remained mysterious until experiments clearly showed a continuous transition between bp iii and the isotropic phase @xcite and the existence of a critical point terminating a line of coexistence of the two phases @xcite .
the same year , lubensky and stark developed a theory for the isotropic to bp iii transition with a chiral liquid - gas like critical point @xcite , which was extended to include scaling theory in ref .
very recently , there has been a renewal of interest in the study of the blue phases with the discovery of the smectic blue phases @xcite , exhibiting isotropic - bp iii coexistence terminating at a critical point @xcite . in section [ sec : theory ] of this paper , we introduce an effective index for the propagation of light in the isotropic phase and in bp iii , which contains pretransitional effects such as the optical activity and the circular dichroism . in section [ sec : long_wavelength ] , we present a standard derivation which is based on the long wavelength approximation @xcite . in section
[ sec : de vries ] , we make a digression on the optical properties of periodic chiral media for comparison with the case of non - periodic chiral media . by evaluating our weak - scattering expression for the optical dielectric constant tensor in a cholesteric phase , we find the well - known de vries formula , which describes the optical activity of light propagating in this medium along the pitch axis . in section [ sec : exact_wavel ] we come back to isotropic chiral media and obtain the optical activity for arbitrary wavelength and chirality .
this approach is new in that it is general , not limited to the long wavelength approximation , and applicable in particular to the resonant region where the wavelength of the light is of the order of the pitch of the cholesteric , a regime which has not been considered in detail in previous studies . in the long wavelength approximation
, we collect contributions to the optical activity up to order @xmath4 .
though various contributions up to this order have appeared separately in the literature @xcite and very recently in @xcite , these references are not fully consistent with each other . we therefore think that it is important to collect and discuss all these contributions in one place .
we then generalize our method and obtain the wavelength dependance of the circular dichroism and of the symmetric part of the index , which to our knowledge have not appeared in the literature and have never been measured .
finally in section [ sec : theory&exp ] , our theoretical predictions are supported by experimental spectra of the optical activity for different values of the chirality in the bp iii phase @xcite .
the agreement between theory and experiments is very good .
experiments confirm the presence of a broad maximum in the magnitude of the optical activity when the wavelength is of the order of the pitch and for positive ( respectively negative ) pitch .
this feature is the pretransitional signature in isotropic chiral media of the divergence arising from the long - range periodicity of the cholesteric phase , bp i and bp ii .
let @xmath5 be the relative complex - valued dielectric constant of the cholesteric liquid crystal in the isotropic phase or in the blue phase bp iii with respect to the average dielectric constant : @xmath6 where @xmath7 is a relative anisotropic small local fluctuation of the dielectric constant : @xmath8 , @xmath9 , and @xmath10 denotes a thermal average . to simplify the notations , we take the velocity of light in the medium to be @xmath11 , and we do not indicate explicitly the dependence of the dielectric constant on the frequency of light in the medium denoted @xmath12 .
maxwell equations in the absence of sources lead to the following helmholtz equation for the electric field @xmath13 : @xmath14 from eq .
( [ helmoltz ] ) , we deduce that the propagation of light in a homogeneous medium in the absence of fluctuations is characterized by the following green s function @xmath15 where @xmath16 is a projector on the space transverse to the wave vector @xmath17 .
the first term of eq .
( [ green_libre ] ) is the transverse part describing travelling wave solutions of maxwell equations and the second term is the longitudinal part which describes non - propagating modes . the scattering by the randomly fluctuating part of the dielectric function @xmath7 is described by the following 4-rank tensor @xcite @xmath18 unless specified otherwise , we consider in this paper only non - absorbing media in which the tensor @xmath19 is real .
the averaged green s function @xmath20 follows from dyson s equation , @xmath21 in the weak scattering approximation , the tensor @xmath22 can be calculated using the following equation @xcite : @xmath23 in terms of the fourier transform @xmath24 of the tensor defined in eq .
( [ def_tensor_b ] ) .
note that eq .
( [ sigma ] ) is general and applies also to periodic media for which @xmath24 has delta - function peaks in which case this equation takes the form of a discrete sum over bloch waves @xcite .
the general form of @xmath25 in an isotropic chiral medium is : @xmath26 if @xmath27 denotes a chiral parameter of the medium , @xmath28 satisfy the general symmetry relation ( valid in any medium in the absence of a magnetic field @xcite ) @xmath29 and the chiral symmetry relation @xmath30 the chiral symmetry relation implies that @xmath31 and @xmath32 are odd functions of @xmath27 and that @xmath33 is an even function of @xmath27 . the dispersion relation for light propagation modified by fluctuations is obtained from @xmath34=0 $ ] . in a basis composed of two vectors perpendicular to @xmath17 , the diagonal elements of @xmath22 are equal to @xmath31 , and the off - diagonal elements are @xmath35 . in such a basis ,
longitudinal terms such as the last term in eq .
( [ def_sigma12 ] ) vanish , and the following relation is obtained @xmath36 this relation determines @xmath37 and @xmath38 as a function of the frequency @xmath12 . writing @xmath39 and expanding eq .
( [ dispersion ] ) to first order in @xmath40 , one finds @xmath41 to first order in @xmath22 , this gives @xmath42/2\omega$ ] .
the optical activity @xmath43 is defined as the angle through which the polarization vector of linearly polarized light has turned when traversing a medium of length @xmath44 .
it is given by @xmath45 .
similarly the circular dichroism , measuring the difference in transmission from left and right circularly polarized waves , is @xmath46 . to first order in @xmath22
, we have therefore @xmath47 and @xmath48 note that these relations are valid at arbitrary frequency @xmath12 when eq .
( [ sigma ] ) is valid .
the fluctuations also modify the symmetric part of @xmath22 , , the average dielectric constant .
the relative change of the average dielectric constant within the same approximation is @xmath49 .
we use in this paper the form of the landau - de gennes free energy of ref .
@xcite , @xmath50 -\beta \epsilon_{ij } \epsilon_{jl } \epsilon_{li } + \gamma ( \epsilon_{ij}^2)^2 \},\ ] ] where @xmath51 and @xmath52 is the levi - civita fully antisymmetric tensor .
the coefficient @xmath53 is proportional to the reduced temperature @xmath54 , whereas all other coefficients are assumed to be temperature independent . as explained in refs .
@xcite and @xcite , it is convenient to write this free energy in dimensionless form , by expressing all lengths in units of the order parameter correlation length @xmath55 .
the chirality is then @xmath56 where @xmath57 is the wave number of the cholesteric phase which determines the pitch @xmath58 .
the reduced temperature is defined by @xmath59 where @xmath60 is the temperature corresponding to the limit of metastability of the isotropic phase @xcite , and @xmath61 is the ratio of the two landau coefficients , which is of order 1 . using the gaussian approximation for the fluctuations @xcite , the dielectric anisotropy correlation tensor defined in eq .
( [ def_tensor_b ] ) takes the following form @xmath62 where @xmath63 is evaluated from the equipartition theorem @xmath64 } .\ ] ] note that @xmath65 denotes the dimensionless wavevector measured in units of @xmath66 , and that eq .
( [ def_gamma ] ) has been corrected from the expression given in ref .
@xcite ( the factor @xmath67 has been replaced by @xmath68 in agreement with ref .
@xcite ) . in eq .
( [ def_gamma ] ) , the parameter @xmath27 represents an effective chirality at the temperature @xmath54 .
except at the critical point , @xmath27 is different from @xmath69 , and the difference @xmath70 is proportional to the order parameter of the bp iii - isotropic phase transition @xcite . the index @xmath71 in eq .
( [ tensor_b ] ) denotes the five independent modes of the symmetric and traceless part of the dielectric tensor @xmath7 .
note that eq .
( [ def_gamma ] ) is equivalent to @xmath72 with @xmath73 , @xmath74 , @xmath75 and @xmath76 .
we have introduced @xmath77 , the transition temperature of the mode @xmath71 , and @xmath78 , the wavevector which minimizes the energy of the mode @xmath71 . denoting @xmath79 , we have explicitly @xmath80 and @xmath81 for the modes @xmath82 , @xmath83 and @xmath84 for the modes @xmath85 , @xmath86 and @xmath87 for the mode @xmath88 . a plot of @xmath63 as a function of @xmath65
is shown in fig .
[ fig : gamma ] for the modes @xmath89 and @xmath90 , a chirality @xmath91 and a temperature @xmath92 .
this particular plot shows the dominance of the mode @xmath93 , because of the choice of the temperature @xmath94 and @xmath54 close to @xmath95 .
more generally it can be shown that as @xmath96 , @xmath63 becomes a dirac function localized at @xmath97 .
the tensor @xmath98 introduced in eq .
( [ def_tensor_b ] ) are eigenvectors of the tensor @xmath24 , with eigenvalues @xmath99 @xcite .
these tensors are @xmath100 @xmath101=t^{-1}(\q)^*,\ ] ] @xmath102 with @xmath103 and @xmath104 forming a right - handed system of orthonormal vectors .
we choose the vectors @xmath105 and @xmath106 such that @xmath107 and @xmath108 , so that @xmath109
in this section the optical activity and the circular dichroism of isotropic chiral media is derived using the method of ref .
@xcite , which is valid in the long wavelength approximation and to linear order in @xmath27 . in section [ sec : exact_wavel ] , we will present another derivation of these results , which does not rely on an expansion in @xmath110 or in @xmath27 .
we present here the method of ref .
@xcite because it has been widely used in the literature .
following ref .
@xcite , we use an expansion to first order in @xmath111 of the green s function @xmath112 with @xmath113 using this expression in eq .
( [ sigma ] ) , we obtain @xmath114 which was defined in eq .
( [ def_sigma12 ] ) .
after integration over @xmath115 , we obtain @xmath116 } { 12{\pi } ^{2}\left({\omega}^2-{q}^2+i0^+ \right)^{2}}},\ ] ] where @xmath117 . in the integration of eq .
( [ sigmar ] ) , we take the upper limit of integration over @xmath65 to be infinity .
the continuous model used in this paper is valid only up to wave vectors @xmath118 , where @xmath53 is an intermolecular distance , but in the case of eq .
( [ sigmar ] ) , the integral is not sensitive to the value of @xmath119 as discussed in refs .
@xcite and @xcite . expanding eq .
( [ sigmar ] ) in powers of @xmath12 and integrating over @xmath65 , we obtain the complex - valued @xmath120 in the long wavelength limit . using eqs .
( [ eq : phi ] ) and ( [ eq : psi ] ) , we deduce the optical activity @xmath121 and the circular dicroism @xmath122 deriving eqs .
( [ ao ] ) and ( [ cd ] ) , we have assumed that @xmath123 and @xmath124 , and the equations are valid for temperatures such that @xmath125 . within these approximations ,
the optical rotation and circular dichroism are proportional to the chirality @xmath27
. the next order will be of order @xmath126 as imposed by the symmetry relation of eq .
( [ chiral_symmetry ] ) . the first term in eq .
( [ ao ] ) is identical to the expression of refs .
@xcite and @xcite for the optical activity in the pretransitional region .
this term is associated with the modes @xmath127 and gives the @xmath128 dependence observed in many experiments . as first noted by filev , the modes @xmath129 need to be considered in addition to the contribution from the modes @xmath130 , in a region very close to the transition in highly chiral liquid crystals @xcite .
this is the origin of the third term in eq .
( [ ao ] ) , whose contribution has the opposite sign with respect to the contribution from the @xmath127 modes .
this term produces a change of sign of the optical activity close to the transition as shown in fig .
( [ fig : aovst ] ) .
there has been some debate in the literature concerning the temperature dependance of this term , whether it should be in @xmath131 or @xmath132 .
the very short paper of ref .
@xcite did not provide enough explanations to answer this point , and that of ref .
@xcite also lacks explanations and contains some errors or misprints ( in particular in eq .
( 4 ) of ref .
@xcite ) . in ref .
@xcite , collings et al .
studied experimentally systems of high chirality and found good agreement with filev s model with a @xmath133 exponent . unfortunately due to a large number of fitting parameters ,
the data could also have been explained with an exponent @xmath134 @xcite .
we shall come back to this point in section [ sec : exact_wavel ] . in a study of the wavelength dependence of the optical activity in the long wavelength regime @xcite , collings et al .
confirmed that the modes @xmath129 contribute to higher order than the modes @xmath127 in agreement with eq .
( [ ao ] ) .
the presence of a second order correction in @xmath135 for the modes @xmath127 with a dependence @xmath136 , which would correspond to the second term in eq .
( [ ao ] ) has not been proved or disproved experimentally , although quite good fits to the data can be obtained with this term included as found in refs .
@xcite and @xcite .
we present in this section a derivation of the de vries formula , which is the well known solution of maxwell s equations for light propagating in a cholesteric liquid crystal along the pitch axis @xcite .
the reason of this digression into the optics of periodic chiral media will become clear in the next section , where we explore the relation between optical activity in isotropic chiral liquid crystals and the de vries formula for the cholesteric phase .
we shall take the @xmath137-axis to be the helical axis of the cholesteric phase , and ( @xmath138 , @xmath139 , @xmath140 ) to be a right handed frame . in the cholesteric phase , the order parameter , the anisotropy in the dielectric constant , is @xmath141 where @xmath142 , and @xmath143 . using eq .
( [ cholesteric_orderparam ] ) and the definitions ( [ def_t0]-[def_t2 ] ) , it is simple to show that @xmath144 - \frac{\epsilon_a}{\sqrt{6 } } t_{ij}^0 ( \ez).\ ] ] from this , the tensor @xmath145 defined in eq .
( [ def_tensor_b ] ) can be constructed . to calculate its fourier transform @xmath24 , it is convenient to use a spherical coordinate system with @xmath146 , @xmath147 which can be simplified using eq .
( [ cholesteric_eps ] ) @xmath148 + a_{ijkl } ( { \bf q } ) , \\ & = & \frac{\omega^4 \epsilon_a^2 \left ( 2\pi \right)^3}{4 } \left [ \delta^3 \left ( { \bf q}+\kappa \ez \right ) t_{ik}^{-2 } t_{jl}^{2 } + \delta^3 \left ( { \bf q}-\kappa \ez \right ) t_{ik}^{2 } t_{jl}^{-2 } \right ] + a_{ijkl } ( { \bf q}),\end{aligned}\ ] ] where @xmath149 , and @xmath150 is a linear combination of tensors of the form @xmath151 with @xmath71 and @xmath152 being @xmath153 or @xmath154 .
the bragg peaks of the cholesteric phase located at @xmath155 correspond to the first two terms of eq .
( [ cholesteric_b_simple ] ) .
note that in a periodic medium considered in this section , @xmath33 is a function of @xmath17 , whereas in an isotropic medium to be considered in the next section , @xmath33 is only a function of @xmath156 . in a periodic medium @xmath157 can be derived generally from @xmath158 using @xmath159 e_{ijl}k_l}{4ik^2}.\ ] ] using eqs .
( [ sigma ] ) , ( [ cholesteric_b_simple ] ) and ( [ delta_sigma ] ) , we find @xmath160 \left [ t_{ik}^{-2 } t_{jl}^{2 } - t_{ik}^{2 } t_{jl}^{-2 } + \tilde{a}_{ijkl } \right],\ ] ] where @xmath161 . in eq . ( [ sigma1_vries ] ) , the tensors @xmath162 can be eliminated using eqs .
( [ def_t0 ] ) , ( [ def_t2 ] ) and ( [ relation1 ] ) .
we now assumes that @xmath17 is along the @xmath137-axis . in this case , it is simple to show that the tensor @xmath163 does not contribute to @xmath33 . according to the analysis of section [ sec : theory ] , to linear order in @xmath22 the optical activity and the circular dichroism are proportional to the real and imaginary part of @xmath157 evaluated at @xmath164 . with these assumptions , eq .
( [ sigma1_vries ] ) gives @xmath165 the real part of eq .
( [ sigma1_vries3 ] ) leads to de vries formula @xcite @xmath166},\ ] ] where @xmath167 and @xmath168 , the pitch of the cholesteric phase , and the circular dichroism @xmath169 is zero within the same approximations .
note the following features @xcite : there is a dispersion anomaly at the bragg reflection @xmath170 , but both eq .
( [ vries_f ] ) and @xmath171 break down near the bragg reflection . indeed close to the bragg reflection , terms of higher order in @xmath22
contribute to the optical rotation and the circular dichroism , and for this reason eqs .
( [ vries_f ] ) and @xmath171 are only valid in the domain @xmath172 @xcite .
the sign of the optical rotation is such that a right - handed helix ( @xmath173 ) produces a positive optical rotation ( the material is laevogyric ) when @xmath174 and a negative one ( the material is dextrogyric ) when @xmath175 . in the long wavelength limit @xmath174 ,
the de vries optical activity @xmath176 is of order @xmath177 and depends on the light propagation direction as opposed to the isotropic - bp iii phases where it is of order @xmath178 and is independent of the light propagation direction .
in this section , we compute the exact wavelength dependance of the optical activity , the circular dichroism and the average index in an isotropic chiral medium , without relying on the long wavelength approximation and at any order in the chirality @xmath27 .
the method has been pioneered by dolganov @xcite in his study of bp i and bp ii , but this reference does not present all the details of the calculation . instead of using the expansion eq .
( [ green_exp ] ) , which is only valid in the long wavelength approximation , we calculate exactly @xmath179 from eqs .
( [ sigma ] ) and ( [ delta_sigma ] ) . separating the contribution from the modes @xmath93 and @xmath180 ,
we obtain ( see appendix [ appendixa ] for details ) @xmath181 and @xmath182 with @xmath183 . from these equations , the contribution of the modes @xmath127 and @xmath184 of eq .
( [ sigmar ] ) of section [ sec : long_wavelength ] is recovered in the limit @xmath185 . according to the analysis of section [ sec : theory ]
however , the correct procedure to obtain the optical activity and the circular dichroism to leading order in @xmath22 is from @xmath179 evaluated at @xmath186 . as will become clear later
, this does not give the same result in general when compared to section [ sec : long_wavelength ] where the limit @xmath187 is taken in eq .
( [ green_exp ] ) .
using @xmath186 , eqs .
( [ vries ] ) and ( [ novries ] ) can be simplified @xmath188 and @xmath189 note that the integrand of eq .
( [ novries2 ] ) vanishes for @xmath190 but is non - zero otherwise .
this means that the modes @xmath127 contribute only if the light is propagating off axis or if the pitch axis and the director are not perpendicular to each other as in smc@xmath191 for instance .
( [ vries2 ] ) is the analog for isotropic chiral media of the de vries formula of eq .
( [ sigma1_vries3 ] ) valid for periodic media .
the two expressions become identical with the replacements @xmath192 since the light is propagating along the pitch axis , and @xmath193 according to eq .
( [ cholesteric_b_simple ] ) . other periodic and chiral media like bp
i and bp ii can be treated as the cholesteric phase in section [ sec : de vries ] .
therefore we expect in bp i and bp ii a divergence of the @xmath184 contribution to the optical activity at the bragg condition , _
i.e. _ when the denominator of eq .
( [ vries2 ] ) is zero , and a change of sign of the optical activity when crossing this point .
there should be no divergence for the modes @xmath127 as can be inferred from eq .
( [ novries2 ] ) .
the recent reference of hunte and singh @xcite contains a derivation of the pretransitional optical activity in isotropic chiral media , which uses a method similar to the one presented in this article although the authors have only applied it to the long wavelength regime .
we believe that the authors have made an error in deriving eqs .
( 30 ) and ( 31 ) of ref .
@xcite which represent respectively the contribution of the modes @xmath127 and of @xmath184 to the optical activity .
the integrand of eq .
( 30 ) of ref .
@xcite should vanish when @xmath194 as is true for the integrand of eq .
( [ novries2 ] ) but it does not , and eq .
( 31 ) of ref .
@xcite should reproduce the de vries as is true for eq .
( [ vries2 ] ) but does not
. therefore we think that the results of the derivation of the optical activity in ref .
@xcite are incorrect although the method used is valid .
a straightforward evaluation of eq . ( [ novries2 ] ) gives @xmath33 for the modes @xmath127 and for an arbitrary frequency @xmath12 : @xmath195 the @xmath130 contribution of the optical activity has a trivial frequency dependance of @xmath196 .
the reason is the following : the modes @xmath127 unlike the modes @xmath129 are longitudinal , therefore these modes are associated with the longitudinal part of the green s function ( also called near - field part ) which is the second term in the r.h.s of eq .
( [ green_libre ] ) .
the only pole of the near - field green s function is at @xmath197 , which is the reason for the absence of a complex wavelength dependance for the modes @xmath127 and also the reason for which these modes give the leading contribution to the optical activity at small @xmath12 .
therefore the most interesting part of the wavelength dependance of the optical activity comes from the modes @xmath129 .
it is also the dominant contribution when @xmath54 is close to @xmath95 .
the integration in eq .
( [ vries2 ] ) can be done analytically for @xmath198 , and the final result ( see appendix [ appendixb ] for details of the derivation ) takes the following form @xmath199 k_b t,\ ] ] where @xmath200 and @xmath201 .
note that eq .
( [ oag ] ) agrees with the symmetry relation of eq .
( [ chiral_symmetry ] ) .
we have introduced the function @xmath202 this function tends to @xmath203 as @xmath204 , is equivalent to @xmath205 as @xmath206 , and diverges at @xmath207 as a result of the divergence of the initial expression of eq .
( [ vries2 ] ) .
( [ oag ] ) and ( [ function_f ] ) fully agree with filev s results in ref .
@xcite quoted for @xmath208 .
similar functions but not exactly identical functions are present in refs .
@xcite , @xcite , @xcite , and @xcite which could indicate misprints or errors regarding this particular point .
it is interesting to consider two particular cases of eq .
( [ oag ] ) , the case where @xmath209 , which defines the long wavelength approximation and the opposite case where @xmath210 .
in the first case , it is easy to deduce from eq .
( [ oag ] ) that @xmath211 .
therefore the optical activity in the long wavelength approximation for a temperature @xmath94 is @xmath212 note that the contribution of the modes @xmath184 goes as @xmath213 if @xmath214 , but should go as @xmath215 otherwise . as discussed before , eq .
( [ ao_lw ] ) is not identical with eq .
( [ ao ] ) because @xmath216 . since the modes @xmath127 and @xmath184 contribute with different signs in eq .
( [ ao_lw ] ) , there is a competition between these two modes , which leads to a maximum in the optical activity as a function of the temperature @xmath54 as shown in fig .
[ fig : aovst ] for a chirality @xmath91 . in the opposite case of @xmath210 , we find using eq .
( [ oag ] ) that @xmath217 .
we now consider the complete wavelength dependance of the optical activity for the modes @xmath129 . in figs .
[ fig : aovslambda]a and [ fig : aovslambda]b , the optical activity and the circular dichroism are shown as a function of the wavelength expressed in units of @xmath218 , as calculated with eqs .
( [ oag ] ) and ( [ im_sigma ] ) . in fig .
[ fig : aovslambda]a @xmath219 and in fig .
[ fig : aovslambda]b @xmath220 .
note that in all these figures , a positive value of @xmath27 has been chosen corresponding to a right - handed helix in the cholesteric phase , but that an opposite optical activity and circular dichroism would have been found with a left - handed helix .
these figures clearly show a maximum in the magnitude of the optical activity when the wavelength is equal to the pitch @xmath0 of the cholesteric phase , which is about @xmath221 for @xmath91 in the case of fig .
[ fig : aovslambda ] .
this maximum at this wavelength is the remains in the isotropic phase of the divergence present in the optical activity of bp i and bp ii and in the de vries formula in eq .
( [ vries_f ] ) .
as @xmath54 gets closer to @xmath95 , @xmath222 becomes peaked at @xmath223 , and the width of the maximum in the magnitude of the optical activity gets smaller as can be seen by comparing figs .
[ fig : aovslambda]a and [ fig : aovslambda]b .
these figures also show the circular dichroism which to our knowledge has never been discussed theoretically or measured directly in isotropic chiral media .
the absence of experiments is probably due to the difficulty separating the contribution from pretransitional fluctuations and the contribution from the absorption bands . according to fig .
[ fig : aovslambda ] , circular dichroism is best observed in the region where @xmath210 as it gets very small when @xmath209 .
the curve shown in the figure is the contribution from pretransitional fluctuations , which is valid away from absorption bands .
note that experimentally this circular dichroism translates in the multiple light scattering regime into a difference in the scattering mean free paths for circularly polarized waves .
this means that the transport properties of light in the multiple light scattering regime should be different for the different states of circular polarization .
such a difference has indeed been observed in static and dynamic light scattering measurements with circularly polarized light in the isotropic phase and in bp iii @xcite .
finally we have applied our method to the modification of the average dielectric constant due to the fluctuations . in section [ sec : theory ] , we have defined this quantity as @xmath224 , where @xmath31 is the isotropic part of @xmath22 .
the effect of the fluctuations on the symmetric part of the dielectric constant has to our knowledge never been calculated or measured in the pretransitional region .
we find that all the five modes contribute to this average dielectric constant . in order to illustrate a case where the corrections can be large , we focus here on the contribution of the modes @xmath184 , which as noted before is dominant when @xmath54 is sufficiently close to @xmath95 .
using eqs .
( [ sigma ] ) and ( [ def_sigma12 ] ) , we find @xmath225 with @xmath226 . after integrating over @xmath227 and @xmath65 ,
one obtains @xmath228 k_b t,\ ] ] with @xmath229,\ ] ] and @xmath230 as imposed by the symmetry relation of eq .
( [ chiral_symmetry ] ) , @xmath231 of eq .
( [ reindice ] ) is an even function of @xmath27 . in fig .
[ fig : indice ] , the real and imaginary part of the average dielectric constant @xmath231 is shown as a function of the wavelength , as calculated using eqs .
( [ reindice ] ) and ( [ imindice ] ) for the modes @xmath184 .
similarly to the case of the optical activity , there is a maximum in @xmath231 as a function of the wavelength when the wavelength is of the order of the pitch of the cholesteric .
in this section , we compare the prediction of our model with the experimental spectra of the optical activity ( ord spectra ) in the bp iii phase taken from ref . @xcite .
this reference is the only one known to us that reports measurements of the optical activity in the isotropic phase and in bp iii which are not limited to the long wavelength regime .
the ord spectra are measured in the bp iii phase for pure cholesteryl myristate ( cm ) , pure cholesteryl nonanoate ( cn ) and mixtures of cn with small quantities of cholesteryl chloride ( cc ) , a compound of opposite chirality .
the sign of the optical rotation agrees with section [ sec : de vries ] : all the samples have a dominant left - handed character ( @xmath232 and @xmath233 ) , and the optical activity is indeed found to be negative when @xmath234 .
the optical activity presents a broad maximum at @xmath235 , which grows smaller and broader and shifts to higher wavelength as @xmath236 is increased .
the curves are fitted with a theoretical expression , which is the sum of eq .
( [ mode1 ] ) ( for the modes @xmath127 ) and eq .
( [ oag ] ) ( for the modes @xmath184 ) , and as can be seen in fig .
[ fig : oa_fit ] , the fits are excellent . the spatial dispersion of the index , which affects the wavelength of the light in the medium @xmath1 , is incorporated in this fit using the measurements on cn reported in ref @xcite .
there are four free parameters in this fit @xmath237,@xmath238,@xmath239 and @xmath240 .
these parameters are @xmath241 , @xmath242 , @xmath243 , and @xmath244 so that @xmath245 where @xmath1 is expressed in nm so that @xmath238 and @xmath239 have units of nm@xmath246 .
the values of these parameters for the different samples are shown in table [ tab : coeff ] .
the parameter @xmath238 obtained from the fits should be identical to the absolute value of the inverse pitch of the cholesteric phase .
this is indeed the case as shown in fig .
[ fig : chirality_fit ] , where the points correspond to the measurements of fig .
[ fig : oa_fit ] and the solid line is a linear fit which gives a slope of @xmath247 and an intercept of @xmath248 .
in this article , we have discussed spatial dispersion effects ( optical activity , circular dichroism and average index ) for light propagation in the isotropic - bp iii phase , by introducing an effective index which describes the fluctuations .
we have obtained the wavelength dependance of the spatial dispersion effects without relying on the long wavelength approximation , on which previous studies on this problem have been based .
this new approach allows us to discuss the optical properties in the resonant region when the wavelength of the light is of the order of the pitch of the cholesteric . in the vicinity of this point
, there is a divergence of the magnitude of the optical activity in the cholesteric , bp i , and bp ii phases and a broad maximum in the case of the bp iii and isotropic phases .
these features are confirmed by measurements of spectra of the optical activity in the bp iii phase for different values of the chirality .
we have also provided predictions for the wavelength dependance of the circular dichroism and for the symmetric part of the effective index .
we hope that this work will motivate experimentalists to study the optical properties of periodic and non - periodic chiral media .
+ we acknowledge many stimulating discussions with b. pansu , p. ziherl and p. galatola .
this work was supported in part by the mrsec program under grant nsf dmr00 - 79909 .
d. lacoste received support by a grant from the french ministry of foreign affairs .
using eqs . ( [ sigma ] ) and ( [ delta_sigma ] ) , we find that @xmath249 g_{kl}^0({\bf k+q},\omega).\ ] ] for the modes @xmath184 , @xmath250 according to eqs .
( [ tensor_b ] ) and ( [ def_t2 ] ) . using the relation @xmath251 together with eq .
( [ m ] ) , eq .
( [ sigma_n ] ) takes a simpler form @xmath252 with the definition eq .
( [ green_libre ] ) , this expression reduces to eq .
( [ vries ] ) .
similarly , @xmath253 for the modes @xmath127 according to eqs .
( [ tensor_b ] ) and ( [ def_t1 ] ) . using the relation @xmath254 eq .
( [ sigma_n ] ) takes the form @xmath255 which after simple algebra reduces to eq .
( [ novries ] ) .
using the change of variables @xmath256 , eq .
( [ vries2 ] ) can be written as @xmath257 where @xmath258 denotes the integral @xmath259 which can be decomposed in @xmath260 and @xmath261 if @xmath262 and @xmath88 otherwise .
the integration over @xmath65 in eq .
( [ res ] ) can be carried out in the complex plane with the function @xmath263 defined in eq .
( [ function_f ] ) which is identical with @xmath264 except for the absence of absolute values and for this reason @xmath263 is analytical .
@xmath222 has only two poles in the upper half plane @xmath265 and @xmath266 , which correspond to @xmath267 and @xmath268 in eq .
( [ oag ] ) . | this paper concerns optical properties of the isotropic phase above the isotropic - cholesteric transition and of the blue phase bp iii .
we introduce an effective index , which describes spatial dispersion effects such as optical rotation , circular dichroism , and the modification of the average index due to the fluctuations .
we derive the wavelength dependance of these spatial dispersion effects quite generally without relying on an expansion in powers of the chirality and without assuming that the pitch of the cholesteric @xmath0 is much shorter than the wavelength of the light @xmath1 , an approximation which has been made in previous studies of this problem .
the theoretical predictions are supported by comparing them with experimental spectra of the optical activity in the bp iii phase . |
cross sections of the reactions @xmath13 and @xmath2 , are determined by strong and electromagnetic interactions of the light quarks @xmath14 and can not be evaluated at present from first principles . but a rather good description of these cross sections is provided by the vector meson dominance model ( vdm ) with phenomenologically adjusted coupling constants ( @xmath15 , @xmath16 , @xmath17 , @xmath18 ) . in the vdm ,
the cross sections are determined by the amplitudes of vector meson v transitions ( @xmath19 , @xmath20 , @xmath21 ) into the final state : @xmath22 , @xmath23 . in the energy range from 980 to 1070 mev ,
the main contributions to the @xmath13 and @xmath2 cross sections come from the @xmath10-meson . therefore the measurement of these cross sections allows us to determine the @xmath10-resonance main parameters and study its interference with other vector mesons .
such studies provide important information about light meson physics .
earlier these processes were studied in several experiments : @xmath24 @xcite , @xmath1 @xcite and @xmath0 @xcite .
the snd ( spherical neutral detector ) experiment continues these efforts at the vepp-2 m collider @xcite . here
we present the results of this investigation .
the snd has been operating at the vepp-2 m collider in the energy range from 360 to 1400 mev since 1995 ( the snd detailed description can be found in ref.@xcite ) .
the detector contains several subsystems .
the tracking system includes two cylindrical drift chambers .
the three - layer spherical electromagnetic calorimeter is based on nai(tl ) crystals .
the muon / veto system consists of plastic scintillation counters and two layers of streamer tubes .
the calorimeter energy and angular resolution depend on the photon energy as @xmath25{e(\mbox{gev})}}$ ] and @xmath26{e(\mathrm{gev } ) } } \oplus 0.63^\circ$ ] .
the tracking system angular resolution is about @xmath27 and @xmath28 for azimuthal and polar angle respectively .
the energy loss resolution @xmath29 in the drift chamber is about 30% good enough to provide charged kaon identification in the region of @xmath10-meson production .
this work is based on the data sample collected in the snd experiments during 1998 .
the data were collected at 32 energy points in the energy range from 984 to 1060 mev ( table [ tab1 ] ) .
the total integrated luminosity accumulated in these experiments is 8.5 pb@xmath30 .
the beam energy was calculated from the magnetic field value in the bending magnets and revolution frequency of the collider . in the vicinity of the @xmath10-meson peak we also used the beam energy calibrated by the cryogenic magnetic detector ( cmd-2 ) , which operated at vepp-2 m at the same time .
this calibration @xcite is based on the measurements of the charged kaon momenta in the drift chamber of cmd-2 .
the energy calibration accuracy in each point is about 0.1 mev , the common shift of the energy scale is estimated to be 0.04 mev , and the energy spread of the beams is about 0.37 mev . for the luminosity measurements , @xmath31 and
@xmath32 processes were used .
the corresponding luminosity values were respectively used for normalizing the events containing charged particles and those with neutral particles only .
they are different by about 10% ( tab.[tab1 ] ) due to the dead time of the trigger , which selected events containing neutral particles only .
the systematic error of the integrated luminosity determination is estimated to be 2% . since luminosity measurements by @xmath31 and @xmath32
reveal a systematic spread of about 1.3% , this was added to the statistical error of the luminosity determination in each energy point .
the statistical accuracy was better than 1% .
the experimental conditions were rather stable during data taking : the snd systems counting rate , average currents and luminosity of the collider were stable at the level of about 20 % and no correlation between different energy points was observed .
the procedure of the experimental data selection consists of several stages . during the experimental runs ,
the first - level trigger was used @xcite .
the first - level trigger selects events of different types : events with photons only and events with charged particles .
the thresholds applied on the energy deposition in the calorimeter were about 200 mev for events containing charged particles and 300 mev for events with neutral particles only . during the processing of experimental data ,
the event reconstruction is performed in the following sequence .
the first step is a search for separated clusters in the calorimeter .
track reconstruction in the drift chambers is then performed .
tracks are linked to the calorimeter clusters .
clusters with energy depositions of more than 20 mev , and not linked to tracks in the drift chambers are considered as photons . for charged tracks no requirements on the energy depositions in the calorimeter
are imposed . for further analysis ,
events with @xmath33 cm and @xmath34 cm for each charged particle were selected . here
@xmath35 is the coordinate of the charged particle production point along the beam axis ( the longitudinal size of the interaction region is @xmath36 cm ) ; @xmath37 is the distance between a charged particle track and the beam axis in the @xmath38 plane . for each process
individual selection criteria were then applied .
the detection efficiency was obtained from monte - carlo ( mc ) simulation .
simulated events were reconstructed and the actual selection criteria were applied . to take into account the apparatus resolution of the first - level trigger thresholds , for simulated and experimental events the trigger conditions were applied with the threshold values 10% higher than during the experimental runs . as a result of uncertainties in the simulation of the distributions over some selection parameters ,
the detection efficiency was multiplied by correction coefficients , which were obtained in the following way .
the experimental events were selected without any conditions on the parameter under study , using the cut parameters uncorrelated with the studied one .
the same selection was applied for simulated events .
then the cut was applied to the parameter and the correction coefficient was calculated : @xmath39 where @xmath40 and @xmath41 are the number of events in experiment and simulation selected without any cuts on the parameter under study ; @xmath42 and @xmath43 is the number of events in experiment and simulation , when the cut on the parameter was applied . as a rule ,
the error in the coefficient @xmath44 determination is connected with the uncertainty of background subtraction .
this systematic error was estimated by varying other selection criteria .
the overlap of the beam background with the events containing charged particles can result in track reconstruction failure and a decrease of detection efficiency . to take into account this effect , background events ( experimental events collected when detector was triggered with an external generator )
were overlapped with the simulated events .
it was found that the detection efficiency decreased by about 3% and therefore the correction coefficient @xmath45 was used for the processes containing charged particles . in the analysis described further simulated events are not overlapped with beam background events . in the studied energy region
the @xmath0 events have the following features : * the charged kaons are rather slow ( @xmath46 @xmath47 ) and therefore have large @xmath29 losses in the drift chamber . * events often contain several photons due to the kaon decays ( @xmath48 , @xmath49,@xmath50 , @xmath51 ) in the tracking system .
* tracks attributed to the kaon decays in the tracking system ( secondary particles ) are also present . to select the @xmath0 events , detection of two charged particles was required , with polar angles in the range @xmath52 , with acollinearity angle in the azimuthal plane @xmath53 and the energy losses in the drift chamber @xmath54 , where @xmath55 is an average @xmath56 loss of a minimum ionizing particle .
the events must also contain two or more photons and at least one secondary particle from kaon decay in the tracking system .
the distribution of the angle @xmath57 for charged particles in the selected events has the typical shape for the @xmath58 reaction ( fig.[kkc_tet ] ) .
after these cuts , the background processes @xmath59 , @xmath60 and also beam and cosmic backgrounds still contribute .
the data accumulated under the @xmath61 production threshold , at the energy point @xmath62{s } = 984 $ ] mev , were used to estimate the number of background events .
after selection criteria were applied , only @xmath63 events were found at this point .
the number of background events in each energy point was calculated as @xmath64 where @xmath65 is the integrated luminosity , and this number was subtracted from the number of selected events .
the detection efficiency was multiplied by correction coefficients @xmath66 , @xmath67 , @xmath68 , @xmath69 . here
@xmath70 is a correction due to the inaccuracy in the simulation of the @xmath29 distributions .
this coefficient does not depend on the energy in the explored region .
the coefficient @xmath68 is a correction to the efficiency of detection of secondary particles and at least two photons .
its value varies in this energy range by @xmath71 10% and in the vicinity of the @xmath10-resonance peak is @xmath72 .
the correction @xmath73 is due to the inaccuracy in the simulation of the average energy deposition of @xmath0 events .
the total systematic error of the detection efficiency determination is 6.8% .
the number of selected @xmath0 events ( after background subtraction ) and the detection efficiency are shown in table [ tab2 ] .
the detection efficiency rises with energy because the probability of kaon absorption in the detector passive material decreases with energy . above 1035 mev
the efficiency goes down since the probability of kaon decay in the detector decreases . the @xmath75 and @xmath76 mesons decay lengths in the studied energy region are @xmath77 @xmath78 cm and @xmath79 @xmath80 m respectively .
the @xmath75 meson decays inside the collider vacuum chamber or tracking system , and the @xmath76 reaches the calorimeter , where it can either produce signals ( `` photons '' ) due to nuclear interactions or decay in flight , or it can pass through the detector without interacting .
the analysis of the @xmath1 reaction is based on the detection of @xmath81 or @xmath82 decays . in the first case ( charged mode )
the events which contained two charged particles and at least one `` photon '' were used .
the polar angles of the charged particles and their acollinearity angle were bounded by the criteria : @xmath83 , @xmath84 and @xmath85 .
the cut @xmath86 rejects the background from @xmath0 events . to suppress the same background ,
the following cuts were also applied : @xmath87 for each charged particle and @xmath88 at least for one .
the @xmath75 meson decay length is about 0.5 cm , so to suppress the background from @xmath24 , @xmath89 , @xmath90 , @xmath91 events , the cut @xmath92 cm @xmath93 was used .
the beam and cosmic backgrounds were suppressed by the criteria : @xmath94 and @xmath95 mev , where @xmath96 and @xmath97 are the total energy deposition and the energy deposition in the third calorimeter layer of any charged particle .
pions in the process under study have energies of about @xmath98 @xmath99 mev and actually give a very low energy deposition in the third layer .
the distribution of the angle @xmath100 between charged particles in the selected events is shown in figure [ kslan12 ] .
it has a characteristic peak at the minimum angle between pions , close to @xmath101 . to estimate
the number of background events , the data accumulated at the energy point @xmath62{s } = 984 $ ] mev , i.e. under the reaction threshold , were used .
after the cuts , @xmath102 events were left .
their production points are uniformly distributed along the beam direction , so these events can be attributed to the beam and cosmic backgrounds .
the number of background events in other energy points was calculated according to the formula ( [ nfon ] ) and then subtracted from the number of selected events .
the detection efficiency was multiplied by correction coefficients @xmath66 , @xmath67 and @xmath103 . here
@xmath104 is a correction due to an inaccuracy in the simulation of the @xmath29 distribution .
the coefficient @xmath103 is a correction due to an inaccuracy in the simulation of the @xmath37 distribution ( fig.[kslr ] ) .
its value varies in this energy region by @xmath71 10% and in the vicinity of the @xmath10-resonance peak is @xmath105 .
the total systematic error of the detection efficiency determination is 3.7% .
the numbers of selected @xmath1 events ( after background subtraction ) and the detection efficiencies are shown in table [ tab21 ] .
the detection efficiency depends on the energy due to the dependence of @xmath106 and @xmath107 distributions on the @xmath75 energy and the dependence of @xmath76 nuclear interactions and decay length on its energy . for analysis of the @xmath1 decay in the neutral mode ( @xmath82 ) , we used events where at least four photons were detected and no charged particles were present . to reject beam and cosmic backgrounds ,
constraints were imposed on the total energy deposition ( @xmath108{s}$ ] ) and total momentum of events ( @xmath109{s}$ ] ) . to suppress cosmic background even further ,
the events where most of the hit crystals could be fitted by a single track , were rejected . for events satisfying selection criteria described above
, the kinematic fit was performed under the following assumptions : the event contains two photon pairs originating from the @xmath110 decay and invariant mass of these four photons is equal to the @xmath75-meson mass .
the value of the likelihood function @xmath111 is calculated . in events with more than four photons ,
extra photons are considered as spurious ones and rejected .
to do this , all possible subsets of four photons were tested and one , corresponding to the maximum likelihood , was selected .
figures [ ksl_mpi ] and [ ksl_mks ] represent invariant masses of photon pairs and of all four photons found in the reconstruction . to suppress the beam background , polar angles of the photons , reconstructed as @xmath112-quanta from the @xmath113 decay ,
were limited to @xmath114 .
the background from the multi - photon events ( pure qed events , @xmath115 and @xmath116 ) was rejected by the following cuts : @xmath117 and @xmath118 mev . here
@xmath119{s/4-m_{k_s}^2}$ ] and @xmath120 is the measured kaon momentum .
the @xmath111 and @xmath120 distributions are shown in figures [ ksl_chi2 ] , [ ksl_mo1 ] and [ ksl_mo2 ] .
after these cuts were applied , only three events were found at the energy point @xmath62{s } = 984 $ ] mev .
the number of background events for each energy point was calculated according to eq.([nfon ] ) and then subtracted from the number of selected events .
the number of background events from the @xmath121 decay was estimated using simulation and even in the vicinity of the @xmath10-resonance peak , their number is less than 0.2 % of the selected @xmath1 events .
the detection efficiency was multiplied by correction coefficients : @xmath122 and @xmath123 .
the @xmath124 is a correction due to an inaccuracy in the simulation of the @xmath111 distribution .
the experimental @xmath1 events can contain false photons , caused by the beam background overlaps .
the energy of these photons is similar to the energy of the @xmath112-quanta from the @xmath75 decay , and they are localized at the polar angles close to the beam direction . during reconstruction , some false photons can be mistaken as originating from the @xmath75 decay .
in other words , the reconstruction program can reject with some probability a valid photon from the @xmath75 decay and substitute it by the false one .
an event reconstructed in such a way will not pass our cut on the photon polar angles ( described above ) because the beam related false photons have small polar angles .
therefore the beam background overlaps lead to the decrease of the detection efficiency
an effect not taken into account in the mc simulation .
hence the necessity of the correction coefficient @xmath125 arises .
the detection efficiency does not depend on the energy and equals 0.068 after all corrections were implemented .
the total systematic error of the detection efficiency determination is 3.5% . the selected event numbers ( after background subtraction )
are shown in table [ tab22 ] .
for analysis of the @xmath2 process , events containing two charged particles and two or more photons were used .
extra photons can appear because of the beam background overlap , splitting of the electromagnetic showers , and nuclear interactions of the charged pions in the calorimeter . under these selection conditions ,
the background sources are @xmath0 , @xmath126 , @xmath89 , @xmath90 , @xmath91 processes and the beam background .
the polar angles of the charged particles were bounded by the criterion : @xmath52 . to suppress the beam background ,
the following cuts on the spatial angle between two charged particles and energy deposition of the neutral particles were applied : @xmath127 , @xmath128{s}$ ] . to suppress the background from the @xmath0 process
, the following cuts were imposed : @xmath129 for each charged particle , @xmath88 at least for one and @xmath53 . to reject the @xmath130 events the energy deposition of the charged particles
was bounded by the criterion : @xmath131{s}$ ] . for events left after these cuts ,
a kinematic fit was performed under the following constraints : the charged particles are considered to be pions , the system has zero momentum , the total energy is @xmath62{s}$ ] , and the photons originate from the @xmath110 decays . the value of the likelihood function @xmath132 is calculated . in events with more than two photons , extra photons
are considered as spurious ones and rejected . to do this , all possible subsets of two photons
were tested and the one , corresponding to the maximum likelihood was selected .
two - photon invariant mass and @xmath132 distributions are shown in figures [ mgg ] and [ chi ] .
after the kinematic fit , the following additional cuts were applied : @xmath133 , @xmath134 , and @xmath135 for energy points with @xmath62{s}<1028 $ ] mev , and @xmath136 for energy points above 1028 mev . here
@xmath137 is polar angle of any photon selected by the reconstruction program as originated from the @xmath138-decay .
@xmath139 is the number of detected photons .
a harder cut on @xmath132 at the five last energy points is necessary to suppress @xmath140 events , where @xmath141 is a photon emitted by initial particles . as experimental data were accumulated in a rather narrow energy region , the detection efficiency for the events without @xmath112-quantum radiation by initial particles does not depend on energy in the entire region .
the detection efficiency dependence on the radiated photon energy , obtained from simulation , is shown in figure [ pi3rad ] .
the detection efficiency was multiplied by correction coefficients : @xmath67 , @xmath104 , @xmath142 and @xmath143 .
the last correction factor is due to inaccuracies in the simulation of extra photons .
the detection efficiency ( without @xmath112-quantum radiation ) equals 0.183 ( under condition the @xmath144 ) and 0.086 ( under condition the @xmath145 ) after all corrections were implemented .
the total systematic error of the detection efficiency determination is 4.5% .
the selected event numbers are shown in table [ tab22 ] .
the number of background events ( @xmath146 , @xmath147 , @xmath91 and @xmath148 ) was estimated using simulation in the following way : @xmath149 where @xmath150 is a background process number , @xmath151 is the cross section of the process taking into account the radiative corrections , @xmath65 is the integrated luminosity , @xmath152 is the detection efficiency and @xmath153 is a correction to the efficiency . for @xmath154 reactions
the @xmath151 were taken according to the snd measurements presented in this article and for the process @xmath146 other snd measurements @xcite were used .
the @xmath1 is the main background in the vicinity of the @xmath10-resonance peak ( @xmath62{s } = 1015.75 $ ]
@xmath155 mev ) and its contribution is @xmath71 90% of all background . outside this region
the background is strongly determined by the @xmath146 and @xmath91 events .
the background from the @xmath0 events is negligible , and even at the @xmath10-meson peak it is only @xmath156 of all background events .
the accuracy of background event numbers determination is less than 15% at all energy points .
the obtained number @xmath157 was subtracted from the selected events .
the numbers of the @xmath24 and background events are shown in table [ tab3 ] .
in the vdm framework the cross sections of the @xmath0 , @xmath7 and @xmath9 process are written as follows @xcite : @xmath158 @xmath159 @xmath160 @xmath161 here @xmath162 denotes the vector mesons @xmath163 ; @xmath164 , @xmath165 are kaon momenta , @xmath166 is a phase space factor including @xmath167 intermediate state @xcite , @xmath168 , @xmath169 and @xmath170 are amplitudes corresponding to contributions additional to the conventional vdm ( for example @xmath21 , @xmath20 and @xmath10 primes ) , and @xmath171 and @xmath172 are the relative interference phases .
the @xmath172 and @xmath171 phases are equal to @xmath173 in the framework of the naive quark model .
the partial decay width dependences on energy are written as : @xmath174 where @xmath175 denotes the pseudoscalar meson @xmath176 or @xmath177 .
the coupling constants are determined through the decay branching ratios in the following way : @xmath178^{1/2 } , \mbox{~~ } |g_{vp\gamma}| = \biggl[{{3\gamma_vb(v \to p\gamma ) } \over { q^3_{p \gamma}(m_v ) } } \biggr]^{1/2 } , \nonumber \\ |g_{v\rho\pi}| = \biggl[{{4\pi\gamma_vb(v \to \rho\pi ) } \over { w(m_v ) } } \biggr]^{1/2 } , \mbox{~~ } |g_{vpp}| = \biggl[{{6\pi m_v^2\gamma_vb(v \to pp ) } \over { q^3_{pp}(m_v ) } } \biggr]^{1/2 } , \nonumber \\
|g_{\omega ( \rho ) k^+k^-}| = { { 1 } \over { \sqrt{2 } } } |g_{\phi k^+k^-}| , \mbox{~~ } |g_{\omega ( \rho ) k_sk_l}| = { { 1 } \over { \sqrt{2 } } } |g_{\phi k_sk_l}| \end{aligned}\ ] ] here @xmath179 . to restrict the growth of the partial width @xmath180 with energy , the form factor is usually included @xcite in the @xmath181 coupling constant : @xmath182 , where @xmath183 ^ 2\end{aligned}\ ] ] we did not use this form factor for the experimental data approximation in this work :
the parameter @xmath184 was set to zero .
the coulomb interaction of the charged kaons in the final state was taken into account by multiplying the coupling constant @xmath185 by the appropriate form factor : @xmath186 , where @xcite : @xmath187 in the vicinity of the @xmath10-meson peak , the value @xmath188 is nearly constant and the influence of the @xmath189 form factor on the resonance width and mass is negligible .
for the experimental data approximation , theoretical cross sections were reduced to the following expressions : @xmath190 @xmath191 @xmath192{w(m_\phi ) \over w(m_\omega ) } { { \gamma_\omega m_\omega^2 \sqrt{\sigma(\omega\to 3\pi)m_\omega } } \over { d_{\omega}(s ) } } \sqrt[]{c_{\omega\rho\pi}(s ) } + a^0_{3\pi } \biggr|^2 , \end{aligned}\ ] ] where @xmath193 and @xmath194 , @xmath195 , @xmath196 are complex constants , corresponding to the contributions of the higher radial excitations of the @xmath21 , @xmath20 and @xmath10 mesons in the cross section @xcite .
the experimental cross section @xmath197 can be calculated as follows : @xmath198 where @xmath199 is the number of selected events , @xmath200 is the integrated luminosity , @xmath201 is the detection efficiency , and @xmath202 is the radiative correction calculated according to ref.@xcite .
the cross sections were fitted with the theoretical functions and the values of parameters and their errors were obtained using the procedure described in ref.@xcite which takes into account the beam energy spread and the error in the beam energy determination . for the approximation of the @xmath203 cross section , to take into account the detection efficiency dependence @xmath204 on the emitted @xmath112-quantum energy ( fig.[pi3rad ] ) , in the formula ( [ aprox ] ) the factor
@xmath205 was replaced with the following expression : @xmath206 where @xmath207 is the emitted photon energy and @xmath208 is the electron `` radiator '' function @xcite .
the fitting of the @xmath209 cross section was performed with the @xmath210 , @xmath211 , @xmath212 , @xmath196 and in some cases with the phase @xmath171 as free parameters .
approximations were performed under the following assumptions on the @xmath171 value : * @xmath213 , * @xmath214 , * @xmath171 is a free parameter . here
the expression for @xmath215 was taken from ref.@xcite and it is equal to @xmath216 , when @xmath217 .
the results of the fit are shown in table [ tab5 ] .
the first three variants correspond to the case when both the real and imaginary parts of the @xmath196 amplitude were free parameters .
the value of the @xmath218 differs from zero by more than three standard deviations .
the value of @xmath219 is consistent with zero
. then the imaginary part was fixed at @xmath220 and only the @xmath218 was taken as a free parameter ( four to six variants in table [ tab5 ] ) .
the values of @xmath210 and @xmath211 do not depend on the theoretical model .
the model dependence of the parameter @xmath221 is about @xmath222 and the difference of its values for variants 4 and 5 is much greater than the errors . the energy dependence of the @xmath24 cross section for the variant 5 is shown in figure [ pi3sig1 ] . at the right side from the resonance peak ,
the cross section has a minimum due to the interference of the @xmath10meson amplitude with the nonresonant part of the cross section .
the cross section value does not actually depend on the model parameters in the energy range @xmath62{s}= 984 $ ]
@xmath223 mev , but in the region of the interference minimum at @xmath62{s}= 1040 $ ] mev , it can differ by a factor of 2 for different models ( fig.[pi3sig2 ] ) .
this uncertainty is connected with the difference in the values of @xmath202 correction in eq.([aprox ] ) , calculated for different models .
the @xmath0 and @xmath7 cross sections were fitted with @xmath224 , @xmath225 , @xmath226 , @xmath227 as free parameters .
the parameters @xmath194 , @xmath195 and @xmath172 were either fixed ( @xmath228 and @xmath229 ) , or were free parameters .
in particular , various fits were performed under the following assumptions : 1 .
@xmath228 , @xmath230 ; 2 .
@xmath172 was free parameter , @xmath230 ; 3 .
@xmath172 and @xmath231 , @xmath232 were free parameters ; 4 .
@xmath228 , @xmath231 , @xmath232 , @xmath233 , @xmath234 were free parameters ; 5 .
@xmath228 , @xmath231 and @xmath232 were free parameters .
the results of the fit are shown in tables [ tab41 ] and [ tab42 ] . for the @xmath0 process
, the phase @xmath172 agrees with the theoretically predicted value @xmath173 .
the imaginary part of the @xmath194 amplitude is consistent with zero , and its real part is different from zero by about only one standard deviation ( variants 4 and 5 in table [ tab41 ] ) .
the values of @xmath210 and @xmath211 do not depend on the applied theoretical model .
the value of @xmath226 varies by 2% for different variants , but these biases are less than its error .
the cross section is almost independent of the applied model ( fig.[kkcsig ] ) . the data for the @xmath1 reaction obtained in charged and neutral modes were fitted simultaneously .
for this process the phase @xmath172 is different from the expected @xmath173 by more then five standard deviations and is close to @xmath235 ( variant 2 in table [ tab42 ] ) .
the imaginary part of the @xmath195 amplitude is consistent with zero , and its real part is different from zero by about two standard deviations ( variants 5 and 4 in table [ tab42 ] ) . the parameters @xmath210 and @xmath211 are slightly dependent on the theoretical assumptions .
the value of @xmath227 varies by 2% for different variants and is shown in fig.[kslsig ] .
the cross section is almost insensitive to the choice of the model . to obtain the @xmath10-meson production cross section and branching ratios of the main decay modes , the combined fitting of the experimental data , taking into account the @xmath236 cross section also ,
was carried out .
as this work does not contain the analysis of the @xmath121 reaction , it was natural to use the cross section @xmath237 from the previous snd measurements presented in ref.@xcite . in the ref .
@xcite the correction @xmath238 to the detection efficiency was not applied .
the value @xmath239 nb , presented there , was increased by 3% and the value @xmath240 nb was used in the fit .
the results of the fit , for the different model assumptions , are shown in table [ tab6 ] . in all
fits @xmath241 , @xmath242 , and @xmath243 were set . using these data ,
the following @xmath10-meson parameters were obtained : the @xmath10-meson production cross section @xmath244 the branching ratios of the main decay modes @xmath245 and the branching ratio of the decay into an @xmath246 pair : @xmath247 the final results are @xmath248 @xmath249 @xmath250 @xmath251 @xmath252 these values correspond to the the fit where @xmath253 , @xmath228 , @xmath220 , @xmath254 and @xmath255 ( variant 2 in table [ tab6 ] ) .
the third error is a model uncertainty .
the presented values are in good agreement with the world average and with other experimental data ( table [ tab10 ] ) . using the snd data ,
the following coupling constants were calculated : @xmath256 mev@xmath257 , @xmath258 , @xmath259 , @xmath260 gev@xmath30 , and @xmath261 gev@xmath30 .
the measured cross sections are given in table [ tab7 ] .
the systematic errors of the @xmath0 , @xmath2 cross sections and of the related parameters @xmath226 , @xmath262 are 7.1% and 5% respectively and include the systematic uncertainties in the detection efficiency and luminosity determinations . for
the @xmath1 cross section , the systematic error is 4.2% in the charged mode and 4.0% in the neutral mode .
the systematic error of the parameter @xmath227 is 3.2 % .
the @xmath10-meson mass and width measured in the three studied main decay modes are consistent with each other .
the measurements of the @xmath10-meson parameters in all channels have high accuracy .
let us mention that all these processes have different energy dependence for detection efficiency .
the dependence of the mass value on the applied theoretical model is less than the measurement accuracy . as a result of the combined approximation ,
we find : @xmath263 where the systematic error of 0.04 mev is attributed to the possible common shift of the collider energy scale .
this value is consistent with the world average : @xmath264 mev @xcite .
the value of the total @xmath10-meson width , obtained by using different models , can vary from 4.18 to 4.21 mev and was found to be @xmath265 ( the systematic error 0.02 mev is due to model dependence ) , and differs from the world average @xmath12 mev @xcite by 4.8 standard deviations and from the most precise measurement @xmath266 mev @xcite by 4.5 standard deviations . the @xmath2 process ,
as the approximation showed , can not be described by only @xmath10 and @xmath20 meson decays and it is necessary to include additional amplitude @xmath196 , which can be explained as a contribution of the higher resonances .
the presence of such an amplitude is also expected from the experimental data on the @xmath2 cross section measured by the snd detector @xcite in the energy region above the @xmath10-resonance , which can not be described with the conventional vdm .
the value of the amplitude @xmath267 mev@xmath268 ( table [ tab6 ] ) is only 2 times lower than the real part of @xmath20-meson amplitude in the vicinity of the @xmath10-meson peak @xmath269 mev@xmath268 and approximately equals its imaginary part @xmath270 mev@xmath268 .
the presence of the @xmath196 amplitude prevents a test of the deviation of the @xmath271 phase from @xmath173 . both @xmath213 and @xmath272 @xcite
are consistent with the experimental data . if the @xmath171 phase is considered to be a free parameter with no energy dependence , then its fitted value is @xmath273 for free real and imaginary parts of @xmath196 ( variant 5 in table [ tab5 ] ) , and @xmath274 , if @xmath220 is set ( variant 3 in table [ tab6 ] ) .
the parameter : @xmath275 has a rather large model error ( @xmath276 nb ) , due to the uncertainty in the choice of the phase @xmath271 . the quoted value of the @xmath262 corresponds to the choice @xmath277 and @xmath220 .
it agrees with other measurements , for example , @xmath278 from ref.@xcite and @xmath279 from ref.@xcite . in case of the @xmath0 and @xmath7 processes , the @xmath280 ,
@xmath281 agree with zero ( variant 4 in tables [ tab41 ] and [ tab42 ] ) and @xmath282 , @xmath283 deviates from zero by about two standard deviations ( variant 5 , table [ tab6 ] ) .
the phase @xmath172 extracted from the combined fit , or from the fitting of the @xmath0 process , is consistent with @xmath173 ( variants 2 and 4 in tables [ tab41 ] and [ tab6 ] ) , but the fitting of the @xmath1 data gives the phase value close to @xmath235 .
such a result can be attributed to the uncertainty of the nonresonant contributions to the cross section and was used to estimate the model error in the cross section determination .
the results @xmath284 @xmath285 agree with the world average @xcite .
isotopic symmetry predicts the equality of the @xmath185 and @xmath286 coupling constants .
the electromagnetic interactions of charged kaons in the final state lead to the relation @xmath287{z(m_\phi)}=g_{\phi k_sk_l}$ ] . using the particle data group ( pdg ) data , @xcite one can find : @xmath288{z(m_\phi ) } } = \sqrt[]{\biggl({{b(\phi \to k^+k^- )
q_{k_sk_l}^3(m_\phi ) } \over { b(\phi \to k_sk_l ) q_{k^+k^-}^3(m_\phi ) } } \biggr ) { 1 \over { z(m_\phi ) } } } = 0.95 \pm 0.01\end{aligned}\ ] ] this value differs considerably ( five standard deviations ) from the expected value of unity .
it is possible that the coulomb correction is less than expected due to finite dimensions of the final particles .
a similar problem in the @xmath289 decay was discussed in refs.@xcite .
quite recently the ratio of the constants @xmath185 and @xmath286 was discussed in ref.@xcite , where it was suggested to study this ratio in @xmath246 annihilation around the @xmath10 resonance peak .
the drawback of this proposal is a very low sensitivity of such studies to the nonresonant contribution to the total cross section .
the extracted coupling constant values depend on the nonresonant contribution to the total cross section from the @xmath21 and @xmath20 mesons and from higher radial excitations ( @xmath290 , @xmath291 , @xmath292 , ... ) .
so one of the possible approaches to the solution of this problem can be a precise study of these contributions to the total cross section , by means of the measurement of the @xmath13 cross section in a wide energy range .
unfortunately the snd data have rather high systematic errors , which preclude such studies , and they are accumulated in a rather narrow energy region . from snd data ( table [ tab6 ] ) one can obtain @xmath293{z(m_\phi ) } } = 0.92 \pm 0.03.\end{aligned}\ ] ] this value agrees with the quoted world average , but is also consistent with the expected value of 1 ( the difference is about 2.5 standard deviations ) .
the main parameters of the @xmath10-resonance were obtained in experiments with the snd detector at the vepp-2 m collider .
the integrated luminosity used was about @xmath294 .
the processes @xmath0 , @xmath1 and @xmath2 were studied .
the measured cross sections were approximated within the vdm , taking into account @xmath21 , @xmath20 and @xmath10-mesons .
possible contributions from higher resonances @xmath290 , @xmath291 , @xmath292 were included as constant amplitudes .
the @xmath10-meson parameters obtained by the approximation ( table [ tab10 ] ) mainly agree with pdg data and have accuracies comparable with the world averages .
the @xmath10-meson total width was found to be @xmath295 mev in contradiction to the pdg value @xmath12 mev . for a good approximation of the @xmath2 cross section the additional contribution
is strongly required , and its value is about 2 times lower than the real part of the @xmath20-meson contribution .
this contribution can be attributed to the higher resonances , for example to the resonance structure found by snd @xcite .
99 g.parrour et al . , phys . lett . * b 364 * , 362 ( 1976 ) .
a.d.bukin et al . , yad . fiz .
* 27 * , 976 ( 1978 ) [ sov .
j. of nucl .
* 27 * , 516 ( 1978 ) ] .
a.cordier et al . , phys . lett . * b 364 * , 13 ( 1980 ) s.i.dolinsky et al .
, phys . rep . *
202 * , 99 ( 1991 ) r.r.akhmetshin et al . ,
phys . lett .
* b 364 * , 199 ( 1995 ) b.r.akhmetshin et al . , phys
* b 434 * , 426 ( 1998 ) l.m.kurdadze et al .
, preprint inp 84 - 07 , novosibirsk , 1984 r.r.akhmetshin et al . , phys
* b 466 * , 385 ( 1999 ) p.m.ivanov et al .
, phys . lett .
* b 107 * , 297 ( 1981 ) v.e.balakin et al . ,
* b 34 * , 328 ( 1971 ) a.n.skrinsky , in proc . of workshop on physics and detectors for da@xmath296ne , frascati , italy , april 4 - 7 , 1995 , p.3 m.n.achasov et al .
instr . and meth . * a 449 * , 125 ( 2000 ) r.r.akhmetshin et al . ,
preprint , budker inp 99 - 11 , novosibirsk , 1999 .
m.n.achasov et al . , phys . lett . * b 499 * , 122 ( 1999 ) n.n.achasov et al . ,
* 54 * , 1097 ( 1991 ) [ sov
. j. nucl .
* 54 * , 664 ( 1991 ) ] ; n.n.achasov , et .
al . , int .
j. of mod .
phys . * a vol.7 no.14 * 3187 ( 1992 ) e.cremmer and m.gourdin , nucl.phys .
* b 9 * , 451 ( 1969 ) n.n.achasov , a.a.kozhevnikov , phys . rev . * d 55 * , 2663 ( 1997 ) ; n.n.achasov , a.a.kozhevnikov , yad . fiz .
* 60 * , 1131 ( 1997 ) [ phys . at
. nucl . * 60 * , 1011 ( 1997 ) ] n.n.achasov , a.a.kozhevnikov , phys . rev .
* d 57 * , 4334 ( 1998 ) + n.n.achasov , a.a.kozhevnikov , yad . fiz .
* 60 * , 2212 ( 1997 ) [ phys . at
. nucl . * 60 * , 2029 ( 1997 ) ] e.a.kuraev , v.s.fadin , yad . fiz . * 41 * , 733 ( 1985 ) [ sov .
. phys . * 41 * , 466 ( 1985 ) ] a.v.bozhenok , et .
al . , preprint , budker inp ( in russian ) 99 - 103 , 1999 particle data group , d.e.groom , et al .
, eur.phys.j .
* c 15 * , 1 ( 2000 ) d.atwood and w.j.marciano , phys.rev .
* d 41 * , 1736 ( 1990 ) g.p.lepage , phys.rev . *
d 42 * , 3251 ( 1990 ) n.n.achasov , a.a.kozhevnikov , phys . rev . *
d 61 * 054005 ( 2000 ) m.n.achasov et al .
, zh . eksp .
. fiz . * 117 * , 22 ( 2000 ) [ jetp * 90 * , 17 ( 2000 ) ] m.n.achasov et al .
, phys . lett .
* b 462 * , 365 ( 1999 ) a.bramon et al .
, phys . lett .
* b 486 * , 406 ( 2000 ) + .experimental main parameters : @xmath62{s}$ ] the energy in the center of mass system , @xmath297 beam energy spread , @xmath298 , @xmath299 integrated luminosities used for normalizing the events containing charged particles and with neutral particles only . [ cols="^,^,^,^,^,^,^,^ " , ] ccccc @xmath62{s}$]&@xmath300&@xmath301&@xmath301 & @xmath302 + & & ( charged mode ) & ( neutral mode ) & + + 984.21 & @xmath303&&@xmath304 + 1003.91 & @xmath305&@xmath306&@xmath307&@xmath308 + 1010.17 & @xmath309&@xmath310&@xmath311&@xmath312 + 1015.75 & @xmath313&@xmath314&@xmath315&@xmath316 + 1016.68 & @xmath317&@xmath318&@xmath319&@xmath320 + 1017.59 & @xmath321&@xmath322&@xmath323&@xmath324 + 1018.78 & @xmath325&@xmath326&@xmath327 & @xmath328 + 1019.79 & @xmath329&@xmath330&@xmath331 & @xmath332 + 1020.65 & @xmath333&@xmath334&@xmath335 & @xmath336 + 1021.68 & @xmath337&@xmath338&@xmath339&@xmath340 + 1023.27 & @xmath341&@xmath342&@xmath343&@xmath344 + 1028.23 & @xmath345&@xmath346&@xmath347 & @xmath348 + 1033.84 & @xmath349&@xmath350&@xmath351 & @xmath352 + 1039.59 & @xmath353&@xmath354&@xmath355 & @xmath356 + 1049.81 & @xmath357&@xmath358&@xmath359 & @xmath360 + 1059.66 & @xmath361&@xmath362&@xmath363 & @xmath364 + + 984.02 & @xmath303&&@xmath365 + 1003.71 & @xmath305&@xmath366&@xmath367&@xmath368 + 1010.34 & @xmath369&@xmath370&@xmath371&@xmath372 + 1015.43 & @xmath373&@xmath374&@xmath375&@xmath376 + 1016.78 & @xmath377&@xmath378&@xmath379 & @xmath380 + 1017.72 & @xmath381&@xmath382&@xmath383 & @xmath384 + 1018.62 & @xmath385&@xmath386&@xmath387 & @xmath388 + 1019.51 & @xmath389&@xmath390&@xmath391 & @xmath392 + 1020.43 & @xmath393&@xmath394&@xmath395 & @xmath396 + 1021.41 & @xmath397&@xmath398&@xmath399&@xmath400 + 1022.32 & @xmath401&@xmath402&@xmath403&@xmath404 + 1027.52 & @xmath405&@xmath406&@xmath407 & @xmath408 + 1033.58 & @xmath409&@xmath410&@xmath411 & @xmath412 + 1039.64 & @xmath413&@xmath414&@xmath415 & @xmath416 + 1049.60 & @xmath417&@xmath418&@xmath419 & @xmath420 + 1059.52 & @xmath421&@xmath422&@xmath423 & @xmath424 + | the cross sections of the processes @xmath0 , @xmath1 and @xmath2 were measured in the snd experiment at the vepp-2 m collider in the energy region near the @xmath3 meson .
these measurements were based on about @xmath4 @xmath5 , @xmath6 @xmath7 and @xmath8 @xmath9 selected events .
the measured cross sections have been analyzed in the framework of the vector meson dominance model and the main parameters of the @xmath10-resonance were obtained , such as its mass , width , the production cross section and branching ratios of the main decay modes .
the measured value of the @xmath10 meson total width , @xmath11 is lower than the present world average of @xmath12 mev .
contributions in addition to the conventional vector meson dominance model were found in the @xmath2 reaction cross section . |
for the last three decades , equations of the type @xmath0 have been studied intensively . here ,
@xmath1 , @xmath2 , @xmath3 , @xmath4 are transition rates on @xmath5 , and @xmath6 is a familiy of independent brownian motions .
the following special cases with very different interpretations and different behaviour are quite common in the literature .
[ ex1 ] the ( wright - fisher ) stepping stone model from mathematical genetics : @xmath7 .
[ ex2 ] the parabolic anderson model ( with brownian potential ) from mathematical physics : @xmath8 .
[ ex3 ] the super random walk from pure probability theory : @xmath9 .
[ ex4 ] the critical ( spatial ) ornstein - uhlenbeck process : @xmath10 . for the super random walk , @xmath11 is the branching rate which in this case is time - space independent . in @xcite , a two type model based on two super random walks with time - space dependent
branching was introduced .
the branching rate for one species is proportional to the value of the other species .
more precisely , the authors considered @xmath12 where now @xmath13 and @xmath14 are families of independent brownian motions
. solutions are called mutually catalytic branching processes . in the following years
, properties of this model were well studied ( see for instance @xcite , @xcite ) .
+ in this paper , we are interested in a variant of mutually catalytic branching , namely the symbiotic branching model introduced in @xcite for continuous space .
the same equations as for the mutually catalytic branching model are considered but additionally the driving noises are correlated in the following way : @xmath15 where @xmath16 $ ] is a correlation parameter . for @xmath17 and as well for general @xmath18 there are basically two approaches to formalize the equations . in @xcite under quite restrictive assumptions on the transition kernel @xmath4 , existence of solutions was obtained in the space of tempered sequences . since their assumptions in particular assume symmetry and exponential decay of @xmath4 , already existence of solutions in cases we are interested in is not clear .
this is why we stick to the setup of @xcite , which as well is more popular for interacting diffusions . for the transition kernel @xmath4
we assume @xmath19 two main examples of interest are the following : [ e1 ] the discrete laplacian is given by @xmath20 obviously , @xmath21 , @xmath22 , @xmath23 are fulfilled .
further , the one - dimensional riemann walk ( see for instance @xcite ) has transition rates @xmath24 with @xmath25 normalising the total rate to @xmath26 .
here @xmath21 , @xmath22 , @xmath23 are also fulfilled but in contrast to the discrete laplacian the assumptions of @xcite are not satisfied . to specify the state space we fix a positive , summable function @xmath27 on @xmath5 satisfying @xmath28 for some finite constant @xmath29 . in @xcite a possible choice ( see their equation ( 4.13 ) ) is given by @xmath30 where @xmath31 is positive and summable , @xmath32 denote the @xmath33-step transition probabilities , and @xmath34 .
this is needed to verify the generalized mytnik self - duality which was introduced for the continuous space analogue model in proposition 5 of @xcite .
the duality for the discrete space is similar . for the duality in @xmath35 in the special case @xmath17 see lemma 4.1 of @xcite
the state space is now defined by pairs of functions of the following liggett - spitzer space @xmath36 the choice of @xmath27 does not influence the results .
[ o ] for @xmath37 and @xmath16 $ ] there is a ( weak ) solution of the symbiotic branching model with almost surely continuous paths and state space @xmath35 . for @xmath38 solutions are unique in law .
the proof of proposition [ o ] is standard .
existence can be proven by finite dimensional approximations as in @xcite . for
@xmath39 uniqueness follows from the generalized mytnik self - duality as in @xcite . for @xmath40 uniqueness
is true since moments increase slowly enough , and for @xmath41 uniqueness of solutions is not known .
+ for this work we restrict ourselves to homogeneous initial conditions @xmath42 this is not necessary but simplifies the notation a lot .
+ the interesting feature of the symbiotic branching model is that it connects examples [ ex1]-[ex3 ] above .
being first established as a time - space inhomogeneous version of a pair of super random walks , the examples from above appear as special cases : @xmath17 obviously corresponds to the mutually catalytic branching model .
the case @xmath40 with the additional assumption @xmath43 corresponds to the stepping stone model as can be seen as follows : since in the perfectly negatively correlated case @xmath44 , the sum @xmath45 solves a discrete heat equation and with the further assumption @xmath43 stays constant for all time .
hence , for all @xmath2 , @xmath46 which shows that @xmath47 is a solution of the stepping stone model with initial condition @xmath48 and @xmath49 is a solution with initial condition @xmath50 .
finally , suppose @xmath51 is a solution of the parabolic anderson model , then , for @xmath41 , the pair @xmath52 is a solution of the symbiotic branching model with initial conditions @xmath53 .
+ the purpuse of this and the accompanying paper @xcite is to understand the nature of the symbiotic branching model better .
how does the model depend on the correlation @xmath18 ?
are properties of the extremal cases @xmath54 inherited by some regions of the parameters ?
since the longtime behaviour of super random walk , stepping stone model , mutually catalytic branching model , and parabolic anderson model are very different , one might guess that for varying @xmath18 different regimes correspond to the different models .
+ the focus of @xcite lies on the longtime behaviour in law ( unifying the classical results for the stepping stone model , mutually catalytic branching model , and parabolic anderson model ) if @xmath4 generates a recurrent markov process , on ( un)boundedness of higher moments @xmath55 $ ] as @xmath56 , and the wave speed for the continuous space analogue .
it was shown that in the recurrent case @xmath55 $ ] is bounded in @xmath57 if and only if @xmath58 for the transient case the behaviour in @xmath18 is open to a large extent .
+ in contrast to @xcite the present paper focuses on second moments .
note that ( [ e ] ) implies that in the recurrent case second moments are bounded if and only if @xmath59 .
this can also be seen and analysed in more detail using a moment - duality which will be explained in section [ sm ] . using this duality
, we show how to reduce second moments of symbiotic branching processes to moment generating functions and laplace transforms of local times of discrete space markov processes , i.e. @xmath60 , \end{aligned}\ ] ] where @xmath61 and @xmath62 denotes the local time @xmath63 in @xmath64
. for simple random walks the behaviour as @xmath65 was partially analysed in @xcite by analytic methods .
here , we present a simple new proof based on a renewal - type equation and tauberian theorems .
the simplicity of the proof has the advantage that no further assumptions on the markov process ( in particular no symmetry and no finite range assumptions ) are needed . for any @xmath3 and @xmath66 the technique yields precise growth rates including all constants . + as an application , intermittency and aging for symbiotic branching processes
are established .
+ the main results on intermittency and aging are collected in section [ sec : results ] . in section [ section2
] we first establish representations for second moments of symbiotic branching processes ( section [ sm ] ) , then prove the results for exponential moments of local times ( section [ sectionlocal ] ) , and , finally , proofs of the main results ( section [ sectionmain ] ) are given .
the first property we address is intermittency ( see for instance @xcite for a discussion of the ideas ) .
the @xmath67-th lyapunov exponent is defined by @xmath68\label{ly}\end{aligned}\ ] ] if the limit exists ( @xmath69 analogously ) . since in this work
we only deal with second moments we further define @xmath70.\end{aligned}\ ] ] in lemma [ duality ] we will see that @xmath71 and hence we abbreviate @xmath72 .
one says the system is intermittent ( or weakly intermittent as recently in @xcite ) if @xmath73 .
intermittency for the parabolic anderson model ( @xmath41 ) is a well - studied property ( see @xcite and @xcite ) .
the existing proofs heavily depend on the linear structure of the system since they employ explicit solutions given as feynman - kac type representations .
such explicit solutions are not known to exist for the symbiotic branching model .
hence , one might ask whether or not the results obtained for @xmath41 can be transferred to some larger regime of correlation values . indeed , this can be done .
+ let us first fix some notation
. in the following @xmath74 denotes a continuous time markov process with transition rates @xmath4 and @xmath75 $ ] . due to the moment - duality for symbiotic branching processes ( see lemma [ lem : moment - duality ] ) the notation of symmetrization is needed . for two independent markov processes
@xmath76 , @xmath77 with transition rates @xmath4 the symmetrization is defined as @xmath78 the transition rates of the symmetrization are given by @xmath79 its transition probabilities are denoted @xmath80 . note that in the symmetric case @xmath81 . the green function of @xmath74 is denoted @xmath82 and we abbreviate @xmath83 .
further , we set @xmath84 and abbreviate @xmath85 .
analogously , we use @xmath86 for the symmetrization @xmath87 . [
t3]let @xmath88 be a solution of the symbiotic branching model with homogeneous initial conditions
. then @xmath88 is intermittent if and only if @xmath89 in particular , there is no intermittency for non - positive @xmath18 .
the previous theorem suggests @xmath64 deviding symbiotic branching into two regimes in which the @xmath90 regime behaves like the parabolic anderson model with respect to intermittency .
+ although after understanding the two - types particle moment - dual ( lemma [ mom ] ) one sees that the problem can be treated as for @xmath41 , we present a new proof . in @xcite results of @xcite for higher moments of the parabolic anderson model were generalized to more general symmetric transitions than the discrete laplacian . here ,
in particular , complete results for the asymptotic behaviour ( exponential and subexponential ) of second moments of the parabolic anderson model with arbitrary transitions @xmath4 are proven . + in the course of the proofs we obtain the expression @xmath91 for the second lyapunov exponent , where @xmath92 is the inverse of the laplace transform of the return probabilities ( see proposition [ prop1 ] ) .
this expression , by tauberian theorems , gives us the explicit asymptotic behaviour for the lyapunov exponents as function of @xmath11 . in the following
, @xmath93 denotes strong asymptotic equivalence , i.e. @xmath94 means @xmath95 .
[ cor]let @xmath88 be a solution of the symbiotic branching model with homogeneous initial conditions .
then for @xmath96 , the map @xmath97 has the following properties : * @xmath72 is strictly convex , * @xmath98 for all @xmath11 , and @xmath99 for @xmath100 , * if @xmath101 as @xmath102 , @xmath103 , we have , as @xmath104 , @xmath105 * if @xmath101 , as @xmath102 , @xmath106 , we have , as @xmath107 , @xmath108 here , @xmath109 denotes the gamma function .
our approach has the further advantage that the growth rates in the critical and subcritical regimes follow directly : [ pro ] let @xmath88 be a solution of the symbiotic branching model with homogeneous initial conditions . if @xmath101 , as @xmath56 , then the following hold : * @xmath90 and @xmath106 * * if @xmath110 , then
@xmath111\sim \frac{1}{\varrho(1-\kappa\varrho \bar g_{\infty})},\quad\text { as } t\to \infty .
\end{aligned}\ ] ] * * if @xmath112 then , as @xmath56 , @xmath113 \sim \begin{cases } \frac{\bar g_{\infty}(\alpha-1)}{c \gamma(2-\alpha)\gamma(\alpha ) } \ , t^{\alpha-1 } & : 1<\alpha<2,\\ \frac{\bar g_{\infty}}{c}\ , \frac{t}{\log t } & : \alpha=2,\\ \frac{\bar g_{\infty}}{\bar h_{\infty}}\ , t & : \alpha>2 .
\end{cases } \end{aligned}\ ] ] * @xmath17 @xmath113 \sim \begin{cases } \frac{\kappa c}{1-\alpha}t^{1-\alpha}&:\alpha<1,\\ \kappa c \log(t)&:\alpha=1,\\ 1+\kappa \bar g_{\infty}&:\alpha>1 , \end{cases}\quad\text { as } t\to \infty .
\end{aligned}\ ] ] * @xmath59 @xmath114 \sim \begin{cases } 1-\frac{1}{\varrho } & : \alpha\leq 1,\\ 1-\frac{1}{\varrho}+\frac{1}{\varrho(1-\varrho\kappa \bar g_{\infty})}&:\alpha>1 , \end{cases}\quad\text { as } t\to \infty .
\end{aligned}\ ] ] for @xmath41 ( parabolic anderson model ) , the subexponential growth was partially analysed for finite range transitions in @xcite ( see their page 15 ) .
an example which was not included is for instance the riemann walk defined in example [ e1 ] . since in this case @xmath115 it serves as a convenient example for the above results which exhibits a precise recurrence / transience transition at @xmath116 .
further , the simple random walk on @xmath5 is contained with @xmath117 combining the intermittency result with the extension of the results of @xcite given in @xcite , we support the unstable behaviour of symbiotic branching for @xmath90 .
it is quite standard ( see for instance @xcite ) that spatial processes being intermittent have a very local property : for large times the mass of the process is concentrated on few sites ( `` islands '' ) . since for symbiotic branching , on each finite box solutions approach each constant configuration infinitely often , the islands do not stabilize . since the diffusion function has the form @xmath118 , we see that @xmath47 will not produce high peaks if @xmath49 is very small and vice versa .
hence , we suspect that @xmath119 are concentrated on the same islands .
understanding the pathwise behaviour better is an ambitious task for the future . recently in @xcite the concept of aging was discussed for certain classes of interacting diffusions .
they say that aging ( for linear test - functions ) appears if the limit @xmath120\end{aligned}\ ] ] depends on the choices of @xmath121 and @xmath57 .
aging does not appear if this is not the case .
the main results of @xcite were formulated with more general test - functions , though , restricted to finite range transitions .
differently , the present technique is restricted to linear test - functions but not to finite range transition .
our results suggest that neither finite range nor the linearity of test - functions is crucial .
symmetry of the transitions is assumed as in @xcite .
+ in @xcite it is shown that no aging appears in the parabolic anderson model ( in our model @xmath41 ) in any dimensions for the discrete laplacian .
further , for the super random walk ( in our model related to @xmath17 ) it was shown that aging appears exactly in dimensions @xmath122 .
this leads to the question if there are different phases for the symbiotic branching model .
we show that the model exhibits three different regimes ; an anderson model like behaviour for @xmath90 , a super random walk like behaviour for @xmath17 , and a stepping stone model like behaviour for @xmath59 .
the new case @xmath59 and corollary [ coro ] suggest that there are three regimes in which the most prominent examples fall .
[ t5 ] let @xmath88 be a solution of the symbiotic branching model with homogeneous initial conditions and @xmath123 . then ,
if @xmath124 , as @xmath56 , the following is true .
* if @xmath90 , then no aging occurs for any @xmath125 . * if @xmath17 , then * * no aging occurs , for any @xmath106 , * * @xmath126=(1-a)_+$ ] , for @xmath127 , * * @xmath128=\frac{(1+\frac{a}{2})^{1-\alpha}-(\frac{a}{2})^{1-\alpha}}{(1+a)^{\frac{1-\alpha}{2}}}$ ] , for any @xmath129 . *
if @xmath59 , then * * no aging occurs , for any @xmath106 , * * @xmath126=(1-a)_+$ ] , for @xmath127 , * * @xmath128= \frac{\int_0 ^ 1 ( 2 r+ a)^{-\alpha } ( 1-r)^{\alpha-1 } d r } { 2^{-\alpha } \gamma(\alpha)\gamma(1-\alpha)}$ ] , for any @xmath129 .
we emphasise that our proof of theorem [ t5 ] can be applied to more general interacting diffusions . in particular , the examples from the introduction are included . for finite range transitions , i ) , ii ) , iii ) of the following proposition
were proven in @xcite .
[ coro]consider solutions of ( [ 111 ] ) with homogeneous initial conditions .
then for * @xmath130 , aging appears as in theorem [ t5 ] i ) , * @xmath131 , aging appears as in theorem [ t5 ] ii ) , * @xmath132 , aging appears as in theorem [ t5 ] ii ) , * @xmath133 , aging appears as in theorem [ t5 ] iii ) . in the cases in which aging occurs ,
the upper and lower limits are bounded by the stated values up to constants depending on @xmath134 .
we start with a discussion on how second moments of symbiotic branching processes can be reduced to exponential moments of local times .
let us first recall the two - types particle moment - dual introduced in section 3.1 of @xcite .
this will be used to calculate second moments explicitly .
since the dual markov process is described formally in @xcite we only sketch the pathwise behaviour . to find a suitable description of the mixed moment
@xmath135 $ ] , @xmath136 particles are located at position @xmath137 .
each particle moves independently as a continuous time markov process on @xmath5 with transition rates given by @xmath4 . at time @xmath64
, @xmath33 particles have type @xmath26 , @xmath138 particles have type @xmath139 .
one particle of each pair changes its type when the time the two particles have spent at same sites with same type exceeds an independent exponential time with parameter @xmath11 .
let @xmath140 note that since there are only @xmath136 many particles the infinite product is actually a finite product and hence well - defined .
[ mom ] let @xmath88 be a solution of the symbiotic branching model with initial conditions @xmath37 and @xmath16 $ ] . then , for any @xmath141 , @xmath2 , @xmath142&=&{\mathbb{e}}\big[(u_0,v_0)^{l_t}e^{\kappa(l_t^=+\varrho l_t^{\neq})}\big ] .
\end{aligned}\ ] ] though the proof for the moment - duality was given in @xcite ( see the proof of their proposition 9 ) only for the discrete laplacian we skip a proof . for general transitions @xmath4 the proof follows along the same lines
. + note that for homogeneous initial conditions @xmath143 the first factor in the expectation of the right - hand side equals @xmath26 .
lemma [ mom ] in the special case @xmath41 , @xmath143 was already stated in @xcite , reproven in @xcite , and used to analyse the lyapunov exponents of the parabolic anderson model . + for @xmath144 , the difficulty of the dual process is based on the two stochastic effects : on the one hand , one has to deal with collision times of random walks which were analysed in @xcite .
additionally , particles have types either @xmath26 or @xmath139 which change dynamically .
+ second moments are special since particles of different types do not change types anymore .
hence , when starting with two particles of same type there is precisely one event of changing types .
this is used to obtain the following representation of second moments .
[ duality ] let @xmath88 be a solution of the symbiotic branching model with homogeneous initial conditions . then , for any @xmath141
, @xmath2 , @xmath145&={\mathbb{e}}[e^{\kappa\varrho l_t}],\\ { \mathbb{e}}[u(t , k)^2]&={\mathbb{e}}[v(t , k)^2]=\begin{cases } 1+\kappa{\mathbb{e}}[l_t]&:\varrho=0,\\ 1-\frac{1}{\varrho}+\frac{1}{\varrho}{\mathbb{e}}[e^{\kappa\varrho l_t}]&:\varrho\neq 0 , \end{cases } \end{aligned}\ ] ] where @xmath62 denotes the local time in @xmath64 of the symmetrization @xmath87 defined in ( [ sy ] ) started in @xmath64 .
the first expression for the mixed second moment follows directly from lemma [ mom ] : there are two particles which start with different types . since pairs of particles of different types
are never forced to change their types , they stay of different type for all time . hence , @xmath146 for all @xmath2 and the assertion follows .
+ for the second expression note that there is only one possible change of types . starting with two particles of same types
one of the types may change and the particles can not change their types again . using independence of the particles and the exponential time we can make this explicit .
let @xmath147 be an exponential variable with parameter @xmath11 , denote by @xmath148 the law of the two independent markov processes , and @xmath62 their collision local time .
integrating out the exponential variable leads to @xmath114 & = { \mathbb{e}}^{x\times y}[e^{\kappa(l_t^=+\varrho l_t^{\neq})}]\\ & = { \mathbb{e}}^{x\times y}[e^{\kappa(l_t^=+\varrho l_t^{\neq})}1_{y < l_t}]+{\mathbb{e}}^{x\times y}[e^{\kappa(l_t^=+\varrho l_t^{\neq})}1_{y\geq l_t}]\\ & = { \mathbb{e}}^{x}\big[\int_0^{l_t}\kappa e^{-\kappa x}e^{\kappa x+\kappa \varrho ( l_t - x)}\,dx\big]+{\mathbb{e}}^x\big[e^{\kappa l_t}{\mathbb{e}}^y[1_{y\geq l_t}]\big]\\ & = \begin{cases } \kappa{\mathbb{e}}[l_t]+{\mathbb{e}}[e^{\kappa l_t}e^{-\kappa l_t}]&:\varrho=0,\\ { \mathbb{e}}\big[e^{\kappa\varrho l_t}\int_0^{l_t}\kappa e^{-\kappa \varrho x}\,dx\big]+{\mathbb{e}}[e^{\kappa l_t}e^{-\kappa l_t}]&:\varrho\neq 0 .
\end{cases } \end{aligned}\ ] ] this proves the assertion .
now we prepare for the proof of the aging result .
[ appli ] let @xmath88 be a solution of the symbiotic branching model with homogeneous initial conditions and symmetric transitions @xmath4 .
then , for any @xmath141 , @xmath2 , @xmath149=1+\kappa\int_0^tp_{2r+s}(k , k){\mathbb{e}}[e^{\kappa \varrho l_{t - r}}]\,dr \end{aligned}\ ] ] and similarly for @xmath49 . the proof is only given for @xmath47 since due to symmetry the same proof works for @xmath49 .
we first employ the standard pointwise representation of solutions @xmath150 yielding @xmath151\\ & = 1+{\mathbb{e}}\big[\sum_{i\in{\mathbb{z}}^d}\int_0^tp_{t - r}(i , k)\sqrt{\kappa u(r , i)v(r , i)}\,dw^1_r(i ) \sum_{j\in{\mathbb{z}}^d}\int_0^{t+s}p_{t+s - l}(j , k)\sqrt{\kappa u(l , j)v(l , j)}\,dw^1_l(j)\big ] .
\end{aligned}\ ] ] further , since martingale increments are orthogonal this equals @xmath152 .
\end{aligned}\ ] ] now using independence of @xmath153 for @xmath154 and it s isometry we continue the chain of equalities as @xmath155\\ & = 1+\int_0^t\sum_{i\in{\mathbb{z}}^d } p_{t - r}(i , k)p_{t+s - r}(i , k)\kappa{\mathbb{e}}[u(r , i)v(r , i)]\,dr , \end{aligned}\ ] ] where we were allowed to change the order of integration since all terms are non - negative . using lemma [ duality ] , which in particular shows for homogeneous initial conditions that second moments do not depend on the spatial variable , symmetry of the transitions , and the chapman - kolmogorov equality , we finish with @xmath156\,dr\\ & = 1+\kappa\int_0^tp_{2t+s-2r}(k , k ) { \mathbb{e}}[e^{\kappa \varrho l_{r}}]\,dr = 1+\kappa\int_0^tp_{2r+s}(k , k ) { \mathbb{e}}[e^{\kappa \varrho l_{t - r}}]\,dr . \end{aligned}\ ] ] since we are going to examine the second lyapunov exponent @xmath72 of solutions we give a simple argument which ensures existence of the exponent .
[ existencegamma ]
let @xmath88 be a solution of the symbiotic branching model with homogeneous initial conditions .
then the lyapunov exponent @xmath72 exists .
note that to ensure existence of the limits @xmath157 \end{aligned}\ ] ] it suffices to show subadditivity of @xmath158 $ ] . using lemma [ duality ] this
is reduced to showing subadditivity of @xmath159 $ ] , where @xmath62 is the local time in @xmath64 of @xmath87 started in @xmath64 .
thus , by conditioning on @xmath160 , we get @xmath161 = \log { \mathbb{e}}^0[e^{\kappa\varrho l_{s}}{\mathbb{e}}^{\bar x_s}[e^{\kappa\varrho l_t } ] ] \leq\log { \mathbb{e}}^0[e^{\kappa\varrho l_{s}}{\mathbb{e}}^{0}[e^{\kappa \varrho l_t } ] ] = \log { \mathbb{e}}^0[e^{\kappa\varrho l_{s}}]+\log{\mathbb{e}}^{0}[e^{\kappa \varrho l_t } ] .
\end{aligned}\ ] ] in lemma [ duality ] , we observed that in order to study second moments of symbiotic branching processes it suffices to study exponential moments of local times of the symmetrization @xmath87 .
we now take up this issue and discuss exponential moments of @xmath62 in greater generality than needed for the symbiotic branching model . for
the following let @xmath74 be a time - homogeneous markov process with countable state space @xmath162 and transition kernel @xmath163 .
in particular , the transition rates are not assumed to be symmetric . + we start with a renewal - type equation for exponential moments of local times . [
l3]let @xmath62 be the local time of @xmath74 in @xmath164 for the process started in @xmath165 .
then for @xmath61 the following equation holds : @xmath60=1+\kappa \int_0^tp_r(i , i){\mathbb{e}}[e^{\kappa l_{t - r}}]\,dr,\quad t\geq 0 \label{eqn : maineqn}. \end{aligned}\ ] ] we use the exponential series to get @xmath60 & = { \mathbb{e}}\big[e^{\kappa \int_0^t \delta_i(x_s)\,ds}\big ] = { \mathbb{e}}\left[\sum_{n=0}^{\infty}\frac{\kappa^n}{n!}\left(\int_0^t \delta_i(x_{s})\,ds\right)^n\right]\\ & = 1+{\mathbb{e}}\left[\sum_{n=1}^{\infty}\frac{\kappa^n}{n!}\int_0^t \cdots \int_0^t
\delta_i(x_{s_1})\ldots \delta_i(x_{s_n})\,ds_n\ldots ds_1\right]\\ & = 1+{\mathbb{e}}\left[\sum_{n=1}^{\infty}\kappa^n\int_0^t \int_{s_1}^{t } \cdots \int_{s_{n-1}}^{t } \delta_i(x_{s_1})\ldots \delta_i(x_{s_n})\,ds_n\ldots d s_2 ds_1\right ]
. \end{aligned}\ ] ] the last step is justified by the fact that the function that is integrated is symmetric in all arguments and , thus , it suffices to integrate over a simplex .
we can exchange sum and expectation and obtain that the last expression equals @xmath166\,ds_n\ldots d s_2 ds_1 .
\end{aligned}\ ] ] due to the markov property , the last expression equals @xmath167 \,ds_n\ldots d s_2 ds_1 \end{aligned}\ ] ] and can be rewritten as @xmath168 \,ds_n\ldots d s_2 \right ) ds_1 .
\end{aligned}\ ] ] using the same line of arguments backwards for the term in parenthesis , the assertion follows .
a similar renewal - type equation as ( [ eqn : maineqn ] ) can be shown with essentially the same proof for a discrete - time markov process .
it reads @xmath169 = 1 + \kappa \sum_{n=0}^{m } p_n(i , i ) { \mathbb{e}}[e^{\kappa l_{m - n}}],\quad m\geq 1 , \end{aligned}\ ] ] where @xmath170 is the return probability after @xmath33 steps and @xmath171 is the number of visits after @xmath33 steps .
+ similar equations were obtained for symmetric markov chains on @xmath5 in @xcite using a completely different technique .
note that neither symmetry nor any structure of the set @xmath162 is needed .
the information on the geometry of @xmath162 is completely encoded in @xmath172 . for the rest of this section we fix the markov process @xmath74 , @xmath164 , and abbreviate @xmath173 .
\end{aligned}\ ] ] the return probabilities @xmath172 are always assumed to be strongly asymptotically equivalent to @xmath174 , as @xmath175 , for @xmath125 , as for instance for simple random walks on @xmath5 and the riemann walk on @xmath176 .
further , @xmath134 is monotone , decreasing , positive with @xmath177 , and @xmath178 is monotone , increasing , positive with @xmath179 .
the laplace transform for a function @xmath180 on @xmath181 is denoted by @xmath182 and the convolution of two functions @xmath183 is denoted @xmath184 . in this notation equation
[ eqn : maineqn ] reads @xmath185 taking the laplace transform of equation ( [ db ] ) leads to @xmath186 obviously , since @xmath134 is bounded by @xmath26 , @xmath187 is always finite for all @xmath188 . a priori this is not true for @xmath178 but if so , we obtain a useful representation from ( [ eqn : kappa0prime ] ) .
[ le ] if @xmath189 , then @xmath190 in the following we proceed in two steps .
first , we use ( [ db ] ) to understand in which cases @xmath191 grows exponentially in @xmath57 and discuss properties of the exponential growth rate .
the following corresponence between exponential growth and finiteness of laplace transforms holds ( existence of the limit was proven in lemma [ existencegamma ] ) : @xmath192 this observation is particularly important for the second step in which we discuss the behaviour of @xmath191 as @xmath65 . in the cases in which @xmath191 grows subexponentially ( [ a ] )
implies that @xmath189 for all @xmath193 .
hence , lemma [ le ] can be used for all @xmath193 .
the strategy in this case is the following : by assumption , the asymptotic behaviour of @xmath194 as @xmath57 tends to infinity is known , namely @xmath174 . using tauberian theorems the asymptotic behaviour of @xmath187 as @xmath195 tends to zero can be deduced . by lemma
[ le ] this determines the asymptotic behaviour of @xmath196 as @xmath195 tends to zero . using tauberian theorems in the opposite direction ,
the asymptotic behaviour of @xmath191 as @xmath57 tends to infinity is obtained .
+ to manage the transfer from the behaviour of @xmath134 to @xmath197 and back from @xmath198 to @xmath178 the following tauberian theorems are used .
they are taken from @xcite ( see theorem 1.7.6 , theorem 1.7.1 , corollary 8.1.7 , and the considerations at the beginning of section 8.1 ,
[ lem : moment - duality ] let @xmath180 be a monotone function on @xmath199 with @xmath200 , then the following hold : 1 . if @xmath129 and @xmath201 , then @xmath202 if and only if @xmath203 as @xmath204 .
2 . if @xmath205 , then @xmath206 as @xmath204 .
3 . if @xmath106 and @xmath207 , then @xmath208 and @xmath209 as @xmath204 . the main point of
the analysis is the following representation of the exponential growth rate which follows directly from lemma [ l3 ] .
[ prop1 ] let @xmath3 , then @xmath210=\hat{f}^{-1}\big(\frac{1}{\kappa}\big).\ ] ] first , ( [ a ] ) implies that @xmath211 moreover , @xmath212 we are done if we can show @xmath213 first we show `` @xmath214 '' . due to lemma [ l3 ]
we obtain @xmath215 which implies @xmath216 . since @xmath217 this shows that @xmath218 .
+ now we show `` @xmath219 '' . first , iterating ( [ db ] ) yields for fixed @xmath33 @xmath220 using @xmath221 and @xmath222\leq e^{\kappa t}$ ] yields @xmath223 as @xmath224 .
hence , for fixed @xmath2 @xmath225 taking laplace transforms we note that @xmath196 is finite if and only if the laplace transform of the right - hand side is finite .
however , using fubini s theorem ( note that only @xmath3 needs to be considered ) we obtain @xmath226 which is finite since we assumed @xmath227 .
in particular , the previous result shows that understanding @xmath228 suffices to understand the exponential growth rates of @xmath229 $ ] .
this is not difficult due to the following observation : @xmath197 is a strictly decreasing , convex function with @xmath230 .
hence , @xmath228 is a strictly decreasing , convex function with @xmath231 and @xmath232 if and only if @xmath233 .
this implies that @xmath234 precisely for @xmath235 .
this and more properties of the exponential growth rate are collected in the following corollary .
[ cor1 ] let @xmath3 and @xmath236 $ ] . then with @xmath237 the following hold
: * @xmath238 and @xmath239 if and only if @xmath240 , * the function @xmath241 is strictly convex for @xmath240 , * @xmath242 for all @xmath11 , and @xmath243 , as @xmath100 , * if @xmath103 , then @xmath244 and , as @xmath104 , @xmath245 * if @xmath106 , then @xmath246 and , as @xmath247 , @xmath248 parts i ) and ii ) are proven as argued above the corollary . + since @xmath249 , the first part of iii ) follows from @xmath250 continuity of @xmath134 and @xmath177
imply that for @xmath251 there is @xmath252 such that @xmath253 for @xmath254 .
hence , @xmath255 since @xmath256 for @xmath257 we obtain @xmath258 the second part of iii ) now follows since as well @xmath259 for all @xmath3 .
+ finally , for iv ) and v ) note that the asymptotics of @xmath197 for @xmath260 are known from lemma [ lem : moment - duality ] .
this translates to @xmath261 and hence to @xmath262 .
so far , we have understood the behaviour of @xmath229=g(t)$ ] as @xmath65 for @xmath263 . in this case @xmath191
grows exponentially and the behaviour of the exponential rates in @xmath11 could be analysed .
we now come to the case @xmath264 .
first , if @xmath265 , there is nothing to be done since the only appearing case is @xmath266 which yields @xmath267 for all @xmath2 .
hence , we can stick to @xmath268 .
[ prop10 ] let @xmath3 and @xmath269 .
then , as @xmath102 , @xmath270 \sim \frac{1}{1-\kappa g_{\infty}}. \end{aligned}\ ] ] since @xmath268 we can apply part iii ) of lemma [ lem : moment - duality ] .
hence , @xmath271 , as @xmath204 . as discussed above ,
since @xmath191 does not grow exponentially , @xmath272 for all @xmath193 , and we can use lemma [ le ] .
this implies @xmath273 as @xmath204 .
going backwards with lemma [ lem : moment - duality ] , part i ) , @xmath274 , the asymptotic of @xmath178 follows .
[ prop11 ] let @xmath3 and @xmath275 . then ,
as @xmath65 , @xmath270 \sim \begin{cases } t^{\alpha-1}\frac{\alpha-1}{\kappa c \gamma(2-\alpha)\gamma(\alpha ) } & : 1<\alpha<2,\\ \frac{t } { \log t } \frac{1}{\kappa c}&:\alpha=2,\\ t\frac{1}{\kappa h_{\infty } } & : \alpha>2 .
\end{cases } \end{aligned}\ ] ] since @xmath268 we can apply lemma [ lem : moment - duality ] , part iii ) .
hence , @xmath276 , as @xmath204 .
as discussed above , since @xmath191 does not grow exponentially , @xmath272 for all @xmath193 , and we can use lemma [ le ] . since @xmath277 , the denominator @xmath278 appearing in lemma [ le ] does not behave like a constant and we can not apply part i ) of lemma [ lem : moment - duality ] with @xmath274 .
instead we use lemma [ lem : moment - duality ] , part iii ) , to obtain @xmath279 as @xmath204 .
this , by lemma [ lem : moment - duality ] , part i ) , implies the assertion .
we now investigate equation ( [ eqn : maineqn ] ) for @xmath66 .
[ prop12 ] if @xmath66 , then , as @xmath102 , @xmath270 \sim \begin{cases } \frac{1}{t^{1-\alpha}}\frac{1}{-\kappa c \gamma(1-\alpha ) \gamma(\alpha ) } & : 0<\alpha<1,\\ \frac{1}{\log t}\frac{1}{-\kappa c } & : \alpha=1,\\ \frac{1}{-\kappa g_{\infty}+1}&:\alpha>1 .
\end{cases } \end{aligned}\ ] ] first note that for @xmath66 , @xmath222<1 $ ] and hence for all @xmath193 , @xmath189 which validates the use of lemma [ le ] .
this implies @xmath280 as @xmath204 .
using lemma [ lem : moment - duality ] in both directions returns the assertion .
these follow from lemma [ duality ] and corollary [ cor1 ] .
this follows from lemma [ duality ] and propositions [ prop10 ] and [ prop11 ] . for this proof
we always denote by @xmath93 the strong asymptotics at infinity and we abbreviate @xmath281 .
lemma [ appli ] implies @xmath282=\frac{\int_0^t p_{2r+s } { \mathbb{e}}[e^{\varrho \kappa l_{t - r } } ] \ , d r}{\sqrt{\int_0^t p_{2r } { \mathbb{e}}[e^{\varrho \kappa l_{t - r } } ] \ , d r \ , \int_0^{t+s } p_{2r } { \mathbb{e}}[e^{\varrho \kappa l_{t+s - r } } ] \ , d r}}. \end{aligned}\ ] ] _ step 1 , @xmath17 : _ first , assume @xmath106 which implies @xmath283 .
since @xmath284 we obtain , independently of the choice of @xmath57 and @xmath121 , @xmath282= \frac{\int_0^t p_{2r+s } \ , d r}{\sqrt{\int_0^t p_{2r } \ , d r \ , \int_0^{t+s } p_{2r}\ , d r } } \stackrel{s , t\to\infty}{\to } 0 .
\end{aligned}\ ] ] here , we used @xmath285 if @xmath286 .
we now come to the case @xmath127 , where we get @xmath287 therefore , we have @xmath282\sim\frac{\log\big(\frac{2 t + s + 1}{s+1}\big)}{\sqrt{\log ( 2 t + 1 ) \ , \log ( 2(t+s)+1)}}. \label{eqn : age1 } \end{aligned}\ ] ] for @xmath288 with @xmath289 this expression behaves asymptotically as @xmath290 on the other hand , for @xmath288 with @xmath291 the term in ( [ eqn : age1 ] ) behaves asymptotically as @xmath292 hence , for @xmath293 , we obtain @xmath294 \sim ( 1-a)_+.$ ] + now suppose @xmath129 . then @xmath295 therefore , we have @xmath282\sim\frac{(2 t+s+1)^{1-\alpha } -(s+1)^{1-\alpha}}{\sqrt{((2 t+1)^{1-\alpha } -1 ) \ , ( ( 2 ( t+s)+1)^{1-\alpha } -1)}}. \end{aligned}\ ] ] for @xmath296 this behaves asymptotically as @xmath297 _ step 2 , @xmath59 : _ let us first consider @xmath106 . since @xmath298
this case is exactly the same as @xmath17 , @xmath106 .
now suppose @xmath127 .
in this case we have by proposition [ prop12 ] @xmath299 we use the scaling @xmath288 with @xmath300 .
let @xmath301 .
the integral above can be split from @xmath302 to @xmath303 and @xmath303 to @xmath304 .
we treat the first integral and show that its order is less than @xmath305 . first note that in the range of integration @xmath306 therefore , @xmath307 on the other hand , the second integral can be treated as follows . in its range of integration
we have @xmath308 therefore , @xmath309 thus , @xmath310 \sim \frac{\frac{1}{\log ( t ) } \ , \frac{1}{2 } \log \left(\frac{2(1 - \theta ) t+s}{2+s}\right)}{\sqrt{\frac{1}{\log ( t ) } \ , \frac{1}{2 } \log ( ( 1 - \theta ) t)\ , \frac{1}{\log ( t ) } \ , \frac{1}{2 } \log ( ( 1 - \theta)(t+s ) ) } } \sim \frac { \log \left(\frac { t}{t^a}\right)}{\sqrt{\log ( t)\log ( t ) } } = 1-a.\end{aligned}\ ] ] analogously , for case @xmath291 .
therefore , we get @xmath294\sim ( 1-a)_+,$ ] whenever @xmath311 .
+ for @xmath59 only @xmath129 is left : here , we have @xmath312 we set @xmath313 .
the integral can be rewritten as @xmath314 the same way one can see that @xmath315 d r
\sim \frac{c}{-\kappa \varrho c \gamma(\alpha)\gamma(1-\alpha ) } \int_0 ^ 1 ( 2 r)^{-\alpha } ( 1-r)^{\alpha-1 } d r.\end{aligned}\ ] ] thus , @xmath310 & \sim \frac{\int_0 ^ 1 ( 2 r+ a)^{-\alpha } ( 1-r)^{\alpha-1 } d r}{\int_0 ^ 1 ( 2 r)^{-\alpha } ( 1-r)^{\alpha-1 } d r}\\ & = \frac{\int_0 ^ 1 ( 2 r+ a)^{-\alpha } ( 1-r)^{\alpha-1 } d r } { 2^{-\alpha } b(\alpha,1-\alpha)}= \frac{\int_0 ^ 1 ( 2 r+ a)^{-\alpha } ( 1-r)^{\alpha-1 } d r } { 2^{-\alpha } \gamma(\alpha)\gamma(1-\alpha)},\end{aligned}\ ] ] when @xmath316 . here , @xmath317 denotes the beta function .
+ _ step 3 , @xmath90 : _ the transient case ( @xmath106 ) with @xmath110 has already appeared in the case @xmath17 for @xmath106 .
@xmath318 $ ] is bounded due to proposition [ prop10 ] .
hence , no aging occurs .
+ for @xmath96 , and for @xmath103 , we proved in corollary [ cor1 ] i ) that @xmath319 $ ] grows exponentially .
this implies that there is a @xmath320 depending on the growth rate , i.e. on @xmath18 and @xmath11 , ( for example @xmath321 does the job ) such that for all @xmath2 and @xmath322 we have @xmath323 \leq e^{-\grw r } { \mathbb{e}}[e^ { \varrho \kappa l_t } ] .
\end{aligned}\ ] ] therefore , @xmath324 \ , d r \leq p_{s } { \mathbb{e}}[e^ { \varrho \kappa l_{t } } ] \int_0^t e^{-\grw r } \ , d r \leq \frac{1}{\grw}\ , p_{s } { \mathbb{e}}[e^ { \varrho \kappa l_{t } } ] . \end{aligned}\ ] ] on the other hand , by the assumption @xmath325 , the renewal - type equation of lemma [ l3 ] , and since @xmath90 , @xmath326 \ , d r \geq c \int_0^t p_{r } { \mathbb{e}}[e^ { \varrho \kappa l_{t - r } } ] \ , d r = \frac{c}{\varrho \kappa } \ , \left ( { \mathbb{e}}[e^ { \varrho \kappa l_{t } } ] -1\right ) \geq c ' { \mathbb{e}}[e^ { \varrho \kappa l_{t } } ] .
\end{aligned}\ ] ] putting these pieces together we obtain that @xmath327 \leq c''\ , \frac{p_{s } { \mathbb{e } } [ e^ { \varrho \kappa l_{t } } ] } { \sqrt{{\mathbb{e } } [ e^ { \varrho \kappa l_{t } } ] { \mathbb{e } } [ e^ { \varrho \kappa l_{t+s } } ] } } = c''p_{s}\ , \sqrt { \frac { { \mathbb{e } } [ e^ { \varrho \kappa l_{t } } ] } { { \mathbb{e } } [ e^ { \varrho \kappa l_{t+s } } ] } } .
\end{aligned}\ ] ] note that , since @xmath90 , the term with the square root is bounded by @xmath26 . since clearly @xmath328 tends to zero , the whole expression must tend to zero independently of how @xmath329 . +
the only case left is @xmath106 and @xmath112 .
first we consider @xmath330 .
@xmath331 \ , d r \approx
\int_0^t ( r+1)^{-\alpha } ( t - r)^{\alpha-1 } \ , d r = \int_0 ^ 1 ( r+1/t)^{-\alpha } ( 1-r)^{\alpha-1 } \ , d r.\end{aligned}\ ] ] this expression tends to infinity for @xmath65 .
the rate is @xmath332 on the other hand , @xmath333 \ , d r \approx \int_0^t ( r+s)^{-\alpha } ( t - r)^{\alpha-1 } \ , d r = \int_0 ^ 1 ( r+s / t)^{-\alpha } ( 1-r)^{\alpha-1 } \ , d r.\end{aligned}\ ] ] this expression is bounded or tends to infinity ( depening on how @xmath329 ) .
it is bounded by @xmath334 putting these pieces together we obtain that @xmath335 \ , d r } { \sqrt{\int_0^t p_{2 r } { \mathbb{e } } [ e^ { \varrho \kappa l_{t - r } } ] \ , d r \int_0^{t+s } p_{2 r } { \mathbb{e } } [ e^ { \varrho \kappa l_{t+s - r } } ] \ , d r } } \leq c ' \,\frac { c_1 + c_2 ( s / t)^{1-\alpha}}{\sqrt{t^{\alpha-1 } ( t+s)^{\alpha-1 } } } \to 0.\end{aligned}\ ] ] the calculation is completely analogous in the cases @xmath336 .
the pointwise representation of ( [ gr ] ) is not restricted to the symbiotic branching model but also holds for the interacting diffusions of this proposition .
hence , the same derivation as for the symbiotic branching model yields @xmath337 = \frac{\int_0^tp_{2r+s}(k , k){\mathbb{e}}[f(u(t - r , k))]\,dr}{\sqrt{\int_0^tp_{2r}(k , k){\mathbb{e}}[f(u(t - r , k))]\,dr\int_0^{t+s}p_{2r}(k , k){\mathbb{e}}[f(u(t+s - r , k))]\,dr}}. \end{aligned}\ ] ] part i ) is contained in theorem [ t3 ] since the parabolic anderson model appears as special case @xmath41 .
for part ii ) we can estimate the expectations from above and below to obtain the same result as in theorem [ t5 ] ii ) except constants .
the same is true for part iii ) since the pointwise representation ( [ gr ] ) implies @xmath338=1 $ ] .
finally , we could interprete part iv ) as a submodel of the symbiotic branching model . instead
, we give a direct proof using use the coalescing particles dual of @xcite .
the dual process consists of two independent particles started in @xmath339 , performing transitions @xmath4 in continuous time . after spending an exponential time @xmath147 with parameter @xmath11 , independent of the particles , at same sites ,
the particles coalesce .
we denote @xmath148 the law of the particles and suppose @xmath340
. then @xmath114&={\mathbb{e}}^{y\times x}[w^{\text{number of non - coalesced particles}}]\\ & = { \mathbb{e}}^{y\times x}[w1_{y\leq l_t}]+{\mathbb{e}}^{y\times x}[w^21_{y > l_t } ] = w(1-{\mathbb{e}}^x[e^{-\kappa l_t}])+w^2{\mathbb{e}}^x[e^{-\kappa l_t } ] , \end{aligned}\ ] ] where @xmath62 denotes the collision time of the particles .
using @xmath338=w$ ] this yields @xmath341={\mathbb{e}}[u(t , k)]-{\mathbb{e}}[u(t , k)^2]=(w - w^2){\mathbb{e}}[e^{-\kappa l_t}]$ ] and we can proceed as for the symbiotic branching model with @xmath59 .
we are grateful to j. grtner from tu berlin for simplifying our original proof of lemma [ l3 ] .
m. b. marcus and j. rosen .
moment generating functions for local times of symmetric markov processes and random walks . in _ probability in banach spaces , 8 ( brunswick , me , 1991 ) _ , volume 30 of _ progr .
_ , pages 364376 .
birkhuser boston , boston , ma , 1992 .
t. shiga .
stepping stone models in population genetics and population dynamics . in _
stochastic processes in physics and engineering ( bielefeld , 1986 ) _ , volume 42 of _ math .
_ , pages 345355 .
reidel , dordrecht , 1988 . | for the symbiotic branching model introduced in @xcite , it is shown that aging and intermittency exhibit different behaviour for negative , zero , and positive correlations .
our approach also provides an alternative , elementary proof and refinements of classical results concerning second moments of the parabolic anderson model with brownian potential . some refinements to more general ( also infinite range ) kernels of recent aging results of @xcite for interacting diffusions are given . , |
attention received by porous silicon ( p - si ) from scientists is known from more than couple of decades since the discovery of its light emission properties [ 1,2 ] .
p - si is still being studied due to its applications in various fields in addition to optoelectronic devices [ 1 - 7 ] .
many reports shows applications of p - si in advanced sensing[8 - 11 ] , in making electrodes for batteries [ 12,13 ] , in carrying out mass spectroscopy [ 14 ] in carrying out mass spectroscopy [ 15 ] and in field effect transistors [ 16 , 17 ] .
different methods to fabricate p - si & it s various applications / properties have been studied in depth and very good reports are available [ 18 , 19 ] .
various techniques used for the formation of p - si include strain etching [ 20 , 21 ] of si wafer in a solution of hno @xmath0 & hf acid , electrochemical etching [ 22 , 23 ] where an electrical bias is used for porosification of si wafer in etching solution of hf acid . a simpler method of photochemical etching also known as laser induced etching ( lie ) was used by choy and cheah [ 24 ] and later used by many others [ 25 - 29 ] .
lie is mainly used for preparing p - si in a localized area on a si wafer without using an electrical bias .
a combination of two techniques ( lie and electrochemical etching ) has also been reported [ 30 , 31 ] for fabrication of p - si .
all these methods have certain advantages and disadvantages over each other and are used depending on the convenience & suitability .
lie method and stain etching method does not require any bias for making p - si whereas a bias is needed while using electrochemical etching .
stain etching is the easiest method to use but is a very slow process and does not give a
reproducible & reliable result [ 32 ] .
in addition to the above - mentioned techniques , method of metal assisted chemical etching or metal induced etching ( mie ) [ 3238 ] is also used for preparing p - si at faster rates as compared to stain etching .
the mie technique , in contrast to stain etching technique , gives good reproducibility in p - si formation without using any bias or laser radiation .
moreover , mie technique does not compromise on the light emitting properties of p - si in contrary to that prepared by stain etching .
p - si prepared by lie technique , which also is an electrodeless technique like mie technique , gives visible photoluminescence ( pl ) at room temperature similar to the p - si prepared by mie .
light emitting p - si , fabricated using these two electrodeless techniques of mie and lie , should be studied to investigate the structural difference in resulting p - si .
this will be helpful in understanding various mechanisms involved in p - si fabrication and will enable scientist to explore the application of p - si beyond the traditional applications . a comparative study will also provide knowledge to control the pore size & other parameter to tailor different properties of p - si . though the basic p - si fabrication mechanism is the same in lie and mie process ( dissolution of si atoms from si wafer as a result of reaction with hf acid ) , there are some similarities and dissimilarities between the two processes and can be summarized as follows . in the lie method , n - type crystalline si substrate , while dipped in hf acid solution , is irradiated with laser light thereby creating excess photoexcited holes on the irradiated surface .
these holes initiate the etching process [ 29 ] and result in the formation of p - si that contains si nanostructures . on the other hand in mie ,
first metal nano particles are grown on si wafer surface before the si wafer is immersed in etchant solution ( hf + h@xmath1o@xmath1 ) .
detailed mechanism of pore formation in lie and mie have been reported in literature [ 29 , 33 ] but a comparative study of pore formation using both of these techniques is not available and is the theme of present paper .
aim of this paper is to present basic properties of p - si fabricated using mie technique and compare its morphological and pl properties with those of p - si prepared by lie technique earlier [ 29 ] .
a comparison of pore size is presented using scanning electron microscopy ( sem ) with explanation for the existing difference in pore size in samples prepared using the two techniques .
a cross- ( x- ) sectional sem , done to investigate the pore growth process shows submicron sized pore formation in mie samples .
a uniformly distributed , well - aligned si nano wire like structures are seen from which a possibility of visible light emission has also been investigated using pl spectroscopy .
p - si samples have been prepared using above - mentioned two techniques namely lie and mie .
the lie samples were prepared from commercially available wafers of n- si(100 ) having resistivity of 35 @xmath2 cm .
these wafers were cleaned in acetone and ethanol to remove impurities prior to starting the porosification process .
mounted on two teflon plates , these wafers were immersed in the hf acid ( 40 % ) in a plastic container .
the lie [ 27 , 29 ] was performed by using an argon ion laser beam ( photon energy of 2.41 ev ) focused onto the si wafer with laser power density of 1.76 kw/@xmath3 .
two samples were prepared by irradiating the wafers for 45 minutes ( _ sample : l45 _ ) and 60 minutes ( _ sample : l60 _ ) with the abovementioned etching laser power density .
the mie samples were prepared by silver ( ag ) -assisted chemical etching of n - si ( 100 ) wafer . the cleaned wafers ( as discussed above )
were immersed in hf solution to remove thin oxide layer formed at surface .
these wafers were dipped in solution containing 4.8 m hf & 5 mm agno@xmath0 for one minute at room temperature to deposit ag nano particles ( agnps ) .
the agnps deposited samples were then kept for etching in an etching solution containing 4.6 m hf & 0.5 m h@xmath1o@xmath1 for 45 minutes ( sample : m45 ) and 60 minutes ( sample : m60 ) .
etched wafers were transferred in hno@xmath0 acid to dissolve ag metal .
then the samples were dipped into hf solution to remove oxide layer induced by nitric acid used in above step .
surface morphology of all the four samples have been characterized using field - emission scanning electron microscopy ( sem ) , supra55 zeiss .
the pl from mie sample was recorded with 325 nm laser excitation source using dongwoo optron 80k pl system at room temperature .
surface morphologies , observed using sem , of all the four samples ( l45 , l60 , m45 & m60 ) have been shown in fig .
1 . figure 1a and 1b shows sem images of samples prepared using lie technique whereas the sem images of mie samples are shown in fig .
1c and 1d .
the scale bars in fig . 1a 1d correspond to 10 @xmath4 m .
porous si with pore size of @xmath510 @xmath4 m can be seen in fig .
1a and 1b corresponding to samples l45
& l60 respectively .
comparatively wider and connected pores [ 29 ] can be seen in fig . 1b as compared to fig .
1a due to increased etching time . for a given laser power density ,
when the etching time is increased , more si atoms gets dissolved in hf solution due to chemical etching reaction to result in deeper & wider pores .
it is important here to mention that in the absence of laser , etching does not take place and porosification is not possible .
an sem image of such a control sample is shown in fig .
s1 in the sm .
effect of etching time and laser power density on lie process have been studied in detail and has been reported earlier [ 27 , 29 ] . on the other hand , figs.1c and 1d show sem images corresponding to mie samples for etching times of 45 minutes ( sample m45 ) and 60 minutes ( sample m60 ) respectively .
submicron sized pores are formed in mie samples in contrast with wider pores for the case of lie samples ( fig .
wider pores are formed during lie because porosofication can take place in lateral as well as vertical direction whereas etching lead by agnps during mie take place faster in vertically downward direction rather than laterally to result in relatively smaller pores in mie .
m , height=396 ] comparison of sem images in fig . 1 corresponding to lie samples ( fig
. 1a-1b ) and mie samples ( fig1c-1d ) also reveals that the pores are densely packed for mie samples as compared to lie samples .
it is also evident that the pores are relatively uniform in size for mie samples whereas the pores formed using lie is not uniform .
this difference can be understood as follows .
as known , the lie process is a laser power dependent process [ 29 ] .
a laser beam having gaussian intensity profile is used for lie of si wafer in hf acid .
the gaussian laser beam with maximum intensity at the center of the beam induces highest etching rate at the center as compared to that near the edge of the laser beam . as a result , a variation in the pores size is expected in lie samples .
however this is not very clear from the sem images shown in fig .
1a-1b , a distribution in size of si nanostructures contained in lie samples has been reported to explain observed raman line - shape earlier [ 3942 ] . on the other hand ,
the mie is initiated by agnps on the si wafer deposited prior to chemical etching in etchant solution as discussed in the experimental section .
it can be understood that the pore size and pore density in the case of mie samples are direct manifestation of agnps size and distribution of agnps on the wafer .
furthermore , lie technique is suitable for localized growth of p - si and mie technique is suitable if one needs p - si over a relatively large area .
size of samples prepared using lie is controlled by the spot size of laser and are typically are @xmath5100 @xmath4 m in size ( sample size ) due to which it is not possible to obtain a x - sectional image of lie samples . on the other hand ,
mie samples give us the option to investigate the pores growth by analyzing the x - sectional images .
x - sectional sem images of the mie samples are shown in fig .
2a-2b corresponding to samples m45 and m60 respectively . a very clear nanowire ( submicron sized ) like features can be seen for both the samples .
the scale bars on figs .
2a-2b are of 20 @xmath4 m and those on figs .
2c-2d are of 4 @xmath4 m in length .
approximate lengths of the nanowires in for sample m45 is 60 @xmath4 m whereas that for m60 it is 80 @xmath4 m .
this increase in nanowires ( or pore ) length is a direct manifestation of longer metal induced chemical etching of si wafer .
the parallel pores in sample m60 are more uniform as compared to sample m45 . to have a closer look at the nanowires formed
, the x - sectional images were magnified as shown in fig .
figures 2c-2d show the magnified view of the portion marked in figs 2a-2b respectively .
it is clear from fig .
2c-2d also that aligned uniform nanowires are formed in sample m60 & sample m45 .
such kind of parallel si nanowires may be used as field emitters [ 43 ] for various applications . as discussed above ,
the nanowires are formed parallel to each other when prepared using mie method because the growth of such nanowires are initiated and governed by the distribution and size of agnps grown prior to chemical etching .
the x - sectional images for lie samples can not be presented here due to sample constraints because it is not possible to prepare the lie samples for x - sectional sem imaging .
the sem results discussed in fig .
2 , indicates that the dimensions of si nanowires prepared as a result of mie might be sufficiently smaller to show quantum confinement effect .
pl measurement was carried out to investigate the light emitting properties of samples prepared by mie technique .
it is important here to mention that samples prepared using lie also show visible pl as reported earlier [ 26 , 38 , 39 , 42 ] .
a visible pl at room temperature form semiconductor nanostructures is used as a signature of quantum confinement effect .
figure 3 shows the room temperature pl spectrum of the mie sample m45 .
figure 3 shows a pl peak centered at 1.96 ev and is @xmath5 370 mev broad .
the pl peak in visible region with 370 mev of broadening is a clear indication of confinement effect in the mie sample .
inset of fig .
3 shows the red color emission corresponding to 1.96 ev pl emission under illumination of 325 nm laser .
the image shown in inset has been captured during the pl recording with uv laser ( 325 nm excitation ) .
similar kind of visible pl has been observed from si nanostructures prepared using lie technique as reported earlier [ 39 ] .
figure 3 shows only one peak which is in contrast with a two peak behavior for lie samples as reported earlier [ 39 ] .
the number of pl peaks in a spectrum provides information about the available sizes in the nanostructures sample .
single pl peak in mie sample as shown in fig .
3 means that the si nanostructures are uniform and does not contain a range of sizes .
in contrast , it is reported [ 39 ] that a double pl peak is observed from lie samples due to two different dominant nanostructure sizes available in lie samples .
lie samples contain different sized si nanostructures because a gaussian beam is used to prepare the sample using lie .
variation in etching laser power density in a gaussian beam results in si nanostructures of different sizes in different region of p - si sample . on the other hand ,
the nanostructures sizes are controlled by metal nanoparticles for mie samples and result in uniform si nanostructure .
in summary , a comparison of p - si prepared using lie and mie techniques reveals that smaller and uniform pores are formed in mie samples as compared to lie ones . during the p - si formation using mie technique , the pores start forming from the sites ( on wafer ) where silver nanoparticles are present .
uniformly distributed silver nanoparticles lead the etching process and pores proceed mainly in the downward direction making etching more favorable in the vertical direction and lesser etching take place in the lateral direction ( parallel to the wafer surface ) . on the other hand , in lie method the etching take place in both the direction ( vertically down the wafer and parallel to the wafer ) , of course with different etching rates . as a result in the same time
the pores formed in lie samples are wider as compared to those formed on mie samples .
in addition , the etching initiates at silver nanoparticles sites distributes on the si surface in the case of mie samples to result in relatively uniform pore distribution as compared to lie samples .
a laser beam having gaussian profile of intensity used for lie gives rise a distribution in pore size in lie samples & si nanostructures are fabricated . a x - sectional sem image from mie samples
show that well aligned si nanowires are formed with increasing sizes when etching time is increased .
these parallel si nanowires may be used for field emission devices and opens scope for further studies .
the si nanowires in the p - si samples prepared by mie show visible pl at 1.96 ev ( red emission ) at room temperature possibly due to weak quantum confinement effect .
the pl spectrum show only single peak for mie samples in contrary with already reported double pl peak for lie samples giving confirmation that the si nanowires are of relatively uniform size in mie samples as compared to lie samples .
authors are thankful to sophisticated instrument centre ( sic ) , iit indore ( iit indore ) for providing photoluminescence and fesem measurement facilities .
authors acknowledge technical support from mr .
farhan babu ( iit indore ) in sem experiment .
canham lt ( 1990 ) silicon quantum wire array fabrication by electrochemical and chemical dissolution of wafers .
appl phys lett 57:10461048 .
doi : doi:10.1063/1.103561 cullis ag , canham lt ( 1991 ) visible light emission due to quantum size effects in highly porous crystalline silicon .
nature 353:335338 .
doi : 10.1038/353335a0 cullis ag , canham lt , calcott pdj ( 1997 ) the structural and luminescence properties of porous silicon .
j appl phys 82:909965 .
doi : doi:10.1063/1.366536 maier - flaig f , rinck j , stephan m , et al . ( 2013 ) multicolor silicon light - emitting diodes ( sileds ) .
nano lett .
doi : 10.1021/nl3038689 soref r ( 2010 ) silicon photonics : a review of recent literature .
silicon 2:16 .
doi : 10.1007/s12633 - 010 - 9034-y soref r ( 2010 ) mid - infrared photonics in silicon and germanium .
nat photonics 4:495497 .
doi : 10.1038/nphoton.2010.171 chen r , li d , hu h , et al . ( 2012 ) tailoring optical properties of silicon nanowires by au nanostructure decorations : enhanced raman scattering and photodetection .
j phys chem c 116:44164422 .
doi : 10.1021/jp210198u .
ltant s , sailor mj ( 2000 ) detection of hf gas with a porous silicon interferometer .
adv mater 12:355359 .
doi : 10.1002/(sici)1521 - 4095(200003)12:5<355::aid - adma355>3.0.co;2-h lin vs - y , motesharei k , dancil k - ps , et al .
( 1997 ) a porous silicon - based optical interferometric biosensor .
science 278:840843 .
doi : 10.1126/science.278.5339.840 .
canham lt ( 1995 ) bioactive silicon structure fabrication through nanoetching techniques .
adv mater 7:10331037 .
doi : 10.1002/adma.19950071215 wang b , cancilla jc , torrecilla js , haick h ( 2014 ) artificial sensing intelligence with silicon nanowires for ultraselective detection in the gas phase .
nano lett 14:933938 .
doi : 10.1021/nl404335p flavel bs , sweetman mj , shearer cj , et al .
( 2011 ) micropatterned arrays of porous silicon : toward sensory biointerfaces .
acs appl mater interfaces 3:24632471 .
10.1021/am2003526 ge m , rong j , fang x , zhou c ( 2012 ) porous doped silicon nanowires for lithium ion battery anode with long cycle life .
nano lett 12:23182323 .
doi : 10.1021/nl300206e zhao y , liu x , li h , et al .
( 2012 ) hierarchical micro / nano porous silicon li - ion battery anodes .
chem commun 48:50795081 .
doi : 10.1039/c2cc31476b wei j , buriak jm , siuzdak g ( 1999 ) desorption ionization mass spectrometry on porous silicon .
nature 399:243246 .
doi : 10.1038/20400 li q , koo s - m , edelstein md , et al .
( 2007 ) silicon nanowire electromechanical switches for logic device application .
nanotechnology 18:315202 .
doi : 10.1088/0957 - 4484/18/31/315202 wang b , stelzner t , dirawi r , et al .
( 2012 ) field - effect transistors based on silicon nanowire arrays : effect of the good and the bad silicon nanowires .
acs appl mater interfaces 4:42514258 .
doi : 10.1021/am300961d fll h , christophersen m , carstensen j , hasse g ( 2002 ) formation and application of porous silicon .
mater sci eng r rep 39:93141 .
doi : 10.1016/s0927 - 796x(02)00090 - 6 korotcenkov g , cho bk ( 2010 ) silicon porosification : state of the art .
crit rev solid state mater sci 35:153260 .
doi : 10.1080/10408436.2010.495446 vzsonyi , szilgyi e , petrik p , et al .
( 2001 ) porous silicon formation by stain etching .
thin solid films 388:295302 .
doi : 10.1016/s0040 - 6090(00)01816 - 2 gonzlez - daz b , guerrero - lemus r , marrero n , et al .
( 2006 ) anisotropic textured silicon obtained by stain - etching at low etching rates .
j phys appl phys 39:631 .
doi : 10.1088/0022 - 3727/39/4/006 shih s , jung kh , hsieh ty , et al .
( 1992 ) photoluminescence and formation mechanism of chemically etched silicon .
appl phys lett 60:18631865 .
doi : 10.1063/1.107162 astrova ev , borovinskaya tn , tkachenko av , et al .
( 2004 ) morphology of macro - pores formed by electrochemical etching of p - type si .
j micromechanics microengineering 14:1022 .
doi : 10.1088/0960 - 1317/14/7/024 choy ch , cheah kw ( 1995 ) laser - induced etching of silicon .
appl phys a 61:4550 .
doi : 10.1007/bf01538209 hadjersi t , gabouze n , yamamoto n , et al .
( 2004 ) photoluminescence from photochemically etched highly resistive silicon .
thin solid films 459:249253 .
doi : 10.1016/j.tsf.2003.12.103 yamamoto n , takai h ( 1999 ) blue luminescence from photochemically etched silicon .
jpn j appl phys 38:57065709 .
doi : 10.7567/jjap.38.5706 mavi hs , prusty s , kumar m , et al .
( 2006 ) formation of si and ge quantum structures by laser - induced etching .
phys status solidi -appl mater sci 203:24442450 .
doi : 10.1002/pssa.200521027 mavi hs , islam ss , kumar r , shukla ak ( 2006 ) spectroscopic investigation of porous gaas prepared by laser - induced etching .
j non - cryst solids 352:22362242 .
doi : 10.1016/j.jnoncrysol.2006.02.046 kumar r , mavi hs , shukla ak ( 2008 ) macro and microsurface morphology reconstructions during laser - induced etching of silicon .
micron 39:287293 .
doi : 10.1016/j.micron.2007.04.005 kang y , jorn j ( 1998 ) photoelectrochemical dissolution of n - type silicon .
electrochimica acta 43:23892398 .
doi : 10.1016/s0013 - 4686(97)10150 - 5 juhasz r , linnros j ( 2002 ) silicon nanofabrication by electron beam lithography and laser - assisted electrochemical size - reduction .
microelectron eng 6162:563568 .
doi : 10.1016/s0167 - 9317(02)00532 - 4 li x , bohn pw ( 2000 ) metal - assisted chemical etching in hf / h2o2 produces porous silicon .
appl phys lett 77:25722574 .
doi : 10.1063/1.1319191 huang z , geyer n , werner p , et al .
( 2011 ) metal - assisted chemical etching of silicon : a review .
adv mater 23:285308 .
doi : 10.1002/adma.201001784 huang z , shimizu t , senz s , et al .
( 2009 ) ordered arrays of vertically aligned [ 110 ] silicon nanowires by suppressing the crystallographically preferred < 100 > etching directions .
nano lett 9:25192525 .
doi : 10.1021/nl803558n huang z , zhang x , reiche m , et al .
( 2008 ) extended arrays of vertically aligned sub-10 nm diameter [ 100 ] si nanowires by metal - assisted chemical etching .
nano lett 8:30463051 .
doi : 10.1021/nl802324y lin l , guo s , sun x , et al .
( 2010 ) synthesis and photoluminescence properties of porous silicon nanowire arrays .
nanoscale res lett 5:1822 .
doi : 10.1007/s11671 - 010 - 9719 - 6 chartier c , bastide s , lvy - clment c ( 2008 ) metal - assisted chemical etching of silicon in hf
electrochimica acta 53:55095516 .
doi : 10.1016/j.electacta.2008.03.009 kumar r , mavi hs , shukla ak , vankar vd ( 2007 ) photoexcited fano interaction in laser - etched silicon nanostructures .
j appl phys 101:064315 .
doi : 10.1063/1.2713367 kumar r , shukla ak ( 2009 ) quantum interference in the raman scattering from the silicon nanostructures .
phys lett a 373:28822886 .
doi : 10.1016/j.physleta.2009.06.005 kumar r , mavi hs , shukla ak ( 2010 ) spectroscopic investigation of quantum confinement effects in ion implanted silicon - on - sapphire films .
silicon 2:2531 .
doi : 10.1007/s12633 - 009 - 9033-z shukla ak , kumar r , kumar v ( 2010 ) electronic raman scattering in the laser - etched silicon nanostructures .
j appl phys 107:014306 .
doi : 10.1063/1.3271586 wu h - c , tsai t - y , chu f - h , et al .
( 2010 ) electron field emission properties of nanomaterials on rough silicon rods .
j phys chem c 114:130133 .
doi : 10.1021/jp908566q | porous silicon ( p - si ) , prepared by two routes ( metal induced etching ( mie ) and laser induced etching ( lie ) ) have been studied by comparing the surface morphologies .
a uniformly distributed smaller ( submicron sized ) pores are formed when mie technique is used because the pore formation is driven by uniformly distributed metal ( silver in present case ) nanoparticles , deposited prior to the porosification step . whereas in p - si samples prepared by lie technique , wider pores with some variation in pore size as compared to mie technique
is observed because a laser having gaussian profile of intensity is used for porosification .
uniformly distribute well - aligned si nanowires are observed in samples prepared by mie method as seen using cross - sectional sem imaging , which shows a single photoluminescence ( pl ) peak at 1.96 ev corresponding to red emission at room temperature .
the single pl peak confirms the presence of uniform sized nanowires in mie samples .
these vertically aligned si nanowires can be used for field emission application . |
controlling crystallization is a subject of considerable importance both from a fundamental and an applied perspective .
chemical physicists want to understand how the interparticle interactions determine the most favoured crystal structure and the ease with which crystallization can occur .
biochemists want to know the best recipe to crystallize the protein in which they are interested , so that they can then determine its structure .
nanotechnologists want to know how to get colloids or nanoparticles to self - assemble into a given target structure , such as a diamond lattice because of its potential importance in photonics .
metallurgists want to be able to predict which alloys will most readily avoid crystallization , and instead form a metallic glass.@xcite in this paper we want to take a theoretical and computer simulation perspective on the factors that control crystallization , and its absence , in these kinds of systems .
so far in the literature , isotropic models have been the starting point for much of the theoretical work in this area , and although there have been considerable successes , this approach has its limits .
for example , the experimentally observed `` crystallization slot '' for globular proteins , where the interprotein interactions , as measured by the second virial coefficient , are sufficiently attractive to encourage crystallization , but not so attractive as to lead to irreversible aggregation,@xcite has been rationalised using isotropic model potentials.@xcite another fruitful idea is that if a protein solution lies near to a metastable critical point associated with separation into protein - rich and solvent - rich phases , the associated concentration fluctuations could enhance crystal nucleation.@xcite however , proteins are not isotropic , but are anisotropic both in their shape and in their interactions .
one reflection of the latter is that protein crystals typically have much lower packing fractions@xcite than the close - packed structures that are favoured by isotropic potentials .
similarly , for colloids and nanoparticles there are now a considerable array of different crystals that have been obtained just using isotropic particles , particularly through the use of binary mixtures .
for example , just by varying the relative sizes of the two types of particles , surprisingly complex crystals can be formed.@xcite very recently , the use of particles with charges of opposite signs has led to colloidal@xcite and nanoparticle @xcite analogues of ionic crystals , even relatively low - density structures such as zinc blende.@xcite however , there are limits to the structures that can be assembled from isotropic particles,@xcite and so there has been considerable recent interest in developing methods to produce colloidal particles that are anisotropic in shape@xcite or in their interactions.@xcite for all these reasons , there has been much theoretical interest in beginning to explore simple anisotropic models and their effects on crystallization@xcite and self - assembly.@xcite here we provide a particular viewpoint on this broad area of controlling crystallization and its absence . in section [ sect : pxtal ] we outline some of our perspectives on protein crystallization , particularly why it is important to take into account the evolutionary origins of proteins when trying to understand their crystallization behaviour . in section [
sect : glass ] we explore how the concept of frustration can provide physical insight into the structural origins of the glass - forming ability for two of the most common models for simulating supercooled liquids .
finally , in section [ sect : patchy ] , we introduce some of our recent results that use model patchy particles to explore how the geometry of the interparticle interactions can both influence and be used to control the crystallization and self - assembly behaviour .
the question concerning protein crystallization that we have been particularly seeking to address@xcite is : why are globular proteins seemingly so hard to crystallize ?
although it is not true of all proteins lysozyme and insulin provide notable examples of proteins that crystallize relatively easily there is plenty of anecdotal evidence that proteins are often very difficult to crystallize .
furthermore , the rise of structural genomics initiatives , has now made it possible to start to quantify this difficulty .
for example , the success rate for producing x - ray quality crystals has been estimated to be roughly 20% for those prokaryotic proteins that can be expressed in soluble form.@xcite moreover , the crystallization approaches used are mainly empirical .
for example , screening kits allow large numbers of solution conditions , which have previously been found to be useful for crystallization , to be quickly tested .
our perspective on this question@xcite is that the need for proteins to function within the crowded cellular environment places major constraints on the surface properties of proteins . in particular , not only must a protein interact correctly with its binding partners , but it must also make sure that it does not stick to anything else in the cell , including other copies of itself . failure to do so is likely to be deleterious to the cell , as in the many protein aggregation diseases , of which the most relevant to this paper are those associated with native state aggregation and crystallization .
for example , sickle cell anaemia involves the ordered aggregation of hemoglobin into fibrils , and some forms of anaemia and cataracts are the result of the crystallization of mutant forms of the haemoglobin@xcite and @xmath0-crystallin@xcite proteins , respectively .
thus , our hypothesis is that the surface properties of proteins have been selected in order to prevent native - state aggregation and crystallization _ in vivo _ ( we term this ` engineering out ' of unwanted properties negative design ) and that , because of the robustness of the mechanisms used , many proteins are difficult to crystallize even in the far - from - physiological conditions explored by the protein crystallographer .
furthermore , there is a relatively simple way to test this hypothesis .
if negative design is present , then random mutagenesis of surface amino acids should on average lead to a protein that is easier to crystallize , and the two such mutagenesis studies in the literature do indeed find such a correlation.@xcite more useful , though , would be to identify the mechanisms used to achieve this negative design , since then this raises the possibility of developing rational strategies to overcome the negative design in order to make protein crystallization easier .
interestingly , bioinformatic analyses have provided a ` smoking gun ' for the potential role of surface lysine residues .
lysine is the most common surface amino acid , but yet is the most underrepresented at functional interfaces@xcite and at contacts between proteins in crystals.@xcite this of course begs the question : what is it doing there , if it is only reluctantly involved in interactions ?
one possible answer is that lysine helps to prevent unwanted interactions .
consistent with this suggestion , the group of zygmunt derewenda have shown that systematically mutating surface lysine residues , particularly to alanine,@xcite almost invariably enhances the crystallizability of a protein.@xcite the long lysine side - chain has substantial conformational entropy , which would be lost if constrained at a protein - protein interface .
thus , it has been suggested that lysine stabilizes the free protein in solution by providing a ` surface entropy shield',@xcite and replacing lysines with the compact alanine is expected to make crystallization thermodynamically more favourable .
one possible objection to our negative design hypothesis is that , given that proteins are irregular objects subject to dynamical fluctuations , it is not surprising that they are hard to crystallize .
however , further evidence that the low crystallizability of most proteins is not an intrinsic property , but one that arises due to selection , comes from instances when protein crystallization does occur _ in vivo _ , because it achieves some functional purpose .
nature has no problem getting proteins to crystallize , even in the ` dirty ' environment of the cell , when there is a positive selection pressure for this .
there are many fascinating examples of this _ in vivo _ protein crystallization , as catalogued in a recent review.@xcite such crystallization represents a rather neglected area of biological self - assembly , and one which is worthy of further study . here
we give just a few examples .
certain genera of viruses that infect insect larvae , coopt the cell to express large quantities of specific proteins during the late stages of infection . these then crystallize around the viral particles to provide a protective environment for the viruses after the death of the insect larvae , as illustrated in fig .
[ fig : invivo](b ) . on ingestion of the crystals by a new insect larva
, the protein crystals dissolve in the alkaline environment of the gut , releasing the virus to infect the new host.@xcite some examples of protein crystallization occurring in the cell .
( a ) a trichocyst attached to the outer membrane of _ paramecium_. ( b ) an encapsulated rod virus of the granulosis virus of _ plodia interpunctella_. ( c ) paracrystalline arrays of an iridescent virus in the epidermis of a larva of _ culicoides variipennis
sonorensis_. the scale bars in ( b ) and ( c ) correspond to 0.1@xmath1 m and 1@xmath1 m , respectively . reproduced with permission from refs .
( a ) and ( b ) and ( c ) .,width=321 ] a number of examples of crystalline proteins in the cell correspond to proteins stored in secretory granules .
[ fig : invivo](a ) illustrates one particularly specialized example from the protist ( a type of single - celled eukaryotic organism ) _
paramecium_.@xcite large numbers of these granules , which are called trichocysts , are attached to the outer membrane of this organism , and are assembled from three different families of closely related polypeptides , each localized to different regions of the trichocyst to form a structure that is of the order of 34@xmath1 m in size . in response to an external stimulus the trichocysts can be simultaneously extruded from the cell . in this process ,
the crystalline trichocysts expand in length by a factor of roughly eight .
the purpose seems to be defensive .
the explosive release of the trichocysts can push _ paramecium _ away from a potential predator , giving it a chance to escape.@xcite in both these examples , a high level of control is asserted over the crystallization process in order to determine the location , size and shape of the crystal .
the mechanical properties are also important for their function with the virus - encapsulating crystals required to be tough in order to provide a protective environment , and the trichocysts able to undergo an irreversible expansion triggered by the presence of calcium ions .
our final example is of the crystallization of virus particles themselves ( fig .
[ fig : invivo](c ) ) , and has perhaps more to do with an absence of a selection pressure to prevent crystallization ( why should it bother the virus ? ) than a functional purpose for the crystals .
the most dramatic examples come from iridoviruses , which have large icosahedral capsids .
this type of virus was first identified due to the iridescent colours imparted to insect larvae ( hence the name)@xcite due to the bragg scattering of visible light by the crystals.@xcite
there has been much work trying to understand the unusual dynamic properties of supercooled liquids as the glass transition approached.@xcite to do this from a simulation perspective , model systems have been developed that are robust glass - formers and show no tendency to crystallize .
such models also allow one to probe the structural determinants of a system s glass - forming ability .
a key notion is that of ` frustration',@xcite which is said to occur when the preferred local order is incompatible with global crystalline order . typically , the liquid structure reflects the preferred local order , and so when frustration is present , significant structural reorganization will be necessary to nucleate a crystalline phase . to apply these ideas
, one needs to be able to identify the preferred local order , but , because of the inherent disorder associated with a liquid , sometimes this can be hard to achieve directly from analysing liquid configurations .
an alternative approach is to instead look at the structures of isolated clusters , since such a cluster can adopt the preferred structure without having to mould itself to any surrounding medium .
this approach was first applied by charles frank to rationalize why small liquid metal droplets could be substantially supercooled.@xcite one of the potentials most commonly used by simulators is the lennard - jones potential .
however , the one - component lennard - jones fluid is not a good system to look at the behaviour of supercooled liquids , since crystallization into a close - packed solid occurs relatively easily . to generate a good glass - former , dzugutov modified this potential by adding a barrier in the potential at roughly @xmath2 times the position of the potential minimum ( fig . [ fig : dz](a ) ) ,
hence energetically disfavouring close - packed crystals because of their octahedral interstices.@xcite instead , polytetrahedral order,@xcite and local icosahedral order in particular , is preferred . as can be seen in fig .
[ fig : dz](a ) , the distances present in the 13-atom icosahedral cluster avoid sampling this bump in the potential .
( a ) the dzugutov ( dz ) and lennard - jones ( lj ) potentials compared .
the vertical lines show the positions of the pair distances in the 13-atom icosahedron and their heights are proportional to the numbers of pairs with that distance .
( b ) some of the particularly stable structures for dzugutov clusters .
each cluster is labelled by its size and point group.,width=321 ] even though there is a stable body - centred - cubic phase,@xcite the resulting dzugutov liquid can be easily supercooled .
interestingly , the dzugutov system is one of the only liquids in which an increase in a structural length scale has been detected on increased supercooling , in particular domains with local icosahedral order grow in size as the temperature is decreased.@xcite however , these domains are not compact , but are ramified in structure .
these results raise the questions : what are the origins of the non - compactness , and might this shape be a generic property of domains of enhanced local order within supercooled liquids ?
an analysis of the structures of isolated clusters is able to provide some answers .
this type of ordering is also clearly seen in the clusters , but an analysis of the energetics shows that it is a consequence of the unusual shape of the dzugutov potential .
therefore , this domain shape is unlikely to be universal .
the lowest - energy dzugutov clusters are all aggregates of interpenetrating and face - sharing icosahedra ( fig .
[ fig : dz](b)).@xcite as the cluster size increases first chains , then discs , then rings , and finally 3-dimensional porous networks of icosahedra are seen .
the energetic causes for this ordering are also clear .
the bump in the potential promotes local icosahedral order . however , icosahedral structures are inherently strained , e.g. the distance between adjacent atoms on the surface of the regurlar 13-atom icosahedron is 5% longer than that to the central atom ( fig .
[ fig : dz](a ) ) .
this strain , and the associated energetic penalty , grows rapidly for compact polyicosahedral structures .
the resulting non - compact icosahedral aggregates represent a compromise that maintains the local icosahedral coordination , but which does not involve excessive strains .
similar behaviour has also been seen for variants of the dzugutov potential.@xcite a second approach used to generate a good glass - former is to introduce two atom types .
partly this is because the formation of a compositionally ordered crystal can require the atoms to diffuse significantly further to find the correct environment than for the one - component case .
however , there is more to it than this , since the interactions between the particles and the composition also need to be tuned to reach those regions of the parameter space where crystallization is particularly difficult.@xcite the most commonly used model of this type is the kob - andersen binary lennard - jones mixture,@xcite i.e. the potential is @xmath3,\ ] ] where @xmath4 and @xmath5 are the atom types of atoms @xmath6 and @xmath7 , and @xmath8 and @xmath9 are the pair well depth and equilibrium pair separation , respectively , for the interaction between atoms @xmath6 and @xmath7 .
the kob - andersen parameters , @xmath10 , @xmath11 , @xmath12 , and @xmath13 , are non - additive and strongly favour mixing . at the canonical composition a@xmath14b the ground state
is a coexisting pure a face - centred - cubic crystal , and an ab cscl - type crystal,@xcite with low - lying layered a@xmath14b crystals also possible.@xcite however , crystallization has never been seen in a simulation .
( a ) the energies of the global minima for four series of kob - andersen blj clusters with different numbers of b atoms . to make particularly stable clusters stand out ,
the energies are measured with respect to @xmath15 , a fit to the energies of the clusters . for clarity , the curves
have also be displaced with respect to each other .
( b ) some of the particularly stable structures identified in ( a ) with all b atoms completely coordinated .
( c ) a particularly stable cluster at near equimolar composition.,width=321 ] here , we use isolated clusters to analyse the preferred coordination environments of a atoms around b atoms , since at the glass - forming composition b atoms are in a minority .
in particular , we locate the particular stable clusters with one , two , three and four b atoms to see how these preferred environments can organize into larger structures ( fig .
[ fig : blj ] ) . for coordination of a single b atom , clusters with 8 , 9 and 10
atoms have similar stability .
all three structures are based on a square antiprism of a atoms , but with possible capping atoms over the two square faces .
this stability of the square antiprism is also seen for a@xmath16b@xmath17 and a@xmath16b@xmath18 , where the additional capping atoms are now b atoms .
one way of combining such coordination environments is illustrated by a@xmath19b@xmath17 and a@xmath20b@xmath18 and involves the sharing of the square - faces of the square antiprisms to form linear aggregates .
similar columns of square anti - prisms are found in the al@xmath17cu crystal .
a@xmath21b@xmath18 illustrates another possibility and involves the sharing of two triangular faces of a@xmath19b@xmath17 with a bicapped square antiprism .
another possible coordination environment for the b atoms is the tricapped trigonal prism .
although this structure is not stable for a@xmath22b ( it can relax to a monocapped square antiprism by a single diamond - square process@xcite ) it becomes a common environment for larger clusters .
this is illustrated by the structures a@xmath23b@xmath18 ,
a@xmath24b@xmath14 and a@xmath25b@xmath14 , which have one , two and four tricapped trigonal prismatic environments , respectively , albeit with b atoms in some of the capping sites .
these results tie in well with studies of the local structure in liquid configurations , which have found that square antiprism , and trigonal prismatic environments increasingly dominate the local structure in the liquid as the temperature is decreased.@xcite so how do these results help to rationalize the system s ability to avoid crystallization .
firstly , these preferred local environments are not present in the lowest - energy crystal structures .
however , there are still a number of crystalline structures that are significantly lower in energy than the liquid that do involve these environments.@xcite secondly , the diversity of environments and ways that these can pack is likely to frustrate the formation of a uniform crystal .
in contrast to the a@xmath14b system , at equimolar compositions the kob - andersen blj model easily crystallizes into a cscl - type structure.@xcite again , isolated clusters can help to understand this behaviour .
examining a@xmath16b@xmath26 clusters , we found a@xmath16b@xmath27 to be particularly stable and to exhibit the bulk crystal structure ( fig .
[ fig : blj](c ) ) , hence showing the absence of frustration at this composition .
in the above models with isotropic interactions , the connection between the preferred local structure and the form and parameters of the potential can be quite subtle . to probe the connection between crystallizability and local structure further
, it would be useful to have a model where the local structure can be more directly controlled . to achieve this goal in a simple and flexible way , here we use ` patchy ' particles , where the particles are spherical and only interact strongly when the patches on adjacent particles are aligned , but we should note that other ways of using anisotropic interactions to control the local structure have recently been explored@xcite
. there has been much recent interest in such patchy models from the perspective of protein crystallization,@xcite self - assembly@xcite ( particularly into monodisperse clusters in a way similar to virus self - assembly@xcite ) and the dynamics of supercooled liquids,@xcite but it is still an area that is very much in its infancy with much to be discovered .
our interest is to probe how the anisotropy of the interparticle interactions can be used to control a system s crystallization and self - assembly behaviour , and , in particular , through its effect on the degree of frustration .
as well as being of fundamental interest , the hope is that the results will also be useful for understanding the crystallization behaviour of proteins and anisotropic colloids . the geometry of the interactions between the model particles . in this example , there are four patches regularly arranged on the discs , with their directions described by the patch vectors , @xmath28 .
patch 4 on particle @xmath6 interacts with patch 2 on particle @xmath7 because they are closest to the interparticle vector.,width=264 ] here , we model the patchy particles with a single - site potential , i.e. each particle is represented by just a single site , but which has both position and orientation . the potential consists of an isotropic repulsion , which is based on the lennard - jones potential @xmath29\ ] ] but where the attraction is modulated by an orientational dependent term , @xmath30 .
thus , the complete potential is @xmath31 where @xmath32 is the orientation of particle @xmath6 .
the patches are specified by a set of patch vectors , @xmath33 , as illustrated in fig .
[ fig : potential ] .
@xmath30 has the form .
@xmath34 where @xmath35 is the standard deviation of the gaussian , @xmath36 is the angle between the patch vector @xmath37 on particle @xmath6 and the interparticle vector @xmath38 , and @xmath39 is the patch that minimizes the magnitude of this angle .
hence , only the patches on each particle that are closest to the interparticle axis interact with each other , and the potential is continuous as a function of the orientations of the particles .
@xmath40 when two patches point directly at each other , but falls off as the patches deviate further from the perfect alignment .
one of the nice features of this single - site potential is that , given a set of patch vectors , the potential has only one parameter @xmath35 , which determines the widths of the patches . furthermore , as @xmath41 the isotropic lennard - jones potential is recovered .
hence , it is possible to systematically study the behaviour of the model as a function of the degree of anisotropy with the well - characterized lennard - jones model as one limit .
we first discuss the application of this model to crystallization in two dimensions to illustrate the effects of the geometrical arrangement of the patches , since visualization is easier than in three dimensions .
we choose to study particles with 4 , 5 or 6 patches arranged regularly on the surface of the disks .
particularly interesting will be the 5-patch system as the local five - fold symmetry of the patches is incommensurate with global crystalline order .
the two - dimensional lennard - jones reference system crystallizes into a close - packed crystal with a hexagonal arrangement of neighbours around each particle .
unsurprisingly , for particles with a regular hexagonal array of patches , the anisotropy reinforces this behaviour .
crystallization is easy ( fig .
[ fig:2dcool](c ) ) and the close - packed crystal is lowest in energy for any combination of pressure and @xmath35 . in the 4-patch system
there is competition between a low - density square crystal in which each patch points directly at a nearest neighbour , and a higher - density hexagonal crystal .
the latter is orientationally ordered with two opposite patches pointing at nearest neighbours , and the other two at next neighbours .
as the square crystal has more pairs of nearest - neighbour patches pointing directly at each other , it is energetically preferred when the patches are sufficiently narrow ( fig .
[ fig:2dphase](a ) ) .
however , because of its more open structure , it becomes destabilised relative to the hexagonal crystal as the pressure is increased .
crystallization to the square crystal occurs relatively easily in the appropriate region of the phase diagram ( fig .
[ fig:2dcool](a ) ) .
structural phase diagram showing how the ground - state structure in two dimensions depends on patch width and pressure ( in units of @xmath42 ) for particles with ( a ) four and ( b ) five regularly spaced patches .
the unit cells of the crystalline structures are depicted with thick lines , and the dashed lines show the periodic boundary conditions .
, width=321 ] for the 5-patch system the situation is more complex .
no crystal phases are possible where all the patches point directly at neighbouring particles . for example , on cooling this system in the region of the phase diagram where the patches prefer to be aligned , the resulting configuration has no overall crystalline order .
however , most of the particles have one of the two local environments highlighted in fig .
[ fig:2dcool](b ) . from these two motifs ,
two crystals can be constructed where every particle has an identical environment ( fig . [ fig:2dphase](b ) ) . in both crystals
each particle has five nearest neighbours , and the crystals can be thought of in terms of tilings of squares and equilateral triangles , where every vertex is surrounded by three triangles and two squares .
the nomenclature for these semi - regular tilings is ( @xmath43 ) and ( @xmath44 ) , and derives from the ordering of polygons around each vertex.@xcite however , we will use the notation ` @xmath35 ' and ` h ' by analogy to the frank - kasper phases of these names that can be envisaged as square - triangle tilings in two of their three dimensions.@xcite when the patches are sufficiently narrow , the h- and @xmath35-phases represent the best compromise between the five - fold symmetry of the particles and crystalline order . however , the patches can not point directly at each other in these crystals , and so the system is frustrated .
the mean deviation of the patches away from the nearest - neighbour interparticle vectors is smaller for the @xmath35-phase ( @xmath45 ) than for the h - phase ( @xmath46 ) , making the former lowest in energy for intermediate values of @xmath35 at low pressure .
however , at small @xmath35 , the energetic penalty associated with these deviations becomes significant , and causes the h - phase to distort so that three of the five patches can point directly at each other .
this distortion reduces the aspect ratio of the unit cell and the triangles deviate significantly from the ideal equilateral geometry .
no such similar distortion is possible for the @xmath35-phase , and so the h - phase becomes lowest in energy at small @xmath35 ( fig .
[ fig:2dphase](b ) ) . the h- and
@xmath35-phase represent only two of the possible tilings of squares and triangles .
these tilings need not be crystalline , and dodecagonal quasicrystalline packings are also possible .
indeed , these quasicrystals have been seen in both metallic alloys@xcite and macromolecular systems,@xcite under circumstances near to where crystalline square - triangle phases are stable .
therefore , a stable quasicrystalline state remains an intriguing possibility for the current system .
the more dense hexagonal crystal is most stable at higher pressure , and closer to the isotropic limit .
although positionally ordered , the crystal is orientationally disordered , because of the incompatibility of the five - fold symmetry of the particles and the six - fold symmetry of the crystal ( fig . [ fig:2dphase](b ) ) .
an interesting comparison to the current system , which has five - fold symmetry in the attractive interactions , is provided by a study of hard pentagons.@xcite although the shape of these particles does frustrate crystallization to some extent , the lattice formed by the particle centres in the only stable orientationally - ordered crystalline phase , is based on a slightly distorted hexagonal lattice . also relevant is a recent study of a two - dimensional system of particles interacting with a lennard - jones potential plus an anisotropic term that favours the formation of isolated five - fold rings of particles@xcite .
the presence of such order again frustrates crystallization , and , interestingly , on varying the strength of the anisotropic term in the potential , the dynamic properties of the liquid , such as the fragility , can be changed substantially . in our applications of this model to three dimensions
, we again look at how low - density crystalline phases can be stabilized by the patchy interactions .
the two systems that we consider are 4- and 6-patch particles with a regular tetrahedral and octahedral arrangement of the patches , respectively .
for these systems , the stable crystalline phases at sufficiently low pressure and @xmath35 would be expected to be a diamond and a simple cubic lattice , respectively , with each patch directly point at a neighbouring particle .
these have maximum packing fractions of 34% and 52% compared to 74% for the face - centred - cubic crystal favoured by the isotropic lennard - jones potential .
one might have expected these two systems to behave quite similarly in the regime where the open structures are preferred , but in fact their crystallization behaviour is quite different .
the 6-patch system is able to crystallize easily , leading to a step - like decrease in the energy on crystallization ( fig .
[ fig:3d](a ) ) , sometimes even giving a perfect defect - free crystal .
by contrast for the 4-patch system at best only partial crystallization is observed , and on ordering the energy does not exhibit a step , but instead changes continuously . as in section [ sect : glass ] , examining the structure of isolated clusters can give us some clues to the physical origins of this behaviour , because the clusters provide a picture of the preferred local order for the system . for both systems , when the patches are sufficiently narrow , as expected , the lowest - energy clusters exhibit open structures where the maximum coordination number is equal to the number of patches . for the octahedral particles , for @xmath47 most of the clusters are cuboidal nanocrystals with the simple cubic structure ( e.g. the @xmath48 cube in fig . [ fig:3d](c ) ) .
the system is unfrustrated and it is unsurprising that crystallization is relatively easy . by contrast , the global minima of the tetrahedral system for @xmath49 do not exhibit the structure of the stable crystalline phase , but instead are based on dodecahedral cages ( the 20-particle dodecahedron is shown in fig .
[ fig:3d](b ) ) . only for @xmath50
are clusters with the diamond structures lowest in energy , e.g. the 26-particle structure in fig .
[ fig:3d](b ) . in diamond
the particles form six - fold rings , but which are puckered , whereas the dodecahedra are characterized by planar five - fold rings , where the 108@xmath51 bond angles are very close to the 109.57@xmath51 angles between the tetrahedral patches .
the effect of this difference is small until the patches become very narrow , and so for intermediate values of @xmath35 , the dodecahedral clusters are preferred because they have a greater number of bonds , e.g.for the 20-atom dodecahedron there is only one unused patch per particle , whereas the particles on the surface of diamond clusters often have two unused patches .
thus , the results for these clusters suggest that the tendency of the liquid to form structures involving five - fold rings is one of the reason underlying the much greater difficulty of crystallization in the tetrahedral system . ( a ) the caloric curves for cooling the 4-patch tetrahedral particles and the 6-patch octahedral particles from high temperature .
the number of particles is 512 , the patch width @xmath35 is 0.3 radians , and the pressure is @xmath52 .
the resulting simple cubic crystal for the 6-patch system is also illustrated ( the dots represent the corners of the simulation box ) .
( b ) and ( c ) some of the global minima of isolated clusters for the 4- and 6-patch systems , respectively
. each cluster is labelled by the number of atoms and point group .
, width=321 ] furthermore , as well as this competition between five- and six - fold rings , in this model the diamond and hexagonal diamond ( the one - component equivalent of wurtzite ) structures are practically degenerate .
the structural difference between these two crystals is that all the six - fold rings in diamond have a form analogous to the chair isomer of cyclohexane , whereas in hexagonal diamond , some of these rings are analogous to the boat form of cyclohexane .
again , this further variety of local structural forms is not going to aid crystallization .
these results are consistent with the work of zhang _
et al . _ who attempted to crystallize diamond using similar patchy particles with the same tetrahedral patch geometry.@xcite crystallization to the diamond structure only reliably occurred when a seed crystal was introduced , or when an additional term in the potential was added that favoured staggered over eclipsed torsional configurations , and hence disfavoured hexagonal diamond .
this work on model anisotropic particles is particularly timely given the rapid recent advances in synthesizing anisotropic colloids,@xcite and because one target for research in this field is to get particles with tetrahedral symmetry to assemble into a diamond - like crystal.@xcite our results indicate that the crystallization of such tetrahedral colloids might not be so straightforward , because of the potentially frustrating effects of the variety of local structures possible in the liquid phase .
however , it is not clear how an additional potential term similar to that used by zhang _
_ could be introduced into the intercolloidal potential to alleviate this . instead
, an alternative strategy would be to use a binary system of oppositely - charged tetrahedral colloids , as this would penalize the formation of rings with an odd number of particles .
in particular , this change would reduce the tendency for pentagons to form , hence removing some of the frustration and making crystallization more likely . an interesting comparison to the present results
are provided by a recent study that took the stillinger - weber silicon potential@xcite and varied the strength of the anisotropic 3-body term , finding that it had a significant effect on the system s glass - forming ability@xcite . for the original silicon potential crystallization into a diamond structure occurred relatively easily , but as the strength of the 3-body term is decreased the system becomes a good glass - former in the region of parameter space where there is a crossover in the stability of the diamond and body - centred - cubic crystals , and it has been suggested that this is partially due to the structural dissimilarity between the liquid and the possible crystals .
similarly , it would be interesting to look at how varying the patch width affected the dynamics of crystallization for our systems .
through the examples of crystallization ( and its absence ) considered in this paper , we hope to have shown the subtle interplay between the interparticle potential , the preferred local structure and the kinetics of crystallization .
in particular , if the interactions favour a local structure that is incompatible with the global crystalline order , or almost equally favour a variety of different local environments , then crystallization is likely to be frustrated .
furthermore , we have provided further empirical support that examining the structures of isolated clusters , as first done by frank,@xcite can potentially provide a clearer picture of the preferred local order , and hence the possible presence of frustration .
an important question concerning the type of structural approach advocated here is how it relates to other approaches to understanding the kinetics of crystallization , such as classical nucleation theory . in the latter ,
one of the key parameters in determining the ease of nucleation is the surface free energy associated with the crystal - liquid interface .
the connection to the current approach is that this interfacial free energy is particularly sensitive to the degree of structural dissimilarity between the crystal and the liquid , i.e. the greater the structurally dissimilarity , the greater the interfacial free energy , and hence the harder crystallization becomes . in such instances where the liquid and crystal are structurally very different
, one way of circumventing the large free - energy barriers to direct nucleation of the crystal is to undergo a two - step nucleation process , i.e. first nucleate a metastable phase , and then nucleate the final crystal form from within that phase@xcite .
for example , for the stillinger - weber silicon potential mentioned earlier , crystallization to the diamond structure only occurs at temperatures below that for a liquid - liquid phase transition , which leads to the nucleation of a lower - density liquid phase that more closely resembles the solid@xcite .
the patchy models introduced in section [ sect : patchy ] are particular useful for studying crystallization , since the geometry of the patches allows the system s structural propensities to be directly controlled in a simple and flexible manner .
one of our original intentions for these models was to also use them to illuminate aspects of the crystallization of proteins .
therefore , a key question is how relevant these models are to real proteins .
clearly , the interactions between proteins are strongly anisotropic in character , i.e. in order for two proteins ( in their native state ) to come together , the regions of their respective surfaces that come into contact , the `` patches '' if you like , must be correctly aligned and oriented .
furthermore , the patchy models , like proteins , are also able to form relatively low - density crystals
. however , actual interprotein interactions are much more complex
proteins are anisotropic in shape , their surfaces are very heterogeneous , and the interactions can depend sensitively on solution conditions , such as ph and the concentrations of other ions.@xcite thus , how much of this complexity one needs to capture in order to reproduce or gain insight into the crystallization of proteins is an open , but important , question .
the examples described in section [ sect : pxtal ] illustrate the exquisite control that biological systems are able to exert over the interaction properties of proteins , in particular their ability to avoid native state aggregation or crystallization in the dense cellular environment , and the ability of particular proteins to form into complex crystalline assemblies .
there are some indications of how this control is achieved , e.g. the role of lysine residues in preventing unwanted interactions , but there is much still to be learnt .
such knowledge will be particularly important if rational methods to make protein crystallization _ in vitro _ easier are to be developed that aim to overcome the evolutionary selection of the surface properties of proteins to prevent native state aggregation . in some ways , our patchy particles are probably somewhat closer to some of the anisotropic colloids now beginning to be generated.@xcite therefore , it will be particularly interesting , once it becomes possible to generate such colloids in sufficient number and quality , and with their surfaces appropriately functionalized , to be able to probe their phase behaviour and phase transformation kinetics .
hopefully , some of the insights gained from our patchy models will help to guide the experimentalists in their quest to create crystals with useful photonic properties .
the authors are grateful to the royal society , the ramn areces foundation and the engineering and physical sciences research council for financial support . | the ability to control the crystallization behaviour ( including its absence ) of particles , be they biomolecules such as globular proteins , inorganic colloids , nanoparticles , or metal atoms in an alloy , is of both fundamental and technological importance .
much can be learnt from the exquisite control that biological systems exert over the behaviour of proteins , where protein crystallization and aggregation are generally suppressed , but where in particular instances complex crystalline assemblies can be formed that have a functional purpose .
we also explore the insights that can be obtained from computational modelling , focussing on the subtle interplay between the interparticle interactions , the preferred local order and the resulting crystallization kinetics .
in particular , we highlight the role played by `` frustration '' , where there is an incompatibility between the preferred local order and the global crystalline order , using examples from atomic glass formers and model anisotropic particles . |
cp@xmath0 models are a class of models in which the fundamental field is a complex @xmath1-component unit vector ( more precisely , an element of the complex projective manifold cp@xmath0 ) , and which are characterized by a global u(@xmath1 ) symmetry and a local u(1 ) gauge symmetry .
they emerge as effective theories of su(@xmath1 ) quantum antiferromagnets @xcite and of scalar electrodynamics with a compact u(1 ) gauge group .
the simplest three - dimensional ( 3d ) cp@xmath0 lattice model is defined by the hamiltonian @xmath5 where the sum is over the nearest - neighbor sites of a cubic lattice , @xmath6 are @xmath1-component complex vectors satisfying @xmath7 .
the model is ferromagnetic for @xmath8 , antiferromagnetic for @xmath9 .
the cp@xmath10 model can be mapped onto the o(3)-symmetric heisenberg model .
indeed , if one defines o(3 ) spins @xmath11 , where @xmath12 are the pauli matrices , one can rewrite the cp@xmath10 hamiltonian as that of the usual 3-vector heisenberg model . as a consequence , the critical properties can be straightforwardly derived by using the wealth of results available for the heisenberg model , see , e.g. , refs .
@xcite . on the other hand ,
several aspects of the phase behavior of cp@xmath0 models with @xmath13 remain unclear and worth being further investigated .
the critical behavior of these models can be investigated by constructing an effective landau - ginzburg - wilson ( lgw ) theory .
this approach requires the identification of the order parameter associated with the critical modes . assuming that the critical behavior is essentially driven by gauge - invariant modes , a plausible choice for ferromagnetic ( @xmath8 ) systems is the gauge - invariant site variable @xcite @xmath14 which is a hermitian and traceless @xmath15 matrix . in the corresponding lgw theory ,
the fundamental field is therefore the most general traceless hermitian matrix @xmath16 , which one can imagine being defined as the average of @xmath17 over a large but finite lattice domain .
the hamiltonian is the most general fourth - order polynomial in @xmath18 consistent with the u(@xmath1 ) symmetry : @xmath19 for @xmath20 one recovers the o(3)-symmetric lgw theory because the cubic term vanishes and the two quartic terms are equivalent @xcite , consistently with the equivalence between the cp@xmath10 and the heisenberg model . because of the presence of the cubic term , on the basis of mean - field arguments , one expects the system to undergo a first - order transition for any @xmath13 , unless the hamiltonian parameters are tuned so that @xmath21 in the effective model .
this prediction is , however , contradicted by recent numerical studies @xcite , which find evidence of continuous transitions in models that are expected to be in the same universality class as that of the 3d cp@xmath3 model . in particular ,
a numerical study of 3d loop models @xcite provided the estimate @xmath22 for the correlation - length critical exponent .
these results imply the existence of a 3d cp@xmath3 universality class characterized by a u(3 ) global symmetry and u(1 ) gauge invariance , with a corresponding fixed point ( fp ) that can not be determined in perturbation theory at fixed @xmath1 . in order to access this fp , the authors of ref .
@xcite proposed a double expansion around @xmath20 ( where the cubic term vanishes ) and @xmath23 , arguing that a continuous transition may be possible for values of @xmath1 sufficiently close to @xmath20 . for larger values of @xmath1 ,
i.e. @xmath24 , the numerical analyses @xcite show evidence of first - order transitions . in the large-@xmath1 limit
, the quantum field theory corresponding to the ferromagnetic cp@xmath0 model gives rise to an effective abelian higgs model and landau - ginzburg ( lg ) theory of superconductivity @xcite , whose renormalization - group ( rg ) flow presents a stable fp for a sufficiently large number of components .
thus continuous transitions are again possible in cp@xmath0 models at large @xmath1 , sharing the same universal critical behaviors of the lg theory of superconductivity @xcite .
these results are again in contrast with the conclusions obtained from the lgw theory ( [ hlg ] ) and suggest that , at least for large values of @xmath1 , critical modes are not exclusively associated with the gauge - invariant order parameter @xmath25 , cf .
( [ qdef ] ) . in this paper
we investigate the critical behavior of antiferromagnetic cp@xmath0 ( acp@xmath0 ) models , such as those described by the hamiltonian ( [ hcpn ] ) with @xmath26 , on a cubic lattice ( we expect a similar behavior on any bipartite lattice ) . for @xmath20 ,
they undergo a critical transition in the same universality class as that of the ferromagnetic cp@xmath10 model . indeed
, the acp@xmath10 model is equivalent to the antiferromagnetic heisenberg model , which in turn can be mapped onto the ferromagnetic one by performing the transformation @xmath27 , where @xmath28 . therefore , the staggered variables @xmath29 or @xmath30 have the same critical behavior as @xmath31 or @xmath25 in the ferromagnetic model .
however , for @xmath32 the behavior of acp@xmath0 models differs from that of ferromagnetic cp@xmath0 models , as we shall show . keeping the assumption that the critical modes can be represented by local staggered gauge - invariant variables ,
we show that the lgw hamiltonian describing the behavior of the critical modes in the acp@xmath0 models is the one given in eq .
( [ hlg ] ) , without the cubic term , that is with @xmath21 .
indeed , the staggered nature of the order parameter gives rise to a symmetry @xmath33 , which prevents the presence of odd terms in @xmath18 , such as the cubic term .
this fact greatly simplifies the rg analysis of the theory .
in particular , it allows us to predict that the critical behavior of the acp@xmath3 model belongs to the universality class of the o(8 ) vector model , with a dynamical enlargement of the symmetry at the critical point .
correspondingly , we predict @xmath34 and @xmath35 for the acp@xmath3 model , which differ from those found in refs .
@xcite for the ferromagnetic cp@xmath3 universality class .
to validate this prediction , we perform monte carlo ( mc ) simulations of the lattice acp@xmath3 and o(8 ) vector models .
the critical exponents and finite - size scaling ( fss ) functions turn out to be the same in the two models , in agreement with the rg argument . finally , we present a general rg study of the lgw theory ( [ hlg ] ) without cubic term .
we compute high - order field - theoretical ( ft ) perturbative series in two different schemes .
the rg analysis does not provide evidence of the existence of stable fp for @xmath24 .
we mention that a critical - point symmetry enlargement analogous to that of acp@xmath3 models occurs in the antiferromagnetic rp@xmath36 model @xcite .
these systems are similar to those considered here : their hamiltonian is also given by eq .
( [ hcpn ] ) , but the site variable is a _ real _ unit vector .
in the rp@xmath36 case , an analogous rg argument based on the corresponding lgw theory shows that the model should have a critical behavior in the universality class of the o(5 ) vector model .
the paper is organized as follows . in sec .
[ lgw ] we construct the lgw theory which is expected to describe the critical modes at continuous transitions of acp@xmath0 models , assuming a staggered gauge - invariant order parameter .
[ acp2 ] is devoted to a numerical study of the acp@xmath3 model .
we show that its continuous transition is in the same universality class as that of the o(8 ) vector model . in sec .
[ hopa ] we study the rg flow relevant for models with more components , i.e. , for @xmath24 , computing and analyzing high - order ft perturbative series for the corresponding lgw theories . finally in sec .
[ conclu ] we summarize our main results and draw some conclusions .
some details are reported in the appendices .
we now derive the lgw theory for the critical modes at the transition in acp@xmath0 models , which is completely specified by the symmetry of the model , the nature of the order parameter , and the symmetry - breaking pattern . in the case at hand ,
the model has a global u(@xmath1 ) symmetry and a local u(1 ) gauge invariance .
we assume that the critical modes are effectively represented by local gauge - invariant variables , such as ( [ qdef ] ) . in the case of antiferromagnetic interactions ( @xmath26 ) ,
the minimum of the hamiltonian ( [ hcpn ] ) is locally realized by taking @xmath37 for any pair of nearest - neighbor sites . in order to construct the lgw hamiltonian
, we should identify the order parameter of the transition . at variance with the ferromagnetic case , in the acp@xmath0 model we should take into account the explicit breaking of translational invariance in the low - temperature phase . to clarify the issue ,
let us consider the antiferromagnetic o(@xmath38 ) vector model with hamiltonian @xmath39 . in this case
the order parameter is @xmath40 , where @xmath41 is the parity of the site @xmath28 defined by @xmath42 . under translations of one site
, we find @xmath43 , hence this parameter allows us to probe the breaking of translational invariance . in the acp@xmath0 model
the natural field variable is the combination @xmath44 , which is invariant under the local u(1 ) gauge transformations of the model .
hence , we define the order parameter @xmath45 it is immediate to verify that @xmath46 is hermitian and traceless [ this follows from the presence of @xmath41 ] , that it changes sign under translations of one site which exchange the two sublattices , and that it coincides with the o(3 ) order parameter for @xmath48 .
then , as usual , in order to construct the lgw model , we replace @xmath46 with a local variable @xmath49 as fundamental variable ( essentially , one may imagine that @xmath49 is defined as @xmath46 , but now the summation extends only over a large , but finite , cubic sublattice ) . then , we write down the most general fourth - order polynomial that is invariant under u(@xmath1 ) transformations and under the @xmath50 transformation @xmath51 , a consequence of the translation invariance of the original theory .
we obtain @xmath52 the original u(@xmath1 ) symmetry corresponds to the symmetry @xmath53 where @xmath54 .
the order parameter @xmath49 is a hermitean traceless matrix as the variable @xmath18 introduced in the ferromagnetic case .
however , because of the presence of the symmetry @xmath55 , which is a specific feature of the antiferromagnetic model , the lgw hamiltonian ( [ ahlg ] ) does not present a cubic term , which instead appears in the @xmath56 hamiltonian ( [ hlg ] ) corresponding to the ferromagnetic case .
note that the model is not only characterized by the symmetry group , but also by the nature of the order parameter .
there are indeed other models with u(@xmath1 ) symmetry ( see , e.g. , ref .
@xcite for an example ) , which , however , have a different order parameter and different symmetry - breaking patterns , leading to different universality classes . the stability domain of @xmath57 can be determined by studying the asymptotic behavior of the potential @xmath58 this analysis can be easily performed by noting that @xmath59 only depends on the @xmath1 real eigenvalues @xmath60 of the hermitian matrix @xmath49 , which satisfy the condition @xmath61 .
we find that the theory is stable if @xmath62 and if @xmath63 where @xmath64 physical systems corresponding to the effective theory ( [ ahlg ] ) with @xmath65 that do not satisfy these constraints are expected to undergo a first - order phase transition .
the analysis of the minima of the potential @xmath59 for @xmath66 gives us information on the symmetry - breaking patterns .
for @xmath67 , the absolute minimum of @xmath59 is realized by configurations with @xmath68 and @xmath69 where @xmath70 indicates the @xmath71 identity matrix .
this gives rise to the symmetry - breaking pattern @xmath72 on the other hand , for @xmath73 and even @xmath1 the minumum is realized by @xmath74 implying the symmetry - breaking pattern @xmath75 for @xmath73 and odd values of @xmath1 , we have instead @xmath76 @xmath77 , so that @xmath78 note that for @xmath2 , the symmetry breaking patterns ( [ vnpat ] ) and ( [ vppat2 ] ) are equivalent , hence the sign of @xmath79 does not play any role .
an important remark is in order .
the derivation of the lgw hamiltonian ( [ ahlg ] ) is based on the assumption that the order parameter is the staggered and traceless hermitian matrix ( [ biab ] ) .
this assumption can be checked for @xmath80 . as we show in app .
[ appmin ] , the minimum - energy configurations of hamiltonian ( [ hcpn ] ) for @xmath26 have a very simple structure .
modulo a global u(3 ) transformation , one can take @xmath81 on one sublattice , and @xmath82 on the other one .
hence , a zero - temperature configuration corresponds to @xmath83 therefore , @xmath46 is nonvanishing in the low - temperature phase and represents the correct order parameter .
the symmetry - breaking pattern is therefore that given in eq .
( [ vnpat ] ) or , equivalently , eq . ( [ vppat2 ] ) . for @xmath24
, we have not been able to identify ordered zero - temperature configurations that are translation invariant at least on one sublattice , hence , we have not been able to check that @xmath46 , as defined in eq .
( [ biab ] ) , is nonvanishing in the low - temperature phase , hence that it can be taken as the order parameter . in the following , we make the working hypothesis that this is the case , determining what this assumption implies for the nature of the transitions in the acp@xmath0 models . for @xmath2 the lgw theory ( [ ahlg ] ) simplifies .
indeed , one can easily prove that @xmath84 for any @xmath85 traceless hermitean matrix .
then , let us define an 8-component real vector field @xmath86 as follows : @xmath87 where @xmath88 . in terms of the field
@xmath86 we have @xmath89 the , we can rewrite @xmath57 as @xmath90 where @xmath91 , proving that the model is equivalent to the o(8 ) vector theory .
since the o(8 ) vector theory has a stable wilson - fisher fp , the above correspondence allows us to predict that the critical behavior of the acp@xmath3 model must share the same universal features as the 3d o(8 ) vector model .
note that the enlargement of the symmetry to o(8 ) is a feature of the lgw theory ( [ ahlg ] ) , i.e. , of the expansion up to fourth powers of @xmath49 .
indeed , one can easily check that the sixth - order terms allowed by the u(3 ) symmetry , such as @xmath92 , do not share the o(8 ) symmetry . since these terms are rg irrelevant at the o(8 ) fp , the contribution of the terms breaking the o(8 ) symmetry is suppressed at the critical point . in this sense
we have a dynamic enlargement of the symmetry to o(8 ) at the critical point .
of course , the fact that only one quartic term is independent for @xmath2 holds also for the hamiltonian ( [ hlg ] ) corresponding to the ferromagnetic cp@xmath3 model .
thus , it has an o(8 ) fp , which is unstable due to the presence of the cubic term .
this cubic term gives rise to a spin-3 perturbation of the o(8 ) model @xcite .
its rg dimension @xmath93 can be estimated using the results of ref .
@xcite for the crossover exponent @xmath94 , i.e. @xmath95 . the crossover exponent @xmath94 at the o(8 ) fp was estimated in ref .
@xcite , obtaining @xmath96 in the fixed - dimension massive - zero - momentum scheme , and @xmath97 in the @xmath98 expansion . since @xmath99
( see sec .
[ mco8 ] ) , we obtain @xmath100 . as @xmath101 , the o(8 )
fp is unstable in the presence of the cubic term , and , therefore , the o(8 ) fp can not be the stable fp in the case of ferromagnetic interactions .
this is confirmed by the results of refs .
the critical exponents , for instance @xmath22 , significantly differ from those of the o(8 ) universality class .
it is worth mentioning that an analogous enlargement of the symmetry at the critical point is also observed @xcite in the antiferromagnetic rp@xmath36 model .
the hamiltonian of rp@xmath0 models is analogous to that of cp@xmath0 models it is given by eq .
( [ hcpn ] ) , with real @xmath1-component spins @xmath102 replacing the complex vectors @xmath6 .
the order parameter should be a symmetric and traceless @xmath15 matrix @xmath103 , analogous to the matrix @xmath46 defined in eq .
( [ biab ] ) .
the corresponding lgw hamiltonian is given in eq .
( [ ahlg ] ) , with @xmath49 replaced by @xmath103 @xcite . for @xmath2 , the two quartic terms are proportional , and one obtains the @xmath56 theory of the o(5 ) vector model ( again high - order terms break the o(5 ) symmetry , but since they are irrelevant , the o(5 ) symmetry breaking is suppressed in the critical limit ) .
the numerical results of refs .
@xcite for the antiferromagnetic rp@xmath3 model confirm this prediction . to identify the nature of a critical transition for @xmath24 ,
if it exists , we must determine the fixed points of the rg flow of the lgw theory ( [ ahlg ] ) .
the absence of a stable fp implies the absence of continuous transitions .
such an analysis is quite complex , due to the presence of two quartic terms .
we will perform it in sec .
we now check the predictions of the previous section for the critical behavior of the 3d acp@xmath3 model , confirming that it undergoes a continuous transition in the o(8 ) universality class . in order to study the critical behavior of the acp@xmath3 lattice model ( [ hcpn ] ) with @xmath104 , we perform monte carlo simulations of cubic systems of linear size @xmath105 with periodic boundary conditions . because of the antiferromagnetic nature of the model we take @xmath105 even . we use a standard metropolis algorithm , hence we are only able to obtain reliable results up to @xmath106 . we use a simple updating algorithm .
if @xmath107 is a six - component vector , the update consists in proposing the new vector @xmath108 , where @xmath109 is a random o(2 ) matrix acting on two randomly chosen components of @xmath110 . in our mc simulations
we compute correlations of the gauge invariant operator @xmath111 its two - point correlation function is defined as @xmath112 due to the staggered nature of the ordered parameter , we should distinguish correlations between points belonging to the same sublattice and points belonging to different sublattices .
here we define the susceptibility and the correlation length by summing only over points with the same parity : @xmath113 where @xmath114 runs over all even points , @xmath115 is the fourier transform of @xmath116 over the even sublattice , @xmath117 , and @xmath118 . finally , we consider the binder parameter @xmath119 ^ 2 \rangle \over \langle \sum_{{x}\ ; { \rm even } } { \rm tr}\ , p_{0}^\dagger p_{x } \rangle^2 } .
\label{binderdef}\ ] ] in the fss limit any rg invariant quantity @xmath109 , such as @xmath120 and @xmath121 , is expected to behave as @xmath122 where @xmath123 and @xmath124 is a universal function apart from a trivial normalization of the argument . in particular , the quantity @xmath125 is universal within the given universality class .
the approach to the asymptotic behavior is controlled by the universal exponent @xmath126 , which is associated with the leading irrelevant rg operator .
around @xmath127 one may expand @xmath124 and @xmath128 in powers of the scaling variable @xmath129 , obtaining @xmath130 the exponent @xmath131 is determined by analyzing the fss behavior of the susceptibility @xmath132 .
\label{chisca}\ ] ] we present a fss analysis of the numerical data of the acp@xmath3 lattice model , up to @xmath133 . in fig .
[ rxi ] we show @xmath120 for several values of @xmath105 .
data have a crossing point , providing evidence of a transition in the interval @xmath134 .
an analogous behavior is shown by the binder parameter @xmath121 .
for the acp@xmath3 lattice model for different lattice sizes @xmath105 .
they show a crossing point at @xmath135 .
the dotted lines are drawn to guide the eye .
the vertical dashed line corresponds to our best estimate ( [ betacacp ] ) of @xmath127 obtained by a fss analysis of the available data assuming the transition to belong to the o(8 ) universality class . ] to determine whether the transition is continuous or of first order , we should estimate the effective exponent @xmath136 that gives the slope of the data at the critical point . at a first - order transition
we expect @xmath137 @xcite , while @xmath138 at continuous transitions .
the numerical data , including those of the specific heat , definitely exclude a first - order transition .
indeed , the increase of @xmath139 at the crossing point is much slower than @xmath140 .
therefore , we conclude that the acp@xmath3 model has a continuous transition . for the o(8 )
vector model .
plot of @xmath141 versus @xmath142 , setting @xmath143 and @xmath144 .
the dashed horizontal line corresponds to @xmath145 . ] to show that the acp@xmath3 lattice model belongs to the o(8 ) universality class , we show that critical exponents and fss curves are the same in the two models .
thus , we begin by computing these quantities in the o(8 ) spin model defined by the hamiltonian @xmath146 where the spin variable @xmath102 is an 8-component unit vector .
we consider cubic systems of linear size @xmath105 with periodic boundary conditions .
we consider the two - point function @xmath147 and compute the corresponding susceptibility and second - moment correlation length .
they are defined as in eqs .
( [ chisusc ] ) and ( [ xidefpb ] ) , but now we sum over all lattice points , as the model is ferromagnetic .
moreover , we consider the binder parameter @xmath121 defined as in eq .
( [ binderdef ] ) : we replace @xmath148 with @xmath149 and sum over all lattice points .
we perform simulations on lattices of size up to @xmath150 ( we use a cluster algorithm ) and estimate @xmath141 , @xmath121 , and @xmath151 . to compute the critical exponents we fit the data to eqs .
( [ rsca ] ) and ( [ chisca ] ) . in the analysis
we take into account the scaling corrections of order @xmath152 , fixing @xmath153 , as predicted by the ft perturbative analyses discussed below .
corrections turn out to be small , hence the analyses of the mc data are unable to provide a more accurate estimate of @xmath154
. fits of @xmath141 to eq .
( [ fitr ] ) ( we take the first terms in the expansions ) give @xmath155 , @xmath156 and the universal critical value @xmath157 .
the quoted uncertainty includes the statistical error and the variation of the estimates as @xmath154 varies between 0.75 and 0.85 , an interval that is larger than that obtained in the ft analyses reported below . in fig .
[ rxiscao8 ] we show @xmath141 versus @xmath142 using the above - reported estimates .
data collapse onto a single curve , confirming the accuracy of the estimates .
consistent results are obtained from the analysis of the binder parameter , which also gives @xmath158 .
the values of the critical exponents can be compared with previous results .
field theory gives @xmath159 and @xmath160 @xcite , while the analysis of strong - coupling expansions gives @xmath161 @xcite ( the two estimates are obtained by means of two different resummation methods ) . within errors ,
they agree with the estimates ( [ o8critexp ] ) .
we have repeated the analysis of the available six - loop series within the massive zero - momentum renormalization scheme @xcite , using the conformal mapping method that exploits the known large - order behavior of the perturbative expansions @xcite .
we obtain @xmath162 , @xmath163 , and @xmath164 , where the errors are related to the change of the estimates with respect to a ( reasonable ) variation of the parameters entering the resummation procedure .
they are substantially consistent with our favorite mc estimates ( [ o8critexp ] ) , although one may suspect that errors are slightly underestimated .
the ft analysis also provides the estimate of @xmath154 that we used ( note that , to be on the safe side , we allowed for a much larger uncertainty in the mc analysis ) .
if the transitions in the acp@xmath3 and o(8 ) models belong to the same universality class , critical exponents and fss curves @xmath124 for rg invariant quantities ( apart from a trivial rescaling of the variable @xmath129 ) should be the same .
this is what we check below .
a first unbiased universality check which does not need an estimate of the critical point @xmath127 , is obtained by plotting @xmath141 versus @xmath121 . indeed , since both @xmath141 and @xmath121 satisfy eq .
( [ rsca ] ) , we must have @xmath165 where @xmath166 is a universal function . in fig . [ rxibi ]
we compare the results for the two models : they appear to approach the same asymptotic curve with increasing the lattice size .
scaling corrections are consistent with the expected @xmath152 behavior with @xmath153 .
they are much smaller for @xmath167 than for @xmath168 . vs @xmath121 .
acp@xmath3 and o(8 ) results approach the same asymptotic curve .
the dotted horizontal and vertical lines correspond to the values @xmath169 and @xmath170 estimated in the o(8 ) vector model . ]
the dependence of the data on the inverse temperature @xmath171 around the crossing point is consistent with the o(8 ) results .
fits of the data around the crossing point to the first few terms of the expansions appearing in eq .
( [ fitr ] ) provide an accurate estimate of the critical point , @xmath172 in fig . [ chirxisca ] we show @xmath173 and @xmath141 versus @xmath129 , using the estimate ( [ betacacp ] ) of @xmath127 and the o(8 ) estimates ( [ o8critexp ] ) of the critical exponents . the data approach asymptotic scaling curves .
scaling corrections are larger for @xmath141 , but definitely compatible with the expected @xmath152 behavior .
note also ( not shown ) that the scaling curves of @xmath141 for the acp@xmath3 and o(8 ) nicely match after a trivial rescaling of the scaling variable @xmath142 : if we define @xmath174 for the acp@xmath3 and @xmath175 for the o(8 ) model , all data fall on the same curve when plotted versus @xmath176 . finally , we consider the specific heat @xmath177 at @xmath127 . we expect @xmath178 with @xmath179 .
the mc data are consistent with this behavior . in conclusion
the numerical analysis of the acp@xmath3 lattice model provides a robust evidence that its continuous transition belongs to the universality class of the o(8 ) vector model , as predicted by the rg arguments of sec .
[ lgw ] . and @xmath141 vs @xmath142 with @xmath180 and the critical exponents of the o(8 ) universality class , cf .
( [ o8critexp ] ) . in the top figure
the dashed horizontal segment indicates the o(8 ) universal value @xmath145 , which is clearly approached by the data of the acp@xmath3 lattice model with increasing @xmath105.,title="fig : " ] and @xmath141 vs @xmath142 with @xmath180 and the critical exponents of the o(8 ) universality class , cf .
( [ o8critexp ] ) . in the top figure
the dashed horizontal segment indicates the o(8 ) universal value @xmath145 , which is clearly approached by the data of the acp@xmath3 lattice model with increasing @xmath105.,title="fig : " ]
in this section we present a ft study of the rg flow of the lgw theory ( [ ahlg ] ) with @xmath24 , i.e. the most general @xmath56 theory with traceless hermitian @xmath15 matrix fields and parity symmetry . the critical behavior at a continuous transition is controlled by the fps of the rg flow , which are determined by the common zeroes of the @xmath171 functions associated with the quartic parameters .
the presence of a stable fp controls the universal features of the critical behavior in the case of continuous transition .
the absence of a stable fp implies the absence of a corresponding universality class , hence a transition characterized by the same symmetry - breaking pattern must be of first order .
we compute the @xmath171 functions of the quartic couplings in the @xmath181 renormalization scheme @xcite , which uses the dimensional regularization around four dimensions , and the modified minimal - subtraction prescription .
thus the rg functions are obtained from the divergences for @xmath182 appearing in the perturbative expansion of the correlation functions of the critical massless theory .
the procedure is straightforward , see , e.g. , ref .
the renormalized couplings are defined from the irreducible four - point correlation function , and the @xmath181 @xmath171 functions are @xmath183 where @xmath184 is the renormalization energy scale of the @xmath185 scheme . here ,
@xmath186 and @xmath187 are the renormalized couplings corresponding to @xmath188 , @xmath79 , defined so that @xmath189 and @xmath190 at the lowest order .
we compute the @xmath171 functions up to five loops .
the complete series for @xmath191 are reported in app .
[ hopert ] .
let us first analyze the one - loop @xmath171 functions .
they read @xmath192 the exact normalization of the renormalized variables can be easily read from these series . since for @xmath20 and @xmath2 the two quartic terms are not independent , an appropriate combination of the above @xmath171 functions must reproduce the @xmath171 functions of the o(3 ) and o(8 ) @xmath56 theories . indeed , using eq .
( [ quar ] ) and setting @xmath193 , we obtain @xmath194 for @xmath20 , and @xmath195 for @xmath2 .
these exact relations provide a stringent check of the five - loop series for the model ( [ ahlg ] ) , which must reproduce the corresponding series of the o(3 ) and o(8 ) vector model @xcite for @xmath20 and 3 . for @xmath24 the fps of the rg flow
are given by the common zeros of the @xmath171 functions ( [ betau ] ) and ( [ betav ] ) .
their stability requires that the eigenvalues of the matrix @xmath196 ( where @xmath197 correspond to @xmath198 ) have positive real part . in the standard @xmath98-expansion scheme @xcite ,
the fps , i.e. , the common zeroes of the @xmath171-functions , are determined perturbatively as expansions in powers of @xmath199 , while exponents are obtained by expanding the corresponding rg functions computed at the fp in powers of @xmath98 .
a straightforward analysis of the one - loop @xmath171 functions ( [ betau ] ) and ( [ betav ] ) finds four different fps .
two of them have @xmath200 and are always unstable .
we have the trivial gaussian fp at @xmath201 , which is always unstable with respect to both quartic perturbations .
there is also an o(@xmath38 ) symmetric fp with @xmath202 at @xmath203 which can be shown , nonperturbatively , to be unstable with respect to the operator @xmath204 .
indeed , such operator contains a spin-4 perturbation with respect to the o(@xmath38 ) group @xcite , which is relevant for any @xmath205 to @xmath206 , and for any @xmath207 in three dimensions @xcite .
there are also two fps with @xmath208 .
one of them is stable , the other is unstable .
however , they only exist for @xmath209 , with @xmath210 for @xmath211 these two fixed points merge and then , for @xmath212 , they become complex . for @xmath2 the stable fixed point merges with the o(8 ) fixed point .
these results show that , for integer values of @xmath1 satisfying @xmath24 , there is no stable fixed point close to four dimensions , hence only first - order transitions are allowed . in order to establish the behavior of the system for @xmath213 ,
we must determine the fate of the stable fixed point that exists for @xmath214 close to four dimensions . for finite @xmath98 , we expect a stable and an unstable fixed point with @xmath215 up to @xmath216 .
the two fixed points merge for @xmath217 and become complex for @xmath218 . in order to compute @xmath219 ,
we expand @xmath220 and require @xmath221 the last equation being a consequence of the coalescence of the two fixed point at @xmath222 .
a straightforward calculation gives finally @xmath223 the expansion alternates in sign , as expected for a borel - summable series .
resummations using the pad - borel method appear to be stable .
we obtain @xmath224 using the series to order @xmath225 and @xmath226 at order @xmath227 ( the number in parentheses indicates how the estimate changes by varying the resummation parameters ) . apparently , @xmath228 varies only slightly as @xmath98 changes from 0 to 1 .
in particular , this analysis predicts the absence of stable fps for any integer @xmath24 in three dimensions .
methods based on the expansion around four dimensions allow us to find only the 3d fps which can be defined , by analytic continuation , close to four dimensions .
other fps , which do not have a 4d counterpart , can not be detected .
however , the extension of this result to the relevant @xmath229 dimension fails in some cases .
for example , this also occurs for the ginzburg - landau model of superconductors , in which a complex scalar field couples to a gauge field : although @xmath98-expansion calculations do not find a stable fp @xcite , thus predicting first - order transitions , numerical analyses of 3d systems described by the ginzburg - landau model show that they can also undergo continuous transitions , see , e.g. , ref .
this implies the presence of a stable fp in the 3d ginzburg - landau theory in agreement with experiments @xcite .
other examples are provided by the lgw @xmath56 theories describing frustrated spin models with noncollinear order @xcite , the @xmath230he superfluid transition from the normal to the planar phase @xcite , and the chiral transitions of the strong interactions in the case the u(1)@xmath231 anomaly effects are suppressed @xcite . therefore , a more conclusive analysis requires a direct study of the 3d flow .
this is achieved by an alternative analysis of the @xmath232 series : the 3d @xmath232 scheme without @xmath98 expansion @xcite .
the rg functions @xmath233 are the @xmath232 functions .
however , @xmath199 is no longer considered as a small quantity , but it is set equal to its physical value ( @xmath234 in our case ) before computing the fps .
this provides a well defined 3d perturbative scheme which allows us to compute universal quantities , without the need of expanding around @xmath235 @xcite .
we look for stable fps of the rg flow , with a finite attraction domain in the space of the renormalized couplings @xmath186 and @xmath187 .
the rg trajectories are determined by solving the differential equations @xmath236 where @xmath237 , with the initial conditions @xmath238 where @xmath239 parametrizes the different rg trajectories in terms of the bare quartic parameters , and the @xmath240 sign corresponds to the rg flows for positive and negative values of @xmath79 . in our study of the rg flow we only consider values of the bare couplings which satisfy eqs .
( [ stabcond1 ] ) and ( [ stabcond2 ] ) .
the physically relevant results are obtained by resumming the perturbative expansions , which are divergent but borel summable in a large region of the renormalized parameters .
the resummation can be done exploiting methods that take into account their large - order behavior , which is computed by semiclassical ( hence , intrinsically nonperturbative ) instanton calculations @xcite .
relevant results for the large - order behavior of the series of the model ( [ ahlg ] ) are reported in app .
[ sesum ] . in the massive zero - momentum ( mzm ) scheme @xcite
one performs the perturbative expansion directly in three dimensions , in the critical region of the disordered phase , in powers of the zero - momentum renormalized quartic couplings .
the theory is renormalized by introducing a set of zero - momentum conditions for the one - particle irreducible two - point and four - point correlation functions of the 2@xmath2412 matrix field @xmath49 : @xmath242,\nonumber \\ & & \gamma^{(4)}_{a_1a_2,b_1b_2,c_1c_2,d_1d_2}(0 ) = z_\psi^{-2 } m^{4-d } \times \label{ren2 } \\ & & \quad \times \left(u u_{a_1a_2,b_1b_2,c_1c_2,d_1d_2 } + v v_{a_1a_2,b_1b_2,c_1c_2,d_1d_2}\right ) , \nonumber \end{aligned}\ ] ] where @xmath243 are appropriate form factors defined so that @xmath244 and @xmath245 at the leading tree order .
the fps of the theory are given by the common zeroes of the callan - symanzik @xmath171-functions @xmath246 the normalization of the zero - momentum quartic variables @xmath247 is such that their one - loop @xmath171 functions read
@xmath248 we compute the mzm perturbative expansions of the @xmath171 functions and of the critical exponents up to six loops , requiring the computation of 1428 feynman diagrams .
the complete expansion for @xmath191 can be found in app .
[ hopert ] .
the large - order behaviors of the series are reported in app .
[ sesum ] .
the rg trajectories are obtained by solving differential equations analogous to eqs .
( [ rgfloweq ] ) and ( [ ini - rgflow ] ) , after resumming the @xmath171 functions as outlined in app .
[ sesum ] .
some rg trajectories in the renormalized coupling space of the lgw theory ( [ ahlg ] ) for @xmath191 are shown in figs .
[ n4trams ] and [ n4tramzm ] , for the @xmath181 and mzm schemes , respectively , for several values of the ratio @xmath249 . in both renormalization schemes
most of the rg trajectories flow towards the region in which the series are no longer borel summable . in the mzm scheme , for @xmath250
some trajectories flow instead towards infinity . in all cases
, we do not have evidence of a stable fp .
analogous results are obtained for @xmath251 .
these results imply that there is no universality class characterized by the symmetry breakings ( [ vnpat ] ) and ( [ vppat ] ) .
this would suggest a first - order transition .
it is also possible that more than one transition is present , each of them associated to a partial decoupling of some degrees of freedom , hence to a different symmetry - breaking pattern , as it happens in two - dimensional frustrated xy models @xcite . in this case
, continuous transition would still be possible .
we have investigated the nature of the phase transitions in 3d acp@xmath0 models , such as the lattice model ( [ hcpn ] ) with @xmath26 , which are characterized by a global u(@xmath1 ) symmetry and a local u(1 ) gauge symmetry . in order to analyze their critical behavior , we construct the corresponding lgw theory , assuming a staggered gauge - invariant order parameter .
this leads to the lgw @xmath252 theory ( [ ahlg ] ) , where @xmath49 is a traceless hermitian @xmath15 matrix , which is symmetric under the global u(@xmath1 ) transformations @xmath253 , and the @xmath50 transformations @xmath51 . for @xmath254 ,
the lgw mode is equivalent to the lgw associated with an 8-component real vector field .
hence , we predict that , if the acp@xmath3 model undergoes a continuous transition , it should belong to the o(8 ) vector universality class . note that , at the critical point , there is an effective symmetry enlargement , u(3)@xmath255o(8 ) , and the same should occur in the low - temperature phase as the critical point is approached .
the low - temperature symmetry @xmath256 should be promoted to o(7 ) .
we confirm the rg predictions by comparing the fss behavior of the o(8 ) vector and acp@xmath3 models , obtained by mc simulations of both lattice models .
we note that the critical behavior , characterized by the o(8 ) critical exponents @xmath257 and @xmath258 , definitely differs from that of ferromagnetic cp@xmath3 models , for which recent studies @xcite have provided numerical evidence of continuous transitions with critical exponents @xmath22 and @xmath259 . in the case of acp@xmath0 lattice models with a higher number of components , i.e. @xmath24 , the identification of the order parameter is more complex .
if the order parameter is a staggered gauge - invariant hermitian matrix as for @xmath20 and @xmath2 , the associated lgw theory is that given in eq .
( [ ahlg ] ) . to determine the possible existence of continuous transitions , we study the rg flow in perturbation theory . we compute ft perturbative series in two different renormalization schemes up to five and six loops , respectively .
the analysis of the rg flow does not provide evidence of stable fps .
this implies that the dynamics of the staggered gauge - invariant modes associated with the @xmath15 hermitian matrices defined in eq .
( [ biab ] ) does not give rise to continuous transitions .
in particular , we do not expect continuous transitions characterized by the symmetry breaking @xmath260 .
thus , if the acp@xmath0 lattice model presents transitions with this symmetry breaking , they must be first order .
note that it is still possible to have continuous transitions if they are associated with a different symmetry breaking , as it may arise from a partial decoupling of some degrees of freedom .
another possible scenario may arise from the relevance of further gauge degrees of freedom which are not taken into account by the lgw theory , analogously to the case of ferromagnetic cp@xmath0 model in the large-@xmath1 limit .
this issue needs further investigation .
] in this section we wish to characterize the structure of the ground state of the antiferromagnetic model for @xmath2 , showing the emergence of a ferromagnetic order on a staggered lattice .
let us first consider a lattice plaquette , see fig .
[ gs_n3](a ) , and let us determine the configurations of the four spins @xmath261 , @xmath262 , @xmath263 , and @xmath264 that minimize the energy
. given the global invariance of the model , it is not restrictive to assume that @xmath265 . since the model is antiferromagnetic , the energy is minimized if neighbouring spins are orthogonal , i.e. , if @xmath266 = 0 for nearest neughbor sites @xmath267 and @xmath268 .
therefore , we have @xmath269 where @xmath270 and @xmath271 are two - dimensional complex unit vectors . if we set @xmath272 and consider the links connecting site 4 with its neighbors , we obtain the conditions @xmath273 if @xmath274 ( @xmath86 is an arbitrary phase ) , this condition implies @xmath275 .
therefore , discarding an irrelevant phase , we obtain @xmath276 . if instead if @xmath277 , eliminating an irrelevant phase we obtain @xmath278 .
this analysis shows , therefore , that in the ground - state configuration two opposite spins in the plaquette are identical .
note that the result holds only for @xmath2 . for @xmath4 , eq .
( [ eq - v2v3v4 ] ) does not imply @xmath275 in the generic case .
we only obtain that @xmath279 belongs to @xmath280-dimensional subspace containing @xmath261 .
this may leave open the possibility of other symmetry breaking patterns for @xmath281 .
let us now assume that we are dealing with a three - dimensional system and let us consider a cube , see fig .
[ gs_n3](b ) .
we wish now to show that the dominant ground - state configurations have the following structure .
one can have @xmath282 ( we properly fix the phases ) , while @xmath262 , @xmath264 , @xmath283 , @xmath284 are arbitrary spins lying in the complex two - dimensional plane orthogonal to @xmath261 .
of course , the equivalent arrangement with @xmath285 is also possible .
note that this configuration is highly degenerate , as four two - dimensional vectors can be arbitrarily chosen . to prove the previous statement
, we consider all other possible arrangements that are consistent with the result obtained for the plaquette and we show that they are less degenerate that the one discussed before .
hence they are entropically disfavored and become irrelevant in the infinite - volume limit .
it is not restrictive to assume that @xmath286 , since two opposite spins of plaquette @xmath287 ( we report in angular brackets the sites belonging to the plaquette ) are necessarily identical .
let us now consider the plaquette @xmath288 .
there are three different possibilities consistent with the result valid for a single plaquette : ( i ) @xmath289 and @xmath290 ; ( ii ) @xmath289 with @xmath291 ; ( iii ) @xmath292 .
we wish now to exclude cases ( ii ) and ( iii ) . in case ( ii )
the orthogonality condition applied to links ( 1,2 ) and ( 2,6 ) implies @xmath293 since @xmath294 , these two conditions completely specify vector @xmath262 ( the phase is of course irrelevant ) . repeating the argument to the spins at sites 2,4,5,7 we end up with @xmath295 it is therefore a configuration of type ( i ) , but now the spins are ordered on the complementary sites 2,4,5,7 .
let us now consider case ( iii ) .
generically , @xmath296 and @xmath297 .
this implies that @xmath298 and @xmath299 are uniquely defined .
therefore , the configuration is defined by specifying @xmath262 , @xmath264 , and @xmath283 , in the two - dimensional complex space orthogonal to @xmath261 .
this configuration is less degenerate than that defined at point ( i ) , hence it is irrelevant in the infinite - volume limit ( it is entropically suppressed ) .
if @xmath300 and @xmath301 , we obtain the same type of degeneracy , since now @xmath264 , @xmath283 , and @xmath298 can be chosen .
the other two cases give the same result .
the result for a cube extends trivially to the whole lattice , proving that in the ground state we observe two different symmetry breakings .
first , lattice translational invariance is broken with the emergence of a staggered symmetry .
second , on one of the two sublattices the system orders ferromagnetically , breaking the @xmath302 symmetry down to @xmath303 .
it is important to note that the discussion only applies in three dimensions . in two dimensions the results
does not hold . indeed , referring to fig .
[ gs_n3](c ) , nothing forbids in the two - dimensional case a configuration with @xmath276 , @xmath304 , and @xmath305 .
in this appendix we report the ft perturbative series of the @xmath171 functions used in our rg analysis of sec .
we only report those for @xmath191 ; the perturbative series for other values of @xmath1 are available on request .
the five - loop @xmath171 functions of the @xmath306 scheme for @xmath191 are @xmath307 @xmath308 the six - loop @xmath171 functions of the mzm scheme for @xmath191 are @xmath309 @xmath310
since perturbative expansions are divergent , resummation methods must be used to obtain meaningful results . given a generic quantity @xmath311 with perturbative expansion @xmath312 , we consider @xmath313 which must be evaluated at @xmath314 .
the expansion ( [ seriesx ] ) in powers of @xmath315 is resummed by using the conformal - mapping method @xcite that exploits the knowledge of the large - order behavior of the coefficients , generally given by @xmath316^{k}\,k^b\,\left [ 1 + o(k^{-1})\right ] .
\label{lobh}\ ] ] the quantity @xmath317 is related to the singularity @xmath318 of the borel transform @xmath319 that is nearest to the origin : @xmath320 .
the series is borel summable for @xmath321 if @xmath319 does not have singularities on the positive real axis , and , in particular , if @xmath322 .
the large - order behavior can be determined using semiclassical computations , based on the computations of appropriate instanton configurations @xcite . for even @xmath1 ,
these semiclassical calculations show that the expansion is borel summable when @xmath323 where @xmath324 is given in eq .
( [ stabcond1 ] ) . for odd @xmath1
we obtain analogously @xmath325 where @xmath326 is given in eq .
( [ stabcond3 ] ) .
note that the conditions for borel summability on the renormalized couplings correspond to the stability conditions ( [ stabcond1 ] ) and ( [ stabcond2 ] ) of the bare quartic couplings . in this borel - summability region
we have for even @xmath1 @xmath327 for odd @xmath1 , we should replace @xmath328 with @xmath329 . under the additional assumption that the borel - transform singularities lie only in the negative axis
, the conformal - mapping method turns the original expansion into a convergent one in the region ( [ brr ] ) .
outside , the expansion is not borel summable .
analogously one can derive the large - order behavior of the mzm scheme , which is again given by eq .
( [ lobh ] ) but with @xmath330 for even @xmath1 . for odd values of @xmath1
, @xmath328 should be replaced with @xmath329 .
we use the conformal mapping method to resum the series taking into account what we know about their large - order behavior .
the method we use is described in refs .
resummations depend on two parameters , which are optimized in the procedure .
the approximants we use depend on two parameters @xmath331 and @xmath332 ; we use here the notations of refs . @xcite . | we investigate the critical behavior of three - dimensional antiferromagnetic cp@xmath0 ( acp@xmath0 ) models in cubic lattices , which are characterized by a global u(@xmath1 ) symmetry and a local u(1 ) gauge symmetry . assuming that critical fluctuations are associated with a staggered gauge - invariant ( hermitian traceless matrix ) order parameter , we determine the corresponding landau - ginzburg - wilson ( lgw ) model . for @xmath2
this mapping allows us to conclude that the three - component acp@xmath3 model undergoes a continuous transition that belongs to the o(8 ) vector universality class , with an effective enlargement of the symmetry at the critical point .
this prediction is confirmed by a detailed numerical comparison of finite - size data for the acp@xmath3 and the o(8 ) vector models .
we also present a renormalization - group ( rg ) analysis of the lgw theories for @xmath4 .
we compute perturbative series in two different renormalization schemes and analyze the corresponding rg flow . we do not find stable fixed points that can be associated with continuous transitions . |
in a standard imaging system , light scattered from an object forms a diffraction pattern which encodes information about the object fourier components .
a lens recombines the scattered rays so that they interfere correctly to form an ) . ,
title="fig:",scaledwidth=20.0% ] , title="fig:",scaledwidth=20.0% ] these ideas , along with the development of powerful light sources , producing collimated beams of coherent x - rays , enabled the development of coherent x - ray diffraction microscopy @xcite ( aka lensless or diffractive imaging ) .
this technique aims at imaging , through coherent illumination , fourier amplitude measurements and adequate sampling , macroscopic objects such as entire cellular organisms @xcite , or nanoporous aerogel structures @xcite .
see @xcite for a review .
diffraction microscopy solves the phase problem using increasingly sophisticated algorithms based on the support constraint , which assumes adequate sampling .
the object being imaged is limited within a support region @xmath0 : @xmath1 the sampling conditions required to benefit from the support constraint have limited the adoption of projection algorithms to other experimental geometries that allow only for sub - nyquist sampling , most notably bragg sampling from periodic crystalline structures .
modern sampling theory however , tells us that that nyquist sampling conditions dictated by the support are the worst case scenario for an arbitrary object . in other words ,
shannon was a pessimist : he did not account for the signal structure .
compressive sensing theory tells us that the number of measurements are dictated by the signal structure rather than it s length . by structured
we mean that the signal has only a few non - zero coefficients when represented in terms of some basis , or can be well approximated well by a few non zero coefficients : they can be described in terms of a few atoms , a few stars , a few wavelet coefficients , or possibly a few protein folds .
in other words , an object of interest is often sparse or concentrated in a small number of non - zero coefficients in a well chosen basis , i.e. it can be compressed with no or almost no loss of information .
the meanings of `` well - chosen '' and `` of interest '' are slightly circular : a basis is well - chosen if it succinctly describes a signal of interest ; likewise , a signal is of interest if it can be described with just a handful of basis elements . since we do not know where these few terms are located , conventional wisdom would indicate that one has to first measure the full sample at the desired resolution , since overlooking an important component of a signal seems almost inevitable if the whole haystack is nt thoroughly searched over .
there would seem to be no alternative to processing each signal in its entirety before we can compress it and store only the desired information ( such as the location of the atoms in a molecule ) . but
a new theory of `` compressive sampling '' has shown how an image of interest or structured signals generally can be reconstructed , exactly , from a surprisingly small set of direct measurements .
cands and colleagues @xcite have defined a notion of `` uniform uncertainty '' that guarantees , with arbitrarily high probability , an exact solution when the signal is sparse and a good approximation when it is compressible or noisy .
their uniform uncertainty condition is satisfied , among others , by fourier measurements of a sparse real space object .
the question of whether modern sensing theory is applicable to fourier amplitude measurements was first raised by moravec , romberg and baraniuk @xcite who provide an upper bound sampling condition for the successful retrieval of a sparse signal autocorrelation , and discuss other conjectures with far reaching consequences for low - resolution undersampled phase retrieval .
since the theory is relatively new and not widely known to the phase retrieval community , modern sampling theory is briefly reviewed .
the notion that a diffraction pattern from a sparse object can be reconstructed at sub - nyquist sampling is not entirely new .
the so called `` direct methods '' are routinely used for atomic resolution imaging of increasing complex molecular structures .
direct methods enforce the condition that the resulting molecule is composed of a finite number of atoms .
the conditions for successful ab - initio phase retrieval using these methods are strict : it requires ( 1 ) atomic resolution and ( 2 ) about 5 strong peaks per atom .
condition ( 2 ) means that the algorithms do not scale well with a large number of atoms since the number of strong reflections decreases rapidly with the number of atoms .
suppose that one collects an incomplete set of frequency samples ( amplitude and phase ) of a discrete signal @xmath2 of length @xmath3 .
the goal is to reconstruct the full signal @xmath4 given only @xmath5 samples in the fourier domain where the `` visible frequencies '' are a subset @xmath6 ( of size @xmath5 ) of the set of all frequencies @xmath7 . at first glance ,
solving the underdetermined system of equations appears hopeless , as it is easy to make up examples for which it clearly can not be done . but suppose now that the signal @xmath4 is compressible , meaning that it essentially depends on a number of degrees of freedom which is smaller than @xmath3 . then in fact , accurate and sometimes exact recovery is possible by solving a simple convex optimization problem .
( candes romberg and tao @xcite ) : assume that @xmath4 is @xmath8 -sparse , ( e.g. @xmath8 atomic charges in real space with @xmath3 resolution elements ) , and that we are given @xmath5 fourier coefficients with frequencies selected uniformly at random .
suppose that the number of observations obeys @xmath9 .
then minimizing @xmath10 reconstructs @xmath4 exactly with overwhelming probability . in particular , writing @xmath11 , then the probability of success exceeds @xmath12 .
the theorem shows that a simple convex minimization will find the exact solution without any knowledge about the support , the number of nonzero coordinates of @xmath13 , their locations , and their amplitudes which we assume are all completely unknown a priori .
following @xcite we formulate this more explicitly . the algorithm that optimizes the @xmath10 norm : latexmath:[\ ] ] although this result describes only local convergence , it shows how a few wrong peaks can be recovered easily .
27 l. pauling and m. d. shappell , zeits .
* 75 * , 128 ( 1930 ) .
bernal , i. fankuchen , m. f. perutz , `` an x - ray study of chymotrypsin and haemoglobin . ''
nature 141 , 523 - 524 ( 1938 ) .
d. sayre , `` on the implication of a theorem due to shannon '' , acta cryst . * 5 * , ( 1952 ) 843 .
y. m. bruck and l. g. sodin .
`` on the ambiguity of the image reconstruction problem . ''
optics communications , * 30*(3):304 - 308 , ( 1979 ) .
`` fourier phase problems are uniquely solvable in more than one dimension .
i : underlying theory '' optik , * 61*(3):247 - 262 , 1982 . m. h. hayes , `` the reconstruction of a multidimensional sequence from the phase or magnitude of its fourier transform , '' ieee trans .
30*(2 ) , 140 - 154 , ( 1982 ) . m. h. hayes and j. h. mcclellan , `` reducible polynomials in more than one variable , '' proc .
70*(2 ) , 197 - 198 , ( 1982 ) .
j. miao , p. charalambous , j. kirz , d. sayre , `` extending the methodology of x - ray crystallography to allow imaging of micrometre - sized non - crystalline specimens , '' nature * 400 * , 342 - 344 ( 1999 ) .
d. shapiro , p. thibault , t. beetz , v. elser , m. howells , c. jacobsen , j. kirz , e. lima , h. miao , a. neiman , d. sayre , `` biological imaging by soft x - ray diffraction microscopy , '' proc .
* 102 * , 1543 - 1546 ( 2005 ) .
nature * 442 * , 63 - 67 ( 2006 ) .
a. barty , s. marchesini , h. n. chapman , c. cui , m. r. howells , d. a. shapiro , a. m. minor , j. c. h. spence , u. weierstall , j. ilavsky , a. noy , s. p. hau - riege , a. b. artyukhin , t. baumann , t. willey , j. stolken , t. van buuren , j. h. kinney , `` three - dimensional coherent x - ray diffraction imaging of a ceramic nanofoam : determination of structural deformation mechanisms , '' * 101 * , 055501 ( 2008 ) , [ arxiv:0708.4035 ] .
p. w. hawkes & j. c. h. spence ( eds . ) , _ science of microscopy _ ( springer , 2007 ) .
e. j. cands , j. romberg and t. tao , `` robust uncertainty principles : exact signal reconstruction from highly incomplete frequency information , '' ieee trans . inform .
theory , * 52 * , 489 - 509 ( 2006 ) [ arxiv : math/0409186 ] .
m. l. moravec , j. k. romberg , r. g. baraniuk , richard , compressive phase retrieval , wavelets xii .
spie * 6701 * , 670120 ( 2007 ) .
e. van den berg and m. p. friedlander , spgl1 : a solver for large - scale sparse reconstruction , http://www.cs.ubc.ca/labs/scl/index.php/main/spgl1 e. van den berg and m. p. friedlander , `` probing the pareto frontier for basis pursuit solutions '' , ubc computer science technical report tr-2008 - 01 , january 2008 .
available at http://www.optimization-online.org/db_f w. h. richardson , bayesian - based iterative method of image restoration " . *
62 * , 55 - 59 ( 1972 ) .
l. b. lucy , an iterative technique for the rectification of observed distributions "
. astronomical journal textbf79 745 - 754 ( 1974 ) .
fienup , `` phase retrieval for undersampled broadband images , '' , * 16 * , 1831 - 1839 ( 1999 ) .
s. marchesini , h. he , h. n. chapman , s. p. hau - riege , a. noy , m. r. howells , u. weierstall , j.c.h .
spence , `` x - ray image reconstruction from a diffraction pattern alone , '' * 68 * , 140101(r ) 1 - 4 , ( 2003 ) , [ arxiv : physics/0306174 ] .
l. d. marks , w. sinkler and e. landree , `` a feasible set approach to the crystallographic phase problem '' , acta cryst . * a55 * , 601 - 612 ( 1999 ) . j. r. fienup , `` reconstruction of an object from the modulus of its fourier transform , '' * 3 * , 27 - 29 ( 1978 ) . j. r. fienup , `` phase retrieval algorithms : a comparison '' , * 21 * , 2758 - 2769 ( 1982 ) . j. p. abrahams , a. w. g. leslie , acta cryst . *
52 * , 30 - 42 ( 1996 ) .
v. elser , `` phase retrieval by iterated projections , '' * 20 * , 40 - 55 ( 2003 ) .
h. h. bauschke , p. l. combettes , and d. r. luke , `` hybrid projection reflection method for phase retrieval , '' * 20 * , 1025 - 1034 ( 2003 ) .
d. r. luke , `` relaxed averaged alternating reflections for diffraction imaging , '' inverse problems * 21 * , 37 - 50 ( 2005 ) . , ( arxiv : math.oc/0405208 ) .
l. combettes , `` the convex feasibility problem in image recovery , in advances in imaging and electron physics , '' ( p. hawkes , ed . ) , vol .
155 - 270 .
( academic press , new york 1996 ) .
s. marchesini , `` a unified evaluation of iterative projection algorithms for phase retrieval , '' rev .
inst . * 78 * , 011301 1 - 10 ( 2007 ) , [ arxiv : physics/0603201 ] . g. oszlnyi and a. sto , `` ab initio structure solution by charge flipping , '' acta cryst . * a60 * , 134 - 141 ( 2004 ) [ arxiv : cond - mat/0308129 ] .
j. wu , k. leinenweber , j. c. h. spence , `` ab initio phasing of x - ray powder diffraction patterns by charge flipping , '' nature materials * 5 * , 647 - 652 ( 2006 ) .
e. j. cands , d. l. donoho , `` new tight frames of curvelets and optimal representations of objects with piecewise c@xmath94 singularities . ''
pure appl . math .
* 57 * , 219 - 266 ( 2004 ) .
s. marchesini , 24,32890 - 3296 ( 2007 ) , [ arxiv : physics/0611233 ] . g. oszlnyi and a. sto , `` ab initio structure solution by charge flipping .
ii . use of weak reflections , '' acta crystallogr . * a61 * , 147 - 152 ( 2005 ) .
i. daubechies , m. defrise , and c. d. mol , `` an iterative thresholding algorithm for linear inverse problems with a sparsity constraint , '' comm .
pure appl . math . * 57*(11 ) , 1413 - 1457 , ( 2004 ) . | any object on earth has two fundamental properties : it is finite , and it is made of atoms . structural information about an object can be obtained from diffraction amplitude measurements that account for either one of these traits .
nyquist - sampling of the fourier amplitudes is sufficient to image single particles of finite size at any resolution .
atomic resolution data is routinely used to image molecules replicated in a crystal structure . here
we report an algorithm that requires neither information , but uses the fact that an image of a natural object is compressible .
intended applications include tomographic diffractive imaging , crystallography , powder diffraction , small angle x - ray scattering and random fourier amplitude measurements . |
starbursts are a well known phenomenon in some dirrs ( e.g. gallagher & hunter @xcite ; for a critical discussion of the exact definition of starbursts see knapen & james @xcite ) .
they imply a mechanism to organise large - scale star formation which can not operate all the time due to fast gas consumption . moreover
, the instability leading to an episodic sfr should work in a small parameter range only , because most of the isolated dirrs seem to evolve in a highly self - regulated manner .
the first theoretical models dealing with the variability of the star formation rate in galaxies were based on closed - box models . for reasonable models of the interstellar medium ( ism ) it turned out that stellar feedback is rather efficient in suppressing instabilities ( e.g. ikeuchi & tomita @xcite , scalo & struck - marcell @xcite ) .
these models only allow for a burst - like behaviour when a long time - delay between star formation and stellar feedback is introduced .
kppen et al .
* hereafter kth95 ) corroborated this result by investigating a simple numerical model motivated by full chemodynamical simulations .
they demonstrated that the effective sfr is almost independent of the detailed recipe for the stellar birth function , provided a negative feedback due to the thermal state of the ism is considered in the stellar birth function .
these box models , however , suffer from neglecting galactic dynamics .
several attempts have been made to study the evolution of isolated dwarf galaxies by taking full dynamics , as well as star formation and stellar feedback , into account ( kruegel et al . @xcite , and later hensler et al.@xcite , pelupessy et al .
@xcite , struck @xcite ) . though these calculations differ in many details , they agree in producing large - scale star formation variations on dynamical timescales which indicates the crucial role dynamics plays for episodic starbursts .
a major disadvantage of numerical simulations is their complexity .
this often prevents a detailed investigation of the parameter space . instead of analysing dwarf galaxies by such detailed models , we apply a set of equations that is as simple as possible , but still incorporating the generic features of the complex numerical simulations .
we use an extended version of the analytical models of kth95 that allows us to deal with the dynamical evolution of a galaxy in a simple way ( for details cf . to theis & kppen @xcite ) . here
we compare the evolution of dwarf galaxies for two different ism models : in the diffuse ism model energy is dissipated by radiation , whereas in the clumpy ism model energy is lost via inelastic cloud collisions .
* diffuse ism models . * fig . [ theisfig1 ] shows the evolution for the diffuse ism model . due to strong cooling , the gas temperature drops almost immediately to values near @xmath0 bringing the system far out of virial equilibrium . in the ongoing collapse
the density increases and the gas temperature drops even more .
the cooling and density enhancement continue , until star formation becomes more prominent after about 100 myr .
the first stars reinject energy to the ism leading to the quasi - equilibrium stage known from the box models of kth95 .
however , different to them , the density still increases due to the ongoing collapse .
though the star formation and , by this , the stellar energy injection grows , the feedback can not stop further collapse , because cooling is too efficient to prevent the gas from reaching virial equilibrium .
the collapse proceeds until the angular momentum barrier becomes dominant and the system expands .
then the first star burst episode finishes and the gas is quickly heated to @xmath0 .
however , due to the strong enhancement of the cooling rate at temperatures beyond @xmath0 , the gas never reaches the virial temperature .
the longterm evolution of this model shows that the previously discussed bursts are just transitory phenomena .
the oscillations are almost completely damped after 2 gyr .
such a behaviour , i.e. only an initial burst , was found in almost all models with radiative cooling as source of dissipation
. * clumpy ism models . * the evolutionary behaviour changes when applying the clumpy ism model .
though the structure of the equations remains identical , the dissipational timescale differs substantially from the previous case in which the cooling timescale was much shorter than other involved timescales . for cloud - cloud collisions ,
the timescale @xmath1 becomes of the order of gyrs .
driven by stellar feedback and the much longer dissipational time the ism expands after the first burst ( fig .
[ theisfig2 ] , upper right ) .
the dark matter halo , however , prevents the gas from being blown - away .
therefore , it collapses again reaching a dynamical equilibrium after about 2 gyrs .
this can be seen from a comparison between the actual kinetic temperature ( velocity dispersion of the cloud system ) and the virial temperature of the system .
the ongoing evolution is governed by a slow loss of energy due to dissipation . during this stage of slow shrinkage
the evolution is close to virial equilibrium .
after 6 gyrs the system lost sufficient energy to start with the next run - away collapse resulting in another starburst event after a long period of quiescence .
finally , the galaxy produces stars on a mean rate of about @xmath2 .
the main difference between the two model scenarios is the ratio @xmath3 between the dissipational timescale @xmath4 and the dynamical timescale @xmath5 . in case of the diffuse ism model ( small @xmath3 )
the energy input by the stars is almost instantaneously radiated away .
therefore , the radial oscillations are only decaying initial virial oscillations acting on a dynamical timescale .
this behaviour turned out to be rather robust against variations of model parameters like stellar heating , gas mass , initial size or stellar birth function . on the other hand ,
when @xmath3 becomes of the order of one or larger ( i.e. for the clumpy ism model ) , two timescales are important : the dynamical timescale and the dissipational timescale .
after a first burst , a long period ( @xmath6 ) of quiescence follows , before the next star formation episode starts . during most of the quiescence phase
the gas is semi - adiabatically accreted with an average infall velocity of only a few kms@xmath7 .
the related extended gas reservoirs are known around many dwarf galaxies ( e.g. ngc 4449 ) .
they are very sensitive to perturbations by other galaxies .
e.g. a substantial amount of the gas might be stripped off , by this reducing the strength of the second burst . after the second burst the galaxy enters a phase with continuous star formation up to the present ( here : 10 gyr ) .
this is in agreement with observations of several dirrs which still show a substantial , more or less continous star formation ongoing over the last one or few gyrs ( e.g. greggio et al . | we study the stability properties of isolated star forming dwarf galaxies which undergo dynamically driven star bursts induced by stellar feedback . here
we focus on the impact of the adopted ism model , i.e. either a diffuse or a clumpy ism .
we apply a one - zone model extended for active dynamical evolution .
this allows for a coupling between the dynamical state of the galaxy and its internal properties like star formation activity or the thermal state ( or dynamical pressure , respectively ) of the interstellar medium ( ism ) .
we found two major types of repetitive star bursts : one set ( type a ) of quasi - periodic star bursts is related to the dynamical timescale of the galaxy . in that case , the star formation follows the variations of the gas density induced by decaying virial oscillations .
the second set ( type b ) of recurrent star bursts is characterized by a long quiescence period given by the sum of the dynamical and the dissiptional timescale : after a first burst , the inserted energy leads to a substantial expansion of the system , by this stopping any significant sf activity .
a next burst might occur , when the gas reaches high densities again , i.e. after the gas recollapsed and the energy injected by stellar feedback is dissipated . in case of a diffuse ism model , type a bursts are the most common type due to the high efficiency of radiative cooling ( no type b bursts are found ) .
bursts occur then mainly during an initial transitory phase . in case of a clumpy ism model ( i.e. dissipation by inelastic cloud - cloud collisions ) ,
the dissipative timescale is of the order of the dynamical time or longer .
this allows for both , type a and type b bursts . whereas initial transitory bursts are quite common , type b bursts are only found in a small mass range for given feedback and dissipation parameters . |
the @xmath0-quark mass for instance , the @xmath8 running mass at renormalization scale @xmath9 @xmath10 or @xmath11 is one of the fundamental parameters of the standard model and therefore its precise knowledge is highly desirable .
the latest edition of the review of particle physics reports @xmath12 @xcite .
a direct way to determine @xmath1 is by means of lattice qcd simulations ; however , since the physical @xmath0-quark is too heavy for current lattice setups , the determination of @xmath1 from pure lattice qcd requires either the extrapolation of the lattice results from lighter simulated masses or the use of the heavy - quark effective theory ( hqet ) formulated on the lattice . using the former approach ,
the values @xmath13 @xcite and @xmath14 @xcite have recently been deduced , while the results @xmath15 @xcite , @xmath16 @xcite , and @xmath17 @xcite have been determined adopting the hqet - based approach .
all above findings have been obtained using unquenched gauge configurations with @xmath18 dynamical flavors in the sea .
recently , more accurate determinations of the @xmath0-quark mass have been performed by exploiting moment sum rules for two - point functions of heavy
heavy currents : low-@xmath19 moment sum rules based on three - loop @xmath6 @xcite and four - loop @xmath20 @xcite fixed - order pqcd calculations combined with the experimental data yield @xmath21 @xcite and @xmath22 @xcite , respectively . the latter finding has been confirmed by a study based on a combination of perturbative qcd and lattice qcd simulations with @xmath23 dynamical flavors in the sea @xcite . combining large-@xmath19 moments obtained within renormalization - group - improved nnll - order @xmath24 sum rules with the experimental data yields @xmath25 @xcite . in this paper , we show that the borel qcd sum rules for heavy
light correlators provide the possibility to extract the bottom - quark mass with comparable accuracy if a precise value for the @xmath2-meson decay constant @xmath5 is adopted as input .
let us first explore what degree of sensitivity of @xmath5 to the precise value of the @xmath0-quark mass can be expected on the basis of simple quantum - mechanical considerations .
nonrelativistic potential models predict the following relationship between the ground - state wave function at the origin , @xmath26 and the ground - state binding energy @xmath27 : latexmath:[\[\begin{aligned } \label{qm1 } exact for any ground state in a purely coulomb or purely harmonic - oscillator potential ( or , to be more precise , in any model where the potential involves only one coupling constant ) .
moreover , this relation proves to be a good approximation for ground states in potentials given by the sum of confining and coulomb interactions @xcite .
now , taking into account that the decay constant is the analogue of the wave function at the origin and incorporating the known scaling behaviour of the decay constant of a heavy meson in the heavy - quark limit @xcite which should work well for the beauty mesons one obtains an approximate relation between the @xmath2-meson mass @xmath29 and the heavy - quark pole mass @xmath30 : @xmath31 keeping the ground - state mass fixed and equal to its experimental value @xmath32 , we can easily pin down the dependence of @xmath5 on small variations @xmath33 of the heavy - quark pole mass near some average value of @xmath30 . taking into account
that @xmath34 for @xmath35 we end up with @xmath36 and @xmath37 or , equivalently , @xmath38 thus , the sensitivity of @xmath5 to the precise value of the heavy - quark mass should be rather high : a variation of the quark mass by @xmath39 entails @xmath40 clearly , a similar effect should be observable in the outcomes of qcd sum rules @xcite . recently , several qcd sum - rule analyses @xcite of beauty - meson decay constants relying on three - loop heavy
light correlators @xcite have been published , see table [ table:1 ] ( note that all results collected in table [ table:1 ] are obtained by applying the qcd sum - rule method to essentially the same analytical expression for the correlator ) .
.some recent qcd sum - rule predictions for @xmath5 from heavy light two - point functions . [ cols="^,^,^,^,^",options="header " , ] figure [ plot :
mb]c presents the extracted values of @xmath1 depending on the number of terms kept in the perturbative part of the correlator . moving from @xmath41 ( lo ) to @xmath42 ( nlo ) accuracy of the perturbative expansion
has two effects : first , the central @xmath1 value decreases sizeably from @xmath43 to @xmath44 and , second , the ope - error also reduces considerably .
adding the @xmath6 ( nnlo ) correction does not , however , entail a sizeable change of the predictions : @xmath45 obviously , the extracted values of @xmath1 exhibit a nice `` convergence '' depending on the accuracy of the perturbative correlation function .
the ope error in the extracted @xmath46 is related to the variations of the ope parameters in the ranges given in ( [ table:2 ] ) and the independent variations of the scales @xmath47 and @xmath48 in the range 3 gev @xmath49 6 gev .
the individual contributions to the ope error read : 14 mev ( @xmath47 , @xmath48 ) , 20 mev ( quark condensate ) , 7 mev ( gluon condensate ) , 8 mev ( @xmath50 ) , and 4 mev ( light - quark mass ) . adding these values in the quadrature
gives 27 mev .
the systematic uncertainty in the extracted value of @xmath1 is found as the spread of the results for different anstze for the effective continuum threshold and amounts to 11 mev . to obtain the final estimate for @xmath1 one should further add the ( gaussian ) error 18 mev , related to the uncertainty in the lattice value of @xmath51 mev .
the @xmath20 correction to the perturbative spectral density is , at present , not known .
nevertheless , on the basis of our findings we do not expect a sizeable shift of the central value of @xmath1 due to the inclusion of the @xmath20 correction .
one , however , might expect a reduction of the sensitivity of the extracted value of @xmath1 to the precise values of the scales @xmath47 and @xmath48 and thus a further increase of the accuracy of the extracted value of the bottom - quark mass .
we performed a detailed qcd sum - rule analysis of the @xmath2- and @xmath52-meson decay constants , with particular emphasis on the study of the errors in the extracted decay - constant values : the ope uncertainty due to the errors of the qcd parameters and the intrinsic error of the sum - rule approach due to the limited accuracy of the extraction procedure .
our main findings may be summarized by the following observations : 1 .
the choice of the renormalization scheme used to define the heavy - quark mass is crucial for the convergence of the perturbative expansion of the two - point function : the latter exhibits in its pole - mass formulation no sign of convergence but develops in its running - mass formulation a clear hierarchy of the perturbative contributions . for the extracted decay constant ,
the pole - mass result is sizeably smaller than its running - mass counterpart , albeit both enjoy a perfect stability in the borel parameter .
borel stability does not imply reliability of sum - rule results .
the extraction of hadronic properties is significantly improved by allowing a borel - parameter dependence for the effective continuum threshold , which then quite naturally increases the accuracy of the duality approximation .
as shown already before in the charmed - meson sector @xcite , considering suitably optimized polynomial anstze for the effective continuum threshold provides an estimate of the intrinsic uncertainty of the method of qcd sum rules .
3 . for beauty mesons , a very strong correlation between the exact @xmath1 value and the sum - rule result for @xmath5
is found : @xmath53 this enables us to revert the problem and to make use of the precise lattice - qcd computations of @xmath5 to extract the value of @xmath1 . combining our sum - rule analysis with the latest results for @xmath5 and @xmath54 from lattice qcd yields
@xmath55 the ope error is related to the uncertainties in the ope input parameters , and the `` exp '' error is induced by the error in the lattice determination of @xmath5 .
news is that the systematic uncertainty of the sum - rule method , estimated from the spread of the results for different anstze of the effective continuum threshold , amounts to @xmath56 and remains under control .
adding all three error in quadrature yields our final estimate @xmath57 with ( [ ourmb1 ] ) , the qcd sum rules for heavy
light correlators evaluated at @xmath6 accuracy yield , for the decay constants , @xmath58 our algorithm enables us to provide both the ope uncertainties and the intrinsic ( systematic ) uncertainty of the sum - rule method related to the limited accuracy of the extraction procedure .
we observe an extreme sensitivity of the decay constant to the input value of the quark mass , but only for beauty mesons .
it is not observed in the charm sector , where one finds @xmath59 @xcite .
therefore , the extracted value of @xmath60 is rather mildly sensitive to the precise value of @xmath61 in the charm sector , on the other hand , one observes a stronger sensitivity of the extracted value of @xmath60 to the algorithm adopted for fixing the @xmath62-dependent effective continuum threshold @xcite .
our value ( [ ourmb ] ) of @xmath1 is extracted from the borel qcd sum rule for the heavy
light correlator known to @xmath6 accuracy . taking into account that the value of @xmath1 is changing only marginally when moving from the @xmath42 to @xmath6 accuracy of the correlator , we do not expect that the inclusion of the presently unknown @xmath20 correction will lead to a substantial change in the extracted value of @xmath1 .
our result is compatible with the result @xcite @xmath63 found from moment sum rules for heavy
heavy correlators known to the same @xmath6 accuracy as in our analysis .
we observe an excellent agreement with the prediction of the rg - improved nnll analysis of the @xmath24 sum rule @xcite , @xmath64 our result agrees within 2@xmath65 with the pdg estimate @xmath66 we realize , however , a pronounced tension with the predictions of @xcite @xmath67 and @xcite @xmath68 based on sum rules for heavy heavy correlators calculated to @xmath20 accuracy .
as already noticed above , it seems unlikely that the @xmath20 correction may bring our result in agreement with the relatively low value of @xcite ; therefore , we expect that this tension will persist .
the origin of this disagreement requires further considerations .
we conclude by emphasizing that the properly formulated borel qcd sum rules for heavy light correlators provide a competitive tool for the reliable calculation of heavy - meson properties and for the extraction of basic qcd parameters by making use of the results from lattice qcd and the experimental data .
we point out that in the context of qcd sum rules based on correlation functions calculated at @xmath6 accuracy , eq .
( [ ourmb ] ) gives the appropriate value of the @xmath0-quark mass .
we thank andre hoang for helpful discussions . d.m .
was supported by the austrian science fund ( fwf ) , project no .
the work was supported in part by a grant for leading scientific schools 3920.2012.2 , and by fasi state contract no .
02.740.11.0244 .
30 j. beringer _
( particle data group ) , phys .
d * 86 * , 010001 ( 2012 ) .
p. dimopoulos _ et al .
_ ( etm collaboration ) , jhep * 1201 * , 046 ( 2012 ) .
n. carrasco _ et al . _
( etm collaboration ) , pos(lattice 2012)104 ( 2012 ) ; pos(ichep2012)428 ( 2012 ) .
v. gimenez , l. giusti , g. martinelli , and f. rapuano , jhep * 03 * , 018 ( 2000 ) . c. mcneile , c. michael , and g. thompson ( ukqcd collaboration ) , phys .
b * 600 * , 77 ( 2004 ) .
f. bernardoni _ et al . _
( alpha collaboration ) , nucl .
* 234 * , 181 ( 2013 ) .
j. h. khn and m. steinhauser , nucl .
b * 619 * , 588 ( 2001 ) .
k. g. chetyrkin _ et al .
_ , phys .
d * 80 * , 074010 ( 2009 ) .
c. mcneile _ et al . _
( hpqcd collaboration ) , phys .
d * 82 * , 034512 ( 2010 ) .
a. hoang , p. ruiz - femenia , and m. stahlhofen , jhep * 10 * , 188 ( 2012 ) .
w. lucha , d. melikhov , and s. simula , phys .
b * 687 * , 48 ( 2010 ) ; phys .
nucl . * 73 * , 1770 ( 2010 ) .
m. neubert , phys .
rept . * 245 * , 259 ( 1994 ) .
m. a. shifman , a. i. vainshtein , and v. i. zakharov , nucl .
b * 147 * , 385 ( 1979 ) .
t. m. aliev and v. l. eletsky , yad .
fiz . * 38 * , 1537 ( 1983 ) .
s. narison , phys .
b * 520 * , 115 ( 2001 ) .
m. jamin and b. o. lange , phys .
d * 65 * , 056005 ( 2002 ) .
w. lucha , d. melikhov , and s. simula , j. phys .
g * 38 * , 105002 ( 2011 ) .
s. narison , phys .
b * 718 * , 1321 ( 2013 ) .
k. g. chetyrkin and m. steinhauser , phys .
b * 502 * , 104 ( 2001 ) ; eur . phys .
j. c * 21 * , 319 ( 2001 ) .
h. na _ et al .
_ , phys .
d * 86 * , 034506 ( 2012 ) . c. mcneile
_ et al . _ ,
d * 85 * , 031503 ( 2012 ) .
a. bazavov _ et al .
_ ( fermilab lattice and milc collaborations ) , phys . rev .
d * 85 * , 114506 ( 2012 ) .
w. lucha , d. melikhov , and s. simula , phys .
d * 76 * , 036002 ( 2007 ) ; phys .
b * 657 * , 148 ( 2007 ) ; phys .
71 * , 1461 ( 2008 ) ; phys .
b * 671 * , 445 ( 2009 ) ; d. melikhov , phys .
b * 671 * , 450 ( 2009 ) .
w. lucha , d. melikhov , and s. simula , phys .
d * 79 * , 096011 ( 2009 ) ; j. phys .
g * 37 * , 035003 ( 2010 ) ; w. lucha , d. melikhov , h. sazdjian , and s. simula , phys .
d * 80 * , 114028 ( 2009 ) .
k. g. chetyrkin and m. steinhauser , phys . rev .
lett . * 83 * , 4001 ( 1999 ) ; nucl .
b * 573 * , 617 ( 2000 ) ; k. melnikov and t. van ritbergen , phys .
b * 482 * , 99 ( 2000 ) .
g. colangelo _ et al .
_ , eur .
j. c * 71 * , 1695 ( 2011 ) .
b. efron and r. j. tibshirani , _ an introduction to the bootstrap _ , monographs on statistics and applied probability , vol . * 57 * , crc press , 1993 .
w. lucha , d. melikhov , and s. simula , phys .
b * 701 * , 82 ( 2011 ) . | we prove that borel qcd sum rules for heavy
light currents yield very strong correlations between the @xmath0-quark mass @xmath1 and the @xmath2-meson decay constant @xmath3 namely , @xmath4 this fact opens the possibility of an accurate sum - rule extraction of @xmath1 by using @xmath5 as input . combining precise lattice qcd determinations of @xmath5 with our sum - rule analysis based on the three - loop @xmath6 heavy
light correlation function leads to @xmath7 |
in this section we provide a proof of eq .
( 2 ) of the main text .
our approach and terminology rely heavily on the classical text @xcite .
let @xmath207 . for @xmath208 , @xmath2
is assumed to have a degenerate steady state manifold , i.e. @xmath209 for small non - zero @xmath210 , some eigenvalues of @xmath211 ( may ) depart from @xmath212 .
the set of these eigenvalues is called the @xmath213-group since they cluster around the unperturbed eigenvalue , in this case @xmath212 , for small @xmath214 @xcite .
let @xmath215 be the projection associated to the @xmath213-group originating from the degenerate @xmath212 eigenvalue of @xmath2 ( whose associate projection is given by @xmath8 ) .
define also the projected liouvillian @xmath216 . a central result of @xcite states that both @xmath217 and @xmath218 are analytic in @xmath210 , i.e. their power series in @xmath210 have a finite radius of convergence .
since @xmath217 commutes with @xmath211 one clearly has @xmath219 .
we can now expand both @xmath217 and @xmath218 around @xmath208 .
accordingly we write @xmath220 where we defined @xmath221 and @xmath222 using @xcite and remembering that @xmath223 , one finds @xmath224 for example , in case the zero eigenvalue has no nilpotent part , as it happens in physical systems , one has @xcite @xmath225 and @xmath226 in eqns .
( [ deg - pert - th ] ) and ( [ eq : dp ] ) above , @xmath44 is the projected resolvent of _
@xmath131 _ related to the @xmath212 eigenvalue .
explicitly , if @xmath131 has the following jordan decomposition @xmath227 with @xmath228 projectors , @xmath229 nilpotents and @xmath230 , the projected resolvent is given by @xmath231.\ ] ] define further @xmath232 .
now we use the inequality @xmath233 with @xmath234 and @xmath235 to obtain @xmath236 therefore @xmath237 the proof is completed using triangle inequality and the bounds ( [ norms ] ) and ( [ eq : bound ] ) ( and setting @xmath40 ) , implying @xmath238 . note that this proof , together with the bound ( 2 ) in the main text , remains valid in a slightly more general setting where the eigenvalues @xmath47 of @xmath239 satisfy @xmath240 .
for example , in the extreme case of unitary dynamics where the eigenvalues are purely imaginary this result become essentially the standard adiabatic theorem as discussed in sec .
[ sec : hamiltonian - example ] , but intermediate cases are accounted for as well .
we now consider a case in which @xmath241 . performing the rescaling @xmath242 one
is led to analyze @xmath243 with @xmath244 .
the bounds in eq .
( [ norms ] ) become now @xmath245 reasoning as previously we now obtain @xmath246 with @xmath247 .
the norm used is the maximum singular value of the maps realized as matrices over @xmath248 parameter used are @xmath249 ( @xmath250 , @xmath251 ) , @xmath252 and @xmath133 .
the system has @xmath253 and @xmath254 .
[ fig : distance - multi ] ] as reminded in the previous section , our projection result eq .
( 2 ) of the main text , holds also when @xmath255 $ ] and in this case it simply amounts to a type of adiabatic theorem for closed quantum systems . to illustrate this fact we consider a system of @xmath256 spins interacting collectively with @xmath257 bosons and hamiltonian of atoms with the radiation field .
we restrict ourself to the space of only one boson or spin excitation hilbert space _ @xmath258 _ , @xmath259 @xmath260 where @xmath261 , @xmath262 has all spins down , @xmath263 one boson in mode @xmath264 and @xmath265 is the boson vacuum .
hamiltonian @xmath266 admits the following dark states ( @xmath267 ) , @xmath268 , with @xmath269 @xcite .
ssm includes all the states built over the dark state manifold all of which are decoherence - free at zero temperature @xcite .
let us now introduce an hamiltonian perturbation which conserves the total number of excitations , such as @xmath270 , and the corresponding superoperator @xmath271 $ ] .
projected hamiltonian over the dark - state manifold turns out to be @xmath272 $ ] with @xmath273 , shows how eq .
( 2 ) of the main text is fulfilled in this unitary case as well . | it has been recently realized that dissipative processes can be harnessed and exploited to the end of coherent quantum control and information processing . in this spirit
we consider strongly dissipative quantum systems admitting a non - trivial manifold of steady states .
we show how one can enact adiabatic coherent unitary manipulations e.g. , quantum logical gates , inside this manifold by adding a weak , time - rescaled , hamiltonian term into the system s liouvillian .
the effective long - time dynamics is governed by a projected hamiltonian which results from the interplay between the weak unitary control and the fast relaxation process .
the leakage outside the manifold entailed by the hamiltonian term is suppressed by an environment - induced symmetrization of the dynamics .
we present applications to quantum - computation in decoherence - free subspaces and noiseless subsystems and numerical analysis of non - adiabatic errors . _
introduction: _ weak coupling to the environmental degrees of freedom is often regarded as one of the essential prerequisites for realizing quantum information processing .
in fact decoherence and dissipation generally spoil the unitary character of the quantum dynamics and induce errors into the computational process . in order to overcome such an obstacle a variety of techniques
have been devised including quantum error correction @xcite , decoherence - free subspaces ( dfss ) @xcite and noiseless subsystems ( nss ) @xcite .
however , it has been recently realized that dissipation and decoherence may even play a positive role to the aim of coherent quantum manipulations .
indeed , it has been shown that , properly engineered , dissipative dynamics can in principle be tailored to enact quantum information primitives such as quantum state preparation @xcite , quantum simulation @xcite and computation @xcite . in this letter
we investigate the regime where the coupling of the system to the environment is very strong and the open system dynamics admits a non - trivial steady state manifold ( ssm ) .
we will show how , in the long time limit , unitary manipulations e.g. , quantum gates , inside the ssm can be enacted by adding a time - rescaled hamiltonian acting on the system only .
this coherent dynamics is governed by a sort of projected hamiltonian which results from the non - trivial interplay between the weak unitary control term and the strong dissipative process .
the latter effectively renormalizes the former by continuously projecting the system onto the manifold and adiabatically decoupling the non steady states .
several of the results of this letter can be regarded as a rigorous formulation and significant extension of ideas first explored in @xcite and @xcite .
we would also like to point out the relation with techniques relying on some type of quantum zeno dynamics @xcite .
the latter can be in fact regarded as a special case of our general result ( [ projection - th ] ) .
this letter is organized as follows : we first set the stage of our analysis and describe the main theoretical ideas and results .
we then discuss in detail , aided by numerical simulations , a few different models demonstrating dissipation - assisted computation over ssms comprising decoherence - free subspaces and noiseless - subsystems . for the reader s convenience we have collected background technical material and all the mathematical proofs in @xcite .
_ evolution of steady state manifolds [ stage ] :
_ in the following @xmath0 will denote the hilbert space of the system and @xmath1 the algebra of linear operators over it .
a time - independent liouvillian super - operator @xmath2 acting on l@xmath3 is given .
the ssm of @xmath4 comprises all the quantum states @xmath5 contained in the kernel @xmath6 of @xmath7 we will denote by @xmath8 ( @xmath9 ) the spectral projection over ker@xmath10 ( the complementary subspace of ker@xmath11 ) .
one has that @xmath12 and @xmath13 notice also that @xmath8 may not be hermitian .
the liouvillian @xmath2 is also assumed to be such that : * a ) * the equation @xmath14 defines a semi - group of trace - preserving positive maps with @xmath15 @xcite ; * b ) * the non zero eigenvalues @xmath16 of @xmath2 have negative real parts i.e. , the ssm is attractive . in this case @xmath17 on top of the process described by @xmath2 we now add a control hamiltonian term @xmath18 $ ] where @xmath19 the time @xmath20 is a scaling parameter that , in the spirit of the adiabatic theorem , will be eventually sent to infinity .
if @xmath21 then @xmath22 the basic dynamical equation we are going to study is the following @xmath23 notice that even if we are _ not _ assuming that @xmath24 is of the lindblad type @xcite i.e. , the @xmath25 being completely positive ( cp ) maps , our basic equation ( [ basic - eq ] ) is time - local and in this sense markovian .
the system is also strongly dissipative in the sense that , for large @xmath26 the dominant process is the one ruled by @xmath24 .
if the system is initialized in one of its steady states , on general physical grounds one expects the system , for small @xmath27 to stay within the ssm with high probability .
however , for @xmath2 with a multi - dimensional ssm a non - trivial internal dynamics may unfold . in order to gain physical insight on this phenomenon we would like first to provide a simple argument based on time - dependent perturbation theory .
eq .
( [ basic - eq ] ) immediately leads to @xmath28 for the evolution semi - group .
we can formally solve this equation by @xmath29 from which , by iteration , it follows the standard dyson expansion with respect the perturbation @xmath30 considering terms up the first order applied to @xmath31 and inserting the spectral resolution @xmath32 one obtains @xmath33 where @xmath34 is a pseudo - inverse of @xmath2 i.e. , @xmath35 the norm of the third term is upper bounded by @xmath36 uniformly in @xmath37 it follows that scaling @xmath38 by @xmath39 over a total evolution time @xmath40 the first and second term above are @xmath41 while the third one the only one involving transitions outside the steady state manifold
is @xmath42 this demonstrates that , at this order of the dyson expansion , the dynamics is ruled by an effective generator @xmath43 whose emergence is basically due to a fermi golden rule mechanism .
moreover , by looking at the structure of the liouvillian pseudo - inverse @xmath44 @xcite , we see that @xmath45 where @xmath46 and the @xmath47 s are the non - vanishing eigenvalues of @xmath2 @xcite .
the meaning of this quantity is that the time - scale @xmath48 sets a lower bound to the relaxation time of the irreversible process described by @xmath11 since @xmath49 if no nilpotent blocks are present in the spectral resolution of @xmath24 @xcite the leakage outside the ssm becomes negligible when @xmath50 namely when the time - scale @xmath20 is much longer that the relaxation time @xmath48 i.e. , dissipation is much faster than the coherent part of the dynamics .
system specific examples of ( [ op - time ] ) will be given later when concrete applications are discussed .
now we present our main technical result on the projected dynamics over ssms ( see @xcite for the proof s details ) : @xmath51 where @xmath52 and @xmath53 denotes the evolution over @xmath54 $ ] generated by @xmath55 it should be stressed that ( [ projection - th ] ) is based just on degenerate perturbation theory for general linear operators @xcite .
in particular , it does _ not _ rely on the assumption that @xmath2 can be cast in lindblad form @xcite or on the ssm structure described in @xcite .
an immediate corollary of ( [ projection - th ] ) is that @xmath56 , namely the probability of leaking outside of the ssm , induced by the unitary term @xmath57 for large @xmath26 is smaller than @xmath58 the constant @xmath59 controls the strength of the deviations from the ideal adiabatic behavior at finite @xmath20 [ the rhs of ( [ projection - th ] ) ] and it can be related to the spectral structure of @xmath44 .
roughly speaking , one expects @xmath60 and therefore violations of adiabaticity , to increase when the dissipative gap @xmath61 decreases .
however , a subtler interplay between the gap with the matrix elements of @xmath62 may play an important role here as well in the information - geometry of ssm @xcite .
let us now turn to the structure of the effective generator @xmath63 of course it crucially depends on the projection @xmath8 that in turn depends on the nature of @xmath11 here below we discuss two ( non mutually exclusive ) cases .
their physical relevance relies on the importance , both theoretical and experimental , of the concepts of decoherence - free subspaces @xcite and noiseless - subsystems @xcite in quantum information . *
i ) * the most general dissipative generator @xmath2 of a markovian quantum dynamical semi - group @xmath64 can be written as @xmath65 where @xmath66 is a cp map @xmath67 is the dual map i.e. , @xmath68 @xcite .
we now assume that @xmath66 is trace - preserving ( @xmath69 ) _ and _ unital ( @xmath70 ) .
under these assumptions whence ker@xmath10 coincides with the set of fixed points of @xmath71 the latter is known to be the commutant @xmath72 @xcite of the _ interaction algebra _
@xmath73 generated by the kraus operators @xmath74 and their conjugates @xcite . from @xcite it follows that the ssm of @xmath2 is @xmath75-dimensional and is given by the convex hull of states of the form @xmath76 where @xmath77 is a state over the noiseless - subsystem factor @xmath78 if , for some @xmath79 @xmath80 one has that the corresponding @xmath81 is a dfs and the ssm contains pure states .
conversely , if @xmath82 then no pure states are in the ssm .
a characterization of the algebraic structure of ssms for @xmath2 s of the lindblad form @xcite is provided in @xcite .
now @xmath8 is the projection onto the commutant algebra @xcite and one can check that @xmath83 $ ] where @xmath84 @xcite . by definition @xmath85=0 $ ] for all the unitaries in @xmath86 namely the effective dynamics
admits as a symmetry group the full - unitary group of the interaction algebra @xmath87 .
this means that the renormalization process @xmath88 amounts to an environment - induced _ symmetrization _ of the dynamics @xcite . from @xcite
it also follows that @xmath89 has a non - trivial action just on the noiseless - subsystems of @xmath90 the symmetrization process dynamically decouples the system from the noise process driven by operators in @xmath87 @xcite .
* ii ) * suppose there exists a subspace @xmath91 such that @xmath92 in particular @xmath93 i.e. , @xmath94 is a dfs @xcite for the unperturbed @xmath11 if also @xmath95 hold for all @xmath96 and @xmath97 a simple calculations shows that @xmath98,$ ] where @xmath99 is the orthogonal projection over @xmath94 @xcite . remarkably , in all cases * i)**ii ) * above we see that the induced ssm dynamics @xmath100 is _ unitary _ and governed by a dissipation - projected hamiltonian .
qualitatively : this coherent dynamics results from the interplay between the weak ( slow ) hamiltonian @xmath101 and the strong ( fast ) dissipative term @xmath11 the former induces transitions out of the ssm while the latter projects the system back into it on much faster time - scale . as a result non of the liouvillian are adiabatically decoupled from the dynamics up to contributions @xmath102 .
we would like now to make a few important remarks .
* 1 ) * by defining @xmath103 eq . ( [ basic - eq ] ) gives rise to a dynamical equation of the form @xmath104 where @xmath105 and @xmath106 namely in this rotated frame @xmath107 evolves in a time - dependent bath described by @xmath108 . this establishes a connection of the present approach to the one with time - dependent baths in @xcite and @xcite .
smallness of @xmath109 in the picture ( [ basic - eq ] ) translates into slowness of the bath time - dependence in the rotated frame . *
2 ) * the environment - induced renormalization @xmath110 is _ not _ an algebra homomorphism ; this implies that the algebraic structure of a a set of projected hamiltonians may differ radically from the algebraic structure of the original ( unprojected ) ones
. in particular commuting ( non - commuting ) @xmath111 s may be mapped onto non - commuting ( commuting ) @xmath112 this implying a potential increase ( decrease ) of their ability to enact quantum control @xcite .
notice that also the hamiltonian locality structure may be affected by the projection e.g. , a @xmath113-local @xmath109 may give rise to a @xmath114-local @xmath115 the dissipative technique here discussed might then be exploited to effectively generate non - local interactions out of simpler ones in a fashion similar to perturbative gadgets @xcite ( see also @xcite ) . *
3 ) * any extra term @xmath116 , either hamiltonian or dissipative , in the liouvillian such that @xmath117 will not contribute to the effective dynamics ( [ projection - th ] ) in the limit in which @xmath2 dominates . for example in the case * ii ) * discussed in the above the projected dynamics does not change by perturbing @xmath109 with any extra hamiltonian term @xmath118 such that @xmath119 and @xmath120 [ here @xmath121 is the projector @xmath122 of the @xmath123th summand in @xcite ] .
the projected dynamics has a degree of resilience against perturbations that are eliminated by the environment - induced symmetrization . *
4 ) * if the interaction algebra @xmath73 in * ii ) * is an abelian then from @xcite one finds @xmath124 this shows that quantum zeno dynamics and the associated control and computation techniques of refs .
@xcite can be regarded as a _ special _ case of the projection phenomenon described by eq .
( [ projection - th ] ) .
* 5 ) * when @xmath125 the dyson series for @xmath126 shows that , @xmath127 this means that the dynamics inside the ssm is now ruled by the second - order effective generator @xmath128 ( up to errors @xmath129 ) .
this dynamics is in general non unitary and its effective relaxation time can be roughly estimated by @xmath130 notice the counterintuitive fact that the stronger the dissipation outside the ssm the _ weaker _ the effective one inside @xcite . left panel : dfs example with @xmath131 given by eq .
( [ l0-four ] ) .
we used @xmath132 and @xmath133 ( see text for details ) .
right panel : example for noiseless subsystem with parameters @xmath134 ( see text ) .
the norm used is the maximum singular value of the maps realized as matrices over @xmath135 .
the linear fits are obtained using the four most significant points .
[ fig : j_zero],title="fig:",width=151,height=113 ] left panel : dfs example with @xmath131 given by eq .
( [ l0-four ] ) .
we used @xmath132 and @xmath133 ( see text for details ) .
right panel : example for noiseless subsystem with parameters @xmath134 ( see text ) .
the norm used is the maximum singular value of the maps realized as matrices over @xmath135 .
the linear fits are obtained using the four most significant points .
[ fig : j_zero],title="fig:",width=151,height=113 ] _ unitaries over a dfs . :
_ we show here how to perform coherent manipulations on a logical qubit built upon the ssm of four qubits which comprises a dfs @xcite .
consider the following unperturbed liouvillian @xmath136 where @xmath137 are collective spin operators and @xmath138 decoherence rates .
the interaction algebra @xmath87 generated by the @xmath139 s is the algebra of permutation invariant operators @xcite .
therefore , from * i ) * , it follows that ker@xmath2 has the structure @xcite where @xmath140 is now a total angular momentum label , @xmath141 and the @xmath142 s are the dimensions of irreps of the permutation group @xmath143 @xcite . for @xmath144
the one - dimensional @xmath145 representation shows with multiplicity two .
if we denote by @xmath94 this two - dimensional subspace the conditions in * ii ) * are met .
let us denote with @xmath99 the projector onto @xmath94 .
it is known that one can construct universal set of gates in similar dfss ( see e.g. @xcite ) when the dynamics is entirely contained in the dfs . here
we show that coherent manipulation is possible also when the dynamics leaks out of the dfs .
consider for example the following hamiltonian perturbations @xmath146 and @xmath147 one can check that in the logical space @xmath94 , such hamiltonians reduce to elementary pauli operations , i.e. @xmath148 .
we now build the perturbed liouvillians @xmath149 $ ] , @xmath150 , let us also denote @xmath151\mathcal{p}_{0}$ ] with @xmath152 free parameter . in fig .
[ fig : j_zero ] left panel we show a numerical experiment confirming our general theorem eq .
( [ projection - th ] ) for such @xmath153 . in the logical qubit space ,
the effective evolution @xmath154 is a unitary evolution @xmath155 with @xmath156 , and one can easily generate any unitary in @xmath157 by concatenating such gates . moreover , the bound in eq .
( [ projection - th ] ) implies that , for any vectors @xmath158 in the logical space @xmath94 , @xmath159 , showing that effectively , one can generate unitary gates on the logical qubit space @xmath94 up to an error @xmath160 in view of remark * 3 ) * one is allowed to add to @xmath153 any perturbation @xmath161 satisfying @xmath162 , and still obtain the same unitary gates @xmath163 within an error @xmath164 albeit with a possibly different @xmath59 @xcite .
in @xcite we show the stability of this dynamics also against certain _ dissipative _ perturbations of @xmath11 fig .
[ fig : j_zero ] ( left panel ) shows that the whole @xmath165-dimensional ssm is evolving unitarily in the long time limit . to illustrate our results
let us consider the experimental dfs system studied in @xcite consisting of a couple of trapped @xmath166 ions subject to collective dephasing [ @xmath167 in ( [ l0-four ] ) ] . in this case
@xmath168 and ( assuming a similar relaxation time for a four qubits system ) eq .
( [ op - time ] ) and fig .
[ fig : j_zero ] show that for @xmath169 one should observe small deviations of the effective dynamics from unitarity .
_ unitaries over noiseless subsystem: _ next we discuss dissipation - assisted computation over noiseless subsystems @xcite .
the liouvillian is in the class previously discussed , @xmath170 , taking @xmath171 where @xmath139 are again collective spin operators . for generic @xmath172 s
the ssm coincides with one of the former examples i.e , rotationally invariant state .
the latter for an odd number @xmath173 of spins , contains only mixed states . as perturbation
we use the following hamiltonian @xmath174 and the full liouvillian reads @xmath175 $ ] . again
one observes an effective unitary evolution , up to an error @xmath176 ( see fig .
[ fig : j_zero ] right panel ) over the full five - dimensional ssm ; in particular , this construction can be seen as a scheme to enact dissipation - assisted control over the noiseless - subsystem @xmath177 factor @xcite . in @xcite noiseless - subsystems
have been realized in a nmr system comprising three nuclear spins subject to collective ( artificial ) noise ; for a relaxation time @xmath178 the noiseless encoding provides an advantage .
fig [ fig : j_zero ] shows that setting the operation time , say at @xmath179 then effective dynamics over the nss becomes very close to a unitary one .
finally we would like to stress that the markovian form ( 1 ) is just sufficient ( and mathematically convenient ) to prove the existence of an effective projected dynamics , _ not _ necessary .
the spin - boson hamiltonian discussed in @xcite indicates that the relevant dynamical mechanism is the existence of a _
strong _ system - bath coupling that adiabatically - decouples non steady - states from the dynamics . _
conclusions: _ [ conclusions ] in this letter we have shown how an effective unitary dynamics can be enacted over the manifold of steady states of a strongly - dissipative system . the strategy is to introduce a small time - rescaled hamiltonian term in the system s liouvillian largely dominated by the dissipative processes . in the long time
limit the dynamics leaves the steady state manifold invariant and becomes unitary up to a small error whose strength is connected to the liouvillian relaxation time , and total operation time .
the effective hamiltonian ruling the long time - dynamics is shaped by the continuous interplay of the weak hamiltonian control with the the fast relaxation process that adiabatically decouples non - steady states .
this effective projected hamiltonian , in some cases , can be seen as a symmetrized form of the bare one and it is robust against all perturbations , dissipative or hamiltonian , that are filtered out by this environment - induced symmetrization . to illustrate these ideas
we have shown how to realize quantum gates on manifolds comprising decoherence - free subspaces @xcite as well as noiseless subsystems @xcite . in all these cases
we have also provided a numerical estimate of the deviations from the ideal long - time unitary behavior and the actual , finite time , one .
agreement with the theoretical prediction ( [ projection - th ] ) is found in all cases .
the results of this letter seem to suggest the intriguing possibility of fighting quantum decoherence by introducing even more quantum decoherence .
this work was partially supported by the aro muri grant w911nf-11- 1 - 0268 and by nsf grant phy- 969969 .
useful input from s. garnerone , j. kaniewski , d. lidar , i. marvian and s. muthukrishnan is gratefully acknowledged .
10 .
d. a. lidar , t. brun eds .
cambridge univesity press ( 2013 ) p. zanardi and m. rasetti , phys .
rev .
lett . * 79 * , 3306 ( 1997 ) ; p. zanardi , phys . rev .
a * 57 * , 3276 ( 1998 ) ; d. a. lidar , i. l. chuang , and k. b. whaley , phys .
rev .
lett . * 81 * , 2594 ( 1998 ) .
d. kielpinski et al , science * 291 * , 1013 ( 2001 ) ; e. knill , r. laflamme , and l. viola , phys . rev
. lett . * 84 * , 2525 ( 2000 ) ; p. zanardi , phys . rev .
a * 63 * , 012301 ( 2000 ) .
l. viola et al , science * 293 * , 2059 ( 2001 ) m. j. kastoryano , f. reiter , and a. s. sorensen , phys .
rev .
lett . * 106 * , 090502 ( 2011 ) .
j. t. barreiro , m. mller , p. schindler , d. nigg , t. monz , m. chwalla , m. hennrich , c. f. roos , p. zoller , and r. blatt , nature ( london ) * 470 * , 486 ( 2011 ) .
k. stannigel , p. hauke , d. marcos , m. hafezi , s. diehl , m. dalmonte , p. zoller , phys .
rev .
lett . * 112 * , 120406 ( 2014 ) f.verstraete , m.m.wolf,and j.i.cirac , nat.phys.*5 * , 633 ( 2009 ) .
a. carollo , m. f. santos , v. vedral , phys .
rev .
lett . * 96 * , 020403 ( 2006 ) o. oreshkov , j. calsamiglia , phys . rev . lett . * 105 * , 050503 ( 2010 ) g. a. paz - silva , a. t. rezakhani , j. m. dominy , and d. a. lidar , phys . rev . lett .
* 108 * , 080501 ( 2012 ) f. schfer , i. herrera , s. cherukattil , c. lovecchio , f.s .
cataliotti , f. caruso and a. smerzi , nat . commun . * 5 * , 3194 ( 2014 ) d. burgarth , p. facchi , v. giovannetti , h. nakazato , s. pascazio , k. yuasa , arxiv:1403.5752 see appendix in this paper , unless otherwise stated the norms for superoperators @xmath180 will be @xmath181 for semigroups
@xmath182 of cp - maps one has @xmath183 @xmath184 will denote the standard operator norm for @xmath185l@xmath186 g. lindblad , commun .
math .
phys . * 48 * , 119 ( 1976 ) t. kato , _ perturbation theory for linear operators _ , springer 1995 b. baumgartner and h. narnhofer , j. phys .
a * 41 * , 395303 ( 2008 ) l. banchi , p. giorda , p. zanardi , phys .
rev .
e * 89 * , 022102 ( 2014 ) by definition @xmath187=0,\,\forall o\in{\cal a}\}.$ ] standard structure - theorems for @xmath188-algebras imply that @xmath189 where @xmath140 labels the irreducible representations of @xmath73 with dimension @xmath190 and multiplicity @xmath142 @xcite . in this case @xmath191 where the haar - measure integral is performed over the unitary group of the algebra @xmath192 d. w. kribs , proc .
edin .
math .
soc .
* 46 * ( 2003 ) if @xmath193 one has : @xmath194)=-i{\cal p}_{0}(\tilde{k}x)+i{\cal p}_{0}(x\tilde{k})=-i{\cal p}_{0}(\tilde{k})x+ix{\cal p}_{0}(\tilde{k})=-i[{\cal p}_{0}(\tilde{k}),\ , x].$ ] where we used e.g. , @xmath195 valid as @xmath196 is in @xmath72 @xcite .
p. zanardi , phys .
lett .
a 258 * 77 * ( 1999 ) ; p. zanardi , phys .
rev .
a * 60 * 729 ( 1999 ) ; l. viola , e. knill , s. lloyd , phys .
rev .
lett . *
82 * , 2417 ( 1999 ) let @xmath197 and consider @xmath198).$ ] one can write @xmath199 where @xmath200 therefore @xmath201 where we have used the properties of @xmath8 assumed in the main text and @xmath202 considering now the @xmath203 term in the commutator in the same way one finds @xmath204)=[\pi\tilde{k}\pi,\ , x]$ ] for all @xmath205 s. p. jordan and e. farhi , phys . rev .
a * 77 * , 062329 ( 2008 ) ; c. m. herdman , k. c. young , v. w. scarola , m. sarovar , and k. b. whaley , phys .
rev .
lett . * 104 * , 230501 ( 2010 ) ; b. antonio and s. bose , phys . rev .
a * 88 * , 042306 ( 2013 ) for @xmath173 qubits @xmath206^{-1}.$ ] j. kempe , d. bacon , d.a .
lidar , and k.b .
whaley , phys .
rev . a , * 63 * , 042307 , ( 2001 ) l. campos venuti _ et
. al , _ in preparation |
parametrically forced surface waves arising as a result of the faraday instability have provided an excellent opportunity to study nonlinear pattern formation .
one of the special features of this system is that the system size relative to the basic correlation length can be varied so that both the large aspect ratio and small aspect ratio limits can be explored . at large aspect ratio ,
all of the classic ordered patterns have been found , including stripes , hexagons , and squares ; additional exotic structures such as quasicrystalline and superlattice patterns have also been found , as well as secondary instabilities giving rise to spatiotemporal chaos .
extensive references can be found in @xcite .
the case of small aspect ratio has also been studied in rectangular and circular containers .
typically the wave patterns found near onset are either normal modes of the container or symmetrical combinations of these modes @xcite .
for example , in the circular case the normal modes are bessel functions of the radius multiplied by sinusoidal functions of the azimuthal angle .
the effects of container shape can be either a nuisance or a benefit depending on one s point of view .
one example of the usefulness of considering container geometry is the study by lane et al .
@xcite in which the conceptual differences between square symmetry and square geometry were elucidated . on the other hand ,
finite size effects impede efforts to utilize amplitude equations to describe the wave dynamics .
the influence of the container shape is also a fundamental issue in the field of quantum chaos .
there is known to be a close correspondence between certain finite quantum systems ( or analogous systems supporting classical waves ) and their particle ( or ray - optic ) counterparts @xcite .
of particular interest are non - integrable quantum systems with classical counterparts that are chaotic , such as the billiard formed from two semicircles separated by two straight edges . for almost all initial conditions , a particle launched inside such
a billiard will exhibit sensitive dependence on initial conditions - the hallmark of chaos .
experimental , numerical and theoretical studies have shown that the statistical behavior of the wavefunctions of the quantum or wave version of this system is distinctly different from the behavior for integrable " geometries such as a square or circle @xcite .
most notably , regions of high amplitude in the wavefunctions - called scars - are found along some of the deleted word periodic orbits of the classical counterpart @xcite .
effects of this type were explored to a limited extent using parametrically forced surface waves by blmel et al .
their experiment utilized water as a working fluid and high frequency excitation .
they reported observations of scarlets " , that are ridge like structures consistent with a random superposition of plane waves @xcite ( but are not located along periodic orbits . ) on the other hand , no clear evidence for the simpler scarred " wavefunctions was given .
the possible effects of hydrodynamic nonlinearity on the utility of the ray optics approach has also not been discussed .
nonlinearity is in principle important , since even infinitesimally above the onset of instability , saturation of the wave amplitude is produced by nonlinear effects . in this paper we first present observations of the spatial modes of faraday waves in a finite non - integrable geometry , close to onset where nonlinearity is as weak as possible , and the waves might be usefully described by quasi - classical ( ray optics ) methods .
a low viscosity fluid in a stadium shaped container is used for this purpose . scarred patterns that resemble the computed eigenfunctions of the stadium geometry
are clearly evident , but some of the linear eigenfunctions are apparently suppressed .
the relative suppression of certain modes has been plausibly explained by agam and altshuler @xcite in terms of higher dissipation rates for those modes near the boundaries in comparison with the modes that are observed .
we then consider the evolution of the wave patterns as the degree of nonlinearity is increased .
transitions between modes are found at some driving frequencies , along with a general increase in spatial complexity .
the scars that are characteristic of the linear eigenmodes are often evident substantially above onset .
finally , we consider the development of spatiotemporal chaos ( stc ) in the stadium geometry . though the onset of stc is strongly dependent on the excitation frequency , the boundaries continue to play a large role , leading for example to coherent oscillations between symmetric and asymmetric states , a phenomenon that we study using the concept of pattern entropy .
a fluid layer with a free surface is subjected to an oscillatory vertical acceleration of amplitude @xmath0 .
the surface is flat until a critical acceleration @xmath1 is reached , at which point the surface becomes unstable and standing wave patterns are observed that oscillate at half the driving frequency .
the threshold acceleration depends on the frequency and the viscosity of the fluid .
it is convenient to define a dimensionless driving parameter @xmath2 that measures the departure from onset and hence the degree of nonlinearity .
the patterns are time - independent for a range of positive @xmath3 but eventually a secondary instability gives rise to spatiotemporal chaos . in this section ,
we briefly discuss the linear inviscid theory , the effect of viscosity , and the role of nonlinearity , as they pertain to the present investigation .
the linear stability theory for faraday waves was first developed by benjamin and ursell @xcite .
we summarize it here because the quantum / classical correspondence occurs for linear waves .
they started from the euler ( inviscid ) equation of motion and the continuity equation for an ideal fluid with a free surface in an oscillating gravitational field , and simplified the equations by retaining only the linear terms appropriate for small amplitude waves .
the surface deformation @xmath4 as a function of spatial coordinate @xmath5 and time @xmath6 may be written as a superposition of normal modes @xmath7 with coefficients
@xmath8 : @xmath9 where @xmath7 is a complete orthogonal set of eigenfunctions of the helmholtz equation , @xmath10 the sidewall boundary condition is imposed by setting the normal component of the velocity of the fluid at the wall to zero .
this leads to a quantization condition on @xmath11 ( the wavenumber ) .
in addition , @xmath11 satisfies the dispersion relation which relates the frequency of oscillation @xmath12 of the fluid to the wavenumber : @xmath13 where , @xmath14 is the fluid density , @xmath15 is the surface tension , @xmath16 is the mean fluid depth , and @xmath17 is the gravitational acceleration . in our experiment , the wave number is sufficiently large so that the surface tension term is much greater than the gravity term .
the hyperbolic tangent factor is close to unity since @xmath18 is large . the time dependent amplitudes @xmath19 of these normal modes satisfy the mathieu equation : @xmath20 an instability occurs and the amplitude @xmath19 grows exponentially in time when the eigenvalue is in a band ( known as the stability tongue ) such that the frequency of oscillation of the fluid is half the driving frequency
. the instability occurs at arbitrarily small driving amplitude in the absence of viscosity .
damping , which is provided by a number of distinct mechanisms in addition to bulk viscosity , can be included by means of a phenomenological linear damping term as reviewed in ref.@xcite .
though treating damping in this way may not be fully adequate , the main effect is to reduce the width of the stability tongue in parameter space and raise the critical threshold to a finite amplitude .
a proper theoretical treatment of instability in the viscous case has been given in ref .
@xcite , where the shapes of the computed stability boundaries were presented .
if the acceleration @xmath0 is slightly higher than @xmath1 , all modes in a band ( @xmath21 ) are accessible and can be excited . the wavenumber width @xmath22 of the stability band for small @xmath3 has been estimated @xcite to be : @xmath23 where @xmath24 is the kinematic viscosity of the fluid . for a suitable choice of @xmath3 , @xmath14 , and @xmath12 , and
assuming no interaction between modes , one then expects to find either single mode patterns or superpositions of a few modes whose thresholds lie in the window @xmath25 .
the cumulative number of eigenvalues of the helmholtz equation @xmath26 is related to the geometry and is given by : @xmath27 where @xmath28 is the area , @xmath29 is the perimeter of the stadium , and the negative or positive sign corresponds to dirichlet or neumann boundary condition respectively @xcite . at high @xmath30 ,
the perimeter term is negligible compared to the area term .
taking @xmath31 and using eqs . 5 , 6 with an area @xmath28 that is appropriate to the experiments reported here , one can estimate that the typical number of accessible modes is about 8 for a driving frequency of 70 hz .
what are the effects of nonlinearity ?
a nonlinear theory that describes regular faraday wave patterns in large containers rather well has been given by zhang and vials @xcite and chen and vials @xcite .
in this theory , an evolution equation is determined for the time derivative of the amplitude of a typical fourier mode @xmath32 of the interfacial deformation .
it may be expressed in the form @xmath33 where @xmath34 is a slow time variable , the linear term is due to the basic instability , the cubic self - interaction term produces saturation of the wave pattern , and the coupling terms to other modes ( which depend on their relative angle @xmath35 ) are also of cubic order .
the constants have been computed , and the ratio @xmath36 is of order unity and independent of @xmath3 .
this implies that coupling effects between the accessible modes may be substantial .
the theory was able to explain the striking cascade of 2n - fold patterns discovered by binks and van de water @xcite .
it also explains semi - quantitatively the appearance of striped , square and hexagonal patterns observed in experiments using viscous fluids in large containers @xcite .
however , the amplitude equation is variational , and is only appropriate near onset .
it can not describe nonuniform patterns , secondary instabilities , or spatiotemporal chaos .
an earlier approach that allowed spatially varying patterns was given by milner @xcite .
the amplitude equations also ignore the effects of the boundary . for slightly viscous fluids in small containers , a large fraction of
the dissipation occurs in the boundary layer and can in fact be the leading cause of dissipation @xcite . in work stimulated by the experiments reported here ,
agam and altshuler @xcite show that the dissipation near the boundary depends strongly on the nature of the mode .
the apparatus is similar to that used by gluckman _
et al . _ in ref .
@xcite . fig
. 1 shows a schematic diagram of the experimental setup .
the stadium shaped container made of delrin has the following dimensions : depth @xmath37 cm , radius of semicircles @xmath38 cm , and length of straight edge @xmath39 cm .
the top and bottom plates of the container are made of glass to allow the transmission of light .
the fluid is silicone oil of kinematic viscosity @xmath40 , chosen for its stable surface tension and good wetting characteristics . to minimize meniscus waves ,
a brim full boundary condition was prepared by machining a ledge in the boundary at the same height as the fluid .
therefore the fluid meets the ledge at @xmath41 . by maintaining the fluid under brim full conditions ,
the fluid surface is pinned to the ledge and boundary dissipation is reduced .
this situation has been modeled as a dirichlet boundary condition ( @xmath42 ) @xcite .
the container is rigidly attached to an electromagnetic shaker ( vibration test systems model 40c ) and the acceleration is measured with an accelerometer .
the apparatus is placed within a temperature - controlled environment .
the driving frequency is selected to be greater than 55 hz to be in the capillary wave limit , but less than 75 hz to prevent the density of modes from becoming too high .
the patterns are imaged with shadowgraph techniques .
the specific implementation is discussed in depth in ref.@xcite .
light from an expanded and collimated incident beam is collected and imaged onto a ccd ( charge - coupled device ) video camera via a large collecting lens and the camera lens .
the resulting images can be interpreted by considering which rays of light reach the ccd plane after passing through the fluid .
the relatively small aperture of the camera lens restricts the rays that reach the ccd .
all the rays that leave the fluid surface at an angle measured from the normal that is greater than a critical angle ( typically about @xmath43 radians ) are blocked .
since the critical angle is so small , light is collected only from the nearly horizontal regions of the wave surface .
therefore , the bright regions in an image corresponds to local extrema or antinodes of the wave pattern .
images are averaged over one video frame , 1/30 s , which is more than a full cycle of the standing waves .
the imaging process is nonlinear in the wave height , but a quantitative model for the measured intensity was presented and tested in ref . @xcite .
we made a survey of the time - independent wave patterns near onset over the range 55 to 65 hz by changing the frequency in 0.1 hz steps . in order to obtain useful statistics for the surface wave patterns ,
a systematic procedure was followed : for each selected frequency , @xmath1 was first measured to within 0.1% and then @xmath3 was raised to 0.01 , the smallest value that could be maintained accurately .
the threshold @xmath1 is @xmath44 at @xmath45 hz and increases weakly with frequency . a sequence of images from the survey for driving frequency @xmath46 between 60.1 hz and 62.8 hz , and with an approximate spacing of 0.4 hz ( i.e. every fourth image ) , is shown in fig .
this spacing is comparable to the experimentally observed increment ( 0.3 hz ) typically required to obtain a distinctly different pattern in this frequency range ; it is greater then the computed mean level spacing ( about 0.1 hz in this frequency range as estimated from eqns . 5 and 6 . )
most of the observed patterns show the reflection symmetries of the stadium .
regions of large amplitude are often located along lines that would form periodic orbits of the classical analog .
since these regions are similar to those found in other numerical and experimental investigations @xcite , we refer to the patterns containing such enhancements as scarred patterns .
we compare the observed patterns with numerically computed eigenstates of the laplace operator for the stadium geometry , obtained for comparable mean wavenumber using an algorithm due to heller @xcite . sample computed eigenstates ( selected from a large number of distinct patterns )
are shown in fig .
we find that some of the computed states resemble observed patterns .
on the other hand , a one - to - one correspondence for sequences of eigenstates was definitely not observed .
furthermore , certain computed eigenstates that occur frequently such as the whispering gallery " mode ( fig .
3d ) were not observed in the full experimental frequency range .
interestingly , most of the observed symmetric patterns resemble one of three basic classes of eigenstates shown in fig .
3(a - c ) .
for instance , figs .
2(a , d ) resemble the bouncing ball eigenstate fig .
3(a ) ; figs . 2(g , h ) are close to the longitudinal eigenstate of fig .
3(b ) ; and figs .
2(b , e ) are a combination of the longitudinal and bowtie eigenstates of figs .
3(b , c ) .
it is noteworthy that among the observed patterns are states such as fig .
2(c ) that do not have the reflection symmetries of the stadium .
we refer to these as disordered patterns .
table i summarizes the percentages of the onset patterns that were visually judged to approximate particular computed scarred eigenstates in the frequency range 55 to 65 hz at @xmath31 .
( visual comparison was used because automated pattern recognition , which we attempted , was not sufficiently reliable . ) since the discovery of scars , there have been a number of theoretical attempts to obtain a quantitative measure for scarring @xcite based on eigenstate overlap , wigner function overlap , and inverse participation ratios for the amplitudes in the vicinity of the scars . to utilize such measures experimentally
, the local wave amplitude is required with high accuracy .
the shadowgraph technique used here is quantitative but nonlinear @xcite and does not provide this information .
development of a quantitative experimental measure of scarring has proven to be difficult even for linear probes .
we use the concept of pattern entropy " as a tool to classify the patterns .
egolf , melnikov , and bodenschatz @xcite have applied this concept successfully to measure the complexity of patterns observed in rayleigh - bnard convection .
the pattern entropy is calculated from the power spectrum of the pattern .
if @xmath47 is the normalized two dimensional power spectrum of the pattern at time t , then the pattern entropy @xmath48 is defined as : @xmath49 here @xmath48 measures the spectral complexity of a pattern .
if the image consists of just one fourier mode of amplitude unity , then @xmath50 ; otherwise @xmath51 . to minimize the effects of experimental noise , we sum contributions only in a band of wavenumbers centered at the mean wavenumber of the pattern with a rage of @xmath52 . in table
ii , the approximate entropy ranges for the various types of patterns observed in the range 55 - 65 hz are given .
note that the patterns are not distinguishable solely by their entropy , since some of the ranges overlap .
however , the pattern entropy can be useful in studies of time dependence farther above onset , as we show in sec .
here we examine the evolution of the wave patterns farther from onset , where the interactions between different fourier components of the waves become increasingly nonlinear and the approximation of eq .
( 2 ) becomes inapplicable .
the patterns were observed to be time - independent while changing adiabatically with @xmath3 for @xmath53 . on the other hand ,
they become weakly time - dependent for @xmath54 at most frequencies . the evolution toward time - dependence with increasing
@xmath3 depends on the excitation frequency .
three examples of this evolution are shown in figs .
4 - 6 . for some driving frequencies ,
the patterns remain reflection - symmetric as @xmath3 is increased , but exhibit transitions from one spatial mode to another prior to the onset of time dependence , as in fig .
5 . in these cases , the transition to spatial disorder ( asymmetry ) tends to coincide with the onset of spatiotemporal chaos ( stc ) .
it is important to note that as @xmath3 increases , the width of the stability tongue grows ( see eq.(5 ) ) : for example at @xmath55hz and @xmath56 the number of accessible modes of the container is approximately 35 .
therefore , the observed mode switching might be a combined effect of the increase in the number of accessible modes and an increase in the degree of nonlinearity that couples them .
it is remarkable that the container boundary continues to influence the patterns even at @xmath57 ( fig .
5d ) , where nonlinearity clearly plays a major role .
the variability of the nonlinear development is evident from examining the examples in figs .
4 - 6 . in fig .
4 , there is a general increase in complexity with @xmath3 , but the dominant mode does not change . in fig .
6 , the near onset pattern is nearly obliterated even at @xmath58 by the growing complexity , and the pattern is also distinctly asymmetric , while remaining time - independent . in one instance , a complex but symmetric time - independent pattern was observed at an unusually high driving amplitude of @xmath59 at a frequency of 65.0 hz , in a regime where spatiotemporal chaos is usually fully developed . the image shown in fig .
7 was averaged over 3000 images taken over a period of 5 minutes to test for time - dependence .
the lack of blurring demonstrates its time - independence .
for driving amplitudes just beyond the frequency - dependent onset of spatiotemporal chaos , the time dependence of the pattern is often intermittent ; the patterns appear to oscillate between states that are relatively ordered and states that are relatively disordered ( see the images in fig . 8 and a corresponding web - based movie @xcite )
. the power spectra of these typical patterns are also shown in fig .
8 and indicate the greater complexity of the disordered case , where the power is distributed more uniformly on the ring corresponding to the dominant wavenumber .
the time dependence and complexity of the patterns are monitored using two quantitative measures : ( i ) the rate of change of the pattern @xmath60 , and ( ii ) the entropy @xmath48 as defined in eq .
the rate of change @xmath60 was calculated by subtracting two consecutive images i(@xmath61 ) and i(@xmath62 ) separated by a time interval of @xmath63 seconds , and calculating @xmath60 according to the following formula : @xmath64 where , @xmath5 is the position , and @xmath65 is a constant scaling factor . in fig .
9(a ) , a section of the resulting quantity @xmath60 is shown as function of time for @xmath66hz and @xmath67 .
the data has been smoothed by averaging over four adjacent points . at every pronounced minimum of this smoothed data
we find that the corresponding pattern is symmetric and appears to have long range order . at all other times
the pattern is asymmetric and disordered .
a graph of the corresponding entropy @xmath48 is shown in fig
. 9(b ) .
examples of the ordered ( x ) and disordered ( y ) states are shown in fig . 8 , where the location in time is indicated by symbols x ( @xmath68s ) and y ( @xmath69s ) respectively in fig .
9 . at x , where the pattern entropy is low
, @xmath60 is small , while at y , where the pattern entropy is high , @xmath60 is large .
this correlation between @xmath60 , and @xmath48 holds true for most of the other strong peaks and valleys . for higher @xmath3 ( @xmath70 )
the oscillations diminish in strength and uniform stc is observed . in this regime , following gluckman et al .
@xcite , we obtained the time averaged pattern after adding 3000 instantaneous images obtained over five minutes .
an example of an instantaneous pattern is shown in fig .
10(a ) . the resultant average pattern is shown in fig .
the average reveals considerable ordered structure including remnants of the bouncing ball states .
this phenomenon is seen at most frequencies and persists up to @xmath71 . beyond this point , scars were visually absent and the average patterns are locally parallel to the boundary , as observed in circular or square patterns by gluckman et al .
in this paper we have discussed the parametrically forced wave patterns formed in a stadium - shaped container containing a low viscosity fluid , as a function of driving frequency and amplitude .
the patterns near onset ( @xmath72 ) were compared to a simple model consisting of linearized equations that reduce to the helmholtz equation with dirichlet boundary conditions ( see section ii ) . while a large proportion of observed patterns resemble the numerically computed eigenstates of the stadium , many of the computed eigenstates ( for instance the whispering gallery modes ) were not observed in a scan with sufficient frequency resolution to detect them .
the observed patterns may be broadly classified into three categories : ( a ) bouncing ball patterns , ( b ) longitudinal patterns , and ( c ) bowtie patterns , which have high amplitudes near corresponding periodic orbits .
in addition , a significant number of disordered patterns ( lacking in symmetry but time - independent ) were observed near onset .
furthermore , the observed mode spacing ( @xmath73hz ) is somewhat greater than the mean eigenvalue separation implied by eq .
( 6 ) ( @xmath74hz ) .
these observations imply that the simplest model is inadequate even close to onset .
recently , agam and altshuler have offered an explanation for the selection of modes at onset @xcite by considering the stability of the periodic orbits corresponding to the scars .
they argue that a threshold for excitation of a particular scarred pattern is given by : @xmath75 where @xmath76 is proportional to the rate that energy is pumped into the system ( i.e. the driving amplitude ) , @xmath77 is the dissipation rate in the bulk of the fluid , @xmath78 is the dissipation near the perimeter , and @xmath79 is the lyapunov exponent of the ray orbit that predominantly scars the pattern .
the bulk dissipation @xmath80 is the same for all the patterns and corresponds to approximately 2 sec@xmath81 in the frequency window used in the experiments .
therefore the appearance of a scarred pattern depends on a combination of the two remaining factors which are orbit - dependent .
stability of a pattern is favored both by a small lyapunov exponent of the associated scarring orbit , and by small perimeter dissipation @xmath78 . in the limit of high wavenumber @xmath30 relevant to our experiments ,
agam and altshuler derive an expression for the damping rate of scars due to boundary effects : @xmath82 where @xmath12 is the angular frequency , @xmath24 is the viscosity , @xmath29 is the length of the periodic orbit , and the sum is over all the collision points of the orbit with the boundary , @xmath83 being the angle between the orbit and a line perpendicular to the boundary at the collision point .
the parameter @xmath84 and the lyapunov exponent have been calculated for most of the shortest periodic orbits and a long ergodic orbit in ref .
the perimeter dissipation @xmath84 , which can be either smaller or larger than @xmath85 , varies between 0.2 sec@xmath81 and 3.0 sec@xmath81 .
the extreme values correspond to patterns scarred by the horizontal orbit and the ergodic orbit respectively .
most of the features observed in the experiments near onset appear to be captured by eq .
( [ criteria ] ) .
the bouncing ball orbit is prominent because the lyapunov exponent is zero in that case , and the longitudinal orbit occurs because of relatively low perimeter dissipation .
the whispering gallery orbits and others with angles that come close to @xmath86 have particularly large boundary dissipation and are suppressed .
[ criteria ] implies that if one increases the dissipation on the boundary so that it dominates , the longitudinal orbit will be the last to survive .
this is precisely what we observe in the experiments when the level of the fluid is lowered , a change that results in higher perimeter dissipation because of motion of the contact line .
the theory just cited @xcite is also able to account for the observed tendency of one scarred pattern to suppress other nearby eigenmodes through nonlinear interactions , as well as the existence of some asymmetric patterns . at some driving frequencies the patterns
are observed to switch modes as @xmath3 is increased ( fig .
this occurs especially when the bowtie state is observed at onset , an observation that may be related to the larger perimeter dissipation and lyapunov exponent of the bowtie mode . at higher driving amplitude ,
additional nonlinear effects occur as indicated by the growth in spatial complexity , and no adequate theoretical treatment exists . however , the onset patterns are often robust , persisting in the presence of increasing spatial complexity .
the boundaries remain influential even beyond the onset of time dependence . at sufficiently high @xmath3 ( @xmath87 ) , the onset patterns are no longer visible , though they persist in the time - average .
strong intermittency in the degree of order of the patterns is observed in the regime of spatiotemporal chaos .
furthermore , the rate of change of the pattern @xmath60 just above the stc onset is strongly correlated with the order as characterized by the entropy @xmath48 . the more ordered patterns
evolve more slowly in time , a striking observation that remains to be explained .
this tendency for ordered patterns to be more stable may be related to the complex but highly symmetric and time - independent pattern observed at atypically large acceleration ( fig .
we thank oded agam and boris altshuler for many useful discussions , and e.j .
heller for software to determine eigenstates of the stadium .
bruce boyes provided technical help .
this work was supported by the national science foundation under grant dmr-9704301 .
a. k. acknowledges a grant from the alfred p. sloan foundation . .
the percentages of patterns of various types observed in a sample of 128 patterns near onset ( @xmath31 ) , for driving frequency between 55 and 65 hz .
a typical example of a bouncing ball pattern is shown fig .
2(a ) , a longitudinal pattern in fig .
2(g ) , a bowtie pattern in fig .
2(e ) , and a disordered pattern in fig .
patterns with several components such as fig .
2(b ) , which contains both the longitudinal and bowtie modes , are counted in both categories .
therefore the sum of the percentages is slightly over 100% .
[ cols="<,^",options="header " , ] | surface wave patterns are investigated experimentally in a system geometry that has become a paradigm of quantum chaos : the stadium billiard .
linear waves in bounded geometries for which classical ray trajectories are chaotic are known to give rise to scarred patterns . here , we utilize parametrically forced surface waves ( faraday waves ) , which become progressively nonlinear beyond the wave instability threshold , to investigate the subtle interplay between boundaries and nonlinearity .
only a subset ( three main types ) of the computed linear modes of the stadium are observed in a systematic scan .
these correspond to modes in which the wave amplitudes are strongly enhanced along paths corresponding to certain periodic ray orbits .
many other modes are found to be suppressed , in general agreement with a prediction by agam and altshuler based on boundary dissipation and the lyapunov exponent of the associated orbit .
spatially asymmetric or disordered ( but time - independent ) patterns are also found even near onset . as the driving acceleration is increased ,
the time - independent scarred patterns persist , but in some cases transitions between modes are noted .
the onset of spatiotemporal chaos at higher forcing amplitude often involves a nonperiodic oscillation between spatially ordered and disordered states .
we characterize this phenomenon using the concept of pattern entropy .
the rate of change of the patterns is found to be reduced as the state passes temporarily near the ordered configurations of lower entropy .
we also report complex but highly symmetric ( time - independent ) patterns far above onset in the regime that is normally chaotic . |
the bardeen - cooper - schrieffer ( bcs ) theory @xcite or its advanced version the hartree - fock - bogoliubov ( hfb ) theory @xcite has long been used to treat pairing correlations @xcite in atomic nuclei . in the mean field
we introduce the bogoliubov quasi - particles and write the ground state as a slater determinant of the latter .
usually the variation principle is used to determine the structure of the quasi - particles . although enjoying great success , the method has disadvantages of breaking the exact particle number and a need for an unphysical minimum pairing - force strength @xcite
. the theory could be improved by the `` variation after particle - number projection '' ( vapnp ) procedure , in which the bcs or hfb wavefunction is projected onto good particle number before the variation is done @xcite . effectively , the pair condensate wavefunction [ eq . ( [ gs ] ) ]
is taken as the variational ground state .
the method has been discussed @xcite and applied with success to ultrasmall metallic grains @xcite in the vapnp+bcs version , and to realistic nuclei @xcite in the vapnp+hfb version .
however , the computing time cost is relatively large and there is only a limited number of nuclei on the nuclear chart that has been calculated by the vapnp + hfb method in large configuration spaces .
recently we proposed @xcite a new criteria to determine the pair condensate wavefunction based on the heisenberg equations of motion for density matrix operators .
the relevant equations have been solved for the bcs - type hamiltonian : each single - particle level has a distinct set of quantum numbers corresponding to the symmetries of the pairing operator , thus a level could only be paired with its time - reversed partner .
the validity of the theory has been proved on a large ensemble with random interactions @xcite . in this work
we consider the general situation of the hfb - type hamiltonian , where several single - particle levels could have the same set of quantum numbers and pairing among them is allowed .
the method should cost much less time in computing compared with the traditional variational calculation . in the algorithm of the current theory
only the one - body density matrices on the pair condensate are calculated , whereas the variation principle needs the two - body density matrices in computing the expectation value of a two - body hamiltonian . in sec .
[ sec_formalism ] we give the general formalism to solve for the particle - number - projected hfb wavefunction with the pairing theory based on the generalized density matrix ( gdm ) formalism .
then the method is tested in the simple two - level model with factorizable pairing interactions and in the semi - realistic model with the zero - range delta interaction , in sec .
[ sec_two_level ] and sec .
[ sec_delta_int ] , respectively .
finally sec .
[ sec_summary ] summarizes the work .
the antisymmetrized ( two - body ) fermionic hamiltonian governing the dynamics of the system is written as @xmath0 in the presence of pairing correlations , the ground state of the @xmath1-particle system is assumed to be an @xmath2-pair condensate , @xmath3 where @xmath4 is the normalization factor , @xmath5 is the pair creation operator @xmath6 in eq .
( [ p_dag_alpha ] ) @xmath7 represents a subspace consisting of `` pair - indices '' whose dimension is half of that of the single - particle space .
we could take , for example , @xmath7 to consist of those single - particle levels with a positive magnetic quantum number .
@xmath8 is the time - reversed level of the single - particle level @xmath9 ( @xmath10 ) .
the pair structure matrix @xmath11 is hermitian and block - diagonal .
the time - reversal invariance of @xmath12 implies that @xmath13 ( @xmath14 and @xmath9 have the same set @xmath15 of quantum numbers related to the symmetries of @xmath5 .
for example , in the spherical shell model @xmath16 is a collection of parity @xmath17 , angular momentum @xmath18 and its projection @xmath19 ; in the deformed nilsson mean field @xmath20 is a collection of parity @xmath17 and angular - momentum projection @xmath21 onto the symmetry axis .
we introduce the unitary transformation @xmath22 between the original single - particle basis @xmath23 and the new single - particle basis @xmath24 as @xmath25 where elements of the transformation matrix @xmath22 are defined through dirac notation .
properties of time - reversal operation imply that @xmath26 .
the matrix @xmath22 is block - diagonal : @xmath27 vanishes unless @xmath28 and @xmath29 have the same quantum number @xmath15 . under the transformation ( [ eta ] )
the operator @xmath5 ( [ p_dag_alpha ] ) becomes @xmath30 with @xmath31 we choose the transformation @xmath22 that diagonalizes the hermitian matrix @xmath11 in eq .
( [ v_12 ] ) , consequently @xmath5 becomes @xmath32 we call this new single - particle basis the `` canonical basis '' . in the following the arabic numerals
@xmath28 , @xmath33 , ... refer to single - particle levels in this basis unless otherwise specified . in the canonical basis
the density matrices for the pair condensate ( [ gs ] ) are `` diagonal '' : @xmath34 the normalization factor @xmath35 ( [ gs ] ) , occupation numbers @xmath36 ( [ rho_def ] ) , and pair - transfer amplitudes @xmath37 ( [ kappa_def ] ) are functions of the pair structures @xmath38 ( [ p_dag_1 ] ) ; their functional forms , as the `` kinematics '' of the system , have already been given in eqs .
( 23 ) and ( 24 ) of ref .
@xcite and are not repeated here .
the hartree - fock mean field @xmath39 and pairing mean field @xmath40 are defined as @xmath41 @xmath39 and @xmath40 are block - diagonal matrix : @xmath42 and @xmath43 vanish unless @xmath28 and @xmath33 have the same quantum number @xmath15 .
the hamiltonian parameters ( @xmath44 and @xmath45 ) in eqs .
( [ f_def ] ) and ( [ delta_def ] ) should be calculated from those in the original single - particle basis ( @xmath46 and @xmath47 ) through the transformation ( [ eta ] ) .
the equation of motion for the density matrix @xmath48 has been derived in eq .
( 14 ) of ref .
@xcite , @xmath49 where @xmath50 and @xmath51 are ground state energies for @xmath12 and @xmath52 , respectively . on the right - hand side , @xmath39 and @xmath40 are the mean fields ( [ f_def ] ) and ( [ delta_def ] ) , @xmath53 is the transpose of @xmath39 , and terms like ` @xmath54 ' are understood as matrix multiplication . in deriving eq .
( [ eom_k0 ] ) we have used the main approximation of the method , @xmath55 which says that on the pair condensate ( [ gs ] ) the two - body density matrix factorizes into products of one - body density matrices in both the particle - hole and particle - particle channels . equation ( [ eom_k0 ] ) is a block - diagonal matrix equation ; its @xmath56 matrix element vanishes unless @xmath28 and @xmath33 have the same quantum number @xmath15 . within each block
we take the @xmath56 matrix element on both sides of eq .
( [ eom_k0 ] ) , when @xmath57 we get after simplification @xmath58 and the @xmath59 matrix element gives @xmath60 equations ( [ main_12 ] ) and ( [ main_11 ] ) are the main equations of the theory .
below we show that the number of constraints from these two equations equals to the number of parameters in the pair structure @xmath11 ( [ p_dag_alpha ] ) ; thus the latter is fixed completely .
we assume a single - particle space of dimension @xmath61 split by the quantum number @xmath15 as @xmath62 . the number of restrictions in eq .
( [ main_12 ] ) is @xmath63 ( there is a factor @xmath64 because the equation has real and imaginary parts ) . equation ( [ main_11 ] ) implies that the right - hand side is independent of the single - particle label @xmath65 , which gives @xmath66 constraints .
hence the total number of constraints is @xmath67 . on the other hand ,
the number of independent parameters in the hermitian block - diagonal pair - structure matrix @xmath11 ( [ p_dag_alpha ] ) is @xmath68 ( there is a `` @xmath69 '' because an overall normalization factor does not matter ) .
the number of restrictions indeed equals to the number of parameters . in practical calculations usually the hamiltonian parameters @xmath46 and @xmath47 ( [ h_f ] ) are real .
in this case the pair structures @xmath11 ( [ p_dag_alpha ] ) could be taken as real numbers and the transformation @xmath22 ( [ eta ] ) is an orthogonal matrix . the number of restrictions from eqs .
( [ main_12 ] ) and ( [ main_11 ] ) is @xmath70 , which equals to the number of independent parameters in the real symmetric block - diagonal pair - structure matrix @xmath11 ( [ p_dag_alpha ] ) . in the following
we assume that this is the case .
an alternative parametrization may be more convenient in practice .
the independent parameters could be taken as those in the orthogonal transformation @xmath22 ( [ eta ] ) and the pair structure in the canonical basis @xmath71 ( [ p_dag_1 ] ) .
the orthogonal transformation for @xmath72 could be parameterized as @xmath73 for @xmath74 , @xmath75 and in general , @xmath76
we test the formalism in the simple model with two single - particle levels and a factorizable pairing interaction .
the theory is solved analytically and the results are compared with the exact shell - model results . we assume the rotational invariance and a single - particle space of two @xmath18-levels each with degeneracy @xmath77 ( degenerate in the magnetic quantum number @xmath19 ) .
the fermionic hamiltonian ( [ h_f ] ) takes the form @xmath78 where @xmath79 is the pairing operator @xmath80 in which @xmath81 and @xmath82 are `` diagonal '' and `` off - diagonal '' pairing strengths , respectively .
the hamiltonian parameters in the canonical single - particle basis are calculated through the transformation ( [ tran2 ] ) , @xmath83 where @xmath84 and @xmath85 . \label{pi_12}\end{aligned}\ ] ] based on eqs .
( [ f_def ] ) and ( [ delta_def ] ) we compute the matrix elements of the mean fields in the canonical single - particle basis , @xmath86 , \nonumber\end{aligned}\ ] ] and @xmath87 , \nonumber \\ \delta_{2\tilde{2 } } = - \omega ( g - p \sin2\theta ) [ ( g + p \sin2\theta ) s_1 + ( g - p \sin2\theta ) s_2 ] , \nonumber \\ \delta_{1\tilde{2 } } = - \omega p \cos2\theta [ ( g + p \sin2\theta ) s_1 + ( g - p \sin2\theta ) s_2 ] .
\nonumber\end{aligned}\ ] ] consequently eqs .
( [ main_12 ] ) and ( [ main_11 ] ) imply @xmath88 \nonumber \\ + p^2 \sin4\theta [ \omega ( n_1 + n_2 - 1 ) ( s_1 - s_2 ) - ( s_1 + s_2 ) ( n_1 - n_2 ) ] , \label{main_2l_12}\end{aligned}\ ] ] and @xmath89 \nonumber \\ - 2 g p \omega \sin2\theta s_1 s_2 + ( \omega - 1 ) ( g + p \sin2\theta)^2 n_1 s_1 s_2 - ( \omega - 1 ) ( g - p \sin2\theta ) ^2 n_2
\label{main_2l_11}\end{aligned}\ ] ] these two equations fix the two independent model parameters @xmath90 and @xmath91 ( @xmath36 and @xmath37 are functions of @xmath90 )
. then the density matrices in the original single - particle basis are calculated through the transformation @xmath22 ( [ tran2 ] ) , @xmath92 and @xmath93 we consider the range of model parameters in the realistic spherical nuclear shell model .
usually the two single - particle levels @xmath29 and @xmath9 belong to different major shells and @xmath94 is bigger than @xmath95 mev in magnitude that is about the energy of two major - shell gaps ( adjacent major shells have opposite parity ) .
based on the empirical pairing strength formula @xmath96 mev ( @xmath97 is the mass number ) @xcite , @xmath98 should be less than @xmath64 mev in medium and heavy nuclei having considered possible deviations from the constant pairing formula .
the off - diagonal pairing strength @xmath82 is usually smaller than the diagonal pairing strength @xmath81 owing to the smaller overlap of the two single - particle wavefunctions involved .
we test the model numerically in an ensemble consisting of examples with different model parameters .
the single - particle angular momentum @xmath18 of the examples takes value from @xmath99 .
the number of pairs is in the range @xmath100 ( from two particles to two holes in the model space ) .
the energy gap between the two single - particle levels is selected to be the energy unit so that @xmath101 and @xmath102 .
based on the estimations in the previous paragraph , we choose the range of the pairing strength to be @xmath103 and @xmath104 with step size @xmath105 .
we scan the whole parameter space and the total number of examples in the ensemble is @xmath106 ( the factor of @xmath107 is the number of possible values of @xmath81 and @xmath82 ) . the results for the density matrices @xmath108 and @xmath48 in the original single - particle basis are shown in figs .
[ fig_rho_line ] and [ fig_kappa_line ] , respectively .
each point in these figures corresponds to one example in the ensemble with the horizontal coordinate being the exact result of the respective quantity and the vertical one being that from the gdm calculation , thus a perfect calculation would have all the points lying on the @xmath109 straight line . from figs .
[ fig_rho_line ] and [ fig_kappa_line ] we see that in general the gdm theory reproduces the exact density matrices well .
the root - mean - square deviations are @xmath110 @xmath111 and @xmath112 are equal because @xmath113 .
the results for the ground state energies are shown in fig .
[ fig_e_line ] .
the horizontal coordinate of each point is the pairing correlation energy for the example @xmath114 , where @xmath115 is the exact ground state energy by the direct diagonalization , and @xmath116 or @xmath117 is the occupation number of the naive fermi distribution . the vertical coordinate is the ground state energy by the gdm calculation measured from the exact one , @xmath118 , where @xmath119 is calculated in the canonical single - particle basis using the hamiltonian ( [ h_12 ] ) together with the recursive formulas derived in refs .
we see that in general the gdm calculation reproduces well the ground state energies : the errors are small compared with the pairing correlation energies .
the average values of the errors and the pairing correlation energies for the ensemble are @xmath120 and @xmath121 , respectively . at last in fig .
[ fig_derivations_p ] we show the accuracy of the method depending on the off - diagonal pairing strength @xmath82 .
the @xmath122 examples in the ensemble are divided into five subgroups according to the value of @xmath82 , and deviations within each subgroup of various quantities are calculated .
we see that in general the errors of the gdm calculation increase with @xmath82 . in realistic spherical nuclear shell model @xmath82 should be smaller than about @xmath123 based on the estimations made in the paragraph below eq .
( [ kappa_ori ] ) . in this region
the gdm theory is rather accurate .
in this section we further test the theory in a semi - realistic model with the pairing hamiltonian matrix elements calculated from the zero - range delta interaction .
the model has only one species of nucleons ( for example , neutrons ) and the single - particle space consists of five levels : three of them ( @xmath124 , @xmath125 , @xmath126 ) are from the @xmath127 major shell and two ( @xmath128 , @xmath129 ) are from the @xmath130 major shell .
the single - particle energies are determined by random number generator to be @xmath131 , @xmath132 , @xmath133 , @xmath134 , and @xmath135 mev ( the gap between the @xmath136 and @xmath130 major shells is taken to be around @xmath95 mev ) . the two - body pairing matrix elements [ @xmath137 with @xmath138 are calculated from the zero - range delta interaction @xmath139 , taking the single - particle wavefunctions ( @xmath29 , @xmath9 ... ) to be the harmonic oscillator ones .
we perform six sets of calculations at different model parameters of the particle number @xmath1 and the pairing strength @xmath140 as listed in table [ tab_setpara ] .
the unit for @xmath140 is @xmath141 , where @xmath142 is the length parameter for the harmonic oscillator wavefunctions ( @xmath143 is the mass of the nucleon ) .
values of @xmath140 are chosen so that magnitudes of the resulting two - body pairing matrix elements are realistic .
the last row of table [ tab_setpara ] shows the average value of the pairing matrix elements of a specific form , @xmath144 = -\sum_\alpha v_{\alpha\tilde{\alpha}\tilde{\alpha}\alpha } / ( 2\omega)$ ] , where @xmath145 runs over the entire single - particle space whose dimension is @xmath61 .
the results are shown in table [ tab_res ] .
we see that in general the gdm calculation reproduces the exact density matrices accurately , except for the very small value of @xmath146 in set5 .
particularly , the abrupt change of @xmath48 values from set3 to set4 is captured very well .
the off - diagonal parts of the density matrices ( @xmath147 , @xmath148 , @xmath149 , and @xmath150 ) , originating from the `` off - diagonal '' parts of the hfb - type pairing hamiltonian , are always well reproduced , ranging from a less - than - one - percent effect to a few .
the errors of the gdm ground state energy relative to the exact ones are shown in the last column .
we see that the errors are small and consistent with the statistical estimate made within a large random ensemble for the case of a bcs - type pairing hamiltonian in ref .
in summary , we considered the solution of the gdm pairing theory with the hfb - type hamiltonian . in the algorithm only the one - body density matrix is computed , thus the method should be much faster than the vapnp + hfb method in which the two - body density matrix is needed in calculating the expectation value of a two - body hamiltonian . with the assumption that the parent and daughter nuclei were represented by pair condensates with the same pair structure , the pair - transfer amplitudes @xmath48 could be calculated in one run together with the occupation numbers @xmath108 .
in contrast , the traditional variation method calculates the parent and daughter nuclei separately , each with a different pair structure .
the formalism is tested in the simple two - level model with factorizable pairing interactions , and the semi - realistic model with the zero - range delta interaction . in both cases
the gdm calculation reproduces quite well the exact density matrices and ground state energy within the physical range of parameters for the realistic spherical nuclear shell model .
the errors are small and consistent with the statistical estimates made within a large random ensemble for the bcs - type hamiltonian in ref .
@xcite . in the current form
the theory is ready for the application to realistic nuclear systems .
support is acknowledged from the hujiang foundation of china ( b14004 ) , and the startup funding for new faculty member in university of shanghai for science and technology .
part of the calculations is done at the high performance computing center of michigan state university .
.particle number @xmath1 and pairing strength @xmath140 as input parameters for the six sets of calculations .
the quantity @xmath151 in the unit of @xmath140 is the length parameter for the harmonic oscillator wavefunctions . also shown
is @xmath152 $ ] , the average value of the pairing matrix elements of a specific form .
see text for details . [
cols="^,^,^,^,^,^,^",options="header " , ] | recently we proposed a particle - number - conserving theory for nuclear pairing [ jia , phys .
rev .
c * 88 * , 044303 ( 2013 ) ] through the generalized density matrix formalism .
the relevant equations were solved for the case when each single - particle level has a distinct set of quantum numbers and could only pair with its time - reversed partner ( bcs - type hamiltonian ) . in this work we consider the more general situation when several single - particle levels could have the same set of quantum numbers and pairing among these levels
is allowed ( hfb - type hamiltonian ) .
the pair condensate wavefunction ( the hfb wavefunction projected onto good particle number ) is determined by the equations of motion for density matrix operators instead of the variation principle .
the theory is tested in the simple two - level model with factorizable pairing interactions and the semi - realistic model with the zero - range delta interaction . |
the avogadro constant , @xmath0 , is determined by dividing the molar volume of a silicon single - crystal highly enriched with @xmath1si by its unit cell volume , obtained from the measured value of the lattice parameter @xcite .
the lattice parameter is measured by means of combined x - ray and optical interferometry .
recently , the effect of surface stress was brought to our attention as a possible error source in our measurement of the lattice parameter @xcite .
indeed , this issue was already taken into account but never definitely investigated @xcite . an intrinsic surface - stress exists in solid , in a similar way
as it occurs in liquids , which was first modelled by gibbs @xcite and shuttleworth @xcite , whose works were seminal in surface thermodynamics .
though it is widely reported in review papers and textbooks on surface science @xcite , their theory is still a matter of debate @xcite .
a continuous mechanics theory of solid surfaces has been developed by gurtin and murdoch @xcite , which is summarized in @xcite .
when the surface relaxes , it compresses or pulls the underlying crystal ; since the resulting lattice strain depends on the crystal size and geometry , the lattice parameter of an x - ray interferometer is different from that of the si spheres used to determine the molar volume . after summarizing the operation of a combined x - ray and optical interferometer
, this paper investigates the effect of surface stress on the interferometer operation .
an elastic membrane model has been used to provide a surface load in a finite element analysis of the lattice strain .
particular emphasis was given to the operation of the @xmath1si interferometer used to determine @xmath0 .
eventually , a measurement of the lattice parameter with a variable - thickness interferometer is proposed in order to evidence the effect of surface stress on the underlying crystal .
a review of the lattice parameter measurement can be found in @xcite . as shown in fig .
[ xint ] , an x - ray interferometer consists of three si crystal slabs so cut that the \{220 } planes are orthogonal to the crystal surfaces .
x - rays from a 17 kev mo k@xmath2 source are split by the first crystal and recombined , via two transmission crystals , by the third , called the analyser . when the analyser is moved in a direction orthogonal to the \{220 } planes , a periodic intensity - variation of the transmitted and diffracted x - rays
is observed , the period being the diffracting - plane spacing , about 192 pm .
the analyser embeds front and rear mirrors , so that its displacement and rotations can be measured by an optical interferometer .
the measurement equation is @xmath3 , where @xmath4 pm is the spacing of the \{220 } planes and @xmath5 is the number of x - ray fringes observed in a displacement of @xmath6 optical fringes having period @xmath7 nm . to ensure the interferometer calibration
, the laser source operates in single - mode and its frequency is stabilized against that of a transition of the @xmath8i@xmath9 molecule . to eliminate the adverse influence of the refractive index of air and to achieve millikelvin temperature uniformity and stability ,
the experiment is carried out in a thermovacuum chamber .
continuous developments led the measurement sensitivity and accuracy to approach @xmath10 ; the most recent results are given in @xcite .
similarly to liquid surfaces , solid surfaces exhibit a stress also if the bulk material is stress - free .
this is the macroscopic consequence of atomic affects : roughly speaking , the atoms on the surface lose half of the bonds they would have in the bulk ; therefore , crystal surfaces are disturbances in the bonding of atoms .
surface stress was first generalized to solids by gibbs and shuttleworth @xcite as the amount of reversible work per unit area needed to stretch a pre - existing surface . since a solid surface may change not only by exchange of matter with the bulk , but also by elastic deformation , the concept of surface stress must be derived from the molecular dynamics of the surface @xcite .
continuum mechanics provides the framework to define the surface properties and the relationships among them . the elasticity theory of solid surfaces to which we refer was developed by gurtin and murdoch @xcite .
a summary of the linearised theory which is valid when both the surface and bulk strains are small is given in @xcite .
the key concept is that , for a surface attached to the underlying solid , a stress - free reference does not exist ; since the bonds between the surface atoms and between the bulk atoms are different , the surface of a stress - free crystal has residual strain , @xmath11 , and stress , @xmath12 .
if the surface tends to shrink ( expand ) with respect to the bulk , the surface stress is positive ( negative ) and it is said to be tensile ( compressive ) . since a fully relaxed surface has no normal stress , surface strain and stress are second rank tensors , which are characterized by two principal orthogonal - components .
if the surface has more than a 2-fold symmetry , the surface stress and strain tensors are isotropic . in this case , @xmath13 and the free surface - energy per unit area is @xcite @xmath14[free ] where @xmath15 are the strain component in the plane tangent to the surface , @xmath16 is the surface energy for a stress - free bulk , @xmath17 and @xmath18 are elastic constants , and the summation over the repeated indices
@xmath19 is implied . the definition of surface stress is @xmath20 ; hence , by taking the derivatives , @xmath21 where @xmath22 is the intrinsic surface stress .
@rrrll @xmath23/(n / m ) & @xmath24/(n / m ) & @xmath25/(n / m ) & year & reference + @xmath26 & @xmath27 & @xmath28 & 1988 & @xcite + @xmath29 & @xmath30 & @xmath31 & 1989 & @xcite + @xmath32 & @xmath33 & @xmath34 & 1990 & @xcite + @xmath35 & @xmath36 & @xmath37 & 1992 & @xcite + @xmath38 & @xmath39 & @xmath40 & 1993 & @xcite + @xmath29 & @xmath41 & @xmath42 & 1994 & @xcite + @xmath43 & @xmath44 & @xmath45 & 2005 & @xcite + @xmath46 & @xmath47 & @xmath48 & 2005 & @xcite + @xmath49 & @xmath50 & @xmath51 & 2008 & @xcite + the determination of the surface stress in solids is a challenge ; the measurement of the lattice parameter in small particles has been the main method @xcite . given the experimental difficulties and since the results are affected by the sample preparation and adsorbed impurities , the values obtained from _ ab - initio _ and molecular dynamics calculations seem more reliable .
the results for the symmetric @xmath52 reconstruction of the si(001 ) surface are summarized in table [ s - stress ] ; the reconstructed surface is formed by domains of anisotropic stress , but , for symmetry reasons , the mean stress is isotropic .
the scatter of data suggests a reassuring small surface stress . because of the anisotropic nature of the crystal , surface stress depends on the surface orientation .
a measurement of the polar dependence of the relative surface - stress at high temperature ( 1373 k ) is given in @xcite , which allows the surface stress of any surface to be determined from that of the ( 001 ) reference surface .
in the case of our x - ray interferometer , the @xmath53 surfaces are those of main interest ; because of their 2-fold symmetry , two values are necessary to represent the surface stress of the two perpendicular @xmath54 $ ] and @xmath55 $ ] components .
as reported in @xcite , at 1373 k , the sought principal components are @xmath56 and @xmath57 .
indicative @xmath25 values are given in table [ s - stress ] .
the crystal s surfaces are normally covered by a few nanometre thick oxide - layer , which can further stress the surface .
the stress changes of the si(001 ) @xmath52 surface during plasma oxidation were measured by using a thin silicon cantilever ; the bending caused by the oxidation of one face was optically measured and related to the surfaces stress difference by elasticity theory @xcite .
initially , the oxidation induces a compressive stress , followed by a tensile stress ; eventually , the stress variation turns to compressive for oxide thickness greater than about 1.5 nm .
the initial stress changes were explained in terms of the electron diffusion into the sample and the shrinkage caused by the oxygen - bridged dimer structure . the final compressive variation , @xmath58
n / m , was explained by the oxidation of si atoms at deep sites ; since the oxygen expands the si
si bond , oxidation is expected to cause a compressive stress .
the surface relaxes the intrinsic stress by compressing or pulling the underlying crystal .
for instance , the ( isotropic ) lattice strain of a sphere of radius @xmath59 is @xmath60 where @xmath61 is the bulk modulus and the young - laplace approximation holds when the elastic constants can be neglected when compared to the @xmath62 product . by choosing @xmath63 n / m , we can use this formula where @xmath64 mm and @xmath65 gpa to estimate the lattice strain of the spheres used to determine @xmath0 .
the result is @xmath66 nm / m , which is irrelevant for the @xmath0 determination .
however , if @xmath67 mm which is a value equal to the thickness of the crystal used to measure the lattice parameter we obtain @xmath68 nm / m , which is significant with respect to the uncertainty of the lattice - parameter measurement .
figure [ analyser ] ( left ) shows the analyser crystal of the @xmath1si x - ray interferometer used for the @xmath0 measurement and its crystallographic orientation .
the end mirrors are parallel to the diffracting planes ; to avoid stress propagation , they are separated by vertical cuts .
the lamella thickness , 1.2 mm , was chosen by a trade off between the increased lattice strain ( due to the surface stress ) of a thin lamella and the increased x - ray absorption of a thick lamella . to estimate the lattice parameter change caused by the surface stress
, we used a commercial finite element analysis software @xcite .
as shown in fig .
[ analyser ] , the ( 220 ) planes are orthogonal to the lamella faces which are parallel to @xmath53 planes .
owing the large uncertainty of the residual stress values , though the @xmath53 surfaces are anisotropic , the analyser surfaces were modelled as stretched by isotropic membranes attached to the underlying crystal lattice . since we are only concerned with infinitesimal strains , though the surface stress @xmath69 depends on the strain according to ( [ stress ] ) , we used an isotropic @xmath63 n / m value .
this surface stress value allows the analysis results to be easily scaled if a different value has to be considered and , hopefully , it is larger than the actual stress value .
hence , according the young - laplace equation , the surface applies the inward pressure @xmath70 , where @xmath71 is the mean curvature . as schematically shown in fig .
[ analyser ] ( right ) , when considering flat surfaces , the surface stress was modelled by shear forces per unit length applied orthogonally to the surface edges .
si interferometer .
m : end mirrors , a : analyser lamella , b : base . the @xmath72 mm@xmath73 reliefs that support the crystal are coloured in black .
the lattice parameter was measured along the indicated line , at @xmath74 mm .
right : surface - stress modelling by shear forces.,title="fig:",width=245 ] si interferometer .
m : end mirrors , a : analyser lamella , b : base . the @xmath72 mm@xmath73 reliefs that support the crystal are coloured in black .
the lattice parameter was measured along the indicated line , at @xmath74 mm .
right : surface - stress modelling by shear forces.,title="fig:",width=245 ] as strains are infinitesimal , we uses a linear model ; the mesh , of about @xmath75 tetrahedral anisotropic elements , was the result of successive refinements .
the not null elements of the stiffness matrix are given table [ s : matrix ] in both the [ 100 ] , [ 010 ] , [ 001 ] and [ 110 ] , [ @xmath7610 ] , [ 001 ] crystal axes .
as regards the model parameters , the si density and gravitational acceleration are 2330 kg / m@xmath73 and 9.81 m / s@xmath77 , respectively .
@ccccccccc + & & + & & + + & @xmath78 & & @xmath79 & & @xmath80 + & 35.2 & & 165.7 & & 50.9 + the analyser rests on a silicon plate ; it stands on three @xmath72 mm@xmath73 reliefs carved out of the crystal base by chemical etching . in order to avoid stresses ,
it is held in position by a thin layer of high - viscosity silicone oil .
since the oil applies only viscous forces , the contact with the support has been simulated by setting to zero the displacement of the contact nodes .
we calculated the strain @xmath81 , where @xmath82 is the horizontal component of the displacement vector .
both the effects of self - weight and surface strain have been considered ; the supports were so chosen as to minimize bending or sagging @xcite .
the three dimensional map of the strain distribution in the analyser and a magnified image of the strained diffracting planes are shown in fig .
[ strain:1 ] .
the @xmath83 strain is trivially related to the lattice parameter by @xmath84 .
hence , the relative variations of the @xmath85 values along measurement lines at different heights are identical to the @xmath83 values calculated along the same lines ; they are shown in fig .
[ strain:2 ] ( left ) .
the figure indicates that , in the measurement horizontal strip a couple of millimetre wide at a @xmath74 mm , the lattice is uniformly strained by about @xmath86 nm / m , as estimated by the previous `` spherical - cow '' model .
in addition to the horizontal strain , there is a vertical strain gradient of about @xmath87 nm m@xmath88 cm@xmath88 , the bottom part being less strained , which could be the subject matter of future experimental verifications .
component of the diffracting - plane strain for the optimal location of the supports , where blue is @xmath90 nm / m and red is @xmath91 nm / m .
right : qualitative magnified image of the diffracting planes .
a realignment of the analyser has been simulated by subtracting the average displacement and tilt .
the displacement magnification factors is shown by the 1 nm bar.,title="fig:",width=245 ] component of the diffracting - plane strain for the optimal location of the supports , where blue is @xmath90 nm / m and red is @xmath91 nm / m .
right : qualitative magnified image of the diffracting planes .
a realignment of the analyser has been simulated by subtracting the average displacement and tilt .
the displacement magnification factors is shown by the 1 nm bar.,title="fig:",width=245 ] variations in @xmath92 plane , at different distances from the analyser bottom - surface .
right : relative @xmath85 variations along the @xmath74 mm measurement baseline ( figure [ analyser ] ) .
different support - to - mirror distances have been considered , as indicated in the graph.,title="fig:",width=245 ] variations in @xmath92 plane , at different distances from the analyser bottom - surface .
right : relative @xmath85 variations along the @xmath74 mm measurement baseline ( figure [ analyser ] ) .
different support - to - mirror distances have been considered , as indicated in the graph.,title="fig:",width=245 ] figure [ strain:2 ] ( right ) updates the corresponding figure 6 of @xcite , which was used to assess the sole effect of self weigh in the lattice parameter measurement .
it shows that , if the analyser surfaces are loaded by a tensile stress of 1
n / m , the @xmath85 value is about @xmath93 smaller than its value in an unstrained crystal .
the figure also shows that the lattice parameter profile given in @xcite can not deliver clues about the surface contribution to the measured @xmath85 value because the only effect of the surface stress is to translate all the curves downwards by @xmath93 . since the uncertainty associated to the measured @xmath85 value is @xmath94 , the lattice strain induced by a 1 n / m stress impairs the measurement accuracy .
the results indicate also that we can cope effectively the surface stress by a numerical determination of its effect .
however , owing the large uncertainty of the sign and value of the surface stress , we do nt yet propose to reconsider the corrections and error - budget contributions given in @xcite .
further investigations are under way to assess and to quantify the effect of surface stress , if any . in the next section
we describe an analyser design to work a lattice parameter measurement out so that there is a measurable effect of the surface stress .
variations are calculated along the indicated lines , at @xmath95 mm.,width=245 ]
a way to evidence if the surface stress strains the analyser crystal up to an extent which is significant to the lattice parameter measurement is to compare the results of measurements carried out in crystals having a large thickness difference .
since the induced @xmath85 changes are small , to avoid they are masked by lattice imperfections , measurements must be carried out in the same crystal . consequently , as shown in fig .
[ new - xint ] , we designed an x - ray interferometer having a variable analyser thickness . the minimum 0.4 mm thickness is set by machining capabilities ; the maximum 1.5 mm thickness is set by x - ray absorption . after the optimization of the finite element analysis and the location of the supports so as to minimize the self - weight bending , we obtained the strain distribution shown in fig .
[ new - strain:1 ] ( left ) ; a magnified image of the strained \{220 } planes is shown in fig .
[ new - strain:1 ] ( right ) .
the expected @xmath85 profile along the horizontal lines at @xmath95 mm height is shown in fig .
[ new - strain:2 ] ( left ) . as shown in fig .
[ new - strain:2 ] ( right ) , the surface stress and self - weight have opposite effects . the surface stress will be estimated from the @xmath85 gap between the thick and thin analyser parts .
if the surface is loaded by a tensile stress of 1 n / m , @xmath85 decreases by about @xmath96 from the central ( thick ) to the outer ( thin ) parts .
this variation is large enough to be detected . with a high crystal perfection ,
the present detection limit of @xmath97 corresponds to a minimum detectable stress of 0.2 n / m . to ensure that self - weight bending does not change in a significant way the @xmath85 gap between the thick and thin analyser parts , the height of the analyser base was increased to 19.8 mm .
figure [ rejection ] shows that , when the analyser supports are randomly displaced ( in @xmath98 plane ) from their optimal positions up to @xmath99 cm , the @xmath85 gap between the thick and thin analyser - parts changes less than what can be experimentally detected . component of the diffracting - plane strain for the variable - thickness analyser , where blue is @xmath90 nm / m and red is @xmath91 nm / m .
right : qualitative magnified image of the diffracting planes .
a realignment of the analyser has been simulated by subtracting the average displacement and tilt .
the displacement magnification factors is shown by the 1 nm bar.,title="fig:",width=245 ] component of the diffracting - plane strain for the variable - thickness analyser , where blue is @xmath90 nm / m and red is @xmath91 nm / m .
right : qualitative magnified image of the diffracting planes .
a realignment of the analyser has been simulated by subtracting the average displacement and tilt .
the displacement magnification factors is shown by the 1 nm bar.,title="fig:",width=245 ] variations in @xmath92 plane , along the @xmath95 mm measurement lines ( figure [ new - xint ] ) .
the supports are optimally located so as to minimize the self - weight bending . left : both the self - weight and surface stress
are considered .
right : only the self - weight is considered.,title="fig:",width=245 ] variations in @xmath92 plane , along the @xmath95 mm measurement lines ( figure [ new - xint ] ) .
the supports are optimally located so as to minimize the self - weight bending .
left : both the self - weight and surface stress are considered .
right : only the self - weight is considered.,title="fig:",width=245 ] variations in @xmath92 plane , along the @xmath95 mm measurement lines , when the analyser supports are displaced ( in @xmath98 plane ) from their optimal positions up to @xmath99 cm.,width=245 ]
the influence of the intrinsic surface - stress on the operation of the x - ray interferometer utilized to measure the lattice parameter of the @xmath1si crystal used to determine the avogadro constant has been investigated by finite element analysis .
the crystal surfaces , modelled as membranes having a tensile stress of 1 n / m , relax by compressing the underlying crystal . though the induced strain is greater than the uncertainty associated to the measured value of the lattice parameter , since the stress magnitude is highly uncertain and the chosen 1 n / m value could be pessimistic , we do not propose a correction of the published value . to verify
if the surface stress contributes or does not contribute to the lattice parameter measurement , a variable - thickness design of the x - ray interferometer has been investigated to induce detectable strain variations in the same interferometer crystal .
the realization of such an interferometer is under way in collaboration at the physikalisch technische bundesanstalt in collaboration with the leibniz - institut fr oberflchenmodifizierung .
we thank petr kren of the czech metrology institute for having brought to our attention this criticality of the lattice parameter measurement .
this work was jointly funded by the european metrology research programme ( emrp ) participating countries within the european association of national metrology institutes ( euramet ) and the european union .
99 andreas b _ et al .
_ 2011 determination of the avogadro constant by counting the atoms in a @xmath1si crystal _ phys . rev .
lett . _ * 106 * 030801 andreas b _ et al .
_ 2011 counting the atoms in a @xmath1si crystal for a new kilogram definition _ metrologia _ * 48 * s1 - 13 massa e , mana g , kuetgens u and ferroglio l 2011 measurement of the 220 lattice - plane spacing of a 28si x - ray interferometer _ metrologia _ * 48 * s37 - 43 kren p private communication ferroglio l , mana g , palmisano c and zosi g 2008 influence of surface stress in the determination of the ( 220 ) lattice spacing of silicon _ metrologia _ * 45 * 110 - 8 gibbs j w 1906 _ the scientific papers of j. willard gibbs , vol . 1 _ ( london : longmans - green ) p 55 shuttleworth r 1950 _ phys . soc . a _ * 63 * 444 - 57 cammarata r c 1994 surface and interface stress effects in thin films _ prog . surf .
sci . _ * 46 * 1 - 38 desjonqures m c and spanjaard d 1996 _ concept in surface physics _ ( springer , berlin ) makkonen l 2012 misinterpretation of the shuttleworth equation _
scripta materialia _ * 66 * 627 - 9 gurtin m e and murdoch i 1975 a continuum theory of elastic material surfaces _ arch .
* 57 * 291 - 323 gurtin m e and murdoch i 1978 surface stress in solids _ int .
j. solids struct .
_ * 14 * 431 - 40 wolfer w g 2011 elastic properties of surfaces on nanoparticles _ acta materialia _ * 59 * 7736 - 43 becker p and mana g 1994 the lattice parameter of silicon : a survey _ metrologia _ * 31 * 203 - 9 miyamoto 1994 comparative study of si(100 ) surface stresses with dimerized group - iv adatoms _ phys .
b _ * 49 * 1947 - 51 hara s , izumi s , kumagai t and sakai s 2005 surface energy , stress and structure of well - relaxed amorphous silicon : a combination approach of ab initio and classical molecular dynamics _ surface science _
* 585 * 17 - 24 alerhand o l , vanderbilt d , meade r d and joannopoulos j d 1988 spontaneous formation of stress domains on crystal surfaces _ phys .
lett . _ * 61 * 1973 - 6 payne m c , roberts n , needs r j , needels m and joannopoulos j d 1989 total energy and stress of metal and semiconductor surfaces _ surf .
sci . _ * 211/212 * 1 - 20 meade r d and vanderbilt d 1990 _ proc .
on the physics of semiconductors _ singapore : world scientific p 123 poon t w , yip s , ho p s and abraham f f 1992 ledge interactions and stress relaxations on si(001 ) stepped surfaces _ phys .
b _ * 45 * 3521 - 31 garca a and northrup j e 1993 stress relief from alternately buckled dimers in si(100 ) _ phys .
b _ * 48 * 17350 - 3 delph t j 2008 near - surface stresses in silicon(001 ) _ surf .
sci . _ * 602 * 259 - 67 mtois j j , sal a and mueller p 2005 measuring the surface stress polar dependence _ nature materials _ * 4 * 238 - 42 itakura a n , narushima t , kitajima m , teraishi k , yamada a and miyamoto a 2000 surface stress in thin oxide layer made by plasma oxidation with applying positive bias _ applied surface science _ * 159 - 160 * 62 - 6 comsol multiphysics@xmath100 4.3a ferroglio l , mana g , and massa e 2011 the self - weight deformation of an x - ray interferometer _
metrologia _ * 48 * s50 - 4 | a stress exists in solids surfaces , similarly to liquids , also if the underlying bulk material is stress - free .
this paper investigates the surface stress effect on the measured value of the si lattice parameter used to determine the avogadro constant by counting si atoms .
an elastic - film model has been used to provide a surface load in a finite element analysis of the lattice strain of the x - ray interferometer crystal used to measure the lattice parameter .
eventually , an experiment is proposed to work a lattice parameter measurement out so that there is a visible effect of the surface stress . |
since its isolation in 2004 @xcite , graphene , a monolayer of carbon atoms arranged in a honeycomb lattice , has received great attention due to both its unique electrical and mechanical properties @xcite . in graphene , the carbon atoms display an sp@xmath0 hybridization , with the out of plane p@xmath1 orbitals forming a @xmath2 band which is responsible for its electrical properties , while the sp@xmath0 orbitals form strong @xmath3 in plane bonds that govern its mechanical properties .
it has been verified both experimentally and theoretically @xcite that graphene is the known material with the largest in plane elastic constants .
in freely suspended graphene samples the vibrations of the lattice can be classified into in plane and out of plane ( flexural ) modes , with the flexural mode lying at lower energies and showing a quadratic dispersion .
anharmonic effects at long wavelengths strongly couple in plane and flexural modes @xcite .
the flexural mode is responsible for the significant temperature dependence of the electronic resistivity at low temperatures @xcite .
anharmonic effects in suspended graphene explain its negative thermal expansion coefficient @xcite and they play a significant role in the thermal conductivity @xcite .
the interaction between a graphene layer and a substrate underneath changes significantly the properties of the out of plane vibrations of the entire system @xcite . coupling to a substrate also leads to heat transfer between the two systems @xcite .
the deformations of the hybrid system made up of the graphene layer and the substrate at small amplitudes and long wavelengths are rigorously described by the theory of elasticity .
this theory fixes the number of independent couplings required , which is determined by the dimensionality and symmetries of the two systems , graphene and substrate , to be studied .
a model fully consistent with the theory of elasticity is described in the next section .
being general , this model should also describe other two dimensional materials supported by a substrate .
we studied how this coupling gives origin to a finite lifetime for the flexural mode and to a hybridization of this with the substrate surface rayleigh mode , comparing our model with experimental data from ref . .
we also studied the thermal expansion of graphene on a substrate and also the effect of the flexural mode , modified by the coupling to the substrate , on the electrical resistivity of doped graphene , focusing on two of the most common substrates : silicon dioxide , sio@xmath4 , and hexagonal boron nitride , hbn .
( a ) representation of the model used for the graphene membrane - substrate coupling .
( b ) density plot of the spectral function @xmath5 for graphene on sio@xmath4 [ in units of @xmath6 . for @xmath7 , it is zero everywhere , except at the dispersion relation of the fr mode , @xmath8 , where it is a dirac delta function with weight @xmath9 shown in the inset . for @xmath10 , @xmath5 is finite , with a peak close to @xmath11 , which becomes very broad for small @xmath12 , indicating that the flexural phonon becomes poorly defined .
vertical lines are @xmath13 and @xmath14 .
we used @xmath15 . ,
title="fig:",width=132](a ) representation of the model used for the graphene membrane - substrate coupling .
( b ) density plot of the spectral function @xmath5 for graphene on sio@xmath4 [ in units of @xmath6 . for @xmath7 , it is zero everywhere , except at the dispersion relation of the fr mode , @xmath8 , where it is a dirac delta function with weight @xmath9 shown in the inset . for @xmath10 , @xmath5 is finite , with a peak close to @xmath11 , which becomes very broad for small @xmath12 , indicating that the flexural phonon becomes poorly defined .
vertical lines are @xmath13 and @xmath14 .
we used @xmath15 .
, title="fig:",width=188 ]
in our model , a flat graphene membrane is supported by a semi - infinite flat substrate , graphene will have regions where it conforms to the substrate , while other regions will become detached from it @xcite .
the detailed analysis of this situation is beyond the scope of the present paper , although the analysis presented provides a reasonable approximation to the regions which are well attached to the substrate . ] that occupies the half - space @xmath16 . in a long wavelength description
, we will use the elastic theory of a crystalline membrane to model graphene and linear elasticity theory to describe the substrate .
therefore , the action describing the membrane - substrate coupled system will be given by @xmath17 .
@xmath18 is the quadratic action for the flexural mode , @xmath19 is the action for the in plane modes of the membrane including anharmonic coupling between in plane and flexural modes , @xmath20 is the linear elasticity action for the substrate and @xmath21 and @xmath22 describe the membrane - substrate coupling . assuming that the fluctuations around the equilibrium positions are small
, we keep @xmath21 and @xmath22 only to quadratic order in the displacement fields@xcite . finally , assuming in plane isotropy , the most general description of the membrane - substrate model consistent with the theory of elasticity is @xmath23 here @xmath24 and @xmath25 are , respectively , the out of plane and in plane displacement fields of the membrane with mass density per unit area @xmath26 , bending rigidity @xmath27 and lam coefficients @xmath28 and @xmath29 .
@xmath30 is the strain tensor to lowest order in @xmath24 and @xmath25 .
@xmath31 is the displacement field of the substrate with mass density per volume @xmath32 and elastic constants @xmath33 ( see table [ tab : material ] ) ( greek indices run from 1 to 2 and latin indices from 1 to 3 ) . for graphene , @xmath34 and @xmath35 @xcite .
the graphene - substrate coupling @xmath36 in @xmath21 is determined by the van der waals interaction between the two systems .
the couplings in @xmath22 depend on short range interatomic interactions , and are significantly weaker than the van der waals interaction for graphite ( graphene on graphene)@xcite . out of plane phonons of graphene
are also the most affected by the presence of a substrate @xcite .
having this in mind , in the following we will ignore @xmath22 .
we expect that intrinsic anharmonic effects of the membrane , important in a free standing membrane @xcite , will be unimportant in the presence of a substrate and we will therefore ignore them unless when discussing thermal expansion . in this approximation , in plane and out of plane modes decouple and we can ignore the term @xmath19 , leaving us with a harmonic theory . for a free standing membrane
, @xmath18 gives a quadratic dispersion relation @xmath37 , @xmath38 .
coupling to a static substrate will gap this dispersion relation and one would obtain @xmath39 , with @xmath40 .
the value of the constant @xmath36 greatly varies from substrate to substrate ( see table [ tab : material ] ) .
it was estimated in ref . to have a value of @xmath41 for graphene on sio@xmath4 .
for graphene on hbn its value can be estimated from density functional theory ( dft ) calculations @xcite to be around @xmath42 , depending on the orientation of graphene on hbn .
for the ( 111 ) surface of transition metal carbides it is of the order of @xmath43 , while for the ( 001 ) face it is approximately zero @xcite , as it is for graphene on platinum ( 111 ) @xcite .
the main object of interest from which all physically relevant quantities can be obtained is the height - height retarded green s function @xmath44\right\rangle,\ ] ] where @xmath45 means thermal and quantum average . ignoring @xmath19 it is possible to solve the theory exactly . as a matter of fact
, the problem reduces to that of two coupled harmonic oscillators .
therefore one obtains @xmath46^{-1},\ ] ] where @xmath47 ^{-1}$ ] is the correction to the free propagator due to coupling to the substrate , with @xmath48 the surface - to - surface substrate propagator similarly defined as in eq .
[ greens ] , with both fields @xmath49 evaluated at the surface of the substrate , @xmath50 , and the index @xmath51 meaning @xmath52 .
since @xmath20 is quadratic , to obtain @xmath48 it suffices to study the classical response of the substrate to an external pressure at the boundary @xmath50 .
this was done for an isotropic substrate in ref . and we generalized the result for the case of an uniaxial substrate ( see appendix [ appendix ] for details ) .
there are two specially relevant cases for the behaviour of @xmath48 : @xmath53 , for which @xmath54 ( replace @xmath55 for the uniaxial case ) ; and @xmath56 , for which @xmath57 , with @xmath58 for an isotropic medium ( where @xmath59 is the transverse / longitudinal sound velocity of the substrate ) .
therefore , at small @xmath12 , coupling to the substrate will lead to a contribution to @xmath60 proportional to @xmath12 , while the first order contribution to @xmath60 arising from intrinsic anharmonic effects in a free standing membrane is proportional to @xmath61@xcite . therefore
, anharmonic effects will be irrelevant when comparing to the effect of the substrate .
this justifies our approximation of neglecting @xmath19 .
a semi - infinite elastic medium supports a continuum of 3d bulk modes for @xmath10 ( replace @xmath62 for the uniaxial case ) . for @xmath7 ,
the substrate supports a 2d surface rayleigh mode with dispersion given by @xmath63 , with @xmath64 the rayleigh velocity .
therefore the effect of coupling to the substrate on the flexural mode will be twofold : ( i ) coupling to the substrate will gap the dispersion relation of flexural mode , @xmath65 , which we will refer to as the flexural - gapped ( fg ) mode , so that it will in general lie within the continuum of substrate bulk modes , which act as a dissipative bath , leading to a broadening of this branch .
( ii ) the flexural mode will also hybridize with the rayleigh mode ( already pointed out in ref . )
giving origin to another , unbroadened , branch , @xmath8 , which we will refer to as the flexural - rayleigh ( fr ) mode .
this information is encoded in the spectral / dissipation function , defined as , as a consequence of the canonical commutation relation @xmath66=i\hbar \delta(x - y)$ ] , where @xmath67 @xmath68 the fr mode appears in @xmath5 as a dirac delta function divergence at @xmath69 , with @xmath8 the solution of @xmath70 for @xmath7 , with a weight @xmath9 given by @xmath71 this situation is illustrated in fig .
[ fig : model]b , where we show a density plot of the spectral function along with the dispersion of the fr mode .
the gap of the fg mode is controlled by @xmath40 , while the broadening is controlled by @xmath72 ( replace @xmath55 for the uniaxial case ) . for @xmath15
one obtains @xmath73 and @xmath74 for graphene on sio@xmath4 .
this value is an overestimation comparing with the inverse relaxation times obtained from molecular dynamics simulations for acoustic flexural phonons on a sio@xmath4 substrate , @xmath75 @xcite .
the hybridization between flexural and rayleigh modes is more relevant for values of the spring constant @xmath36 such that @xmath76 , being maximum in this situation for @xmath77 and being suppressed for @xmath78 . [
cols="^,^,^,^,^,^,^,^,^,^,^",options="header " , ] ( a ) plot of @xmath79 for graphene on a sio@xmath4 substrate .
its behaviour changes from the one of a free membrane at large @xmath12 , @xmath80 , to the one of the substrate at low @xmath12 , @xmath81 .
( b ) plot of @xmath82 as a function of temperature for sio@xmath4 and hbn substrates ( solid lines ) and high temperature limits ( dashed lines ) .
@xmath15 was used for both substrates . , title="fig:",width=245 ] ( a ) plot of @xmath79 for graphene on a sio@xmath4 substrate .
its behaviour changes from the one of a free membrane at large @xmath12 , @xmath80 , to the one of the substrate at low @xmath12 , @xmath81 .
( b ) plot of @xmath82 as a function of temperature for sio@xmath4 and hbn substrates ( solid lines ) and high temperature limits ( dashed lines ) .
@xmath15 was used for both substrates .
, title="fig:",width=245 ]
it is interesting to compare our model with experimental data from ref . of phonon dispersion relations of graphene on different substrates .
this is shown in fig .
[ fig : exp ] .
although our continuous model fails at large momenta , it semi - quantitatively explains the lack of experimental data for the flexural mode at low momenta for graphene on light substrates , since the phonons become ill defined as quasi - particles .
also notice that experimentally there are indications of a rayleigh mode .
our model predicts that just by probing the carbon layer it is possible to detect the hybrid fr mode .
if this is the case or if what is experimentally seen comes from the fact that the first few layers of the substrate are also being probed is not clear .
the equal time height - height correlation function is given by @xmath83 which at high temperature reduces to @xmath84 .
interestingly , the dynamics of the substrate now makes @xmath85 at low momenta ( see fig . [
fig : height - height]a ) , since @xmath86 , while it would tend to a constant for a static substrate . for large distances
one therefore obtains @xmath87 , a result that is independent of @xmath36 and coincides with the result obtained for the surface out of plane displacement field of the bare substrate . for the quadratic dispersion relation of flexural phonons ,
it is known that @xmath88 diverges at any finite temperature , indicating the absence of crystalline order .
coupling to a substrate makes this result finite , as can be seen in fig .
[ fig : height - height]b . at high temperature one can obtain the following approximate expressions : @xmath89 it is interesting to notice that for small @xmath36 , the previous result coincides with the one that is obtained if one ignores the dynamics of a substrate @xcite . also of experimental interest
is the areal thermal expansion coefficient , @xmath90 , of graphene on a substrate .
it can be written in terms of the free energy @xmath91 , with @xmath92 the partition function , as @xmath93 where @xmath94 is the bulk modulus . under an isotropic expansion of the membrane @xmath95 , with a relative change of area given by @xmath96 ,
the term @xmath19 will generate a new quadratic term in the action @xmath97 . in a quasi - harmonic treatment
we will keep this new term while still ignoring anharmonic terms in the action . writing the partition function as a path integral in imaginary time , @xmath98e^{-s_{e}}$ ] with @xmath99 the euclidean action @xcite
, we can obtain @xmath90 from @xmath100 where @xmath101 is the matsubara height - height green s function .
performing the matsubara sum over bosonic frequencies @xmath102 one obtains a generalization of the result found in ref .
@xmath103 the obtained value is negative , since @xmath104=-\text{sgn}(\omega)$ ] , and tends to a constant at high temperature , approximately given by @xmath105 where @xmath106 is the debye momentum .
close to room temperature one obtains a value in the order of @xmath107 to @xmath108 , a value that is smaller in absolute value than the one obtained for a suspended membrane @xcite ( see fig .
[ fig : exp_resist]a ) . knowing @xmath109 one can also study the contribution to the electrical resistivity of the flexural phonons on doped supported graphene .
we compute the resistivity from the known formula @xmath110 where @xmath111 is graphene density of states at the fermi level , with @xmath112 and @xmath113 the fermi velocity and momentum respectively , and @xmath114 the transport scattering time . in order to compute @xmath114
, one must describe the electron - phonon interaction in graphene .
assuming that graphene is electronically weakly coupled to the substrate , the electron - phonon interaction in graphene should have the same form as the one in free standing graphene @xcite : @xmath115 where @xmath116 is the electron annihilation operator in the sublattice basis , @xmath117 is the bare deformation potential , @xmath118 describes the change in electron hopping with bond stretching @xcite , @xmath119 is the 2d pauli vector and @xmath120 is the vector potential induced by the distortion , @xmath121 notice that since the deformation potential is a coupling to the electronic density , it will be subject to screening @xcite .
focusing on the electron - flexural phonon interaction , after doing a fourier transform , writing the electron operator in the chiral basis and focusing only in scattering in the conduction band ( + ) , we can write @xmath122 with @xmath123 where @xmath124 . to compute the transport scattering time , @xmath125 , we first compute the electron lifetime , @xmath126 , in second order in the electron - flexural phonon interaction .
this can be obtained from the imaginary part of the ( retarded ) self energy , computed from the diagram in fig . [ fig : feydiagram ] .
considering only scattering in the conduction band one obtains @xmath127 to compute the transport scattering time , the sum in momentum must be weighted by the factor @xmath128 that appears due to vertex corrections when computing the conductivity @xcite . in the quasi - elastic approximation for
acoustic phonon scattering one ignores the phonon energy in the energy conserving delta function and sets @xmath129 , the electron fermi energy .
we finally obtain a generalization of the result from refs .
@xmath130 where @xmath131 is the squared electron - flexural phonon coupling , with @xmath132^{2}/\epsilon(q)^{2 } + \hbar^{2}v_{f}^{2}\beta^{2}/(8a^{2}))$ ] the generalized deformation potential @xcite and @xmath133 is the static dielectric function of graphene @xcite .
we have defined @xmath134 .
the electron scattering due to flexural phonons is a two phonon process .
therefore there are three contributions : ( i ) scattering by two fg modes , ( ii ) scattering by two fr modes , and ( iii ) a mixed process with scattering by one fr and one fg mode .
we see in fig .
[ fig : exp_resist]b that the contribution from fr modes is smaller in hbn than in sio@xmath4 .
this is explained by the smaller hybridization with the rayleigh mode in hbn due to a smaller rayleigh velocity when compared to sio@xmath4 .
we see that at room temperature the obtained resistivities are of order @xmath135 , a value that is much smaller than the expected contribution from the flexural mode in suspended samples ( @xmath136 ) @xcite and the contribution from in plane phonons ( @xmath137 ) @xcite .
second order sunset diagram contributing to the electron self energy .
the solid lines represent electron propagators , while wiggly lines represent height - height membrane propagators.,height=75 ]
we have modelled the dynamics of the flexural mode of a membrane coupled to a dynamical substrate , with the aim of understanding the role that the modified flexural mode might have in the physics of graphene on a substrate .
since a half - space elastic medium supports both 3d bulk modes and a 2d surface rayleigh mode , coupling of the membrane modes to the substrate leads to a splitting of the spectral weight of the flexural mode in two parts : one that will hybridize with the substrate rayleigh mode , acquiring an almost linear dispersion relation for @xmath138 , and a second branch which resembles the original flexural mode , which becomes gapped and is broadened by the continuum of substrate bulk modes .
this picture seems to be confirmed by experimental data @xcite .
as expected , coupling to the substrate leads to a stabilization of the membrane and all correlation functions become finite , while for a free membrane they are known to have infrared divergences in the harmonic theory .
it is worthwhile noticing that at high temperature the low momentum behaviour of the height - height correlation function changes from the @xmath80 of a free membrane to @xmath81 of the substrate .
this implies that for large distances the height - height function will go as @xmath139 .
we also explored the behaviour of the areal thermal expansion coefficient of graphene on substrate . at room temperature
we obtained a value of the order of @xmath107 to @xmath108 .
finally we studied the contribution of flexural modes on the electrical resistivity of doped graphene supported by a substrate .
we found that coupling of the membrane to the substrate strongly suppresses the contribution of the flexural phonons , even if one takes into account the contribution coming from the hybridized flexural - rayleigh mode .
note that the model describes a flat graphene layer on a flat surface .
this is a good approximation for substrates such as hbn . in the case of corrugated substrates , such as sio@xmath140
, our analysis is expected to describe the regions where the graphene layer and the substrate are flat and the two systems are in close contact .
the authors would like to thank j. schiefele , h. ochoa , e. cappelluti , r. roldan , n.m.r .
peres and m.i .
katsnelson for useful discussions .
b.a . acknowledges financial support from fundao para a cincia e a tecnologia , portugal , through grant no .
sfrh / bd/78987/2011 .
f.g . acknowledges financial support from mineco , spain , through grant fis2011 - 23713 , and the european research council advanced grants program , through grant 290846 .
( a ) thermal expansion coefficient of graphene on sio@xmath4 and hbn substrates ( solid lines ) and high temperature limits ( dashed lines ) . also shown
is the estimated high temperature thermal expansion coefficient for free standing graphene ( ) .
( b ) electrical resistivity due to flexural phonons in graphene on sio@xmath4 and hbn for a electronic density @xmath141 .
dashed lines show the individual contribution of scattering by two fr modes .
@xmath15 was used for both substrates . , title="fig:",width=245 ] ( a ) thermal expansion coefficient of graphene on sio@xmath4 and hbn substrates ( solid lines ) and high temperature limits ( dashed lines )
. also shown is the estimated high temperature thermal expansion coefficient for free standing graphene ( ) .
( b ) electrical resistivity due to flexural phonons in graphene on sio@xmath4 and hbn for a electronic density @xmath141 . dashed lines show the individual contribution of scattering by two fr modes .
@xmath15 was used for both substrates .
, title="fig:",width=245 ]
in a quadratic theory , the quantum mechanical retarded green s function coincides with the classical green s function . therefore , in order to determine the green s function of a semi - infinite elastic medium , @xmath142\right\rangle\ ] ] ( where @xmath143 ) , we study the classical response of the substrate to an external pressure , @xmath144 , at the boundary @xmath145 : @xmath146 where @xmath147 and we have used the fact that there is translational invariance in the @xmath148-plane and in time . the displacement field obeys the bulk equations of motion @xmath149 and the boundary conditions @xmath150 obtaining the solution for @xmath31 in the presence of the external pressure , we can read of @xmath151 .
we look for solutions of the form @xmath152 , with @xmath153 , where @xmath154 and @xmath155 are determined by the bulk equations of motion and the coefficients @xmath156 are fixed by the boundary conditions . in order to obtain a retarded response , we must pick the solutions for @xmath154 such that the real part has an opposite sign from @xmath157 ; and to obtain a finite response we pick the solutions that have a positive imaginary part .
we are interested in the response of the substrate at the boundary @xmath50 .
this can be written in matrix form as @xmath158 , from which one can read the coefficient @xmath159 which was simply written as @xmath48 in sec [ sec : model ] .
an uniaxial material has a plane of isotropy , having @xmath160 independent elastic constants . in voigt notation ,
the elastic constants tensor is given by : @xmath161.\ ] ] we can set @xmath162 without loss of generality .
the final result from the calculation is given by @xmath163,\nonumber \\
\delta_{0}^{zy}(q,\omega ) & = im^{-1}\left[c_{33}(p_{2}-p_{1})+c_{13}q(f_{2}-f_{1})\right],\nonumber \\ \delta_{0}^{zz}(q,\omega ) & = im^{-1}c_{44}\left(f_{1}p_{1}-f_{2}p_{2}\right),\end{aligned}\ ] ] where we have defined @xmath164 with @xmath165 43ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ] + 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty http://www.jstor.org/stable/3839379 [ , * * , ( ) ] link:\doibase 10.1103/revmodphys.81.109 [ * * , ( ) ] link:\doibase 10.1103/physrevb.71.205214 [ * * , ( ) ] link:\doibase 10.1126/science.1157996 [ * * , ( ) ] , link:\doibase 10.1103/physrevlett.102.046808 [ * * , ( ) ] @noop _ _ ( , , ) link:\doibase 10.1103/physrevlett.100.076801 [ * * , ( ) ] link:\doibase 10.1103/physrevb.82.195403 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.105.266601 [ * * , ( ) ] link:\doibase 10.1103/physrevb.83.235416 [ * * , ( ) ] link:\doibase 10.1016/j.physe.2011.03.017 [ * * , ( ) ] , link:\doibase 10.1103/physrevb.86.144103 [ * * , ( ) ] link:\doibase 10.1103/physrevb.84.075471 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.64.768 [ * * , ( ) ] link:\doibase 10.1103/physrevb.42.11469 [ * * , ( ) ] link:\doibase 10.1063/1.4712041 [ * * , ( ) ] http://stacks.iop.org/0295-5075/91/i=5/a=56001 [ * * , ( ) ] link:\doibase 10.1140/epje / i2010 - 10543 - 1 [ * * , ( ) ] http://stacks.iop.org/0953-8984/23/i=4/a=045009 [ * * , ( ) ] link:\doibase 10.1021/la990503 m [ * * , ( ) ] , @noop * * , ( ) link:\doibase 10.1016/0039 - 6028(90)90531-c [ * * , ( ) ] link:\doibase 10.1016/0039 - 6028(92)90046 - 9 [ * * , ( ) ] link:\doibase 10.1002/sia.1948 [ * * , ( ) ] link:\doibase 10.1021/nl101657v [ * * , ( ) ] , http://stacks.iop.org/0953-8984/24/i=10/a=104025 [ * * , ( ) ] link:\doibase 10.1103/physrevb.76.073103 [ * * , ( ) ] link:\doibase 10.1063/1.1388626 [ * * , ( ) ] link:\doibase 10.1103/physrevb.73.041402 [ * * , ( ) ] link:\doibase 10.1063/1.1727602 [ * * , ( ) ] link:\doibase 10.1063/1.1736249 [ * * , ( ) ] http://stacks.iop.org/0295-5075/43/i=3/a=355 [ * * , ( ) ] @noop _ _ ( , , ) link:\doibase 10.1103/physrevb.65.235412 [ * * , ( ) ] link:\doibase 10.1103/physrevb.76.045430 [ * * , ( ) ] link:\doibase 10.1016/j.physrep.2010.07.003 [ * * , ( ) ] @noop _ _ ( , , ) http://stacks.iop.org/1367-2630/8/i=12/a=318 [ * * , ( ) ] link:\doibase 10.1103/physrevb.77.115449 [ * * , ( ) ] link:\doibase 10.1103/physrevb.83.165405 [ * * , ( ) ] | out of plane vibrations are suppressed in graphene layers placed on a substrate .
these vibrations , in suspended samples , are relevant for the understanding of properties such as the electrical resistivity , the thermal expansion coefficient , and other .
we use a general framework to study the properties of the out of plane mode in graphene on different substrates , taking into account the dynamics of the substrate .
we discuss broadening of this mode and how it hybridizes with the substrate rayleigh mode , comparing our model with experimental observations .
we use the model to estimate the substrate induced changes in the thermal expansion coefficient and in the temperature dependence of the electrical resistivity . |
in the first four decades since their discovery , pulsars were almost the exclusive domain of radio ( and to a lesser extent x - ray ) astronomy .
indeed , of the @xmath12000 known pulsars , the majority were discovered in radio . since june 2008 ,
however , with the launch of the lat , on the _ fermi _ satellite , @xmath0-rays have become a viable means of discovering ( and studying ) pulsars .
more important than the number of lat - detected pulsars ( a small fraction of the overall population ) , the lat sample is subject to different biases than the radio sample .
lat pulsars are typically nearby ( @xmath1few kpc ) and energetic ( @xmath2 erg s@xmath3 ) , and a large fraction are _ radio - quiet_. furthermore , @xmath0-rays , unlike the radio beams , carry a significant fraction of the rotational energy of pulsars , thus providing a powerful probe into these extreme objects .
finally , lat pulsars provide a crucial input into our understanding of the overall neutron star population of the galaxy .
the lat is a pair conversion telescope consisting of a tracker , a calorimeter , and a segmented anti - coincidence detector , along with a programmable trigger and a complex data acquisition system . by incorporating the latest advances from particle physics '' . ] , including the use of silicon - strip detectors , the lat has achieved a giant leap in capabilities compared to its predecessor , egret .
the lat extends to higher energies ( @xmath4300 gev vs @xmath110 gev ) , has a larger effective area and field of view , improved angular and energy resolutions , and much lower deadtime ( ( * ? ? ?
* atwood et al . 2009 ) ) . in its first 4 years of operations ,
the lat has collected @xmath4200 million `` source '' class events , compared to @xmath11.5 million photons collected by egret , in its 9-year lifetime .
regarding pulsars , the progress made by the lat has been spectacular .
the @xmath0-ray pulsar population has grown from 7 firm detections ( plus a few candidates ) at the time of the _ fermi _ launch ( ( * ? ? ? * thompson 2008 ) ) , to 117 detections in three years of lat survey observations ( the second _ fermi_-lat catalog of @xmath0-ray pulsars , _ in preparation _ ) . beyond the jump in the number of pulsars , the increased statistics enable detailed ( e.g. phase - resolved ) studies on individual pulsars , previously out of reach .
the lat has also opened up the unexplored 10100 gev window .
egret detected a mere handful of @xmath410 gev photons from the brightest pulsars ( ( * ? ? ? * thompson 2005 ) ) , whereas the lat detects significant emission from over two dozen pulsars in this energy range and even @xmath425 gev events from the brightest ones ( e.g. crab , vela , geminga , see ( * ? ? ?
* saz parkinson 2012 ) ) . of the 117 lat detections ,
61 pulsars were known prior to _
fermi_. pulsations were obtained by folding the @xmath0-rays with a radio ( or x - ray ) timing model .
of these 61 , 20 are msps ( e.g. ( * ? ? ?
* abdo et al . 2009c ) ) , a hitherto unknown class of strong @xmath0-ray emitters .
more surprisingly , a large number of msps ( @xmath440 ) were discovered in radio searches of lat unassociated sources ( ( * ? ? ?
* ray et al . 2012 ) ) .
most ( 20 so far , e.g. ( * ? ? ?
* ransom et al . 2011 ) ) will likely exhibit @xmath0-ray pulsations - ray sources selected precisely for their _ pulsar - like _ qualities .
] , once enough data and/or a precise timing model is obtained ( usually from radio observations ) .
finally , 36 of the 117 @xmath0-ray pulsars were discovered directly in blind searches of lat data ( ( * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* abdo et al .
2009b , saz parkinson et al .
2010 , saz parkinson 2011 , pletsch et al .
2012a , pletsch et al . 2012b ) ) .
the rest of this paper discusses the challenges , results , and prospects of these searches .
two factors make @xmath0-ray searches for pulsars particularly challenging , compared to radio searches .
the first involves the scarcity of events .
the lat detects @xmath11 @xmath0-ray per day from a typical ( bright ) pulsar .
this means that lat searches for pulsars span months to years ( see figure 1 ) .
the second complication involves the broad , energy - dependent psf of the instrument ( from @xmath5 at 100 mev to @xmath6 at 100 gev , 68% containment , normal incidence ) .
this results in significant source confusion , especially in high background regions like the galactic plane , where the diffuse @xmath0-ray emission makes it hard to resolve individual sources .
essentially , it is impossible to select , with certainty , events coming from a source , so maximum likelihood techniques must be employed . using these techniques ,
it is possible to determine the probability that each event is coming from the source of interest and improve the sensitivity of the search by assigning a weight to each event equal to this probability ( * ? ? ?
* ( kerr 2011 ) ) .
-ray data ( right ) .
* left * psr j1745 + 1017 , an msp discovered in radio searches of _ fermi _ unassociated sources with the effelsberg telescope ( barr et al .
2012 , mnras , _ submitted_. figure credit : e. barr ) .
* right * psr j1957 + 5033 , a _ radio - quiet _ @xmath0-ray pulsar discovered in a lat blind search .
note that , although the pulse profiles and significances are similar , the @xmath0-ray observation is @xmath110,000 times longer , and the number of @xmath0-ray events ( 582 ) is approximately one per day ( @xmath1200,000 pulsar rotations ! ) . on the other hand , unlike in radio , there is no need to scan in dm in a @xmath0-ray search.,title="fig : " ] -ray data ( right ) . *
left * psr j1745 + 1017 , an msp discovered in radio searches of _ fermi _ unassociated sources with the effelsberg telescope ( barr et al .
2012 , mnras , _ submitted_. figure credit : e. barr ) .
* right * psr j1957 + 5033 , a _ radio - quiet _ @xmath0-ray pulsar discovered in a lat blind search .
note that , although the pulse profiles and significances are similar , the @xmath0-ray observation is @xmath110,000 times longer , and the number of @xmath0-ray events ( 582 ) is approximately one per day ( @xmath1200,000 pulsar rotations ! ) . on the other hand , unlike in radio , there is no need to scan in dm in a @xmath0-ray search.,title="fig : " ] searches over long ( sparse ) data sets make standard fft techniques impractical . to tackle this problem , ( * ? ? ?
* atwood et al . ( 2006 ) ) developed the `` time - differencing technique '' and applied it successfully to egret data ( ( * ? ? ?
* ziegler et al . 2008 ) ) .
the core of the technique involves computing the fft of the differences between the times of events ( up to a maximum , sliding , time window ) , rather than using the original time series . using a time window of @xmath1weeks significantly alleviates the coherence requirements of the search . within weeks of the launch of _ fermi _ , the first _ radio - quiet _
@xmath0-ray pulsar was discovered in the supernova remnant cta 1 ( ( * ? ? ?
* abdo et al . 2008 ) ) .
this was followed by many more discoveries ( e.g. ( * ? ? ?
* ; * ? ? ?
* abdo et al .
2009b , saz parkinson et al .
many of these early @xmath0-ray pulsars discovered by the lat are coincident with old egret unidentified sources .
many , like the pulsar in cta 1 , are also coincident with known supernova remnants or pulsar wind nebulae , and were thus long suspected of hosting pulsars ( e.g. psr j1836 + 5925 , j2021 + 4026 ) .
radio follow - up observations of the lat - discovered pulsars showed that most of them are _ radio - quiet _
( or extremely radio - faint ) , suggesting that the @xmath0-ray emission originates far from the neutron star surface , resulting in broad beams .
it is now clear that these radio - quiet @xmath0-ray pulsars represent a significant fraction of the neutron star population of the galaxy ( see guillemot et al .
( 2012 ) , in these proceedings , for a detailed review of the radio observations of lat @xmath0-ray pulsars ) .
all 36 pulsars found in lat blind searches to date are young ( @xmath7 yr ) energetic ( @xmath8 erg s@xmath9 erg s@xmath3 ) pulsars with frequencies below @xmath120 hz .
these pulsars often exhibit timing irregularities , in the form of _ timing noise _ and glitches .
since most are not detected in radio , the lat is the only instrument capable of timing them ( ( * ? ? ?
* ray et al . 2011 ) ) .
while it is possible to time these noisy ( or `` glitchy '' ) pulsars , a good timing model often requires many frequency derivatives and other whitening terms , making their _ discovery _ over long data spans extremely challenging .
searches for these pulsars may only be possible over lat observing periods lasting @xmath1months , rather than years ( see figure 2 ) . as mentioned above , the discovery of @xmath0-ray msps ( ( * ? ?
* abdo et al . 2009c ) ) was largely unexpected .
equally unexpected was the large number of msps discovered by radio telescopes searching lat unassociated sources ( ( * ? ? ?
* ray et al .
this fact , combined with the large fraction of _ radio - quiet _ young pulsars , raises the question of whether large large numbers of _ radio - quiet _ @xmath0-ray msps remain to be discovered in blind searches of lat data .
the answer , from examining the identified @xmath0-ray sources in the _ almost _ flux - limited _
fermi_-lat bright source list ( ( * ? ? ?
* abdo et al . 2009a ) ) , appears to be no . whereas at least two thirds of young @xmath0-ray pulsars are _
radio quiet _
, at most one third of @xmath0-ray msps are ( ( * ? ? ?
* romani 2012 ) ) .
blind searches for msps are vastly more complicated than searches for young pulsars .
firstly , a majority ( @xmath180% ) of msps are in binary systems , and a full blind search over unknown orbital parameters is out of reach given current computer capabilities .
searches for isolated msps are possible , but the high frequencies increase the memory and cpu requirements .
positional uncertainties become more relevant with increasing frequencies and the typical tolerance of a blind search using several years of data is of order a fraction of an arc second .
this means that one must either perform a fine scan over the lat positions , or else identify a precise position of a plausible counterpart , using multi - wavelength ( e.g. x - ray ) observations .
a number of lat sources have been identified as promising pulsar candidates , by virtue of their variability and spectral properties ( ( * ? ? ?
* ackermann et al . 2012 ) ) .
more recently , x - ray and optical studies of some of the brightest of these have identified some strong `` black widow '' candidates : eclipsing binary msps with very compact , almost circular orbits , where the pulsar is destroying its low - mass companion ( ( * ? ? ?
* ; * ? ? ?
* romani & shaw 2011 , romani 2012 ) ) .
these studies derive extremely precise positions ( to @xmath10 ) and stringent constraints on two out of the three requisite orbital parameters ( leaving only the projected semi - axis relatively unconstrainted ) .
thus , for the first time , searches for such binary msps using @xmath0-ray data are possible .
a deep lat search of 2fgl j2339.6 + 0532 ( the most promising of these `` black widow '' candidates ) unfortunately produced no significant candidate , but further efforts are in progress ( see belfiore ( 2012 ) for details ) .
understanding the @xmath0-ray emission from the galactic center ( gc ) region is both challenging and controversial .
claims that @xmath0-ray emission from the gc may be related to dark matter ( e.g. ( * ? ? ? * ; * ? ? ?
* hooper & goodenough 2011 , weniger 2012 ) ) should be measured up against more conventional explanations , such as a possible origin from @xmath0-ray pulsars .
as a site of massive star formation , it likely contains thousands of pulsars , but their radio detection is hampered by the large amount of interstellar scattering .
nevertheless , some pulsars have been discovered fairly close to the gc , and predictions for the number of radio pulsars that could be associated with the gc range in the thousands ( ( * ? ? ?
* deneva et al . 2009 ) ) .
it is possible that some of the undiscovered pulsars in the region are _ radio quiet _
, like the majority of young pulsars found in lat blind searches ( e.g. psr j1732 - 3131 , see figure 3 ) .
blind searches for @xmath0-ray pulsars around the gc are affected by low fluxes ( due to the large distance ) , and high levels of diffuse emission . indeed , most lat pulsars are likely nearby ( @xmath1few kpc ) .
the lat has , however , detected pulsations from msp j1823 - 3021a , in the globular cluster ngc 6624 , at 8.4 kpc ( roughly the distance to the gc ) .
it is possible to estimate how far blind searches might be sensitive out to . from scaling arguments , considering that the @xmath0-ray flux of the crab is several thousand times brighter than the faintest @xmath0-ray pulsar discovered in a blind search ,
we conclude that it is possible to discover a _
crab - like _ @xmath0-ray pulsar in a lat blind search out to at least @xmath115 kpc .
perhaps the biggest complication in searching for faint pulsars around the gc is the fact that such a pulsar may be young and _ noisy _ ( or `` glitchy '' ) .
as described above , the loss of coherence of the signal limits the amount of data that can be effectively searched ( see figure 2 ) .
thus , regardless of the number of years of lat data available , it may only be possible to search several months at a time .
thus , the sensitivity of searches in this region may only improve significantly with improvements in reconstruction ( e.g. pass 8 , see baldini ( 2011 ) , _ fermi _ symposium ) , or a change in the observing mode . for the first four years of its mission ,
_ fermi _ has operated mostly in survey mode .
this has many advantages for most of the scientific goals of the mission , including pulsar searches ( and timing ) . modifying the observing profile , however ,
could enhance the sensitivity to detection of faint signals ( both from pulsars and/or dark matter ) from the gc .
figure 3 ( right ) shows the relative gain in exposure ( compared to survey mode ) with a modified mode in which the lat points to a location slightly offset from the gc . whenever the gc is occulted , the lat could go back into survey mode , thus maintaining some level of exposure over the entire sky . 3 , while the exposure for other parts of the sky decreases to roughly one third , in the worst case . , title="fig : " ] 3 , while the exposure for other parts of the sky decreases to roughly one third , in the worst case . , title="fig : " ]
the nasa senior review recently recommended that _ fermi _ operations continue through 2016 . in principle , _ fermi _ could continue to operate well beyond this date .
as the lat accumulates more data , it will detect @xmath0-ray pulsations from ever fainter _ radio - loud _ pulsars .
as for blind searches of @xmath0-ray data , they too will continue to produce new discoveries , although the increasing computational demands will require additional resources ( e.g. einstein@home[multiblock footnote omitted ] ) and efficient computational techniques to exploit them .
finally , multi - wavelength observations ( radio , x - rays , and optical ) , will continue to play a crucial role in lat pulsar studies , especially in facilitating the search for more exotic pulsar systems , such as `` black widow '' systems , young binary pulsars , and pulsars around the gc . in this last case , a modified observing profile enhancing the exposure to this region ( by a factor of @xmath13 )
could play a crucial role in future discoveries .
the @xmath11 lat collaboration acknowledges support from a number of agencies and institutes for both development and the operation of the lat as well as scientific data analysis .
these include nasa and doe in the united states , cea / irfu and in2p3/cnrs in france , asi and infn in italy , mext , kek , and jaxa in japan , and the k. a. wallenberg foundation , the swedish research council and the national space board in sweden . additional support from inaf in italy and cnes in france for science analysis during the operations phase is also gratefully acknowledged .
i acknowledge the support of the american astronomical society and the national science foundation in the form of an international travel grant , which enabled me to attend this conference .
abdo , a. a. , ackermann , m. , atwood , w. b. , et al .
2008 , science , 322 , 1218 abdo , a. a. , ackermann , m. , ajello , m. , et al . 2009a , , 183 , 46 abdo , a. a. , ackermann , m. , ajello , m. , et al .
2009b , science , 325 , 840 abdo , a. a. , ackermann , m. , ajello , m. , et al . 2009c , science , 325 , 848 ackermann , m. , et al .
2012 , , 753 , 83 atwood , w. b. , ziegler , m. , johnson , r. p. , & baughman , b. m. 2006 , , 652 , l49 atwood , w. b. , abdo , a. a. , ackermann , m. , et al .
2009 , , 697 , 1071 belfiore , a. , phd thesis , universit degli studi di pavia , 2012 deneva , j. s. , cordes , j. m. , & lazio , t. j. w. 2009 , , 702 , l177 hooper , d. , & goodenough , l. 2011 , physics letters b , 697 , 412 kerr , m. 2011 , , 732 , 38 pletsch , h. j. , guillemot , l. , allen , b. , et al .
2012a , , 744 , 105 pletsch , h. j. , guillemot , l. , allen , b. , et al . 2012b , , 755 , 20 ransom , s. m. , phd thesis , harvard university , 2001 ransom , s. m. , ray , p. s. , camilo , f. , et al .
2011 , , 727 , l16 ray , p. s. , kerr , m. , parent , d. , et al .
2011 , , 194 , 17 ray , p. s. , abdo , a. a. , parent , d. , et al .
2012 , arxiv:1205.3089 romani , r. w. , & shaw , m. s. 2011 , , 743 , l26 romani , r. w. 2012 , , 754 , l25 saz parkinson , p. m. , dormody , m. , ziegler , m. , et al .
2010 , , 725 , 571 saz parkinson , p. m. 2011 , arxiv:1101.3096 saz parkinson , p. m. 2012 , proceedings , @xmath0 2012 , heidelberg , germany , 9 - 13 july 2012 thompson , d. j. 2005 , , 157 , 324 , arxiv : astro - ph/0412376 thompson , d. j. 2008 , reports on progress in physics , 71 , 116901 weniger , c. 2012 , , 8 , 7 ziegler , m. , baughman , b. m. , johnson , r. p. , & atwood , w. b. 2008 , , 680 , 620 | the large area telescope ( lat ) on the _ fermi _ satellite is the first @xmath0-ray instrument to discover pulsars directly via their @xmath0-ray emission .
roughly one third of the 117 @xmath0-ray pulsars detected by the lat in its first three years were discovered in blind searches of @xmath0-ray data and most of these are undetectable with current radio telescopes .
i review some of the key lat results and highlight the specific challenges faced in @xmath0-ray ( compared to radio ) searches , most of which stem from the long , sparse data sets and the broad , energy - dependent point - spread function ( psf ) of the lat .
i discuss some ongoing lat searches for @xmath0-ray millisecond pulsars ( msps ) and @xmath0-ray pulsars around the galactic center .
finally , i outline the prospects for future @xmath0-ray pulsar discoveries as the lat enters its extended mission phase , including advantages of a possible modification of the lat observing profile . |
the flavor mixing pattern in the lepton sector is quite different from the mixing pattern in the quark sector . in the quark sector ,
one of the mixing angles is around @xmath1 and the other two mixing angles are very small whereas in the lepton sector two of the mixing angles ( atmospheric mixing angle @xmath2 and solar mixing angle @xmath3 ) are large and the third mixing angle ( reactor mixing angle @xmath4 ) which was recently measured in a number of neutrino oscillation experiments @xcite is around @xmath5 . to explain the mixing pattern in the lepton sector ,
a number of theoretical ideas have been proposed .
a particular approach for explaining the lepton flavor mixing pattern is based on non - abelian discrete symmetries which predict values of mixing matrix elements independent of the lepton masses and are known as mass independent textures .
some typical mixing patterns obtained using the above approach are tribimaximal mixing @xcite , bimaximal mixing @xcite , golden ratio - i @xcite , golden ratio - ii @xcite , hexagonal mixing @xcite all of which predict a vanishing reactor mixing angle . however , in the light of recent experimental data @xcite , these mixing patterns need modifications .
the other approaches used to explain lepton flavor mixing relate mixing matrix elements to lepton masses include texture zeros @xcite , vanishing minors @xcite , hybrid textures @xcite , equalities between the elements @xcite .
+ two texture zeros in the effective neutrino mass matrix ( @xmath6 ) in a basis where the charged lepton mass matrix ( @xmath7 ) is diagonal have been extensively studied in the past @xcite . out of the 15 possible cases of two texture zeros in @xmath6 ,
only 7 are compatible with the present neutrino oscillation data .
the seven allowed cases of two texture zeros in the nomenclature of ref .
@xcite are listed in table [ tab1 ] .
+ .[tab1]viable two texture zero neutrino mass matrices .
@xmath8 denotes the non - zero elements . [ cols="^,^,^,^",options="header " , ] the numerically estimated mass matrices for classes @xmath9 and @xmath10 are given in table [ tab3 ] .
these mass matrices are obtained for the best fit values of @xmath11 and @xmath4 .
the numerical mass matrices for @xmath12 and @xmath13 patterns can be obtained from table [ tab3 ] with the operation of @xmath14 [ eq . ( [ eq14 ] ) ] on the mass matrices of patterns @xmath9 and @xmath10 , respectively .
in this section , we show how the texture zero structures discussed in this work can be obtained using discrete abelian family symmetries .
general guidelines for the symmetry realization of texture zeros in both the quark and the lepton mass matrices have been enunciated in ref .
@xcite which outlines procedures for enforcing texture zeros at any place in the fermion mass matrices by imposing discrete abelian family symmetries . to obtain the desired texture structures for class @xmath15 , we use the framework of type - i+ii seesaw mechanism @xcite .
+ the type - i seesaw @xcite contribution to the effective neutrino mass matrix is given by @xmath16 where @xmath17 and @xmath18 are the dirac and the right - handed neutrino mass matrices , respectively .
for the type - i seesaw mechanism to work , the standard model ( sm ) is extended by adding three right - handed neutrinos ( @xmath19 ) . for classes @xmath10 and @xmath13 , we will also need one @xmath20 singlet scalar ( @xmath21 ) .
for the type - ii seesaw @xcite contribution to the effective neutrino mass matrix , we need a scalar @xmath20 triplet higgs ( @xmath22 ) .
the effective neutrino mass matrix containing both type - i+ii seesaw contributions is given by @xmath23 where @xmath24 denotes the type - ii seesaw contribution .
the symmetry realization of all the presently allowed classes of two texture zeros in @xmath6 in the flavor basis has been presented in @xcite in the context of type - ii seesaw mechanism and in ref .
@xcite considering both type - i+ii seesaw contributions and using a minimal cyclic symmetry group .
classes @xmath9 and @xmath12 have also been realized in ref .
@xcite using @xmath25 or its @xmath0 subgroup in the context of type - i+ii seesaw mechanism .
also , classes @xmath10 and @xmath13 have been obtained in ref .
@xcite by softly breaking the @xmath26 symmetry .
+ for the symmetry realization , we use the discrete cyclic group @xmath0 . for class @xmath9 , we assume the following transformation properties of the leptonic fields under the action of @xmath0 symmetry : @xmath27 where @xmath28 = @xmath29 .
@xmath30 @xmath31 denotes @xmath20 doublets and @xmath32 denote the right - handed @xmath20 singlet charged lepton and neutrino fields , respectively . according to the above transformations of the leptonic fields , the bilinears @xmath33 , @xmath34 and @xmath35 relevant for @xmath7 , @xmath17 and @xmath18 respectively , transform as @xmath36 the sm higgs doublet is assumed to remain invariant under @xmath0 leading to diagonal @xmath7 , @xmath17 .
@xmath18 has the following form : @xmath37 this leads to the following type - i seesaw contribution to the effective neutrino mass matrix : @xmath38 we add a higgs triplet @xmath22 which transforms as @xmath39 under @xmath0 , leading to the following type - ii seesaw contribution to the effective neutrino mass matrix : @xmath40 the effective neutrino mass matrix after the complete type - i+ii seesaw contributions has the form of class @xmath9 viz .
@xmath41 in table [ tab4 ] , we have summarized the transformation properties of lepton and scalar fields under the action of @xmath0 for classes @xmath9 , @xmath12 , @xmath10 and @xmath13 .
+ for classes @xmath10 and
@xmath13 we need an @xmath20 singlet scalar for obtaining the desired form of @xmath18 ( see table [ tab5 ] ) corresponding to these two classes .
we have closely followed our earlier work @xcite for the symmetry realization of classes @xmath10 and @xmath13 . also , for the symmetry realization of classes @xmath9 and @xmath12 we need to softly break the scalar potential by allowing the @xmath0 forbidden , dimension three term @xmath42 . this term is required to obtain a non - zero and small vev of the scalar triplet ( @xmath22)@xcite . + from table [ tab3 ] , we can see that for each class there is a hierarchy between the magnitudes of the non - zero elements of @xmath6 .
the magnitude of @xmath43 and @xmath44 elements is at least an order of magnitude greater than the rest of the non - zero elements of @xmath6 . to account for the hierarchy between the non - zero elements of the neutrino mass matrices of classes @xmath9 , @xmath12 , @xmath10 and @xmath13
, we assume the dominance of type - ii seesaw contribution . as can be seen from table [ tab5 ] , for all the classes ( @xmath9 , @xmath12 , @xmath10 , @xmath13 ) considered in this work , the dominant type - ii seesaw contributes to @xmath43 and @xmath44 elements of @xmath6 in each case , thus , accounting for the hierarchy between the non - zero elements of @xmath6 .
all the above classes of texture zeros are realized at the seesaw scale which poses the question whether the texture zeros realized in this work survive when the rg evolution of @xmath6 from the seesaw to the electroweak scale is taken into account . at one loop ,
the rg running of the neutrino mass matrix below the seesaw scale is described by the rg equation @xcite : @xmath45 where @xmath46 denotes the effective dimension five neutrino mass operator @xcite , @xmath47 is the yukawa coupling matrix for the charged leptons , @xmath48 in the sm and @xmath49 denotes the contribution from gauge interactions .
the renormalization scale @xmath50 enters through @xmath51 . from eq .
( [ eq27 ] ) one can see that in the flavor basis where the charged lepton yukawa coupling matrix is diagonal , the radiative corrections to each element of the effective neutrino mass matrix are multiplicative , so that a zero entry remains zero .
+ the situation changes at energies larger than the mass scale of the lightest right - handed neutrino . above this threshold ,
the neutrino yukawa couplings @xmath52 also contribute to the rg equations . above the highest seesaw scale ,
the effective neutrino mass matrix is defined as @xmath53 where @xmath54 denotes the higgs doublet vacuum expectation value ( vev ) . between the mass thresholds ,
the singlet neutrinos are successively integrated out which leads to modifications in running . in the full type - i+ii seesaw scenario ,
the running of the effective neutrino mass matrix @xmath6 above and between the seesaw scales is given by the running of following three different contributions to @xmath6 : @xmath55 where @xmath56 .
the one loop @xmath57-functions for the effective neutrino mass matrix in various effective theories can be summarized as @xcite @xmath58^t m_\nu^{(i ) } + m_\nu^{(i ) } [ c_l y_l y_l^\dagger + c_\nu y_\nu y_\nu^\dagger + c_\triangle y_\triangle y_\triangle^\dagger ] + \xi m_\nu^{(i)}\ ] ] where @xmath59 denotes any of the three contributions to the effective neutrino mass matrix .
+ for all the yukawa coupling matrices realized in this work , the hermitian products @xmath60 ( @xmath61 ) which are relevant for the rg evolution of @xmath6 come out to be diagonal so that rg corrections are again multiplicative on the effective neutrino mass matrix elements leaving zero elements intact .
thus , although , the values of neutrino masses and neutrino mixing parameters change due to rg corrections while running down from the seesaw scale to the electroweak scale , the correlations induced between neutrino masses and mixing parameters by texture zeros remain unchanged due to the stability of texture zeros .
the two texture zero neutrino mass matrices in the flavor basis have been fairly successful in describing neutrino masses and mixings . we have studied the implications of large effective majorana neutrino mass for a class of two texture zero neutrino mass matrices .
we found that classes @xmath9 , @xmath12 , @xmath10 and @xmath13 all predict near maximal atmospheric neutrino mixing angle when supplemented with the assumption of large effective majorana neutrino mass .
moreover , the near maximality of the atmospheric neutrino mixing angle is independent of the values of the solar and reactor neutrino mixing angles .
furthermore , we have shown how one can obtain such texture structures in the context of type - i+ii seesaw using the discrete abelian group @xmath0 .
assuming type - ii seesaw dominance one can explain the hierarchy between the non - zero elements of the effective neutrino mass matrix .
the hermitian products @xmath60 for various yukawa coupling matrices realized in this work are all diagonal , leading to the stability of texture zeros against running from the seesaw to the electroweak scale .
the assumption of large @xmath62 is motivated by the fact that there are a number of forthcoming and presently ongoing experiments searching for neutrinoless double beta decay .
these experiments are capable of confirming or ruling out large @xmath62 .
p. f. harrison , d. h. perkins and w. g. scott , _ phys . lett .
_ * b 530 * , 167 ( 2002 ) , hep - ph/0202074 ; p. f. harrison and w. g. scott , _ phys . lett . _ * b 535 * , 163 ( 2002 ) , hep - ph/0203209 ; zhi - zhong xing , _ phys
* b 533 * , 85 ( 2002 ) , hep - ph/0204049 .
f. vissani , hep - ph/9708483 ; v. d. barger , s. pakvasa , t. j. weiler and k. whisnant , _ phys .
lett . _ * b 437 * , 107 ( 1998 ) , hep - ph/9806387 ; a. j. baltz , a. s. goldhaber and m. goldhaber , _ phys .
lett . _ * 81 * , 5730 ( 1998 ) , hep - ph/9806540 .
a. datta , f. s. ling and p. ramond , _ nucl .
* b 671 * , 383 ( 2003 ) , hep - ph/0306002 ; y. kajiyama , m. raidal and a. strumia , _ phys .
* d 76 * , 117301 ( 2007 ) , arxiv:0705.4559 [ hep - ph ] ; l. l. everett , a. j. stuart , _ phys .
* d 79 * , 085005 ( 2009 ) , arxiv:0812.1057 [ hep - ph ] ; f. feruglio , a. paris , _ jhep _ * 1103 * , 101 ( 2011 ) , arxiv:1101.0393 [ hep - ph ] .
paul h. frampton , sheldon l. glashow and danny marfatia , _ phys .
* b 536 * , 79 ( 2002 ) , hep - ph/0201008 .
zhi - zhong xing , _ phys .
* b 530 * , 159 ( 2002 ) , hep - ph/0201151 ; bipin r. desai , d. p. roy and alexander r. vaucher , _ mod .
_ * a 18 * , 1355 ( 2003 ) , hep - ph/0209035 ; a. merle , w. rodejohann , _ phys
* d 73 * , 073012 ( 2006 ) , hep - ph/0603111 ; s. dev , sanjeev kumar , s. verma and s. gupta , _ nucl .
* b 784 * , 103 ( 2007 ) , hep - ph/0611313 ; g. ahuja , s. kumar , m. randhawa , m. gupta , s. dev , _ phys .
rev . _ * d 76 * , 013006 ( 2007 ) , hep - ph/0703005 ; s. kumar , _ phys .
_ * d 84 * , 077301 ( 2011 ) , arxiv:1108.2137 [ hep - ph ] ; d. meloni , g. blankenburg , _ nucl . phys . _ * b 867 * , 749 ( 2013 ) , arxiv:1204.2706 [ hep - ph ] ; w. grimus , p. o. ludl , _ j. phys . _
* g 40 * , 055003 ( 2013 ) arxiv:1208.4515 [ hep - ph ] ; manmohan gupta , gulsheen ahuja , _ int . j. mod .
phys . _ * a , 27 * , 1230033 ( 2012 ) , arxiv:1302.4823 [ hep - ph ] ; j. liao , d. marfatia , k. whisnant , arxiv:1311.2639 [ hep - ph ] ; d. meloni , a. meroni , e. peinado , _ phys . rev .
_ * d 89 * ( 2014 ) 053009 , arxiv:1401.3207 [ hep - ph ] .
l. lavoura , _ phys .
* b 609 * , 317 ( 2005 ) , hep - ph/0411232 ; e. i. lashin , n. chamoun , _ phys .
* d 80 * , 093004 ( 2009 ) , arxiv:0909.2669 [ hep - ph ] ; s. dev , s. verma , s. gupta and r. r. gautam , _ phys .
* d 81 * , 053010 ( 2010 ) , arxiv:1003.1006 [ hep - ph ] ; s. dev , s. gupta and r. r. gautam , _ mod .
lett . _ * a 26 * , 501 ( 2011 ) , arxiv:1011.5587 [ hep - ph ] ; t. araki , j. heeck and j. kubo , _ jhep _ * 1207 * , 083 ( 2012 ) , arxiv:1203.4951 [ hep - ph ] ; jiajun liao , danny marfatia , kerry whisnant , _ phys .
_ * d 88 * , 033011 ( 2013 ) , arxiv:1306.4659 [ hep - ph ] ; weijian wang , arxiv:1311.6944 [ hep - ph ] ; weijian wang , arxiv:1402.6808 [ hep - ph ] .
s. kaneko , h. sawanaka and m. tanimoto , _ jhep _ * 0508 * , 073 ( 2005 ) , hep - ph/0504074 ; s. dev , s. verma and s. gupta , _ phys . lett .
_ * b 687 * , 53 - 56 ( 2010 ) , arxiv:0909.3182 [ hep - ph ] ; s. dev , s. gupta and r. r. gautam , _ phys .
* d 82 * , 073015 ( 2010 ) arxiv:1009.5501 [ hep - ph ] ; ji - yuan liu , shun zhou , _ phys .
_ * d 87 * , 093010 ( 2013 ) , arxiv:1304.2334 [ hep - ph ] ; weijian wang , _ eur .
j. _ * c 73 * , 2551 ( 2013 ) , arxiv:1306.3556 [ hep - ph ] ; s. dev , r. r. gautam and lal singh , _ phys .
_ * d 88 * , 033008 ( 2013 ) , arxiv:1306.4281 [ hep - ph ] .
f. t. avignone iii , s. r. elliott , j. engel , _ rev .
phys . _ * 80 * , 481 ( 2008 ) , arxiv:0708.1033 [ nucl - ex ] ; j. j. gomez - cadenas , j. martin - albo , m. mezzetto , f. monrabal , m. sorel , _ riv .
nuovo cim . _ * 35 * , 29 ( 2012 ) , arxiv:1109.5515 [ hep - ex ] ; s. m. bilenky , c. giunti , _ mod .
lett . _ * a 27 * , 1230015 , arxiv:1203.5250 [ hep - ph ] .
p. minkowski , _ phys .
* b 67 * , 421 ( 1977 ) ; t. yanagida , _ proceedings of the workshop on the unified theory and the baryon number in the universe _
( o. sawada and a. sugamoto , eds . ) , kek , tsukuba , japan , 1979 , p. 95 : m. gell - mann , p. ramond , and r. slansky , _ complex spinors and unified theories in supergravity _ ( p. van nieuwenhuizen and d. z. freedman , eds . ) , north holland , amsterdam , 1979 , p.315 ; r. n. mohapatra and g. senjanovic , _ phys .
* 44 * , 912 ( 1980 ) .
w. konetschny and w. kummer , _ phys .
* b 70 * , 433 ( 1977 ) ; t. p. cheng and l. f. li , _ phys .
* d 22 * , 2860 ( 1980 ) ; j. schechter and j. w. f. valle , _ phys .
_ * d 22 * , 2227 ( 1980 ) ; g. lazarides q. shafi and c. wetterich , _ nucl .
_ * b 181 * , 287 ( 1981 ) ; r. n. mohapatra and g. senjanovic , _ phys . rev . _ * d 23 * , 165 ( 1981 ) .
p. h. chankowski , z. pluciennik , _ phys .
* b 316 * , 312 ( 1993 ) , hep - ph/9306333 ; k. s. babu , c. n. leung , j. pantaleone , _ phys .
* b 319 * , 191 ( 1993 ) , hep - ph/9309223 ; s. antusch , m. drees , j. kersten , m. lindner , m. ratz , _ phys .
* b 519 * , 238 ( 2001 ) , hep - ph/0108005 .
s. f. king , n. n. singh , _ nucl .
* b 591 * , 3 - 25 ( 2000 ) , hep - ph/0006229 ; s. antusch , j. kersten , m. lindner , m. ratz , _ phys .
lett . _ * b 538 * , 87 ( 2002 ) , hep - ph/0203233 ; s. antusch , j. kersten , m. lindner , m. ratz , m. a. schmidt , _ jhep _ * 0503 * , 024 ( 2005 ) , hep - ph/0501272 ; shamayita ray , _ int . j. mod
. phys . _ * a 25 * , 4339 - 4384 ( 2010 ) , arxiv:1005.1938 [ hep - ph ] . | the implications of a large value of the effective majorana neutrino mass for a class of two texture zero neutrino mass matrices have been studied in the flavor basis .
it is found that these textures predict near maximal atmospheric neutrino mixing angle in the limit of large effective majorana neutrino mass .
it is noted that this prediction is independent of the values of solar and reactor neutrino mixing angles .
we present the symmetry realization of these textures using the discrete cyclic group @xmath0 .
it is found that the texture zeros realised in this work remain stable under renormalization group running of the neutrino mass matrix from the seesaw scale to the electroweak scale , at one loop level . |
in deriving the partition function for a desired ensemble , the most common approach is to maximize an entropy function with constraints appropriate to the thermodynamic condition . while equivalent to the approach proposed below , such a method ( called the traditional approach hereafter ) does not make clear to students the explicit role of the assumption of equal _ a priori _ states and the corresponding role of the entropy in the thermodynamic potential for the microcanonical ensemble .
indeed , @xmath2 is often taken as a postulate@xcite and its connection to the statistical formula @xmath4 ( appearing on boltzmann s tombstone ) is not obvious .
further , in the traditional approach , the role of the entropy in understanding equilibrium in non - isolated , open ensembles can be confusing .
we note in passing that concerns over the rigor of the method of most probable distribution prompted darwin and fowler to develop a derivation of the partition function based upon complex analysis.@xcite also , infrequently stressed is the gibbs entropy , @xmath1 , where @xmath5 is the probability of finding a system in a given state , which can be invoked for any equilibrium ensemble and associated state probabilities.@xcite it is a direct consequence of the statistical entropy formula , @xmath4 , in conjunction with the gibbs construction of an ensemble that contains a large number of macroscopic subsystems , each consistent with the desired thermodynamic variables ; @xmath6 gives the number of possible realizations within the gibbs construction for the ensemble under consideration .
the gibbs entropy also permits the derivation of the connection between the characteristic thermodynamic function and the partition function for a given ensemble without further appeal to thermodynamic expressions , as is required in the traditional approach . in the present approach ,
first , the connections between the statistical formula @xmath7 , the gibbs entropy , @xmath1 , and the microcanonical entropy expression @xmath2 are clarified .
the condition for microcanonical equilibrium , and the associated role of the entropy in the thermodynamic potential then arises from the postulate of equal _ a priori _ states .
the derivation of the canonical partition function follows by invoking the gibbs construction and the first and second law of thermodynamics _ via _ the fundamental equation , @xmath8 , that incorporates the conditions of conservation of energy and composition without the needs for explicit constraints .
the role of the temperature ( coming from the constraint of total energy and an appeal to appropriate thermodynamic relationships in the traditional approach ) is immediately apparent and also introduced _ via _ the fundamental equation .
the need for explicit maximization of any function is thus also avoided .
legendre transforming a particular thermodynamic function to include desired thermodynamic control variables for an ensemble of interest and invoking equilibrium leads to the corresponding partition function . using the resulting probabilities in the gibbs entropy expression
directly connects the partition function to the thermodynamic potential .
the central role of the entropy in establishing equilibrium for a given ensemble emerges naturally from the current approach .
connections to generalized ensemble theory also arise and are presented in this context .
the present approach is novel in providing clarity as to the roles played by the different formulas and physical quantities of interest .
further , it makes explicit the assumptions inherent in deriving the partition function for an ensemble and provides its direct connection to the relevant thermodynamic potential in a systematic fashion .
this approach also makes deriving the partition function for a given ensemble a simplified , straight - forward process , even for more challenging examples such as the isothermal - isobaric ensemble .
using this approach in the classroom has led to better retention and understanding of the foundations of statistical mechanics and an ability for students to confidently apply the machinery to problems that arise in their subsequent work .
we begin by introducing the concept of an ensemble of replicas that describe the molecular states corresponding to a given macrostate ; this picture is referred to as the gibbs construction " herein , due to it s original introduction by gibbs@xcite who addressed many of the subtleties inherent@xcite in the formulation of statistical mechanics .
consider a collection of macroscopic molecular subsystems " of @xmath9 molecules within a volume @xmath10 , each of which is part of the larger gibbs construction , the totality of which is known as the system " .
no other constraints have yet been imposed , _
i.e. _ the system s macrostate is otherwise unspecified .
it is desirable to define the microscopic statistics of this system as thoroughly as possible and then apply any other constraints at the end .
let the total number of subsystems in our collection be known as @xmath11 .
then let @xmath12 , the occupation number , denote the number of subsystems from this collection that are in the same thermodynamic state .
these occupations will thus take on a large value in the thermodynamic limit and they obey a sum rule , @xmath13 .
note , technically the energy is course - grained , _
i.e. _ specified to within a small but otherwise arbitrary range ( these arguments are presented in detail elsewhere@xcite ) and the results are insensitive to this choice . first , consider the following combinatoric formula : @xmath14 @xmath15 is the number of ways in which the set of occupations @xmath16 may be arranged consistent with the given macrostate .
first , it is to be shown that when evaluated at fixed energy , this quantity @xmath17 may be identified with the thermodynamic entropy of the ensemble of systems at equilibrium , with each systems entropy given by s = @xmath18 .
note , the expression necessarily involves the logarithm of the combinatoric expression to make the entropy an extensive property ; for two independent systems the possible number of arrangements is the product of those for the individual systems , @xmath19 .
next the connection between the combinatoric formula @xmath20 and the gibbs entropy is presented ; the details of this have been given elsewhere.@xcite applying the sterling approximation@xcite to the factorial function gives the entropy : @xmath21 further substituting @xmath22 where @xmath23 is identified as the probability of a particular state and using the property of the natural log gives a system s entropy as:@xcite @xmath24 this is the gibbs entropy in an as yet unspecified ensemble with its associated probabilities ; the gibbs entropy is an entirely general definition that , for any equilibrium ensemble , specifies the relationship between the partition function and the associated characteristic thermodynamic function .
now , specializing to a set of microcanonical subsystems , and invoking the equilibrium principle of equal _ a priori _ states , _
i.e. _ @xmath25 , gives the well known result : @xmath26 it is also simple and useful to show that the gibbs entropy , and thus the thermodynamic entropy , is maximized microcanonically@xcite by the state - independent probabilities @xmath27 .
proceeding , taking the derivative of equation [ eq : ge ] and setting it to zero as @xmath28 gives @xmath29 , a constant value independent of the summation index .
thus , normalizing the probabilities as , @xmath30 immediately yields @xmath31 .
thus , for an isolated system , the assumption of equal _ a priori _ states leads to a probability @xmath5 that is independent of index , _
i.e. _ every subsystem has energy @xmath32 by construction
. further , the characteristic maximum entropy in the microcanonical equilibrium ensemble also follows . then applying the gibbs entropy expression leads to the identification of the thermodynamic entropy as the characteristic function of the microcanonical ensemble and gives its relationship to the @xmath33 partition function , @xmath34 , which can also be interpreted as the density of states@xcite at that energy
specializing the gibbs construction from the previous section to include temperature , we have a collection of subsystems all possessing the same @xmath35 .
this can be thought of by placing the subsystems in contact with a large heat bath of temperature @xmath36.@xcite we now imagine that each subsystem ( after having achieved equilibrium with the heat bath by definition ) is to be insulated and the energy of the @xmath37 subsystem is measured as @xmath38 , and for which there is also an associated macroscopic entropy @xmath39 . of great importance , we also note that the thermodynamic energy @xmath38 is exactly equal to the microscopic configurational energy of the subsystem upon insulation . furthermore , the details and/or rates involved in the insulation process are irrelevant for an equilibrium ensemble . using the earlier result , the entropy for a collection of subsystems with a specified energy @xmath38 is @xmath40 , where @xmath11 is the number of subsystems with energy @xmath38 in the ensemble .
consider the ratio of the density at energies @xmath41 : @xmath42 the fundamental equation of thermodynamics@xcite is now invoked : @xmath43 the canonical ensemble is given by a state with well defined thermodynamic variables , @xmath44 .
so the energy function , @xmath45 is legendre transformed to a new thermodynamic function , the helmholtz free energy , @xmath46 _ via _ : @xmath47 where the condition for canonical equilibrium is that @xmath48 and @xmath49 are constant , giving : @xmath50 integrating between two state points gives : @xmath51 note , the constraints of fixed temperature , particle number and volume have been explicitly enforced by using the fundamental equation , equation [ eq : fe ] and dropping differential terms that are fixed canonically . substituting equation [ eq : ds ] into equation [ eq : ratio ] gives : @xmath52 where @xmath53 and @xmath54 ( _ i.e. _ the probability of choosing the @xmath37 state from the entire ensemble at equilibrium with the heat bath ) .
the normalization factor , @xmath55 , may be readily recognized as the canonical partition function .
most importantly , we note that the insulation procedure applied to each subsystem has allowed us to identify the macroscopic energy ( and entropy ) of that microcanonical system , with the microscopic energy of the molecular configuration present at the time of insulation .
we can now proceed to use the gibbs entropy expression , equation [ eq : ge ] , substituting the canonical expression for @xmath5 to obtain : @xmath56 above , @xmath57 represents the canonical average energy that is identified with the thermodynamic energy , @xmath32.@xcite thus , the relationship @xmath58 is obtained directly from the gibbs entropy .
note , the gibbs entropy is defined for any set of probabilities and , as was shown above , is simply a consequence of the combinatoric formula @xmath59 interpreted in the context of the gibbs construction .
thus , an entropy can be associated even with nonequilibrium probabilities .
however , in that case , the entropy does not play the role of being the constrained maximized quantity that it does at equilibrium and its utility , in such circumstances , is unclear .
further note , the role of temperature is introduced _ via _ the fundamental equation without further appeal to thermodynamic relationships .
this emphasizes the role of temperature as the system is in contact with a heat bath different energy ranges are now accessible with canonical probabilities .
the ability of a diathermal system to exchange energy with its surroundings also clarifies how the concepts of work and entropy make sense for an open system and provides their relationship to the temperature .
gibbs construction for the canonical ensemble .
the subsystems of identical @xmath60 are in thermal equilibrium with a large bath at temperature @xmath36.,width=3 ]
using the gibbs construction from the previously derived canonical ensemble , the constraint that all subsystems possess identical @xmath9 can now be relaxed .
we now consider the ratio of subsystem , having chosen a particular system from the energy level @xmath38 with @xmath61 molecules : @xmath62 as before , the fundamental equation of thermodynamics can be relied upon to relate the change in entropy to the other variables of our ensemble . in this case , the probabilities that will generate the macrostate corresponding to constant @xmath63 are desired .
after doing so , the entropy _ via the boltzmann law _ is used to determine the microscopic states .
legendre transforming the canonical thermodynamic equation to substitute @xmath64 for @xmath9 : @xmath65 where at equilibrium @xmath66 and with constant @xmath63 : @xmath67 upon integrating , the difference equation for the entropy is found : @xmath68\end{aligned}\ ] ] the ratio of microstates then becomes : @xmath69 and since @xmath70 @xmath71 where the normalization factor @xmath72 is the _ grand canonical partition function _ :
@xmath73 it may be noted that in this case the constraint of constant @xmath9 was merely relaxed , legendre transformed to the corresponding macrostate , and the partition function then followed quite naturally and simply .
the resulting probabilities can now be substituted into the gibbs entropy and the relationship between the thermodynamic potential and partition function is thus directly established .
gibbs construction for the grand canonical ensemble .
@xmath64 has been legendre transformed to replace @xmath9 as the macroscopic constant , and so the set of subsystems includes those of differing @xmath9 values.,width=3 ]
having successfully derived the grand canonical partition function by relaxing the constant @xmath9 constraint , it can now be shown that the isothermal - isobaric ensemble is generated by starting with constant @xmath74 and relaxing the condition of constant @xmath10 .
consider the number of subsystems with both volume @xmath75 and energy @xmath76 : @xmath77 legendre transforming our desired variables to a new characteristic function @xmath78 , and then applying the condition of equilibrium @xmath79 and our constant differential terms : @xmath80 and so the ratio of observable subsystems becomes : @xmath81 } \nonumber \\
= \frac{e^{\beta e_i}e^{\beta pv_j}}{e^{\beta e_{i+1}}e^{\beta pv_{j+1}}}\end{aligned}\ ] ] and since @xmath70 @xmath82 where the normalization factor , @xmath83 is the _ isothermal - isobaric partition function_. again , the associated probabilities can now be substituted into the gibbs entropy and the relationship between the thermodynamic potential and partition function is thus directly established .
gibbs construction for the isothermal - isobaric ensemble .
@xmath84 has been legendre transformed to replace @xmath10 as the macroscopic constant , and so the set of subsystems includes those of differing @xmath10 values.,width=3 ]
based on the previous sections , it may be noted that if a partition function for _ any _ macrostate in thermal equilibrium is to be derived , one shall _ always _ arrive at an expression that involves an exponential function ; this follows as a consequence of the boltzmann law . furthermore , the partition function being the normalization factor to express this as a probability means that one will always have a discrete sum .
therefore , our partition functions will always be some variation on a theme amounting to a
sum of exponentials . " in the continuum limit , it can be shown that the canonical partition function @xmath85 ( written as a sum over energy levels ) transforms as : @xmath86 or , in other words , the canonical partition function is the laplace transform of the microcanonical partition function @xmath11 .
how does it work the other way ?
let s apply the inverse laplace transform to @xmath85 : @xmath87 where the phase space differential form @xmath88 .
now , @xmath89 and because no singularity is present in the right - half of the complex plane , the contour may be taken vertically through @xmath90 . since @xmath91 along the integration , the substitution @xmath92
can be made : @xmath93 where indeed equation [ eq : delta_function ] can be identified as the microcanonical partition function . thus , any constant energy shell ensemble may be laplace transformed to an ensemble of a new intensive variable . as the partition functions are related to one another through the laplace transform ,
this is isomorphic to the thermodynamic potentials ( to each of which may be associated a particular partition function ) being related through the legendre transform.@xcite the quantum harmonic oscillator is an illustrative example of how the canonical partition function may be transformed to the microcanonical case : @xmath94 \label{eq : nve_qho}\end{aligned}\ ] ] the reason that the aforementioned recipe " for generating the partition function ( as outlined in sections [ sec : canonical],[sec : grand_canonical ] and [ sec : isobaric ] ) in an arbitrary ensemble works is due to the thermodynamic relations and the boltzmann law . the underlying mathematical structure that allows this has also been previously formulated as generalized ensemble theory.@xcite
an approach is presented for deriving partition functions that is an alternative to more common methods .
it emphasizes the central role that ( maximizing ) the boltzmann entropy plays in connecting the molecular states of the system to the observable thermodynamics . using this technique in a classroom setting for a beginning graduate class in statistical mechanics
has led to systemization and demystification of the derivation for useful ensembles . also , the role of legendre transforms to introduce thermodynamic control variables appears naturally and is tied directly to both the derivation of the ensemble and corresponding partition function . within this formalism , students are clear on how the thermodynamic potential relates to a given ensemble and the role of equal _ a priori _ states .
further , relating the partition function to the thermodynamic potential using the gibbs entropy is straightforward and no further appeal to thermodynamic expressions is required as the relevant thermodynamic connection was included from the start of the derivation .
finally , the similarities between the derivation method demonstrated and the relations known from generalized ensemble theory have been noted . it is our hope that the formulaic approach presented here will be of utility in both research and pedagogy .
the authors acknowledge funding from the u.s .
department of energy , basic energy sciences ( grant no .
de0gg02 - 07er46470 ) .
lawrence livermore national laboratory is operated by lawrence livermore national security , llc , for the u.s .
department of energy , national nuclear security administration under contract de - ac52 - 07na27344 . | a pedagogical approach for deriving the statistical mechanical partition function , in a manner that emphasizes the key role of entropy in connecting the microscopic states to thermodynamics , is introduced .
the connections between the combinatoric formula @xmath0 applied to the gibbs construction , the gibbs entropy , @xmath1 , and the microcanonical entropy expression @xmath2 are clarified .
the condition for microcanonical equilibrium , and the associated role of the entropy in the thermodynamic potential is shown to arise naturally from the postulate of equal _ a priori _ states .
the derivation of the canonical partition function follows simply by invoking the gibbs ensemble construction at constant temperature and using the first and second law of thermodynamics ( _ via _ the fundamental equation @xmath3 ) that incorporate the conditions of conservation of energy and composition without the needs for explicit constraints ; other ensemble follow easily .
the central role of the entropy in establishing equilibrium for a given ensemble emerges naturally from the current approach .
connections to generalized ensemble theory also arise and are presented in this context . |
-input multiple - output ( mimo ) techniques offer transmit diversity and high data rates through the use of multiple antennas at both transmitter and receiver sides @xcite-@xcite .
a key component of a mimo system is the mimo detector at the receiver , which , in practice , is often the bottleneck for the overall performance and complexity .
mimo detectors including sphere decoder and several of its variants @xcite-@xcite achieve near maximum likelihood ( ml ) performance at the cost of high complexity .
other well known detectors including zf ( zero forcing ) , mmse ( minimum mean square error ) , and zf - sic ( zf with successive interference cancellation ) detectors @xcite,@xcite are attractive from a complexity view point , but achieve relatively poor performance .
for example , the zf - sic detector ( i.e. , the well known v - blast detector with ordering @xcite,@xcite ) does not achieve the full diversity in the system . the mmse - sic detector has been shown to achieve optimal performance @xcite . however , these detectors are prohibitively complex for large number of antennas of the order of tens to hundreds . with small number of antennas , the high capacity potential of mimo
is not fully exploited . a key issue with using large number of antennas , however , is the high detection complexities involved . in this paper
, we focus on large mimo systems , where by ` _ large _ '
we mean number of transmit and receive antennas of the order of tens to hundreds .
such large mimo systems will be of immense interest because of the very high spectral efficiencies possible in such systems .
for example , in a v - blast system , increased number of transmit antennas means increased data rate without bandwidth increase . however , major bottlenecks in realizing such large mimo systems include @xmath4 physical placement of large number of antennas in communication terminals , @xmath5 lack of practical low - complexity detectors for such large systems , and @xmath6 channel estimation issues . in this paper
, we address the second problem in the above ( i.e. , low - complexity large mimo detection ) . specifically , we present a low - complexity detector for large mimo systems , including v - blast as well as high - rate non - orthogonal space - time block codes ( stbc ) @xcite .
the proposed low - complexity detector has its roots in past work on hopfield neural network ( hnn ) based algorithms for image restoration @xcite,@xcite , which are meant to handle large digital images .
hnn based image restoration algorithms in @xcite are applied to multiuser detection ( mud ) in cdma systems on awgn channels in @xcite .
this detector , referred to as the likelihood ascent search ( las ) detector , essentially searches out a sequence of bit vectors with monotonic likelihood ascent and converges to a fixed point in finite number of steps @xcite .
the power of the las detector for cdma lies in @xmath4 its linear average per - bit complexity in number of users , and @xmath5 its ability to perform very close to ml detector for large number of users . taking the cue from las detector s complexity and performance superiority in large systems , we , in this paper ,
successfully adopt the las detector for large mimo systems and report interesting results .
we refer to the proposed detector as mf / zf / mmse - las detector depending on the initial vector used in the algorithm ; mf - las detector uses the matched filter output as the initial vector , and zf - las and mmse - las detectors employ zf and mmse outputs , respectively , as the initial vector .
our major findings in this paper are summarized as follows : * in an uncoded v - blast system with bpsk , the proposed detector achieves _ near - exponential diversity _ for hundreds of antennas ( i.e. , achieves near siso awgn performance ) .
for example , the proposed detector nearly renders a @xmath7 mimo fading channel into 200 parallel , non - interfering siso awgn channels .
the detector achieves this excellent performance with an average per - bit complexity of just @xmath0 , where @xmath1 and @xmath2 denote the number of transmit and receive antennas , respectively . * with an outer turbo code , the proposed detector achieves good coded bit error performance as well .
for example , in a 600 transmit and 600 receive antennas v - blast system with a high spectral efficiency of 200 bps / hz ( using bpsk and rate-1/3 turbo code ) , our simulation results show that the proposed detector performs close to within about 4.6 db from the theoretical capacity .
we note that performance with such closeness to capacity has not been reported in the literature so far for such large number of antennas using a practical complexity detector .
* we have adopted the proposed detector for the low - complexity decoding of large full - rate non - orthogonal stbcs from division algebras ( da ) in @xcite .
we decode the @xmath3 full - rate non - orthogonal stbc from da ( which has 256 data symbols in one stbc matrix ) using the proposed detector and show that it performs close to within about 5.5 db from capacity using 4-qam and rate-3/4 turbo code at a spectral efficiency of 24 bps / hz . *
we point out that because of the high complexities involved in the decoding of large non - orthogonal stbcs using other known detectors , the ber performance of such high - rate large non - orthogonal stbcs have not been reported in the literature so far .
the very fact that we could show the simulated ber performance plots ( both uncoded as well as turbo coded ) for a @xmath3 full - rate non - orthogonal stbc with 256 complex symbols in one stbc matrix in itself is a clear indication of the superior low - complexity attribute of the proposed detector . to our knowledge , this is the first time that simulated ber plots and nearness to capacity results for a full - rate @xmath3 stbc from da are reported in the literature ; this became feasible due to the low - complexity attribute of the proposed detector .
* we also illustrate the applicability of the proposed detector in the low - complexity detection of large multicarrier cdma ( mc - cdma ) systems . in large mc - cdma systems with hundreds of users ,
the proposed detector is shown to achieve near single - user performance , at an average per - bit complexity linear in number of users , which is quite appealing for its use in practical cdma systems .
the rest of the paper is organized as follows . in section [ sec2 ] ,
we present the proposed las detector for v - blast systems and its complexity .
the simulated uncoded and coded ber performance of the proposed detector for v - blast is presented in section [ sec3 ] .
decoding of non - orthogonal stbcs and ber performance results are presented in section [ sec4 ] . the las detector for mc - cdma and
the corresponding ber performance results are presented in section [ sec5 ] .
conclusions are presented in section [ sec6 ] .
in this section , we present the proposed las detector for v - blast and its complexity . consider a v - blast system with @xmath1 transmit antennas and @xmath2 receive antennas , @xmath8 , where @xmath1 symbols are transmitted from @xmath1 transmit antennas simultaneously .
let @xmath9 be the symbol - qam/@xmath10-pam as well .
] transmitted by the @xmath11th transmit antenna .
each transmitted symbol goes through the wireless channel to arrive at each of @xmath2 receive antennas .
denote the path gain from transmit antenna @xmath11 to receive antenna @xmath12 by @xmath13 . considering a flat - fading mimo channel with rich scattering , the signal received at antenna @xmath12 , denoted by @xmath14 , is given by @xmath15 the @xmath16 , @xmath17 , @xmath18 ,
are assumed to be i.i.d .
complex gaussian r.vs ( i.e. , fade amplitudes are rayleigh distributed ) with zero mean and @xmath19 = e\big[\big(h_{kj}^{q}\big)^2\big]=0.5 $ ] , where @xmath20 and @xmath21 are the real and imaginary parts of @xmath13 .
the noise sample at the @xmath12th receive antenna , @xmath22 , is assumed to be complex gaussian with zero mean , and @xmath23 , @xmath24 , are assumed to be independent with @xmath25=n_0=\frac{n_te_s}{\gamma}$ ] , where @xmath26 is the average energy of the transmitted symbols , and @xmath27 is the average received snr per receive antenna @xcite . collecting
the received signals from all receive antennas , we write^t$ ] , @xmath28^*$ ] , and @xmath28^h$ ] denote transpose , conjugate , and conjugate transpose operations , respectively . @xmath29 and @xmath30 denote the real and imaginary parts of the complex argument . ]
@xmath31 where @xmath32^t$ ] is the @xmath2-length received signal vector , @xmath33^t$ ] is the @xmath1-length transmitted bit vector , * h * denotes the @xmath34 channel matrix with channel coefficients @xmath16 , and @xmath35^t$ ] is the @xmath2-length noise vector . *
h * is assumed to be known perfectly at the receiver but not at the transmitter .
the proposed las algorithm essentially searches out a sequence of bit vectors until a fixed point is reached ; this sequence is decided based on an update rule . in the v - blast system
considered , for ml detection @xcite , the most likely @xmath36 is taken as that @xmath37 which maximizes @xmath38 the likelihood function in ( [ lamda1w ] ) can be written as @xmath39 where @xmath40 _ update criterion in the search procedure : _ let @xmath41 denote the bit vector tested by the las algorithm in the search step @xmath42 .
the starting vector @xmath43 can be the output vector from any known detector . when the output vector of the mf detector is taken as the @xmath43 , we call the resulting las detector as the mf - las detector
. we define zf - las and mmse - las detectors likewise .
given @xmath41 , the algorithm obtains @xmath44 through an update rule until a fixed point is reached .
the update is made in such a way that the change in likelihood from step @xmath42 to @xmath45 , denoted by @xmath46 , is positive , i.e. , @xmath47 an expression for the above change in likelihood can be obtained in terms of the gradient of the likelihood function as follows .
let @xmath48 denote the gradient of the likelihood function evaluated at @xmath41 , i.e. , @xmath49 where @xmath50 using ( [ 3eq5w ] ) in ( [ deltaw ] ) , we can write @xmath51 now , defining @xmath52 and @xmath4 observing that @xmath53 , @xmath5 adding & subtracting the term @xmath54 to the rhs of ( [ dxw ] ) , and @xmath6 further observing that @xmath55 , we can simplify ( [ dxw ] ) as @xmath56 where @xmath57 now , given @xmath58 , @xmath59 , and @xmath41 , the objective is to obtain @xmath44 from @xmath41 such that @xmath60 in ( [ 3eq8w ] ) is positive . potentially any one or several bits in @xmath41 can be flipped ( i.e. , changed from + 1 to -1 or vice versa ) to get @xmath44 .
we refer to the set of bits to be checked for possible flip in a step as a _ check candidate set_. let @xmath61 denote the check candidate set at step @xmath42 . with the above definitions
, it can be seen that the likelihood change at step @xmath42 , given by ( [ 3eq8w ] ) , can be written as @xmath62 , \label{3eq12w}\end{aligned}\ ] ] where @xmath63 , @xmath64 , and @xmath65 are the @xmath11th elements of the vectors @xmath41 , @xmath48 , and @xmath66 , respectively .
as shown in @xcite for synchronous cdma on awgn , the following update rule can be easily shown to achieve monotonic likelihood ascent ( i.e. , @xmath67 if there is at least one bit flip ) in the v - blast system as well .
_ las update algorithm : _ given @xmath68 and an initial bit vector @xmath69 , bits in @xmath41 are updated as per the following update rule : @xmath70 where @xmath71 is a threshold for the @xmath11th bit in the @xmath42th step is taken to be @xmath72 where @xmath73 is the element in the @xmath11th row and @xmath74th column of the matrix @xmath75 .
it is noted that different choices can be made to specify the sequence of @xmath76 .
one of the simplest sequences correspond to checking one bit in each step for a possible flip , which is termed as a sequential las ( slas ) algorithm with constant threshold , @xmath77 the sequence of @xmath78 in slas can be such that the indices of bits checked in successive steps are chosen circularly or randomly .
checking of multiple bits for possible flip is also possible .
let @xmath79 denote the set of indices of the bits flipped according to the update rule in ( [ 3eq18w ] ) at step @xmath42 .
then the updated bit vector @xmath44 can be written as @xmath80 where @xmath81 is the @xmath74th coordinate vector . using ( [ 3eq20w ] ) in ( [ 3eq9w ] )
, the gradient vector for the next step can be obtained as @xmath82 where @xmath83 denotes the @xmath74th column of the matrix @xmath75 .
the las algorithm keeps updating the bits in each step based on the update rule given in ( [ 3eq18w ] ) until @xmath84 for some @xmath85 , in which case @xmath86 is a fixed point , and it is taken as the detected bit vector and the algorithm terminates . in terms of complexity , given an initial vector , the the las operation part alone has an average per - bit complexity of @xmath0 .
this can be explained as follows .
the complexity involved in the las operation is due to three components : @xmath4 initial computation of @xmath87 in ( [ 3eq9w ] ) , @xmath5 update of @xmath48 in each step as per ( [ 3eq21w ] ) , and @xmath6 the average number of steps required to reach a fixed point .
computation of @xmath87 requires the computation of @xmath88 for each mimo fading channel realization @xmath89see eqns .
( [ 3eq9w ] ) , ( [ 3eq10w ] ) , and ( [ 3eq5bw])@xmath90 , which requires a per - bit complexity of order @xmath0
. update of @xmath48 in the @xmath42th step as per ( [ 3eq21w ] ) using sequential las requires a complexity of @xmath91 , and hence a constant per - bit complexity .
regarding the complexity component @xmath6 , we obtained the average number of steps required to reach a fixed point for sequential las through simulations .
we observed that the average number of steps required is linear in @xmath1 , i.e. , constant per - bit complexity where the constant @xmath92 depends on snr , @xmath1 , @xmath2 , and the initial vector ( see fig .
[ fig_bfr ] ) . putting the complexities of @xmath4 , @xmath5 , and @xmath6 in the above together
, we see that the average per - bit complexity of las operation alone is @xmath0 .
in addition to the above , the initial vector generation also contributes to the overall complexity .
the average per - bit complexity of generating initial vectors using mf , zf , and mmse are @xmath93 , @xmath0 , and @xmath0 , respectively .
the higher complexity of zf and mmse compared to mf is because of the need to perform matrix inversion operation in zf / mmse .
again , putting the complexities of the las part and the initial vector generation part together , we see that the overall average per - bit complexity of the proposed mf / zf / mmse - las detector is @xmath0 .
this complexity is an order superior compared to the well known zf - sic detector with ordering , whose per - bit complexity is @xmath94 .
in this section , we present the uncoded / coded ber performance of the proposed las detector in v - blast obtained through simulations , and compare with those of other detectors .
the las algorithm used is the sequential las with circular checking of bits starting from the first antenna bit .
we also quantify how far is the proposed detector s turbo coded ber performance away from the theoretical capacity .
the snrs in all the ber performance figures are the average received snr per received antenna , @xmath27 , defined in sec .
[ sec2 ] @xcite .
_ mf / zf - las performs increasingly better than zf - sic for increasing @xmath95 _ : in fig .
[ fig_vb2 ] , we plot the uncoded ber performance of the mf - las , zf - las and zf - sic detectors for v - blast as a function of @xmath95 at an average received snr of 20 db with bpsk .
the performance of the mf and zf detectors are also plotted for comparison . from fig .
[ fig_vb2 ] , we observe the following : * the ber at @xmath96 is nothing but the siso flat rayleigh fading ber for bpsk , given by @xmath97 $ ] which is equal to @xmath98 for @xmath99 db @xcite . while the performance of mf and zf degrade as @xmath95 is increased , the performance of zf - sic improves for antennas up to @xmath100 , beyond which a flooring effect occurs .
this improvement is likely due to the potential diversity in the ordering ( selection ) in zf - sic , whereas the flooring for @xmath101 is likely due to interference being large beyond the cancellation ability of the zf - sic . *
the behavior of mf - las and zf - las for increasing @xmath95 are interesting . starting with the mf output as the initial vector , the
mf - las always achieves better performance than mf .
more interestingly , this improved performance of mf - las compared to that of mf increases remarkably as @xmath95 increases .
for example , for @xmath100 , the performance improves by an order in ber ( i.e. , @xmath102 ber for mf versus @xmath103 ber for mf - las ) , whereas for @xmath104 the performance improves by four - orders in ber ( i.e. , @xmath105 ber for mf versus @xmath106 ber for mf - las ) .
this is due to the large system effect in the las algorithm which is able to successfully pick up much of the diversity possible in the system .
this large system performance superiority of the las is in line with the observations / results reported in @xcite for a large cdma system @xmath89large number of antennas in our case , whereas it was large number of users in @xcite@xmath107 .
* while the zf - las performs slightly better than zf - sic for antennas less than 4 , zf - sic performs better than zf - las for antennas in the range 4 to 24 .
this is likely because , for antennas less than 4 , the ber of zf is small enough for the las to clean up the zf initial vector into an output vector better than the zf - sic output vector .
however , for antennas in the range of 4 to 24 , the ber of zf gets high to an extent that the zf - las is less effective in cleaning the initial vector beyond the diversity performance achieved by the zf - sic . a more interesting observation
, however , is that for antennas greater than 25 , the large system effect of zf - las starts showing up .
so , in the large system setting ( e.g. , antennas more than 25 in fig .
[ fig_vb2 ] ) , the zf - las performs increasingly better than zf - sic for increasing @xmath95 .
we found the number of antennas at which the cross - over between zf - sic and zf - las occurs to be different for different snrs . * another observation in fig .
[ fig_vb2 ] is that for antennas greater than 50 , mf - las performs better than zf - las .
this behavior can be explained by observing the performance comparison between mf and zf detectors given in the same figure . for more than 50 antennas ,
mf performs slightly better than zf .
it is known that zf detector can perform worse than mf detector under high noise / interference conditions @xcite ( here high interference due to large @xmath1 ) .
hence , starting with a better initial vector , mf - las performs better than zf - las . ) for v - blast at an average received snr = 20 db .
bpsk , @xmath1 bps / hz spectral efficiency .
, width=338 ] [ fig_vb2 ] v - blast system .
bpsk , 200 bps / hz spectral efficiency .
zf - las achieves higher order diversity ( near - exponential diversity ) than zf - sic at a much lesser complexity.,width=319 ] [ fig_vb3 ] _ zf - las outperforms zf - sic in large v - blast systems both in complexity & diversity : _ in fig .
[ fig_vb3 ] , we present an interesting comparison of the uncoded ber performance between zf , zf - las and zf - sic , as a function of average snr for a @xmath108 v - blast system . this system being a large system , the zf - las has a huge complexity advantage over zf - sic as pointed out before in sec .
[ vb_comp ] .
in fact , although we have taken the effort to show the performance of zf - sic at such a large number of antennas like 200 , we had to obtain these simulation points for zf - sic over days of simulation time , whereas the same simulation points for zf - las were obtained in just few hours .
this is due to the @xmath94 complexity of zf - sic versus @xmath0 complexity of zf - las , as pointed out in sec .
[ vb_comp ] .
more interestingly , in addition to this significant complexity advantage , zf - las is able to achieve a much higher order of diversity ( in fact , near - exponential diversity ) in ber performance compared to zf - sic ( which achieves only a little better than first order diversity ) .
this is clearly evident from the slopes of the ber curves of zf - las and zf - sic .
_ note that the ber curve for zf - las is almost the same as the uncoded ber curve for bpsk on a siso awgn channel , given by @xmath109 @xcite .
this means that the proposed detector nearly renders a @xmath108 mimo fading channel into 200 parallel , non - interfering siso awgn channels .
_ _ las detector s performance with hundreds of antennas : _ as pointed earlier , obtaining zf - sic results for more than even 50 antennas requires very long simulation run times , which is not the case with zf - las .
in fact , we could easily generate ber results for up to 400 antennas for zf - las , which are plotted in fig .
[ fig_vb4 ] . the key observations in fig .
[ fig_vb4 ] are that @xmath4 the average snr required to achieve a certain ber performance keeps reducing for increasing number of antennas for zf - las , and @xmath5 increasing the number of antennas results in increased orders of diversity achieved ( close to siso awgn performance for 200 and 400 antennas ) .
we have also observed from our simulations that for large number of antennas , the las algorithm converges to almost the same near - ml performance regardless of the initial vector chosen .
for example , for the case of 200 and 400 antennas in fig .
[ fig_vb4 ] , the ber performance achieved by zf - las , mf - las , and mmse - las are almost the same ( although we have not explicitly plotted the ber curves for mf - las and mmse - las in fig . [ fig_vb4 ] ) .
so , in such large mimo systems setting , mf - las may be preferred over zf / mmse - las since zf / mmse - las require matrix inverse operation whereas mf - las does not .
observation @xmath4 in the above paragraph is explicitly brought out in fig .
[ fig_vb5 ] , where we have plotted the average received snr required to achieve a target uncoded ber of @xmath110 as a function of @xmath95 for zf - las and zf - sic .
it can be seen that the snr required to achieve @xmath110 with zf - las significantly reduces for increasingly large @xmath95 .
for example , the required snr reduces from about 25 db for a siso rayleigh fading channel to about 7 db for a @xmath111 v - blast system using zf - las . as we pointed out in fig .
[ fig_vb4 ] , this @xmath111 system performance is almost the same as that of a siso awgn channel where the snr required to achieve @xmath110 ber is also close to 7 db @xcite , i.e. , @xmath112 db . .
bpsk , @xmath1 bps / hz spectral efficiency . for large number of antennas ( e.g. , @xmath113 ) ,
the performance of zf - las , mf - las , and mmse - las are almost the same.,width=326 ] in v - blast for increasing values of @xmath95 .
zf - las versus zf - sic .
zf - las achieves near siso awgn performance.,width=326 ] for different snrs and initial vectors ( mf , zf , mmse ) .
results obtained from simulations.,width=321 ] [ fig_bfr ] in this subsection , we present the turbo coded ber performance of the proposed las detector .
we also quantify how far is the proposed detector s performance away from the theoretical capacity . for a @xmath114 mimo system model in sec .
[ sec2 ] with perfect channel state information ( csi ) at the receiver , the ergodic capacity is given by @xcite @xmath115 , \label{cap}\end{aligned}\ ] ] where @xmath116 is the @xmath117 identity matrix and @xmath27 is the average snr per receive antenna . we have evaluated the capacity in ( [ cap ] ) for a @xmath118 mimo system through monte - carlo simulations and plotted it as a function of average snr in fig .
figure [ fig2 ] shows the simulated ber performance of the proposed las detector for a @xmath118 mimo system with bpsk and rate-1/3 turbo code ( i.e. , spectral efficiency = 200 bps / hz ) .
figure [ fig4 ] shows similar performance plots for rate-3/4 turbo code at a spectral efficiency of 450 bps / hz . from the capacity curve in fig .
[ fig1 ] , the minimum snrs required at 200 bps / hz and 450 bps / hz spectral efficiencies are -5.4 db and -0.8 db , respectively .
the following interesting observations can be made from figs .
[ fig2 ] and [ fig4 ] : mimo system with receive csi.,width=312 ] v - blast system .
200 bps / hz spectral efficiency .
proposed mf / zf / mmse - las detectors performance is away from capacity by 4.6 db.,width=321 ] v - blast system .
450 bps / hz spectral efficiency .
proposed mf / zf / mmse - las detectors performance is away from capacity by 5.6 db.,width=321 ] * in terms of uncoded ber , the performance of mf , zf , and mmse are different , with zf and mmse performing the worst and best , respectively . but the performance of mf - las , zf - las , and mmse - las are almost the same ( near - exponential diversity performance ) with the number of antennas being large ( @xmath119 ) . * with a rate-1/3 turbo code ( fig .
[ fig2 ] ) , all the las detectors considered ( i.e. , mf - las , zf - las , mmse - las ) achieve almost the same performance , which is about 4.6 db away from capacity db , where @xmath120 is the turbo code rate , which amounted to a pessimistic prediction of nearness to capacity . here , we have corrected those plotting errors .
figures [ fig2 ] , [ fig4 ] , [ fig_qam ] and the nearness to capacity results given in table - i in this paper are the corrected ones . ]
( i.e. , near - vertical fall of coded ber occurs at about -0.8 db ) .
turbo coded mf / mmse without las also achieve good performance in this case ( i.e. , less than only 2 db away from turbo coded mf / zf / mmse - las performance ) .
this is because the uncoded ber of mf and mmse at around 0 to 2 db snr are small enough for the turbo code to be effective .
however , this is not the case with turbo coded zf without las .
as can be seen , in the range of snrs shown , the uncoded ber of zf without las is so high ( close to 0.5 ) that the vertical fall of coded ber can happen only at very high snrs , because of which we have not shown the performance of turbo coded zf without las .
& at capacity & proposed las & zf & mf & mmse + rate-1/3 , & -5.4 db & -0.8 db & high & 1.2 db & -0.3 db + 200 bps / hz & & & & & + rate-1/2 & -3.2 db & 1.5 db & high & high & 3 db + 300 bps / hz & & & & & + rate-3/4 & -0.8 db & 4.8 db & high & high & high + 450 bps / hz & & & & & + in table i , we summarize the performance of various detectors in terms of their nearness to capacity in a @xmath121 v - blast system using bpsk , and rate-1/3 , rate-1/2 and rate-3/4 turbo codes . from table - i
, it can be seen that there is a clear superiority of the proposed mf / zf / mmse - las over mf / mmse without las in terms of coded ber ( nearness to capacity ) when high - rate turbo codes are used .
for example , when a rate-3/4 turbo code is used the mf / zf / mmse - las performs to within about 5.6 db from capacity , whereas the performance of rate-3/4 turbo coded mf / mmse without las are much farther away from capacity . _ performance of @xmath10-pam/@xmath10-qam : _ although the las algorithm in sec .
[ sec2 ] is presented assuming bpsk , it can be adopted for @xmath10-ary modulation including @xmath10-pam and @xmath10-qam . in the case of bpsk ,
the elements of the data vector take values from @xmath122 .
@xmath10-pam symbols take discrete values from @xmath123 where @xmath124 , @xmath125 , and @xmath10-qam is nothing but quadrature pam .
we have adopted the las algorithm for @xmath10-pam/@xmath10-qam and evaluated the performance of the las detector for 4-pam/4-qam and 16-pam/16-qam without and with coding . in @xmath10-pam/@xmath10-qam
also , we have observed large system behavior of the proposed detector similar to those presented for bpsk . as an example , in fig .
[ fig_qam ] , we present the uncoded and coded performance of the mmse - las detector in a @xmath121 v - blast system for 16-pam/16-qam with rate-1/2 and rate-1/3 turbo codes at spectral efficiencies of 1200 bps / hz and 800 bps / hz , respectively . it can be observed that the las detector achieves performance close to within about 13 db from the theoretical capacity . _
effect of channel estimation errors : _ as we pointed out earlier , another key issue in large mimo systems is channel estimation @xcite,@xcite .
we have evaluated the effect of channel estimation errors on the performance of the proposed detector in v - blast by considering an estimation error model , where the estimated channel matrix , @xmath126 , is taken to be @xmath127 where @xmath128 is the estimation error matrix , the entries of which are assumed to be i.i.d .
complex gaussian with zero mean and variance @xmath129 .
our simulation results showed that in a @xmath108 v - blast system with bpsk , rate-1/2 turbo code and las detection , the coded ber degradation compared to perfect channel estimation is only 0.2 db and 0.6 db for channel estimation error variances of @xmath130 and @xmath131 , respectively .
the investigation of estimation algorithms and efficient pilot schemes for accurate channel estimation in large mimo systems as such are important topics for further research . v - blast system for 16-pam and 16-qam with rate-1/2 and rate-1/3 turbo codes.,width=321 ]
v - blast with large number of antennas can offer high spectral efficiencies , but it does not provide transmit diversity . on the other hand , well known orthogonal stbcs
have the advantages of full transmit diversity and low decoding complexity , but suffer from rate loss for increased number of transmit antennas @xcite,@xcite-@xcite . _
full - rate non - orthogonal stbcs from division algebras ( da ) _
@xcite , on the other hand , are attractive for achieving high spectral efficiencies in addition to achieving full transmit diversity , using large number of transmit antennas .
construction of full - rate non - orthogonal stbcs from da for arbitrary number of transmit antennas @xmath42 is given by the matrix in ( 20.a ) at the bottom of this page @xcite . in ( 20.a ) , @xmath132 , @xmath133 , and @xmath134 , @xmath135 are the data symbols from a qam alphabet .
note that there are @xmath136 data symbols in one stbc matrix .
when @xmath137 and @xmath138 , the stbc in ( 20.a ) achieves full transmit diversity ( under ml decoding ) as well as information - losslessness @xcite . when @xmath139 , the code ceases to be of full - diversity ( fd ) , but continues to be information - lossless ( ill ) @xcite .. \hspace{10 mm } ( \mbox{20.a})\ ] ] ] high spectral efficiencies with large @xmath42 can be achieved using this code construction .
for example , with @xmath140 transmit antennas , the @xmath3 stbc from ( 20.a ) with 16-qam and rate-3/4 turbo code achieves a spectral efficiency of 48 bps / hz . this high spectral efficiency is achieved along with the full - diversity of order @xmath141 .
however , since these stbcs are non - orthogonal , ml detection gets increasingly impractical for large number of transmit antennas , @xmath42 .
consequently , a key challenge in realizing the benefits of these full - rate non - orthogonal stbcs in practice is that of achieving near - ml performance for large number of transmit antennas at low detection complexities . here , we show that near - ml detection of large mimo signals originating from several tens of antennas using full - rate non - orthogonal stbcs is possible at practically affordable low complexities ( using the proposed las detector ) , which is a significant new advancement that has not been reported in the mimo detection literature so far .
we have adopted the proposed las detector for the decoding of full - rate non - orthogonal stbcs . in fig .
[ fig_n1 ] , we present the uncoded ber of the las detector in decoding @xmath142 full - rate non - orthogonal stbcs from da in ( 20.a ) for @xmath143 , @xmath139 , and 4-qam . it can be observed that as the stbc code size @xmath42 increases , the las performs increasingly better such that it achieves close to siso awgn performance ( within 0.5 db at @xmath110 ber and less ) with the @xmath3 stbc .
we point out that due to the high complexities involved in decoding large size stbcs using other known detectors , the ber performance of stbcs with large @xmath42 has not been reported in the literature so far .
the very fact that we could show the simulated ber plots ( both uncoded as well as turbo coded ) for a @xmath3 stbc with 256 complex symbols in one stbc matrix in itself is a clear indication of the superior low - complexity attribute of the proposed las detector . to our knowledge , we are the first to report the simulated ber performance of a @xmath3 stbc from da ; this became feasible because of the low - complexity feature of the proposed detector .
in addition , the achievement of near siso awgn performance with @xmath3 stbc is a significant result from an implementation view point as well , since 16 antennas can be easily placed in communication terminals of moderate size , which can make large mimo systems practical . in fig .
[ fig_n2 ] , we show the coded ber performance of the @xmath3 stbc using different turbo code rates of 1/3 , 1/2 , and 3/4 . with 4-qam , these turbo code rates along with the @xmath3 stbc from da correspond to spectral efficiencies of 10.6 bps / hz , 16 bps / hz and 24 bps / hz , respectively .
the minimum snrs required to achieve these capacities are also shown in fig .
[ fig_n2 ] .
it can be observed that the proposed detector performs to within about 5.5 db of the capacity , which is an impressive result . in all the turbo coded ber plots in this paper ,
we have used hard decision outputs from the las algorithm . in @xcite , we have proposed a method to generate soft decision outputs from the las algorithm for the individual bits that form the qam / pam symbols . with the proposed soft decision las outputs in @xcite ,
the coded performance is found to move closer to capacity by an additional 1 to 1.5 db compared to that achieved using hard decision las outputs reported in this paper .
full - rate non - orthogonal stbcs from da for @xmath143 .
mmse initial vector , 4-qam , @xmath144 .
@xmath3 stbc with 256 complex symbols in each stbc matrix achieves close to siso awgn performance .
, width=321 ] full - rate non - orthogonal stbc from da .
mmse initial vector , 4-qam .
rates of turbo codes : 1/3 , 1/2 , 3/4 .
proposed las detector performs close to within about 5.5 db from the theoretical capacity . , width=321 ]
in this section , we present the proposed las detector for multicarrier cdma , its performance and complexity . consider a @xmath146-user synchronous multicarrier ds - cdma system with @xmath10 subcarriers .
let @xmath147 denote the binary data symbol of the @xmath12th user , which is sent in parallel on @xmath10 subcarriers @xcite,@xcite .
let @xmath148 denote the number of chips - per - bit in the signature waveforms .
it is assumed that the channel is frequency non - selective on each subcarrier and the fading is slow ( assumed constant over one bit interval ) and independent from one subcarrier to the other .
let @xmath149^t$]denote the @xmath146-length received signal vector on the @xmath74th subcarrier ; i.e. , @xmath150 is the output of the @xmath12th user s matched filter on the @xmath74th subcarrier .
assuming that the inter - carrier interference is negligible , the @xmath146-length received signal vector on the @xmath74th subcarrier @xmath151 can be written in the form @xmath152 where @xmath153 is the @xmath154 cross - correlation matrix on the @xmath74th subcarrier , with its entries @xmath155 s denoting the normalized cross correlation coefficient between the signature waveforms of the @xmath156th and @xmath11th users on the @xmath74th subcarrier .
@xmath157 represents the @xmath158 channel matrix , given by @xmath159 where the channel coefficients @xmath160 , @xmath161 , are assumed to be i.i.d .
complex gaussian r.vs ( i.e. , fade amplitudes are rayleigh distributed ) with zero mean and @xmath162=e\big[\big(h_{kq}^{(i)}\big)^2\big]=0.5 $ ] , where @xmath163 and @xmath164 are the real and imaginary parts of @xmath160 .
the @xmath146-length data vector @xmath37 is given by @xmath165^t,\ ] ] and the @xmath154 diagonal amplitude matrix @xmath166 is given by @xmath167 where @xmath168 denotes the transmit amplitude of the @xmath12th user .
the @xmath146-length noise vector @xmath169 is given by @xmath170^t,\ ] ] where @xmath171 denotes the additive noise component of the @xmath12th user on the @xmath74th subcarrier , which is assumed to be complex gaussian with zero mean with @xmath172=\sigma^2 $ ] when @xmath173 and @xmath172=\sigma^2\rho_{kj}^{(i)}$ ] when @xmath174 .
we assume that all the channel coefficients are perfectly known at the receiver .
we note that once the likelihood function for the mc - cdma system in the above is obtained , it is straightforward to adopt the las algorithm for mc - cdma .
accordingly , in the multicarrier system considered , the most likely @xmath36 is taken as that @xmath36 which maximizes @xmath175 the likelihood function in ( [ lamda1 ] ) can be written in a form similar to eqn .
( 4.11 ) in @xcite as @xmath176 where @xmath177 now observing the similarity between ( [ 3eq5 ] ) and ( [ 3eq5w ] ) in sec . [ sec3aa ] , the las algorithm for mc - cdma can be arrived at , along the same lines as that of v - blast in the previous section , with @xmath58 , @xmath59 and @xmath75 replaced by @xmath178 , @xmath179 , and @xmath180 , respectively , with all other notations , definitions , and procedures in the algorithm remaining the same .
the complexity of the proposed detector for mc - cdma can be analyzed in a similar manner as done for v - blast in sec .
first , given an initial vector , the las operation part alone in mc - cdma has an average per - bit complexity of @xmath181 , which is due to @xmath4 initial computation of @xmath87 in ( [ 3eq9w ] ) , which requires @xmath181 complexity per bit , @xmath5 update of @xmath48 in each step as per ( [ 3eq21w ] ) , which requires @xmath182 complexity for sequential las , and hence constant per - bit complexity , and @xmath6 the average number of steps required to reach a fixed point , which , through simulations , is found to have a constant per - bit complexity .
next , the initial vector generation using mmse or zf has a per - bit complexity of @xmath183 for @xmath184 . finally , combining the above complexities involved in the las part and the initial vector generation part , the overall average per - bit complexity of the mmse / zf - las detector for mc - cdma is @xmath183 .
the initial vector generation using mf has a per - bit complexity of only @xmath185 .
hence , if the mf output is used as the initial vector , then the overall average per - bit complexity of the mf - las is the same as that of the las alone , which is @xmath181 . for large @xmath146 ,
the performance of mf - las , zf - las , and mmse - las are almost the same ( see fig . [ fig_mc2 ] ) , and hence mf - las is preferred because of its linear complexity in number of users , @xmath146 , for a given @xmath10 .
we evaluated the ber performance of the proposed las detector for mc - cdma through simulations .
we evaluate the uncoded ber performance of the proposed las detector as a function of average snr , number of users ( @xmath146 ) , number of subcarriers ( @xmath10 ) , and number of chips per bit ( @xmath148 ) .
we also evaluate the ber performance as a function of _ loading factor _
, @xmath186 , where , as done in the cdma literature @xcite , we define @xmath187 .
we call the system as underloaded when @xmath188 , fully loaded when @xmath189 , and overloaded when @xmath190 . random binary sequences of length @xmath148 are used as the spreading sequences on each subcarrier . in order to make a fair comparison between the performance of mc - cdma systems with different number of subcarriers , we keep the system bandwidth the same by keeping @xmath191 constant . also ,
in that case we keep the total transmit power to be the same irrespective of the number of subcarriers used . in the simulation plots we show in this section ,
we have assumed that all users transmit with equal amplitude .
the las algorithm used is the slas with circular checking of bits starting from the first user s bit . , @xmath192 , @xmath193 , i.e. , @xmath194.,width=360 ] , for single carrier cdma ( @xmath195 ) in rayleigh fading for a fixed @xmath194 and average snr = 15 db .
@xmath148 varied from 15 to 1500.,width=336 ] first , in fig .
[ fig_mc1 ] , we present the ber performance of mf / zf - las detectors as a function of average snr in a single carrier ( i.e. , @xmath195 ) _ underloaded system _ , where we consider @xmath194 by taking @xmath192 users and @xmath193 chips per bit . for comparison purposes
, we also plot the performance of mf and zf without las .
single user ( su ) performance , which corresponds to the case of no multiuser interference ( i.e. , @xmath196 ) , is also shown as a lower bound on the achievable multiuser performance . from fig .
[ fig_mc1 ] , we can observe that the performance of mf and zf detectors are far away from the su performance .
whereas , the zf - las as well as mf - las detectors almost achieve the su performance .
we point out that , like zf detector , other suboptimum detectors including mmse , sic , and pic detectors @xcite also do not achieve near su performance for the considered loading factor of 2/3 , whereas the mf - las detector achieves near su performance , that too at a lesser complexity than these other suboptimum detectors .
next , in fig . [ fig_mc2 ]
, we show the ber performance of the mf / zf - las detectors for @xmath195 as a function of number of users , @xmath146 , for a fixed value of @xmath194 at an average snr of 15 db .
we varied @xmath146 from 10 to 1000 users .
su performance is also shown ( as the bottom most horizontal line ) for comparison .
it can be seen that , for the fixed value of @xmath194 , both the mf - las as well as the zf - las achieve near su performance ( even in the presence of 1000 users ) , whereas the zf and mf detectors do not achieve the su performance . in fig .
[ fig_mc3 ] , we show the ber performance of the mf / zf - las detectors as a function of average snr for different number of subcarriers , namely , @xmath197 , keeping a constant @xmath198 , for a _ fully loaded system _ ( i.e. , @xmath189 ) with @xmath199 . keeping @xmath189 and @xmath199 for all cases means that @xmath4 @xmath200 for @xmath195 , @xmath5 @xmath201 for @xmath202 , and @xmath6 @xmath203 for @xmath204 .
the su performance for @xmath195 ( 1st order diversity ) , @xmath202 ( 2nd order diversity ) , and @xmath204 ( 4th order diversity ) are also plotted for comparison .
these diversities are essentially due to the frequency diversity effect resulting from multicarrier combining of signals from @xmath10 subcarriers .
it is interesting to see that even in a fully loaded system , the mf / zf - las detectors achieve all the frequency diversity possible in the system ( i.e. , mf / zf - las detectors achieve su performance with 1st , 2nd and 4th order diversities for @xmath205 and 4 , respectively ) . on the other hand , zf detector is unable to achieve the frequency diversity in the fully loaded system , and its performance is very poor compared to mf / zf - las detectors .
next , in fig .
[ fig_mc4 ] , we present the ber performance of zf / mf - las detectors in a mc - cdma system with @xmath204 as a function of loading factor , @xmath186 , where we vary @xmath186 from @xmath206 to 1.5 .
we realize this variation in @xmath186 by fixing @xmath207 , @xmath204 , and varying @xmath148 from 300 to 5 .
the average snr considered is 8 db . from fig .
[ fig_mc4 ] , it can be observed that as @xmath186 increases all detectors loose performance , but the mf / zf - las detectors can offer relatively good performance even at _ overloaded conditions _ of @xmath190 .
another observation is that at @xmath190 , mf - las performs slightly better than zf - las .
this is because @xmath190 corresponds to a high interference condition , and it is known in mud literature @xcite that zf can perform worse than mf at low snrs and high interference . in such cases , starting with a better performing mf output as the initial vector , mf - las performs better . , @xmath189 , @xmath199 , @xmath198.,width=336 ] , for multicarrier cdma in rayleigh fading .
@xmath204 , @xmath207 , @xmath148 varied from 300 to 5 , average snr = 8 db .
, width=360 ] further to our present work on the application of mf / zf - las detectors for mc - cdma , several extensions are possible on the practical application of these detectors in cdma systems .
two such useful extensions are @xmath4 mf / zf / mmse - las for frequency selective cdma channels with rake combining ; we point out that a similar approach and system model adopted here for mc - cdma is applicable , by taking a view of equivalence between frequency diversity through mc combining and multipath diversity through rake combining , and @xmath5 mf / zf / mmse - las for asynchronous cdma systems , which can be carried out once the system model is appropriately written in a form similar to ( [ eqna ] ) .
these two extensions can allow mf / zf - las detectors to be practical in cdma systems ( e.g. , 2 g and 3 g cdma systems ) , with potential for significant gains in system capacity .
current approaches to mud considered in practical cdma systems appear to be mainly pic and sic .
however , the illustrated fact that mf - las can easily outperform pic / sic detectors both in performance and complexity for large number of users suggests that mf - las can be a powerful mud approach in practical cdma systems .
we presented a near - capacity achieving , low - complexity detector for large mimo systems having tens to hundreds of antennas , and showed its uncoded / coded ber performance in the detection of v - blast and in the decoding of full - rate non - orthogonal stbcs from da .
the proposed detector was shown to have excellent attributes in terms of both low complexity as well as nearness to theoretical capacity performance , achieving high spectral efficiencies of the order of tens to hundreds of bps / hz . to our knowledge , our reporting of the decoding of a large full - rate non - orthogonal stbc like @xmath3 stbc from da and its ber / nearness to capacity results is for the first time in the literature .
we further point out that the proposed detector has good potential for application in practical mimo wireless standards , e.g. , the low - complexity feature of the proposed detector can allow the inclusion of @xmath208 , @xmath209 , @xmath3 non - orthogonal stbcs from da into mimo wireless standards like ieee 802.11n and ieee 802.16e , which , in turn , can achieve higher spectral efficiencies than those are currently possible in these standards .
we conclude this paper by pointing to the following remark made by the author of @xcite in its preface in 2005 : _ `` it was just a few years ago , when i started working at at&t labs research , that many would ask ` who would use more than one antenna in a real system ? ' today , such skepticism is gone . ''
_ extending this sentiment , we believe large mimo systems would be practical in the future , and the practical feasibility of low - complexity detectors like the one we presented in this paper could be a potential trigger to create wide interest in the implementation of large mimo systems .
for example , antenna / rf technologies and channel estimation for large mimo systems could open up as new focus areas .
potential large mimo applications include inter - base station / base station controller back - haul connectivity using large mimo links , and wireless iptv .
other interesting large mimo applications can be thought of as well .
h. vikalo and b. hassibi , `` on the sphere - decoding algorithm ii .
generalizations , second - order statistics , and applications to communications , '' _ ieee trans . signal process .
53 , no . 8 , pp . 2819 - 2834 , august 2005 . h. d. zhu , b. farhang - boroujeny , and r .-
chen , `` on the performance of sphere decoding and markov chain monte carlo detection methods , '' _ ieee signal proc .
letters _ , vol .
669 - 672 , october 2005 .
k. higuchi , h. kawai , n. maeda , h. taoka , and m. sawahashi , `` experiments on real - time 1-gb
/ s packet transmission using mld - based signal detection in mimo - ofdm broadband radio access , '' _ ieee jl .
areas in commun .
1141- 1153 , june 2006 .
p. w. woliniansky , g. j. foschini , g. d. golden , and r. a. valenzuela , `` v - blast : an architecture for realizing very high data rates over the rich - scattering wireless channel , '' _ proc .
295 - 300 , september - october 1998 .
g. d. golden , g. j. foschini , r. a. valenzuela , and p. w. wolniansky , `` detection algorithm and initial laboratory results using v - blast space - time communication architecture , '' _ electron .
14 - 16 , january 1999 .
y. sun , `` hopfield neural network based algorithms for image restoration and reconstruction
part i : algorithms and simulations , '' _ ieee trans . on signal process .
48 , no . 7 , pp .
2105 - 2118 , july 2000 .
k. vishnu vardhan , saif k. mohammed , a. chockalingam , and b. sundar rajan , `` a low - complexity detector for large mimo systems and multicarrier cdma systems , '' _ ieee jsac spl .
iss . on multiuser detection for advanced communication systems and networks _ , vol . 26 , no .
473 - 485 , april 2008 .
saif k. mohammed , k. vishnu vardhan , a. chockalingam , and b. sundar rajan , `` large mimo systems : a low - complexity detector at high spectral efficiencies '' _ accepted for presentation in ieee icc2008 _ , beijing , china , may 2008 .
saif k. mohammed , a. chockalingam , and b. sundar rajan , `` a low - complexity near - ml performance achieving algorithm for large mimo detection , '' _ accepted for presentation in ieee isit2008 _ , toronto , canada , july 2008 .
v. tarokh , n. seshadri , and a. r. calderbank , `` space - time codes for high data rate wireless communications : performance criterion and code construction , '' _ ieee trans .
inform . theory _
744 - 765 , march 1998 .
b. a. sethuraman , b. s. rajan and v. shashidhar , `` full - diversity , high - rate space - time block codes from division algebras , '' _ ieee trans . inform . theory _
2596 - 2616 , october 2003 .
t. l. marzetta , `` blast training : estimating channel characteristics for high capacity space - time wireless , '' _ proc .
37th annual allerton conf . on communication , control , and computing _ , pp .
958966 , september 1999 s. manohar , v. tikiya , r. annavajjala , and a. chockalingam , `` ber - optimal linear parallel interference cancellation for multicarrier ds - cdma in rayleigh fading , '' _ ieee trans .
commun . , _ vol
2560 - 2571 , july 2007 .
k. vishnu vardhan was born in andhra pradesh , india .
he received the undergraduate degree in electronics and communication engineering from pondicherry university , pondicherry , india , in 2005 .
he received the postgraduate degree in telecommunication engineering from indian institute of science , bangalore , india , in 2007 . since july 2007
, he has been with cisco systems ( india ) private limited , bangalore , india .
his research interests include multiuser detection and low - complexity detectors for cdma and mimo systems .
saif khan mohammed received his b.tech degree in computer science and engineering from the indian institute of technology , new delhi , india , in 1998 . from 1998 to 2000 , he was employed with philips inc . ,
bangalore , india , as an asic design engineer . from 2000 to 2003 , he worked with ishoni networks inc . , santa clara , ca , as a senior chip architecture engineer .
from 2003 to 2007 , he was employed with texas instruments , bangalore as systems and algorithms designer in the wireless systems group .
he is currently pursuing his doctoral degree in electrical and communication engineering at the indian institute of science , bangalore , india .
his research interests include low - complexity detection , estimation and coding for wireless communications systems .
a. chockalingam was born in rajapalayam , tamil nadu , india .
he received the b.e .
( honors ) degree in electronics and communication engineering from the p. s. g. college of technology , coimbatore , india , in 1984 , the m.tech degree with specialization in satellite communications from the indian institute of technology , kharagpur , india , in 1985 , and the ph.d .
degree in electrical communication engineering ( ece ) from the indian institute of science ( iisc ) , bangalore , india , in 1993 . during 1986 to 1993 , he worked with the transmission r & d division of the indian telephone industries limited , bangalore . from december 1993 to may 1996 , he was a postdoctoral fellow and an assistant project scientist at the department of electrical and computer engineering , university of california , san diego . from may 1996 to december 1998
, he served qualcomm , inc . , san diego , ca , as a staff engineer / manager in the systems engineering group . in december 1998 , he joined the faculty of the department of ece , iisc , bangalore , india , where he is an associate professor , working in the area of wireless communications and networking .
dr . chockalingam is a recipient of the swarnajayanti fellowship from the department of science and technology , government of india .
he served as an associate editor of the ieee transactions on vehicular technology from may 2003 to april 2007 .
he currently serves as an editor of the ieee transactions on wireless communications .
he is a fellow of the indian national academy of engineering .
b. sundar rajan was born in tamil nadu , india .
he received the b.sc .
degree in mathematics from madras university , madras , india , the b.tech degree in electronics from madras institute of technology , madras , and the m.tech and ph.d .
degrees in electrical engineering from the indian institute of technology , kanpur , india , in 1979 , 1982 , 1984 , and 1989 respectively .
he was a faculty member with the department of electrical engineering at the indian institute of technology in delhi , india , from 1990 to 1997 .
since 1998 , he has been a professor in the department of electrical communication engineering at the indian institute of science , bangalore , india .
his primary research interests include space - time coding for mimo channels , distributed space - time coding and cooperative communication , coding for multiple - access and relay channels , with emphasis on algebraic techniques .
rajan is an associate editor of the ieee transactions on information theory , an editor of the ieee transactions on wireless communications , and an editorial board member of international journal of information and coding theory .
he served as technical program co - chair of the ieee information theory workshop ( itw02 ) , held in bangalore , in 2002 .
he is a fellow of the indian national academy of engineering and recipient of the iete pune center s s.v.c aiya award for telecom education in 2004 .
rajan is a member of the american mathematical society . | we consider large mimo systems , where by ` _ large _ '
we mean number of transmit and receive antennas of the order of tens to hundreds .
such large mimo systems will be of immense interest because of the very high spectral efficiencies possible in such systems .
we present a low - complexity detector which achieves uncoded near - exponential diversity performance for hundreds of antennas ( i.e. , achieves near siso awgn performance in a large mimo fading environment ) with an average per - bit complexity of just @xmath0 , where @xmath1 and @xmath2 denote the number of transmit and receive antennas , respectively . with an outer turbo code ,
the proposed detector achieves good coded bit error performance as well .
for example , in a 600 transmit and 600 receive antennas v - blast system with a high spectral efficiency of 200 bps / hz ( using bpsk and rate-1/3 turbo code ) , our simulation results show that the proposed detector performs close to within about 4.6 db from theoretical capacity .
we also adopt the proposed detector for the low - complexity decoding of high - rate non - orthogonal space - time block codes ( stbc ) from division algebras ( da ) .
for example , we have decoded the @xmath3 full - rate non - orthogonal stbc from da using the proposed detector and show that it performs close to within about 5.5 db of the capacity using 4-qam and rate-3/4 turbo code at a spectral efficiency of 24 bps / hz . the practical feasibility of the proposed high - performance low - complexity detector could potentially trigger wide interest in the implementation of large mimo systems .
we also illustrate the applicability of the proposed detector in the low - complexity detection of large multicarrier cdma ( mc - cdma ) systems . in large mc - cdma systems with hundreds of users ,
the proposed detector is shown to achieve near single - user performance at an average per - bit complexity linear in number of users , which is quite appealing for its use in practical cdma systems .
large mimo systems , v - blast , non - orthogonal stbcs , low - complexity detection , high spectral efficiency , multicarrier cdma . |
recent research results indicate that radiation from gamma - ray bursts ( grbs ) may be constrained mostly in narrow beams @xcite , which suggests that grbs may be standard candles at cosmological distances @xcite .
previous studies of the grb space density distribution suggested that grb production rate may follow the observed star formation rate for small values of redshift @xcite , but may increase monotonically for very high values of redshift @xcite ; however neither the lower limit nor the upper limit of redshift for grb production could be determined reliably previously because these studies are either limited by statistics or have to rely on some empirical relationships between some statistical properties of selected samples of grbs . because both the space density distribution and beaming effect play important roles for the observed grb intensity distribution , here we make use of the batse / cgro grb intensity distribution , which is the largest sample of grbs collected so far , to determine directly the grb space density and opening angle distribution
for the grb space density distribution , we assume the following model , under the cosmological parameters of @xmath2 , @xmath3 km s@xmath4 mpc@xmath4 and @xmath5 , in order to compare with previous star formation rate measurements @xcite .
@xmath6 the functional form is taken from the measured star formation rate @xcite with @xmath7 and @xmath8 .
our main goal in this paper is to determine @xmath9 and @xmath10 for grbs from the observed grb intensity distribution , in order to answer the question where in the universe grbs are produced predominantly . for the grb opening angle distribution , we study three cases when the grb opening angle @xmath11 follows gaussian , exponential , or power - law distributions . for a gaussian distribution of @xmath11 ,
the probability that the opening angle equals to a given angle is , @xmath12 we assume that the total energy @xmath13 is the same for for every long grb , i.e. , they can be used as standard candles@xcite and the opening angle @xmath11 is very small , the probability that the observed fluence @xmath14 is higher than a specified value @xmath15 is then , @xmath16 @xmath17 where @xmath18 is the luminosity distance .
thus the number of bursts detected is , @xmath19 similarly , for an exponential distribution with the parameter @xmath20 , @xmath21 and for a power - law distribution with the parameter @xmath22 , @xmath23 if we assume the peak luminosity of grbs is the standard candle , we only need to replace @xmath13 and @xmath15 by the standard candle luminosity @xmath24 and the peak flux @xmath25 in equations ( 3 ) , ( 4 ) , ( 5 ) and ( 6 ) , respectively .
we number these slightly modified equations as ( 3)@xmath26 , ( 4)@xmath26 , ( 5)@xmath26 and ( 6)@xmath26 , respectively ; for brevity we do not list these equations explicitly .
we apply equation ( 1 ) for the grb space density distribution and equations ( 4 ) , ( 4)@xmath26 , ( 5 ) , ( 5)@xmath26 , ( 6 ) and ( 6)@xmath26 for the grb opening angle distribution to fit the observed grb intensity distribution .
we focus our study on long grbs ( @xmath27 ) @xcite because currently only long bursts are known to be originated at cosmological distances from the direct redshift measurements of the optical afterglows of some long grbs .
all long grbs with peak flux values and detection probability above 5% included in the grb 4b+ catalog @xcite are selected , resulting in 1727 grbs as our sample .
the fluence values of these grbs are taken also directly from the catalog .
for the peak flux distribution , the detection efficiency correction is done directly using the detection probability of each peak flux value given in the catalog .
because the fluence and peak flux does not have a good linear correlation , for each fluence range we average the detection probability of all grbs included in this fluence range and then correct for the detection efficiency of this fluence range .
total energy as standard candle . .@xmath28_dof_.
@xmath24 is in units of @xmath29 _ ergs / s_. @xmath30 in units of gpc@xmath31year .
all errors quoted are for 68.3% confidence . for both cases
the exponential distribution of the opening angle offers the best fits ; however the residuals for the combination of peak luminosity as the standard candle and the exponential distribution is the only statistically acceptable one . [ cols="^,^,^,^,^",options="header " , ] our fitting results are shown in @xmath32 and @xmath33 . because our fittings are done with the integral distributions , the numbers of grbs in different bins are not completely independent . to address this problem , we carried out extensive bootstrap simulations . taking the 1727 grbs as the seed distribution
, we randomly generated a large numbers of peak flux or fluence distributions and fitted the sampled distributions in the same way .
the best values and errors of all parameters are then calculated from these fittings ; the bootstrapped parameter values and errors are consistent with that listed in table 1 .
we therefore believe the correction between the different points does not have any significant impact to our results . from the fitting results , we conclude : * for long grbs , despite the different fitting residuals of the two standard candle models and three opening angle distributions assumed , the cosmological parameters of grbs inferred are strikingly similar , indicating the robustness of the grb cosmological model . * the power - law distribution for the grb opening angles can not fit the data and is thus rejected confidently . *
the gaussian distributions for the grb opening angles fit the data only marginally , and may be rejected too . * for both standard candle models
the exponential distribution for the grb opening angles provides the best fit to the data . *
the combination of the peak luminosity as standard candle and the exponential distribution for the grb opening angles is the only one with statistically acceptable fit to the data ; we therefore accept this as the accepted model for long grbs . under this model
the peak luminosity of a grb is @xmath34 _ ergs / s _ , the exponential constant of the opening angle distribution is @xmath35 ( radian@xmath4 ) , the lower bound to the grb redshift is @xmath36 , and the upper break redshift is @xmath37 ( no lower limit to the break is detected because not enough sufficiently faint grbs have been observed with batse ) .
first we compare our study with previous studies of grb cosmological models with batse intensity distribution @xcite + @xcite .
all these studies concluded that the dimmest grbs are at redshift around unity , probaly following the measured star formation rate which peaks also at redshift around unity @xcite .
the main differences between our approach and previous studies are the following : * we take the grb opening angle distribution as part of the cosmological model .
this is very important , since the observed grb flux depends strongly on the opening angle , as well as the redshift . *
we separate the long and short grbs as two different classes in our study .
our conclusions are only applicable to long grbs . for short grbs
, no conclusion can be drawn due to the limited statistical quality of data .
* one simplification of our study is that we did not make the energy spectral correction to the fluence or peak flux as a function of @xmath38 ( k - correction ) ; we just take the fluence or peak flux values as listed in the catalog .
technically this correction is not straightforward , because we can no longer uniquely convert the fluence or peak flux directly to redshift , because we believe that grb radiation is beamed , and the opening angle follows a certain distribution .
therefore the observed fluence or peak flux is due to both the redshift and the opening angle of the grb .
the good fit of the assumed simple cosmological model with only several free - parameters indicates that this correction may not impact our conclusions significantly .
in fact , since the energy spectra of grbs in the batse band follows approximately a power - law with an index of @xmath392 , the k - correction is not important . frail _ et al . _
( 2001 ) suggested that the total energy of grbs may be a constant , based on the opening angles of several grbs determined from their afterglow light curve break .
our result is not in major conflict with their suggestion , but favors the peak luminosity as the standard candle .
perhaps the direction of the grb beam is not completely stable , as indicated from observations of other relativistic jets and outflows from stellar mass and supermassive black holes ; only when the beam is pointed directly at us can we observe the maximum flux from the grb .
however it has been argued that for some grbs the lower limit of the total energy released is @xmath40 _ ergs _ as determined from the redshifted x - ray emission lines @xcite .
however the reality of the detection of the emission lines is still under debate , and an alternative interpretation of the detected emission lines also exists @xcite .
the opening angles of some grbs with afterglow detections have been determined previously from the grb afterglow data @xcite ; we call this frail - distribution . in @xmath41
we compare the frail - distribution with the opening angle distributions determined with the observed grb intensity distribution data , just as a sanity check .
the power - law distribution agrees with the frail - distribution for large opening angles , but fails to predict the absence of grbs detected with opening angles smaller than 0.04 radian .
no grbs with opening angles narrower than 0.04 radian was directly observed , but the recent report on the polarization of grb021206 suggests that it is possible for some grbs to have opening angles narrower than 0.01 radian @xcite .
the gaussian distribution agrees with the peak of the frail - distribution distribution reasonably well ; however this model can not predict enough grbs with opening angles greater than 0.2 radian .
the exponential distribution is consistent with the frail - distribution distribution for both standard candle models , but favors slightly the peak luminosity standard candle model .
it was suggested that for the model in which the grbs follow the star formation rate strictly , the most possible opening angle should be around 0.1 radian @xcite .
however , if the majority of faint grbs originate at higher redshift than the peak of star formation rate as our model predicts , then a narrower average opening angle is required .
we therefore conclude that grb opening angles follow an exponential distribution with a mean opening angle of @xmath42 radian , and the grb space density follows a power - law with an index of @xmath393.5 between @xmath36 and @xmath43 .
the general trend of the grb space density distribution determined here as a function of @xmath38 is consistent with previous studies @xcite ; however our result is statistically much more robust , and in particular @xmath9 is determined for the first time and the lower limit to @xmath10 is constrained more strictly than before .
our result therefore demonstrates that grbs are no longer produced frequently in the nearby and present universe at @xmath44 , and grbs are predominantly produced in the early universe at @xmath1 .
indeed most grbs with direct redshift measurements have redshifts @xmath45 @xcite , with the exceptions of grb 980425 ( @xmath38=0.0085 ) and grb 030329 ( @xmath38=0.1687 ) which may be associated with two peculiar but similar supernovae sn1998bw @xcite , and sn2003dh @xcite .
it remains to be seen if these grbs belong to a subset of grbs which are different from the majority of grbs which seem to be produced at much larger redshift .
it has been suggested that grbs follow star formation rate , because it is commonly believed that grbs are produced as the final gravitational collapse of massive stars @xcite .
the grb space density determined here follows the measured star formation rate between @xmath0 to @xmath46 .
the determination of @xmath36 suggests that the massive stars capable of producing the majority of long grbs are no longer formed frequently today .
future grb experiments capable of measuring the redshift of many more grbs and thus providing direct distance measurements of grbs , such as the _ swift _ mission to be launched in late 2003 @xcite , will test this prediction firmly .
the monotonic increase of the grb space density up to at least @xmath47 suggests that the formation rate of massive stars capable of producing grbs should peak at significantly higher @xmath38 than the measured star formation rate .
this implies that the dark ages " of the universe ended much earlier than previously believed at @xmath48 based on the measured space density distribution of high - redshift qsos @xcite .
there are also some previous independent works showing that an essential amount of long grbs should be at @xmath49 @xcite or even up to @xmath50@xcite . a recently detected grb031203 by integral was reported to be located at @xmath51 ( formally @xmath52 ) @xcite , consistent with our model prediction .
therefore grbs may be a probe of the epoch of reionization of the early universe @xcite .
currently the grb intensity distribution data can not constrain the exact upper limit of @xmath10 , because not enough extremely faint grbs have been observed .
future grb experiments , such as the proposed next generation grb mission "
@xcite would detect many more grbs originated in the extremely early universe , which may determine the exact ending time of the dark ages " of the universe in order to test the prediction of some numerical simulations @xcite which suggests that the first stars may have been formed at @xmath53 .
grbs may prove to be the most sensitive and ultimate probes into the early epoch of the universe when the first structures were formed .
* acknowledgement : * we thank drs .
zigao dai , tan lu , yongfeng huang and daming wei for valuable comments to the manuscript .
this study is supported in part by the special funds for major state basic research projects ( 10233010 ) and by the national natural science foundation of china .
snz also acknowledges supports by nasa s marshall space flight center and through nasa s long term space astrophysics program . | it has been known that at least some of the observed gamma - ray bursts ( grbs ) are produced at cosmological distances and the grb production rate may follow the star formation rate .
we model the batse detected intensity distribution of long grbs in order to determine their space density distribution and opening angle distribution .
our main results are : the lower and upper distance limits to the grb production are @xmath0 and @xmath1 , respectively ; the grb opening angle follows exponential distribution and the mean opening angle is about 0.03 radian ; the peak luminosity appears a better standard candle than the total energy of a grb . |
this paper is the second part of a big work dedicated to diffraction of a plane wave by a thin infinite impedance strip . in @xcite ( which will be referred to as part
i hereafter ) some preliminary steps were made .
namely , the diffraction problem was formulated and symmetrized .
functional problems of the wiener
hopf class with entire functions were introduced . using the method of embedding formula these problems
were reduced to two auxiliary problems .
finally , two riemann hilbert problems were formulated .
the rimann
hilbert problems are formulated on the complex plane with cuts @xmath0 .
the cuts depend on the impedance of the segment @xmath1 . due to energy absorption / conservation principle
the impedance should obey the condition @xmath2 \le 0 $ ] .
then , if @xmath3 > 0 $ ] the contours @xmath0 coincide with the undeformed contours @xmath4 shown in fig .
[ fig00 ] ( left ) .
these contours correspond to the trajectory of the square root @xmath5 as @xmath6 takes real values ( we remind that @xmath7 has a small positive imaginary part ) . if @xmath3 \le 0 $ ] the contours @xmath0 are obtained from @xmath4 as the result of deformation shown in fig .
[ fig05 ] .
points @xmath8 in the figure are zeros of @xmath9 : @xmath10 the cuts @xmath0 are assumed to be symmetrical : @xmath11 . the aim of the deformation shown in fig .
[ fig05 ] is to make zeros of @xmath9 not belonging to the plane cut along @xmath12 . for the antisymmetrical auxiliary problem the riemann
hilbert has form : find a matrix function @xmath13 such that * it is regular on the complex plane cut along the lines @xmath12 ( see fig .
[ fig00 ] , left ) ; * it obeys the following functional equations connecting the values on the shores of the cuts : @xmath14 @xmath15 with coefficients @xmath16 @xmath17 * it obeys the following growth restrictions : @xmath18 \le \pi/2 , \label{eq0609c}\ ] ] @xmath19 \le \pi/2 , \label{eq0610c}\ ] ] @xmath20 \le \pi/2 , \label{eq0807}\ ] ] @xmath21 \le \pi/2 .
\label{eq0808}\ ] ] * functions @xmath22 grow no faster than a constant near the points @xmath23 .
[ whh ] notations @xmath24 correspond to the values of @xmath25 taken on the left and right shores of the cuts ( see fig .
[ fig00 ] ) ; @xmath26 square root is equal to @xmath7 at the point @xmath27 and then continued to @xmath12 , along the contours shown in fig .
[ fig00 ] , right . for the symmetrical case the riemann
hilbert problem has form : find a matrix function @xmath28 such that * it is regular on the plane cut along the lines @xmath12 ; * it obeys functional equations @xmath29 @xmath30 with coefficients @xmath31 @xmath32 on the cuts ; * it obeys the following growth restrictions : @xmath33 \le \pi/2 , \label{eq0609s}\ ] ] @xmath34 \le \pi/2 , \label{eq0610s}\ ] ] @xmath35 \le \pi/2 , \label{eq0807sym}\ ] ] @xmath36 \le \pi/2 .
\label{eq0808sym}\ ] ] * functions @xmath37 grow no faster than @xmath38 near the points @xmath23 .
[ whh_sym ] if we manage to find a solution of problem [ whh ] , we can recover a antisymmetrical part of the solution of original problem using following procedure .
first , functions @xmath39 are calculated : @xmath40 then function @xmath41 is found by the embedding formula : @xmath42 @xmath43 finally , the antisymmetrical part of the directivity is found : @xmath44 for the symmetrical case ( problem [ whh_sym ] ) the following formulae are used : @xmath45 @xmath46 @xmath47 the directivity related to the initial problem is a sum of the antisymmetrical and symmetrical part : @xmath48 in the present paper we solve problem [ whh ] and problem [ whh_sym ] .
we use for this the method of oe equation proposed recently .
the plan of the research is as follows .
first , a family of riemann hilbert problems indexed by an artificial parameter @xmath49 is formulated .
then , an ordinary differential equation with respect to @xmath49 ( ode1 ) is introduced .
this equation is supplemented with initial conditions .
an oe equation ( an equation for the coefficients of ode1 ) is formulated .
this equation is solved numerically ; the results are compared with solutions obtained by the integral equation method .
a crucial step of the method is introducing of the family of riemann - hilbert problems to which problems [ whh ] and [ whh_sym ] belong as an element .
before we introduce such a family it is necessary to reformulate the riemann
hilbert problems ( problem 1 and 2 ) in such a way that the connection matrices @xmath50 and @xmath51 have eigenvalues tending to 1 as @xmath52 .
one can see that matrices @xmath53 satisfy this condition ( so no reformulation is needed ) , while matrices @xmath50 have one eigenvalue tending to 1 , and the other tending to @xmath54 . to reformulate the antisymmetrical problem make the following variable change : @xmath55 the growth restrictions for the new functions become as follows : @xmath56 \le \pi/2 , \label{eq0810}\ ] ] @xmath57 \le \pi/2 , \label{eq0811}\ ] ] @xmath58 \le \pi/2 , \label{eq0812}\ ] ] @xmath59 \le \pi/2 .
\label{eq0813}\ ] ] the connection formulae for @xmath60 on the cuts become as follows : @xmath61 @xmath62 where @xmath63 @xmath64 we can formulate now a functional problem for @xmath60 , which replaces problem [ whh ] : find a matrix function @xmath65 of elements ( [ eq0809 ] ) such that * it is regular and has no zeros of determinant on the plane cut along the lines @xmath0 ; * it obeys functional equations ( [ eq0814 ] ) , ( [ eq0815 ] ) with coefficients ( [ eq0816 ] ) , ( [ eq0817 ] ) on the cuts ; * it obeys growth restrictions ( [ eq0812 ] ) , ( [ eq0813 ] ) , ( [ eq0814 ] ) , ( [ eq0815 ] ) ; * components of @xmath60 grow no faster than @xmath66 near the points @xmath23 .
[ whh_v ] consider the antisymmetrical case , i. e.problem [ whh_v ]
. represent contours @xmath67 as @xmath68 , @xmath69 , where @xmath70 is a contour going from @xmath71 to @xmath72 . here
@xmath73 or @xmath74 means a shift of the contour .
let @xmath75 , @xmath76 , be a contour going from @xmath71 to @xmath49 along @xmath70 .
i. e. @xmath75 is a part of @xmath70 .
let be @xmath77 the family of the riemann
hilbert problems is built based of problem [ whh_v ] .
the key step is to replace the contours @xmath0 with @xmath78 .
the growth conditions at infinity and the connection matrices ( [ eq0816 ] ) , ( [ eq0817 ] ) remain the same as for problem [ whh_v ] , while the growth restrictions at the ends of the contours should be changed ( since the ends of the contours change from @xmath23 to @xmath79 ) . to formulate these restrictions
we should study behavior of a solution of the riemann
hilbert problem near the end of one of the contours .
for example , consider contour @xmath80 .
let equation ( [ eq0815 ] ) with coefficient ( [ eq0817 ] ) be fulfilled on its shores .
then , obviously , at the vicinity of the end point @xmath81 the solution has form @xmath82 where @xmath83 is an arbitrary matrix analytical near @xmath84 , @xmath85 and @xmath86 are eigenvalues of @xmath87 , @xmath88 is the matrix of eigenvectors of @xmath87 .
the branch of the square root at @xmath89 is chosen according to the explanation above .
an appropriate choice of the logarithms in ( [ eq0906 ] ) determines the growth restrictions near @xmath84 .
choose @xmath90 ( this corresponds to a regular component of the solution ) .
then , consider the function @xmath91 .
obviously @xmath92 , @xmath93 .
introduce an important value @xmath94 which will be called _ the index _ of the riemann
hilbert problem discussed here .
the notation above denotes the continuous change of the logarithm value along the contour @xmath95 .
obviously , @xmath96 for some integer @xmath97 .
it is not difficult to show ( see appendix a ) that under the restriction @xmath2 < 0 $ ] @xmath98 define also the value @xmath99 this function should be continuous on @xmath95 , and besides @xmath100 according to ( [ eq0903 ] ) , @xmath101 . introduce a family of @xmath102 matrix functions @xmath103 such that for any fixed @xmath49 the function @xmath103 taken as the function of @xmath6 is a solution of the following functional problem : find a matrix function @xmath103 with elements denoted as ( [ eq0809 ] ) such that * it is regular and has no zeros of determinant on the plane cut along the lines @xmath104 ; * it obeys functional equations ( [ eq0814 ] ) , ( [ eq0815 ] ) with coefficients ( [ eq0816 ] ) , ( [ eq0817 ] ) on the cuts @xmath104 ; * it obeys growth restrictions ( [ eq0810 ] ) , ( [ eq0811 ] ) , ( [ eq0812 ] ) , ( [ eq0813 ] ) at infinity ; * near @xmath105 components @xmath106 behave as regular functions , components @xmath107 behave as @xmath108 , where @xmath109 and @xmath110 are some functions regular near zero ; * near @xmath84 components @xmath111 behave as regular functions , components @xmath106 behave as @xmath112 , where @xmath113 and @xmath114 are some functions regular near zero .
[ family ] the definition of @xmath115 is mathematically correct , since uniqueness of @xmath103 can be proven for each @xmath49 .
the proof is based on the determinant technique introduced in part i. problem [ whh_v ] and problem [ family ] are connected via the relation @xmath116 in the symmetrical case introduce a family of functions @xmath117 such that for any fixed @xmath49 the function @xmath117 taken as the function of @xmath6 is a solution of the following functional problem : find a matrix function @xmath117 such that * it is regular and has no zeros of determinant on the plane cut along the lines @xmath104 ; * it obeys functional equations ( [ eq0801sym ] ) , ( [ eq0802sym ] ) with coefficients ( [ eq0806sym ] ) , ( [ eq0805sym ] ) on the cuts @xmath104 ; * it obeys growth restrictions ( [ eq0609s ] ) , ( [ eq0610s ] ) , ( [ eq0807sym ] ) , ( [ eq0808sym ] ) at infinity ; * near @xmath105 components @xmath118 behave as regular functions , components @xmath119 behave as @xmath120 , where @xmath121 and @xmath122 are some functions regular near zero ; * near @xmath84 components @xmath123 behave as regular functions , components @xmath118 behave as @xmath124 , where @xmath125 and @xmath126 are some functions regular near zero .
[ familysym ] problem [ whh_sym ] and problem [ familysym ] are connected via the relation @xmath127
we are looking for an ordinary differential equation ( ode1 ) in the form @xmath128 where @xmath129 is the coefficient of the equation . indeed , this equation is useful only if the coefficient @xmath130 has structure simpler than that of @xmath60 .
the form of the coefficient is given by the following theorem .
function @xmath103 , which is a solution of a family of functional problems introduced as problem [ family ] , obeys equation ( [ eq1001 ] ) with the coefficient @xmath131 where @xmath132 , @xmath76 is a @xmath102 matrix function ( not depending on @xmath6 ) ; @xmath133 is connected with @xmath134 via the relation : @xmath135 i. e. to obtain @xmath133 one has to interchange first the rows and then the columns of @xmath134 .
[ ode1 ] * proof * construct the coefficient of ode 1 as follows : @xmath136 consider this combination for fixed @xmath49 as the function of @xmath6 . according to problem
[ family ] , @xmath137 has no singularities on the complex plane @xmath6 cut along the contours @xmath138 .
moreover , since the functions @xmath139 do not depend on @xmath49 , the values of @xmath130 on the left and right shores of @xmath104 are equal : @xmath140 @xmath141 thus , function @xmath130 is single - valued on the plane @xmath6 .
the only singularities it can have are the ends of the contour @xmath104 , i. e. the points @xmath142 . consider the vicinity of the point @xmath143 . to study the singularity of @xmath130 at this point , represent the solution @xmath60 in the form @xmath144 where @xmath145 is a regular function in the area considered
, @xmath146 is the same function as ( [ eq0908 ] ) . substituting ( [ eq1004 ] ) into ( [ eq1003 ] ) , obtain that the coefficient @xmath130 can only have a simple pole at @xmath147 .
let @xmath132 be the residue of @xmath129 at @xmath143 .
similar consideration can be performed with respect to the point @xmath148 .
consider the geometrical symmetry of the problem , namely @xmath149 .
this symmetry transforms the matrix of solutions as follows : @xmath150 due to this transformation , the coefficient @xmath129 has form of ( [ eq1002 ] ) .
@xmath151 similarly to the antisymmetrical case one can prove the following theorem .
function @xmath117 , which is a solution of a family of functional problems introduced as problem [ familysym ] , obeys equation @xmath152 with the coefficient @xmath153 where @xmath154 , @xmath76 is a @xmath102 matrix function ( nor depending on @xmath6 ) ; operator @xmath155 is as introduced above [ ode1s ] initial conditions for ode1 ( [ eq1001 ] ) and ( [ eq1001s ] ) are as follows : @xmath156 @xmath157 [ initialconditions ] * proof * consider the antisymmetrical case , i. e.consider problem [ family ] for some large positive imaginary @xmath49 .
the functional problem can be reduced to a system of integral equations as follows .
introduce the matrix @xmath158 then introduce the functions @xmath159 @xmath160 and @xmath161 @xmath162 , such that @xmath163 @xmath164 assume that contours @xmath165 go from @xmath166 to @xmath167 .
note that for @xmath168 @xmath169 where the integral has sense of the main value . according to the functional equation ( [ eq0802 ] ) , the following equation is valid : @xmath170
@xmath171 according to the geometrical symmetry , @xmath172 here @xmath173 and @xmath174 are corresponding elements of matrix @xmath175 .
note that for large @xmath176 $ ] , @xmath177 , the values of @xmath178 are close to 0
( this is the reason for variable change from @xmath25 to @xmath60 ) .
also , under the same condition @xmath179 is close to zero .
if @xmath180 $ ] is large enough , the system ( [ eq1009a ] ) can be solved by iterations . for large @xmath180 $ ] only the zero - order approximation can be left , i.e. @xmath181 can be set to zero .
this corresponds to ( [ eq1006 ] ) . in the symmetrical case
the proof is similar .
introduce the following notation . consider matrix ode @xmath182 taken on a contour @xmath183 with starting point @xmath184 and ending point @xmath185 .
let the initial condition have form @xmath186 by definition , @xmath187 \equiv { \rm x } ( \tau_2 ) .
\label{eq1102}\ ] ] the following properties are obvious : * if @xmath188 is the contour @xmath183 passed in the opposite direction , then @xmath189 = \left ( { \rm oe}_{h}\left [ { \rm k } ( \tau)\ , d\tau \right ] \right)^{-1 } \label{eq1103}\ ] ] * if @xmath183 is a concatenation of @xmath190 and @xmath191 ( @xmath190 is the first ) then @xmath192 = { \rm oe}_{h_2}\left [ { \rm k } ( \tau)\ , d\tau \right ] { \rm oe}_{h_1}\left [ { \rm k } ( \tau)\ , d\tau \right ] \label{eq1104}\ ] ] according to theorems [ ode1 ] and [ initialconditions ] , the solution of problem [ whh_v ] can be written as @xmath193 \pi(k ) , \label{eq1201}\ ] ] we remind that contour @xmath70 goes from @xmath71 to @xmath72 . a detailed study based on the continuation of matrices @xmath139 near the contours @xmath194 ( see @xcite ) shows that the coefficient @xmath130 is analytical with respect to the variable @xmath49 in a narrow strip surrounding contour @xmath70 .
thus , the contour can be slightly deformed without changing the result , provided that the starting and the ending points of the contour remain the same .
draw contours @xmath195 and @xmath196 as it is shown in fig .
[ fig06 ] .
these contours are needed to calculate the values @xmath197 and @xmath198 , @xmath199 without allowing singulatities of the coefficient of ode 1 .
namely , @xmath200 \pi(k ) , \label{eq1202}\ ] ] @xmath201 \right)^{-1 } \pi(k ) , \label{eq1203}\ ] ] define contour @xmath202 as a concatenation of contours @xmath195 and @xmath196 ( @xmath195 is the first ) . according to functional equation ( [ eq0815 ] ) ,
the following relation is valid : @xmath203 \pi(k ) = \tilde { \rm m}_2 ( k ) .
\label{eq1204}\ ] ] this is the oe
equation for the considered problem . due to the geometrical symmetry ,
equation ( [ eq0814 ] ) will be also valid .
formulate the problem for the oe
equation , to which the antisymmetrical problem of diffraction by an impedance strip becomes reduced .
find function @xmath132 for @xmath76 analytical in a narrow strip surrounding @xmath70 such that equation ( [ eq1204 ] ) is valid for each @xmath199 .
[ oe ] the following problem should be solved in the symmetrical case .
find function @xmath204 analytical in a narrow strip surrounding @xmath70 such that equation @xmath205 \pi(k ) = { \rm n}_2 ( k ) \label{eq1204sym}\ ] ] be valid for each @xmath206 .
solving of the diffraction problem by means of the proposed technique comprises the following steps : * contour @xmath70 ( see fig . [ fig06 ] ) is discretized .
problem [ oe ] is solved numerically ( the procedure is described below ) . as the result , the coefficient @xmath132 becomes known in a set of points @xmath207 covering contour @xmath70 densely .
* points of interest are selected in the @xmath6-plane .
a good choice is the set of points @xmath208 densely covering the segment @xmath209 , since such points enable one to construct the directivity of the field . for these points
the values @xmath210 are found by formula ( [ eq1201 ] ) , i. e.by solving a linear ode with known coefficients and known initial conditions . *
matrix @xmath211 is found at the points @xmath212 by inverting formula ( [ eq0809 ] ) .
* functions @xmath213 are found for @xmath212 by formula ( [ u0tildej ] ) . *
functions @xmath213 are substituted into the embedding formula ( [ embedding_a ] ) to get the function @xmath214 . *
the directivity is found using formula ( [ eq0617 ] ) for the points @xmath215 .
one can see that all steps of this procedure can be done easily except the first one . to solve problem [ oe ] we use the technique introduced in @xcite . here
we describe it .
matrix @xmath216 can be represented in the form : @xmath217 where @xmath91 , @xmath218 are introduced by ( [ eq0901 ] ) and ( [ eq0908 ] ) , respectively .
the left hand side of ( [ eq1204 ] ) can be rewritten as follows : @xmath219\ ] ] @xmath220{\rm f}^{-1}(k),\ ] ] where @xmath221 is a loop of a small radius @xmath222 encircling point @xmath223 in the positive direction , and @xmath224.\ ] ] here @xmath225 is contour going from @xmath226 to the start of the loop @xmath221 along @xmath195 ( see fig .
[ fig06b ] ) .
let @xmath132 be represented in the form @xmath227 the columns of @xmath228 are the eigenvectors of @xmath229 .
almost everywhere matrix @xmath228 can be parametrized as follows : @xmath230 one can see that as @xmath231 @xmath232 \to\ ] ] @xmath233 it follows from ( [ oerez ] ) that the eigenvalues of @xmath132 are connected with the eigenvalues of @xmath234 : @xmath235 thus , to find @xmath229 one needs only to find @xmath236 . for @xmath237 define the function @xmath238\ ] ] where contour @xmath239 is shown in fig .
[ fig06a ] .
function @xmath240 obeys the equation @xmath241 - \frac{1}{k + ( k_0 + \beta)}\left[{\rm r}^*(\beta),{\rm k}(\beta , k)\right ] , \label{voe}\ ] ] where @xmath242 $ ] is a commutator . since @xmath243 is adjoint to @xmath232,\ ] ]
its eigenvalues are the same , it can be written as follows : @xmath244 where @xmath245 taking @xmath246 obtain the relation @xmath247 , and thus @xmath248 taking @xmath249 obtain the relation @xmath250 , and thus @xmath251 where @xmath252 is introduced by ( [ eq0908 ] ) .
elementary calculations demonstrate that equation ( [ voe ] ) is equivalent to the following system of two independent riccati equations : @xmath253 @xmath254 thus we have to find @xmath255 such that there exist solutions @xmath256 of ( [ riccati ] ) on the part of @xmath257 contained between the points @xmath258 and @xmath71 obeying boundary conditions ( [ riccond1 ] ) , ( [ riccond2 ] ) .
this can be done using following numerical procedure .
first , contour @xmath70 should be meshed , i. e.an array of nodes @xmath259 should be taken on it .
starting point @xmath260 should have the imaginary part big enough .
the end point @xmath261 is equal to @xmath72 . at each point @xmath262 matrix @xmath263
is represented in the form ( [ eigtildem2 ] ) , i. e. the values @xmath91 and @xmath264 are computed .
the value @xmath265 is computed by applying formula ( [ ksi12 ] ) . for the `` infinity '' point
@xmath260 the following values are assigned : @xmath266 this is a natural choice for the asymptotics of the unknown coefficient , since @xmath267 tends to identity matrix , as @xmath268 .
the loop over @xmath269 is performed . at the @xmath270th step of the loop
the values @xmath271 are computed . thus on the @xmath270th step all values @xmath272 are already found . on the @xmath270th step of the loop equations ( [ riccati ] )
are solved on the contour @xmath273 for @xmath274 using runge
kutta 4 method .
the initial conditions are set as @xmath275 following from ( [ riccond2 ] ) .
the values @xmath276 are found
. then equations ( [ riccati ] ) are solved on the segment @xmath277 ( only one step is performed ) by euler method .
this method does not require the values of the right hand side at the end of the segment to be known .
the values @xmath278 are found .
the assignment @xmath279 is performed , following from ( [ riccond1 ] ) .
thus matrix @xmath280 becomes known .
it just remains to solve ode1 ( [ eq1001 ] ) and then calculate antisymmetrical part of the directivity using the procedure described above .
this is performed easily .
numerical results are compared with solution obtained by the method of boundary integral equations ( see appendix b ) .
the dependence of @xmath281 on @xmath282 for @xmath283 and @xmath284 is presented in fig . [ fig07 ] .
solid line corresponds to the method of integral equation and dotted line corresponds to the method of oe equation .
one can see that agreement is reasonable .
the solution procedure in the symmetrical case is similar .
here we just present the final results .
they are showed in fig .
[ fig08 ] . dependence of @xmath285 on @xmath282 for @xmath283 and @xmath284 is displayed .
solid line corresponds to the method of integral equation and dotted line corresponds to the method of oe equation .
in the current paper we present a new approach to matrix riemann
hilbert problems related to the problem of diffraction by an impedance strip .
the problems are of a quite general nature , so the methods proposed here can potentially be applied to a wide class of problems .
the technique is based on an analytical result expressed in theorem 1 and 2 .
the initial problem is embedded into a family of similar problems indexed by parameter @xmath49 , and it is shown that the dependence of the solution on @xmath49 is described by an ordinary differential equation with a relatively simple coefficient . then , the riemann hilbert problem is reformulated as a problem for an oe equation , i. e. a problem of reconstruction of the coefficients of an ode by using the boundary data .
there is no analytical solution for the oe problem in the general case , however some analytical technique is available in the commutative case @xcite .
it is also worth to note that numerical solution of the oe problem can be very efficient since the problem is of volterra nature ( the unknown function on a contour is found step by step ) . to demonstrate the practical value of the analytical results obtained here we performed some computations of the directivities for an impedance strip and compared the results with the integral equation method .
the agreement is nice , and this fact means for us mainly the validity of the method in general and the absence of mistakes in the main formulae .
here we do not pursue the aim to establish a robust and accurate numerical procedure based on the new method .
the work is supported by the grants rfbr 14 - 02 - 00573 , scientific schools 283.2014.2 , and the government grant 11.g34.31.0066 .
the authors are grateful to participants of the seminar on wave diffraction held in s.pb .
branch of steklov mathematical institute of ras ( the chairman is prof . v.
m. babich ) for interesting discussions .
let us prove formula ( [ eq0903 ] ) .
obviously , @xmath286|^{{i}\infty}_{0},\ ] ] where @xmath287 $ ] is the argument of the function .
introduce @xmath288 @xmath289 thus @xmath290|^{{i}\infty}_{0 } - { i}{\rm arg}[f_2(b)]|^{{i}\infty}_{0}.\ ] ] here we consider only the case @xmath3 > 0 $ ] for which no deformation of contour @xmath70 is needed , due to the rules of analytical continuation our proof is valid for all @xmath1 lying in the lower half plane .
one can notice(see fig .
[ fig09 ] ) that the argument of @xmath291 changes from @xmath292 $ ] to @xmath293 while @xmath49 goes from @xmath72 to @xmath226 along @xmath70 .
therefore @xmath294 ) - i(\pi/2 - { \rm arg}[-\eta ] ) = i({\rm arg}[-\eta ] - arg[\eta ] ) = { i}\pi.\ ] ]
* the antisymmetrical case . * to verify the results obtained by the oe
equation method we also solved the problem of diffraction by an impedance strip using the integral equation method . in the antisymmetrical case one can obtain following equation with the help of double layer potential : @xmath295 where @xmath296 @xmath297 is a hankel function of the first kind , @xmath298 is a double layer potential : @xmath299 @xmath300 is the antisymmetrical part of the scattered field @xmath301 . on the strip
, @xmath298 is connected with @xmath302 by a simple formula : @xmath303 antisymmetrical part of the directivity can be calculated using formula @xmath304 * symmetrical case . * in the symmetrical case it is natural to use a single layer potential .
one can obtain the following integral equation : @xmath305 where @xmath97 is a single layer potential : @xmath306 here @xmath307 is the symmetrical part of the scattered field @xmath301 .
the normal derivative of the field on the strip is connected with @xmath308 as follows : @xmath309 symmetrical part of the directivity can be calculated using the formula : @xmath310 99 a.v.shanin , a.i.korolkov , diffraction by an impedance strip i. symmetrization , embedding formula , and riemann hilbert problem _ submitted to qjmam _
a.v.shanin , solution of riemann - hilbert problem related to wiener - hopf factorization problem using ordinary differential equations in the commutative case , _ quart .
_ * 66 * ( 2013 ) 533555 .
a.v.shanin , an ode - based approach to some riemann hilbert problems motivated by wave diffraction , 2012 , arxiv:1210.1964 . | the current paper is the second part of a series of two papers dedicated to 2d problem of diffraction of acoustic waves by a segment bearing impedance boundary conditions . in the first part
some preliminary steps were made , namely , the problem was reduced to two matrix riemann
hilbert problem . here
the riemann
hilbert problems are solved with the help of a novel method of oe equations .
each riemann hilbert problem is embedded into a family of similar problems with the same coefficient and growth condition , but with some other cuts .
the family is indexed by an artificial parameter .
it is proven that the dependence of the solution on this parameter can be described by a simple ordinary differential equation ( ode1 ) .
the boundary conditions for this equation are known and the inverse problem of reconstruction of the coefficient of ode1 from the boundary conditions is formulated .
this problem is called the oe equation . the oe
equation is solved by a simple numerical algorithm . |
clusters consisting of some 10@xmath0 to 10@xmath1 atoms , which can be produced for instance in the course of the condensation of vapour , have attracted great interest during the last 20 years .
these mesoscopic clusters were found to exhibit classical as well as quantum mechanical features . in the special case of metallic clusters @xcite ,
the conduction electrons , which move in the average potential produced by the positively charged ions , exhibit shell effects . as a consequence ,
the binding energy of the cluster becomes a function of the number of free electrons ( and thus also of the number of ions ) . in turn , this implies measurable variations of the formation probability of a cluster as a function of its mass in a thermal vapour - gas equilibrium .
other interesting types of mesoscopic systems are the so - called fullerenes which represent a crystal of atoms covering a closed , usually spherical surface @xcite .
a number of review articles exist on this exciting new field of physics ( see for instance ref .
@xcite and ref .
@xcite and books mentioned at the end of ref .
@xcite and ref .
@xcite ) . in the present paper
, we consider a mesoscopic system which consists of a spherical metallic core of radius @xmath2 carrying @xmath3 positive charges and surrounded by an insulating layer of external radius @xmath4 ( see fig . 1 ) .
we determine analytically the bound states of a single electron in the potential @xmath5 which is produced by the positive surface charge density of the metallic core and the insulating dielectric surface layer . as in the case of an ordinary atom , these single electron wave functions describe also approximately the case that several electrons are bound in this potential . for @xmath3 bound electrons ,
the system has total charge 0 .
because of the obvious similarity of this system with an atom or ion , we refer to it as a mesoscopic atom or mesoscopic ion .
as one can surmise , there are two different types of bound electronic states : electrons localized in the metallic core inside of the insulating layer ( region ii , fig . 1 ) and electrons localized in the vacuum outside of the insulating layer ( region 0 , fig . 1 ) .
the bound states localized outside are reminiscent of the bound states of an electron in the coulomb potential of the atomic nucleus .
we are mainly interested in these externally localized electronic states ( class 1-states ) and their decay into lower - lying bound states predominantly localized within the metallic core ( class 2-states ) .
the paper is organized as follows : in sec . 2
, we formulate the single - particle hamiltonian and determine the eigenfunctions and eigenenergies of bound electrons .
in doing so , we take into account the polarization of the insulating surface layer which is produced by the uniform positive charge distribution on the surface @xmath6 of the metal , but we neglect the additional polarization effects produced by the externally localized electrons . in sec .
3 we determine these additional polarization effects within the classical maxwellian electrodynamics .
they lead to an additional potential acting on the electrons which is of importance close to the surface @xmath7 of the insulator .
so far , we have not yet studied the effect of this additional potential on the externally localized electrons .
the sec . 4
is dedicated to a presentation of the results . in subsec .
4.1 , we describe the procedure which we used in order to obtain the eigenvalues and eigenfunctions and we present typical examples of spectra and wavefunctions . in subsec . 4.2
, we discuss the electromagnetic decay of a predominantly externally localized bound electron into a lower - lying internally localized state .
we show that the transition rate and the corresponding life - time may vary by several orders of magnitude depending on the parameters of the system .
in general , the transition is hindered by the small overlap between externally and internally localized states ( class 1-states and class 2-states ) if the insulating surface layer is sufficiently thick .
in particular , we discuss the special role in this decay of class 3-states which exhibit non - negligible amplitudes both in the external region 0 and the internal region ii . finally , in sec . 5 , we summarize the results and discuss some related open questions .
[ fig1 ] , surface @xmath6 ) covered by an insulating layer ( outer radius @xmath4 , surface @xmath7 ) .
states of bound electrons predominantly localized in the surrounding vacuum ( @xmath8 , region 0 ) are called class 1-states in distinction from states localized in the metallic core ( class 2-states).,width=264 ]
in the single - particle hamiltonian for an electron bound to a positively charged and insulated metallic core ( @xmath9 = mass of the electron ) @xmath10 the potential @xmath5 contains the interaction of the electron with the positive surplus charge @xmath11 ( @xmath12 = elementary charge ) and with the ionic background ( jellium ) of the metal .
furthermore , it contains the effects of the insulating surface layer .
the radius @xmath2 of the metal and the thickness ( @xmath13 ) of the surface layer are assumed to be of the order of 5 to 20 (1 = 10@xmath14 m ) and the number @xmath3 of positive charges of the metal is considered to be 1 to 5 . for these dimensions of the system we may expect that , on the one hand , the dynamics of the electron is fully quantum - mechanical and , on the other hand , we may use macroscopic concepts in formulating the potential @xmath5 , at least as a first approximation .
so we assume that the positive surplus charge is uniformly distributed over the surface @xmath6 of the metal @xmath15 the coulomb potential produced by this simple charge distribution acting upon an electron ( charge @xmath16 ) is given by @xmath17 where @xmath18 is the heaviside function @xmath19 inside of the metal , the conduction electrons feel not only the constant potential produced by the positive surface charge but also the attractive interaction with the ionic background which we represent by a constant negative potential @xmath20 @xmath21 ) .
on the other hand , the average energy gap above the fermi level in the insulator may act upon the electron like a repulsive barrier of a height @xmath22 .
last not least , there are polarization effects produced by the positive surplus charge and also by the charge distribution @xmath23 of the externally ( i.e. predominantly in region 0 ) localized electrons .
the polarization effects are investigated in sec . 3 on the basis of maxwell s equations for the electric field @xmath24 and for the displacement field @xmath25 in the presence of homogeneous dielectric media .
in the case that we only take into account the polarization effects produced by the positive surplus charge ( 2 ) , the following modifications of the coulomb potential ( 3 ) are to be considered : within region i , i.e. within the insulating layer , the coulomb potential @xmath26 is to be replaced by @xmath27 and within region ii , i.e. within the metallic core , the constant potential @xmath28 is to be supplanted by @xmath29 within region 0 , i.e. in the vacuum surrounding the insulating layer , the value @xmath30 of the coulomb potential ( 3 ) is unchanged .
one should note that the potential in region i assumes the value @xmath31 at the outer surface @xmath32 of the insulating layer .
let us denote the sum of the constant parts of the potentials in region i and ii by @xmath33 and @xmath34 if @xmath3 is equal to 1 or 2 and the dielectric constant @xmath35 exceeds 1 only slightly , and if the radii are in the range of values treated in sec . 4 , the 2@xmath36 terms on the right - hand sides of eqs . ( [ eq2.4 ] ) and ( [ eq2.5 ] )
are much smaller than the 1@xmath37 terms .
so far , it is only these cases which we treated numerically . altogether , the potential @xmath5 felt by the electron has the form @xmath38 \theta_0(r_2 - r ) + \left[b - { ze^2_0\over \varepsilon_{\rm i}r}\right ] \theta_0(r_1 - r ) \theta_0(r - r_2 ) - { ze^2_0\over r } \theta_0(r - r_1)\,\ , .
\label{eq2.5.1}\ ] ] the constants @xmath39 , @xmath35 and @xmath40 are empirical parameters .
the potential @xmath5 is shown in fig . 2 for a specific choice of the parameters . [ fig2 ] for the following choice of parameters : @xmath41 ; @xmath42 ; @xmath43 ; @xmath44 ev ; @xmath45 ev ; @xmath46 ev ; @xmath47 ; @xmath48,width=264 ] the eigenfunctions @xmath49 of the schrdinger equation @xmath50 separate @xmath51 and the radial functions @xmath52 and @xmath53 have to satisfy the differential equations @xmath54 @xmath55 respectively .
, the potential @xmath5 is constant and thus the eq .
( [ eq2.8 ] ) represents the differential equation for spherical bessel functions .
of the 2 linearly independent solutions , only the regular one , @xmath56 , qualifies as a physical solution .
up to a normalization factor which will be determined later on , the radial function @xmath52 in region ii is simply given by @xmath57 where @xmath58 is defined by @xmath59 and @xmath60 is the spherical bessel function of the 1@xmath37 kind ( see chapt
. 9 of ref .
@xcite ) . as bound states have an energy between the well depth @xmath61 and 0 , the parameter @xmath58 is assured to be real .
let us postpone the treatment of the general case and first turn to the simpler situation when the @xmath62-dependent term @xmath63 is negligible in comparison to the constant @xmath64 in the entire region i. this is so , if the inequality @xmath65 holds .
for the case shown in fig . 2 , the condition is fulfilled .
generally speaking , the barrier heights @xmath40 range from 2 to 5 ev , and the dielectric constant @xmath66 may have values between slightly more than 1 up to more than 100 ( examples : @xmath35 ( caoutchouc ) = 4 ; @xmath35 ( ti0@xmath67-crystal ) = 115 ) .
let us now investigate the case that the potential ( [ eq2.5.1 ] ) can be approximated by @xmath68 \theta_0(r_2 - r ) + b\theta_0 ( r_1 - r ) \theta_0(r - r_2 ) - { ze^2_0\over r}\theta_0(r - r_1)\,\ , .
\label{eq2.13}\ ] ] the radial schrdinger equation ( [ eq2.8 ] ) in region i then assumes again the form of bessel s differential equation .
but as the energy @xmath69 is lower than the barrier height @xmath64 , the solution in region i is given by a linear combination of two independent modified spherical bessel functions
( see chapt .
10.2 of ref .
@xcite ) .
we choose the pair @xmath70 , where @xmath71 and @xmath72 are the spherical bessel - function and spherical hankel - function of the 1@xmath37 kind , resp .
, with imaginary arguments .
the dimensionless real variable @xmath73 is related to the radial coordinate @xmath62 by @xmath74 @xmath75 the two factors @xmath76 and @xmath77 are chosen so that the pair of solutions are simply related to the modified bessel function of integer order @xmath78 and @xmath79 for which a number of rapidly converging series expansions exist ( see ref .
@xcite , chapt .
10.2 ) @xmath80 the radial function @xmath52 in region i thus has the general form @xmath81 , it is more convenient to deal with the differential eq .
( [ eq2.9 ] ) for @xmath53 which has the form @xmath82 we have to find a normalizable solution of ( [ eq2.18 ] ) .
it turns out that the only normalizable solution of ( [ eq2.18 ] ) has the form @xmath83 where the dimensionless variable @xmath84 is related to @xmath62 by @xmath85 with @xmath86 and where @xmath87 is a particular solution kummer s differential equation @xmath88 the parameters @xmath89 and @xmath90 in ( [ eq2.22 ] ) are related to the parameters in eq .
( [ eq2.18 ] ) by @xmath91 @xmath92 @xmath93 and the solution @xmath94 of ( [ eq2.22 ] ) is defined as follows : + we write @xmath90 in the form @xmath95 which implies @xmath96 then the function @xmath97 , where @xmath98 reads ( see ref .
@xcite , eq .
( 13.1.6 ) or ref .
@xcite , eq .
( 13 ) of chapt .
6.7.1 ) @xmath99\right\}\nonumber \\ & + & { ( n-1)!\over \gamma(a_0 ) } \sum^{n-1}_{\nu=0 } { ( a_0-n)_\nu\over ( 1 - n)_\nu } \cdot { z^{\nu - n}_0\over \nu ! } \label{eq2.25}\end{aligned}\ ] ] in ( [ eq2.25 ] ) , the function @xmath100 is the logarithmic derivative of the @xmath101-function @xmath102 and @xmath103 is pochhammer s symbol @xmath104 furthermore , @xmath105 is the confluent hypergeometric function @xmath106 we note in passing that the complicated form of the solution ( [ eq2.25 ] ) is due to the fact that the parameter @xmath90 in @xmath94 is an integer . for non - integer @xmath90
, the function @xmath94 depends in a simple way on two confluent hypergeometric functions ( see ref .
@xcite , eq .
( 13.1.3 ) ) .
this form of @xmath94 becomes singular if @xmath90 tends to an integer value .
the limiting process which leads to the form ( [ eq2.25 ] ) is discussed in ref .
@xcite . for any value of @xmath90 ,
the asymptotic behaviour of @xmath107 for @xmath108 is given by @xmath109 \label{eq2.29}\ ] ] ( see ref .
@xcite , eq .
( 13.1.8 ) ) . from ( [ eq2.29 ] ) and ( [ eq2.19 ] )
it is clear that @xmath53 is a normalizable function irrespectively of the value of the energy @xmath69 .
however , it turns out that the continuity conditions can only be met for certain discrete values of @xmath69 .
the radial function @xmath110 in region 0 is thus defined by @xmath111 before we proceed to the explicit formulation of the continuity conditions , we return to the form of the solution in region i in the general case that the @xmath62dependent term @xmath63 is neglected . then the differential equation for @xmath53 in region i has the form @xmath112 eq .
( [ eq2.31 ] ) resembles eq .
( [ eq2.18 ] ) .
in fact , with the replacements @xmath113 @xmath114 eq . ( [ eq2.18 ] ) becomes eq .
( [ eq2.31 ] ) .
+ with the ansatz @xmath115 we obtain again kummer s differential equation for @xmath116 @xmath117 where @xmath90 is again given by ( [ eq2.23.2 ] ) or ( [ eq2.24.1 ] ) , but @xmath118 is obtained from @xmath89 by the replacements ( [ eq2.32.1 ] ) , ( [ eq2.32.2 ] ) : @xmath119 @xmath120 and the dimensionless variable @xmath121 is related to @xmath62 by @xmath122 the parameter @xmath123 defined in ( [ eq2.15 ] ) agrees with @xmath124 because @xmath125 and @xmath126 . the general solution @xmath127 of eq .
( [ eq2.34 ] ) can be written as a linear combination of the function @xmath128 ) and , as linearly independent 2@xmath36 solution , the confluent hypergeometric function @xmath129 : @xmath130 we note that we had to suppress the solution @xmath131 in region 0 because it would lead to a non - normalizable solution .
the radial solution @xmath52 in region i thus assumes the form @xmath132 @xmath133 \,\ , .
\label{eq2.37p}\ ] ] we now turn to the continuity conditions which enable us to determine the amplitudes @xmath134 , @xmath135 ( or @xmath136 , @xmath137 ) , @xmath138 , and the energy eigenvalues @xmath69 . once these parameters will be obtained , we shall normalize the entire radial function to 1 .
the radial wave function @xmath52 and its 1@xmath37 derivatives must be continuous at the limits @xmath139 and @xmath32 of the regions ii and i : @xmath140 @xmath141 @xmath142 @xmath143 we proceed as follows : + eqs .
( [ eq2.38 ] ) and ( [ eq2.39 ] ) are 2 linear inhomogeneous equations for the amplitudes @xmath134 , @xmath135 ( or @xmath136 , @xmath137 ) .
we solve them for these amplitudes and substitute the result into the equation @xmath144 i.e. the ratio of the eqs .
( [ eq2.41 ] ) and ( [ eq2.40 ] ) .
( [ eq2.42 ] ) represents a transcendental equation for the energy .
it is solved numerically
. the amplitude @xmath138 can subsequently be obtained from ( [ eq2.40 ] ) .
the resulting wavefunction @xmath52 is not yet normalized to 1 .
this can be easily achieved by the replacement @xmath145 with @xmath146 let us present this procedure in more detail : + we introduce the abbreviations : @xmath147_{x=\kappa_{\rm i}r } = \left[\sqrt{{\pi\over 2x } } i_{l+{1\over 2}}(x)\right]_{x=\kappa_{\rm i}r } \,\ , , \label{eq2.45.1}\ ] ] @xmath148_{x=\kappa_{\rm i}r } : = \left[\sqrt{{\pi\over 2x } } k_{l+{1\over 2}}(x)\right]_{x=\kappa_{\rm i}r } \,\ , , \label{eq2.45.2}\ ] ] @xmath149_{y=2\kappa_{\rm i}r } : = \left[y^l e^{-{y\over 2 } } m(a_{\rm i } , 2l+2,y)\right]_{y=2\kappa_{\rm i}r}\,\ , , \label{eq2.46.1}\ ] ] @xmath150_{y=2\kappa_{\rm i}r } : = \left[y^l e^{-{y\over 2 } } u(a_{\rm i } , 2l+2,y)\right]_{y=2\kappa_{\rm i}r}\,\ , .
\label{eq2.46.2}\ ] ] furthermore , we use the prime to denote the derivative with respect to the variable given as argument : @xmath151_{x=\kappa_{\rm i}r_2 } \,\,,\ ] ] @xmath152_{y=2\kappa_{\rm i}r_2 } \,\,,\ ] ] @xmath153_{x=\kappa_{\rm ii } r_2 } \,\,,\ ] ] a.s.o .
+ from the continuity equations ( [ eq2.38 ] ) , ( [ eq2.39 ] ) we find the amplitudes @xmath134 , @xmath135 and @xmath136 , @xmath137 to be given by very similar relations @xmath154_{\kappa_{\rm i}r_2 } } \,\ , , \label{eq2.47.1}\ ] ] @xmath155_{x=\kappa_{\rm i}r_2 } } \,\ , .
\label{eq2.47.2}\ ] ] the wronskian in the denominators is given by ( see ref .
@xcite , eq .
( 10.2.8 ) ) @xmath156 @xmath157_{y=2\kappa_{\rm i}r_2 } } \,\ , , \label{eq2.48.1}\ ] ] @xmath158_{y=2\kappa_{\rm i}r_2 } } \,\ , .
\label{eq2.48.2}\ ] ] the wronskian in the denominators of ( [ eq2.48.1 ] ) and ( [ eq2.48.2 ] ) can be related to the wronskian @xmath159 : @xmath160 using the differential eq .
( [ eq2.34 ] ) , one finds for @xmath161 ( see ref .
@xcite , eq .
( 13.1.22 ) ) @xmath162 and thus for @xmath163 : @xmath164 substituting ( [ eq2.47.3 ] ) into ( [ eq2.47.1 ] ) and ( [ eq2.47.2 ] ) and , analogously , substituting ( [ eq2.48.5 ] ) into ( [ eq2.48.1 ] ) and ( [ eq2.48.2 ] ) we obtain the following forms of the amplitudes in region i : @xmath165 \,\ , , \label{eq2.49.1}\ ] ] @xmath166 \,\ , , \label{eq2.49.2}\ ] ] @xmath167 \,\ , , \label{eq2.50.1}\ ] ] @xmath168 \,\ , .
\label{eq2.50.2}\ ] ] with ( [ eq2.49.1 ] ) , ( [ eq2.49.2 ] ) or ( [ eq2.50.1 ] ) , ( [ eq2.50.2 ] ) the amplitudes of the 2 linearly independent solutions in region i are represented as a function of the energy @xmath69 .
the eigenvalues can now be obtained from the ratio of the two continuity conditions ( [ eq2.41 ] ) and ( [ eq2.40 ] ) , i.e. from the equation @xmath169 this is a transcendental equation for the eigenenergies .
it has infinitely many solutions close to the energy @xmath170 due to the infinite range of the coulomb potential . in spite of the complicated nature of the wavefunctions
it is not difficult to find the numerical solutions of ( [ eq2.51 ] ) .
more explicitly , the logarithmic derivatives in ( [ eq2.51 ] ) look as follows : @xmath171_{r = r_1 } = \left[{{d\over dr}((2\lambda r)^l u(a_0 , 2l+2 , 2\lambda r ) e^{-\lambda r } ) \over ( 2\lambda r)^l u(a_0 , 2l+2 , 2\lambda r ) e^{-\lambda r}}\right]_{r = r_1 } \label{eq2.52}\ ] ] @xmath172_{r = r_1 } = \left[{{d\over dr } ( a_{\rm i}f^{(1)}_l(r ) + b_{\rm i } f^{(2))}_l(r ) ) \over a_{\rm i } f^{(1)}_l(r ) + b_{\rm i } f^{(2)}_l(r)}\right]_{r = r_1 } \label{eq2.53.1}\ ] ] @xmath172_{r = r_1 } = \left[{{d\over dr }
( \tilde a_{\rm i}\tilde f^{(1)}_l(r ) + \tilde b_{\rm i}\tilde f^{(2))}_l(r ) ) \over \tilde a_{\rm i } \tilde f^{(1)}_l(r ) + \tilde b_{\rm
i } \tilde f^{(2)}_l(r)}\right]_{r = r_1 } \label{eq2.53.2}\ ] ] eq .
( [ eq2.53.2 ] ) and ( [ eq2.53.1 ] ) deal with the case of the exact potential and of the approximate constant potential in region i. in calculating the derivative @xmath173 or @xmath174 one can use properties of the different functions involved .
having determined the eigenvalues from eq .
( [ eq2.51 ] ) one obtains the amplitude @xmath138 from eq .
( [ eq2.40 ] ) and thereafter the correct normalization of the entire function must be achieved by numerical calculation of the normalization constant ( [ eq2.44 ] ) .
in sec . 2 , only the polarization of the insulator which is produced by a constant positive surface charge density @xmath175 on the surface @xmath6 of the metal was taken into account .
however , the charge distribution @xmath23 of the externally localized electrons produces not only additional polarization effects in the insulating surface layer but also modifies the distribution of the positive surplus charge on the surface @xmath6 of the metal . in this section
, we present a simple macroscopic description of the polarization effects which is based on maxwell s equations in the presence of homogeneous dielectric media . outside of the metallic core ,
i.e. in the region i ( @xmath176 ) of the insulator , and in the vacuum 0 ( @xmath177 ) surrounding the insulator the electric field @xmath24 and the displacement @xmath25 are related by : @xmath178 where @xmath35 and @xmath179 are the dielectric constant and the electric susceptibility of the insulator , respectively .
the fields satisfy maxwell s equations @xmath180
@xmath181 for the sake of simplicity , we assume that the distribution @xmath23 of the externally located electrons and , as a consequence , also the positive surface charge distribution @xmath182 on @xmath6 , are , i.e. that they depend only on the polar angle @xmath183 and the radial coordinate @xmath62 . they can thus be written in the form @xmath184 integrating the surface charge density over the surface @xmath6 must yield the total positive surface charge @xmath11 .
this implies the condition @xmath185 or , with ( [ eq77 ] ) @xmath186 the components @xmath187 of the electronic charge distribution are related to the total electronic charge distribution by @xmath188 the maxwell equations ( [ eq74 ] ) and ( [ eq75 ] ) are equivalent to the poisson equation @xmath189 in region i and to @xmath190 in region 0 . here
, the potential @xmath191 is related to the electric field @xmath24 by @xmath192 the following boundary conditions must be fulfilled at the surfaces @xmath6 and @xmath7 : @xmath193 @xmath194
@xmath195 @xmath196 furthermore , the potential @xmath197 must satisfy the asymptotic condition @xmath198 the conditions ( [ eq86 ] ) and ( [ eq88 ] ) are obtained from the maxwell equation ( [ eq74 ] ) , and the conditions ( [ eq87 ] ) and ( [ eq89 ] ) result from ( [ eq75 ] ) and the requirement that the classical electric field should be everywhere finite . the solutions @xmath199 and @xmath200 of the poisson equations ( [ eq81 ] ) and ( [ eq82 ] ) in the regions i and 0 , respectively , can be written in the form : @xmath201 where @xmath202 is the coulomb potential produced by the electronic charge distribution @xmath203 @xmath204 the charge distribution @xmath23 itself is to be calculated from the wavefunctions of the bound and predominantly externally located ( i.e. in region 0 , @xmath177 ) electrons .
conduction electrons which are located internally ( i.e. in region ii , @xmath205 ) do not contribute to @xmath23 because they are already taken into account by the number @xmath3 of positive surplus charges .
the density @xmath23 is thus mainly located in region 0 and it is very small in region i , i.e. inside of the insulating surface layer , because the eigenenergies of the electrons are far below the positive barrier in region i. the function @xmath206 is the general azimuthally symmetric solution of the laplace equation in region i @xmath207 and @xmath208 the corresponding most general solution of the laplace equation in region 0 satisfying in addition the asymptotic condition ( [ eq90 ] ) : @xmath209 the coulomb potential @xmath202 can be written as @xmath210 where the functions @xmath211 and @xmath212 are related to the components @xmath187 of the electronic charge distribution by @xmath213 substituting the explicit forms ( [ eq94 ] ) to ( [ eq96 ] ) into the boundary conditions ( [ eq86 ] ) , ( [ eq88 ] ) , and ( [ eq89 ] ) , we obtain the relations @xmath214 furthermore , using the identity ( see for instance ref .
@xcite ) for @xmath215 @xmath216 \label{eq102}\ ] ] together with ( [ eq91 ] ) , ( [ eq92 ] ) and ( [ eq94 ] ) , the condition ( [ eq87 ] ) can be shown to be equivalent to the relation @xmath217 where @xmath218 .
( [ eq103 ] ) is seen to imply that the potential @xmath219 is constant on the surface @xmath6 . for @xmath220
, we obtain from ( [ eq99 ] ) @xmath221 \label{eq104}\ ] ] and from ( [ eq100 ] ) and ( [ eq101 ] ) the relations @xmath222 hence , with ( [ eq104 ] ) , we find @xmath223 \label{eq107}\ ] ] and @xmath224 \,\ , .
\label{eq108}\ ] ] for @xmath218 , we obtain from ( [ eq99 ] ) and ( [ eq103 ] ) : @xmath225 \,\ , , \label{eq109}\\ & & d^{\rm i}_l = { 4\pi\over ( 2l+1)\varepsilon_{\rm i } } \left[-b_l(r_2 ) + \rho^{s_2}_l r^{l+1}_2\right ] \,\ , . \label{eq110}\end{aligned}\ ] ] for the coefficients @xmath226 , we find the relations @xmath227 and @xmath228 \label{eq112}\ ] ] from the eqs .
( [ eq101 ] ) and ( [ eq100 ] ) , which hold for @xmath229 .
the equality of the righthand sides of ( [ eq111 ] ) and ( [ eq112 ] ) together with ( [ eq109 ] ) and ( [ eq110 ] ) yields a condition for the surface charge components @xmath230 for @xmath218 . from this equation one
finds the explicit form @xmath231 for @xmath218 .
the component @xmath232 is determined by the conservation of the total positive charge on @xmath6 and is given by eq .
( [ eq79 ] ) . for @xmath218 ,
one obtains from ( [ eq109 ] ) , ( [ eq110 ] ) and ( [ eq112 ] ) @xmath233 as we have mentioned already , the charge distribution @xmath234 of the externally located electrons is vanishingly small for @xmath235 .
consequently , we have @xmath236 and @xmath237 the potentials @xmath238 , @xmath239 , @xmath240 , and the charge distribution @xmath241 on the surface @xmath6 are thus fully determined for a given azimuthally symmetric distribution @xmath234 of the externally located electrons . the potential @xmath242 acting on an electron in our mesoscopic atom is related to the potential functions as follows : @xmath243 the potential @xmath244 represents the interaction of a conduction electron with the average ionic background charge in the metal and has to be added to the constant potential @xmath245 produced by the positive surplus charge on @xmath6 .
the potential barrier @xmath40 to be added in the insulating surface layer is related to the average energy gap between the electronic bands .
since it is of quantum mechanical origin , it is not contained in the classically determined term @xmath246 .
the schrdinger equation with the potential @xmath247 can no longer be solved analytically .
one could diagonalize the hamiltonian @xmath248 within the set of analytical eigenfunctions which we have determined in sec .
2 . this represents a selfconsistency problem , because the eigenfunctions of the hamiltonian ( [ eq120 ] ) depend on the charge distribution @xmath23 which is calculated from them .
the solution thus requires an iteration process .
let us consider two special cases for which the results of this section simplify considerably : * if we neglect the polarization effects produced by the external electrons , that is if we put @xmath249 the quantities @xmath211 , @xmath212 vanish for @xmath229 , and , as a consequence , so do the components @xmath250 .
the positive surplus charge density on @xmath6 is then spherically symmetric and given by @xmath251 it is easily seen that , in this case , the potential @xmath252 becomes equal to the potential @xmath5 we considered in sec . 2 .
* if we assume that the charge density @xmath23 of the external electrons is spherically symmetric @xmath253 only the component @xmath254 in ( [ eq76 ] ) is unequal 0 @xmath255 and , consequently , the quantities @xmath211 , @xmath212 , and @xmath230 vanish for @xmath218 . in this case , the potentials @xmath256 , @xmath199 and @xmath200 take the following form @xmath257 the case of a spherically symmetric distribution of the externally located electrons is realistic at finite temperature , because magnetic substates of a given electronic level are expected to be occupied with the same probability whenever the mesoscopic atom is in contact with a thermal reservoir of finite temperature . the functions @xmath258 and
@xmath259 depend themselves on the electronic distribution . for @xmath260 ,
@xmath258 decreases to zero as one can see from ( [ eq97 ] ) .
the quantity @xmath261 represents the total charge of the externally located electrons inside the sphere of radius @xmath62 @xmath262 using ( [ eq76 ] ) in ( [ eq128 ] ) we find @xmath263 for @xmath264 , @xmath265 tends to the total charge @xmath266 of the externally located electrons . if the mesoscopic atom is neutral , i.e. if there are as many externally bound electrons as there are positive surplus charges on the surface @xmath6 of the metallic core , the term @xmath267 in eq .
( [ eq127 ] ) cancels the potential @xmath268 for @xmath260 , as it should .
let us study the asymptotic behaviour of the function @xmath258 and @xmath259 for @xmath269 : for simplicity , let us assume that only the lowest external electron state is occupied by two electrons , one with intrinsic spin up and the other one with spin down .
the external state of lowest energy belongs necessarily to orbital angular momentum ( @xmath220 ) and to zero number of radial modes .
we assume that the wavefunctions we determined in sec .
2 are realistic for @xmath270 . in region 0 , the electronic density distribution is then given by @xmath271 or , substituting the explicit wavefunction ( [ eq2.19 ] ) , by @xmath272 here , @xmath273 is the overall normalization constant introduced by eq .
( [ eq2.44 ] ) . for @xmath274
, the function @xmath275 can be replaced by ( see ref .
@xcite , eq .
( 13.5.2 ) ) @xmath276 retaining only the 1@xmath37 term in ( [ eq132 ] ) , we obtain for the asymptotic density ( @xmath277 ) @xmath278 and for the function @xmath258 @xmath279 the integral in ( [ eq134 ] ) is the incomplete @xmath101-function ( see ref .
@xcite , eq .
( 6.5.3 ) , @xmath280 ) @xmath281 with the asymptotic expansion ( ref .
@xcite , eq .
( 6.5.32 ) @xmath282 from ( [ eq134])([eq136 ] ) , we obtain @xmath258 in the form @xmath283 for @xmath284 . of course , one could raise the accuracy of the result ( [ eq137 ] ) by incorporating one or two more terms in ( [ eq132 ] ) and ( [ eq136 ] ) .
the important point , however , is that @xmath258 tends to zero exponentially for large values of @xmath260 .
this justifies the assumption that the electronic charge density @xmath285 in the asymptotic regime @xmath260 is correctly described by the wavefunctions of sec . 2 , which do not contain the electronic polarization effects .
the function @xmath259 can be written in the form ( see ( [ eq98 ] ) ) @xmath286 with @xmath287 the 2@xmath36 term one the r.h.s . of ( [ eq138 ] )
is easily seen to have the approximate form @xmath288 for @xmath284 .
thus the function @xmath259 is asymptotically @xmath289 given by @xmath290 which means that it approaches exponentially the asymptotic value @xmath291 . in the general case of a @xmath183-dependent electronic charge density @xmath234 ,
the quantities @xmath292 and @xmath293 are different from zero and their asymptotic behaviour can be investigated in an analogous way .
one finds that @xmath294 tends exponentially to zero and @xmath295 exponentially to constant values as @xmath62 goes to infinity .
therefore , at large values of @xmath62 , the dominant part of the potential @xmath247 is the potential @xmath5 given by eq .
( [ eq2.5.1 ] ) .
thereby , our assumption that the wavefunctions determined in sec .
2 are essentially correct at large values of @xmath296 is seen to be justified . for non - asymptotic values of @xmath62 , the electronic charge density and
the resulting functions @xmath258 and @xmath259 must be obtained from the eigenfunctions of the hamiltonian ( [ eq119 ] ) which contains the potential @xmath247 defined in ( [ eq117 ] ) to ( [ eq119 ] ) .
let us formulate the special case of a spherically symmetric charge density and the resulting spherically symmetric potential @xmath5 in somewhat more detail : the electronic charge density at a finite temperature @xmath297 is given by @xmath298 or , using the relation ( see for instance ref .
@xcite , eq .
( 3.69 ) ) @xmath299 by @xmath300 here , @xmath301 is the @xmath302-th eigenvalue for given angular momentum @xmath303 and @xmath304 is the corresponding eigenfunction of the radial schrdinger equation .
the parameter @xmath305 represents the chemical potential which corresponds to the number of externally located electrons .
+ the potential @xmath247 has the form @xmath306 where @xmath307 , @xmath308 , and @xmath309 are given by the eq .
( [ eq104 ] ) , ( [ eq107 ] ) , and ( [ eq108 ] ) , resp .
the functions @xmath258 and @xmath259 are defined by ( [ eq97 ] ) and ( [ eq98 ] ) , where @xmath310 . the wavefunctions @xmath304 are solutions of the radial schrdinger equation with @xmath311 given by ( [ eq145 ] ) - ( [ 147 ] ) @xmath312 since @xmath311 depends itself on the functions @xmath313 , one has to solve ( [ eq148 ] ) by iteration . one way to find the solutions @xmath313 of ( [ eq148 ] ) is to expand the solutions @xmath304 in terms of the radial eigenfunctions @xmath314 defined in sec . 2 , where the potential @xmath5 is given by eq .
( [ eq2.5 ] ) and related to @xmath311 by @xmath315 @xmath316 @xmath317 are the eigenvalues obtained in sec . 2 .
( [ eq150 ] ) we left away the basis functions @xmath318 corresponding to continuous eigenvalues @xmath319 of eq .
( [ eq2.8 ] ) .
substitution of ( [ eq150 ] ) into ( [ eq148 ] ) leads to a nonlinear system of equations for the expansion coefficients @xmath320 which must be solved by iteration .
we present results on the spectrum of eigenenergies , the eigenfunctions , and on the electromagnetic transition rate between externally and internally localized states for a range of values of the parameters @xmath4 , @xmath2 , and @xmath321 .
so far , we have not taken into account the effects presented in sec . 3 , i.e. the polarization of the insulating dielectric material by externally localized electrons .
the additional potential produced by this polarization is expected to modify the wave functions in the vicinity of the surface @xmath7 , perhaps even rather substantially . at large distances ( say a couple of ngstrm outside of the surface @xmath7 ) , the additional polarization potential tends rapidly to zero .
consequently , the wavefunctions presented in this article are expected to be trustworthy at distances of a couple of ngstrm from the surface @xmath7 .
we studied only the case that the coulomb potential @xmath322 within the insulator ( @xmath323 ) is small compared to the barrier height @xmath64 which is related to the gap parameter @xmath40 by eq .
( [ eq2.4 ] ) .
this means that we neglected the coulomb potential in region i. .the energies @xmath324 , @xmath325 , and the barrier height @xmath64 as a function of the radii @xmath4 and @xmath2 and of the dielectric constant @xmath35 .
energies are given in mev and radii in .
the parameter @xmath40 is chosen equal to 2500 mev or 4000 mev .
[ tab1 ] [ cols="<,^,^,^,^,^,^,^,^",options="header " , ] augmenting @xmath4 for fixed value of @xmath2 enlarges the width of the barrier and , consequently , decreases the amplitude which an externally located electron has in the inner region ii . as the decay of a class 1-state into a class-2 state depends strongly on this tail , the transition probability must decrease and the lifetime must increase as a function of rising @xmath4 . the results presented in table 4 support this expectation .
the lifetime @xmath326 is found to rise roughly exponentially with @xmath327 .
this is so because the amplitude of the tail in region ii depends exponentially on @xmath327 as one can see from wkb .
the dependence of @xmath326 on @xmath4 is also shown in fig . 6 in logarithmic scale . [ fig6 ] .
apart from @xmath4 , all the other parameters are chosen as in fig .
3.,width=264 ] in fig . 7
, we show the dependence of the lifetime @xmath326 on the barrier height @xmath64 in logarithmic scale .
the lifetime increases as a function of rising @xmath64 .
the increase is somewhat smaller than exponential as can be expected from wkb .
[ fig7 ] in sec and logarithmic scale as a function of the barrier height @xmath64 .
apart from @xmath64 , all the other parameters are chosen as in fig .
3.,width=264 ] [ fig8 ] in sec and logarithmic scale as a function of @xmath2 .
apart from @xmath2 , all the other parameters are chosen as in fig .
3.,width=264 ] if one changes @xmath2 with all the other parameters fixed , one modifies predominantly the internally located class 2 states and the corresponding energies .
as one can see from fig .
8 , the lifetime @xmath326 as a function of rising @xmath2 decreases on the average as one would expect from the fact that the thickness @xmath327 of the surface layer decreases
. however , beside the average decrease of the lifetime , one observes a kind of resonance structure : within a change of @xmath2 by 1 , the lifetime @xmath326 can change by 10 orders of magnitude . + how can we understand these huge fluctuations of the lifetime ?
[ fig9 ] ) of lowest external electron to be found in the inner region ii ( @xmath328 ) as a function of @xmath2 .
apart from @xmath2 , all the other parameters are chosen as in fig .
3.,width=264 ] in fig .
9 , we show the probability of an externally located electron to be found in the inner region ii .
this localization probability in region ii , which is of crucial importance for the transition rate , is seen to show pronounced peaks at precisely those values of @xmath2 , where the lifetime has very small values as one can see by comparing fig .
the large variations of the transition rate as a function of @xmath2 ( and of @xmath3 ) can be simply explained as follows : + as one changes @xmath2 ( or @xmath3 ) , the eigenvalues of class 2-states move whereas the ones of class 1-states remain essentially the same .
more specifically , the eigenvalues @xmath329 of class-2 states are lowered , if the inner radius @xmath2 or the positive surplus charge @xmath3 is increased .
thereby , it happens regularly that a class 2-state which has the same angular momentum @xmath303 as the decaying class 1-state , crosses the energy of this class 1-state .
if one augments @xmath2 or @xmath3 on a sufficiently long scale , such crossings occur several times .
close to the crossing , the external class 1-state develops a non - negligible amplitude within region ii and the class 2-state develops an appreciable amplitude in the external region 0 .
we call such mixed states class 3-states . if the decaying class 1-state happens to have the nature of a class 3-state , its decay probability to lower - lying class 2-states is considerably enhanced as compared to the ordinary situation where no class 2-state of the same angular momentum is available
. we note in passing that this enhancement of the transition probability to lower - lying class 2-states does not only occur for dipole transitions but also for transitions of higher multipolarity .
so far we have not yet calculated the transition rates of higher multipolarity , because they would only matter if the decaying external state is a class 3-state .
the phenomenon can be considered as an example of the mechanism which landau and zener had studied in famous papers @xcite .
this is demonstrated in appendix .
[ fig10 ] ( sec , log.scale ) as a function of radius @xmath4 and @xmath2 ( in ) in a 3-dimensional plot .
except for @xmath4 and @xmath2 , all the parameters are chosen as in fig .
3.,width=377 ] in fig .
10 , the dependence of the lifetime @xmath326 on the radii @xmath4 and @xmath2 is shown in a 3-dimensional plot .
again one notices that the dependence of @xmath326 on @xmath4 is smooth whereas the one on @xmath2 contains resonances .
[ fig11 ] ( ) .
energies are given in units of @xmath330 mev .
except for @xmath2 , all the other parameters are chosen as in fig .
3.,width=377 ] in fig .
11 , the energy difference between the lowest - lying class 1 state and the closest lying class 2-state of the same angular momentum is shown .
the plot supports the explanation of the resonance phenomenon in terms of a coupling between close - lying states . in the numerical results presented in this section , the values @xmath64 of the barrier in region
i and the well depth @xmath39 of the average potential in region ii and the dielectric constant @xmath35 were chosen arbitrarily , and not determined for given choices of the material .
in this paper , we determined the quantum states and eigenenergies of an electron in the average potential produced by a positively charged metal cluster which is covered by an insulating surface layer . because of the similarity of the externally localized electronic states with the bound states of an electron in the coulomb potential of an atom
, we occasionally referred to the system as a mesoscopic atom or , in case of a non - zero total charge , as a mesoscopic ion .
in particular , we investigated the electromagnetic decay of an externally located electron ( class 1-state ) to a lower - lying unoccupied electronic state located within the metal ( class 2-state ) .
it turned out that the life - time depends sensitively on the radius @xmath2 of the metal core and on the number @xmath3 of positive surplus charges of the metal . in the model we studied , we found life - times varying between 10@xmath14 s and years depending on the value of @xmath2 or of @xmath3 . the origin of the large fluctuations was found to be the appearance of states with non - negligible amplitudes both outside inside of the insulating surface layer ( class 3-states ) .
the mechanism leading to these class 3-states is described in appendix .
the wavefunctions and eigenenergies were calculated neglecting the polarization of the insulating surface layer by externally located electrons .
the potential produced by these electronic polarization effects was determined in sec .
it was shown that this additional potential matters mainly in the vicinity of the outer surface @xmath7 of the insulator and decreases exponentially outside of this surface .
the polarization potential could be taken into account by diagonalizing the total hamiltonian including the polarization potential in the basis of the wavefunctions determined in this paper .
apart from this physical simplification of the problem we neglected the coulomb potential produced by the positive surplus charge within the insulator .
this approximation is justified whenever the change of the coulomb potential within the insulating surface layer is much smaller than the barrier height @xmath40 .
this condition was fulfilled in all the cases we studied numerically .
the 1@xmath37 improvement we already mentioned is to incorporate the additional potential between the external electron and the polarization cloud it produces in the insulator .
the form of this electronic polarization potential was derived in sec .
a 2@xmath36 improvement would be to replace the constant potentials @xmath20 in the metal and @xmath331 in the insulator by periodic potentials defined by the crystal structure of these materials .
of course , this is much more complicated and , to our knowledge , it has rarely been done so far for mesoscopic systems .
furthermore , the crystal structure in mesoscopic aggregates is in general poorly known . as a result
, the electronic eigenvalues of the total system are expected to exhibit a band structure similarly to a macroscopic solid , the insulating surface layer being characterized by a completely filled valence band and an empty conduction band .
the empty conduction band in the insulator can be at negative or at positive energy if the energy of the neutral system plus one electron at infinity is put equal to 0 .
describing the average potential in the insulator by a positive barrier @xmath22 is only meaningful if the conduction band of the insulator is at positive energy . as a further restrictive condition for our model to be a reasonable 1@xmath332 approximation there should be no localized traps for an additional electron in the uncharged insulator
this implies that the individual atoms or molecules of the insulator should not have bound states with an additional electron as a monomer @xcite . in general , bound states of an electron with a neutral atom or molecule do not exist for atoms or molecules with completely filled shells .
therefore , one should use insulators which exhibit this property .
presumably , al@xmath670@xmath333 fulfills this condition . finally , we neglected effects of the electron - phonon coupling altogether .
we presume that they tend to reduce the lifetime , but perhaps not drastically .
as long as one can only produce a limited number of mesoscopic atoms , one could just investigate experimentally the effects which we studied in the present paper , especially the amusing strong dependence of the lifetime on the radius and the charge of the metallic core .
although we would be highly pleased by an experimental scrutiny of our predictions , still more exciting open questions could be asked , if it were possible to produce longlived systems of the kind we presented in this paper , and in large quantities . in this case
, one would have access to the study of molecules formed from mesoscopic atoms and to condensed matter consisting of mesoscopic atoms as building blocks .
the presence of externally localized electrons would represent a fundamental difference from the case of a condensed system of uncharged mesoscopic clusters .
on the one hand , one may expect that , in a condensed matter phase of mesoscopic atoms , the weakly bound external electrons would partly dissociate from their mother atoms and move freely in the average field produced by the charged insulated cores .
we would thus surmise that such a system would be a ( good ) conductor . on the other hand
, the fact that the building blocks of such a condensed matter carry charge , will have a profound influence on the equilibrium state of the material .
probably , its density is lower than for uncharged constituents , where the cohesive forces are of van der waals type and , consequently , of much shorter range .
as we learnt from this paper , a mesoscopic system of the kind we investigated can only be stable in a strict sense , if there is no empty internally located electronic state below the energy of the lowest externally located state which is at an energy slightly above @xmath335 .
the positive surplus charge @xmath11 on the surface @xmath6 of the metal leads to a decrease of the well - depth @xmath336 as we have shown in sec . 3 : @xmath337
if , for the uncharged metal , the fermi level is at the energy @xmath338 ( with @xmath339 , @xmath340 ) , it lies at the modified energy @xmath341 in the charged metal .
so the condition for strict stability is that there is no level of an internally located electron between @xmath342 and @xmath335 .
as the separation energy @xmath343 of an electron from the uncharged metal and the fermi energy @xmath344 are related to @xmath339 by @xmath345 we can state that strict stability is achieved whenever there is no level between the energies @xmath346\right\}$ ] and @xmath335 .
this means that , metal with small @xmath347 is favourable for stability . in an ordinary metal ,
the separation energy @xmath347 is of the order of a couple of ev ( examples : @xmath348 ev ; @xmath349 ev ) . in the case of a radius @xmath350
of the metal core , the distance between neighbouring electronic levels is of the order of 100 mev .
this means that in the case we chose for our numerical investigations , there are many unoccupied levels below the energy of the lowest external electron . however , if we consider a system with an inner radius @xmath351 , the spacing of internally located electron levels of given @xmath303 approaches already 1 ev .
it is therefore conceivable that one might succeed to model a strictly stable , , mesoscopic atom . a further trick to widen the spectrum of internally localized electrons would be to replace the metallic sphere with radius @xmath2 by a metallic bubble of outer radius @xmath2 and a thickness of 1 to 2
in this case the levels differing by the number of radial nodes would have a much larger spacing .
this is perhaps possible , if one replaces the charged metal layer by a ( charged ) fullerene @xmath352 , which has a radius of about 3,5 .
the separation energy @xmath347 of an electron from graphite is known to be @xmath353 ev .
we surmise that the separation energy from the fullerene is about the same . of course
, it would no longer be justified to describe free electrons in the fullerene as moving in a simple square well perpendicular to the sphere @xmath6 .
but the spacing of electronic levels is expected to be in the order of ev .
finally , one is not limited to the spherical geometry .
we may also envisage to consider charged nanotubes consisting of a cylindrical surface crystal of @xmath354-atoms covered by an insulating surface layer .
the charge could in this case be generated by applying an external voltage to the ends of the nanotubes . as there is a lot of activity spent on nanotubes presently
, we hope that one could succeed to produce small macroscopic amounts of insulated charged nanotubes in a not too far future .
of course , the problem of the stability of externally localized electrons with respect to electromagnetic decay would be similar for nanotubes as for the case of the system we investigated .
so we hope that our investigation stimulates the interest of the experts in the field of mesoscopic physics .
* acknowledgement : * + klaus dietrich gratefully acknowledges the hospitality and support on the occasion of his stay at the theoretical physics institute of the umcs in april 2003 .
krzysztof pomorski is indebted to the physics department of the technical university of munich for granting of the honorarium of a visiting professor position in february 2002 .
let us consider the simplified potential @xmath355 we decompose this potential as follows @xmath356 @xmath357 \,\ , , \label{eqa3}\ ] ] @xmath358 and define the two hamiltonians @xmath359 where @xmath360 is the kinetic energy operator of radial motion @xmath361 the eigenstates @xmath362 , @xmath363 of the hamiltonians ( [ eqa5 ] )
@xmath364 @xmath365 form a non - orthogonal set of basis functions .
we consider an eigenstate @xmath366 of the hamiltonian @xmath367 with eigenenergy @xmath69 @xmath368 @xmath369 for the special case that the angular momentum @xmath303 is the same for all the states and that two specific eigenenergies @xmath370 and @xmath371 are very close to the energy @xmath69 in ( [ eqa10 ] ) .
then , we may expect that the expansion of @xmath366 in terms of the basis states @xmath362 and @xmath363 may be limited to the two states with the energies @xmath370 , @xmath371 very close to @xmath372 : @xmath373 in principle , we would have to sum over all eigenstates @xmath362 and @xmath363 on the righthand side of ( [ eqa11 ] ) , but of all the expansion coefficients @xmath374 , @xmath375 only the ones corresponding to the energies @xmath370 , @xmath371 closest to @xmath69 are expected to be large and all the others to be much smaller . henceforth , we leave away the subscript @xmath376 for simplifying the notation . it is convenient to introduce the dual basis states @xmath377 , @xmath378 defined by the orthogonality conditions @xmath379 where @xmath380 and @xmath381 can be 1 or 2 .
@xmath382 and @xmath383 can be obtained by diagonalizing the inverse of the overlap matrix @xmath384 it is easily seen that the schrdinger eq .
( [ eqa10 ] ) together with the assumption ( [ eqa11 ] ) leads to the two coupled linear equations @xmath385 where the matrix elements are defined as follows : @xmath386 introducing the abbreviations @xmath387 @xmath388 the solutions @xmath389 of the eigenvalue equation @xmath390 are given by @xmath391 writing the two states @xmath392 in the form @xmath393 where @xmath394 one obtains @xmath395 from the eq .
( [ eqa14 ] ) and the amplitudes @xmath396 from the normalisation of @xmath397 .
the physically interesting quantities are the probabilities @xmath398 and @xmath399 to find the system described by @xmath397 in the substate @xmath400 or @xmath401 .
these probabilities are given by the mean - values of the projection operators @xmath402 it is easily seen that the probabilities are given by the following equations @xmath403
@xmath404 @xmath405 @xmath406 where @xmath407 is the overlap of the two basis states @xmath408 and the quantities @xmath409 by @xmath410 where @xmath411 it is seen that for @xmath412 i.e. far away from the crossing of the energies @xmath413 and @xmath414 , one of the ratios @xmath409 is much larger than the other .
for @xmath415 one finds @xmath416 @xmath417 \label{eqa33}\ ] ] and for @xmath418 the reversed result @xmath419 @xmath420 this shows that a predominantly externally localized state turns into a predominantly internally localized one as the energies @xmath413 and @xmath414 cross as a function of @xmath2 or @xmath3 .
the coupled true states @xmath421 are thus seen to change their nature as a result of the crossing of the energies @xmath413 , @xmath414 of a class 1 and a class 2 state .
close to the crossing point the states @xmath421 are seen to be delocalized ,
containing comparable contributions of the two basis states @xmath422 .
w. d. knight , k. clemenger , w. a. de heer , w. a. saunders , m. y. chou , m. l. cohen : phys .
lett . * 52 * , ( 1984 ) 2141 ; w. d. knight , w. a. de heer , k. clemenger , w. a. saunders : solid state commun . *
53 * , ( 1985 ) 44 ; hellmut haberland ( ed ) : clusters of atoms and molecules " , vol . 1 and 2 , springer verlag , 1994
. s. frank , n. malinowski , f. tast , m. heinebrodt , i.m.l .
billas , t.p .
martin : z. phys . *
d40 * , ( 1997 ) 250 ; m. springborg , s. satpathy , n. malinowski , t.p .
martin , u. zimmermann : phys .
lett . * 77 * , ( 1996 ) 1127 ; h. kuzmany , j. fink , m. mehring , s. roth ( eds ) : electronic properties of fullerenes " , springer - verlag , 1993
. m. brack : rev .
phys . * 65 * , 677 ( 1993 ) .
w. a. de heer , rev .
phys . * 65 * , 611 ( 1993 ) .
m. abramowitz and j.a .
stegun ( eds ) : `` handbook of mathematical functions '' , nat . bureau of standards , applied mathematics , series 55 , issued june 1964 .
a. erdelyi , higher transcendental functions " , mcgraw - hill , 1953 .
jackson : `` classical electrodynamics '' , john wiley , 1967 .
bransden and c.j .
joachain , `` physics of atoms and molecules '' , longman group limited , 1983 .
m.e . rose , `` elementary theory of angular momentum '' , new york : dover , 1995 l.d .
landau , phys .
* 2 * ( 1932 ) 46 ; c. zener , proc .
( london ) * a137 * ( 1932 ) 696 .
h. haberland and k.h .
bowen : `` solvated electron clusters '' , chapt .
2.5 in `` cluster of atoms and molecules ii '' , vol .
56 of the springer series in chemical physics , 1994 . | we determine the wavefunctions of electrons bound to a positively charged mesoscopic metallic cluster covered by an insulating surface layer .
the radius of the metal core and the thickness of the insulating surface layer are of the order of a couple of ngstrm .
we study in particular the electromagnetic decay of externally located electrons into unoccupied internally located states which exhibits a resonance behaviour .
this resonance structure has the consequence that the lifetime of the `` mesoscopic atoms '' may vary by up to 6 orders of magnitude depending on the values of the parameters ( from sec to years ) . |
over the last several decades the study of interplay of noise and nonlinearity has become a major subject of research in various multidisciplinary areas@xcite .
these studies include the noise induced directed transport , stochastic resonance , noise induced stability of unstable states , noise induced phase transition , noise induced ordering etc . in most of the above mentioned cases the presence of noise is very much essential . in the literature ,
the noise induced directed transport in periodic extended structures in the absence of overall net bias has been extensively studied .
two basic ingredients are required for the posilibity of such directed transport , namely , the system should be driven out of equilibrium and there should be some asymmetry ( either temporal or spatial ) along with nonlinearity in the system .
the systems consisting of brownian particles and operating with the minimal conditions for such directed transport are known as brownian ratchets / brownian motors @xcite .
there is increasing interest from physicists , biologists and engineers on the study of so called ratchets or brownian motors @xcite .
based on the various ways of introducing the asymmetry or nonlinearity in the systems , there are various kinds of brownian ratchet models like flashing ratchets , rocking ratchets , time - asymmetric ratchets and inhomogeneous ratchets@xcite . the main motivation in the study of performance characteristics of brownian ratchet is the notion of efficiency of energy transduction from the thermal fluctuations @xcite . with the help of stochastic energetic formalism @xcite proposed by sekimoto , efficiency can be defined in a wide class of ratchet models @xcite .
different kinds of efficiency ( thermodynamic , stoke s and generalized ) for ratchet models have been studied in detail both analytically and numerically . since the brownian ratchet operates out of equilibrium , there is always an unavoidable heat transfer to the medium / environment , which makes it to be less efficient .
it has been noticed that brownian motors based on the principle of flashing ratchet models result in low efficiency . on the other hand , brownian motors based on the adiabatically changing potentials , e.g. , rocked ratchet models exhibit high value in efficiency @xcite . in this work ,
we are mainly interested in the case of spatial asymmetric rocked ratchet models . in all kinds of ratchet models ,
it is to be emphasized that the brownian particle moves in a periodic potential system and hence it ends up with the same potential energy even after crossing over to the adjacent potential minimum .
hence there is no extra energy stored in the particle which can be used for a given purpose . in order to extract work / energy out of it s motion ,
it is necessary to apply an external load force against which the particle moves and stores energy in the form of potential energy @xcite . in this context the efficiency of the ratchet can be defined as the output work against the load force per input energy given to the system .
this definition of efficiency is called as thermodynamic efficiency and it s value is zero when there is no applied load in the ratchet @xcite .
the ratchets are not designed to lift the load all the time . alternatively , there are various ways of defining the efficiency in the ratchet based on the purpose of the work given to the ratchet . in this work ,
we are mainly interested in the notion of stoke s efficiency , which is defined as the ratio of the rate of work done against the mean viscous drag and the rate of input energy given to the ratchet @xcite .
the brownian ratchets operate at microscopic length scale . for the systems operating at this length scale , the energy exchanged ( i.e. , in the form of mechanical energy , chemical energy etc . ) between the system and it s environment is of the order of thermal energy ( few @xmath2 , @xmath3 being the boltzmann constant and @xmath4 being the temperature ) . thermal fluctuations play a predominant role , thereby exhibiting distinctly different behavior from that of macro systems @xcite . in the present work ,
we are mainly interested in the study of fluctuating properties of stoke s efficiency in case of a spatial and temporal asymmetric rocked ratchet model .
recently the enhancement of stoke s efficiency by noise was studied in a microscale domain @xcite . in this study , we are mainly interested on the stochastic properties of stoke s efficiency and it s probability distribution from adiabtic to nonadiabatic limit . we show that in most of the parameter space , fluctuations dominate the mean values .
as we increase the frequency of the drive the mean stoke s efficiency and the average current decreases . and the external time asymmetric driving force @xmath5 as a function of @xmath6 . , width=288,height=240 ]
we consider the dynamics of an overdamped brownian particle in a periodic spatial asymmetric ratchet potential @xmath7 in the presence of an external time asymmetric driving force @xmath5 .
the motion is governed by the overdamped langevin equation @xcite @xmath8 where @xmath9 denotes the instantaneous position of the particle and @xmath10 being the friction coefficient .
@xmath11 is the randomly fluctuating gaussian thermal noise satisfying the properties , @xmath12 and @xmath13=0 .
the angular bracket @xmath14 denote the ensemble average over all the realizations of noise . a schematic diagram of the ratchet potential @xmath7 along with the external drive , @xmath5 is shown in fig.1 .
@xmath15 where @xmath16 is the height of the ratchet potential .
@xmath17 , is the spatial periodicity of the ratchet potential , which is set to unity .
@xmath18 represents the spatial asymmetric parameter which characterizes the asymmetry of the ratchet potential .
+ the temporal asymmetric driving force @xmath5 is given by @xmath19 here @xmath20 represents the temporal asymmetric parameter in the driving force and ranges from @xmath21 to @xmath22 .
@xmath23(@xmath24 denotes time period of the external driving force with @xmath25 being the frequency of the drive .
@xmath26 is an integer .
this model has been studied extensively in the adiabatic limit in @xcite . in this reference ,
the nature of the averaged current and average efficiency is studied in detail . in the adiabatic limit and at temperature , @xmath27 limit , the analytical expression for the time averaged current , @xmath28 is obtained in @xcite .
@xmath29 where @xmath30 and @xmath31 are the fraction of currents in the positive and negative direction over a time period of @xmath32 and @xmath33 respectively .
@xmath30 is the current when the external driving force , @xmath34 and @xmath31 is the current when the external driving force , @xmath35 . in our present work ,
we study the unidirectional current as well as the stoke s efficiency in the adiabatic as well as nonadiabatic regimes .
we show that in general the current decreases as we shift from adiabatic to nonadiabatic regime .
for the first time we have also studied the nature of the stochastic stoke s efficiency , @xmath1 and it s fluctuations .
we show that in general average stoke s efficiency , @xmath36 is not a good physical variable because fluctuations dominate the mean values .
large spatial asymmetry in the ratchet potential decrease the fluctuations in @xmath1 . to this end
we have calculated the probability distribution , @xmath37 of stochastic stoke s efficiency , @xmath1 .
recently stochastic effeciency in heat engines has been studied in great detail in ref .
the stoke s efficiency is related to the rate of work done against the average frictional force to the rate of input energy .
formally stochastic stoke s efficiency , @xmath1 at any instant of time can be defined as @xmath38 where @xmath39 is the instantaneous velocity and @xmath40 being the average velocity of the brownian particle .
@xmath41 is the rate of input energy .
@xmath42 is the rate of work done against the viscous force , @xmath43 . in this work
we are mainly interested in the nature of average stoke s efficiency , @xmath36 as well as on it s distribution @xmath37 .
stoke s efficiency @xmath1 , over a cycle @xmath23 is given by @xmath44 this is a stochastic quantity and for a given period of the drive .
it depends on a realization of @xmath39 over the interval .
averaged stoke s efficiency is now defined as @xmath45
we have simulated the overdamped langevin dynamics using the algorithim of huens method @xcite . from the solution of langevin equation , we evaluate the averaged current , input energy and stoke s efficiency . the averaged current is given by @xmath46 where @xmath47 represents the instantaneous position of the particle at time @xmath6 and @xmath48 being the position of the particles at time @xmath49 .
we have run the simulation for @xmath50 , @xmath23 being one period of the drive .
the steady state averages were taken just after ignoring the initial transients upto @xmath51 .
each time step is taken to be @xmath52 .
this particular ratchet model has been studied theoretically in ref .
@xcite in the adiabatic limit .
the nature of average stoke s efficiency , @xmath36 in the adiabatic regime has been studied in detail . in the present work ,
we are mainly interested on the stochastic properties of stoke s efficiency , @xmath1 and hence on its probability distribution , @xmath37 in the adiabatic limit as well as for different frequency of driving , @xmath25 . in this particular ratchet model ,
when the noise strength is very low , i.e. , in the deterministic limit there exists potential barriers for the motion of the brownian particle in both forward as well as backward directions .
the potential barrier in the forward direction disappears only when the driving force in the forward direction is larger than the force exerted by the barrier towards the right of the potential , i.e. , when @xmath53 .
similarly the barrier in the backward direction disappears only when @xmath54
. hence there are critical forces , @xmath55 and @xmath56 beyond which the barrier disappears in both forward as well as backward direction .
it implies that a finite current exists in the ratchet only when the amplitude of driving , @xmath57 is in between @xmath58 and @xmath59 .
when barrier disappears in the backward direction , there is a finite current in the backward direction resulting a negative contribution in the net current , @xmath28 in the ratchet . in the deterministic limit ,
the averaged current , @xmath28 shows a peaking behavior as a function of amplitude of driving force , @xmath57 . the current increases from @xmath60 and shows a maximum at around @xmath61 and then decreases with @xmath57 .
we have simulated the averaged current , @xmath28 and the average stoke s efficiency , @xmath36 as a function of @xmath57 and plotted in fig.2 and fig.3 respectively for different values of temporal asymmetry parameter , @xmath20 for a symmetric potential and for temperature @xmath62 ( close to the deterministic limit ) .
the inset of both the figures shows the same plot with larger temperature @xmath63 .
these results are for @xmath64 corresponding to the adiabatic limit . in this limit
the forcing @xmath5 is assumed to change very slowly .
it s frequency is smaller than any other frequency related to the relaxation rate in the problem , i.e. , the system remains in the steady state at each instant of time . from fig.2 and fig.3
, we notice that our simulation is in perfect agreement with the analytical obtained in @xcite .
the stoke s efficiency follows the nature of average current as expected . in our work we have scaled all the lengths with respect to @xmath65 and energies with respect to the height of the ratchet potential , @xmath16 .
all physical quantities are in the dimensionless form .
we have set @xmath66 . from the simulation results obtained in fig.2 , we have calculated the values of @xmath58 for three different values of @xmath20 , i.e. , @xmath67 for @xmath68 , @xmath69 for @xmath70 and @xmath71 for @xmath72 respectively in the deterministic limit ( @xmath62 ) .
@xmath73 remains same for all @xmath20 values since @xmath56 is independent of @xmath20 .
these values are in exact agreement with analytically obtained values in @xcite . as a function of amplitude of driving , @xmath57 for a symmetric potential @xmath74 for different values of @xmath20 .
all other fixed parameters are temperature @xmath62 , frequency of driving @xmath64 .
the inset shows the same plot for temp @xmath63.,width=307,height=240 ] as a function of amplitude of driving , @xmath57 for a symmetric potential @xmath74 for different values of @xmath20 .
all other fixed parameters are temperature @xmath62 , frequency of driving @xmath64 .
the inset shows the same plot for temp @xmath63 .
, width=307,height=240 ] fig.4(a ) shows the average current , @xmath28 as a function of amplitude of driving , @xmath57 . as anticipated earlier in the deterministic limit ,
current exhibits a peaking behaviour at @xmath75 . the average stoke s efficiency as a function of @xmath57 follows the nature of currents shown in fig.4(b ) .
in fig.4(c ) , we have plotted the relative variance ( fluctuations ) of stochastic stoke s efficiency , @xmath76 as a function of @xmath57 .
@xmath77 shows a decreasing behaviour with @xmath57 and shows a minimum .
however , the minimum is larger than @xmath22 , indicating that the fluctuations , @xmath78 in the stoke s efficiency always dominate their mean values , @xmath36 .
hence the average stoke s efficiency is not a good physical variable in this parameter regime .
the distribution of stochastic stoke s efficiency is studied in fig.4(d ) for various values of amplitude of driving , @xmath57 . for lower @xmath57 values ( @xmath79 )
, the distribution shows a single peak .
this is because of the presence of potential barrier in both forward as well as backward directions . as a result
the particle makes random movement across any one minimum of the ratchet potential . with further increase in @xmath80 values ,
the barriers in the forward direction disappears and the probability of moving the particle towards right increases .
the distribution exhibits multipeaked structure and the mean value shifts towards right indicating that the average stoke s efficiency , @xmath36 increases with @xmath57 .
the multipeaked structure in the distribution is related to the distribution of local average velocities ( which we have checked separately in fig.5 ) and is due to the presence of barriers .
the distribution for negative values of @xmath1 has a finite weight .
this results from atypical realizations of input energy ( fig.6 ) which show negative tails @xcite . for larger @xmath57 , the distribution becomes narrow and the variance decreases . at the same time the mean value shifts towards left . in this regime , the barriers in both forward as well as backward directions disappear and the particle follows straight paths in both the directions . as a function of @xmath57 in ( a ) , @xmath36 as a function of @xmath57 in ( b ) ,
relative fluctuation @xmath77 as a function of f in ( c ) and the probability distribution , @xmath37 for different @xmath57 values in ( d ) .
the other fixed parameters are @xmath64 , @xmath74 , @xmath70 and @xmath63 respectively.,width=384,height=336 ] , @xmath81 for different @xmath57 values .
the other fixed parameters are @xmath64 , @xmath74 , @xmath70 and @xmath63 respectively.,width=336,height=288 ] next in fig.7 , we have studied the average current , @xmath28 , average stoke s efficiency , @xmath36 , relative fluctuations , @xmath77 in @xmath1 and the probability distribution , @xmath37 with temperature for a symmetric potential in ( a ) , ( b ) , ( c ) and ( d ) respectively .
we have fixed the other parameters as @xmath82 and @xmath64 . in this parameter regime , in order to have finite current in the ratchet , the minimum value of @xmath20 should be greater than @xmath83 .
hence we have kept fixed the value of @xmath20 as @xmath84 . here
also we observe that the average current , @xmath28 and average stoke s efficiency @xmath36 show peaking behaviour with temperature , @xmath4 .
the stoke s efficiency @xmath36 follows the same behaviour of average current @xmath28 .
however , from fig.7(a ) and 7(b ) , we notice that there exists a finite current and finite stoke s efficiency even at zero temperature .
this is because of the absence of potential barrier in the forward direction even at zero temperature . in this fixed parameter regime , the crictical value of @xmath57 , @xmath85 and hence @xmath86 .
therefore the barrier disappears in the forward direction , as a result we get a finite current and finite stoke s efficiency even at @xmath27 limit .
the relative fluctuations in stoke s efficiency @xmath77 with temperature is studied in fig.7(c ) .
@xmath77 shows a minimum in the limit @xmath87 limit and then increases with increase in temperature , @xmath4 .
it is always larger than @xmath22 for all values of @xmath4 .
this indicates that fluctuation in stoke s efficiency , @xmath88 always dominate the mean value , @xmath36 confirming that @xmath1 is not a self averaging quantity in this parameter regime .
the nature of the distribution function , @xmath37 for various temperature values is studied in fig.7(d ) . for lower temperature values ,
i.e. , close to the deterministic limit , the distribution shows a single peaked structure with few minor peaks . with further increase in temperature , the probability of moving the particle to the nearest neighbouring potential minima in the forward direction increases , resulting a finite positive contribution towards the net current , @xmath28 and towards the average stoke s efficiency , @xmath36 . for larger temperature , the barriers disappear in both the directions . as a result
the particle spreads over in both the direction .
this results decrease in the net current , @xmath28 and average stoke s efficiency , @xmath36 . at the same time
the distribution becomes narrow and shifts towards left .
, @xmath89 for different @xmath57 values .
the other fixed parameters are @xmath64 , @xmath74 , @xmath70 and @xmath63 respectively.,width=336,height=288 ] as a function of @xmath4 in ( a ) , @xmath36 as a function of @xmath4 in ( b ) , relative fluctuation @xmath77 as a function of @xmath4 in ( c ) and the probability distribution , @xmath37 for different @xmath4 values in ( d ) .
the other fixed parameters are @xmath64 , @xmath74 , @xmath70 and @xmath82 respectively.,width=384,height=336 ] the average current , @xmath28 , average stoke s efficiency , @xmath36 , the relative fluctuations in @xmath1 , @xmath77 and the distribution function @xmath37 with spatial asymmetry , @xmath90 is studied in fig.8(a ) , ( b ) , ( c ) and ( d ) respectively .
the average current , @xmath28 ( fig.8(a ) ) and average stoke s efficiency , @xmath36 ( fig.8(b ) ) increase with increase in spatial asymmetry , @xmath90 .
this is due to the fact that with increase in asymmetry of the ratchet potential , the barriers in the forward direction starts decreasing and the probability of moving the particle towards the next nearest neighbouring potential minima in the forward direction increases .
this results in a finite positive contribution to the net current in the ratchet and hence to the average stoke s efficiency , @xmath36 .
therefore , in this given parameter regime , for a spatial asymmetry ratchet potential , the magnitude of average current is quite larger than that in case of a symmetric ratchet potential and hence the spatial asymmetry in the ratchet helps in enhancing the stoke s efficiency .
however , the relative fluctuations , @xmath77 ( fig .
8(c ) ) shows a decreasing behaviour with increase in spatial asymmetry , @xmath90 , and become less than @xmath22 for larger spatial asymmetry in the potential .
this result reflects that for a larger spatial asymmetry in the ratchet potential , the fluctuations in stoke s efficiency do not dominate the mean values . as a result @xmath36 behaves as a good physical variable in this parameter regime and a self - averaging physical quantity .
the probability distribution , @xmath37 is studied in fig .
8(d ) for various @xmath90 values . for a symmetric ratchet potential , i.e. , for @xmath74 , the distribution shows a single peaked structure with few minor peaks in it .
the inclusion of spatial asymmetry in the potential decreases the barrier height towards the right of the potential . as a result
the local velocities in the forward direction increases and the minor peaks in the distribution disappears .
therefore , with increase in @xmath90 values , the spreading in the distribution decreases at the same time the distribution with mean value shifts towards right indicating that the @xmath36 increases with @xmath90 . as a function of @xmath90 in ( a ) , @xmath36 as a function of @xmath90 in ( b ) ,
relative fluctuation @xmath77 as a function of @xmath90 in ( c ) and the probability distribution , @xmath37 for different @xmath90 values in ( d ) .
the other fixed parameters are @xmath64 , @xmath82 , @xmath70 and @xmath63 respectively.,width=384,height=336 ] finally we have studied the average current , @xmath28 , average stoke s efficiency , @xmath36 , relative fluctuations , @xmath77 and the probability distribution , @xmath37 with the frequency of driving , @xmath25 in fig.9(a ) , ( b ) , ( c ) and ( d ) respectively .
the average current , @xmath28 ( fig.9(a ) ) and the average stoke s efficiency , @xmath36 ( fig.9(b ) ) shows a decreasing behaviour with frequency of drive , @xmath25 .
this is because with increase in frequency of drive , the forcing amplitude varies so fast that , the particle is not able to take the advantage of force in either of the direction and the movement of the particle towards the nearest neighbouring potential minima is suppressed .
this decreases the net current as well as the average stoke s efficiency .
for very high frequency of driving , the forcing amplitude varies so fast that the particle does nt feel the influence of forcing .
the particle as if stays at one position only and hence the net current , @xmath28 and the average stoke s efficiency approaches to zero .
however , the relative fluctuations , @xmath77 ( fig.9(c ) ) increases with increase in the frequency of driving , @xmath25 and is always larger than @xmath22 . the probability distribution function , @xmath37 with frequency of driving , @xmath25 is shown in fig.9(d ) .
for low frequency of driving , i.e. , in the adiabatic limit , the distribution shows a maximum and with increase in frequency , @xmath25 , the distribution spreads over towards left as well as the mean value shifts towards left . with further increase in @xmath25 , a finite peak developes across zero with few minor peaks towards the right side of the distribution . at sufficiently high value of @xmath25 ( i.e. , in nonidiabatic regime ) , we get a sharp peak across zero and the minor peaks in the distribution get suppressed .
this clearly confirms that there is no effective transport in the parameter regime with high frequency of driving . as a function of @xmath25 in ( a ) , @xmath36 as a function of @xmath25 in ( b ) ,
relative fluctuation @xmath77 as a function of @xmath25 in ( c ) and the probability distribution , @xmath37 for different @xmath25 values in ( d ) .
the other fixed parameters are @xmath74 , @xmath82 , @xmath70 and @xmath63 respectively.,width=384,height=336 ]
in conclusion , we have studied the stoke s efficiency and it s stochastic properties in the case of a temporal asymmetric rocked ratchet potential using overdamped langevin dynamics simulation .
our main results indicate that over larger parameter space , stochastic stoke s efficiency behaves as a non - self averaging physical quantity .
the mean current and stochastic efficiency monotonically decrease as we go from adiabatic to the non - adiabatic regime .
all the above mentioned quantities have been studied for different values of forcing amplitude , temperature , asymmetry parameter and driving frequency .
we are presently investigating the nature of stochastic thermodynamic efficiency in these ratchet systems .
ms acknowledges the funding from the inspire faculty award by the department of science and technology , govt . of india and thanks
iop , bhubaneswar for the hospitality where a part of the work was done .
amj thanks dst , india for financial support .
0 d.dan , m. c. mahato and a. m jayannavar , int .
b * 14 * , 1585 ( 2000 ) ; physica a , * 296 * , 375 ( 2001 ) ; ; phys .
rev e * 63 * , 56307 ( 2001 ) , d. dan and a. m. jayannavar , phys . rev . *
e66 * , 41106 ( 2002 ) . | we study the stoke s efficiency and it s fluctuating properties in the case of a spatial asymmetric ratchet potential with a temporal asymmetric driving force from adiabatic to nonadiabatic regime .
our numerical investigations show that the average stoke s efficiency and the average current decrease with the frequency of driving .
for low frequency of driving , i.e. , in the case of an adiabatic regime , we reproduced the analytical results supporting our numerical simulations . by evaluating the probability distribution , @xmath0 for stoke s efficiency , @xmath1 we focus on the stochastic properties of stokes efficiency .
we find that in most of the parameter space , fluctuations in @xmath1 are comparable to or larger than the mean values . in such a situation
one has to study the full probability distribution of @xmath1 .
with increase in frequency of driving , the distribution becomes multipeaked . at the same time the average stoke s efficiency decreases with increase in frequency of drive . for high frequency of driving ,
the distribution develops a peak across zero .
further increase in frequency this peak gets sharper . and finally at sufficiently high frequency we get a strong peak across zero indicating that there is no effective transport in this regime .
.5 cm _ keywords_:ratchets , efficiency , fluctuations |
there are two very important issues in the loop approach to non - perturbative canonical quantum gravity@xcite , which we have to address in our quest for a complete theory .
the first is the problem of evolution of the gravitational quantum states and the second one is the recovery of the classical continuous space from the discrete structures which appear at very short scale .
the problem of evolution in canonical quantum gravity is quite non - trivial .
the underlying general covariance principle of the theory leads not to a true hamiltonian , but to a sum of constraints .
the physical states of the theory are then the ones which are annihilated by these constraints . imposing the gauge and the diffeomorphism constraints has led to a hilbert space
whose basis is given by spin networks space .
the hamiltonian constraint , which generates infinitesimal translations in the time direction , has to be imposed in this space of states in order to obtain the physical space of states .
there are different approaches which have been proposed in order to study the action of the hamiltonian constraint . in the canonical approach ,
one relies on regularizing the hamiltonian constraint as a quantum operator and investigating its action on the states of the theory .
various regularization schemes have been suggested@xcite , the most commonly accepted being the one proposed by thiemann@xcite .
however , all of these suffer from various problems ( see for example @xcite ) .
another approach which has been suggested is to look at the finite action of the hamiltonian constraint as opposed to the infinitesimal one@xcite .
this has led to the study of spin foams .
an alternative approach is to find a consistent set of rules for quantum evolution which , in the classical limit , recover the action of the hamiltonian constraint@xcite .
it is this approach that we study in this paper . while the idea is very intriguing , it is highly non - trivial to define an appropriate set of rules which lead to a good classical limit .
the problem of finding the appropriate rules for evolution of the kinematical states is now related to the problem of defining the classical limit of the theory . on the other hand
the investigation of the classical correspondence is also quite subtle .
one of the main results in canonical quantum gravity is that geometric operators such as area and volume have discrete spectra@xcite .
this means that space has a discrete structure at the planck scale .
thus , in taking the classical limit , we should also recover the continuum space - time we see around us . according to a proposal in @xcite , the problem of recovering
the continuum space is similar to some problems of non - equilibrium critical phenomena in statistical physics .
the long range behavior characteristic of classical general relativity could arise at the critical values of some parameters of the discrete theory .
this is already seen in certain discrete approaches to quantum gravity such as dynamical triangulations .
the important aspect of the proposal is that , since there is no external agent to tune the parameters of the universe , discrete space should be a self - organizing critical system .
the idea of self - organized criticality was introduced by bak , tang and wiesenfeld in @xcite ( for reviews see @xcite ) .
the self - organized criticality paradigm is that certain systems , consisting of many parts interacting via local rules , self - organize themselves in a critical state .
usually this critical state is characterized by the lack of specific scale of propagation of perturbations . in this paper , we study various models whose dynamics is suggested by canonical quantum gravity .
there are certain aspects of the action of the hamiltonian constraint , which are common to all of the proposed regularizations .
these aspects guide us in the introduction of rules for graphical evolution of the spin networks . with the evolution rules in hand ,
we develop a model along the lines of the sandpile models of self - organized criticality .
first we briefly introduce the spin network states . in section 3
, we discuss the various regularizations of the action of the hamiltonian constraint and the corresponding graphical rules for evolution .
we then introduce our model and discuss our results .
the basic variables in canonical quantum gravity can be taken to be an su(2 ) connection @xmath0 and a densitized triad @xmath1 which takes values in the dual of @xmath2 . a convenient basis for the space of gauge invariant states is given by the spin network states , introduced in canonical quantum gravity by rovelli and smolin@xcite and by baez@xcite .
a spin network state @xmath3 is defined by the following elements : * a closed graph @xmath4 , embedded in the spatial manifold @xmath5 , consisting of finite number of oriented edges @xmath6 incident at vertices @xmath7 .
a vertex is called @xmath8-valent or of valence @xmath8 if there are @xmath8 edges incident at that vertex . *
each edge @xmath9 of the graph @xmath4 is labeled by an irreducible representation @xmath10 of su(2 ) .
further on , @xmath11 is referred to as the color of the edge .
* at each vertex @xmath12 there is an intertwining operator from the tensor product of the representations carried by the incoming edges to the tensor product of representations labeling the outgoing edges .
when considering only the combinatorial aspects of a spin network state it is very convenient to think of the spin network on which the state is based as the dual 1-skeleton of a simplicial triangulation of the spatial manifold @xmath5 . in this picture ,
a spin network is equivalent to a colored triangulation of the spatial manifold . in the rest of this paper
, we shall consider a spin network or its dual triangulation as equivalent objects .
to we shall now schematically describe the graphical action of the hamiltonian constraint of canonical quantum gravity .
essentially the action of the hamiltonian constraint can be thought of as adding a loop to the original spin network with some extra requirements .
first , the added loop should coincide with two of the edges of the original spin network joined at one vertex in the vicinity of that vertex .
second , the added loop can coincide with any other ( portions of ) edges . whenever the added loop runs along already existing edge , the recoupling theory of angular momentum determines the way the original spins change . according to
the regularization described in @xcite , the hamiltonian acts only in a vicinity of the vertices of the spin network . for simplicity
let us consider a trivalent vertex .
at such a vertex the action of the hamiltonian constraint can be thought of as a sum of three terms , each one corresponding to the old spin network with an added new edge of color one connecting two of the original edges ( see figure [ ham1 ] ) .
the colors of the original edges in their portion between the newly created vertices and the original vertex are either increased or decreased by one .
consecutive action of the hamiltonian constraint keeps adding new edges only to the original ones , thus producing a web - like structure at each vertex .
the problem with such a definition of the action of the hamiltonian constraint is that it is ultra - local in the sense that the action of the constraint at one vertex does not propagate any information to the neighboring vertices of the spin network .
hence , there are no long range correlations .
a modification which was aimed at fixing this problem was introduced in @xcite .
this new version of the constraint action allows not only the addition of new edges but also the addition of whole loops coinciding completely with the already present edges . in such cases
the original graph does not change at all but the edges of coincidence have their colors changed .
this is shown in figure [ ham2 ] . to in our work
, we model the action of the hamiltonian constraint by introducing a generalization of this second type of regularization .
we assume that loops can be added not only to a single closed cycle of edges , but also can run along any collection of the original edges . also , a very important assumption is that the added loop can run many times along one and the same edge .
since the effect of adding a loop segment to an existing edge simply changes the color of that edge , we can move away from the loop picture and consider simply gauge invariant changes of the colors of the edges of the original spin network .
for simplicity , we study only planar trivalent spin network states in a 2-dimensional space with boundary ( topology of a disk ) . it is convenient to take this space to be a triangle .
the spin networks associated with that triangle are based on the graph , dual to the triangulation of the triangle .
the triangulation is performed by subdividing the sides of the triangle into @xmath13 parts .
the triangulated triangle is shown in figure [ space ] .
the dual graph on which the spin network states are based is obtained by assigning a vertex to each 2-cell in the triangulation and an edge crossing each 1-cell and connecting the vertices in the neighboring 2-cells ( shown using dashed lines in figure [ space ] ) .
the colors carried by the edges of the spin network can be thought of as being lengths of the corresponding crossed 1-cells of the triangulation@xcite .
more precisely , the length of a side of the triangulation which is intersected by a spin network edge of color @xmath14 is @xmath15 where @xmath16 is the planck length . ] .
thus , we shall refer to the triangulation and lengths or its dual spin network and the corresponding colors interchangeably .
to the requirement for gauge invariance at the vertices of the spin network state is essentially a triangle inequality that the lengths of the sides of a triangle should satisfy together with the requirement that the sum of these lengths be an even number .
we call the triangle inequality together with the requirement for an even sum of the site s lengths a gauge invariance condition .
if @xmath17 , @xmath18 and @xmath14 are the colors of the three edges at a node , the gauge invariance requires that : @xmath19 further , we will assume that the colors of the edges of the spin network crossing the boundary are unconstrained by the evolution process .
as an initial state for our system , we construct a spin network based on the dual graph and assign random colors to the edges of the spin network in a gauge invariant fashion .
the evolution of spin networks in our model corresponds to changing a set of colors such that the gauge invariance condition remains satisfied at each node of the spin network .
we use the following procedure to implement this evolution .
choose , at random , one of the edges of the spin network and change its color by an amount @xmath20 .
this will , in general , affect the gauge invariance of the two nodes on which the edge is incident . to restore the gauge invariance at the nodes , which are no longer gauge invariant
, we change the color of one of the remaining incident edges .
we continue this process until gauge invariance is restored at every node of the graph . depending upon the value of @xmath20 , there are various possibilities for the exact implementation of the described procedure . for an odd @xmath20 ,
say @xmath21 , the gauge invariance condition is always violated at the two nodes on which the initial edge is incident , because of ( [ even ] )
. the restoration of the gauge invariance of the first pair of nodes will necessary induce violation of the gauge condition at two new nodes .
there are two cases for this pattern to end either at a certain step a non - invariant node contains an edge which crosses the boundary of the triangle or the spin changes meet somewhere on the spin network and neutralize each other .
these two cases can be viewed from a loop perspective as corresponding to
( i ) adding an open line of color 1 , whose both ends lie on the boundary or ( ii ) adding a closed loop of color 1 .
to the case of even @xmath20 , say @xmath22 , is more involved .
it is possible that the change of the color of a particular edge by @xmath23 leaves both the original nodes gauge invariant .
actually this is true in most of the cases requirement ( [ even ] ) is not violated and it is up to the conditions ( [ gge ] ) to determine the gauge invariance
. we will mostly be interested in this latter case .
let us now discuss the case @xmath24 in more detail .
let us choose an edge of the graph at random and change the length of this edge by @xmath25 .
the only case in which this can violate gauge invariance is if the triangle was `` flat '' initially , that is , if one of the three inequalities in equation ( [ gge ] ) is saturated .
we call such triangles ( vertices ) critical .
if , for such a critical triangle , we increase the longest side or decrease one of the two shorter sides , we will violate gauge invariance . to compensate for this
, we have to adjust the length of one of the other two sides .
the adjusted side also belongs to one of triangle s neighbors .
if this change violates the gauge invariance condition in the neighboring triangle , we adjust the colors there as well .
this process will continue until all the triangles in the triangulation satisfy gauge invariance again .
thus , one change of @xmath25 can lead to a series of spin changes ( see figure [ aval ] for an example ) .
these spin changes are analogous to the topplings in the sandpile models .
drawing further analogy with the sandpile models , we refer to the series of spin changes required to restore gauge invariance after the initial change as an avalanche .
since we allow arbitrary spins to start with , the number of critical triangles will be negligible in comparison to the total number of triangles .
thus if we change the spins in an arbitrary way almost no avalanches will occur . in the sandpile model
the analogous situation is when we add and remove grains of sand with equal probabilities .
hence , we introduce a probability @xmath8 for flattening " of a triangle .
if the arbitrary chosen side has the biggest length in the triangle , we increase the length by 2 with a probability @xmath8 and decrease it with a probability @xmath26 .
similarly if the arbitrary chose side is the smallest one or the one in the middle , we decrease the length by 2 with probability @xmath8 and increase it with probability @xmath26 .
thus it is clear that for @xmath8 close to 1 , we will be pushing the triangles in the triangulation towards criticality . the rest of the avalanche ( if any ) proceeds a deterministic way as described above . as a further modification of the above update rule
, we introduce the probabilistic update into the entire avalanche . then , there is a probability @xmath8 that a gauge non - invariant node is restored to gauge invariance and probability @xmath26 for it to go further away from gauge invariance .
we performed simulations for system sizes ranging from 400 to 10,000 triangles with a typical run covering a million iterations of the system .
to to start with , we simulated the system with the deterministic evolution rule for system size of 400 triangles .
the results of these simulations are shown in figure [ t20 - 2 ] .
as can be seen from graph ( a ) in the figure , the fraction of critical triangles in the system rises very rapidly from an initial value of 0.24 to reach an equilibrium value of around @xmath27 .
the average color per edge in the system increases linearly as the simulation proceeds .
the distribution of avalanches @xmath28 as function of avalanche size , @xmath29 is shown in graph ( b ) of figure [ t20 - 2 ] .
this distribution can be described by an exponential decay @xmath30 to a good accuracy .
the results for a larger system size ( 10,000 triangles ) are shown in figure [ t100 - 2 ] .
the fraction of critical triangles reaches its equilibrium value more slowly as compared to the smaller system .
the equilibrium value of @xmath31 is similar to that for 400 triangles .
as expected , the average color per edge increases at a slower rate as well . the avalanche distribution can again be described well by a decaying exponential .
thus , our results scale with the system size . to the simulations for the case where we introduce the probability @xmath8 into the first move were performed next . depending upon the value of @xmath8 ,
a certain fraction of the triangles in the system become critical .
once reached , the fraction of critical triangles remained almost constant .
the avalanche distribution was again described by an exponential decay with the decay constant decreasing with decreasing probability .
the data remained qualitatively similar to figures [ t20 - 2 ] and [ t100 - 2 ] . finally , we simulated the case where the probabilistic update was also included in the avalanche phase of the dynamics .
the frequency distribution of the avalanches was still described by an exponential decay .
the decay constant however did not decrease as we decreased the probability from 1 .
it initially increases till @xmath32 and then decreases again ( see figure [ prob ] ) .
to figure [ prob ] seems to indicate that our system undergoes a phase transition around @xmath32 .
however , if we look at the avalanche distribution more carefully , we find that while we get some extremely large avalanches , the number of sites involved in these avalanches is very small .
for example , the largest avalanche that we got for the 400 triangles case consisted of 819 topplings which affected only 12 edges .
this implies that while we seem to have a phase transition , it does not lead to long range correlations .
we developed a model of evolving spin networks in which we keep the underlying graph fixed and consider changes of the colors of the edges . in the case
when the colors are changed by @xmath33 , these changes involve either sets of edges which form closed cycles or run between the boundaries of the space . when the colors are changed by @xmath24 , the response of the system is more involved .
our aim was to prescribe a set of rules for the spin changes and then check if such spin changes would propagate in a scale free behavior , characteristic for a soc system .
the answer to this question to a very large extent depends on the specific features of the model we proposed .
it turned out that with the rules we proposed , the system does not exhibit soc behavior .
instead we found that for particular values of the parameters in the system we got a phase transition . at this point , it can not be decided using a different set of rules will lead to the kind of self - organized behaviour that we were seeking .
there are various possibilities for modification of the evolution rules which we are exploring currently .
these include : * mixed color updates for each vertex , at which the gauge condition is violated by @xmath34 introduce with certain probabilities gauge invariance restoration either by update of the color of one of the edges by @xmath34 or of two of the edges by @xmath33 .
* conserved total color after adding a color of 2 to a particular edge , we continue by redistributing the added color among the adjacent edges , thus preserving the total color of the spin network .
* one directional propagation starting from an edge , one of the nodes of which remains gauge invariant after the spin change .
this will avoid the problem of order of overlapping present in the case when the spin changes propagate in two directions . * spin changes starting at the boundaries and propagating into the system .
further work is needed to show if an appropriate set of evolution rules can result in the definition of the spin network states as a self - organizing critical system .
we would like to thank lee smolin for the helpful discussions , which led us to this paper and for his important comments .
we are also grateful to fotini markopoulou and maya paczuski for the useful comments and suggestions .
this work was supported in part by nsf grant phy-9514240 to the pennsylvania state university . | one of the main results in canonical quantum gravity is the introduction of spin network states as a basis on the space of kinematical states .
to arrive at the physical state space of the theory though we need to understand the dynamics of the quantum gravitational states . to this aim
we study a model in which we allow for the spins , labeling the edges of spin networks , to change according to simple rules .
the gauge invariance of the theory , restricting the possible values for the spins , induces propagating modes of spin changes .
we investigate these modes under various assumptions about the parameters of the model . |
in a recent high accuracy experimental study of electric quadrupole and magnetic dipole moments of a sequence of odd - mass cd isotopes @xcite it was noted that the electric quadrupole moments of the @xmath0 states in the cd isotopes vary linearly with neutron number and are very simply and very well described in terms of a @xmath1 orbital sequentially occupied in the seniority scheme with a large neutron effective charge @xmath2 . if this is so , the same simple physics should be manifest in the other presumed single quasi - particle states in cd and , a - fortiori , in the the sn isotopes which have a closed proton shell .
the idea that the single neutron occupation probabilities can be so simply extracted from electromagnetic properties is sufficiently attractive to warrant a more detailed second look .
the aim of this paper is to show that this can be done using a minimalist theoretical framework .
ultimately of course , large scale theoretical calculations , shell model or beyond - mean - field model , are required to describe the structure of these nuclei , including their electromagnetic properties .
the aim of this paper is much more modest , but the picture is nevertheless useful for its simplicty .
as theoretical framework , i use the spherical bcs model with a weak - coupling approximation to include quadrupole collectivity . in this model , @xcite the states in the odd - even isotopes with dominantly single quasi - particle parentage can be written as : @xmath3 where @xmath4 and @xmath5 with the usual relation between the quasi - particle operators @xmath6 and the particle operators @xmath7 @xmath8 the bcs ground state is @xmath9 , and @xmath10 represents the first ( collective ) @xmath11 state which can usually be interpreted as a coherent superposition of two quasi - particle states .
the coefficient @xmath12 is given in first - order perturbation theory by @xmath13 i take the interaction potential @xmath14 to be a quadrupole - quadrupole force @xcite @xmath15 where p , n refer to protons and neutrons respectively and the one - body quadrupole operator is @xmath16 for a neutron quasi - particle , the expression for @xmath12 then becomes @xmath17
to first order in @xmath12 , the electric quadrupole moment of the weak coupling state in equation [ qp_state ] is @xmath18 \nonumber \\ & = & \sqrt{\frac{16\pi}{5 } } ( 2,0,j , j \mid j , j)(u^2_j - v^2_j)\frac{1}{\sqrt{2j+1 } } \langle \psi_j \mid\mid r^2 y_2 \mid\mid \psi_j \rangle \nonumber \\ & & \left [ e_n + \frac{2}{5 } \frac{1}{e_2-e_0 } \langle 2^{+ } \mid\mid e_n q(n ) + e_p q(p ) \mid\mid 0 \rangle \langle 2^{+ } \mid\mid \chi_{nn } q(n ) + \chi_{pn } q(p ) \mid\mid 0 \rangle \right ] \end{aligned}\ ] ] here @xmath19 and @xmath20 are the neutron and proton effective charges . in shell model calculations ,
their values are roughly @xmath21 and @xmath22 @xcite@xcite , attributed mainly to the giant quadrupole resonance that is not explicitly included in the usual shell model .
there are different ways of estimating the strength of the effective quadrupole - quadrupole force , however @xmath23 is always substantially smaller than @xmath24 .
since a ratio @xmath25 is obtained from hartree - fock calculations with a skyrme interaction @xcite,@xcite , i feel justified in making the approximation that @xmath26 for a neutron quasi - particle state , the quadrupole moment can then be written as @xmath27 where the last equation defines the total effective charge @xmath28 and the shell model quadrupole moment is @xmath29 note that the spectroscopic quadrupole moment normally reported in experiments is @xmath30 .
provided that the parameter @xmath31 can be estimated , equ .
[ q_moment ] can be used to simply extract the neutron occupation probabilities from experimental data on quadrupole moments and e2 transition probabilities .
the spectroscopic quadrupole moments and @xmath32 values for the sn and cd isotopes have been measured over a range of masses between the magic neutron numbers @xmath33 and @xmath34 .
table 1 summarizes the experimental excitation energy and @xmath35 value of the @xmath36 states in the even - even isotopes and table 2 lists the measured spectroscopic quadrupole moments of the of the odd - neutron isotopes .
the radial integrals @xmath37 are calculated using the shell model potential of reference @xcite .
the strength of the quadrupole - quadrupole interaction is taken to be constant @xmath38 .
this is chosen to yield reasonable values of @xmath39 for the @xmath40 states in the lightest sn and cd isotopes where the @xmath41 is expected to be mostly empty and near @xmath42 where it should be filled .
the factor @xmath43 is taken to be the average of the experimental values in the even neutron isotopes above and below the odd neutron isotope being studied .
when both are not available , then the nearest @xmath32 value and the average energy denominator is used .
the extracted occupation functions @xmath39 are given in table 2 and shown in figs.1 and 2 . the uncertainty in @xmath39
is estimated from the uncertainties in the experimental @xmath32 and @xmath44 and does not include the uncertainty in @xmath31 , nor the uncertainty in @xmath19 .
.[tab:1/tc ] experimental @xmath45 of a sequence of cd and sn isotopes .
[ cols= " < , < , < , > , < , < , < , < , > , < " , ] ( color online ) occupation function @xmath46 for the cd isotopes for @xmath47 . ] ( color online ) occupation function @xmath46 for the sn isotopes for @xmath48 . ]
the extracted occupation probabilities are consistent with expectations with the exception of @xmath49 .
since the @xmath1 orbital can not be less than half filled in @xmath49 , the measured negative quadrupole moment of the @xmath40 level , even with a very large error bar , would imply a major change in structure compared to the other sn isotopes but there is no other evidence for this happening.leaving out @xmath49 , the total effective charge @xmath50 of the @xmath1 quasiparticle states in the sn isotopes , varies relatively weakly , ranging from @xmath51 to @xmath52 , and the occupation probabilities are therefore a roughly linear function of the neutron number n. this is very similar to the model advocated by yordanov et al .
. however , the cd isotopes are more collective as evidenced by the lower energy and larger @xmath32 of the @xmath36 states and consequently the @xmath1 total effective charge @xmath50 is much larger and more variable , ranging from @xmath53 to @xmath54 near mid - shell .
the occupation probabilities have a more complicated n dependence than assumed by yordanov et al .
@xcite despite the remarkably linear n dependence of the quadrupole moments .
it is interesting that the data for the @xmath55 quasi - particle state in the cd isotopes clearly shows that this orbital fills more rapidly that the @xmath56 orbital .
the sn data are consistent with this .
it is a limitation of laser spectroscopy that it can provide information only on the properties of the ground state and isomeric states .
but a great strength is that it can do so with a consistent methodology for long chains of isotopes . since occupation probablities are monotonic though non - linear functions of n in pretty much all models , the data on each isotope propagates information along the whole chain.for example , in the cd isotopes the occupation probability of the @xmath55 orbital is close to @xmath57 by n=63 and this value will only increase with increasing n ; similarly , the occupation probability of the @xmath58 orbital is down to @xmath59 by @xmath60 and will only decrease with decreasing n. since the coefficients @xmath61 turn out to be small ( always less than 0.03 for the sn isotopes and less than 0.15 for the cd isotopes and much less when @xmath62 is small ) , the weak coupling approximation is reasonable in the present case .
however , calculating the quadrupole moments to first order in @xmath12 also assumes that the quadrupole moment of the collective @xmath36 state in the even isotopes is not too large .
fortunately , this is so . where experimental data are available , @xmath63 ranges from -0.28 b to -0.45 b in cd and is substantially smaller in sn
. the resulting correction to the first order calculation would be less than @xmath64 in the cd isotopes and less than @xmath65 in the sn isotopes .
the value of @xmath31 has a significant uncertainty but it does nt affect the shape of the curves in figs.1 and 2 though it controls the scale of the vertical axis .
the fact that a constant value works well for both isotope chains provides evidence for the validity of the approximation that @xmath66 since the @xmath36 state is expected to be mainly a two neutron quasi - particle state in sn but a mixed proton and neutron two quasi - particle state in cd .
however , the noticeable discontinuity at @xmath67 just below the @xmath34 neutron shell closure may indicate the limitation of the approximation .
it may also be noted that the self - consistent estimate @xcite of the quadrupole interaction strengths gives @xmath68 and @xmath69 . our value is smaller than the self consistent @xmath70 by a factor of roughly four , but this is perhaps not surprising since it multiplies an experimental @xmath32 in equ .
[ q_moment ] which can include various renormalization effects .
i have shown that the measured electic quadrupole moments of a long chain of odd cd and sn isotopes can be simply understood in terms of a weak coupling model . in the model ,
the quadrupole moments are proportional to the occupation probability of the relevant orbitals which can therefore be extracted from the data .
the formula for the quadrupole moment requires the empirical determination one strength parameter ( which is held constant for the two isotope chains ) , but otherwise it involves experimental data and standard values for parameters such as effective charges .
the deduced occupation probabilities agree with expectations except for @xmath49 .
it would be interesting to repeat the experiments on the heavy sn isotopes near the doubly closed shell @xmath71 in order to decrease the large error bars for @xmath72 and to measure @xmath73 .
discussions with j. crawford and f. buchinger are gratefully acknowledged .
this work was supported by an internal research grant from mcgill university .
20ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo
secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ] + 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty @noop * * , ( ) @noop _ _ ( , ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop @noop * * , ( ) @noop * * , ( ) http://cds.cern.ch/record/1602593/files/intc-p-383.pdf [ ] ( ) @noop * * , ( ) http://www.nipne.ro/indico/getfile.py/access?contribid=242&sessionid=3&resid=0&materialid=slides&confid=0 [ `` , '' ] ( ) , @noop * * , ( ) , @noop * * , ( ) , @noop _ _ , ( , ) @noop * * , ( ) @noop _ _ ( , ) | it is shown that a simple bcs model with a quadrupole - quadrupole interaction provides a consistent description of the measured quadrupole moments of a sequence of odd mass sn and cd isotopes and allows the extraction of the neutron single particle occupation probabilities . |
semiclassical theories have become very important for a deeper understanding of quantum systems , and gutzwiller s trace formula @xcite has become a powerful tool for classically chaotic systems .
it provides a semiclassical approximation of the quantum level density in terms of classical periodic orbits . in a systematic expansion of the level density in powers of
@xmath0 it can be considered as the leading order .
higher orders of this asymptotic expansion have been developed in several studies @xcite , but for a long time were only tested for billiard systems , i.e. , systems with hard walls instead of smooth potentials . by extending an expansion which was derived by gaspard et al .
@xcite , grmaud @xcite developed @xmath0 corrections to gutzwiller s trace formula for quantum systems with a smooth potential .
he presented a first - order @xmath0 correction term to gutzwiller s trace formula and obtained numerical results for the diamagnetic hydrogen atom .
comparisons with values extracted from exact quantum calculations showed that the semiclassical results can be computed with very high accuracy . however , in the theory presented in @xcite some important topics have not yet been considered .
for example , the @xmath0 corrections can not be calculated for classical orbits which have a turning point ( i.e. , the velocity vanishes completely at this point ) .
furthermore , discrete and continuous symmetries have not been included . if there are discrete symmetries the eigenstates of the quantum system split up into several subspaces . in these subspaces , classical orbits which are not periodic without a symmetry transformation contribute to the level density .
preliminary results for the diamagnetic hydrogen atom , which is an example system with discrete symmetries , were published in @xcite without explaining which modifications in the numerical calculations the introduction of discrete symmetries entails . in particular , an analysis of the differential equations which have to be solved to obtain the correction terms is not presented . in this paper
we want to give a transparent derivation of the modifications necessary for the calculation of the correction terms in systems with discrete symmetries .
we will show in detail how such symmetries of the hamiltonian have to be taken into account .
it will be explained in which equations it is necessary to introduce symmetry operations and which transformations lead to the correct boundary conditions of the classical greens function , which is an essential part of the correction terms .
furthermore , we want to give a deeper understanding of the significance of the symmetry of the system for the quantum spectrum and the connection to periodic orbits .
we will apply the method to the diamagnetic hydrogen atom .
the three - dimensional hydrogen atom in a uniform magnetic field has a continuous symmetry , namely the rotational invariance around the magnetic field axis .
continuous symmetries have a substantial influence on the correction terms .
actually one can not obtain the correct results for the first - order @xmath0 correction for the three - dimensional hydrogen atom with the formulas presented because the rotational invariance leads to additional contributions to the first - order corrections .
therefore , we will consider the hydrogen atom as a pure two - dimensional system as it was done before in @xcite . if one uses semiparabolic coordinates , the potential of the two - dimensional diamagnetic hydrogen atom exhibits a discrete @xmath1-symmetry .
the @xmath0 corrections will be calculated for selected periodic orbits taking into account that discrete symmetry .
the semiclassical results will be compared with the analysis of exact quantum calculations .
the agreement between the results of both methods turns out to be very good .
the outline of the paper is as follows . in section 2
we will first give a summary of the derivation of the @xmath0 corrections without taking into account discrete symmetries .
in section 3 , we will introduce the hydrogen atom in a uniform magnetic field with all aspects relevant for the calculation of the correction terms .
we will also discuss the influence of the rotational invariance of the three - dimensional hydrogen atom .
then we will extend the @xmath0 corrections to discrete symmetries , calculate results for the two - dimensional hydrogen atom and compare them with exact quantum mechanical calculations in section 4 .
as mentioned in the introduction the @xmath0 corrections to gutzwiller s trace formula for systems with a smooth potential were derived by grmaud @xcite based on earlier work by gaspard et al .
@xcite . in this section , for the reader s convenience we will briefly summarize all major steps of this derivation by following the way described in @xcite in order to be able to include discrete symmetries in section [ sec : symmetries_influence ] .
the starting point for the derivation of the semiclassical level density which is used in this theory is a discrete version of feynman s path integral representation of the propagator @xmath2 , \label{eq : propagator_pathintegral}\end{gathered}\ ] ] for a system with @xmath3 degrees of freedom and a time independent hamiltonian : @xmath4 @xmath5 represents the propagation of a particle going from @xmath6 to @xmath7 in time @xmath8 with @xmath9 .
@xmath10 denotes the classical lagrangian of the system
. for the calculation of the level density one has to know the trace of the propagator , which can be obtained by setting @xmath11 in and by integrating over this variable : @xmath12.\end{gathered}\ ] ] after expanding the exponential function and including all terms which produce contributions to the leading order in @xmath0 and to the first - order correction , the integral is evaluated as described in @xcite .
the result in the limit @xmath13 is given by : @xmath14 in this formula the sum runs over all periodic orbits of the classical system described by the hamiltonian with period @xmath8 .
@xmath15 represents the time period of the primitive periodic orbit , i.e. , if the orbit consists of multiple repetitions of a periodic orbit , one considers only the basic traversal in this case .
@xmath16 stands for the leading order of the semiclassical approximation .
it is identical with gutzwiller s @xcite result , and is given by : @xmath17 .
\end{gathered}\ ] ] in this expression one can find the @xmath18-dimensional stability matrix @xmath19 , which represents the stability properties of the periodic orbit .
the determinant @xmath20 is related to the @xmath21-dimensional monodromy matrix .
this relation is shown in section [ sec : derivatives_c_0 ] .
a further term is the classical action @xmath22 @xmath23 is the maslov index , and @xmath24 is the derivative of the time period with respect to the energy @xmath25 . in equation
@xmath26 is the abbreviation of the first - order correction in @xmath0 . as one can see in one has to use the average of @xmath27 , where @xmath28 parametrizes the periodic orbit .
the integral over @xmath28 runs from @xmath29 to @xmath15 , i.e. , one has to integrate over the whole primitive periodic orbit .
the explicit expression of @xmath27 needs lengthy calculations , which are discussed in @xcite in detail .
the result reads : @xmath30 where @xmath31 are derivatives of the potential evaluated at the point @xmath32 on the classical orbit .
@xmath33 are the components of the classical green s function , which is a solution of the linearized equation of motion : @xmath34 due to the factor @xmath35 in equation , the correction term is singular for vanishing velocities , and because of the integral over @xmath28 in equation , the @xmath0 correction term diverges for orbits with a turning point .
the classical green s function is an essential part of the first - order @xmath0 correction to the trace of the propagator . in order to be able to look at the symmetry transformations of @xmath36 in equation
, one has to know its structure . in this section
we discuss all important parts of this structure . from the derivation of the correction term @xmath37
, it follows that the classical green s function has to fulfil the boundary conditions @xcite @xmath38 with the projection operator @xmath39 along the direction of the classical orbit at time @xmath28 , which has the form @xmath40 furthermore , @xmath41 , with the @xmath3-dimensional unity matrix @xmath42 .
if we use the notation @xmath43 it is possible to write the classical green s function as a product @xcite @xmath44 with the @xmath45 monodromy matrix @xmath46 and the four @xmath47 matrices @xmath48 and @xmath49 .
the monodromy matrix is a symplectic matrix , which can be obtained by solving the linearized hamiltonian equations of motion @xmath50 here , @xmath51 is the matrix @xmath52 and @xmath42 is the @xmath53 unity matrix . for a hamiltonian of the form it has the structure : @xmath54 where @xmath55 and @xmath56 are @xmath47 matrices .
the boundary conditions at time @xmath57 are @xmath58 exploiting the formulation of the classical green s function and the symplecticity of the monodromy matrix @xmath46 , one can formulate the boundary condition as follows : @xmath59 if this expression is combined with the condition @xmath60 which is identical with the first and third condition of , it leads to the matrix equation : @xmath61 \left ( \begin{array}{c } { \ensuremath{\boldsymbol{a}}}_-(t ' ) \\ { \ensuremath{\boldsymbol{b}}}_-(t ' ) \end{array } \right ) = \left ( \begin{array}{cc } { \ensuremath{\boldsymbol{1 } } } & { \ensuremath{\boldsymbol{0 } } } \\ { \ensuremath{\boldsymbol{0 } } } & \mathcal{q}_{t_0 } \end{array } \right ) { \ensuremath{\boldsymbol{m}}}(t ) \left ( \begin{array}{r } -{\ensuremath{\boldsymbol{j}}}_1(t')^\mathrm{t } \\ { \ensuremath{\boldsymbol{j}}}_2(t')^\mathrm{t } \end{array } \right ) .
\label{eq : matrix_equation_for_a_b}\ ] ] this equation has to be solved in order to obtain the matrices @xmath62 and @xmath63 .
when @xmath62 and @xmath63 are known one has the solution for the classical green s function .
unfortunately , the matrix @xmath64\ ] ] is singular but , as was shown in @xcite , there is a solution for classical trajectories which have no turning point ( i.e. the velocity @xmath65 never vanishes completely along the orbit ) , viz .
@xmath66 where @xmath67 represents a particular solution of equation , which can be determined by a singular value decomposition ( see e.g. @xcite ) .
the second condition of equation has been used to derive this result . in summary , it is possible to obtain the classical green s function by solving the linearized equation of motion of the monodromy matrix and by calculating @xmath62 and @xmath63 from equation . then the correction term @xmath27 follows from equation . in equation
, the factor @xmath68 appears in the construction of @xmath27 a second time .
the reason for this lies in the application of the projection operator @xmath39 in the second condition of .
the integrals and double integrals in equation can be transformed into a set of ordinary differential equations , which can be computed effectively .
the transformation of the integrals is discussed in @xcite in detail .
the next step on the way to the level density is the trace of the quantum green s function , which can be obtained from the trace of the propagator via the semi - sided fourier transform @xmath69 k_\ell(t ) .
\label{eq : semisided_fourier}\ ] ] one has to include all contributions to the leading order as well as the first - order @xmath0 correction in the semiclassical approximation of the integral .
the methods which are used to evaluate this integral are discussed in @xcite .
we do not want to repeat the calculation but present the results obtained in @xcite and @xcite , which can be summarized as follows : @xmath70 \\
\times \left \ { 1 + { \mathrm{i}}\hbar \left ( c_{1\,\ell}(t_0 ) + c_{1\,\ell}^{t \to e}(t_0 ) \right ) + { \mathrm{o}}\left ( \hbar^2 \right ) \right \ } , \label{eq : semicl_greens_function}\end{gathered}\ ] ] where the leading order , belonging to the 1 in the curly brackets , is known from gutzwiller s trace formula @xcite .
the reduced action @xmath71 is used in this expression , and @xmath23 represents the maslov index .
the first - order @xmath0 correction consists of two terms .
the first , @xmath72 is the same as in equation but now with fixed period @xmath73 , and the second is the contribution from integral to the @xmath0 correction , and is given by : @xmath74 @xmath75 and @xmath76 are the derivatives @xmath77 of the action @xmath78 and of the logarithm of the amplitude of the trace of the propagator @xmath79 respectively . in this section
we want to summarize the contributions to the second correction term @xmath80 which are important for the understanding of the symmetry properties of this term .
as we will see , all parts of equation can be obtained as numerical solutions of differential equations of some `` new '' coordinates . because of the relation @xmath81 all higher derivatives of the action @xmath82 with respect to the time period of the orbit
can be expressed as derivatives of the classical hamiltonian , where @xmath83 represents the phase space vector of the orbit . introducing the derivatives
@xmath84 which appear as coefficients of the taylor expansion @xmath85 of the orbit @xmath86 around a reference orbit @xmath87 with period @xmath73 , the higher derivatives of @xmath82 are : @xmath88 from hamilton s equations of motion for the phase space coordinates @xmath89 one can infer the equations which govern the motion of the derivatives @xmath90 : @xmath91 these equations are inhomogeneous differential equations .
their solutions are of the type @xmath92 where @xmath93 is the monodromy matrix , which was introduced in section [ sec : class_green_trace_prop ] .
together with the condition @xmath94 for periodic orbits , the solutions lead to the following equations for the initial values : @xmath95
\vec{x}^{(1)}(0,t_0 ) & = \vec{\dot{x}}(t_0,t_0 ) , \\ \left [ { \ensuremath{\boldsymbol{1 } } } - { \ensuremath{\boldsymbol{m}}}(t_0,t_0 ) \right ] \vec{x}^{(2)}(0,t_0 ) & = 2 \vec{\dot{x}}^{(1)}(t_0,t_0 ) + \vec{\ddot{x}}(t_0,t_0 ) + \vec{f}^{(2)}(t_0,t_0 ) , \\ \left [ { \ensuremath{\boldsymbol{1 } } } - { \ensuremath{\boldsymbol{m}}}(t_0,t_0 ) \right ] \vec{x}^{(3)}(0,t_0 ) & = 3 \vec{\dot{x}}^{(2)}(t_0,t_0 ) + 3 \vec{\ddot{x}}^{(1)}(t_0,t_0 ) + \vec{\dddot{x}}(t_0,t_0 ) \\ & \quad+ \vec{f}^{(3)}(t_0,t_0 ) . \end{aligned } \label{eq : starting_values_singular_matrix}\ ] ] the matrix @xmath96 $ ] is singular .
a particular solution can be found via a singular value decomposition . for a general unstable orbit
the kernel is one - dimensional , i.e. , the solution space is in general one - dimensional ( see @xcite ) but , as also shown in @xcite , @xmath97 can directly be computed if a particular solution @xmath98 from is known : @xmath99 similar relations can be found for @xmath100 and @xmath101 , in which also only a particular solution of equation is required .
if one looks at the definition of @xmath102 in equation , one can easily convince oneself that @xmath103 the derivatives of the action are already known from section [ sec : derivatives_w ] . in the next step one has to find an expression for the derivatives of @xmath104 .
following again the method presented in @xcite we can introduce a new matrix @xmath105 whose determinant is identical with @xmath106 .
it is given by @xmath107 where @xmath108 and @xmath109 are the projection operators parallel to the flow of the classical orbit and perpendicular to the energy shell , respectively .
these relations lead to the expressions @xmath110 @xmath111 with @xmath112 by using the derivatives @xmath90 and the definition of the projection operator @xmath113 , one obtains : @xmath114 the last pieces which are needed for calculating @xmath102 are the first and the second derivative of the monodromy matrix @xmath115 with respect to the period @xmath8 : @xmath116 similar to the procedure which was used in section [ sec : derivatives_w ] to arrive at the equations of motion for the @xmath90 , it is possible to obtain the differential equations governing the evolution of the derivatives @xmath117 from differential equation of the monodromy matrix .
the equations read @xmath118 with initial values @xmath119 now we have obtained all results necessary for the calculation of the first - order @xmath0 corrections to gutzwiller s trace formula without taking into account symmetries . in the next sections we will apply this theory to the diamagnetic hydrogen atom .
the diamagnetic hydrogen atom was often used as an example for a quantum system whose classical dynamics is chaotic ( see e.g. @xcite or @xcite for an overview ) . as a real physical system
it was the topic of studies in experimental physics @xcite .
it has even been used for the numerical test of the @xmath0 correction terms @xmath120 and @xmath121 @xcite . because of its simple scaling property , which is also fulfilled for the @xmath0 corrections , it is possible to compare the semiclassical results for individual orbits with exact quantum mechanical calculations .
the hydrogen atom in a uniform magnetic field has a continuous symmetry , namely the rotational invariance around the magnetic field axis .
this symmetry can be used to formulate the dynamics in a two - dimensional coordinate system .
this method works very well for the leading order but it leads to new difficulties for the first - order @xmath0 correction .
this aspect is discussed in section [ sec : rot_invariance ] before we will look at the `` two - dimensional hydrogen atom '' with its discrete symmetries . written in
the cylindrical coordinates @xmath122 , @xmath123 and @xmath124 and atomic units , the classical hamiltonian of the hydrogen atom in a uniform magnetic field in @xmath124-direction has the form : @xmath125 where @xmath126 in this expression @xmath127 is the @xmath124 component of the angular momentum , which is a constant of motion .
the paramagnetic term is not considered in equation because it is constant .
if one exploits the classical scaling property @xmath128 the scaled hamiltonian @xmath129 only depends on the scaled energy @xmath130 , but not on the energy @xmath25 or the magnetic field strength @xmath131 separately .
the regularization of the coulomb singularity can be achieved by introducing semiparabolic coordinates @xcite @xmath132 and the scaled time @xmath133 @xmath134 this transformation leads to the new hamiltonian @xmath135 where the momenta are defined by @xmath136 the scaled time @xmath133 is the parameter for the integration of the differential equations of the classical values . in the case
@xmath137 , the @xmath138 and @xmath139 coordinates , which are only defined as positive coordinates in equation , can be extended to negative values .
then , @xmath140 and @xmath141 have the same structure as the momenta of cartesian coordinates .
this structure of the momenta is very important because it is often used in the derivation of the @xmath0 corrections ( see e.g. equation for the monodromy matrix ) .
however , it can not be found in the quantum mechanical analogue of equation .
starting again with the hamiltonian in cylindrical coordinates @xmath142 and using the transformations and , one arrives at the schrdinger equation for the @xmath143-part of the wave function : @xmath144 in this quantum mechanical expression , the momentum operators for @xmath138 and @xmath139 differ from those in cartesian coordinates even for @xmath145 , which is the quantum analogue of @xmath146 . as was mentioned above , this difference does not have any effect on the leading order of the semiclassical level density .
the results of the classical calculations agree very well with values which were extracted from quantum calculations .
but there is no agreement at all between the classically evaluated correction terms @xmath120 and @xmath147 and quantum mechanical calculations .
this behaviour can be understood , when one introduces the new wave function @xmath148 which leads to the following schrdinger equation : @xmath149 now , the momentum operators agree with the structure in the classical case but an additional term @xmath150 appears , which leads to a non - vanishing centrifugal part of the potential even for @xmath145 .
the additional term has also to be inserted in the classical hamiltonian : @xmath151 since the new centrifugal part is proportional to @xmath152 , which was set to one when atomic units were introduced , it contributes to the first - order @xmath0 correction in the expansion of the level density as one can easily see when one repeats the calculations from section [ sec : derivation ] . that is the reason why the influence of the continuous symmetry does not appear in the leading order but in the first - order corrections .
the centrifugal term consists of an integral over the classical periodic orbit with the kernel @xmath153 , which diverges at the coordinate axes .
this problem does not arise if one regards the hydrogen atom as a purely two - dimensional system because in this case there is no centrifugal term .
the momenta have the structure of the momenta of cartesian coordinates in the classical dynamics as well as in the quantum case .
therefore , in the following sections we only look at this _ two - dimensional hydrogen atom _ for the calculation of the first - order @xmath0 corrections , as it was done in @xcite before .
the two - dimensional diamagnetic hydrogen atom can be formulated similar to the three - dimensional one . for a magnetic field @xmath154
, the classical hamiltonian is given by : @xmath155 if one uses the classical scaling property and semiparabolic coordinates for the two - dimensional case @xcite @xmath156 the transformed hamiltonian reads : @xmath157 the schrdinger equation associated with the classical hamiltonian can be obtained by the same procedure as in the three - dimensional case .
the result is : @xmath158 in this case , the momentum operators have exactly the same structure as for cartesian coordinates .
note that @xmath159 takes the place of @xmath0 , which is equal to one in atomic units and which is often called `` effective @xmath0 '' .
the `` potential '' of the two - dimensional diamagnetic hydrogen atom in semiparabolic coordinates ( see equation ) @xmath160 has a @xmath1-symmetry .
this symmetry can be seen in figure [ fig : symmetry ] ( a ) , in which a few equipotential contours of the potential are plotted . for the leading order ( gutzwiller s trace formula ) , it is known , how the symmetry of the system can be exploited in the calculation of the classical values . in section [ sec : symmetries_influence ] we look at the influence of the discrete @xmath1-symmetry on classical values which contribute to the first - order @xmath0 correction terms @xmath120 and @xmath121 . [ cols="^,^ " , ]
in this paper we extended the theory presented in @xcite to systems with discrete symmetries and the influence of these symmetries on the correction terms was discussed in detail .
the symmetry transformations presented in this paper made possible the calculation of the correction terms for a number of orbits which could not be included without symmetry operations .
nevertheless , these orbits , which are not periodic in the plane of semiparabolic coordinates without a symmetry operation , contribute to some of the subspaces of the quantum spectrum .
the classical results for the correction terms could be compared with values which were extracted from exact quantum calculations .
an excellent agreement between the results of both methods was found . in spite of this success
, it must be noted that before semiclassical spectra including first - order @xmath0 corrections can be calculated over the complete spectrum a number of problems still remain to be solved . on the one hand , as was already mentioned in @xcite , and was pointed out in section [ sec : derivation ] , the correction term @xmath120 can not be calculated in the form presented for orbits which have a turning point , but the inclusion of these orbits is essential for the @xmath0 correction of the level density .
the correction term @xmath120 diverges for orbits with turning points .
however , the extraction of the @xmath0 correction terms from quantum spectra with the method presented in section [ sec : quantum_spectra ] leads to results which are on the same order of magnitude as the values of orbits without turning points .
thus we can assume that the reason lies in an insufficiency of the theory , and is not a physical property . on the other hand
, besides the successfully implemented discrete symmetries , physical systems often have a continuous symmetry .
for example , it is necessary to take into account the rotational invariance around the magnetic field axis for the complete calculation of all first - order @xmath0 corrections to the semiclassical level density of the three - dimensional diamagnetic hydrogen atom . as was mentioned in section [ sec : rot_invariance ] , the problem can be considered as an additional centrifugal term in the potential .
this term leads to diverging integrals if the course of the orbit is not changed .
a regularization of these integrals suggested in @xcite leads to good results for a few individual orbits , however , a mathematical justification is lacking .
furthermore , for the hydrogen atom it would be interesting to look at the @xmath161 level density , @xmath162 , where @xmath161 is the scaling parameter introduced in section [ sec : hinv_quantum_spectra ] , in contrast to the `` energy '' level density . in this case
, there is no need to calculate the quantum mechanical matrix elements which in our case are required in equations and .
this was already done before for the leading order ( see e.g. @xcite ) .
this means that one has to find a semiclassical expression for a level density with weighting factors , @xmath163 .
eckhardt et al .
@xcite presented an extension of gutzwiller s theory which allows the calculation of the leading order of the level density @xmath164 .
it would be desirable to include higher orders of the @xmath0 expansion in that theory . | we extend a theory of first order @xmath0 corrections to gutzwiller s trace formula for systems with a smooth potential to systems with discrete symmetries and , as an example , apply the method to the two - dimensional hydrogen atom in a uniform magnetic field .
we exploit the @xmath1-symmetry of the system in the calculation of the correction terms .
the numerical results for the semiclassical values will be compared with values extracted from exact quantum mechanical calculations .
the comparison shows an excellent agreement and demonstrates the power of the @xmath0 expansion method . , , and semiclassical theories , gutzwiller s trace formula , @xmath0 expansions , discrete symmetries , diamagnetic hydrogen atom 03.65.sq , 05.45.mt |
designing the electromagnetic response of an artificial medium is one of the main target of modern photonics and metamaterial science is probably the most important research field based on such skill .
basically the design is made possible by the physical fact that the electromagnetic field , when traveling within a nonhomogeneous medium with sub - wavelength features , is not able to follow its spatial rapidly - varying details so that the field only experiences the effect of an averaged or effective medium .
a number of different homogenization approaches have been developed for predicting the effective medium electromagnetic response and they exploit different and suitable approximation schemes . the simplest homogenization technique deals with the retrieval of the effective parameters from the scattering properties of the medium @xcite and it is based on postulating the equivalence between a complex metamaterial array and a uniform slab of same thickness with unknown constitutive parameters .
another homogenization technique is the field averaging method which is based on the averaging of the electromagnetic field in a metamaterial unit cell @xcite and , in analogy with the retrieval technique , it is a numerical method for the determination of effective parameters .
in addition to numerical methods , mean - field homegenization theories are available where the effective parameters are evaluated from the distribution of the underlying metamaterial inclusions .
examples of such techniques are those exploiting lorentz @xcite , clausius - mossotti @xcite , maxwell - garnett @xcite approximations , or based on multipolar expansion @xcite and source - driven approach @xcite . spatial periodicity and rapidly - varying spatial scales are the two basic ingredients of each metamaterial homogenization approach .
two different homogenization techniques are obtained by assuming one of these two ingredients and subsequently incorporating the other .
therefore , starting from the photonic - crystal description of the structure where the spatial periodicity is fully taken into account , the effective medium response can be extracted in the long wavelength regime @xcite .
conversely , the spatial rapidly - varying metamaterial features allows an asymptotic multi - scale analysis of the sample electromagnetic response which combined with the array periodicity yields the effective medium response @xcite .
even though the metamaterial inclusions pattering has a spatial scale much smaller than the radiation wavelength , such two scales are generally not so different to allow a description of the effective medium response only comprising effective dielectric permittivity and magnetic permeability tensors . for this reason
, the effective medium generally shows an additional nonlocal respose @xcite yielding spatial dispersion .
it is well - known that the nonlocal first order contribution ( i.e. containing first order spatial derivatives of the electric field ) is equivalent to a reciprocal bianisotropic response @xcite whereas the second order contribution can be partially interpreted as a correction to the effective magnetic permeability tensor @xcite , a phenomenon which is known as artificial or optical magnetism @xcite .
if the effective nonlocality is weak , the effective medium response is adequately described by reciprocal bianisotropic constitutive relations where the chirality tensor accounts for the strength of magnetic and electric polarizations coupling . an efficient way for observing the effect of electromagnetic chirality is considering metamaterials whose underlying constituents pattering exhibits chiral asymmetry , i.e. its mirror image can not be superposed onto it , and theoretical and experimental investigations have been performed both in three - dimensional @xcite and in two - dimensional configurations @xcite .
chiral metamaterials have attracted a good deal of attention since they can yield giant optical activity , asymmetric transmission @xcite , repulsive casimir force @xcite , backward waves @xcite and negative refractive index @xcite . in this paper
we theoretically investigate the electromagnetic response in the long wavelength regime of a periodic composite medium , whose inclusions are non - magnetic , and we carry out the analysis by including spatial dispersion up to the second order .
specifically , by using the ratio @xmath2 between the composite periodicity and the wavelength as an asymptotic expansion parameter , we exploit a general multiscale technique for separating the fast matter scale from the slow radiation one and to extract the average effect of the composite on the electromagnetic field .
we fully develop the analysis up to the second order in @xmath2 and we find that the obtained effective medium response has a @xmath3 order reproducing the know results of standard homogenization approaches and @xmath4 and @xmath5 orders accounting for first and second order effective medium nonlocality , respectively .
the effective dielectric permittivity tensor ( of rank two ) arising from the zeroth order is obtained by averaging over the metamataterial unit cell specific fast - varying fields satisfying electrostatic - like equations .
the main result of the present paper is that this very simple procedure still works for evaluating the two tensors ( of ranks three and four , respectively ) describing effective medium nonlocality up to the second order and the corresponding electrostatic - like equations for the suitable fast - varying fields to be averaged are here derived and discussed . as a consequence we obtain a simple scheme for evaluating the effective medium response up to the second order which is based on solving a small number of poisson equations on the metamaterial unit cell .
in particular we obtain a very compact expression for the three - rank tensor describing first order spatial nonlocality which remarkably reveals that it can be evaluated by using the same fast - varying fields appearing in the zeroth order description , thus avoiding the requirement of solving additional electrostatic - like equations . therefore the ensuing procedure for deducing the medium response up to
the first order is remarkably simple and compact . for metamaterials with negligible second - order spatial dispersion , we combine the obtained description with the equivalence of first - order spatial dispersion and reciprocal bianisotropic electromagnetic response to obtain a compact and simple expression for the effective medium chirality tensor .
using such expression we investigate the impact of the composite patterning symmetry on electromagnetic chirality .
specifically we obtain that if the composite does not show chiral asymmetry , i.e. its mirror image can be superposed on it , the effective chiral tensor vanishes with two remarkable exceptions corresponding to the specific cases where the rotation needed to superpose the mirror image onto the structure is either a @xmath0 or @xmath1 rotation around an axis orthogonal to the reflection plane . in these two situations
the composite is geometrically achiral and it nonetheless shows electromagnetic chirality whose chiral tensor turns out to be that of pseudo - chiral - omega medium .
such general result encompasses in particular the relevant cases of two - dimensional and recently considered one - dimensional metamaterials showing chiral response . as an example admitting full analytical description ,
we discuss one - dimensional metamaterials whose single chirality parameter is shown to be directly related to the metamaterial dielectric profile by quadratures .
the paper is organized as follows . in sec .
ii we discuss our approach to the effective medium description of a periodic composite in the long wavelength regime including spatial nonlocality .
this is done examining the general multiscale technique in sec .
ii.a and applying it to the zeroth , first and second orders in sections ii.b , ii.c and ii.d , respectively . in sec .
iii we sum up the results obtained in sec .
ii and we stress the simplicity and compactness of the proposed approach . in sec .
iv we consider media whose second order contribution to spatial nonlocal response can be neglected and we focus on electromagnetic chirality . specifically in sec .
iv.a we adopt the bianisotropic description of the medium to deduce a simple expression for the chirality tensor , in sec .
iv.b we examine the relation between geometric and electromagnetic chirality of metamaterials and in sec .
iv.c we specialize such general analysis to the cases of two - dimensional and one - dimensional chiral metamaterials . in sec .
v we focus on one - dimensional chiral metamaterials and we obtain a closed - form expression for the single parameter ruling electromagnetic chirality of such media . in sec .
vi we draw our conclusions .
let us consider propagation of a monochromatic electromagnetic field through an unbounded metamaterial whose underlying non - magnetic inclusions ( both metal and dielectric ) are arranged on a lattice whose period is much smaller than the radiation wavelength @xmath6 .
the electric @xmath7 and magnetic @xmath8 field amplitudes satisfy maxwell equations @xmath9 where time dependence @xmath10 has been assumed and the medium relative dielectric constant @xmath11 is a periodic complex function having the same periodicity of the metamaterial .
due to the rapidly varying dielectric spatial modulation it is convenient to introduce the parameter @xmath12 where @xmath13 is the largest of the lattice basis vector lengths . following the standard multiscale technique , it is possible to exploit the condition @xmath14 to develop an asymptotic analysis of electromagnetic propagation .
accordingly , we introduce the fast spatial coordinates @xmath15 , regarded as independent from the slow coordinates @xmath16 .
any field @xmath17 can be decomposed as @xmath18 where the overline denote the spatial average over the metamaterial unit cell , i.e @xmath19 where @xmath20 is the unit cell and @xmath21 is its volume scaled by @xmath22 , and the tilde denote the rapidly varying zero mean residual , i.e. @xmath23 . in our scheme
the relative dielectric permittivity depends on the fast coordinates only so that we set @xmath24 and it can be decomposed as @xmath25 . following the general multiscale technique @xcite , we set for the electromagnetic field amplitudes @xmath26 @xmath27 \eta^n,\ ] ] or , in other words , the field amplitudes are expanded in powers of @xmath2 and , for each order , the average on the unit cell ( terms with an overbar and depending on the slow coordinates only ) is separated from the zero mean residual ( terms with an overtilde and depending on both slow and fast coordinates ) . after substituting eqs.([two - scale ] ) into eqs.([maxwell ] ) and noting that @xmath28 it is possible to extract equations for each order and , for each of them , to separately balance the averaged contributions and zero mean residuals . as a result , for the averaged equations we obtain @xmath29 whereas for the zero mean residual equations we have @xmath30\end{aligned}\ ] ] where @xmath31 .
equations ( [ n - slowly ] ) do not contain the fast coordinates @xmath32 and are coupled to the rapidly varying field orders only through the averaged term @xmath33 .
multiplying each of eqs.([n - slowly ] ) for @xmath34 and summing over @xmath35 we obtain @xmath36 where @xmath37 and @xmath38 is the overall averaged electromagnetic field and @xmath39 where @xmath40 therefore the slowly varying electric and magnetic field amplitudes satisfy the macroscopic maxwell equations ( [ macroscopic ] ) with the slowly varying displacement vector @xmath41 of eq.([eff - disp ] ) which has two contributions , the former due to the spatial average of the dielectric profile and latter due to the dielectric modulation .
the latter contribution is obtained by summing the spatial average of the rapidly varying fields @xmath42 multiplied by the dielectric permittivity arising from each order .
therefore , in order to obtain an effective medium description of the metamaterial response , such rapidly varying fields have to be related to the slowly varying ones through eqs.([n - rapid ] ) . from the structure of eqs.([n - rapid ] ) we note that the fields @xmath43 and @xmath44 have to be evaluated recursively from the knowledge of @xmath45 and @xmath46 .
in order to accomplish this task it is convenient to deduce equations involving the fields divergence . after applying the operator @xmath47 to both the third and fourth of eqs.([n - rapid ] ) we obtain @xmath48\end{aligned}\ ] ] where we have used the identity @xmath49 .
setting @xmath50 and using the first and the second of eqs.([n - rapid ] ) , eqs.([div ] ) become @xmath51 relabelling @xmath52 in eqs.([div ] ) and substituting the expressions for @xmath53 and @xmath54 from the third and fourth of eqs.([n - rapid ] ) we obtain @xmath55 \nonumber \\ & & - \left(\nabla_{\textbf{r}}\varepsilon \right ) \cdot \overline{\textbf{e}}_{n+1}\end{aligned}\ ] ] for @xmath56 .
equations ( [ n - rapid ] ) together with eqs.([div0 ] ) and ( [ divn ] ) can be used to evaluate the rapidly varying fields of order @xmath57 once those of order @xmath35 are known and these fields are linearly dependent on the slowly - varying fields . therefore , the discussed homogenization technique allows to obtain an effective medium description of the metamaterial respose up to the desired order in @xmath2 . in this paper
we will provide an effective medium description up to the second order . from the second of eqs.([n - rapid ] ) and the first of eqs.([div0 ] ) we note that the field @xmath58 has no sources and therefore it vanishes . on the other hand the first of eqs.([n - rapid ] ) and the second of eqs.([div0 ] ) state that the field @xmath59 is conservative and it is produced by the spatial modulation of @xmath60 and the slowly varying field @xmath61 .
therefore we have @xmath62 where the sum is hereafter understood over repeated indices , @xmath63 is the unit vector along the @xmath64-th direction , @xmath65 is the partial derivative along @xmath66 , @xmath67 and the functions @xmath68 satisfy the equations @xmath69 where @xmath70 . note that the source term of eq.([f0 ] ) has the same metamaterial periodicity and its spatial average vanishes , i.e. @xmath71 , so that the fields @xmath68 are periodic functions with the same metamaterial periodicity and they are defined up to an arbitrary constant ( see appendix a ) . inserting the second of eqs.([0-order ] ) into eq.([dn ] ) with @xmath50 we obtain @xmath72 where @xmath73 so that the zeroth order contribution to the displacement vector of the effective medium amounts to a tensor correction to the average dielectric permittity .
this result agrees with ref.@xcite where the authors consider the homogenization of a two - dimensional structure up to the zeroth order in @xmath2 .
note that , as shown in appendix b , the rank - two tensor @xmath74 satisfies the relation @xmath75 and this is the correct symmetry property of the correction to the dielectric permittivity tensor , in agreement with the onsager symmetry principle @xcite .
after symmetrizing the pair of indices @xmath76 , i.e. by setting @xmath77 , we get @xmath78 where we have set @xmath79 for later convenience .
it is worth noting that the expression for @xmath80 only involves spatial derivatives of the fields @xmath68 so that , at the zeroth order , the undefined additive constant in @xmath68 ( see appendix a ) do not affect the effective correction to the permittivity tensor of eq.([depeff ] ) .
in addition , as a consequence of eq.([f0 ] ) we have the divergenceless condition @xmath81 in order to obtain the equations for the fast varying fields of the first order we use the third and fourth of eqs.([n - rapid ] ) and eqs.([divn ] ) with @xmath50 thus getting @xmath82 , \nonumber \\
\nabla_{\textbf{r}}\cdot \left ( \varepsilon \tilde{\textbf{e}}_1 \right ) & = & - \left(q_{ij}-\overline{q_{ij}}\right ) \frac{\partial \overline{e}_{0j}}{\partial x_i } -(\partial_i \varepsilon ) \overline{e}_{1i } , \nonumber \\
\nabla_{{\textbf{r } } } \times \left ( \tilde{\textbf{e}}_1 - \hat{\textbf{e}}_i f_j \frac{\partial \overline{e}_{0j}}{\partial x_i } \right ) & = & \textbf{0}\end{aligned}\ ] ] where we have used eqs.([0-order ] ) and ( [ q ] ) and we have exploited the evident relation @xmath83 . note that we use the symbol @xmath84 to label the partial derivative operators with respect the slowly varying coordinates @xmath85 in order to avoid confusion with @xmath86 which we use for partial derivative operators with respect the fast varying coordinates .
the first and the second of eqs.([1-order - eqs ] ) provide the field @xmath87 since @xmath88 where eq.([div - less ] ) has been used , is the compatibility condition of the system .
therefore we set @xmath89 \overline{e}_{0j}\ ] ] where the functions @xmath90 satisfy the magnetostatic equations @xmath91 where @xmath92 , which have to be solved with the prescritions that @xmath90 has the metamaterial periodicity and it is defined up to an arbitrary constant ( see appendix a ) .
note that if @xmath90 satisfies the second of eqs.([magneto ] ) then @xmath93 where eq.([div - less ] ) has been used and this implies that @xmath94 since the constant is the only harmonic function which has the metamaterial periodicity . in addition @xmath95 has vanishing average since it is the derivative of a periodic function and therefore @xmath96 .
therefore the first of eqs.([magneto ] ) is automatically satisfied and has not to be additionally required .
the third and the fourth of eqs.([1-order - eqs ] ) provide the field @xmath97 and it is easily seen that it can be expressed as @xmath98\ ] ] where the functions @xmath99 satisfy the equation @xmath100 it is important noting that field @xmath97 has , by its very definition , zero average and , from eq.([e1 ] ) this forces the undefined additive constant of @xmath68 to be set equal to zero , i.e. in such a way that @xmath101 note that the source term of eq.([f1 ] ) , in its right hand side , has vanishing spatial average so that the equation is structurally equivalent to eq.([f0 ] ) and therefore @xmath99 has the metamaterial periodicity and it is defined up to an additive constant ( see appendix a ) . inserting eq.([e1 ] ) into eq.([dn ] ) with @xmath102 we obtain @xmath103\ ] ] where @xmath104 so that the first order contribution to the displacement vector of the effective medium provides a tensor correction to the average dielectric permittivity which has the same structure as its zeroth order counterpart ( see eq.([d0 ] ) ) and a contribution which is linear in the derivatives of the slowly varying zeroth order electric field .
this former contribution is responsible for the first order nonlocal part of the effective medium dielectric response .
the rank - three tensor @xmath105 can be conveniently rewritten as ( see appendix b ) @xmath106 which reveals that the nonlocal contribution of the effective medium response up to the first order can be predicted directly from the knowledge of the functions @xmath107 thus avoiding to solve eqs.([f1 ] ) . note that eq.([alpha ] ) also shows that @xmath108 which is the correct antisymmetric property , in agreement with the onsager symmetry principle , for the third order tensor associated with first order spatial nonlocality of the effective medium response @xcite .
the equations for the fast varying electric field of the second order are obtained from the third of eqs.([n - rapid ] ) and the second of eqs.([divn ] ) with @xmath102 so that @xmath109 \nonumber \\\end{aligned}\ ] ] where we have used eqs.([h1 ] ) and ( [ e1 ] ) , we have set @xmath110 and we have used the evident relation @xmath111 $ ] . note that , since we are investigating the effective medium response up to the second order @xmath5 , we are not considering the equations for the second order magnetic field @xmath112 since it provides contributions to the effective medium response only of order @xmath22 .
it is easily seen that the field @xmath113 satisfying eqs.([2-order - eqs ] ) is @xmath114 \nonumber \\\end{aligned}\ ] ] where the functions @xmath115 and @xmath116 satisfy the equations @xmath117 it is important noting that field @xmath113 has , by its very definition , zero average and , from eq.([e2 ] ) this forces the undefined additive constants of @xmath90 and @xmath99 to be set equal to zero , i.e. in such a way that @xmath118 note that the source terms of eqs.([r ] ) have vanishing spatial average so that such equations are structurally equivalent to eq.([f0 ] ) and therefore @xmath115 and @xmath116 have the metamaterial periodicity and they are defined up to an arbitrary constant ( see appendix a ) . inserting eq.([e2 ] ) into eq.([dn ] ) with @xmath119 we obtain @xmath120\end{aligned}\ ] ] where @xmath121 and @xmath122 .
note that in the fourth term in eq.([d2 ] ) , @xmath123 has been symmetrized with respect the indices @xmath124 since it is contracted with the symmetric mixed partial derivatives tensor . from eq.([d2 ] ) we note that second order contribution to the displacement vector of the effective medium contains a correction to the effective dielectric permittivity tensor and a correction to the first order nonlocal response which which have the same structures as their zeroth and first order counterpart ( see eqs.([d0 ] ) and ( [ d1 ] ) ) . in addition it provides a novel contribution to the effective dielectric permittivity which is physically due to the effect of the rapidly varying first order magnetic field and a contribution which is linear in the mixed second order derivatives of the slowly varying zeroth order electric field , the latter being responsible for the second order nonlocal part of the effective medium dielectric response .
the rank - two tensor @xmath125 and the rank - four tensor @xmath123 can be rewritten as ( see appendix b ) @xmath126}. \nonumber \\\end{aligned}\ ] ] so that the second order contribution to the effective medium response can be predicted without solving eqs.([r ] ) .
it is worth noting that the relations @xmath127 and @xmath128 hold and they are the correct symmetric properties , in agreement with the onsager symmetry principle , for contributions to the effective dielectric tensor and for the rank - four tensor associated with second order spatial nonlocality @xcite , respectively .
the slowly varying electric and displacement fields up the second order are @xmath129 and @xmath130 so that , inserting eqs.([d0 ] ) , ( [ d1 ] ) and ( [ d2 ] ) into eq.([eff - disp ] ) ( whose series is truncated up to the second order ) and adding suitable higher order terms for restoring the field @xmath131 in the nonlocal contributions , we obtain @xmath132 where , from eqs.([depeff ] ) , ( [ alpha ] ) and ( [ gam - bet ] ) , we have @xmath133}. \nonumber \\\end{aligned}\ ] ] where @xmath134 in order to sum up the results obtained so far for clarity purposes , we here report eq.([f0 ] ) , the second of eq.([magneto ] ) and eq.([f1 ] ) for the fields @xmath68 , @xmath90 and @xmath99 necessary for evaluating the effective medium tensors , namely @xmath135 whose solutions have to be determined with periodic boundary conditions on the edges of the metamaterial unit cell and with vanishing spatial average . equations ( [ eff - med ] ) , ( [ eff - tens ] ) and ( [ eff - eqs ] ) are the main result of this paper and they show that the effective dielectric tensor @xmath136 and the tensors @xmath137 and @xmath138 describing first and second order spatial dispersion can easily be evaluated after solving eqs.([eff - eqs ] ) for a given periodic dielectric profile @xmath139 . in turn
these equations share a common electrostatic structure and therefore they can be faced through a number of well known theoretical and numerical schemes .
in addition we note that , to the best of our knowledge , the second and the third of eqs.([eff - tens ] ) contain the simplest expressions for the tensors @xmath137 and @xmath138 since they are easily obtained by averaging suitable mesoscopic fields .
we conclude that , the discussed homogenization scheme provides a particularly simple way for predicting and analyzing , in addition to the local dielectric response , the spatial dispersion properties of an arbitrary tridimensional periodic dielectric medium in the long wavelength regime up to the second order .
let us focus here on the situation where the contributions of order @xmath5 in eq.([eff - med ] ) can be neglected .
it is well known that in this case the effective medium first order nonlocal response is equivalent to a reciprocal bianisotropic response which is referred to as a electromagnetic chirality @xcite . in order to evaluate the chiral tensor
let us note that maxwell equations for the effective medium are left invariant by the transformation @xmath140 where @xmath141 is an arbitrary vector field , since eqs .
( [ macroscopic ] ) , after defining @xmath142 , can be rewritten as @xmath143 this shows at the same time that the fields @xmath144 and @xmath8 are not uniquely defined ( as opposed to the fields @xmath7 and @xmath145 ) and that , for each possible vector field @xmath141 , the constitutive relations @xmath146 characterize the electromagnetic behavior of the effective medium . choosing @xmath147 for the arbitrary vector field @xmath141 , where @xmath148 and @xmath149 is the levi - civita tensor , and exploiting the crucial antisymmetric property @xmath150 , eqs.([constitutive ] ) become ( see appendix c ) @xmath151 where @xmath152 .
equations ( [ constitutive1 ] ) shows that the effective medium has a reciprol magneto - electric coupling or electromagnetic chirality described by chiral tensor @xmath153 of eq.([kappa ] ) . inserting the second of eqs.([eff - tens ] ) into eq.([kappa ] ) we obtain @xmath154.\ ] ] this relation is among the main results of the present paper and to the best of our knowledge it is the simplest expression for the effective medium chiral tensor since it is directly evaluated using the functions @xmath107 obtained by solving the first of eqs.([eff - eqs ] ) .
since we are dealing with a very general and tridimensional situation it is convenient to express the chirality tensor as @xcite @xmath155 where @xmath156 is the trace of the chirality tensor , @xmath157 is the identity tensor , @xmath158 is a symmetric trace - free tensor and @xmath159 is an antisymmetric tensor .
such decomposition allows to classify the reciprocal bianisotropic response @xcite of the effective medium and the three main classes are @xmath160 for chiral media , @xmath161 for pseudochiral media and @xmath162 for omega media . for the chirality tensor of eq.([kappa1 ] ) we obtain after some straightforward algebra @xmath163 \overline{\varepsilon f_m \partial_q f_n } , \nonumber \\
j_{ij } & = & \eta k_0 \left [ \epsilon_{imj } \overline{\varepsilon f_m } + \frac{1}{2 } \left(\epsilon_{imn } \delta_{jq } - \epsilon_{jmn } \delta_{iq } \right ) \right ] \overline{\varepsilon
f_m \partial_q f_n}. \nonumber \\\end{aligned}\ ] ] a structure is chiral if it is distinguishable from its mirror image or , in other words , if its mirror image can not be rigidly superposed onto it
. if the structure is not chiral it is necessary invariant under a set of geometrical symmetries which affect the structure of the chirality tensor @xmath153 . in order to investigate the relation between geometrical and electromagnetic chirality it is necessary to identify the symmetries of the dielectric profile which either entail the vanishing or select a specific structure of the chiral tensor .
let us suppose that the metamaterial is achiral or , in other words , that the mirror image of the dielectric profile through a plane orthogonal to the unit vector @xmath164 can be rigidly superposed onto the original profile through the composition of a rotation of an angle @xmath165 around the unit vector @xmath166 and a translation @xmath167 .
this implies that the dielectric permittivity is left invariant by the composition of these three transformations or @xmath168 where @xmath169 is the image of the point @xmath170 obtained through the spatial tranformation @xmath171 where @xmath172 are the components of the translation vector and @xmath173 ( or in matrix form @xmath174 ) is the improper rotation obtained by composing the reflection @xmath175 and the rotation @xmath176 whose matrices are @xmath177 in appendix d we fully investigate the way the symmetry of eq.([sym ] ) affects electromagnetic chirality and as a result we obtain that such symmetry implies the vanishing of the chirality tensor apart from two specific situations where it imposes restrictions on the chirality tensor structure .
with reference to fig.1 , these cases are : ( a ) the rotation is trivial @xmath178 or the rotation unit vector is orthogonal to the reflection unit vector @xmath179 ; ( b ) the rotation and the reflection are such that @xmath180 . in order to discuss such symmetric situations where
the chirality tensor does not wholly vanish it is convenient to consider the orthonormal basis @xmath181 note that for @xmath182 the rotation is the identity transformation so that the rotation vector @xmath183 is in this case an arbitrary unit vector . in case ( a ) , the dielectric profile is invariant under a pure reflection through a plane orthogonal to @xmath164 or a suitable rotation around an axis belonging to the reflection plane is in order for restoring the dielectric profile after the reflection . as it is shown in appendix d , in this case
the unit vectors of eqs.([uuuu ] ) are such that @xmath184 , @xmath185 and @xmath186 so that the improper rotation @xmath187 is a pure reflection through the plane spanned by the unit vectors @xmath188 and @xmath189 and a rotation around @xmath190 of an angle @xmath191 . in other words case ( a ) deals with the situation where the medium admits a plane of reflection symmetry . in this case
the structure of the chirality tensor allowed by this symmetry is ( see appendix d ) @xmath192 where @xmath193 are four independent complex scalars . from eq.([kappa12 ] ) it is evident that @xmath194 so that if the metamaterial admits a plane of reflection symmetry it is a pseudo - chiral - omega medium . followed by a rotation ( circular arrows ) of an angle @xmath165 around the unit vector @xmath183 is equivalent to a reflection through the plane spanned by the unit vectors @xmath188 and @xmath189 and a rotation around the unit vector @xmath190 of an angle @xmath195 such that @xmath196 .
if the medium dielectric profile is invariant under the composition of the reflection and the rotation , three different situations are relevant .
( a ) the rotation is trivial @xmath178 or @xmath197 so that @xmath198 . ( b )
the rotation and the reflection are such that @xmath180 so that @xmath199 or @xmath200 .
( c ) conditions ( a ) and ( b ) are not satisfied . in case ( c )
the chirality tensor vanishes whereas in cases ( a ) and ( b ) it has a reduced structure.,scaledwidth=48.0% ] in case ( b ) the medium profile is restored , after the reflection , by a specific rotation in such a way that @xmath180 . as it is shown in appendix d , in this case the unit vectors of eqs.([uuuu ] ) are such that @xmath201 , @xmath202 and @xmath186 so that the improper rotation @xmath187 is a reflection through the plane spanned by the unit vectors @xmath188 and @xmath189 and a rotation around the unit vector @xmath190 of an angle @xmath203 . in other words case ( b ) deals with the situation where the medium is left invariant ( modulus a translation ) by a reflection through a plane and a @xmath1 rotation around an axis orthogonal to such plane . in this case
the structure of the chirality tensor allowed by this symmetry is @xmath204.\end{aligned}\ ] ] where @xmath205 are two independent complex scalars . from eq.([kappasa ] ) it is evident that @xmath194 so that if the metamaterial has the symmetry of case ( b ) it is pseudo - chiral - omega medium .
we conclude that if the medium can be superposed to its mirror image the chirality tensor vanishes unless the rotation needed to restore the dielectric profile is around an axis orthogonal to the reflection plane and the rotation angle is either @xmath0 or @xmath1 and in both cases it is a pseudo - chiral - omega medium . as an application of the above symmetry analysis
let us consider two relevant situations where the medium is geometrically achiral and has a chiral electromagnetic response nonetheless .
the first situation is that of a medium which is invariant under translations along an axis , say the @xmath206axis , so that its underlying dielectric profile is @xmath207 .
this implies that the structure is left invariant by a reflection though the plane @xmath208 followed by a rotation of zero angle @xmath182 around any axis orthogonal to the @xmath209-axis , so that we set @xmath210 and @xmath211 . using such unit vectors , the basis vectors of eqs.([uuuu ] )
becomes @xmath212 so that since the medium symmetry is of the kind ( a ) ( see the last paragraph ) , from eqs.([kappa12 ] ) we have @xmath213 so that , in agreement with the considerations of the last paragraph , the metamaterial behaves as a pseudo - chiral - omega medium .
since such geometrically achiral medium is translationally invariant we conclude that eq.([kappa2d ] ) contains the most general expression for the chirality tensor of a two - dimensional medium , i.e. it describes two - dimensional electromagnetic chirality .
this situation also encompasses the case of a planar metamaterial @xcite ( i.e. a very thin metamaterial sheet ) since , after the reflection through a plane parallel to the medium plane , the sheet can be made to superpose the original sheet through a zero - angle rotation and a translation ( which does not affect the chiral tensor structure ) along the axis normal to the medium plane .
the second situation is that of a medium which is invariant under rotations around an axis , say the @xmath206axis , so that its underlying dielectric profile is @xmath214 where @xmath215 .
this implies that the structure is left invariant by a reflection though any plane containing the @xmath209 axis followed by a rotation of zero angle @xmath216 around the @xmath209-axis .
therefore we set @xmath217 , where @xmath218 is an arbitrary angle , and @xmath219 .
using such unit vectors , the basis vectors of eqs.([uuuu ] ) becomes @xmath220 so that since the medium symmetry is of the kind ( a ) ( see the last paragraph ) , from eqs.([kappa12 ] ) we have @xmath221 where @xmath222 , @xmath223 , @xmath224 and @xmath225 are functions of @xmath218 .
the entries of the chirality tensor of eq.([kappa1dd ] ) can not depend on the arbitrary angle @xmath218 so that imposing that each @xmath226 is constant we get @xmath227 and @xmath228 and therefore eq.([kappa1dd ] ) yields @xmath229 therefore , a rotationally invariant metamaterial around the @xmath209-axis behaves as an omega - medium whose chirality tensor depends on a single parameter @xmath230 .
this situation encompasses the case of a one - dimensional metamaterial for which @xmath231 so that we conclude that eq.([kappa1d ] ) contains the most general expression for the chirality tensor of a one - dimensional metamaterial , i.e. it describes one - dimensional electromagnetic chirality .
the description of the effective medium response , i.e. the evaluation of the tensors of eqs.([eff - tens ] ) , is achieved by solving eqs.([eff - eqs ] ) for a specific dielectric distribution and numerical integration is usually unavoidable .
however there is a situation where electromagnetic chirality can be fully discussed without resorting to numerical analysis and it is the case of one - dimensional metamaterials .
let us consider a metamaterial whose dielectric profile is such that @xmath232 , i.e. it is a periodic function of the single cartesian coordinate @xmath233 , whose period @xmath13 is such that @xmath234 for the homogenization theory to be applicable . after introducing the fast spatial coordinates
@xmath235 where @xmath236 is the homogenization parameter , the dielectric profile @xmath237 of eq.([dielectric ] ) is a periodic function of period @xmath6 . in appendix
e we show that for such dielectric distribution @xmath238 , the first of eqs.([eff - eqs ] ) can be analytically solved so that the effective permittivity of the first of eqs.([eff - tens ] ) and the chirality tensor of eq.([kappa1 ] ) can be analytically evaluated and they are @xmath239^{-1 } \end{array } \right ) , \hspace{0.5 cm } \kappa = \left(\begin{array}{ccc } 0 & \eta \kappa_0 & 0 \\ -\eta \kappa_0 & 0 & 0 \\ 0 & 0 & 0 \end{array } \right).\ ] ] where @xmath240^{-1 } \frac{2 \pi } { \lambda^2 } \int_0^\lambda dz_1 \int_0^\lambda dz_2 \frac{\varepsilon(z_1)}{\varepsilon(z_2 ) } \left [ \left(\frac{z_1-z_2}{\lambda}\right ) \right .
\nonumber \\ & & - \left .
\frac{1}{2 } \textrm{sign } \left(\frac{z_1-z_2}{\lambda}\right ) \right].\end{aligned}\ ] ] the effective dielectric tensor in the first of eqs.([1d - eff ] ) coincides with the well - known result of the zeroth - order homegenization theory of layered media .
the structure of the chirality tensor in the second of eqs.([1d - eff ] ) shows that the effective medium behaves as an omega - medium and this agrees with the discussion of the last paragraph , see eq.([kappa1d ] ) , since the considered one - dimensional medium is in particular rotationally invariant around the @xmath209-axis .
it is also remarkable that the parameter @xmath230 of eq.([1d - kappa0-body ] ) governing electromagnetic chirality in the present case can be directly evaluated from the dielectric profile by quadratures .
to sum up we have investigated the electromagnetic response of a metamaterial with non - magnetic inclusions by a multiscale analysis and we have included spatial nonlocality up to the second order .
the resulting description is very simple and compact since the three tensors characterizing the effective medium response are shown to be averages ( on the unit cell ) of suitable fast - varying fields which , in turn , are obtained from the dielectric distribution by solving a small number of electrostatic equations . if second order nonlocal response can be neglected , the medium response can be recast into the standard bianisotropic form and this has allowed us to obtain a compact expression for the chirality tensor of the effective medium .
we have exploited our approach to investigate the relation between geometrical and electromagnetic chirality by proving that the latter is shown , in addition to chiral media , even by an achiral medium whose mirror image can be superposed onto it by means of a @xmath0 or @xmath1 rotation around an axis orthogonal to the reflection plane .
we have deduced , as specific relevant examples , the chirality tensor structure of two - dimensional and one - dimensional media and , for the formers , we have obtained a closed form expression for the chirality parameter which can be evaluated from the inclusions dielectric profile by means of quadratures .
the authors thank the u.s .
army international technology center atlantic for financial support ( grant no .
w911nf-14 - 1 - 0315 ) .
let us consider the equation @xmath241 whose source term @xmath242 has the metamaterial periodicity with vanishing spatial average @xmath243 and whose solution @xmath244 is required to have the metamaterial periodicity .
equation ( [ fund ] ) can be written as @xmath245 where the operator @xmath246 is @xmath247 so that solving eq.([fund ] ) amounts to the problem of inverting the operator @xmath246 on the space of functions having the metamaterial periodicity .
note that the relation @xmath248 holds for any function @xmath249 and therefore , if @xmath249 satisfies the homogeneous equation @xmath250 and it has the metamaterial periodicity we have @xmath251 where green theorem has been used together with the periodicity of both @xmath60 and @xmath249 .
the real and imaginary parts of this equation read @xmath252 and therefore , since we are considering dielectrics without gain for which @xmath253 , we obtain @xmath254 or @xmath255 .
this implies that the function @xmath256 satisfies eq.([fund ] ) and the boundary conditions if @xmath244 does and therefore eq.([fund ] ) and the periodic boundary conditions determines @xmath257 up to an arbitrary constant . setting @xmath258 in eq.([fund ] )
we obtain the poisson equation @xmath259 for which the above considerations holds so that @xmath257 has to be determined with the metamaterial periodicity up to an arbitrary constant .
a straightforward calculation shows that the relation @xmath260,\ ] ] where @xmath246 is the operator defined in eq.([operator ] ) , holds for any pair of functions @xmath257 and @xmath261 so that , if both these functions have the metamaterial periodicity after using green theorem we obtain @xmath262 let us consider a function @xmath244 which has the metamaterial periodicity .
the average @xmath263 can be rewritten as @xmath264 where integration by parts and the periodicity of both @xmath60 and @xmath257 have been used . using eq.([f0 ] )
we obtain @xmath265 so that eq.([ident ] ) with @xmath266 yields @xmath267 using eq.([basic ] ) with @xmath268 , the rank - two tensor @xmath73 can be written as @xmath269 where eq.([ident ] ) has been used and therefore @xmath75
. setting @xmath270 , eq.([basic ] ) becomes @xmath271\end{aligned}\ ] ] where eqs.([f1 ] ) and ( [ q ] ) have been used together with the property @xmath272 . after integrating by parts the first term we get @xmath273
therefore the rank - three tensor @xmath274 becomes @xmath275 where use has been made of the fact that @xmath276 .
setting @xmath277 eq.([basic ] ) becomes @xmath278 where the second of eqs.([r ] ) has been used and integration by parts has been performed .
noting that eq.([q ] ) and the second of eqs.([magneto ] ) can be combined to yield @xmath279 we obtain @xmath280 where the property @xmath281 has been used and integration by parts has been performed . therefore , for the rank - two tensor @xmath121 we have @xmath282 setting @xmath283 , eq.([basic ] ) becomes @xmath284 \nonumber \\ & & - \frac{1}{v } \int_c d^3 { \textbf{r}}\left [ \varepsilon \left ( f_i \partial_r w_{sj } + \delta_{rs } f_i f_j \right ) \right]\end{aligned}\ ] ] where the first of eqs.([r ] ) and eq.([p ] ) have been used together with the property @xmath272 and an integration by parts has been performed . from eq.([f1 ] ) we have @xmath285 which inserted in eq.([second ] ) , after using the property @xmath286 and integrating by parts , yields @xmath287 . \nonumber \\ & & \end{aligned}\ ] ] with the help of this equation , the fourth - rank tensor @xmath122 becomes @xmath288 using eq.([ident ] ) it is straightforward to show that @xmath289 , i.e. the tensor is symmetric with respect to the pair of indices @xmath76 .
after symmetrizing these indices , i.e. setting @xmath290 , and using eq.([f1 ] ) we obtain @xmath291 or , equivalently @xmath292}. \nonumber \\\end{aligned}\ ] ]
in order to show that eqs.([constitutive1 ] ) hold , let us first prove that the relation @xmath293 is satisfied . using eq.([kappa ] ) we have @xmath294 which , after exploiting the relation @xmath295 and the antisymmetric property @xmath150 , becomes @xmath296 the tensor @xmath297 is completely antisymmetric since , using again the antisymmetric property @xmath150 , it is straighforward proving that @xmath298 . since every completely antisymmetric three - rank tensor in a tridimensional space is proportional to the levi - civita tensor , the relation @xmath299 holds and the scalar @xmath300 is given by @xmath301 since the relation @xmath302 holds . using the antisymmetric property @xmath150
we obtain @xmath303 so that @xmath304 which , inserted into eq.([proof ] ) , shows that eq.([alf - kappa ] ) is satisfied . inserting the expression @xmath305 into the second of eqs.([constitutive ] ) and using the relation @xmath306 we obtain @xmath307 which coincides with the second of eqs.([constitutive1 ] ) . inserting the expression @xmath305 into the first of eqs.([constitutive ] ) and using the relation @xmath306 we obtain @xmath308 \nonumber \\ & = & \hat{\textbf{e}}_i \varepsilon_0 \left [ \varepsilon^{(eff)}_{ij } \overline{e'}_j - \frac{1}{k_0 } \kappa_{ki } \epsilon_{krj } \frac{\partial \overline{e'}_j}{\partial x_r } \right]\end{aligned}\ ] ] where eq.([alf - kappa ] ) has been used together with the antisymmetry of the levi - civita tensor . exploiting the first of eqs.([macroscopic - primed ] ) this equation becomes @xmath309 which , using eq.([b ] ) , yields @xmath310\end{aligned}\ ] ] which in turn coincides with the first of eqs.([constitutive1 ] ) .
in order to discuss the impact of the symmetry of the dielectric profile on the chirality tensor , it is essential to investigate preliminary the transformation rule of the fields @xmath107 induced by the dielectric symmetry . differentiating the permittivity @xmath311 with respect @xmath312 and using eq.([sym ] ) we obtain @xmath313 where @xmath314 .
let us now consider the functions @xmath315 for which , using again eq.([sym ] ) , we have @xmath316_{{\textbf{r } } } & = & \frac{\partial}{\partial x_n } \left [ \epsilon({\textbf{r } } ' ) \frac{\partial f_i ( { \textbf{r}}')}{\partial x_n } \right ] \nonumber \\ & = & \tilde{r}_{rn } \tilde{r}_{sn } \left [ \partial_r \left(\epsilon \partial_s f_i \right ) \right]_{{\textbf{r } } ' } \nonumber \\ & = & \left [ \nabla_{\textbf{r}}\cdot \left(\epsilon \nabla_{\textbf{r}}f_i \right ) \right]_{{\textbf{r}}'}\end{aligned}\ ] ] where the relation @xmath317 , stating the orthogonality of the improper rotation @xmath187 , have been used . combining eq.([inv - oper ] ) , the first of eqs.([eff - eqs ] ) evaluated at @xmath318 and eq.([par - ep ] ) we obtain @xmath319 = -\partial_j \epsilon\ ] ] which structurally coincides with the first of eqs.([eff - eqs ] ) and which shows , by the uniquity of its solution , that @xmath320 . from eq.([fp ] ) , we thus conclude that the symmetry of the dielectric permittivity entails for the functions @xmath107 the property @xmath321 from eq.([kappa1 ] ) we note that the chirality tensor depends on the rank - one tensor @xmath322 and rank - three tensor @xmath323 whose geometrical properties have to be separately considered . using this equation , exploting the fact that the improper rotation is orthogonal ( so that @xmath324 ) and noting that the transformation in eqs.([transf ] ) superposes a metamaterial unit cell @xmath20 onto another metamaterial unit cell @xmath325 we have @xmath326 after differentiation , eq.([fpp ] ) yields @xmath327 so that , reasoning as in eq.([vec ] ) we have @xmath328 equations ( [ vec ] ) and ( [ tens ] ) state that the vector @xmath322 and the tensor @xmath323 are geometrical invariants of the improper rotation @xmath187 .
since @xmath187 is an improper rotation ( @xmath329 and @xmath330 ) , it is diagonalizable and its eigenvalues are @xmath331 , @xmath332 , @xmath333 ( where @xmath334 ) . in order to relate @xmath195 to the original reflection and rotation we note that @xmath335 whereas from eqs.([sr ] ) we obtain @xmath336 so that , equating these two expressions , we get @xmath337 the matrix @xmath187 admits three eigenvectors @xmath338 , such that @xmath339 , and they satisfy the orthogonality and completeness relations @xmath340 therefore the matrix @xmath187 admits the spectral decomposition @xmath341 so that , after setting @xmath342 with the help eqs.([mm ] ) , eqs.([vec ] ) and ( [ tens ] ) become @xmath343 where no summation is performed over the indices . from eqs.([spectr ] ) it is evident that all the components of @xmath344 and @xmath345 vanish unless the coefficients @xmath346 and @xmath347 are equal to @xmath348 .
since the allowed values for @xmath346 are @xmath349 and those for @xmath347 are @xmath350 , it is evident that the only possible cases where some of the components of @xmath344 and @xmath345 do not vanish are for @xmath351 . as a consequence these values of @xmath195 , corresponding to classes of original reflections and rotations , leads to nonvanishing chiral tensors .
let us now discuss the various classes of original reflections and rotations corresponding to the allowed values of @xmath195 . in order to simplify the treatment
let us consider the three vectors @xmath352 which are easily seen to have real components and to form an orthonormal and positively oriented basis , i.e. @xmath353 the second of eqs.([vers1 ] ) implies that the vectors of eqs.([uuu ] ) satisfy the relation @xmath354 which is particularly useful for the developing of the present analysis .
* case * @xmath355 . from eq.([psi ] ) we deduce that this case geometrically occurs both for a trivial rotation ( @xmath178 ) and for a non trivial rotation whose unit vector is orthogonal to the reflection unit vector ( @xmath356 , @xmath357 ) .
since a trivial rotation is the identity transformation its rotation axis is arbitrary and therefore even in the first subcase we set @xmath357 . in both subcases , it is straightforward proving that the eigenvectors of @xmath187 are the vectors of eqs.([uuu ] ) , i.e. @xmath358 , and that , for @xmath357 , they reduce to @xmath359 since for @xmath355 we have @xmath360 , @xmath361 , @xmath362 , from eqs.([spectr ] ) we deduce that @xmath363 and @xmath364 if only one or three of the indices @xmath365 are equal to @xmath366 . using eqs.([vers1 ] ) , ( [ vers2 ] ) and ( [ bas1 ] ) , from eqs.([vec - tens ] ) we get after some algebra @xmath367 inserting these quantities into the definition of the chirality tensor of eq.([kappa1 ] ) and noting that @xmath368 since @xmath369 we get @xmath370 where @xmath371 , @xmath372 , @xmath373 and @xmath374 are four scalars which are mutually independent since they depends on ten independent scalars . in matrix form
the above chirality tensor reads @xmath375 * case * @xmath376 . from eq.([psi ] ) we deduce that this case geometrically occurs if the reflection and rotation units vectors and the rotation angle are such that @xmath180 . in this case
it is straightforward proving that the eigenvectors are @xmath377 since for @xmath378 we have @xmath379 , @xmath380 , @xmath362 , from eqs.([spectr ] ) we deduce that @xmath381 and @xmath364 if it does not occur that one of the indices @xmath365 is equal to @xmath366 and the other two are both equal to @xmath348 or @xmath382 . using eqs.([vers1 ] ) , ( [ vers2 ] ) and ( [ bas2 ] ) , from eqs.([vec - tens ] ) we get after some algebra @xmath383 inserting these quantities into the definition of the chirality tensor of eq.([kappa1 ] ) we get @xmath384\end{aligned}\ ] ] where @xmath385 and @xmath386 are two scalars which are mutually independent since they depends on four independent scalars . in matrix form
the above chirality tensor reads @xmath387.\end{aligned}\ ] ] * case * @xmath388 . from eq.([psi ] ) we deduce that this case geometrically occurs if the reflection and rotation units vectors and the rotation angle are such that @xmath389 .
since for @xmath388 we have @xmath390 , @xmath391 , @xmath362 , from eqs.([spectr ] ) we deduce that @xmath381 and @xmath364 if it does not occur that @xmath392 or @xmath393 ( i.e. only the components @xmath394 and @xmath395 survive ) . from eqs.([vec - tens ] ) we directly get @xmath396 so that from the definition of the chirality tensor of eq.([kappa1 ] ) we obtain @xmath397 i.e. the chirality tensor vanishes .
let us consider a metamaterial whose dielectric profile is @xmath238 , i.e. it is dependent on a single cartesian coordinate .
the dielectric profile @xmath238 and its inverse @xmath398^{-1}$ ] can be represented through the fourier series @xmath399^{-1 } & = & \sum_{n=-\infty}^{+\infty } e^{i n k_0 z } b_n\end{aligned}\ ] ] where @xmath400 and @xmath401^{-1}\end{aligned}\ ] ] are their fourier coefficients . due to the generality of the fourier series representation ,
the situation we are here considering allows us to investigate even discontinuous dielectric distributions , a relevant situation occurring when the metamaterial is a stratified layered structure , without explicitly resorting to matching condition on the surfaces of the inclusions .
note also that @xmath402 for the one - dimensional dielectric distribution @xmath238 , the functions @xmath68 satisfying the first of eqs.([eff - eqs ] ) are evidently dependent solely on @xmath209 , i.e. @xmath403 , so that such equations yields @xmath404 which have to be solved with the requirements that the functions @xmath68 have the same periodicity as @xmath238 and have vanishing average @xmath405 .
after an integration eqs.([syst1 ] ) become @xmath406 where @xmath407 are constants . averaging eqs.([syst2 ] ) over the metamaterial period , using eqs.([averages ] ) and noting that the average of their left hand sides are zero since they are derivatives of periodic functions we get @xmath408 which , assuming @xmath409 , imply that @xmath410 and @xmath411 .
accordingly , from the first of eqs.([syst2 ] ) we obtain that @xmath412 and @xmath413 are constants so that , requiring their average to be zero , we get @xmath414 using the fourier espansion of @xmath415 of the second of eqs.([1d - die ] ) , the second of eqs.([syst2 ] ) becomes @xmath416 which , after an integration and the requirement for @xmath417 to have vanishing average , yields @xmath418 combining eqs.([1d - f12 ] ) and ( [ 1d - f3 ] ) we obtain for the functions @xmath68 @xmath419 which are fully determined by the dielectric profile @xmath238 and they are unique as a consequence of the chosen constraints ( periodicity and vanishing average ) . using eq.([1d - fj ] ) , the first of eqs.([eff - qp ] ) becomes @xmath420 which , inserted into the first of eqs.([eff - tens ] ) and neglecting the @xmath5 contribution , yields @xmath421 averaging the relation @xmath422 and using eqs.([1d - die ] ) the relation @xmath423 easily follows so that @xmath424 and eq.([1d - eps1 ] ) becomes @xmath425 or in matrix form @xmath426^{-1 } \end{array } \right)\ ] ] where eqs.([averages ] ) have been used . in order to evaluate the chirality tensor we note that eqs.([1d - fj ] ) and the first of eqs.([1d - die ] ) yield @xmath427 so that , from eq.([kappa1 ] ) we obtain @xmath428 where @xmath429 in matrix form , eq.([1d - kappa ] ) reads @xmath430 inserting eqs.([1d - fourier ] ) into eq.([1d - kappa0 ] ) we obtain @xmath431 after noting that the even contribution in the numerator within the series does not contribute and using the relation @xmath432 $ ] valid for @xmath433 , eq.([1d - kappa0 - 1 ] ) yields @xmath434^{-1 } \frac{2 \pi } { \lambda^2 } \int_0^\lambda dz_1
\int_0^\lambda dz_2 \frac{\varepsilon(z_1)}{\varepsilon(z_2 ) } \left [ \left(\frac{z_1-z_2}{\lambda}\right ) \right .
\nonumber \\ & & - \left .
\frac{1}{2 } \textrm{sign } \left(\frac{z_1-z_2}{\lambda}\right ) \right]\end{aligned}\ ] ] where the first of eqs.([averages ] ) has been used .
aa s. obrien , and j. b. pendry , j. phys .
matter * 14 * , 6383 ( 2002 ) .
d. r. smith , s. schultz , p. markos and c. m. soukoulis , phys .
b * 65 * , 195104 ( 2002 ) .
d. r. smith , d. c. vier , t. koschny , and c. m. soukoulis , phys .
e * 71 * , 036617 ( 2005 ) .
x. chen , b. wu , j. kong , and t. m. grzegorczyk , phys .
e * 71 * , 046610 ( 2005 ) .
c. menzel , c. rockstuhl , t. paul , and f. lederer , phys .
b * 77 * 195328 ( 2008 ) .
e. plum , j. zhou , j. dong , v. a. fedotov , t. koschny , c. m. soukoulis , and n. i. zheludev phys .
b * 79 * 035407 ( 2009 ) . c. r. simovski , and s. a. tretyakov , phys
b * 75 * , 195111 ( 2007 ) .
t. d. karamanos , s. d. assimonis , a. i. dimitriadis , and n. v. kantartzis , photon .
* 12 * , 291 ( 2014 ) .
j. b. pendry , a. j. holden , d. j. robbins , and w. j. stewart ieee trans .
theory tech . * 47 * , 2075 ( 1999 ) .
d. r. smith , d. c. vier , n. kroll , and s. schultz , appl .
. lett . * 77 * , 2246 ( 2000 ) .
d. r. smith , and j. b. pendry , j. opt .
b * 23 * , 391 ( 2006 )
. i. tsukerman , j. opt .
b * 28 * , 577 ( 2011 ) .
o. ouchetto , h. ouchetto , s. zouhdi , and a. sekkaki , ieee trans .
antennas propag .
* 61 * , 4214 ( 2013 ) .
a. ishimaru , s. lee , y. kuga , and v. jandhyala , ieee trans .
antenna propag .
* 51 * , 2551 ( 2003 ) .
p. belov , and c r. simovski , phys .
e * 72 * , 026615 ( 2005 ) .
w. lamb , d. m. wood , and n. w. ashcroft , phys .
b * 21 * , 2248 ( 1980 ) .
a. alu , phys .
b * 84 * , 075153 ( 2011 ) .
m. g. silveirinha , phys .
b * 75 * , 115104 ( 2007 ) .
r. tao , z. chen , and p. sheng , phys .
b * 41 * , 2417 ( 1990 ) .
s. datta , c. t. chan , k. m. ho , and c. m. soukoulis , phys .
b * 48 * , 14936 ( 1993 ) .
p. halevi , a. a. krokhin , and j. arriaga , phys .
lett . * 82 * , 719 ( 1999 ) .
g. p. ortiz , b. e. martnez - zrega , b. s. mendoza , and w. l. mochn , phys .
b * 79 * , 245132 ( 2009 ) .
j. a. reyes - avendao , u. algredo - badillo , p. halevi , and f. prez - rodrguez , new j. phys .
* 13 * , 073041 ( 2011 ) .
j. s. perez - huerta , g. p. ortiz , b. s. mendoza and w. l. mochn , new j. phys .
* 15 * , 043037 ( 2013 ) j. a. reyes - avendao , m. p. sampedro , e. jurez - ruiz , and f prez - rodrguez , j. opt . * 16 * , 065102 ( 2014 ) .
d. felbacq , and g. bouchitt , phys .
lett . * 94 * , 183902 ( 2005 ) .
o. ouchetto , s. zouhdi , a. bossavit , g. griso , b. miara , and a. razek , j. mat .
181 * , 225 ( 2007 ) .
d. felbacq , g. bouchitt , b. guizal , and a. moreau , j. nanophoton .
* 2 * , 023501 ( 2008 ) .
y. chen , and r. lipton , photon .
11 * , 442 ( 2013 ) . c. rizza , and a. ciattoni , phys .
110 * , 143901 ( 2013 ) . c. rizza , and a. ciattoni , opt .
lett . * 38 * , 3658 ( 2013 ) . c. rizza , e. palange , and a. ciattoni , photon
. res . * 2 * , 121 ( 2014 ) .
p. a. belov , r. marques , s. i. maslovski , i. s. nefedov , m. silveirinha , c. r. simovski , and s. a. tretyakov , phys .
b * 67 * , 113103 ( 2003 ) .
m. gerken , and d. a. b. miller , appl . opt . *
42 * , 1330 ( 2003 ) .
m. g. silveirinha , phys .
e * 73 * , 046612 ( 2006 ) . j. elser , v. a. podolskiya , i. salakhutdinov , and i. avrutsky , appl .
lett . * 90 * , 191109 ( 2007 ) .
r. j. pollard , a. murphy , w. r. hendren , p. r. evans , r. atkinson , g. a. wurtz , a.v .
zayats , and v. a. podolskiy , phys .
lett . * 102 * , 127405 ( 2009 ) .
a. v. chebykin , a. a. orlov , a. v. vozianova , s. i. maslovski , yu
. s. kivshar , and p. a. belov , phys .
b * 84 * , 115438 ( 2011 )
. a. a. orlov , p. m. voroshilov , p. a. belov , and y. s. kivshar , phys .
b * 84 * , 045424 ( 2011 ) .
serdyukov , i.v .
semchenko , s.a .
tretyakov , a. sihvola , _ electromagnetics of bi - anisotropic materials : theory and applications . _
( gordon and breach science publishers , amsterdam 2001 ) . c. menzel
, t. paul , c. rockstuhl , t. pertsch , s. tretyakov , and falk lederer , phys . rev .
b * 81 * , 035320 ( 2010 ) m. iwanaga , phys . stat . sol .
( b ) * 245 * , 2684 ( 2008 ) . c. tserkezis , n. papanikolaou , g. gantzounis , and n. stefanou , phys .
b * 78 * , 165114 ( 2008 ) .
r. l. chern , and d. felbacq , phys .
b * 79 * , 075118 ( 2009 ) .
j. c. ginn , and i. brener , phys .
lett . * 108 * , 097402 ( 2012 ) .
l. jelinek , r. marqus , f. mesa , j. d. baena , phys .
b * 77 * , 205110 ( 2008 ) . j. k. gansel , m. thiel , m. s. rill , m. decker , k. bade , v. saile , g. von freymann , s. linden , m. wegener , science * 325 * , 1513 ( 2009 )
. e. plum , j. zhou , j. dong , v. a. fedotov , t. koschny , c. m. soukoulis , n. i. zheludev , phys .
b * 79 * , 035407 ( 2009 ) .
b. wang , j. zhou , t. koschny , c. m. soukoulis , appl .
. lett . * 94 * , 151112 ( 2009 ) a. papakostas , a. potts , d.m .
bagnall , s. l. prosvirnin , h. j. coles , and n. i. zheludev phys .
90 * , 107404 - 1 ( 2003 ) . v.
a. fedotov , p. l. mladyonov , s. l. prosvirnin , a. v. rogacheva , y. chen , n. i. zheludev , phys .
lett . * 97 * , 167401 ( 2006 ) .
r. singh , e. plum , c. menzel , c. rockstuhl , a. k. azad , r. a. cheville , f. lederer , w. zhang , n. i. zheludev , phys .
b * 80 * , 153104 ( 2009 ) .
s. zhang , y. s. park , j. li , x. lu , w. zhang , x. zhang , phys .
lett . * 102 * , 023901 ( 2009 ) .
y. ye , s. he , appl .
96 * , 203501 ( 2010 ) . c. menzel , c. helgert , c. rockstuhl , e. b. kley , a. tunnermann , t. pertsch , and f. lederer , phys .
lett . * 104 * , 253902 ( 2010 ) .
z. li , m. mutlu , e. ozbay , j. opt .
* 15 * , 023001 ( 2013 ) .
r. zhao , j. zhou , th .
koschny , e. n. economou , c. m. soukoulis , phys .
lett . * 103 * , 103602 ( 2009 ) .
s. tretyakov , a. sihvola , and l. jylha , photon .
. appl . * 3 * , 107 ( 2005 ) .
r. l. chern , j. phys .
phys . * 46 * , 125307 ( 2013 ) .
j. b. pendry , science * 306 * , 1353 - 1355 ( 2004 ) . s. tretyakov , i. nefedov , a. sihvola , s. maslovski , c. simovski , j. electromagn .
waves appl . *
17 * , 695 ( 2003 ) .
s. tretyakov , a. sihvola , l. jylha , photon .
. appl . * 3 * , 107 ( 2005 ) .
c. monzon , d. forester , phys . rev .
lett . * 95 * , 123904 ( 2005 ) .
j. a. sanders and f. verhulst , averaging methods in nonlinear dynamical systems ( springer - verlag , berlin , 1985 ) .
landau and e.m .
lifshitz , _ electrodynamics of continuous media . _
( pergamon press , 1960 ) .
f. capolino , _ theory and phenomena of metamaterials _ ( taylor and francis group , boca raton , fl , 2009 ) . | we develop , from first principles , a general and compact formalism for predicting the electromagnetic response of a metamaterial with non - magnetic inclusions in the long wavelength limit , including spatial dispersion up to the second order . specifically , by resorting to a suitable multiscale technique ,
we show that medium effective permittivity tensor and the first and second order tensors describing spatial dispersion can be evaluated by averaging suitable spatially rapidly - varying fields each satysifing electrostatic - like equations within the metamaterial unit cell . for metamaterials with negligible second - order spatial dispersion
, we exploit the equivalence of first - order spatial dispersion and reciprocal bianisotropic electromagnetic response to deduce a simple expression for the metamaterial chirality tensor .
such an expression allows us to systematically analyze the effect of the composite spatial symmetry properties on electromagnetic chirality .
we find that even if a metamaterial is geometrically achiral , i.e. it is indistinguishable from its mirror image , it shows pseudo - chiral - omega electromagnetic chirality if the rotation needed to restore the dielectric profile after the reflection is either a @xmath0 or @xmath1 rotation around an axis orthogonal to the reflection plane .
these two symmetric situations encompass two - dimensional and one - dimensional metamaterials with chiral response . as an example admitting full analytical description ,
we discuss one - dimensional metamaterials whose single chirality parameter is shown to be directly related to the metamaterial dielectric profile by quadratures . |
major western boundary currents in the ocean are meandering jets separating water masses with different physical and biogeochemical characteristics .
the prominent examples are the gulf stream in the atlantic ocean and the kuroshio in the pacific ocean . these and similar `` heat engines '' define the climate in large regions of the planet .
similar jets in the stratosphere play important role in transport and distribution of chemical substances . from the hydrodynamic point of view
, they may be considered as jet flows with running waves of different wave lengths and phase velocities imposed .
the simplest kinematic model of such a flow is a two - dimensional jet of an ideal fluid with a given velocity profile that is perturbed by an amplitude - modulated wave traveling from the west to the east .
the problem of transport and mixing of passive scalars in meandering jets has been considered by many authors in the context of atmospheric and oceanic physics @xcite .
the typical phase portrait ( fig . [ fig1 ] ) consists of two chains of circulations with a zigzag - like jet between them and resembles the phase portrait of a particle in the field of two running waves . in the frame
moving with the velocity of one of the waves , the problem is topologically equivalent to the motion of a periodically perturbed nonlinear physical pendulum that is known to demonstrate chaotic oscillations @xcite .
different aspects of chaotic mixing of passive particles in meandering jets in the atmosphere and the ocean have been studied in the papers @xcite . in the paper we focus on topological and statistical aspects of the chaotic transport and mixing in a specific kinematic model of an eastward meandering jet which has been introduced in refs .
@xcite some years ago .
we are motivated by the desire to get a more deep insight into the evolution of material lines in the flow and to establish a connection between dynamical , topological and statistical characteristics of the flow .
the equations of motion of passive particles advected by a planar incompressible flow is known to have a hamiltonian form @xmath0 with the streamfunction @xmath1 playing the role of a hamiltonian and the coordinates @xmath2 of a particle being canonically conjugated variables .
thus , nonstationary two - dynamical advection is equivalent to a hamiltonian system with one and half degrees of freedom whose phase space coincides with its configuration space .
this property is very useful in visualizing geometric invariant sets in real and numerical experiments .
to be specific we consider a two - dimensional bickley jet with the velocity profile @xmath3 whose argument is modulated by a zonal running wave @xcite .
the streamfunction in the fixed frame reference is the following : @xmath4 where @xmath5 , @xmath6 and @xmath7 are amplitude , wavenumber and phase velocity respectively , @xmath8 is a measure of the jet s width . after introducing the following notations : @xmath9 and @xmath10 we get the advection equations ( [ 1 ] ) in the frame moving with the phase velocity @xmath7 : @xmath11 ) in the frame moving with the meander s phase velocity @xmath7 . ]
the respective streamfunction @xmath12 has three normalized control parameters : @xmath13 , @xmath14 and @xmath15 are the jet s width , meander s amplitude and its phase velocity .
the scaling chosen results in translational invariance of the phase portrait along the @xmath16-axis with the period @xmath17 . the detailed analysis of stationary points and bifurcations of eqs . ( [ 11 ] ) has been done in ref .
stationary points may exist only under the condition @xmath18 .
there are four stationary points , two of them are always stable and the other ones are stable under the condition @xmath19 . if the additional condition @xmath20 is fulfilled , there are additional four unstable saddle points . resuming one gets : 1 .
@xmath21 , there are no stationary points . 2 .
@xmath22 and @xmath23 , there are two centers and two saddles with two separatrices connecting the saddles .
there are two separatrices each of which passes through its own saddle .
the jet between the separatrices is westward .
@xmath24 and @xmath25 , the jet is eastward with the same stationary points as in the case 2 .
@xmath26 , there are eight stationary points and two separatrices .
there are two separatrices , each of which connects two saddle points .
the jet is westward .
@xmath27 and @xmath25 , the jet is eastward with the same stationary points as in the case 4 .
therefore , from the point of view of existence and stability of stationary states , there are three possibilities : 1 ) there are no stationary points ; 2 ) there are four stationary points , two centers and two saddles ; 3 ) there are eight stationary points , four centers and four saddles .
a bifurcation between the first and second regimes consists in arising two pairs `` saddle - center '' .
a bifurcation between the second and third regimes consists in arising two saddles and a center between them instead of one saddle ( a fork - type bifurcation ) .
there is one more bifurcation that does not change the number and stability of stationary points but changes the topology of the flow .
the values of the streamfunction on the separatrices are equal on modulo but of opposite signs .
there is a critical value of the phase velocity @xmath28 under which the separatrices coincide and the respective streamfunction is equal to zero . if @xmath29 , a free flow between the separatrices is westward , whereas with @xmath30 it is eastward .
it is difficult to find @xmath28 analytically but it may be shown @xcite that @xmath31 , if @xmath32 otherwise @xmath33 .
the respective phase portraits are typical with hamiltonian systems with running waves in shear flows @xcite . in dependence on the values of the phase velocity @xmath15
one can get different topologies : a homoclinic connection , a heteroclinic connection and a separatrix reconnection .
being motivated by eastward jet currents in the ocean and atmosphere , we deal in this paper with the case 3 ( see the phase portrait in fig .
streamlines with the streamfunction ( [ 12 ] ) , that is time - independent in the moving frame , are shown in fig .
1 . the plot demonstrates three different regions of the flow : a central eastward jet ( j ) , north and south circulations ( c ) and peripheral westward currents ( p ) .
the centers of the circulations are at critical lines to be defined by the condition @xmath34 , @xmath35 , and they are divided by two separatrices connecting saddle points .
no exchange between the north and south circulations is possible in the unperturbed system .
even the simplest periodic perturbation of the meander s amplitude , @xmath36 , results in spitting stable and unstable manifolds of the saddle points .
there arise stochastic layers with chaotic mixing and transport of water masses fluxes of which depend on the width of a layer , a number of overlapping resonances and other factors . as an example
, we plot in fig .
2 the poincar section of the perturbed flow .
the control parameters throughout the paper are chosen as follows : @xmath37 , @xmath38 , @xmath39 , @xmath40 , @xmath41 and @xmath42 .
due to the zonal and meridian symmetries it is enough to consider mixing in the northern part of the first eastern frame only @xmath43 . in fig .
3a we plot the respective poincar section .
the vortex core , that survives under the perturbation , is immersed into a stochastic sea where one can see 6 small islands belonging to the same resonance .
particles , belonging to these islands , rotate around the elliptic point of the vortex core . zoom of the section nearby two of the islands is shown in fig .
the feature we want to pay attention to is a stickiness to the boundaries of the vortex core and of the islands that is visualized by increasing the density of points near the respective boundaries .
it is a typical phenomenon with hamiltonian systems @xcite which influences essentially the transport of passive particles .
chaotic trajectories for @xmath44 steps .
the parameters are : @xmath40 and @xmath45 .
( b ) for short eastward flights , @xmath46 , the pdf decays exponentially .
( c ) the tails of both the pdfs decay as a power law with the slope @xmath47 obtained with westward ( squares ) long flights @xmath48 . ] without perturbation , the transport properties are very simple : particles either rotate in circulations c or move eastward in the jet j and westward in the peripheral currents p. under a perturbation , the motion in the stochastic layers become extremely sensitive to small variations in initial conditions , and one is forced to use an statistical approach to describe transport .
a commonly used statistical measure of transport is the variance @xmath49 , where the averaging is supposed to be done over an ensemble of particles .
long time behavior of @xmath50 and of the probability density function of particle s displacement @xmath51 may be anomalous with typical chaotic hamiltonian systems @xcite .
when @xmath52 and @xmath51 is a gaussian distribution , the advection is known to be normal with a well - defined diffusion coefficient @xmath53 .
when @xmath54 , with @xmath55 , @xmath56 is either zero or infinite @xcite , and one gets either subdiffusion ( @xmath57 ) or super - diffusion ( @xmath58 ) . if the trajectories are dominated by sticking regions nearby boundaries of islands , where particles spend a long time , subdiffusion results .
superdiffusion occurs when particles in the jet travel long distances between sticking events .
the respective length and time pdfs are expected to be non gaussian .
we will call `` a flight '' any event between two successive changes of signs of the particle s zonal velocity . in this terminology
a sticking consists of a number of flights with approximately equal flight times .
the poincar section of the flow with the perturbation strength @xmath40 and frequency @xmath45 is shown in fig .
the flight pdfs are computed with @xmath59 particles ( initially placed in the first east frame inside a stochastic layer ) up to the time @xmath60 .
the pdf of the lengths of flights is shown in fig .
[ fig4]a for both the directions .
the asymmetry between the eastward ( `` positive '' ) and westward ( `` negative '' ) flights is evident .
both the pdfs can be roughly split into three distinctive regions .
the very short flights with small values of @xmath61 ( @xmath62 ) are supposed to be dominated by sticking to the boundaries of the vortex core and oscillatory islands .
the pdf for eastward flights with the lengths in the range @xmath46 decays exponentially ( fig .
[ fig4]b ) .
the tails of both the pdfs are close to a power - law decay @xmath63 . in fig .
[ fig4]c we estimate the exponent for long westward flights , @xmath48 , with corresponding error by least - square fitting of the straight line to the log - log plot of the data to be @xmath47 . plane , ( b ) dependence of the zonal position on time .
stickiness and flight events are evident with the green chaotic trajectory.,scaledwidth=80.0% ] besides resonant islands with particles moving around the elliptic point in the same frame ( fig .
3 ) , we have found so - called ballistic islands which were situated both in the chaotic sea and in the peripheral jets ( fig .
ballistic modes @xcite correspond to the stable periodic motion of particles from one frame to another .
particles , belonging to a ballistic island , move with a constant zonal velocity from one frame to another .
when mapping their positions at the moments @xmath64 @xmath65 onto the first frame , we see a chain of islands which are visible in fig .
5a . in fig . 5b
a zoom of the ballistic island is shown .
stickiness to the island s border is evident . in fig .
6 we demonstrate two ballistic trajectories corresponding to the ballistic island shown in fig . 5b .
if a particle at @xmath66 is placed inside the island , it travels to the west in a regular way ( the upper blue trajectory in fig . 6a ) .
its zonal position @xmath16 grows linearly with time ( the lower blue trajectory in fig .
6b ) . the lower green trajectory in fig .
6a corresponds to a particle placed initially nearby the border of the same ballistic island from outside .
intermittent flight and sticking events are evident in fig . 6b
( the upper green curve in that figure ) .
of particles that reach the lines @xmath67 after @xmath68 turns around the elliptic points .
@xmath69 is a time particles need to reach the lines @xmath67 .
indices @xmath70 and @xmath71 mean particles moving in the eastward and westward directions , respectively.,scaledwidth=80.0% ] are marked by the respective letters . ]
poincar sections provide good impression about the structure of the phase space but not about geometry of mixing . in this section
we consider the evolution of a material line consisting of a large number of particles distributed initially on a straight line that transverses the stochastic layer at @xmath72 .
a typical stochastic layer consists of an infinite number of unstable periodic and chaotic orbits with islands of regular motion to be imbeded .
all the unstable invariant sets are known to possess stable and unstable manifolds . when time progresses particle s trajectories nearby a stable manifold of an invariant set tend to approach the set whereas the trajectories close to an unstable manifold go away from the set .
because of such a very complicated heteroclinic structure , we expect a diversity of particle s trajectories .
some of them are trapped forever in the first eastern frame @xmath43 rotating around the elliptic point along heteroclinic orbits .
other ones quit the frame through the lines @xmath72 or @xmath73 , and then either are trapped there or move to the neighbor frames ( including the first one ) , and so on to infinity . to get a more deep insight into the geometry of chaotic mixing
we follow the methodology of our works @xcite and compute the time @xmath69 , particles spend in the neighbor circulation zones @xmath74 before reaching the critical lines @xmath75 , and the number of times @xmath68 they wind crossing the lines @xmath76 . in the upper panel in fig .
7 the functions @xmath77 and @xmath78 are shown .
the upper parts of each function ( with @xmath79 and @xmath80 ) represent the results for the particles with initial positive zonal velocities which they have simply due to their locations on the material line at @xmath72 .
these particles enter the eastern frame and may change the direction of their motion many times before leaving the frame through the lines @xmath72 or @xmath73 .
the time moments of those events we fix for all the particles with @xmath81 .
the lower `` negative '' parts of the functions @xmath77 and @xmath78 represent the results for the particles with initial negative zonal velocities ( @xmath82 ) which move initially to the first western frame ( @xmath83 ) .
in fact , @xmath84 and @xmath85 are the time moments when a particle with the initial position @xmath86 quits the eastern or western frames , respectively .
both the functions consist of a number of smooth u - like segments intermittent with poorly resolved ones .
border points of each u - like segments separate particles belonging to stable and unstable manifolds of the heteroclinic structure .
the corresponding initial @xmath87-positions is a set ( of zero measure ) of particles to be trapped forever in the respective frame .
a fractal - like structure of chaotic advection in both the frames is shown in the upper panel in fig .
7 , and its fragments for the first levels are shown in the middle panel for the eastern and the western fragments separately .
particles with even values of @xmath88 quit one the frames through the border @xmath72 , those with odd @xmath88 through the border @xmath73 for the eastern frame and @xmath89 for the western one .
let us consider in detail the fractal - like structure in the eastern frame keeping in mind that the results are similar with any other frame .
the @xmath90-dependence is a complicated hierarchy of sequences of segments of the material line .
following to the authors of the paper @xcite , we call as an epistrophe a sequence of segments of the @xmath91-level , converging to the ends of a segment of a sequence of the @xmath88-th level , whose length decrease in accordance with a law . at @xmath92 we see in fig . 7 an epistrophe with segment s length a , b , c , d and so on decreasing as @xmath93 with @xmath94
. letters a and b in fig .
7 denote the first segments of the epistrophes at the level @xmath95 , whereas d and c the first segments of the epistrophes at the level @xmath96 .
the respective laws for all those epistrophes are not exponential . in fig .
8 we demonstrate fragments of the evolution of the material line in the first eastern frame at the moments indicated in the figure . letters on the line mark the corresponding segments of the @xmath90 and @xmath84 functions in fig .
7 . as an example
, let us explain formation of the epistrophe abcd at the level @xmath92
. with the period of perturbation @xmath97 , a portion from the north end of the material line leaves the frame through its eastern border .
look at the segments a and b at @xmath98 and @xmath99 .
they quit the first frame as a fold through the period @xmath100 .
the other segments c , and d ( not shown in fig .
8) do the same job .
the epistrophe s segments at the odd levels ( @xmath101 ) quit the frame with the period of perturbation @xmath102 one by one being folded ( c and d segments ) .
the folds of the segments of the @xmath103-level are exterior with respects to the folds of the segments of the @xmath104-level .
the following empirical law is valid : @xmath105 , where @xmath106 is a time when the first segments of the epistrophes at the level @xmath103 ( a with @xmath92 ) reach the line @xmath73 , and @xmath107 the respective time for the first segments of the epistrophes at the level @xmath108 ( c and d segments with @xmath96 )
. segments of the epistrophes of the even levels ( @xmath109 ) leave the frame with the period @xmath102 as well but through the border @xmath72 moving to the west .
we show the evolution of some of them at the moments @xmath110 and @xmath111 in fig .
thus , the material line evolves by stretching and folding , and folds quit the frame in both directions with the period of perturbation .
we have treated the problem of mixing and transport of passive particles in a kinematic model of a meandering oceanic jet current from the point of view of dynamical system s theory .
a careful simulation of the hamiltonian equations of advection has shown a complicated character of mixing under a time - dependent perturbation of the meander s parameters . both the oscillatory and ballistic resonant islands and sticking of trajectories to their boundaries have been found .
the transport has benn shown to be anomalous .
the geometry of mixing has been shown to be fractal - like .
the trapping time of advected particles and the number of their rotations around elliptical points have been found to have a hierarchical fractal structure as functions of initial particle s positions .
a correspondence between the evolution of material lines in the flow and elements of the fractal has been established .
this work was supported by the russian foundation for basic research ( project no.06 - 05 - 96032 ) , by the program `` mathematical methods in nonlinear dynamics '' of the russian academy of sciences and by the program for basic research of the far eastern division of the russian academy of sciences .
99 j. sommeria , s.d .
meyers , y.l .
swinney , laboratory model of a planetary eastward jet , nature , 337 ( 1989 ) 5862 .
bower , a simple kinematic mechanism for mixing fluid parcels across a meandering jet , j. phys .
oceanogr . , 21 ( 1989 ) 173 - 180 .
samelson , fluid exchange across a meandering jet , j. phys .
oceanogr . , 22 ( 1992 ) 431440 .
solomon , e.r .
weeks , h.l .
swinney , observation of anomalous diffusion and levy flights in a two - dimensional rotating flow , phys .
lett . , 71 ( 1993 ) 39753978 .
d. del - castillo - negrete , p.j .
morrison , chaotic transport by rossby waves in shear flow , phys .
fluids a , 5 ( 1993 ) 948965 . j.b .
weiss , e. knobloch , mass transport by modulated traveling waves , phys .
rev.a . , 42 ( 1989 ) 25792589 .
k. ngan , t. shepherd , chaotic mixing and transport in rossby - wave critical layers , j. fluid mech . , 334 ( 1997 ) 315351 .
duan , s. wiggins , fluid exchange across a meandering jet with quasi - periodic time variability , j. phys . oceanogr .
, 26 ( 1996 ) 1176 - 1188 .
chirikov , a universal instability of many - dimensional oscillator systems , phys . rep .
, 52 ( 1979 ) 263379 .
zaslavsky , hamiltonian chaos and fractional dynamics , oxford , oxford university press , 2005 .
uleysky , m.v .
budyansky , s.v .
prants , chaotic advection in a meandering jet current , nonlinear dynamics , 1 ( 2006 ) is.2 [ in russian ] .
howard , s.m hohs , stochasticity and reconnection in hamiltonian systems , phys .
rev.a , 29 ( 1984 ) 418421 .
shlesinger , g.m .
zaslavsky , j. klafter , strange kinetics , nature , 363 ( 1993 ) 3138 .
v. rom - kedar and g. zaslavsky , chaos , 9 ( 1999 ) 697705 .
a. iomin , d. gangardt , and s. fishman , phys .
e. 57,(1998 ) 40544062 .
m. budyansky , m. uleysky , s. prants , hamiltonian fractals and chaotic scattering by a topographical vortex and an alternating current , physica d , 195 ( 2004 ) 369 - 378 .
budyansky , m.yu .
uleysky , s.v .
prants , chaotic scattering , transport , and fractals in a simple hydrodynamic flow , zh .
fiz . , 126 ( 2004 ) 1167 - 1179 [ jetp , 99 ( 2004 ) 1018 - 1027 ] .
mitchell , j.p .
handley , b. tighe , j.b .
delos and s.k .
knudson , geometry and topology of escape , chaos , 13 ( 2003 ) 880 - 891 . | we model lagrangian lateral mixing and transport of passive scalars in meandering oceanic jet currents by two - dimensional advection equations with a kinematic stream function with a time - dependent amplitude of a meander imposed .
the advection in such a model is known to be chaotic in a wide range of the meander s characteristics .
we study chaotic transport in a stochastic layer and show that it is anomalous .
the geometry of mixing is examined and shown to be fractal - like .
the scattering characteristics ( trapping time of advected particles and the number of their rotations around elliptical points ) are found to have a hierarchical fractal structure as functions of initial particle s positions .
a correspondence between the evolution of material lines in the flow and elements of the fractal is established .
chaotic advection , meandering jet , fractals 47.52.+j,47.53.+n , 92.10.ty |
sitting- and hanging - drop vapor diffusion methods are among the most popular techniques for protein crystallization ( hampton research co. , 2003 ; mcpherson , 1999 ) . in the sitting drop technique , protein crystals can sediment onto the glass or plastic surface supporting the drop , frequently adhering so strongly that they are damaged during retrieval .
adhered crystals are much more likely to crack during growth because of stresses associated with contact with the support .
the hanging drop method largely eliminates these problems .
crystals sediment away from the supporting surface toward the liquid - air interface ( although some may still nucleate and grow on the supporting surface ) .
freely suspended crystals often have unperturbed facets , show less cracking and have smaller mosaicities , and ( in the absence of protein `` skins '' ) are easily retrieved for diffraction studies .
consequently , hanging drops are preferred for final crystallization trials to obtain the largest , highest - quality crystals for data collection
. however , obtaining well - formed hanging drops with well - defined surface - to - volume ratios can be difficult .
depending on mother liquor composition , the drop can move relatively freely over the hydrophobic siliconized glass cover slides used to minimize drop spreading .
while the cover slide is being inverted , larger drops can slide off entirely , and smaller drops can become distended , changing their shape and rate of equilibration with the well solution .
similar problems occur for drops dispensed onto polymer ( e.g. teflon ) films .
we have demonstrated a simple method to fix the drop shape and position .
chemical patterning of a circular area increases the strength of pinning of the drop s contact line , and thereby defines the drop shape .
drops can be flipped without sliding or permanently distorting , and the drop area and curvature for given drop volume can be customized .
furthermore , the drop size and shape for a given drop volume - which determines the equilibration time with the well - is now much more reproducible .
this eliminates an important variable factor affecting nucleation rates and subsequent growth in both hanging and sitting drop vapor diffusion growth .
the boundary of a liquid drop on the surface of a flat substrate is a line separating liquid , solid ( substrate ) and vapor ( air ) , called the contact line .
the angle between the tangent to the vapor - liquid interface at the contact line and the substrate gives the contact angle @xmath0 ( figure 1 ) . on a horizontal , perfectly uniform substrate
, the contact angle would have a unique value determined by the liquid - vapor , liquid - substrate and vapor - substrate surface energies .
if such a substrate were tilted from the horizontal , the drop would immediately slide off . on a real substrate ,
the contact line is pinned by interaction with disorder at the substrate surface , so that the contact angle is no longer unique but depends on the drop s history ( e.g. , how it was dispensed and how it has been tilted ) . the strength of the contact line pinning is reflected in the range of contact angles that can be supported before the contact line and thus the drop depin and move .
when a substrate supporting a drop is rotated from the horizontal , the drop initially distorts while its contact line remains pinned and fixed , as shown in figure 1(a , b ) .
as the substrate is rotated toward the vertical , the net of the gravitational and surface tension forces may locally exceed the pinning force .
the contact line may then become locally unstable and distort from its initially circular form , producing a `` downhill '' spreading of the drop .
upon further rotation , the gravitational force may exceed the maximum pinning force on the drop as a whole .
the entire drop then becomes unstable and slides off the substrate . to understand this process more formally ,
consider a simplified , essentially two - dimensional ( 2d ) model shown in figure 1(b ) ( macdougall and ockent , 1942 ; frenkel , 1948 ) , which can be extended to three dimensions ( furmidge , 1962 ; dussan and chow , 1983 ) .
we assume the drop s contact line is pinned by some specific but isotropic interaction with the substrate . in this simplified geometry
there are only two contact angles @xmath1 and @xmath2 defining the angle of the tangent to the air - liquid interface at the advancing ( downhill ) and receding ( uphill ) edges of the contact line . for a true 3d drop as in figure 1(a ) , the contact angle @xmath0 varies along the ( initially circular ) contact line between extreme values at its uphill and downhill edges ( brown et al.,1980 ; rotenberg et al .
, 1984 , el sherbini and jacobi , 2004 ) . when the substrate is horizontal , @xmath3 .
when the substrate is tilted from the horizontal by an angle @xmath4 , the contact line remains pinned , the advancing contact angle @xmath1 increases and the receding contact angle @xmath2 decreases . for some critical tilt @xmath5 ( corresponding to values of @xmath6 and @xmath7 ) , a sufficiently large drop will depin and slide down the substrate . balancing the downhill component @xmath8 of the force due to gravity @xmath9 with the uphill force due to capillarity @xmath10
provides a boundary condition that must be satisfied by a stationary pinned drop ( frenkel , 1948 ) .
the drop remains stationary and pinned if for a given inclination , drop mass , and contact conditions we are able to find a drop shape consistent with this boundary condition ( frenkel , 1948 ; lawal and brown , 1982 ) . beyond a critical inclination angle @xmath11 ,
no solution can be found and @xmath12 .
the net capillary force is very roughly given by @xmath13 , where @xmath14 is the characteristic width of the drop , @xmath15 is the vapor - liquid surface tension , and @xmath16 .
analogous to a static friction force , @xmath17 increases as the tilt angle increases so as to keep the drop stationary .
the maximum value @xmath18 that can be sustained by the contact line pinning is called the contact angle hysteresis ( de gennes , 1985 ) . for @xmath19 ,
the drop slides and @xmath17 depends upon the flow conditions ( dussan and chow , 1983 ) .
the downhill component of the drop s weight @xmath9 is @xmath20 , where @xmath21 is the density of the liquid , @xmath22 is acceleration due to gravity , and @xmath23 is the drop volume . for the 2d drop of figure 1(b ) , @xmath24 where @xmath25 is the drop s diameter and @xmath26 is its characteristic height . combining the formulas for @xmath10 and @xmath8 yields the condition for drop stability @xmath27 on a surface inclined at angle @xmath4 : @xmath28 for any angle @xmath4 between 0 and @xmath11
the drop is able to adjust its surface shape by varying the contact angle values @xmath1 and @xmath2 and thus @xmath17 .
the last contact angles @xmath29 and @xmath30 the drop reaches at @xmath11 are determined by the contact angle hysteresis @xmath31 . at that critical point
, @xmath32 corresponds to a surface energy difference that is an invariant of the drop shape . for hanging drop crystallization , we are particularly interested in the case @xmath33 when the drop is oriented vertically .
for such a drop to be stable against sliding , @xmath34 . a drop satisfying this criterion is absolutely stable .
it may still move , however , if accelerations during flipping or other handling - which produce forces on the drop that add to the gravitational force - are large enough .
the stability criterion depends on the properties of the drop , which are affected by the presence of protein .
proteins have hydrophobic and hydrophilic parts and thus can behave as surfactants ( mobius and miller , 1998 ) .
protein present at the vapor / liquid interface decreases the solution s surface tension .
in addition , protein adsorption on the substrate surface can change the contact angles , surface pinning properties , and contact angle hysteresis . in protein crystallization
, we often want to minimize the drop surface - to - volume ratio to produce slow equilibration with well solutions favorable for nucleation . for hanging drop crystallization
, we would also like to maximize the stability of the drop , equivalent to obtaining the largest possible drop volume that is stable at @xmath33 .
increasing the drop volume will decrease the drop surface - to - volume ratio until drop flattening by gravity becomes important . increasing the contact angle hysteresis ( i.e.
, the strength of contact line pinning ) will increase the maximum size of an absolutely stable drop . by chemically patterning a substrate to produce hydrophobic and hydrophilic areas so as to increase the strength of contact line pinning and @xmath31 , we show in the following that the maximum absolutely stable drop volume can be increased .
strong pinning of the contact line at the hydrophobic / hydrophilic boundary also yields much more accurate control over drop shape and surface - to - volume ratio .
six times recrystallized and lyophilized hen egg white lysozyme from seikagaku america ( falmouth , ma ) was dissolved in a sodium acetate ( naac ) aqueous buffer , prepared by adding concentrated acetic acid to a 50 mm sodium acetate solution to adjust the ph to 5.0 , where m=[mol / l ] .
protein concentration was measured using a spectronic genesys tm 5 spectrophotometer ( spectronic instrument , ny ) .
solution surface tension as a function of protein concentration was measured using the pendant drop counting method .
a pipette tip is filled with protein solution and then held vertically , and the number of drops that fall from the tip in a given time and the total mass of these drops is measured . using these values , the drop radius
as it emerges from the tip , and the liquid density , the vapor - liquid surface tension can be calculated .
for the protein solutions studied the drop radius was measured to be roughly @xmath35 larger than the external tip diameter .
siliconized and initially flat glass slides hr3 - 231 with diameter 22 mm were purchased from hampton research ( laguna niguel , ca ) . on a freshly unpackaged slide ,
a 40 @xmath36l buffer drop formed a reproducible contact angle of @xmath37 . to increase contact angle hysteresis and thus drop stability under inclination , a hydrophilic circular region was fabricated on some of these hydrophobic siliconized glass slides .
a drop of 1 m naoh was placed on the surface at room temperature , and then the slide was heated on a hot plate in air to @xmath38c until all liquid evaporated .
the base removes the silanol layer and reacts with the glass , producing water - soluble silicates .
these were washed away using a jet of distilled deionized water , and the slide then dried using pure @xmath39 gas .
naoh drop volumes of 25@xmath36l , 50@xmath36l , 75@xmath36l , 100@xmath36l were used to produce etched areas with diameters of @xmath40 mm , @xmath41 mm , @xmath42 mm , and @xmath43 mm . following this treatment ,
the etched areas were highly hydrophilic , with contact angles for pure water of less than @xmath44 .
when a drop was dispensed onto the etched region , the large difference in contact angle between the etched and unetched regions strongly pinned the drop s contact line at the boundary between these regions .
this strong pinning produced a contact angle hysteresis for pure water of @xmath45 , compared with value of @xmath46 for a uniform siliconized glass slide .
the patterned and unpatterned slide surfaces were analysed by contact mode afm ( di multimode iii , santa barbara , ca ) using nsc 1215 tips from mikromasch .
a typical surface profile in figure 3 shows that the treated area is etched to a depth of about 1 @xmath36 m . to determine drop stability on a given unpatterned or patterned slide ,
a drop was dispensed onto a horizontal slide using a 100 @xmath36l micropipette ( pipetman co. , france ) .
the slide was then slowly rotated in @xmath47 steps , allowing roughly one minute after each rotation to allow transient relaxations of the drop to die out .
the contact angles of the drop were measured using a vertical goniometer , and its shape and position recorded using a digital camera .
dispensed drop volumes were accurate to @xmath48 , and tilt and contact angle measurements were accurate to @xmath49 .
each measurement on an `` unpatterned slide '' used a fresh slide , and each measurement on a `` patterned slide '' used a new fabricated slide , to avoid surface contamination problems between measurements .
drop stability on patterned and unpatterned slides was investigated as a function of drop orientation angle , drop volume , drop area , and protein concentration .
as the inclination angle was incremented upward , the contact angles and contact line position varied around the drop circumference , in part through local instabilities ( similar to avalanches ) that caused abrupt jumps in contact line position . for larger drops ,
the drop became absolutely unstable at a critical angle @xmath50 and slid off the slide , as shown in figures 2 and 4 . to construct a drop stability phase diagram , the patterned area ( radius ) and protein concentration
were fixed , and the critical angle measured as a function of drop volume . repeating these measurements with different protein concentrations and patterned areas provided a complete characterization of drop stability .
figure 4 shows a typical drop stability diagram for patterned and unpatterned slides with fixed patterned radius and fixed protein concentration . for a given inclination @xmath51 ,
a drop remains stable ( even though its contact line may undergo local displacements around the drop s circumference ) for volumes up to a critical volume @xmath52 . drops smaller than the critical volume @xmath53 will be absolutely stable ; they will not slide off even at @xmath33 . drops larger than @xmath54 become unstable and slide off at a critical inclination @xmath55 as follows from equation 1 . for the conditions in figure 4 ,
substrate patterning increases the volume range of absolute stability by @xmath56 . for hanging drop crystallization , we want to maximize the range of volumes for which a drop is absolutely stable and thus the critical volume @xmath57 .
@xmath54 depends on the protein concentration and on the area of the patterned region , which determines the drop s area .
figure 5 shows how the critical volume for absolute stability @xmath57 varies with protein concentration .
drop volumes below each curve are absolutely stable . for both unpatterned and patterned slides ,
the critical volume decreases beyond a protein concentration of roughly 1 mg / ml .
this behavior is consistent with the measured increase in drop surface tension with increasing protein concentration ( mobius and miller , 1998 ) . for pure , protein - free buffer on an unpatterned slide ,
the critical volume @xmath54 is 12.5 @xmath36l , a factor of 3.2 smaller than for the smallest protein concentration ( 0.01 mg / ml ) shown by solid circles in figure 5 .
thus , variation of @xmath54 with protein concentration on unpatterned slides is non - monotonic .
this complex behaviour likely results because protein adsorbs to the slide surface , modifying the contact angle and contact angle hysteresis .
even very small solution concentrations of protein are sufficient to modify these properties . on the other hand ,
the critical volume @xmath54 for the patterned slides ( open circles ) is constant between 0 mg / ml and 0.01 mg / ml , indicating that contact line pinning dominates the drop - substrate interaction . over the whole concentration range
, patterning provides large increases in the critical volume for absolute stability , from @xmath58 at c=0.01 mg / ml to @xmath59 at 100 mg / ml . at concentrations that yield the largest protein crystals ( near 2 mg / ml ) , the increase is @xmath60 .
figure 6 shows how the critical volume @xmath57 for absolute stability varies with the drop diameter @xmath61 ( measured when the slide is horizontal ) for pure buffer and for solutions with different protein concentrations . to simplify representation of the data , figure 6 uses dimensionless variables @xmath62 and @xmath63 defined by the formulas @xmath64 and @xmath65 , respectively . here
@xmath36l and @xmath67 mm are the critical volume and corresponding wetting diameter of the drop measured for protein - free buffer solution on an unpatterned , hydrophobic siliconized slide , indicated by point a. points b and c are also for unpatterned slides , but with protein concentrations of 1 mg / ml and 100 mg / ml , respectively . at 1 mg / ml ( point b ) , the critical volume is 41.4 @xmath36l , 3.3 times larger than for protein - free drops .
increasing the protein concentration to 100 mg / ml ( point c ) decreases the critical volume to 32.5 @xmath36l , 2.6 times larger than for protein - free drops . on patterned slides , the wetted drop diameter is equal to the patterned diameter , and can be made much larger than on unpatterned slides . for protein - free buffer ( solid circles in figure 6 ) ,
increasing the wetted diameter from 3.5 mm to 9.5 mm increases the critical volume by a factor of 4.6 . for 1
mg / ml and 100 mg / ml protein , increasing the diameter from 6.1 and 5.1 mm , respectively , to 9.5 mm increases the critical volume by roughly @xmath68 .
this increase comes at the cost of a factor of 0.94 decrease in drop surface to volume ratio .
we have demonstrated that the overall stability of drops on an inclined flat substrate can be significantly increased by patterning the substrate surface so as to strongly pin the drop s contact line .
this allows larger drops to be used in hanging drop crystallization .
more important , patterning precisely defines the drop shape and thus its surface - to - volume ratio , independent of how the drop is dispensed and flipped .
patterned substrates should thus provide more repeatable drop equilibration rates with well solutions and thus more reproducible crystal nucleation and growth .
the largest increases in drop stability and volume occur for protein - free solutions .
this suggests that surface patterning can be used to define `` well '' drops on the same substrate as the protein drop , and with volumes 10 - 100 times the protein drop volume .
consequently , with surface patterning wells can be eliminated , yielding a crystallization platform that can be used in any orientation .
this work was funded by the nasa ( nag8 - 1831 ) and by the nih ( r01 gm65981 - 01 )
. 99 brown , r. a. , orr , f. m. and scriven , l.e .
j. colloid and interface science , 73 , 76 - 87 .
. research tools .
13(1 ) , hampton research co. de gennes , p.g .
57 , 827 - 863 .
dussan , v. , e.b . and chow , r.t.p .
j. fluid mech .
137 , 1 - 29 .
el sherbini , a. i. and jacobi , a.m. ( 2004 ) .
j. colloid and interface science , 273 , 556 - 565 .
frenkel , y. i. ( 1948 ) .
( ussr ) , 18 , 659 - 667 ; english translation available at http://arxiv.org/abs/physics/0503051 .
furmidge , c. g. l. ( 1962 ) .
j. colloid sci .
17 , 309 - 324 .
lawal , a. and brown , r.a .
j. colloid and interface science , 89 , 346 - 352 .
macdougall , g. and ockent , c. ( 1942 ) .
london ser .
a , 180 , 151 - 173 .
mcpherson , a. ( 1999 ) .
crystallization of biological macromolecules .
new york : cold spring harbor laboratory press .
mobius , d. and miller , r. ( 1998 ) . editors .
proteins at liquid interfaces , amsterdam : elsevier .
rotenberg , y. , boruvka , l. , and neumann , a. w. ( 1984 ) .
j. colloid and interface science , 102 , 424 - 434 .
figure 1 . + photograph of a protein - containing drop on a siliconized glass slide .
( b ) simplified 2d model for a pinned drop on a flat inclined substrate .
+ figure 2 .
+ effect of substrate patterning on stability of protein - free buffer drops .
the numbers in each frame indicate the tilt angle @xmath4 in degrees .
substrates are hydrophobic siliconized glass slides .
patterned slides have circular hydrophilic areas with diameters of 8 mm . + figure 3 . + afm image of the boundary between untreated and naoh - treated parts of the slide .
the average etched depth of the naoh - treated hydrophilic area is about 1 @xmath36 m . + figure 4 .
+ stability diagram for drops on unpatterned siliconized slides ( open squares ) and siliconized slides patterned with an 8 mm diameter hydrophobic region ( solid squares ) , for drops containing 1 mg / ml lysozyme in 50 mm naac buffer at ph=5 .
the critical volumes beyond which drops are unstable at @xmath33 are @xmath69 @xmath36l and @xmath70 @xmath36l for the unpatterned and patterned slides , respectively
. drops with smaller volumes are absolutely stable .
dashed lines represent the function @xmath71 . + figure 5
. + critical drop volume @xmath57 for stability on vertically oriented substrates versus protein concentration .
the open and solid circles show data for unpatterned siliconized slides and siliconized slides patterned with a 9.5 mm diameter hydrophobic region , respectively .
the approximate concentration range over which crystal growth can be induced by addition of nacl precipitant is indicated .
+ figure 6 .
+ enhancement of drop critical volume @xmath62 by patterning a circular hydrophilic region of diameter @xmath63 .
the points a , b , and c correspond to drops on unpatterned slides .
the raw data is scaled by the value of the point a ( @xmath72 = 12.5 @xmath36l and @xmath73 = 3.5 mm ) as described in the text .
solid circles correspond to protein free buffer solution , and open circles and squares correspond to protein concentrations of 1mg / ml and 100mg / ml , respectively . | motion of protein drops on crystallization media during routine handling is a major factor affecting the reproducibility of crystallization conditions .
drop stability can be enhanced by chemical patterning to more effectively pin the drop s contact line . as an example
, a hydrophilic area is patterned on an initially flat hydrophobic glass slide .
the drop remains confined to the hydrophilic area , and the maximum drop size that remains stable when the slide is rotated to the vertical position increases .
this simple method is readily scalable and has the potential to significantly improve outcomes in hanging and sitting drop crystallization . |
in @xcite , ap , slovk , and souek construct sequences of invariant differential operators on _ parabolic geometries _ of any type @xmath0 , one for each finite - dimensional representation @xmath1 of @xmath2 .
( here , @xmath2 is a semisimple lie group and @xmath3 a parabolic subgroup . )
these sequences are known as _ bernstein - gelfand - gelfand _ ( bgg ) sequences since , for the homogeneous model @xmath0 of such a geometry , these sequences are complexes , which are dual to a parallel construction due to these authors @xcite on the level of verma modules . in @xcite calderbank and diemer
simplify the construction of bgg sequences in @xcite .
in addition they provide @xcite , for regular parabolic geometries , alternative bgg sequences , which only coincide with the ones in @xcite if the geometry is _ torsion - free_. the latter sequences not only appear to be more natural , they also have the advantage that if @xmath1 is taken to be the trivial representation , then they form complexes , providing fine resolutions of the locally constant sheaf @xmath4 ( as one sees by suitably modifying ( * ? ? ? * proposition 5.5(iv ) ) ) . for the sequences of @xcite
this is only true if the geometry is torsion - free and , in this case , the two sequences are anyway the same . in combination with the construction of canonical cartan connections given in @xcite
, this shows that one can find alternatives to the de rham resolution for any parabolic geometry defined in terms of a regular infinitesimal flag structure @xcite .
a hallmark of these resolutions is that the ranks of the bundles involved are diminished as compared to the de rham complex .
the price one pays is that the operators may be higher than first order .
the construction of these resolutions in @xcite , entails firstly constructing the cartan connection as described in @xcite and this is not at all straightforward . in this article
we present some examples constructed by a more elementary route .
as we show , our method extends to certain non - parabolic geometries , namely arbitrary contact and symplectic geometries .
we shall use the spectral sequence of a filtered complex @xcite without comment and merely as a replacement for tedious diagram chasing .
for our first example we shall construct the rumin complex @xcite .
it is defined on an arbitrary contact manifold but , for simplicity , we shall present the @xmath5-dimensional case , which is typical .
so let @xmath6 be a @xmath5-dimensional smooth manifold with @xmath7 a contact distribution .
equivalently , the contact structure may be defined by @xmath8 , a line sub - bundle of the bundle of @xmath9-forms @xmath10 .
if we define a rank @xmath11 vector bundle @xmath12 as the quotient @xmath13 , then there are induced short exact sequences @xmath14 and the spectral sequence of the de rham complex filtered in this way reads , at the @xmath15-level , @xmath16[r]{${\mathcal{l}}$}\uparrow&&\uparrow&&\uparrow\\ & & 0&&l&&\lambda_h^1\otimes l&&\lambda_h^2\otimes l & & \lambda_h^3\otimes l&&\lambda_h^4\otimes l \end{array}$ } } \put(0,0){\vector(1,0){40 } } \put(0,0){\vector(0,1){45 } } \put(30,5){$p$ } \put(3,35){$q$ } \end{picture}\]]where @xmath17 is the composition @xmath18 .
the leibniz rule shows that @xmath17 is linear over the functions and is , therefore , a homomorphism of vector bundles .
it is called the _
levi form_. by definition of contact manifold , the range of @xmath17 is non - degenerate as a skew form on @xmath19 , defined up to scale .
equivalently , we can choose local co - framings @xmath20 with @xmath21 a section of @xmath22 such that @xmath23 notice that @xmath24whence the @xmath15-differential @xmath25 is an isomorphism of vector bundles
. similar reasoning shows that @xmath26 is surjective . hence ,
at the @xmath27-level we obtain @xmath28 & & 0&&0&&\!\!0\!\!&&\lambda_{h\perp}^2\otimes l & \!\!\!\stackrel{d_\perp}{\longrightarrow}\!\!\ ! &
\lambda_h^3\otimes l & \!\!\!\stackrel{d_\perp}{\longrightarrow}\!\!\!&\lambda_h^4\otimes l \end{array}$ } } \put(0,0){\vector(1,0){40 } } \put(0,0){\vector(0,1){45 } } \put(30,5){$p$ } \put(3,35){$q$ } \end{picture}\]]and deduce that there is a complex @xmath29where @xmath30 denotes the sub - bundle of @xmath31 , trace - free with respect to the levi - form . the operator @xmath32 is second order and , because the spectral sequence converges to the local cohomology of the de rham complex , it follows that this complex is exact on the level of sheaves .
already , the rumin complex goes beyond parabolic geometry .
notice that , although a convenient co - framing was chosen to perform some calculation , the construction itself and the resulting complex are independent of any such choice .
this is a repeated theme in this article .
in this section we shall be concerned with a smooth @xmath11-manifold @xmath6 equipped with a generic distribution @xmath7 of rank @xmath33 .
genericity entails that @xmath34 $ ] has rank @xmath35 and that @xmath36=tm$ ] .
dually , if we let @xmath37 and @xmath38^\perp$ ] then the @xmath9-forms are filtered @xmath39 by the line - bundle @xmath22 and rank @xmath33 bundle @xmath40 . in fact , there is a canonically defined finer filtration constructed as follows .
one easily checks that the levi form @xmath41 , defined as the composition @xmath42 , is a surjective homomorphism of vector bundles with @xmath22 as kernel .
it follows that the other levi form , defined as the composition @xmath43 has range in the kernel of @xmath44 .
however , the short exact sequence @xmath45 identifies this kernel as @xmath46 . in other words
, we have a canonically defined inclusion @xmath47 the range of which defines a line sub - bundle @xmath48 of @xmath12 .
the result is that we can write @xmath49 meaning that @xmath10 is filtered with composition factors being line bundles as indicated ( ordered so that @xmath50 is a canonical quotient and @xmath22 is a canonical sub - bundle ) .
all in all , if we write @xmath51 for @xmath50 and untangle the identifications found above , then we conclude that @xmath52 equivalently , we can work locally with @xmath53 , an _ adapted _ co - framing such that @xmath54 noting that the freedom in such a co - framing comprises exactly the triangular endomorphisms of the filtration ( [ filtration ] ) , where @xmath55 \xi+k={\mathrm{span}}\{\omega^1,\omega^2\,\omega^3\}.\end{array}\ ] ] so far , this is the structure of an _
engel manifold_. as with the rumin complex , it is clear that the first order operator @xmath56 defined as the composition @xmath57 has the locally constant functions as its kernel .
we now seek differential conditions on a section of @xmath12 in order that it be in the range of the operator @xmath58 . starting with any @xmath9-form @xmath59 , @xmath60 the structure equations ( [ structure ] ) show that @xmath61 is well - defined and one easily checks that the equivalence class @xmath62\in \frac{{\mathrm{span}}\{\omega^2\wedge\omega^4,\omega^2\wedge\omega^3 , \omega^1\wedge\omega^4,\omega^1\wedge\omega^3,\omega^1\wedge\omega^2\ } } { { \mathrm{span}}\{\omega^2\wedge\omega^3 , \omega^1\wedge\omega^4,\omega^1\wedge\omega^3,\omega^1\wedge\omega^2\ } } \cong\lambda^2\xi\ ] ] depends only on the equivalence class @xmath63\in\lambda_h^1 $ ] and is independent of choice of co - framing .
we have a well - defined second order differential operator @xmath64\stackrel{{\mathcal{p}}}{\longmapsto } [ p\,\omega^2\wedge\omega^4]\in\lambda^2\xi,\ ] ] giving what we shall call the _ primary _ obstruction to @xmath63 $ ] being in the range of @xmath58 . in a chosen co - frame
, one can easily proceed to find a _
obstruction @xmath65 as follows .
define @xmath66 by ( [ primary_recipe ] ) and then @xmath67 if @xmath61 vanishes , then @xmath68 so @xmath69 vanishes . if , in addition @xmath65 vanishes , then @xmath70 so @xmath71 vanishes .
hence , if both @xmath61 and @xmath65 vanish , then @xmath72 . by the poincar lemma
, it follows that @xmath63 $ ] is locally in the range of @xmath58 , as required .
if the primary obstruction vanishes , then the equivalence class @xmath73\in \frac{{\mathrm{span}}\{\omega^1\wedge\omega^3,\omega^1\wedge\omega^2\ } } { { \mathrm{span}}\{\omega^1\wedge\omega^2\ } } \cong\lambda\xi^3\ ] ] is independent of choice of co - framing .
otherwise , the change @xmath74 induces severe complications with @xmath65 changing by @xmath69 and its derivatives .
if one wants to avoid these complications , it suffices to prohibit ( [ badchange ] ) to arrive at an invariantly defined differential operator @xmath75 whose kernel is locally the range of @xmath58 .
more precisely , we may eliminate ( [ badchange ] ) by choosing a complement to the line sub - bundle @xmath76 .
in other words , we choose a splitting @xmath77 .
an adapted co - framing yields such a splitting and , conversely , a fixed choice of splitting restricts the choice of adapted co - framings precisely by preventing the addition of any multiple of @xmath78 to @xmath79 .
the forms on an engel manifold endowed with this extra structure are filtered as follows .
@xmath80 \lambda^2&=&\lambda\xi+(\lambda^2\xi\oplus\lambda\xi^2)+ ( \lambda^2\xi^2\oplus\lambda\xi^3)+\lambda^2\xi^3\\[2pt ] \lambda^3&=&\lambda^2\xi^2+\lambda^2\xi^3+(\lambda^3\xi^3\oplus\lambda^2\xi^4 ) \end{array}\ ] ] and the spectral sequence of the de rham complex filtered in this way reads , at the @xmath15-level , @xmath81the @xmath15-differentials are easily computed in our adapted co - frame . for example
@xmath82deals with the two rightmost differentials . consequently ,
at the @xmath27-level we obtain @xmath83 & \!&0&&0&&\lambda^2\xi&&0&&0&&0\\[4pt ] & \!&0&&0&&0&&\lambda\xi^3&&0&&0\\[4pt ] & \!&0&&0&&0&&0&&0&&\lambda^3\xi^3\oplus\lambda^2\xi^4&\!&\lambda^4\ , .
\end{array}$ } } \put(0,0){\vector(1,0){40 } } \put(0,0){\vector(0,1){45 } } \put(30,5){$p$ } \put(3,35){$q$ } \end{picture}\ ] ] the bundles @xmath84 and @xmath85 may be identified with @xmath12 and @xmath86 , respectively
. the line bundles @xmath87 and @xmath88 combine to give a rank @xmath33 vector bundle @xmath89 but , in fact , this bundle canonically splits as can readily be seen in our adapted co - frame : @xmath90 independent of choice of co - frame .
we conclude that there is a complex of differential operators ( cf .
@xcite ) @xmath91 resolving the locally constant sheaf @xmath4 . following through the spectral sequence more explicitly as a diagram chase shows that @xmath92 is given by our previous recipe .
we shall see later in [ engelrevisited ] , that ( [ fullengel ] ) is a bgg complex for an appropriate parabolic geometry .
since a contact manifold with no extra structure is not a parabolic geometry , the rumin complex lies outside the realm of parabolic geometry .
nevertheless , there is a parabolic geometry in which the rumin complex finds its genesis .
let us denote by @xmath93 the simple lie group of linear automorphisms of @xmath94 preserving a fixed non - degenerate symplectic form .
viewing the @xmath95-sphere @xmath96 as @xmath97 ( i.e. the space of rays emanating from the origin in @xmath98 ) , the group @xmath99 acts smoothly and transitively on @xmath96 .
the stabiliser subgroup @xmath100 of this action is parabolic .
parabolic geometries modelled on this particular homogeneous space @xmath101 are known as _ contact projective _ @xcite . in any case , when viewed in this way , the sphere @xmath96 inherits a @xmath2-invariant contact structure from the symplectic form on @xmath98 . as in [ rumin ] , let us now consider the case @xmath102 . adopting the notation from @xcite , this homogeneous space is written as ( 42,5 ) ( 20,.1)(1,0)16 ( 20,3.3)(1,0)16 ( 20,1.5)(0,0)@xmath103 ( 36,1.5)(0,0)@xmath103 ( 28,1.5)(0,0)@xmath104 ( 4,1.5)(0,0)@xmath105 ( 4,1.5)(1,0)15 and the bernstein - gelfand - gelfand complex corresponding to the trivial representation of @xmath106 is @xmath107 \mbox { } \hspace{100pt}\to \begin{picture}(42,5 ) \put(20,.1){\line(1,0){16 } } \put(20,3.3){\line(1,0){16 } } \put(20,1.5){\makebox(0,0){$\bullet$ } } \put(36,1.5){\makebox(0,0){$\bullet$ } } \put(28,1.5){\makebox(0,0){$\langle$ } } \put(20,8){\makebox(0,0){$\scriptstyle 0 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 1 $ } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(4,1.5){\line(1,0){15 } } \put(4,8){\makebox(0,0){$\scriptstyle -5 $ } } \end{picture}\to \begin{picture}(42,5 ) \put(20,.1){\line(1,0){16 } } \put(20,3.3){\line(1,0){16 } } \put(20,1.5){\makebox(0,0){$\bullet$ } } \put(36,1.5){\makebox(0,0){$\bullet$ } } \put(28,1.5){\makebox(0,0){$\langle$ } } \put(20,8){\makebox(0,0){$\scriptstyle 1 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 0 $ } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(4,1.5){\line(1,0){15 } } \put(4,8){\makebox(0,0){$\scriptstyle -6 $ } } \end{picture}\to \begin{picture}(42,5 ) \put(20,.1){\line(1,0){16 } } \put(20,3.3){\line(1,0){16 } } \put(20,1.5){\makebox(0,0){$\bullet$ } } \put(36,1.5){\makebox(0,0){$\bullet$ } } \put(28,1.5){\makebox(0,0){$\langle$ } } \put(20,8){\makebox(0,0){$\scriptstyle 0 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 0 $ } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(4,1.5){\line(1,0){15 } } \put(4,8){\makebox(0,0){$\scriptstyle -6 $ } } \end{picture}\to 0.\end{array}\ ] ] this coincides with the rumin complex ( [ rumincomplex ] ) .
the reason for the notation is fully explained in @xcite . here ,
suffice it to say that @xmath108 where @xmath109 denotes the bundle induced by the irreducible representation @xmath110 of @xmath111 ( meaning that its highest weight is @xmath112 $ ] with respect to the standard bourbaki - ordered basis of fundamental weights ) . in summary , there is a homogeneous contact geometry @xmath0 , with @xmath2 simple and @xmath100 parabolic , for which the bgg complex coincides with the rumin complex .
let @xmath6 be a @xmath5-manifold equipped with @xmath7 , a generic distribution of rank @xmath33 .
equivalently , let @xmath113 be a pfaffian system of rank @xmath35 that is generic in cartan s sense , i.e. the first derived system @xmath114 has rank @xmath33 and the second derived system @xmath115 is zero .
we have a filtration of the tangent bundle @xmath116\subset tm\quad\mbox{by vector bundles of ranks}\quad 2,3,5\ ] ] and a dual filtration of the cotangent bundle , which we shall write as @xmath117 where @xmath22 is the line - bundle @xmath118 .
there are locally defined co - framings @xmath119 so that the following congruences hold @xmath120 d\omega^3\equiv\omega^4\wedge\omega^5\bmod\omega^1,\omega^2,\omega^3 \end{array}\ ] ] with @xmath121 and @xmath122 .
we shall refer to such co - framings as _
adapted_. the levi form for @xmath123 , defined as the composition @xmath124 , has @xmath114 as its kernel ( by definition of @xmath114 or by viewing this form in an adapted co - frame ) . hence , the line - bundle @xmath22 may be canonically identified with @xmath125 .
similarly , the levi form for @xmath114 , namely @xmath126 , canonically identifies @xmath114 with @xmath127 .
therefore , we may rewrite ( [ filter ] ) as @xmath128 to proceed , it is useful to have a more compact notation for the bundles induced by @xmath12 . following a common convention for
the irreducible representations of @xmath129 , let us write @xmath130 where @xmath131 means symmetric tensor product .
then ( [ betterfilter ] ) becomes @xmath132 and the induced filtration on @xmath33-forms is @xmath133 \oplus\\[-2pt ] ( 2,2 ) \end{array}+(2,3)+(3,3).\ ] ] without further ado , we may now consider the spectral sequence of the de rham complex filtered in this way . at the @xmath15-level we obtain @xmath134the @xmath15-level
differentials are easily computed from the structure equations ( [ g2structure ] ) , the @xmath27-level is @xmath135 & & 0&0&0&0&0&0\\[2pt ] & & & 0&0&\!(1,3)\!&0&0&0\\[2pt ] & & & & 0&0&0&\!(2,4)\!&0&0\\[2pt ] & & & & & 0&0&0&0&0&0\\[2pt ] & & & & & & \;0\;&\;0\;&\;0\;&\;0\;&\!(4,5)\!&\!\to\!&\!(5,5)\ , , \end{array}$ } } \put(0,0){\vector(1,0){40 } } \put(0,0){\vector(0,1){45 } } \put(30,5){$p$ } \put(3,35){$q$ } \end{picture}\]]and we have shown that there is a differential complex @xmath136 resolving the constant sheaf @xmath4 , where @xmath137 simply indicates a differential operator of order @xmath138 . if necessary , the structure equations ( [ g2structure ] ) can be used to compute the operators precisely . to compare with the usual bgg complex , we follow cartan @xcite in realising the flat model for this geometry as a homogeneous space @xmath0 where @xmath2 is the exceptional non - compact lie group @xmath139 and @xmath100 is a parabolic subgroup . specifically , following the notation of @xcite , the homogeneous space is @xmath140 .
the levi factor of the parabolic subgroup is @xmath141 but it is useful to identify its lie algebra with @xmath142 where we have graded the lie algebra of @xmath139 @xmath143 in accordance with the parabolic subalgebra @xmath144 ( see @xcite ) . with these conventions , the cotangent bundle is @xmath145 and the bgg complex is @xmath146 more generally , in dynkin diagram notation the bundle @xmath147 is written as @xmath148 .
in fact , there are several other complexes that can be created from the de rham complex by choosing to carry out only some of the diagram chasing involved in creating the bgg complex .
we now explain two of these complexes and their motivation . keeping the dynkin diagram notation ,
the filtration of the @xmath33-forms induced from ( [ cotangent ] ) is @xmath149 \oplus\\[0pt ] \begin{picture}(24,5 ) \put(5.6,0){\line(1,0){14.4 } } \put(4,1.6){\line(1,0){16 } } \put(5.6,3.2){\line(1,0){14.4 } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(20,1.4){\makebox(0,0){$\bullet$ } } \put(12,1.5){\makebox(0,0){$\langle$ } } \put(4,8){\makebox(0,0){$\scriptstyle -2 $ } } \put(20,8){\makebox(0,0){$\scriptstyle 0 $ } } \end{picture}\end{array}+ \begin{picture}(24,5 ) \put(5.6,0){\line(1,0){14.4 } } \put(4,1.6){\line(1,0){16 } } \put(5.6,3.2){\line(1,0){14.4 } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(20,1.4){\makebox(0,0){$\bullet$ } } \put(12,1.5){\makebox(0,0){$\langle$ } } \put(4,8){\makebox(0,0){$\scriptstyle -4 $ } } \put(20,8){\makebox(0,0){$\scriptstyle 1 $ } } \end{picture}+ \begin{picture}(24,5 ) \put(5.6,0){\line(1,0){14.4 } } \put(4,1.6){\line(1,0){16 } } \put(5.6,3.2){\line(1,0){14.4 } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(20,1.4){\makebox(0,0){$\bullet$ } } \put(12,1.5){\makebox(0,0){$\langle$ } } \put(4,8){\makebox(0,0){$\scriptstyle -3 $ } } \put(20,8){\makebox(0,0){$\scriptstyle 0 $ } } \end{picture},\ ] ] which suggests that one might cancel @xmath150 from the exterior derivative @xmath151 .
but , as a sub - bundle of @xmath10 , this is precisely the original pfaffian system @xmath123 .
so , we are trying to cancel from @xmath152 , the homomorphism defined as the composition @xmath153 we can accomplish this explicitly by means of an adapted co - frame . specifically , we define a differential operator @xmath154 by the following steps . starting with any @xmath9-form @xmath59 , *
define @xmath155 by @xmath156 , * define @xmath157 by @xmath158 .
this is possible according to the structure equations ( [ g2structure ] ) , which also imply that @xmath159 in other words that @xmath160 one checks easily that this definition of @xmath161 is independent of choice of adapted co - framing .
moreover , if @xmath59 is actually a section of @xmath123 , say @xmath162 , then @xmath163 , @xmath164 , and @xmath165 , whence @xmath166 . in other words
, the differential operator @xmath167 descends to @xmath12 , as claimed in ( [ euler ] ) .
[ thm1 ] the sequence @xmath168 is a locally exact complex .
this is just a matter of unravelling definitions , bearing in mind that the de rham complex is itself locally exact .
suppose , for example , that @xmath59 is a @xmath9-form representing a section of @xmath12 that is annihilated by @xmath167 .
locally , we need to find a smooth function @xmath169 such that @xmath170 is a section of @xmath123 . by construction of @xmath167 we know @xmath171 for some smooth functions @xmath172 .
thus , by exactness of the de rham complex , locally we can write @xmath173 and then @xmath174 is a section of @xmath123 , as required .
the remaining verifications are similarly straightforward .
the operator @xmath175 has a geometric meaning : a section @xmath169 of @xmath12 can be regarded as a lagrangian for an integral curve of @xmath123 . from this point of view
@xmath176 are the euler - lagrange equations associated to this lagrangian . from its construction
, one can easily verify that @xmath167 is third order . more specifically , by construction , its symbol @xmath177 \oplus\\[0pt ] \begin{picture}(24,5 ) \put(5.6,0){\line(1,0){14.4 } } \put(4,1.6){\line(1,0){16 } } \put(5.6,3.2){\line(1,0){14.4 } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(20,1.4){\makebox(0,0){$\bullet$ } } \put(12,1.5){\makebox(0,0){$\langle$ } } \put(4,8){\makebox(0,0){$\scriptstyle -2 $ } } \put(20,8){\makebox(0,0){$\scriptstyle 0 $ } } \end{picture}\end{array}+ \begin{picture}(24,5 ) \put(5.6,0){\line(1,0){14.4 } } \put(4,1.6){\line(1,0){16 } } \put(5.6,3.2){\line(1,0){14.4 } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(20,1.4){\makebox(0,0){$\bullet$ } } \put(12,1.5){\makebox(0,0){$\langle$ } } \put(4,8){\makebox(0,0){$\scriptstyle -4 $ } } \put(20,8){\makebox(0,0){$\scriptstyle 1 $ } } \end{picture}+ \begin{picture}(24,5 ) \put(5.6,0){\line(1,0){14.4 } } \put(4,1.6){\line(1,0){16 } } \put(5.6,3.2){\line(1,0){14.4 } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(20,1.4){\makebox(0,0){$\bullet$ } } \put(12,1.5){\makebox(0,0){$\langle$ } } \put(4,8){\makebox(0,0){$\scriptstyle -3 $ } } \put(20,8){\makebox(0,0){$\scriptstyle 0 $ } } \end{picture}\ ] ] composes with the projection to @xmath178 as the homomorphism @xmath179 \qquad\to\bigodot^3\!\big(\,\begin{picture}(24,5 ) \put(5.6,0){\line(1,0){14.4 } } \put(4,1.6){\line(1,0){16 } } \put(5.6,3.2){\line(1,0){14.4 } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(20,1.4){\makebox(0,0){$\bullet$ } } \put(12,1.5){\makebox(0,0){$\langle$ } } \put(4,8){\makebox(0,0){$\scriptstyle -2 $ } } \put(20,8){\makebox(0,0){$\scriptstyle 1 $ } } \end{picture}\big)\otimes \begin{picture}(24,5 ) \put(5.6,0){\line(1,0){14.4 } } \put(4,1.6){\line(1,0){16 } } \put(5.6,3.2){\line(1,0){14.4 } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(20,1.4){\makebox(0,0){$\bullet$ } } \put(12,1.5){\makebox(0,0){$\langle$ } } \put(4,8){\makebox(0,0){$\scriptstyle -2 $ } } \put(20,8){\makebox(0,0){$\scriptstyle 1 $ } } \end{picture}\to \bigodot^2\!\big(\,\begin{picture}(24,5 ) \put(5.6,0){\line(1,0){14.4 } } \put(4,1.6){\line(1,0){16 } } \put(5.6,3.2){\line(1,0){14.4 } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(20,1.4){\makebox(0,0){$\bullet$ } } \put(12,1.5){\makebox(0,0){$\langle$ } } \put(4,8){\makebox(0,0){$\scriptstyle -2 $ } } \put(20,8){\makebox(0,0){$\scriptstyle 1 $ } } \end{picture}\big)\otimes \lambda^2\big(\,\begin{picture}(24,5 ) \put(5.6,0){\line(1,0){14.4 } } \put(4,1.6){\line(1,0){16 } } \put(5.6,3.2){\line(1,0){14.4 } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(20,1.4){\makebox(0,0){$\bullet$ } } \put(12,1.5){\makebox(0,0){$\langle$ } } \put(4,8){\makebox(0,0){$\scriptstyle -2 $ } } \put(20,8){\makebox(0,0){$\scriptstyle 1 $ } } \end{picture}\big)\\[5pt ] \mbox { } \hspace{145pt}{}=\begin{picture}(24,5 ) \put(5.6,0){\line(1,0){14.4 } } \put(4,1.6){\line(1,0){16 } } \put(5.6,3.2){\line(1,0){14.4 } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(20,1.4){\makebox(0,0){$\bullet$ } } \put(12,1.5){\makebox(0,0){$\langle$ } } \put(4,8){\makebox(0,0){$\scriptstyle -4 $ } } \put(20,8){\makebox(0,0){$\scriptstyle 2 $ } } \end{picture}\otimes \begin{picture}(24,5 ) \put(5.6,0){\line(1,0){14.4 } } \put(4,1.6){\line(1,0){16 } } \put(5.6,3.2){\line(1,0){14.4 } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(20,1.4){\makebox(0,0){$\bullet$ } } \put(12,1.5){\makebox(0,0){$\langle$ } } \put(4,8){\makebox(0,0){$\scriptstyle -1 $ } } \put(20,8){\makebox(0,0){$\scriptstyle 0 $ } } \end{picture}= \begin{picture}(24,5 ) \put(5.6,0){\line(1,0){14.4 } } \put(4,1.6){\line(1,0){16 } } \put(5.6,3.2){\line(1,0){14.4 } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(20,1.4){\makebox(0,0){$\bullet$ } } \put(12,1.5){\makebox(0,0){$\langle$ } } \put(4,8){\makebox(0,0){$\scriptstyle -5 $ } } \put(20,8){\makebox(0,0){$\scriptstyle 2 $ } } \end{picture}.\end{array}\ ] ] furthermore , not only does the symbol have no component in @xmath180 but , in fact , the range of the operator @xmath167 is entirely contained in the sub - bundle where this component vanishes .
this is easily seen in an adapted co - frame : since @xmath181 and since @xmath182 spans @xmath183 , any exact @xmath33-form in @xmath184 has vanishing component in @xmath183 .
if we denote by @xmath185 the rank @xmath186 sub - bundle of @xmath184 defined as the kernel of the natural projection @xmath187 , and by @xmath188 the rank @xmath189 sub - bundle of @xmath190 generated by @xmath191 , then we have cancelled @xmath183 from the complex of theorem [ thm1 ] and demonstrated the following improvement .
the sequence @xmath192 is a locally exact complex .
the ranks of the bundles and the orders of the differential operators in ( [ basiccomplex ] ) are @xmath193 but if we consider @xmath194 in more detail @xmath195 \oplus\\[0pt ] \begin{picture}(24,5 ) \put(5.6,0){\line(1,0){14.4 } } \put(4,1.6){\line(1,0){16 } } \put(5.6,3.2){\line(1,0){14.4 } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(20,1.4){\makebox(0,0){$\bullet$ } } \put(12,1.5){\makebox(0,0){$\langle$ } } \put(4,8){\makebox(0,0){$\scriptstyle -3 $ } } \put(20,8){\makebox(0,0){$\scriptstyle 0 $ } } \end{picture}\end{array } \!\!+\begin{picture}(24,5 ) \put(5.6,0){\line(1,0){14.4 } } \put(4,1.6){\line(1,0){16 } } \put(5.6,3.2){\line(1,0){14.4 } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(20,1.4){\makebox(0,0){$\bullet$ } } \put(12,1.5){\makebox(0,0){$\langle$ } } \put(4,8){\makebox(0,0){$\scriptstyle -5 $ } } \put(20,8){\makebox(0,0){$\scriptstyle 1 $ } } \end{picture}+ \begin{picture}(24,5 ) \put(5.6,0){\line(1,0){14.4 } } \put(4,1.6){\line(1,0){16 } } \put(5.6,3.2){\line(1,0){14.4 } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(20,1.4){\makebox(0,0){$\bullet$ } } \put(12,1.5){\makebox(0,0){$\langle$ } } \put(4,8){\makebox(0,0){$\scriptstyle -4 $ } } \put(20,8){\makebox(0,0){$\scriptstyle 0 $ } } \end{picture}=b^3,\ ] ] then it suggests that we should be able to eliminate @xmath196 from both bundles .
this is , indeed , the case as can be seen in an adapted co - frame : writing a general section of @xmath185 as @xmath197 we may define a differential operator @xmath198 by the following familiar steps .
[ cols= " < , < " , ] this is possible according to the structure equations ( [ g2structure ] ) , which also imply that @xmath199 lies in the sub - bundle @xmath200 of @xmath201 and that it descends to the quotient @xmath202 of @xmath185 .
it is easily verified that this definition of @xmath203 is independent of choice of co - framing and that , if we denote by @xmath204 the composition @xmath205 then the expected theorem follows : the sequence @xmath206 is a locally exact complex .
the ranks of the bundles and the orders of the differential operators in ( [ secondcomplex ] ) are @xmath207 writing ( [ secondcomplex ] ) as @xmath208 \mbox { } \hspace{152pt}\begin{picture}(24,5 ) \put(5.6,0){\line(1,0){14.4 } } \put(4,1.6){\line(1,0){16 } } \put(5.6,3.2){\line(1,0){14.4 } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(20,1.4){\makebox(0,0){$\bullet$ } } \put(12,1.5){\makebox(0,0){$\langle$ } } \put(4,8){\makebox(0,0){$\scriptstyle -5 $ } } \put(20,8){\makebox(0,0){$\scriptstyle 1 $ } } \end{picture}+\begin{picture}(24,5 ) \put(5.6,0){\line(1,0){14.4 } } \put(4,1.6){\line(1,0){16 } } \put(5.6,3.2){\line(1,0){14.4 } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(20,1.4){\makebox(0,0){$\bullet$ } } \put(12,1.5){\makebox(0,0){$\langle$ } } \put(4,8){\makebox(0,0){$\scriptstyle -4 $ } } \put(20,8){\makebox(0,0){$\scriptstyle 0 $ } } \end{picture}+\begin{picture}(24,5 ) \put(5.6,0){\line(1,0){14.4 } } \put(4,1.6){\line(1,0){16 } } \put(5.6,3.2){\line(1,0){14.4 } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(20,1.4){\makebox(0,0){$\bullet$ } } \put(12,1.5){\makebox(0,0){$\langle$ } } \put(4,8){\makebox(0,0){$\scriptstyle -6 $ } } \put(20,8){\makebox(0,0){$\scriptstyle 1 $ } } \end{picture}\to \begin{picture}(24,5 ) \put(5.6,0){\line(1,0){14.4 } } \put(4,1.6){\line(1,0){16 } } \put(5.6,3.2){\line(1,0){14.4 } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(20,1.4){\makebox(0,0){$\bullet$ } } \put(12,1.5){\makebox(0,0){$\langle$ } } \put(4,8){\makebox(0,0){$\scriptstyle -5 $ } } \put(20,8){\makebox(0,0){$\scriptstyle 0 $ } } \end{picture},\end{array}\ ] ] suggests one final cancellation , specifically of @xmath209 from @xmath210 and @xmath211 . the reader can readily verify that this gives the bgg complex ( [ g2bgg ] ) .
it is interesting to note that the ranks of the bundles and orders of differential operators in the bgg complex are @xmath212 in particular , the order of the differential operator in the middle has gone down from @xmath35 to @xmath33 . since our filtering on the de rham complex is , by construction , compatible with the tautological hodge isomorphisms @xmath213 , and since we have run the spectral sequence to its end , it follows that the bgg complex is formally self - adjoint .
let @xmath6 be a @xmath186-manifold equipped with @xmath7 , a generic distribution of rank @xmath35 .
equivalently , let @xmath113 be a pfaffian system of rank @xmath35 that is generic in cartan s sense , i.e. the first derived system @xmath114 is zero .
locally there are co - framings @xmath214 so that @xmath215 span @xmath123 and the following congruences hold .
@xmath216 d\omega^3\equiv\omega^4\wedge\omega^5\bmod\omega^1,\omega^2,\omega^3 .
\end{array}\ ] ] in the terminology of @xcite , these co - framings are _ @xmath9-adapted_. as usual , let us write @xmath12 for @xmath217 .
then the levi form @xmath218 defined as the composition @xmath219 is an isomorphism and we can canonically identify @xmath123 with @xmath125 as vector bundles . indeed , this isomorphism is apparent in our @xmath9-adapted co - framing ( [ spin7structure ] )
. we may mimic ( [ clebsch ] ) to write , up to isomorphism , the general schur - irreducible bundle induced by @xmath12 as @xmath220 with @xmath221 for the bundle @xmath222 where @xmath223 as a subscript means to take the trace - free part .
these observations mean that we may write the filtration @xmath224 and decompose the induced filtrations on the higher forms as @xmath225 \oplus\\[-2pt ] ( 1,1,1)\end{array } + ( 1,1,2)\\[20pt ] \lambda^3= ( 1,1,1)+\begin{array}c(0,2,2)\\[-2pt ] \oplus\\[-2pt ] ( 1,1,2)\end{array}+ \begin{array}c(1,1,3)\\[-2pt ] \oplus\\[-2pt ] ( 1,2,2)\end{array}+(2,2,2)\\[20pt ] \lambda^4= ( 1,2,2)+\begin{array}c(1,2,3)\\[-2pt ] \oplus\\[-2pt ] ( 2,2,2)\end{array}+ ( 2,2,3)\enskip\quad \lambda^5=(2,2,3)+(2,3,3)\\[12pt ] \mbox { } \hspace{140pt}\lambda^6=(3,3,3 ) .
\end{array}\ ] ] from the structure equations ( [ spin7structure ] ) for a @xmath9-adapted co - framing it is easily verified that all expected cancellations at the @xmath15-level of the associated spectral sequence actually take place and we have proved the following result .
[ prebgg ] there is a canonically defined locally exact differential complex @xmath226\to\quad\\ ( 1,2,3)\to(2,3,3)\to(3,3,3)\to 0 \end{array}\ ] ] on any smooth @xmath186-manifold equipped with a generic @xmath35-distribution .
for the moment , the bundle @xmath227 $ ] is a canonically defined sub - quotient of @xmath190 but , in fact , one can improve matters as the following theorem shows ( the analogous step was not necessary in [ five ] ) . [ preferred ] a splitting of the short exact sequence @xmath228 gives rise to a homomorphism of vector bundles defined as the composition @xmath229 and there is a preferred class of splittings characterised by requiring that this induced homomorphism vanish .
this preference canonically splits the bundle @xmath227 $ ] .
certainly , a splitting of the @xmath9-forms splits all the other forms and so , from ( [ higherforms ] ) , one may consider the composition ( [ hom ] ) obtained by splitting the @xmath33-forms and @xmath35-forms . to see that it is a homomorphism , rather than the differential operator it might appear to be , notice that if @xmath230 is a @xmath33-form in @xmath231 , then @xmath232 and it is clear that @xmath233 has components only in @xmath234 inside @xmath235 \oplus\\[-2pt ] ( 1,1,2)\end{array } \oplus\begin{array}c(1,1,3)\\[-2pt ] \oplus\\[-2pt ] ( 1,2,2)\end{array } \oplus(2,2,2).\ ] ] in particular , if we project to @xmath236 as in ( [ hom ] ) , then @xmath233 does not contribute and , from ( [ leibniz ] ) , the result is linear over the functions . now suppose we change the splitting of the @xmath9-forms .
the freedom in doing so lies in @xmath237 this same freedom shows up in splitting the first part of @xmath190 : @xmath238 \oplus\\[-2pt ] ( 1,1,2)\end{array}+\cdots{}\ ] ] and @xmath239 \oplus\\[-2pt ] ( 1,1,2)\end{array}\right)= \begin{array}c(-1,1,1)\\[-2pt ] \oplus\\[-2pt ] ( 0,0,1)\end{array}.\ ] ] bearing in mind that the composition @xmath240 is an isomorphism , independent of choice of splitting ( it is responsible for one of the cancellations occurring at the @xmath15-level of the spectral sequence ) , we conclude that we can spend the @xmath241-freedom in splitting precisely in setting the homomorphism ( [ hom ] ) to zero .
now let us consider how this impacts on the sub - quotient @xmath227 $ ] of @xmath190 .
the freedom in splitting this sub - quotient lies in @xmath242 and one sees that the only way that ( [ basicfreedom ] ) can enter is through @xmath241 . having eliminated this freedom by a preferred choice of splittings for @xmath10 , it is thereby eliminated from ( [ morefreedom ] ) and we have obtained our canonical splitting .
the preferred splittings of @xmath10 afforded by theorem [ preferred ] can be conveniently expressed in terms of our @xmath9-adapted co - framings satisfying ( [ spin7structure ] ) .
if such a co - framing is used to split @xmath10 , then the resulting sub - bundle @xmath231 of @xmath243 is spanned by @xmath244 and , following through its proof , the preferred splittings of theorem [ preferred ] are characterised by requiring that @xmath245 in the terminology of @xcite , co - framings satisfying this extra congruence are called _
@xmath33-adapted_. the lie algebra @xmath246 admits a grading of the form @xmath247 \oplus\\[2pt ] \begin{picture}(42,5 ) \put(20,.1){\line(1,0){14.3 } } \put(20,3.2){\line(1,0){14.3 } } \put(20,1.5){\makebox(0,0){$\bullet$ } } \put(36,1.6){\makebox(0,0){$\times$ } } \put(28,1.5){\makebox(0,0){$\rangle$ } } \put(4,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\line(1,0){15 } } \put(4,8){\makebox(0,0){$\scriptstyle 0 $ } } \put(20,8){\makebox(0,0){$\scriptstyle 0 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 0 $ } } \end{picture } \end{array}\!\oplus \begin{picture}(42,5 ) \put(20,.1){\line(1,0){14.3 } } \put(20,3.2){\line(1,0){14.3 } } \put(20,1.5){\makebox(0,0){$\bullet$ } } \put(36,1.6){\makebox(0,0){$\times$ } } \put(28,1.5){\makebox(0,0){$\rangle$ } } \put(4,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\line(1,0){15 } } \put(4,8){\makebox(0,0){$\scriptstyle 0 $ } } \put(20,8){\makebox(0,0){$\scriptstyle 1 $ } } \put(36,8){\makebox(0,0){$\scriptstyle -2 $ } } \end{picture}\oplus \begin{picture}(42,5 ) \put(20,.1){\line(1,0){14.3 } } \put(20,3.2){\line(1,0){14.3 } } \put(20,1.5){\makebox(0,0){$\bullet$ } } \put(36,1.6){\makebox(0,0){$\times$ } } \put(28,1.5){\makebox(0,0){$\rangle$ } } \put(4,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\line(1,0){15 } } \put(4,8){\makebox(0,0){$\scriptstyle 1 $ } } \put(20,8){\makebox(0,0){$\scriptstyle 0 $ } } \put(36,8){\makebox(0,0){$\scriptstyle -2 $ } } \end{picture}\end{array}\ ] ] and one can see from this grading that the corresponding @xmath186-dimensional homogeneous space @xmath0 is equipped with a canonical @xmath35-dimensional distribution .
the corresponding infinitesimal flag structure @xcite is exactly the geometry of such @xmath35-distributions and the irreducible bundles are related in our two notations by @xmath248 recall in the proof of theorem [ preferred ] that we reduced the freedom in splitting @xmath10 to @xmath249 in ( [ basicfreedom ] ) . in the dynkin diagram notation
this remaining freedom lies in ( 42,5 ) ( 20,.1)(1,0)14.3 ( 20,3.2)(1,0)14.3 ( 20,1.5)(0,0)@xmath103 ( 36,1.6)(0,0)@xmath105 ( 28,1.5)(0,0)@xmath250 ( 4,1.5)(0,0)@xmath103 ( 4,1.5)(1,0)15 ( 4,8)(0,0)@xmath251 ( 20,8)(0,0)@xmath252 ( 36,8)(0,0)@xmath253 , which is exactly the action of @xmath254 on @xmath255 as can be seen in ( [ grading ] ) .
the geometric import of this observation is that theorem [ preferred ] reduces the structure group of the tangent bundle from general @xmath19-preserving and levi - form - preserving automorphisms to the subgroup of @xmath256 defined by the adjoint action of @xmath100 , namely the group @xmath257 with lie algebra @xmath258 .
dually , the @xmath33-adapted co - framings are preserved by exactly this group .
finally , we can take the complex of theorem [ prebgg ] , use the splitting of @xmath227 $ ] afforded by theorem [ preferred ] , and write the result in dynkin diagram notation to obtain the following . on any smooth @xmath186-manifold equipped with a generic @xmath35-distribution ,
there is a canonically defined locally exact differential complex @xmath259 \oplus\\[4pt ] \begin{picture}(42,5 ) \put(20,.1){\line(1,0){14.3 } } \put(20,3.2){\line(1,0){14.3 } } \put(20,1.5){\makebox(0,0){$\bullet$ } } \put(36,1.6){\makebox(0,0){$\times$ } } \put(28,1.5){\makebox(0,0){$\rangle$ } } \put(4,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\line(1,0){15 } } \put(4,8){\makebox(0,0){$\scriptstyle 0 $ } } \put(20,8){\makebox(0,0){$\scriptstyle 2 $ } } \put(36,8){\makebox(0,0){$\scriptstyle -6 $ } } \end{picture } \end{array } \!\!\begin{array}{c}\searrow\\ \nearrow\end{array}\qquad\\[20pt ] \begin{picture}(42,5 ) \put(20,.1){\line(1,0){14.3 } } \put(20,3.2){\line(1,0){14.3 } } \put(20,1.5){\makebox(0,0){$\bullet$ } } \put(36,1.6){\makebox(0,0){$\times$ } } \put(28,1.5){\makebox(0,0){$\rangle$ } } \put(4,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\line(1,0){15 } } \put(4,8){\makebox(0,0){$\scriptstyle 1 $ } } \put(20,8){\makebox(0,0){$\scriptstyle 1 $ } } \put(36,8){\makebox(0,0){$\scriptstyle -6 $ } } \end{picture } \!\to\ !
\begin{picture}(42,5 ) \put(20,.1){\line(1,0){14.3 } } \put(20,3.2){\line(1,0){14.3 } } \put(20,1.5){\makebox(0,0){$\bullet$ } } \put(36,1.6){\makebox(0,0){$\times$ } } \put(28,1.5){\makebox(0,0){$\rangle$ } } \put(4,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\line(1,0){15 } } \put(4,8){\makebox(0,0){$\scriptstyle 1 $ } } \put(20,8){\makebox(0,0){$\scriptstyle 0 $ } } \put(36,8){\makebox(0,0){$\scriptstyle -6 $ } } \end{picture } \!\to\ !
\begin{picture}(42,5 ) \put(20,.1){\line(1,0){14.3 } } \put(20,3.2){\line(1,0){14.3 } } \put(20,1.5){\makebox(0,0){$\bullet$ } } \put(36,1.6){\makebox(0,0){$\times$ } } \put(28,1.5){\makebox(0,0){$\rangle$ } } \put(4,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\line(1,0){15 } } \put(4,8){\makebox(0,0){$\scriptstyle 0 $ } } \put(20,8){\makebox(0,0){$\scriptstyle 0 $ } } \put(36,8){\makebox(0,0){$\scriptstyle -6 $ } } \end{picture } \!\to 0.\end{array}\ ] ] this is the bgg complex in standard notation .
although the complex @xmath260 constructed in [ engel ] used nothing beyond an engel structure , for the full - blown resolution ( [ fullengel ] ) it was necessary to choose some extra structure , namely a complement to @xmath261 ( equivalently , a complement to @xmath262 , the _ engel line field _
@xcite ) . as pointed out to us by boris doubrov
, there is a unique homogeneous space of the form @xmath0 , for @xmath2 semisimple and @xmath100 parabolic , that carries a @xmath2-invariant engel structure . specifically , if @xmath263 and @xmath100 is its borel subgroup , then @xmath264 which , in dynkin diagram notation , reads @xmath265 the @xmath9-forms on this homogeneous space @xmath0 are filtered @xmath266 and the corresponding regular infinitesimal flag structure is exactly that of an engel manifold equipped with a choice of splitting @xmath77 as discussed in [ engel ] .
the bgg complex ( [ fullengel ] ) in dynkin diagram notation reads @xmath267 \oplus\\[4pt ] \begin{picture}(24,5 ) \put(5.6,0){\line(1,0){12.8 } } \put(5.6,3.2){\line(1,0){12.8 } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(20,1.5){\makebox(0,0){$\times$ } } \put(12,1.5){\makebox(0,0){$\langle$ } } \put(4,8){\makebox(0,0){$\scriptstyle -2 $ } } \put(20,8){\makebox(0,0){$\scriptstyle 1 $ } } \end{picture}\end{array}\!\ !
\begin{array}c\longrightarrow\\ \mbox{\large\begin{picture}(0,0 ) \put(0,0){\makebox(0,0){$\nearrow$ } } \put(0,0){\makebox(0,0){$\searrow$ } } \end{picture}}\\[4pt ] \longrightarrow\end{array}\!\ !
\begin{array}c\begin{picture}(24,5 ) \put(5.6,0){\line(1,0){12.8 } } \put(5.6,3.2){\line(1,0){12.8 } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(20,1.5){\makebox(0,0){$\times$ } } \put(12,1.5){\makebox(0,0){$\langle$ } } \put(4,8){\makebox(0,0){$\scriptstyle 2 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -3 $ } } \end{picture}\\[2pt ] \oplus\\[4pt ] \begin{picture}(24,5 ) \put(5.6,0){\line(1,0){12.8 } } \put(5.6,3.2){\line(1,0){12.8 } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(20,1.5){\makebox(0,0){$\times$ } } \put(12,1.5){\makebox(0,0){$\langle$ } } \put(4,8){\makebox(0,0){$\scriptstyle -4 $ } } \put(20,8){\makebox(0,0){$\scriptstyle 1 $ } } \end{picture}\end{array}\!\ !
\begin{array}c\longrightarrow\\ \mbox{\large\begin{picture}(0,0 ) \put(0,0){\makebox(0,0){$\nearrow$ } } \put(0,0){\makebox(0,0){$\searrow$ } } \end{picture}}\\[4pt ] \longrightarrow\end{array}\!\ !
\begin{array}c\begin{picture}(24,5 ) \put(5.6,0){\line(1,0){12.8 } } \put(5.6,3.2){\line(1,0){12.8 } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(20,1.5){\makebox(0,0){$\times$ } } \put(12,1.5){\makebox(0,0){$\langle$ } } \put(4,8){\makebox(0,0){$\scriptstyle 0 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -3 $ } } \end{picture}\\[2pt ] \oplus\\[4pt ] \begin{picture}(24,5 ) \put(5.6,0){\line(1,0){12.8 } } \put(5.6,3.2){\line(1,0){12.8 } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(20,1.5){\makebox(0,0){$\times$ } } \put(12,1.5){\makebox(0,0){$\langle$ } } \put(4,8){\makebox(0,0){$\scriptstyle -4 $ } } \put(20,8){\makebox(0,0){$\scriptstyle 0 $ } } \end{picture}\end{array}\!\ ! \begin{array}{c}\searrow\\ \nearrow\end{array } \begin{picture}(24,5 ) \put(5.6,0){\line(1,0){12.8 } } \put(5.6,3.2){\line(1,0){12.8 } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(20,1.5){\makebox(0,0){$\times$ } } \put(12,1.5){\makebox(0,0){$\langle$ } } \put(4,8){\makebox(0,0){$\scriptstyle -2 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -2 $ } } \end{picture}\to 0.\ ] ]
recall that an engel manifold is a @xmath11-dimensional manifold equipped with a generic @xmath33-dimensional distribution . the geometry considered in [ engel ] and [ engelrevisited ] was defined on an engel manifold by a choice of splitting of @xmath12 , the bundle of @xmath9-forms along @xmath19 ( rather than the filtration that is canonically present ) .
the geometry to be considered in this section will very much resemble this case .
let us consider a pfaffian system @xmath113 of rank @xmath33 on a smooth @xmath5-manifold . as usual
, we define the _ levi form _ @xmath17 as the composition @xmath268 where @xmath269 .
notice that @xmath125 has rank @xmath35 and we shall suppose that @xmath17 is injective , as is generically the case . under the canonical identification @xmath270
we see that the rank @xmath33 sub - bundle @xmath271 gives rise to a rank @xmath33 sub - distribution @xmath272 . in @xcite it is observed that @xmath273 is the unique rank @xmath33 sub - bundle of @xmath19 such that @xmath274\subseteq h$ ] .
it is not necessarily the case , however , that @xmath274=h$ ] ( in which case we would be back in the five variables geometry of [ five ] ) . to proceed further ,
let us write @xmath22 for the line sub - bundle @xmath275 and choose a complementary rank @xmath33 sub - bundle @xmath276 so that we now have a splitting @xmath277 .
this completes the definition of the structure to be considered in this section .
equivalently , we are considering a @xmath5-manifold @xmath6 equipped with a pair of transverse distributions @xmath273 and @xmath278 of ranks @xmath33 and @xmath9 , respectively , such that @xmath279\subseteq \ell\oplus d\qquad\mbox{and}\qquad[\ell\oplus d,\ell\oplus d ] = tm.\ ] ] this is precisely the regular infinitesimal flag structure associated with the grading @xmath280 \oplus\\[1pt ] \begin{picture}(42,5 ) \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(4,1.5){\line(1,0){32 } } \put(4,8){\makebox(0,0){$\scriptstyle -1 $ } } \put(20,8){\makebox(0,0){$\scriptstyle 1 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 1 $ } } \end{picture } \end{array}\oplus \begin{array}{c } \begin{picture}(42,5 ) \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(4,1.5){\line(1,0){32 } } \put(4,8){\makebox(0,0){$\scriptstyle 0 $ } } \put(20,8){\makebox(0,0){$\scriptstyle 0 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 0 $ } } \end{picture}\\[-2pt ] \oplus\\[1pt ] \begin{picture}(42,5 ) \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(4,1.5){\line(1,0){32 } } \put(4,8){\makebox(0,0){$\scriptstyle 0 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -1 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 2 $ } } \end{picture}\\[-2pt ] \oplus\\[1pt ] \begin{picture}(42,5 ) \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(4,1.5){\line(1,0){32 } } \put(4,8){\makebox(0,0){$\scriptstyle 0 $ } } \put(20,8){\makebox(0,0){$\scriptstyle 0 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 0 $ } } \end{picture } \end{array}\oplus \begin{array}{c } \begin{picture}(42,5 ) \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(4,1.5){\line(1,0){32 } } \put(4,8){\makebox(0,0){$\scriptstyle 1 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -2 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 1 $ } } \end{picture}\\[-2pt ] \oplus\\[1pt ] \begin{picture}(42,5 ) \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(4,1.5){\line(1,0){32 } } \put(4,8){\makebox(0,0){$\scriptstyle -2 $ } } \put(20,8){\makebox(0,0){$\scriptstyle 1 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 0 $ } } \end{picture } \end{array}\oplus \begin{picture}(42,5 ) \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(4,1.5){\line(1,0){32 } } \put(4,8){\makebox(0,0){$\scriptstyle -1 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -1 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 1 $ } } \end{picture}\end{array}\ ] ] of @xmath281 .
the bundles @xmath282 and the @xmath9-forms are filtered @xmath283 \oplus\\[-2pt ] l\end{array}$}+\,i= \begin{array}{c } \begin{picture}(42,5 ) \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(4,1.5){\line(1,0){32 } } \put(4,8){\makebox(0,0){$\scriptstyle 1 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -2 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 1 $ } } \end{picture}\\[-2pt ] \oplus\\[1pt ] \begin{picture}(42,5 ) \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(4,1.5){\line(1,0){32 } } \put(4,8){\makebox(0,0){$\scriptstyle -2 $ } } \put(20,8){\makebox(0,0){$\scriptstyle 1 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 0 $ } } \end{picture } \end{array}+ \begin{picture}(42,5 ) \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(4,1.5){\line(1,0){32 } } \put(4,8){\makebox(0,0){$\scriptstyle -1 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -1 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 1 $ } } \end{picture},\ ] ] as expected .
this induces filtrations on the higher forms as follows .
\oplus\\[1pt ] \begin{picture}(42,5 ) \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(4,1.5){\line(1,0){32 } } \put(4,8){\makebox(0,0){$\scriptstyle -1 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -1 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 1 $ } } \end{picture } \end{array}+ \begin{array}{c } \begin{picture}(42,5 ) \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(4,1.5){\line(1,0){32 } } \put(4,8){\makebox(0,0){$\scriptstyle 0 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -3 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 2 $ } } \end{picture}\\[-2pt ] \oplus\\[1pt ] \begin{picture}(42,5 ) \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(4,1.5){\line(1,0){32 } } \put(4,8){\makebox(0,0){$\scriptstyle 0 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -2 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 0 $ } } \end{picture}\\[-2pt ] \oplus\\[1pt ] \begin{picture}(42,5 ) \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(4,1.5){\line(1,0){32 } } \put(4,8){\makebox(0,0){$\scriptstyle -3 $ } } \put(20,8){\makebox(0,0){$\scriptstyle 0 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 1 $ } } \end{picture } \end{array}+ \begin{picture}(42,5 ) \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(4,1.5){\line(1,0){32 } } \put(4,8){\makebox(0,0){$\scriptstyle -2 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -1 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 0 $ } } \end{picture}\ ] ] and @xmath285 \oplus\\[1pt ] \begin{picture}(42,5 ) \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(4,1.5){\line(1,0){32 } } \put(4,8){\makebox(0,0){$\scriptstyle -2 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -1 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 0 $ } } \end{picture}\\[-2pt ] \oplus\\[1pt ] \begin{picture}(42,5 ) \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(4,1.5){\line(1,0){32 } } \put(4,8){\makebox(0,0){$\scriptstyle -2 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -2 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 2 $ } } \end{picture } \end{array}+ \begin{array}{c } \begin{picture}(42,5 ) \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(4,1.5){\line(1,0){32 } } \put(4,8){\makebox(0,0){$\scriptstyle -1 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -3 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 1 $ } } \end{picture}\\[-2pt ] \oplus\\[1pt ] \begin{picture}(42,5 ) \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(4,1.5){\line(1,0){32 } } \put(4,8){\makebox(0,0){$\scriptstyle -4 $ } } \put(20,8){\makebox(0,0){$\scriptstyle 0 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 0 $ } } \end{picture } \end{array}\\ \lambda^4= \begin{picture}(42,5 ) \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(4,1.5){\line(1,0){32 } } \put(4,8){\makebox(0,0){$\scriptstyle -1 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -3 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 1 $ } } \end{picture}+ \begin{array}{c } \begin{picture}(42,5 ) \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(4,1.5){\line(1,0){32 } } \put(4,8){\makebox(0,0){$\scriptstyle 0 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -4 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 0 $ } } \end{picture}\\[-2pt ] \oplus\\[1pt ] \begin{picture}(42,5 ) \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(4,1.5){\line(1,0){32 } } \put(4,8){\makebox(0,0){$\scriptstyle -3 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -2 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 1 $ } } \end{picture } \end{array}\qquad\qquad\qquad \lambda^5=\begin{picture}(42,5 ) \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(4,1.5){\line(1,0){32 } }
\put(4,8){\makebox(0,0){$\scriptstyle -2 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -3 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 0 $ } } \end{picture } \end{array}\ ] ] one can readily verify using an adapted co - frame @xmath286 with @xmath287 q+i={\mathrm{span}}\{\omega^1,\omega^2,\omega^4,\omega^5\ } \end{array}\ ] ] and such that @xmath288 that the expected cancellations in the @xmath15-level of the associated spectral sequence actually take place and we have found a differential complex as follows .
[ protobggcomplex ] on any @xmath5-dimensional manifold equipped with a geometric structure defined by transverse distributions @xmath273 and @xmath278 of ranks @xmath33 and @xmath9 , respectively , and satisfying ( [ anothergeometry ] ) , there is a canonically defined locally exact differential complex @xmath289 \begin{picture}(42,5 ) \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(4,1.5){\line(1,0){32 } } \put(4,8){\makebox(0,0){$\scriptstyle -2 $ } } \put(20,8){\makebox(0,0){$\scriptstyle 1 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 0 $ } } \end{picture } \end{array}\!\to\ !
\begin{array}{c } \begin{picture}(42,5 ) \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(4,1.5){\line(1,0){32 } } \put(4,8){\makebox(0,0){$\scriptstyle 2 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -3 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 0 $ } } \end{picture}\\ + \\[3pt ] \begin{picture}(42,5 ) \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(4,1.5){\line(1,0){32 } } \put(4,8){\makebox(0,0){$\scriptstyle 0 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -3 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 2 $ } } \end{picture}\oplus \begin{picture}(42,5 ) \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(4,1.5){\line(1,0){32 } } \put(4,8){\makebox(0,0){$\scriptstyle -3 $ } } \put(20,8){\makebox(0,0){$\scriptstyle 0 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 1 $ } } \end{picture } \end{array}\!\!\to\hspace{77pt}\\[30pt ] \begin{array}{c } \begin{picture}(42,5 ) \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(4,1.5){\line(1,0){32 } } \put(4,8){\makebox(0,0){$\scriptstyle -2 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -2 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 2 $ } } \end{picture}\oplus \begin{picture}(42,5 ) \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(4,1.5){\line(1,0){32 } } \put(4,8){\makebox(0,0){$\scriptstyle 1 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -4 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 1 $ } } \end{picture}\\ + \\[3pt ] \begin{picture}(42,5 ) \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(4,1.5){\line(1,0){32 } } \put(4,8){\makebox(0,0){$\scriptstyle -4 $ } } \put(20,8){\makebox(0,0){$\scriptstyle 0 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 0 $ } } \end{picture } \end{array}\!\!\to\ !
\begin{array}{c } \begin{picture}(42,5 ) \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(4,1.5){\line(1,0){32 } } \put(4,8){\makebox(0,0){$\scriptstyle 0 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -4 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 0 $ } } \end{picture}\\ \oplus\\[3pt ] \begin{picture}(42,5 ) \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(4,1.5){\line(1,0){32 } } \put(4,8){\makebox(0,0){$\scriptstyle -3 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -2 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 1 $ } } \end{picture } \end{array}\!\to \begin{picture}(42,5 ) \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(4,1.5){\line(1,0){32 } } \put(4,8){\makebox(0,0){$\scriptstyle -2 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -3 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 0 $ } } \end{picture}\to 0.\end{array}\ ] ] as in [ threeinsix ] , one can make a further normalisation in order to split the two bundles that have arisen from the spectral sequence , or from the equivalent diagram chasing , only as filtered bundles .
for the first of these we note that the freedom in its splitting lies in @xmath290 whereas , from ( [ oneforms ] ) , the freedom in splitting @xmath10 lies in @xmath291 \qquad{}=\big(\begin{picture}(42,5 ) \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(4,1.5){\line(1,0){32 } } \put(4,8){\makebox(0,0){$\scriptstyle -1 $ } } \put(20,8){\makebox(0,0){$\scriptstyle 1 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 1 $ } } \end{picture}\oplus\begin{picture}(42,5 ) \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(4,1.5){\line(1,0){32 } } \put(4,8){\makebox(0,0){$\scriptstyle 2 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -1 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 0 $ } } \end{picture}\big)\otimes\begin{picture}(42,5 ) \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(4,1.5){\line(1,0){32 } } \put(4,8){\makebox(0,0){$\scriptstyle -1 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -1 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 1 $ } } \end{picture}\\[6pt ] \qquad\qquad{}=\begin{picture}(42,5 ) \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(4,1.5){\line(1,0){32 } } \put(4,8){\makebox(0,0){$\scriptstyle -2 $ } } \put(20,8){\makebox(0,0){$\scriptstyle 0 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 2 $ } } \end{picture}\oplus\begin{picture}(42,5 ) \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(4,1.5){\line(1,0){32 } } \put(4,8){\makebox(0,0){$\scriptstyle -2 $ } } \put(20,8){\makebox(0,0){$\scriptstyle 1 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 0 $ } } \end{picture}\oplus\begin{picture}(42,5 ) \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(4,1.5){\line(1,0){32 } } \put(4,8){\makebox(0,0){$\scriptstyle 1 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -2 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 1 $ } } \end{picture}.\end{array}\ ] ] we see that only @xmath292 is common to both .
therefore , it is only this freedom that need be eliminated from the freedom to split @xmath10 .
in fact , once this freedom is eliminated , then theorem [ protobggcomplex ] is improved as follows .
[ fullbggcomplex ] on any @xmath5-dimensional manifold equipped with a regular infinitesimal flag structure defined by ( 42,5 ) ( 20,1.5)(0,0)@xmath105 ( 36,1.5)(0,0)@xmath103 ( 4,1.5)(0,0)@xmath105 ( 4,1.5)(1,0)32 , there is a canonically defined locally exact differential complex @xmath293 \begin{picture}(42,5 ) \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(4,1.5){\line(1,0){32 } } \put(4,8){\makebox(0,0){$\scriptstyle -2 $ } } \put(20,8){\makebox(0,0){$\scriptstyle 1 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 0 $ } } \end{picture } \end{array}\!\to\ !
\begin{array}{c } \begin{picture}(42,5 ) \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(4,1.5){\line(1,0){32 } } \put(4,8){\makebox(0,0){$\scriptstyle 2 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -3 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 0 $ } } \end{picture}\\ \oplus\\[3pt ] \begin{picture}(42,5 ) \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(4,1.5){\line(1,0){32 } } \put(4,8){\makebox(0,0){$\scriptstyle 0 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -3 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 2 $ } } \end{picture}\\ \oplus\\[3pt ] \begin{picture}(42,5 ) \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(4,1.5){\line(1,0){32 } } \put(4,8){\makebox(0,0){$\scriptstyle -3 $ } } \put(20,8){\makebox(0,0){$\scriptstyle 0 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 1 $ } } \end{picture } \end{array}\!\to\ !
\begin{array}{c } \begin{picture}(42,5 ) \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(4,1.5){\line(1,0){32 } } \put(4,8){\makebox(0,0){$\scriptstyle 1 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -4 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 1 $ } } \end{picture}\\ \oplus\\[3pt ] \begin{picture}(42,5 ) \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(4,1.5){\line(1,0){32 } } \put(4,8){\makebox(0,0){$\scriptstyle -2 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -2 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 2 $ } } \end{picture}\\ \oplus\\[3pt ] \begin{picture}(42,5 ) \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(4,1.5){\line(1,0){32 } } \put(4,8){\makebox(0,0){$\scriptstyle -4 $ } } \put(20,8){\makebox(0,0){$\scriptstyle 0 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 0 $ } } \end{picture } \end{array}\!\to\ !
\begin{array}{c } \begin{picture}(42,5 ) \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(4,1.5){\line(1,0){32 } } \put(4,8){\makebox(0,0){$\scriptstyle 0 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -4 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 0 $ } } \end{picture}\\ \oplus\\[3pt ] \begin{picture}(42,5 ) \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(4,1.5){\line(1,0){32 } } \put(4,8){\makebox(0,0){$\scriptstyle -3 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -2 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 1 $ } } \end{picture } \end{array}\enskip\quad\\ { } \to\begin{picture}(42,5 ) \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\makebox(0,0){$\times$ } } \put(4,1.5){\line(1,0){32 } } \put(4,8){\makebox(0,0){$\scriptstyle -2 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -3 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 0 $ } } \end{picture}\to 0.\end{array}\ ] ] as already remarked , to complete the proof we should find a preferred class of splittings of the @xmath9-forms ( [ oneforms ] ) so that the @xmath292-freedom present in the general splitting is eliminated . as in [ threeinsix ] , this can be achieved by restricting a particular component of the exterior derivative @xmath294 defined via an arbitrary splitting . in this case , we may consider the composition @xmath295 using an adapted co - framing ( [ yetanothercoframe ] ) , one may readily verify that * this is actually a homomorphism of vector bundles , * insisting that it vanish reduces the freedom in splitting @xmath10 + exactly as desired , * this also eliminates the freedom in splitting the filtered + occurring bundles in theorem [ protobggcomplex ] , which completes the proof .
the differential complex in theorem [ fullbggcomplex ] is our bgg complex for this parabolic geometry .
furthermore , one can easily check that in case @xmath274=\ell\oplus d$ ] , equivalently if the ( 42,5 ) ( 20,1.5)(0,0)@xmath105 ( 36,1.5)(0,0)@xmath103 ( 4,1.5)(0,0)@xmath105 ( 4,1.5)(1,0)32 ( 4,8)(0,0)@xmath296 ( 20,8)(0,0)@xmath297 ( 36,8)(0,0)@xmath251 -component of curvature does not vanish , then further cancellations may be effected and one reduces to the bgg complex for the five variables geometry previously discussed in [ five ] .
let @xmath6 be a @xmath298-manifold endowed with a generic distribution @xmath7 of rank @xmath11 .
equivalently , let @xmath113 be a pfaffian system of rank @xmath35 that is generic in cartan s sense , meaning that its first derived system @xmath299 is zero .
we write the corresponding filtration of the cotangent bundle as @xmath300 genericity says that the levi form , defined as the composition @xmath301 is injective .
it turns out that there exactly two types of generic rank @xmath11 distributions in dimension @xmath298 , corresponding to the two open orbits of the action of @xmath302 on the space of linear maps @xmath303 called elliptic , respectively hyperbolic ; see @xcite .
we shall treat these two cases simultaneously .
the lie algebra @xmath304 admits a grading of the form @xmath305 \qquad\begin{picture}(42,5 ) \put(21.4,.1){\line(1,0){14.3 } } \put(21.4,3.2){\line(1,0){14.3 } } \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.6){\makebox(0,0){$\bullet$ } } \put(28,1.5){\makebox(0,0){$\langle$ } } \put(4,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\line(1,0){16 } } \put(4,8){\makebox(0,0){$\scriptstyle 2 $ } } \put(20,8){\makebox(0,0){$\scriptstyle 0 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 0 $ } } \end{picture}\oplus \begin{picture}(42,5 ) \put(21.4,.1){\line(1,0){14.3 } } \put(21.4,3.2){\line(1,0){14.3 } } \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.6){\makebox(0,0){$\bullet$ } } \put(28,1.5){\makebox(0,0){$\langle$ } } \put(4,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\line(1,0){16 } } \put(4,8){\makebox(0,0){$\scriptstyle 1 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -1 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 1 $ } } \end{picture}\oplus \begin{array}{c } \begin{picture}(42,5 ) \put(21.4,.1){\line(1,0){14.3 } } \put(21.4,3.2){\line(1,0){14.3 } } \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.6){\makebox(0,0){$\bullet$ } } \put(28,1.5){\makebox(0,0){$\langle$ } } \put(4,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\line(1,0){16 } } \put(4,8){\makebox(0,0){$\scriptstyle 2 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -1 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 0 $ } } \end{picture}\\[-1pt ] \oplus\\[2pt ] \begin{picture}(42,5 ) \put(21.4,.1){\line(1,0){14.3 } } \put(21.4,3.2){\line(1,0){14.3 } } \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.6){\makebox(0,0){$\bullet$ } } \put(28,1.5){\makebox(0,0){$\langle$ } } \put(4,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\line(1,0){16 } } \put(4,8){\makebox(0,0){$\scriptstyle 0 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -2 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 2 $ } } \end{picture}\\[-1pt ] \oplus\\[2pt ] \begin{picture}(42,5 ) \put(21.4,.1){\line(1,0){14.3 } } \put(21.4,3.2){\line(1,0){14.3 } } \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.6){\makebox(0,0){$\bullet$ } } \put(28,1.5){\makebox(0,0){$\langle$ } } \put(4,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\line(1,0){16 } } \put(4,8){\makebox(0,0){$\scriptstyle 0 $ } } \put(20,8){\makebox(0,0){$\scriptstyle 0 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 0 $ } } \end{picture } \end{array}\oplus \begin{picture}(42,5 ) \put(21.4,.1){\line(1,0){14.3 } } \put(21.4,3.2){\line(1,0){14.3 } } \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.6){\makebox(0,0){$\bullet$ } } \put(28,1.5){\makebox(0,0){$\langle$ } } \put(4,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\line(1,0){16 } } \put(4,8){\makebox(0,0){$\scriptstyle 1 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -2 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 1 $ } } \end{picture}\oplus \begin{picture}(42,5 ) \put(21.4,.1){\line(1,0){14.3 } } \put(21.4,3.2){\line(1,0){14.3 } } \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.6){\makebox(0,0){$\bullet$ } } \put(28,1.5){\makebox(0,0){$\langle$ } } \put(4,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\line(1,0){16 } } \put(4,8){\makebox(0,0){$\scriptstyle 2 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -2 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 0 $ } } \end{picture}.\end{array}\ ] ] there are two real forms of this grading , namely @xmath306 and the split real form @xmath307 .
one can see that these gradings give rise to an elliptic generic rank @xmath11 distribution on the corresponding @xmath298-dimensional homogeneous space @xmath0 in the first case and to a hyperbolic generic rank @xmath11 distribution on the corresponding homogeneous space @xmath0 in the second case .
the parabolic geometries based on these particular @xmath0 are known as _
quaternionic contact _
@xcite and _ split quaternionic contact _
@xcite , respectively .
regular infinitesimal flag structures of these types correspond exactly to generic rank @xmath11 distributions on @xmath298-manifolds and the irreducible bundles of these geometries can be written as @xmath308 where @xmath309 are non - negative integers .
accordingly , we can write the filtration of the cotangent bundle of a generic rank @xmath11 distribution as @xmath310 and the filtration of the higher forms as @xmath311 \oplus\\[2pt ] \begin{picture}(42,5 ) \put(21.4,.1){\line(1,0){14.3 } } \put(21.4,3.2){\line(1,0){14.3 } } \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.6){\makebox(0,0){$\bullet$ } } \put(28,1.5){\makebox(0,0){$\langle$ } } \put(4,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\line(1,0){16 } } \put(4,8){\makebox(0,0){$\scriptstyle 0 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -3 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 2 $ } } \end{picture } \end{array } + \begin{array}{c } \begin{picture}(42,5 ) \put(21.4,.1){\line(1,0){14.3 } } \put(21.4,3.2){\line(1,0){14.3 } } \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.6){\makebox(0,0){$\bullet$ } } \put(28,1.5){\makebox(0,0){$\langle$ } } \put(4,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\line(1,0){16 } } \put(4,8){\makebox(0,0){$\scriptstyle 3 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -4 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 1 $ } } \end{picture}\\[-1pt ] \oplus\\[2pt ] \begin{picture}(42,5 ) \put(21.4,.1){\line(1,0){14.3 } } \put(21.4,3.2){\line(1,0){14.3 } } \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.6){\makebox(0,0){$\bullet$ } } \put(28,1.5){\makebox(0,0){$\langle$ } } \put(4,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\line(1,0){16 } } \put(4,8){\makebox(0,0){$\scriptstyle 1 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -3 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 1 $ } } \end{picture } \end{array } + \begin{picture}(42,5 ) \put(21.4,.1){\line(1,0){14.3 } } \put(21.4,3.2){\line(1,0){14.3 } } \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.6){\makebox(0,0){$\bullet$ } } \put(28,1.5){\makebox(0,0){$\langle$ } } \put(4,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\line(1,0){16 } } \put(4,8){\makebox(0,0){$\scriptstyle 2 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -3 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 0 $ } } \end{picture}\ ] ] @xmath312 \oplus\\[2pt ] \begin{picture}(42,5 ) \put(21.4,.1){\line(1,0){14.3 } } \put(21.4,3.2){\line(1,0){14.3 } } \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.6){\makebox(0,0){$\bullet$ } } \put(28,1.5){\makebox(0,0){$\langle$ } } \put(4,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\line(1,0){16 } } \put(4,8){\makebox(0,0){$\scriptstyle 2 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -3 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 0 $ } } \end{picture } \\[-1pt ] \oplus\\[2pt ] \begin{picture}(42,5 ) \put(21.4,.1){\line(1,0){14.3 } } \put(21.4,3.2){\line(1,0){14.3 } } \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.6){\makebox(0,0){$\bullet$ } } \put(28,1.5){\makebox(0,0){$\langle$ } } \put(4,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\line(1,0){16 } } \put(4,8){\makebox(0,0){$\scriptstyle 0 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -2 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 0 $ } } \end{picture } \\[-1pt ] \oplus\\[2pt ] \begin{picture}(42,5 ) \put(21.4,.1){\line(1,0){14.3 } } \put(21.4,3.2){\line(1,0){14.3 } } \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.6){\makebox(0,0){$\bullet$ } } \put(28,1.5){\makebox(0,0){$\langle$ } } \put(4,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\line(1,0){16 } } \put(4,8){\makebox(0,0){$\scriptstyle 2 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -5 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 2 $ } } \end{picture } \end{array } + \begin{array}{c } \begin{picture}(42,5 ) \put(21.4,.1){\line(1,0){14.3 } } \put(21.4,3.2){\line(1,0){14.3 } } \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.6){\makebox(0,0){$\bullet$ } } \put(28,1.5){\makebox(0,0){$\langle$ } } \put(4,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\line(1,0){16 } } \put(4,8){\makebox(0,0){$\scriptstyle 3 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -5 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 1 $ } } \end{picture}\\[-1pt ] \oplus\\[2pt ] \begin{picture}(42,5 ) \put(21.4,.1){\line(1,0){14.3 } } \put(21.4,3.2){\line(1,0){14.3 } } \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.6){\makebox(0,0){$\bullet$ } } \put(28,1.5){\makebox(0,0){$\langle$ } } \put(4,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\line(1,0){16 } } \put(4,8){\makebox(0,0){$\scriptstyle 1 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -4 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 1 $ } } \end{picture } \end{array } + \begin{picture}(42,5 ) \put(21.4,.1){\line(1,0){14.3 } } \put(21.4,3.2){\line(1,0){14.3 } } \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.6){\makebox(0,0){$\bullet$ } } \put(28,1.5){\makebox(0,0){$\langle$ } } \put(4,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\line(1,0){16 } } \put(4,8){\makebox(0,0){$\scriptstyle 0 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -3 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 0 $ } } \end{picture}\ ] ] @xmath313 \oplus\\[2pt ] \begin{picture}(42,5 ) \put(21.4,.1){\line(1,0){14.3 } } \put(21.4,3.2){\line(1,0){14.3 } } \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.6){\makebox(0,0){$\bullet$ } } \put(28,1.5){\makebox(0,0){$\langle$ } } \put(4,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\line(1,0){16 } } \put(4,8){\makebox(0,0){$\scriptstyle 1 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -4 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 1 $ } } \end{picture } \end{array } + \begin{array}{c } \begin{picture}(42,5 ) \put(21.4,.1){\line(1,0){14.3 } } \put(21.4,3.2){\line(1,0){14.3 } } \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.6){\makebox(0,0){$\bullet$ } } \put(28,1.5){\makebox(0,0){$\langle$ } } \put(4,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\line(1,0){16 } } \put(4,8){\makebox(0,0){$\scriptstyle 4 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -5 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 0 $ } } \end{picture}\\[-1pt ] \oplus\\[2pt ] \begin{picture}(42,5 ) \put(21.4,.1){\line(1,0){14.3 } } \put(21.4,3.2){\line(1,0){14.3 } } \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.6){\makebox(0,0){$\bullet$ } } \put(28,1.5){\makebox(0,0){$\langle$ } } \put(4,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\line(1,0){16 } } \put(4,8){\makebox(0,0){$\scriptstyle 2 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -4 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 0 $ } } \end{picture}\\[-1pt ] \oplus\\[2pt ] \begin{picture}(42,5 ) \put(21.4,.1){\line(1,0){14.3 } } \put(21.4,3.2){\line(1,0){14.3 } } \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.6){\makebox(0,0){$\bullet$ } } \put(28,1.5){\makebox(0,0){$\langle$ } }
\put(4,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\line(1,0){16 } } \put(4,8){\makebox(0,0){$\scriptstyle 0 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -3 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 0 $ } } \end{picture}\\[-1pt ] \oplus\\[2pt ] \begin{picture}(42,5 ) \put(21.4,.1){\line(1,0){14.3 } } \put(21.4,3.2){\line(1,0){14.3 } } \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.6){\makebox(0,0){$\bullet$ } } \put(28,1.5){\makebox(0,0){$\langle$ } } \put(4,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\line(1,0){16 } } \put(4,8){\makebox(0,0){$\scriptstyle 2 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -6 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 2 $ } } \end{picture } \end{array } + \begin{picture}(42,5 ) \put(21.4,.1){\line(1,0){14.3 } } \put(21.4,3.2){\line(1,0){14.3 } } \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.6){\makebox(0,0){$\bullet$ } } \put(28,1.5){\makebox(0,0){$\langle$ } } \put(4,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\line(1,0){16 } } \put(4,8){\makebox(0,0){$\scriptstyle 1 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -5 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 1 $ } } \end{picture}\ ] ] @xmath314 \oplus\\[2pt ] \begin{picture}(42,5 ) \put(21.4,.1){\line(1,0){14.3 } } \put(21.4,3.2){\line(1,0){14.3 } } \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.6){\makebox(0,0){$\bullet$ } } \put(28,1.5){\makebox(0,0){$\langle$ } } \put(4,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\line(1,0){16 } } \put(4,8){\makebox(0,0){$\scriptstyle 1 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -5 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 1 $ } } \end{picture } \end{array } + \begin{array}{c } \begin{picture}(42,5 ) \put(21.4,.1){\line(1,0){14.3 } } \put(21.4,3.2){\line(1,0){14.3 } } \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.6){\makebox(0,0){$\bullet$ } } \put(28,1.5){\makebox(0,0){$\langle$ } } \put(4,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\line(1,0){16 } } \put(4,8){\makebox(0,0){$\scriptstyle 2 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -5 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 0 $ } } \end{picture}\\[-1pt ] \oplus\\[2pt ] \begin{picture}(42,5 ) \put(21.4,.1){\line(1,0){14.3 } } \put(21.4,3.2){\line(1,0){14.3 } } \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.6){\makebox(0,0){$\bullet$ } } \put(28,1.5){\makebox(0,0){$\langle$ } } \put(4,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\line(1,0){16 } } \put(4,8){\makebox(0,0){$\scriptstyle 0 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -6 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 2 $ } } \end{picture } \end{array}\ ] ] @xmath315 choosing an adapted co - framing of the pfaffian system , one can explicitly verify that all the expected cancellations at the @xmath15-level of the associated spectral sequence take place and , therefore , one obtains the following .
there is a canonically defined locally exact differential complex @xmath316 \begin{picture}(42,5 ) \put(21.4,.1){\line(1,0){14.3 } } \put(21.4,3.2){\line(1,0){14.3 } } \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.6){\makebox(0,0){$\bullet$ } } \put(28,1.5){\makebox(0,0){$\langle$ } } \put(4,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\line(1,0){16 } } \put(4,8){\makebox(0,0){$\scriptstyle 3 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -4 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 1 $ } } \end{picture } \end{array } \!\!\rightarrow\!\ !
\begin{array}{c } \begin{picture}(42,5 ) \put(21.4,.1){\line(1,0){14.3 } } \put(21.4,3.2){\line(1,0){14.3 } } \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.6){\makebox(0,0){$\bullet$ } } \put(28,1.5){\makebox(0,0){$\langle$ } } \put(4,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\line(1,0){16 } } \put(4,8){\makebox(0,0){$\scriptstyle 2 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -5 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 2 $ } } \end{picture}\\ \oplus\\[3pt ] \begin{picture}(42,5 ) \put(21.4,.1){\line(1,0){14.3 } } \put(21.4,3.2){\line(1,0){14.3 } } \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.6){\makebox(0,0){$\bullet$ } } \put(28,1.5){\makebox(0,0){$\langle$ } } \put(4,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\line(1,0){16 } } \put(4,8){\makebox(0,0){$\scriptstyle 4 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -4 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 0 $ } } \end{picture } \end{array } \!\!\rightarrow\!\ !
\begin{array}{c } \begin{picture}(42,5 ) \put(21.4,.1){\line(1,0){14.3 } } \put(21.4,3.2){\line(1,0){14.3 } } \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.6){\makebox(0,0){$\bullet$ } } \put(28,1.5){\makebox(0,0){$\langle$ } } \put(4,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\line(1,0){16 } } \put(4,8){\makebox(0,0){$\scriptstyle 4 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -5 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 0 $ } } \end{picture}\\ \oplus\\[3pt ] \begin{picture}(42,5 ) \put(21.4,.1){\line(1,0){14.3 } } \put(21.4,3.2){\line(1,0){14.3 } } \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.6){\makebox(0,0){$\bullet$ } } \put(28,1.5){\makebox(0,0){$\langle$ } } \put(4,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\line(1,0){16 } } \put(4,8){\makebox(0,0){$\scriptstyle 2 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -6 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 2 $ } } \end{picture } \end{array } \!\!\rightarrow\\ \qquad\begin{array}{c } \begin{picture}(42,5 ) \put(21.4,.1){\line(1,0){14.3 } } \put(21.4,3.2){\line(1,0){14.3 } } \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.6){\makebox(0,0){$\bullet$ } } \put(28,1.5){\makebox(0,0){$\langle$ } } \put(4,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\line(1,0){16 } } \put(4,8){\makebox(0,0){$\scriptstyle 3 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -6 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 1 $ } } \end{picture}\\ + \\[3pt ] \begin{picture}(42,5 ) \put(21.4,.1){\line(1,0){14.3 } } \put(21.4,3.2){\line(1,0){14.3 } } \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.6){\makebox(0,0){$\bullet$ } } \put(28,1.5){\makebox(0,0){$\langle$ } } \put(4,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\line(1,0){16 } } \put(4,8){\makebox(0,0){$\scriptstyle 0 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -6 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 2 $ } } \end{picture } \end{array } \!\!\rightarrow \begin{picture}(42,5 ) \put(21.4,.1){\line(1,0){14.3 } } \put(21.4,3.2){\line(1,0){14.3 } } \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.6){\makebox(0,0){$\bullet$ } } \put(28,1.5){\makebox(0,0){$\langle$ } } \put(4,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\line(1,0){16 } } \put(4,8){\makebox(0,0){$\scriptstyle 1 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -6 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 1 $ } } \end{picture } \rightarrow \begin{picture}(42,5 ) \put(21.4,.1){\line(1,0){14.3 } } \put(21.4,3.2){\line(1,0){14.3 } } \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.6){\makebox(0,0){$\bullet$ } } \put(28,1.5){\makebox(0,0){$\langle$ } } \put(4,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\line(1,0){16 } } \put(4,8){\makebox(0,0){$\scriptstyle 0 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -5 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 0 $ } } \end{picture } \rightarrow 0\end{array}\ ] ] on any smooth @xmath298-manifold equipped with a generic distribution of rank @xmath11 .
the bundles @xmath317 and @xmath318 are sub - quotients of @xmath243 , respectively @xmath319 . however , there is a preferred class of splittings of the filtration of @xmath10 that canonically splits these bundles as the following result shows .
[ anotherpreferredsplitting ] the splittings of the short exact sequence @xmath320 are acted freely upon by @xmath321 there is a preferred class of splittings in which the ( 42,5 ) ( 21.4,.1)(1,0)14.3 ( 21.4,3.2)(1,0)14.3 ( 20,1.5)(0,0)@xmath105 ( 36,1.6)(0,0)@xmath103 ( 28,1.5)(0,0)@xmath104 ( 4,1.5)(0,0)@xmath103 ( 4,1.5)(1,0)16 ( 4,8)(0,0)@xmath322 ( 20,8)(0,0)@xmath323 ( 36,8)(0,0)@xmath252 -freedom is eliminated .
this restricted choice of splittings canonically splits the bundles ' '' '' @xmath317 and @xmath318 .
the only difficulty is in restricting the class of splittings and , as usual , one looks to the exterior derivative @xmath294 in the presence of a chosen splitting .
more specifically , one checks ( e.g. , in an adapted co - frame ) that , having chosen a splitting of @xmath10 , the resulting component of @xmath294 mapping @xmath324 to @xmath325 is actually a homomorphism .
however , @xmath326 & & \|\\[5pt ] & & \begin{picture}(42,5 ) \put(21.4,.1){\line(1,0){14.3 } } \put(21.4,3.2){\line(1,0){14.3 } } \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.6){\makebox(0,0){$\bullet$ } } \put(28,1.5){\makebox(0,0){$\langle$ } } \put(4,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\line(1,0){16 } } \put(4,8){\makebox(0,0){$\scriptstyle 5 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -4 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 1 $ } } \end{picture}\oplus\begin{picture}(42,5 ) \put(21.4,.1){\line(1,0){14.3 } } \put(21.4,3.2){\line(1,0){14.3 } } \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.6){\makebox(0,0){$\bullet$ } } \put(28,1.5){\makebox(0,0){$\langle$ } } \put(4,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\line(1,0){16 } } \put(4,8){\makebox(0,0){$\scriptstyle 3 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -3 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 1 $ } } \end{picture}\end{array}\ ] ] and one checks ( again using an adapted co - frame or by arguing with irreducible bundles and schur s lemma ) that the original freedom in splitting @xmath10 can be used to eliminate the @xmath327-component .
the remaining freedom in splitting @xmath10 is thereby restricted in the appropriate manner ( with the stated knock - on effect on the previously identified subquotients of @xmath243 and @xmath319 ) . of course
, one could rephrase theorem [ anotherpreferredsplitting ] as defining the notion of a @xmath33-adapted co - framing and observe that the effect is to reduce the structure bundle of @xmath10 to @xmath257 . in any case , combining the two theorems above we immediately obtain the following improved complex . on any smooth @xmath298-manifold endowed with a generic @xmath11-distribution , there is a canonically defined locally exact differential complex @xmath328 \begin{picture}(42,5 ) \put(21.4,.1){\line(1,0){14.3 } } \put(21.4,3.2){\line(1,0){14.3 } } \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.6){\makebox(0,0){$\bullet$ } } \put(28,1.5){\makebox(0,0){$\langle$ } } \put(4,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\line(1,0){16 } } \put(4,8){\makebox(0,0){$\scriptstyle 3 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -4 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 1 $ } } \end{picture } \end{array } \!\!\rightarrow\!\ !
\begin{array}{c } \begin{picture}(42,5 ) \put(21.4,.1){\line(1,0){14.3 } } \put(21.4,3.2){\line(1,0){14.3 } } \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.6){\makebox(0,0){$\bullet$ } } \put(28,1.5){\makebox(0,0){$\langle$ } } \put(4,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\line(1,0){16 } } \put(4,8){\makebox(0,0){$\scriptstyle 2 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -5 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 2 $ } } \end{picture}\\ \oplus\\[3pt ] \begin{picture}(42,5 ) \put(21.4,.1){\line(1,0){14.3 } } \put(21.4,3.2){\line(1,0){14.3 } } \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.6){\makebox(0,0){$\bullet$ } } \put(28,1.5){\makebox(0,0){$\langle$ } } \put(4,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\line(1,0){16 } } \put(4,8){\makebox(0,0){$\scriptstyle 4 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -4 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 0 $ } } \end{picture } \end{array } \!\!\rightarrow\!\ !
\begin{array}{c } \begin{picture}(42,5 ) \put(21.4,.1){\line(1,0){14.3 } } \put(21.4,3.2){\line(1,0){14.3 } } \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.6){\makebox(0,0){$\bullet$ } } \put(28,1.5){\makebox(0,0){$\langle$ } } \put(4,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\line(1,0){16 } } \put(4,8){\makebox(0,0){$\scriptstyle 4 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -5 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 0 $ } } \end{picture}\\ \oplus\\[3pt ] \begin{picture}(42,5 ) \put(21.4,.1){\line(1,0){14.3 } } \put(21.4,3.2){\line(1,0){14.3 } } \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.6){\makebox(0,0){$\bullet$ } } \put(28,1.5){\makebox(0,0){$\langle$ } } \put(4,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\line(1,0){16 } } \put(4,8){\makebox(0,0){$\scriptstyle 2 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -6 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 2 $ } } \end{picture } \end{array } \!\!\rightarrow\\ \qquad\begin{array}{c } \begin{picture}(42,5 ) \put(21.4,.1){\line(1,0){14.3 } } \put(21.4,3.2){\line(1,0){14.3 } } \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.6){\makebox(0,0){$\bullet$ } } \put(28,1.5){\makebox(0,0){$\langle$ } } \put(4,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\line(1,0){16 } } \put(4,8){\makebox(0,0){$\scriptstyle 3 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -6 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 1 $ } } \end{picture}\\ \oplus\\[3pt ] \begin{picture}(42,5 ) \put(21.4,.1){\line(1,0){14.3 } } \put(21.4,3.2){\line(1,0){14.3 } } \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.6){\makebox(0,0){$\bullet$ } } \put(28,1.5){\makebox(0,0){$\langle$ } } \put(4,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\line(1,0){16 } } \put(4,8){\makebox(0,0){$\scriptstyle 0 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -6 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 2 $ } } \end{picture } \end{array } \!\!\rightarrow \begin{picture}(42,5 ) \put(21.4,.1){\line(1,0){14.3 } } \put(21.4,3.2){\line(1,0){14.3 } } \put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.6){\makebox(0,0){$\bullet$ } } \put(28,1.5){\makebox(0,0){$\langle$ } } \put(4,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\line(1,0){16 } } \put(4,8){\makebox(0,0){$\scriptstyle 1 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -6 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 1 $ } } \end{picture } \rightarrow \begin{picture}(42,5 ) \put(21.4,.1){\line(1,0){14.3 } } \put(21.4,3.2){\line(1,0){14.3 } }
\put(20,1.5){\makebox(0,0){$\times$ } } \put(36,1.6){\makebox(0,0){$\bullet$ } } \put(28,1.5){\makebox(0,0){$\langle$ } } \put(4,1.5){\makebox(0,0){$\bullet$ } } \put(4,1.5){\line(1,0){16 } } \put(4,8){\makebox(0,0){$\scriptstyle 0 $ } } \put(20,8){\makebox(0,0){$\scriptstyle -5 $ } } \put(36,8){\makebox(0,0){$\scriptstyle 0 $ } } \end{picture } \rightarrow 0.\end{array}\ ] ] of course , on the homogeneous model @xmath0 this is the standard bgg complex .
finally , as mentioned already in the introduction , let us reiterate that our construction does not see the torsion of this parabolic geometry .
in fact , the torsion lies in @xmath329 , which is exactly the component of ( [ torsionishere ] ) that we have not eliminated .
although not a replacement for the de rham complex in resolving the constants , we take the opportunity here to describe another natural differential complex , the rumin - seshadri complex @xcite , the construction of which follows the same general technique .
this complex is defined on any symplectic @xmath330-manifold @xmath6 as follows . denoting by @xmath331 the symplectic @xmath33-form ,
let us consider the filtered differential complex @xmath332 defined by @xmath333 with differentials @xmath334 notice that this complex has local cohomology at both @xmath335 and @xmath336 .
specifically , the kernel of @xmath337 is @xmath338 and the cohomology at @xmath336 is generated by @xmath339 , where @xmath340 is any local potential for the symplectic form @xmath331 , meaning that @xmath341 .
evidently , this is a filtered complex : the de rham complex is a sub - complex .
the associated spectral sequence immediately gives rise to the rumin - seshadri differential complex on @xmath6 .
it has the form @xmath342 & & & & & & & & & & \big\downarrow\makebox[0pt][l]{\scriptsize$d_\perp^{(2)}$}\\ \lambda^0&\stackrel{d_\perp}{\longleftarrow}&\lambda^1 & \stackrel{d_\perp}{\longleftarrow}&\lambda_\perp^2 & \stackrel{d_\perp}{\longleftarrow}&\lambda_\perp^3 & \stackrel{d_\perp}{\longleftarrow}&\cdots & \stackrel{d_\perp}{\longleftarrow}&\lambda_\perp^{n } \end{array}\ ] ] where @xmath343 denotes the @xmath61-forms that are trace - free with respect to @xmath331 .
the conclusion is as follows . on any symplectic manifold ,
there is a differential complex ( [ rscomplex ] ) with local cohomology in degrees @xmath344 and @xmath345 . on the level of sheaves , in both these degrees
the cohomology is the locally constant sheaf @xmath346 . in all other degrees
it is locally exact . on a compact symplectic manifold of dimension @xmath347 ,
@xmath348 the construction of the complex and the identification of its local cohomology are immediate form the spectral sequence . to see ( [ firstcohomology ] ) , note that for a @xmath9-form @xmath59 to be in the kernel of @xmath349 is to say that @xmath350 for some smooth function @xmath155 but then @xmath351 however , the symplectic form can not be exact for @xmath6 compact so @xmath352 and thus @xmath353 . in four dimensions , the complex ( [ rscomplex ] ) is due to r.t .
smith @xcite . in higher dimensions ,
it was also found by l .- s .
tseng and s .- t .
yau @xcite who show that it is elliptic and go on to study its cohomology on compact manifolds .
the complex of first order operators after the second - order operator in the middle , was introduced by t. bouche @xcite and who dubbed it the _ coeffective _ complex ( he regarded it as a subcomplex of the second half of the de rham complex @xmath354 ) .
the coeffective cohomology was further studied by m. fernndez , r. ibez , and m. de len ( see , for example , @xcite ) .
rlb gratefully acknowledges nsf support from grants dms-8352009 and dms-8905207 ( from the 1980s when some of this work was done ) , and current nsf support from dms-1105868 .
mge , arg , and kn would like to thank the erwin schrdinger institute for hospitality in july 2011 during which this work was crucially advanced .
mge gratefully acknowledges support from the australian research council .
arg gratefully acknowledges support from the royal society of new zealand via marsden grant 10-uoa-113 .
we thank boris doubrov for pointing out to us the homogeneous engel manifold of [ engelrevisited ] , jean - pierre demailly for drawing our attention to @xcite , and li - sheng tseng for drawing our attention to @xcite .
bernstein , i.m .
gelfand , and s.i .
gelfand , differential operators on the base affine space and a study of @xmath355-modules , in : lie groups and their representations , halsted , new york , 1975 , pp .
2164 .
bryant , conformal geometry and @xmath35-plane fields on @xmath186-manifolds , in : proceedings of the rims symposium ` developments of cartan geometry and related mathematical problems ' ( 2427 october 2005 ) , kokyuroku , * 1502 * ( 2006 ) , pp . 115 . | for smooth manifolds equipped with various geometric structures , we construct complexes that replace the de rham complex in providing an alternative fine resolution of the sheaf of locally constant functions . in case
that the geometric structure is that of a parabolic geometry , our complexes coincide with the bernstein - gelfand - gelfand complex associated with the trivial representation .
however , at least in the cases we discuss , our constructions are relatively simple and avoid most of the machinery of parabolic geometry . moreover
, our method extends to contact and symplectic geometries ( beyond the parabolic realm ) . |
the pierre auger observatory in malarge , argentina , the world s largest facility to detect ultrahigh energy cosmic rays with energies above @xmath0ev , is currently nearing completion .
data collection started in january 2004 , and the observatory has already accumulated an ultrahigh energy cosmic ray data set of unprecedented size . due to their extremely low flux , primary cosmic rays at ultrahigh energies can not be observed directly .
rather , they enter the atmosphere and interact with air molecules , inducing a cascade of secondary particles , called an extensive air shower .
the properties of the primary particle ( energy , arrival direction , and chemical composition ) have to be deduced from indirect measurements of the extensive air shower .
two detector types have traditionally been used to measure air showers : surface detectors , which detect the particles of the air shower cascade that reach the ground ; and air fluorescence detectors , which make use of the fact that the particles in the air shower excite air molecules , causing uv fluorescence .
observing this uv light from air showers with photomultiplier cameras allows us to image the air shower development and obtain a nearly calorimetric energy estimate of the shower . for both detector types , the atmosphere acts as the calorimeter and provides the large detector volume needed to measure the small cosmic ray flux at the highest energies .
the pierre auger observatory is a hybrid detector which combines a large surface detector ( sd ) array of area @xmath1 and a system of four fluorescence telescopes ( fd ) at the same site . while the sd provides the well - defined aperture and the @xmath2 duty cycle needed to achieve high statistics , the fd provides , for a subset of showers , the calorimetric information needed to calibrate the sd . for the calibration to be meaningful , the properties of the calorimeter , _
i.e. _ the atmosphere , must be well - known .
particulates in the atmosphere , cherenkov radiation and the weather need to be factored into this calibration .
fluorescence light from air showers is affected by the absorption and scattering on molecules and aerosols , and , in addition , some fraction of the cherenkov radiation from the relativistic particles in the air shower is scattered into the detector .
this cherenkov light can not be separated from the fluorescence light , and so the contamination by cherenkov light needs to be modeled and subtracted to give an accurate energy estimate for the cosmic ray shower . moreover , the light transmission properties of the atmosphere are not constant over the large detector volume of the pierre auger observatory , and equally important , they are not constant in time .
changing weather conditions inevitably mean changing transmission properties .
consequently , the properties of the atmosphere need to be monitored continuously .
the pierre auger observatory has an extensive program to monitor the atmosphere within the fd aperture and measure atmospheric attenuation and scattering properties in the 300 to 400 nm sensitivity range of the fds .
a summary of the atmospheric monitoring system can be found in @xcite . within this system , a central role
is played by a system of four elastic backscatter lidar ( _ li_ght _ d_etection _ a_nd _ r_anging ) stations , one at each fluorescence site .
system is currently under construction , with three out of four stations operating . at each lidar station ,
a high - repetition uv laser sends short laser light pulses into the atmosphere in the direction of interest .
the backscattered signal is detected as a function of time by photomultiplier tubes at the foci of parabolic mirrors .
both the laser and the mirrors are mounted on a steering frame that allows the lidar to cover the full azimuth and elevation of the sky . during each hour of fd data taking ,
the four lidars perform a routine scan of the sky over each fd .
the data provide information about the height and coverage of clouds as well as their depth and opacity , and the local aerosol scattering and absorption properties of the atmosphere .
in addition to this routine operation , the lidar system is used for real time monitoring of the atmospheric homogeneity between the fds and selected cosmic ray events .
for example , if a high energy `` hybrid '' event is observed with the sd and one or more fds , the routine scan is interrupted and , within 2 to 4 minutes of the event detection , the lidar scans the atmosphere in the vicinity of the air shower reported by the fd .
this procedure is called `` shoot - the - shower '' ( sts ) , and allows for a rejection of events where the light profile from the track is distorted by clouds or other aerosol non - uniformities that are not characterized well by the average hourly aerosol measurements . both light reflection and opacity
can distort the light profile .
this paper describes the design , standard operation procedure , and performance of the lidar system of the pierre auger observatory .
it is organized as follows . after a brief description of the sd and fd detectors ( section[sec : auger ] ) ,
we give a summary of the relevant atmospheric parameters ( section[sec : atmo ] ) .
section[sec : hardware ] describes the current lidar hardware . in section[sec : ops ] , the daily operating procedure is summarized , including a description of the routine scan and the shoot - the - shower operation .
first results of the analysis of atmospheric parameters like cloud height and extinction are summarized in section[sec : analysis ] . a section on inelastic raman lidar measurements and a summary
conclude the paper .
in its final stage , the surface detector array of the pierre auger observatory will comprise 1600 water cherenkov detector tanks , deployed over an area of @xmath1 on a triangular grid with 1500 m grid spacing .
each of these cylindrical plastic tanks is 1.5 m tall with a @xmath3 base , and filled with 12,000liters of purified water within a tyvek liner .
three @xmath4 photomultiplier tubes in the tanks measure the cherenkov light produced by the particles of the extensive air shower cascade traversing the tank water .
the sd measures the particle density as a function of distance to the shower core , defined as the location where the primary particle would have hit the ground had it not showered in the atmosphere .
the sd array therefore measures the lateral profile of the shower at ground level .
the sd operates with a nearly @xmath2 duty cycle . for a fraction of the showers ,
the sd measurement is complemented by a measurement of the longitudinal profile of the shower in the atmosphere .
this information is provided by four fluorescence detectors . unlike cherenkov light , which is strongly forward - peaked ,
fluorescence light is emitted isotropically .
the fluorescence signal from the air shower is small compared to other light sources , but during dark , moonless nights , about @xmath5 of the total year , the fluorescence light can be distinguished from the night sky background .
the fd comprises four detector stations overlooking the sd from the periphery ( fig.[f : observatory ] ) . at present , the sites at los leones , coihueco , and los morados are completed and fully operational , while the fourth site at loma amarilla is under construction .
each site contains six bays , and each bay encloses a uv telescope comprised of a spherical light - collecting mirror , a photomultiplier camera in the focal surface , and a uv transmitting filter in the aperture .
the mirrors have a radius of curvature of 3.4 m and an area of about @xmath6 .
the camera consists of 440 photomultipliers with a hexagonal bialkaline photocathode , arranged in a @xmath7 array .
each camera has a field of view of @xmath8 in azimuth and @xmath9 in elevation , covering an elevation angle range from @xmath10 to @xmath11 above horizon . to reduce optical aberrations , the fd telescopes use schmidt optics with a circular diaphragm of diameter 2.2 m placed at the center of curvature of the mirror .
coma aberrations are minimized by use of a refractive corrector ring at the telescope aperture . the fd measures the amount of fluorescence light as a function of slant depth in the atmosphere .
this is then converted to the longitudinal profile of the shower , _
i.e. _ the number of charged particles as a function of slant depth .
the integral under the longitudinal profile is related to the electromagnetic energy of the shower , which can be converted to the total energy of the primary particle .
this method of energy determination is nearly calorimetric and mostly independent of the assumed particle interaction model .
the observatory started operation in hybrid mode in january 2004 .
first physics results , among them a first energy spectrum , were published based on data taken between january 2004 and may 2005 ( see @xcite for a summary and further references ) .
the strength of the observatory is its ability to calibrate the sd using fd - sd hybrid events to find an empirical relationship between the energy determined by the fd and the particle density on the ground measured by the sd .
detailed knowledge of the atmosphere is a crucial requirement to make this energy calibration work , as the accuracy of the energy calibration depends on an accurate measurement of the molecular atmosphere and the atmosphere s aerosol content .
the fd observes fluorescence light produced along the trajectory of the air shower in the atmosphere .
the fluorescence signal is roughly proportional to the number of particles in the shower cascade and therefore to the energy of the primary cosmic ray particle .
the fluorescence light traverses a path @xmath12 from the point of origin to the detector , and the detected light intensity @xmath13 is weakened due to scattering and absorption on molecules or aerosols in the atmosphere .
it is related to the light intensity at the place of its production , @xmath14 , by @xmath15 \mathrm{d}r = p_{0 } \exp\left[-\tau(l)\right ] \hspace{10pt},\ ] ] where @xmath16 is the volume extinction coefficient , and @xmath17 is the optical depth to the shower point at distance @xmath12 . to obtain @xmath17 ,
@xmath18 is integrated along the fluorescence light path between the detector and a given point on the shower . for atmospheric fluorescence measurements , we need to account for two scattering processes : rayleigh scattering from atmospheric molecules and scattering from low - altitude aerosols . both molecules and aerosols in the atmosphere predominantly scatter , rather than absorb , uv photons .
some absorption does occur due to the presence of ozone in the atmosphere and because the single scatter albedo of the aerosols is typically slightly less than unity , but these effects are small . in general , fluorescence photons that are scattered by the atmosphere do not contribute to the reconstructed fluorescence light signal , although a small multiple scattering correction can be made for those that do @xcite . in this paper we will use the term `` attenuation '' when photons are scattered in such a way that they do not contribute to the light signal recorded at an fd or lidar .
the attenuation can then be described in terms of optical depths @xmath19 for the molecular component and @xmath20 for the aerosols . neglecting multiple scattering ,
the transmission of light through a vertical column of atmosphere can be expressed as @xmath21 \hspace{10pt}.\ ] ] the parameters @xmath19 and @xmath20 need to be measured to relate the measured light intensity @xmath13 to @xmath14 and thus the shower energy . while absorption and scattering processes of the molecular atmosphere are well understood @xcite ,
the influence of the aerosol component is not . in principle , if one assumes that the aerosols are spherical with a known or assumed size distribution , then the light scattering can be described analytically by mie scattering theory @xcite . in practice , however , aerosols vary in size and shape , and the aerosol content changes on short time scales as wind lifts up dust , weather fronts pass through , or rain removes dust from the atmosphere . the aerosol optical depth @xmath20 therefore needs to be monitored on at least an hourly basis during each night of fd observation over the entire detection volume , which for the pierre auger observatory corresponds to a ground area of @xmath1 up to a height of about 15 km , the limit of the detector aperture . at each fd ,
the local lidars provide the aerosol scattering coefficient , @xmath22 , as a function of height @xmath23 .
the value of @xmath24 determines the amount of cherenkov light scattered from the extensive air showers .
the integral of @xmath22 from the fd height to @xmath23 gives the vertical aerosol optical depth , @xmath25 , which determines the transmission loss of light from each segment of the cosmic ray track to the fd .
a method to obtain @xmath20 from lidar scans is described in detail in @xcite .
in addition , the lidars provide the height and location of clouds and the optical depth inside the lowest cloud .
this analysis is described in section[sec : analysis ] .
there is some overlap between the primary tasks of the lidars and a separate device for atmospheric monitoring at the pierre auger observatory , the so - called central laser facility ( clf ) .
the clf is situated in the center of the sd array . from the clf
, a pulsed uv laser beam is directed into the sky , providing a test beam which can be observed by the fd . the design and performance of the clf
is described in more detail in @xcite . choosing a one - hour time bin for the characterization of the atmosphere is to some extent an arbitrary decision and a compromise , given that the operation of lasers at the site interferes with the fd measurement and causes detector deadtime ( see section [ subsec : opstypical ] ) .
measurements of atmospheric parameters over the first years of data taking indicate that parameters do not change on much shorter time scales .
however , this assumption needs to be tested with more data .
crucial for this test will be data from the shoot - the - shower mode of lidar operation , taken within minutes of selected cosmic ray events , which can be compared to the hourly measurement .
diameter parabolic mirrors with a focal length of @xmath26 . _ _ ]
as of march 2006 , three of the four lidar stations of the pierre auger observatory have been fully operational .
the first lidar was installed at the los leones site in march 2002 and started data taking soon after , mainly to test the impact on fd operations and define optimal running conditions .
the lidar at the coihueco site began operation in march 2005 , and the lidar at the los morados site started operation in march 2006 .
the fourth fd at loma amarilla is currently under construction and is expected to be completed by december 2006 , with the corresponding lidar station to be operational at the same time
. a schematic diagram of a lidar station is shown in fig.[f : lidarscheme ] .
fig.[f : lidarphoto ] shows a photograph of the los leones lidar setup .
each lidar station has at its core a fully steerable alt - azimuth frame built originally for the eas - top experiment @xcite .
two dc servomotors steer the frame axes with a maximum speed of @xmath27 .
the absolute pointing direction is known to @xmath28 accuracy .
the frame is mounted on a @xmath29 shipping container and is protected from the weather by a fully retractable motorized cover during periods when the lidar is not operating .
frame - steering and cover movements are controlled by an mc-204 motion controller from control techniques , which allows the system to be operated both locally at the site and remotely via ethernet .
each mount is equipped with a uv laser source and mirrors for the detection of the backscattered light .
the choice of a laser for the lidar system is dictated by the following requirements : the wavelength of the laser has to roughly match the dominant wavelength of air fluorescence photons ; the laser power should be low to minimize interference with the fd ; and the repetition rate should be high to reduce data collection time . to meet these requirements ,
the lidars are operated with diode pumped solid state lasers of type dc30 - 351 manufactured by photonics industries .
this laser generates the third harmonic of nd : ylf at 351 nm and is operated at a repetition rate of 333hz and a per - pulse energy of roughly @xmath30 .
the laser wavelength of 351 nm is at the center of the nitrogen fluorescence line spectrum , which extends from about 300 nm to 400 nm , with three main spectral lines at 337 nm , 357 nm and 391 nm @xcite . for the collection of the backscattered light , each lidar telescope uses a set of three
@xmath31 diameter parabolic mirrors with a focal length of @xmath26 ( see fig.[f : lidarscheme ] ) .
the mirrors were produced using bk7 glass coated with aluminum , and the reflective surface is protected with @xmath32 coating to ensure the necessary surface rigidness as well as good uv transmittance .
the average spot size at the focus is @xmath33 fwhm .
each of the mirrors is mechanically supported by a kevlar frame which is in turn fixed to the telescope frame using a three point system .
this allows fine adjustment of the field of view direction to ensure collinearity of the mirrors and the laser beam .
a hamamatsu r7400u series photomultiplier is used for backscatter light detection .
each mirror has its own photomultiplier , so each lidar telescope comprises three independent mirror / photomultiplier systems .
the photomultiplier reaches a gain of @xmath34 at the maximum operation voltage of @xmath35 . to facilitate the light collection from the mirror , the whole active @xmath36-diameter photomultiplier window is used .
background is suppressed by the means of a broadband ug-1 filter with @xmath37 transmittance at @xmath38 and fwhm of @xmath39 .
the use of far more selective interference filters is unfortunately not possible because the extreme speed @xmath40 ( @xmath41 ) of the mirrors leads to a large spread of possible incident angles .
as interference filters are very sensitive to the light impact angle , light has to hit the filter almost orthogonally , or else the transmitted wavelength can shift considerably .
however , one must bear in mind that the lidar is constructed to operate during fd data taking , which is only possible on moonless nights .
a simple absorption filter is therefore sufficient for effective background suppression . due to our specific design requirements ,
a rather long ( 12 m ) signal cable between the photomultipliers and digitizers has to be used . to minimize the signal dispersion as well as rf interference ,
uvf-303 series military standard cables are used .
the signals are digitized using a licel tr40 - 160 three - channel transient recorder . for analog detection
the signal is amplified and digitized by a 12bit 40mhz a / d converter with 16k trace length ( _ current mode _ ) . at the same time
a fast 250mhz discriminator detects single photon events above a selected threshold voltage ( _ photon counting mode _ ) .
a combination of current and photon counting measurement is used in the subsequent analysis to increase the dynamic range of the whole system .
the licel recorder is operated using a pc - linux system through a national instruments digital input - output card ( pci - dio-32hs ) with the comedi interface within the root framework .
the lidar is connected to the fd data acquisition ( fd daq ) system by means of three optical fibers . whenever the lidar system wants to start a measurement , a trigger request
is issued to the fd daq . in response ,
a logic pulse of frequency @xmath42 is generated by the fd gps clock and transmitted to the laser , which fires a single laser pulse for every trigger .
the frequency of @xmath42 corresponds to the maximum acquisition rate of the digitizer for the given memory depth ( @xmath43 ) and sample rate ( @xmath44 ) .
the lidar daq is triggered by the laser synchronization signal generated at every successful laser shot . whenever the lidar direction of measurement comes close or into the fd field of view , a veto signal which prevents fd data acquisition can be generated .
the lidar daq software is organized in several layers to allow remote or unattended operation as well as integration into the central auger daq system .
a run - control program sends the hardware settings and run parameters to the daq program through a communication server .
the daq program controls the licel and photomultiplier settings ( tube gain via high voltage level , photon scaler discriminator level ) , triggering system , telescope steering and cover operation . through the daq program
, the user also controls the shooting directions and the number of laser shots per shooting angle .
current and photon counting traces are summed for 1000 laser shots in the licel , stored in a root file and sent to the online analysis framework for monitoring . .
clouds are seen as strong scattering regions .
the first cloud starts at @xmath45 and shows the maximum peak at @xmath46 . in this example
, a second cloud starts at @xmath47 , shows a peak at @xmath48 and ends at @xmath49 .
the grey curve is a simulated lidar response for a purely molecular atmosphere.__,scaledwidth=70.0% ] as mentioned above , the licel module records backscattered light measurements in current mode and photon counting mode .
current mode maximizes the near - field spatial resolution , but it is suitable only for the first few kilometers where the signal is sufficiently high . as the signal decreases as the square of the distance , photon counting is required in order to obtain information about the atmosphere at large distances . on the other hand , the photon counting saturates in the near - field due to limitations of the licel . at distances less than 5
km from the lidar station , the light level causes a rate greater than 10mhz and the deadtime starts to be an issue ( @xmath50 .
however , the combination of current and photon counting mode covers the full dynamic range of the return signal from near the detector out to a distance of 25 km .
fig.[f : signal ] shows an example for a signal in current and photon counting mode . in order to combine the current and photon counting signals ,
the ratio of the two signals is calculated . a range
in which this ratio is almost constant is identified , usually when the photon count rate is under 10mhz , and signals are merged in that region . in the first kilometers
this condition is not met due to the fact that the photon counting mode saturates for high photon rates ( see fig.[f : signal ] ) . typically , the optimal merging region is 5 - 10 km from the detector , where both the current and photon counting signals are valid .
the derived signal can be parameterized by the so - called lidar equation : @xmath51 where @xmath52 is the signal received at time @xmath53 from photons scattered at a distance @xmath54 from the lidar , @xmath55 is the transmitted laser power , @xmath56 is the laser pulse duration , @xmath57 is the backscattering coefficient , @xmath58 is the optical depth , @xmath16 is the extinction coefficient , and @xmath45 is the effective receiving area of the detector .
@xmath45 is proportional to the overlap of the telescope field of view with the laser beam ( shown in fig.[f : overlap ] ) . over the range of distances where the laser beam and mirror viewing field only partially overlap , it is possible to experimentally determine an overlap function from horizontal scans @xcite .
however , since we restrict our analysis to a range starting at @xmath59 m , where the overlap is complete , we do not apply an overlap correction .
the far limit @xmath60 of our range is defined as the distance at which the signal is negligible as compared to the background , typically 25 km . in the region from @xmath61 to @xmath62 , it is convenient to express the return signal as a function of height @xmath23 in terms of a range - corrected and normalized auxiliary function , @xmath63 : @xmath64 in this equation , @xmath65 is the signal at height @xmath23 , @xmath66 is the optical depth calculated in the range @xmath67 $ ] , and @xmath68 is the lidar inclination angle from the zenith . the normalization height @xmath69 is a fixed height to normalize @xmath70 , chosen such that at @xmath69 , the entire signal is in the field of view of the mirrors .
clouds are visible as strong localized scattering sources .
fig.[f : scloud ] shows an example of @xmath63 in the presence of a cloud layer starting at a height of about 830 m .
also shown is the simulated lidar response for a purely molecular atmosphere .
the figure illustrates that it is possible to resolve several layers of clouds .
a second cloud layer can be seen at an altitude of about 1000 m .
the cloud finding algorithm applied in this analysis is described in detail in section [ subsec : anclouds ] .
the lidar stations were designed to be operated remotely from the observatory s central campus in malarge . there , a computer is used to centralize the operation and issue all the startup commands to the three existing lidar stations and also to monitor the quality of the data being collected . the remote operation of systems with this level of complexity presents a number of challenges . in order to achieve a safe handling of the telescopes , various software routines and hardware devices
have been installed to monitor the performance and status of lidar operations .
these monitoring subsystems include programs used to collect weather related information ( mainly rain and wind speed data ) .
the presence of ambient light and the status of the power supply are monitored as well . in the occurrence of an external event such as rain that could jeopardize the lidar equipment , these subsystems assume control of the station by parking the telescope and closing the cover .
lidar operation starts at astronomical twilight .
after the telescope cover is opened , an initialization procedure is executed to calibrate the incremental encoders used to determine the telescope position .
a webcam located in the interior of the telescope cover is used to confirm visually that these tasks are executed correctly . in this way , before starting a run , the operator has information about the status of the telescope in real time and about the weather conditions at each site through the information being sent to the lidar web site .
following initialization , the system enters an operational mode called _ autoscan_. in autoscan mode , the telescope performs a cycle of steering scripts unless otherwise interrupted until the end of the night .
when the laser is fired , the telescope position is determined by the coordinates contained in these scripts .
there are four main steering strategies : three making up the autoscan pattern and a fourth , shoot - the - shower , that periodically interrupts the autoscan .
these strategies are discussed below , and fig.[f : scandir ] shows , in an azimuthal equatorial projection of the sky , the firing pattern for a typical night of lidar activities at coihueco . * _ continuous scans : _ in this scan , the telescope is moved between two extreme positions with a fixed angular speed while the laser is fired .
the telescope sweeps the sky along two orthogonal paths with fixed azimuthal angle , one of which is along the central fd azimuth ( @xmath71 ) . along both paths ,
the maximum zenith angle is @xmath72 .
the continuous sweeps are constrained to take 10 minutes per path from start to finish . along each path , the lidar performs on the order of 100 measurements with 1000 shots per measurement .
the purpose of these scans is to provide useful data for simple cloud detection techniques ( see section [ subsec : anclouds ] for details ) and to probe the atmosphere for horizontal homogeneity .
an example of the data produced by this kind of scan is shown in fig.[f : contscan ] . * _ discrete scans : _ in this scan , the telescope is positioned at a set of particular coordinates to accumulate larger statistics at a few locations . as indicated in fig.[f : scandir ] , these measurements are performed at 6 discrete zenith angles for 4 different azimuth angles , and directly overhead ( zenith angle @xmath73 ) . to accumulate large statistics ,
12 measurements are performed at each location .
each measurement consists of 1000 laser shots run at 333hz .
the combined duration of the two discrete sweeps is about 30 minutes .
the data obtained in this mode are useful to determine the vertical distribution of aerosols in the atmosphere . * _ horizontal and clf shots : _ in this mode , the laser fires horizontally in 3 different directions towards the location of the clf .
three measurements with 1000 shots per measurement are performed .
the data collected in this scanning mode are used to detect low - lying aerosols and also to determine the horizontal attenuation between the clf and the fd telescopes for comparison with measurements made by other atmospheric monitoring systems .
the total duration is about 3 minutes . *
_ shoot - the - shower ( sts ) : _ this rapid response mode is used to measure the atmospheric attenuation in the line of sight between the fd telescopes and a detected cosmic ray shower .
this scanning mode suspends any of the previously mentioned sweeps .
it will be described in further detail in section [ subsec : opstypical ] .
a complete scanning cycle , excluding sts , takes about 60 minutes to complete .
all scans are therefore performed on an hourly basis .
the maximum length of a lidar running night depends on the length of astronomical twilight and varies over the course of the year from less than five hours during the summer to almost fourteen hours during the winter . as shown in fig.[f : scandir ] , some shooting positions are very close to or inside the field of view of the fd telescopes . in order to prevent the detection of a large number of spurious fd events generated by the lidar shooting activity
, buffer zones have been delimited around the fd fields of view .
every time the laser is fired inside this buffer zone , the fd daq is inhibited in order to avoid any interference .
this is accomplished by sending a veto signal from the lidar to the fd when the laser is ready to fire .
the total fd deadtime introduced by all lidar operations is less than @xmath74 .
to @xmath75 in elevation ) . up
to 60 pointing directions , with 1000 laser shots per pointing , are allowed per sts.__,scaledwidth=55.0% ] a primary design requirement of the lidar system is that it probe the atmosphere along the tracks of cosmic rays observed by the fds .
this function , called shoot - the - shower ( sts ) , exists to recognize unusual and highly localized atmospheric conditions in the vicinity of individual air showers of high interest
. conditions that can affect fd observations at different times of the year include the presence of low and fast clouds , and low - level aerosols due to fog , dust , or land fires .
the basic operation of sts is depicted in fig.[f : lidar - sts ] .
the axis of a cosmic ray air shower , when projected onto the field of view of an observing air fluorescence detector , defines a plane called the shower - detector plane , or sdp . when a lidar station shoots the shower , it performs a series of laser shots within this plane , determining the atmospheric transmission between the shower segment and the fd .
for a given shower , up to 60 pointing directions with 1000 laser shots per pointing are allowed , all within the fd field of view . to operate effectively
, the lidar must shoot showers of importance with a minimum of delay after each shower occurs and with minimum disruption to the normal operation of the fd .
the showers of primary interest for sts are hybrid events involving the fds and the sd , since these set the energy scale of the observatory .
fd stereo events , which are of high energy and typically at large distances from the fds , are also good sts candidates .
physics events are made available to the lidar by picking off the sd and fd data stream sent continuously to the observatory s central data acquisition system ( cdas ) , located on the central campus .
a program running in cdas monitors incoming data , checks for hybrid and stereo events , and calculates sts trajectories .
sts requests are sent to the lidar station data acquisition and control pcs ( lidarpcs ) , which initiate the laser shots .
communication between the lidar stations and cdas , shown schematically in fig.[f : lidar - soft ] , occurs over the network links between the central campus and the four fd subnetworks .
each lidarpc listens on an open tcp port for sts requests from the fd / sd monitor running in cdas .
as the lidar shoots , it returns basic diagnostic information , such as shooting directions and pmt signal peaks , back to cdas for display . a tcp socket server operating in cdas manages the bidirectional flow of data packets between the central campus and the lidar stations .
lidars receiving sts requests from cdas automatically stop the default shooting operation ( autoscan ) , move to the fd field of view , initiate the sts , and then resume the autoscan when sts is complete .
if the lidar receives an sts request while shooting another shower , the request is pushed into a queue for later processing . to minimize the fd dead - time introduced by sts , the lidar software cuts on the number of sd tanks participating in the event , which is a rough measure of the primary cosmic ray energy .
this cut reduces the number of requests to several per fd site per night .
a further reduction in the rate is achieved by rejecting events caused by known artificial light sources in the detector , such as other lidar stations and the clf .
in addition , the intensity of the high - repetition laser means that the lidar must carefully avoid incidents of cross fire into other unvetoed fds during sts .
therefore , angular windows in azimuth and in zenith are defined around each fd ; the lidar is forbidden from entering these windows during sts .
the operation of sts in the field of view of the photomultiplier cameras raises the issue of possible long term effects on the phototubes themselves .
although fd data acquisition is inhibited during sts , the photomultipliers continue to operate at high voltage during their exposure to powerful nearby laser shots .
however , since the shooting rate is one to two shots per fd per night , the effect of the sts is not significant in comparison to other strong and persistent light sources .
these sources include the typical night sky background with its large number of bright stars , and heavy lightning activity during the summer months .
in addition , a comparison of tube noise directly before and after sts events shows no significant effect of the shooting activity on the tube noise . after subtraction of a simulated return that assumes a purely molecular atmosphere.__,scaledwidth=70.0% ]
in standard operation mode , the two goals of the lidar system are ( 1 ) to provide hourly measurements of vertical aerosol optical depth and backscatter coefficient as a function of height at each fd site , and ( 2 ) to provide hourly information on cloud coverage , cloud height , and the optical depth of cloud layers .
the analysis of lidar scans to extract the vertical aerosol optical depth is discussed in detail in @xcite . here
, we will describe the analysis of cloud coverage and optical depths of cloud layers .
these measurements are particularly important for the accurate estimation of the fd detection aperture .
while they do not affect the sd aperture , clouds within the fields of view of the fds distort the light profile of showers and can compromise the energy calibration . as described in section[subsec : signaltreatment ] , clouds are detected as strong localized scattering sources , and the timing of the backscattered light gives the distance from the cloud to the lidar .
the return signal is expressed in terms of the @xmath76 function ( eq.[eqs ] ) . to automatically detect clouds
, it is useful to calculate a new function @xmath77 by subtracting the @xmath76 function for a simulated purely molecular atmosphere from the measured @xmath76 function : @xmath78 } + 2\tau_{mol}(h;h_n)\sec(\theta ) \nonumber \\ & = & \ln{\left[\frac{\beta(h)}{\beta(h_n ) } \frac{\beta_{mol}(h_n)}{\beta_{mol}(h)}\right ] } -2\tau_{aer}(h;h_n)\sec(\theta ) \hspace{10pt } , \label{eqssub}\end{aligned}\ ] ] where @xmath69 is the normalization height ( as in eq.[eqs ] ) , @xmath79 is the backscattering coefficient , @xmath66 is the optical depth calculated in the range @xmath67 $ ] , and @xmath68 is the lidar inclination angle from the zenith .
the properties of the molecular atmosphere at the pierre auger observatory site are very well - known thanks to an extensive balloon - launching program which has produced a detailed database including temperature , pressure , and density profiles over the site @xcite .
data from these balloon flights , typically performed every 5 days , are used to create monthly models of the molecular atmosphere at the malarge site . in the lidar analysis we use these monthly models for the calculation of @xmath77 . by using the monthly models , seasonal variations in the molecular atmosphere
are accounted for , but day - to - day variations , which can be substantial @xcite , are not .
these short term variations do not affect the cloud finding algorithm , and their effect on the calculation of the optical depth of the cloud layer is negligible .
as shown in fig.[f : smolcloud ] , @xmath77 is approximately constant before the cloud , and has a non - zero slope inside the cloud .
we apply a second - derivative method @xcite to identify cloud candidates and obtain the cloud thickness . in cases where several cloud layers are present
, the influence of partial overlaps between cloud layers causes an inaccurate estimation of the optical depth of the first cloud . for this reason ,
very near clouds separated by less than 10 m are treated as a single cloud layer , and the optical depth is calculated only for the combination of the layers . in order to reduce the possibility of a spurious cloud detection , we take advantage of the fact that each lidar has three individual mirrors and photomultipliers , so the same signal is detected independently several times . to discriminate against noise
, we require a cloud to be detected in multiple mirrors .
signals from different mirrors of the same lidar are therefore compared before a cloud is stored and a cloud detection is reported .
the application of the cloud detection algorithm to lidar scans can provide a variety of useful data .
collecting all the scans during an hour , the lowest layer of clouds is identified and the proportional sky coverage is calculated .
fig.[f : allheights ] shows the distribution of the height of the lowest detected cloud layer ( upper plot ) and the height of the lowest layer as a function of time ( lower plot ) for data taken between 24 october 2005 and 7 march 2006 with the coihueco lidar .
typical cloud layers can be identified , in particular a recurrent cloud layer between 2000 m and 3000 m above sea level .
once the lowest clouds have been identified , their effect on the light propagation is estimated .
the turbidity of a layer of thickness @xmath80 can be described by a transmission factor @xmath81 where @xmath82 is the total optical depth .
consider , for instance , a cloud at a height @xmath83 that ends at a height @xmath84 . from a lidar scan , the auxiliary function @xmath63 , given by eq.[eqs ] , is obtained , and @xmath77 is calculated by using eq.[eqssub ] .
the difference between the values assumed by @xmath77 in @xmath83 and @xmath84 is : @xmath85 } -2\tau_{aer}(h_f;h_a)\sec{\theta } \hspace{10pt } , \label{eqsdiff}\ ] ] where @xmath86 is due to scattering and absorption of the light by the cloud . at heights greater than 2 km above ground level
a quasi - molecular atmosphere can be assumed in the proximity of clouds .
therefore , @xmath87 at @xmath83 and @xmath84 , and eq . [ eqsdiff ] becomes : @xmath88 in this way , the cloud optical depth can be estimated directly from the signal : @xmath89 using this information , the mean optical depth inside the lowest cloud layer is calculated for every hour of fd data taking . as an example of the capability of the lidar system to detect and characterize clouds , fig.[f : odvsheight ] shows the height of the lowest cloud layer versus the optical depth inside the lowest cloud layer observed with the coihueco lidar between 24 october 2005 and 7 march 2006 .
the plot indicates that low altitude clouds tend to cover a wider range of optical depths than clouds at higher altitude .
the atmospheric monitoring facilities of the pierre auger observatory also include customized infrared cloud cameras at each fd site @xcite .
these cameras scan the entire sky every 15 minutes .
the pictures are processed and used to fill a database with information on the cloud coverage . to first order ,
the images from the cloud camera provide a crude cross - check of the efficiency of the lidar cloud finding algorithm . in practice ,
the two cloud analyses are complementary , with the lidar providing the height and optical properties of cloud layers and the cloud cameras providing a complete two - dimensional image of the cloud coverage .
in addition to operating the elastic backscatter lidars , the pierre auger collaboration is currently also evaluating the usefulness of a raman lidar operated in conjunction with the elastic lidar at los leones .
the raman technique is based on inelastic raman scattering , a secondary component of molecular scattering in the atmosphere .
the inelastic component is suppressed compared to elastic rayleigh scattering since the raman scattering cross section is about three orders of magnitude smaller than the corresponding rayleigh cross section . in raman scattering ,
the scattered photon suffers a frequency shift that is characteristic of the stationary energy states of the irradiated molecule . with raman spectroscopy ,
it is therefore possible to identify and quantify traces of molecules in a gas mixture .
raman lidars operate on the same principle as elastic lidars , except that raman lidars feature a spectrometer - type receiver to discriminate the lidar returns according to wavelength .
our receiver has 3 channels to detect the light intensity at various wavelengths .
one channel collects the elastic lidar return , while the others correspond to the atmospheric oxygen and nitrogen raman lidar backscatter , at about 375 nm and 387 nm , respectively .
the raman lidar technique overcomes a major weakness of the elastic lidar measurements , in which the aerosol scattering contribution is present both in the transmission and backscattering components . extracting the aerosol attenuation from elastic lidar data
therefore requires some assumptions regarding aerosol optical properties , mainly the relationship between aerosol backscatter and extinction . for the raman lidar ,
the aerosol scattering contribution is only present in the transmission , so the only assumption to be made is the wavelength dependence of the aerosol transmission coefficient .
if the raman lidar returns are measured for several wavelengths simultaneously , for example for molecular nitrogen and oxygen , even the wavelength dependence can in principle be retrieved without any assumptions .
in addition , if the elastic signal at the laser wavelength is measured simultaneously , a comparison between the elastic rayleigh and inelastic raman return can be used to give an independent estimate of the aerosol content as a function of height .
a practical disadvantage of the raman lidar technique is the small raman molecular cross section . as a consequence ,
the laser source has to be operated at high power and care has to be taken not to disturb the fd operation .
currently , raman lidar runs at the pierre auger observatory are only performed right before and after the regular fd operation .
the raman lidar system consists of a nd : yag laser ( ultra cfr ) manufactured by big sky laser technologies , inc .
the emitted pulses at 355 nm have a line width of @xmath90 , a pulse energy of about 10mj at a repetition rate of 20hz , and a pulse duration of about 6ns .
the full angle divergence is less than 1mrad .
the emitted laser pulses are directed in the atmosphere ( zenith direction ) with a @xmath91 steering mirror , which has a pointing sensitivity of about 0.1mrad .
the raman lidar has to meet a few general requirements
. it must be capable of sampling a range of altitudes from the ground to 7000 m , be optimized to detect weak raman backscattering from atmospheric nitrogen and oxygen , and efficiently suppress background light and cross - talk between the different channels .
these conditions impose restrictions on the choice of the receiving optics , and on the design of the dichroic beam splitters and interference filters to be used in the wavelength discriminating detectors .
they also impose the use of notch filters providing an additional suppression of the strong elastically backscattered light in the raman channels . in the setup of the raman lidar system ,
the backscattered light is collected by a zenith pointing telescope ( 50 cm diameter parabolic mirror , focal length 150 cm ) through a 10 m long multi mode optical fiber coupled to the telescope with an aspherical lens .
the fiber transports the return light to the 3-channels beam separator : a combination of beam splitters , interference filters , notch filters , and neutral density filters . after passing the final stage of the beam separator unit ,
the light is collected by hamamatsu r1332 photomultipliers .
the voltage pulses at the output of the photomultipliers are amplified by a fast pre - amplifier ( eg&g ortec , model 535 ) . a discriminator ( caen n224 ) reduces the dark counts and forms 30ns wide nim pulses .
these pulses are detected by a pxi-5620 digitizer ( national instruments , maximum sampling rate 64msample / s ) usually in a 0.200ms time window ( 30 km range ) .
the number of pulses in each of 1000 time windows ( 200ns wide , corresponding to a distance resolution of 30 m ) are counted , stored and displayed in a personal computer , after summing over a number of laser shots .
the maximum counting rate is kept below 10mhz so that pulse pile - up effects are negligible .
the data production is presently quite small : 576 kbyte for 40 minutes of data taking .
fig.[f : raman ] shows the typical returns in the elastic and nitrogen / raman channels .
note that the presence of a cloud layer at about 4.8 km height appears as an attenuation in the raman lidar return , since , as described above , the aerosol scattering is only present in the transmission . in the elastic lidar signal , on the other hand
, the cloud appears as an increase in the signal due to the enhanced backscattering .
the raman lidar is remotely operated , sharing some services with the elastic lidar already operating at los leones .
currently , the main goal of the raman lidar is to measure the vertical aerosol optical depth . since the raman measurements have to be taken outside fd operations , the measurement obtained with this technique is not directly usable for the analysis of fd data .
however , it can be used to cross - check and validate the estimates of the aerosol content made by the subsequent elastic lidar operation .
since march 2006 , three of the four lidar stations are routinely operated each night of fd data taking by a designated lidar shift taker at the central campus .
the fourth site is currently under construction and will start data taking in 2007 .
the three operating lidars perform continuous and discrete scans of the local atmosphere , and the data are analyzed for cloud coverage , cloud height , and the optical depth of cloud layers .
an analysis of the aerosol scattering and absorption parameters will be added in the near future . as the part of the lidar program that has the largest potential to interfere with the fd operation , the shoot - the - shower program is still in a testing phase .
algorithms to maximize the efficiency of sts are currently under development .
this includes the implementation of software to filter background events that can erroneously trigger the sts , such as lightning .
another future upgrade currently under consideration is the installation of new mirrors with a larger focal length to reduce the speed of the lidar optics .
the large field of view of the current system implies that the background rate is high and the laser beam enters the field of view of the mirrors very early . to lower the amount of background light and overcome saturation at short range ,
a segmented mirror with a 1 m diameter and 1.1 m focal length is currently under development .
we are grateful to the following agencies and organizations for financial support : comisin nacional de energa atmica and gobierno de la provincia de mendoza , argentina ; infn ( istituto nazionale di fisica nucleare ) ( italy ) ; slovenian research agency ; national science foundation ( usa ) ( contract numbers nsf - phy-0500492 , nsf - phy-0134007 ) ; department of energy ( doe ) office of science ( usa ) ( de - fg03 - 92er40732 ) .
r. cester ( for the pierre auger collaboration ) , _ atmospheric monitoring at the pierre auger observatory _ , in proc .
@xmath92 int .
cosmic ray conference , pune , india ( 2005 ) .
p. mantsch ( for the pierre auger collaboration ) , _ the pierre auger observatory status and progress _ , in proc .
@xmath92 int .
cosmic ray conference , pune , india ( 2005 ) ( astro - ph/0604114 ) .
roberts , j. phys .
g : nucl . part .
phys 31(2005)1291 .
a. bucholtz , appl .
g. mie , ann .
van de hulst , `` light scattering by small particles , '' wiley , new york ( 1957 ) .
mccartney , `` optics of the atmosphere : scattering by molecules and particles , '' wiley , new york ( 1976 ) .
a. filipi , m. horvat , d. veberi , d. zavrtanik , and m. zavrtanik , astropart .
b. fick et al . , journal of instrumentation 1(2006)p11003 .
m. aglietta et al . , nuovo cimento 16(1993)813 .
f. kakimoto et al . , nucl .
meth . a 372(1996)527 .
y. sasano , h. shimizu , n. takeuchi , and m. okuda , appl .
. j. blmer et al .
( for the pierre auger collaboration ) , _ atmospheric profiles at the southern pierre auger observatory and their relevance to air shower measurement _ , in proc . @xmath92 int . cosmic ray conference ,
pune , india ( 2005 ) .
b. wilczyska et al . ,
_ variability of atmospheric depth profiles _ , in proc .
@xmath92 int .
cosmic ray conference , pune , india ( 2005 ) .
a. tonachini , phd thesis , university of torino , in preparation . | the pierre auger observatory in malarge , argentina , is designed to study the origin of ultrahigh energy cosmic rays with energies above @xmath0ev .
the energy calibration of the detector is based on a system of four air fluorescence detectors . to obtain reliable calorimetric information from the fluorescence stations ,
the atmospheric conditions at the experiment s site need to be monitored continuously during operation .
one of the components of the observatory s atmospheric monitoring system is a set of four elastic backscatter lidar stations , one station at each of the fluorescence detector sites .
this paper describes the design , current status , standard operation procedure , and performance of the lidar system of the pierre auger observatory . , , , , , , , , , , , , , , , , , , , , , , , , , , , , , and ultrahigh energy cosmic rays ; air fluorescence detectors ; atmospheric monitoring ; lidar 07.60.-j , 42.68.wt , 92.60.-e , 96.50.sd |
the x - ray luminous nuclei of active galaxies ( agn ) are believed to be powered by accretion of gas onto supermassive black holes . there is now growing evidence that many nearby _ non - active _ galaxies harbor quiescent , weakly or non - accreting black holes ( see reviews by kormendy & gebhardt 2001 ; richstone 2002 ) .
studies of the abundance of black holes , of their masses and their spins , shed light on the mechanisms of black hole growth and of galaxy and agn formation and evolution .
possibly the most direct means of detecting supermassive black holes at the centers of galaxies , and an unavoidable consequence of their existence , would be occasional tidal disruption of stars and subsequent accretion of their debris by these supermassive black holes ( e.g. , hills 1975 ; gurzadyan & ozernoi 1979 ; carter & luminet 1982 ; rees 1988 , 1990 ; wang & merritt 2004 ) .
the events would appear as luminous flares of radiation emitted when the stellar debris is accreted by the black hole .
stellar capture and disruption is apart from accretion of gas and black hole merging one of the three major processes studied in the context of black hole growth ( e.g. , frank & rees 1976 ) .
the relative importance of these three processes in feeding black holes is still under investigation .
zhao , haehnelt , & rees ( 2002 ) and merritt & poon ( 2003 ) recently pointed out that tidal capture may play an important role in explaining the notable @xmath1 relation of galaxies .
given the intense theoretical attention the topic of stellar tidal disruption receives ( see 3 of komossa 2002 , and references therein ) , it is of great interest to see whether these events do occur in nature , and how frequent they are . making use of the unique capability provided by the all - sky survey ( rass , voges et al .
1999 ) of the x - ray satellite _ rosat _ , giant x - ray flares from the directions of a few nearby galaxies were discovered ( komossa & bade 1999 ; komossa & greiner 1999 ; grupe et al .
1999 ; greiner et al .
these flares were characterized by huge peak x - ray luminosities reaching @xmath2 ergs @xmath3 , large amplitudes of decline , ultrasoft x - ray spectra , and absence of seyfert activity in ground - based optical spectra ( see komossa 2002 for a review ) .
the target of the present study , rxj12421119 , was first detected by _ rosat _ in 1992 ( komossa & greiner 1999 ) during a pointed observation with the position sensitive proportional counter ( pspc ) . at that time it showed a very soft x - ray spectrum with @xmath4 kev , and an x - ray luminosity of @xmath5 ergs @xmath3 , which is exceptionally large
given the absence of any signs of seyfert activity of rxj12421119 in ground - based ( komossa & greiner 1999 ) and _ hubble space telescope _
( _ hst _ , gezari et al . 2003 ) optical spectra .
the association with one of the two , previously unstudied non - active galaxies at redshift @xmath6 located in the x - ray position error circle ( named rxj12421119a and rxj12421119b for lack of a better designation ) however , remained uncertain because of the large _ rosat _ positional uncertainty , in this case 40@xmath7 ( figure 1 ) . considering the extraordinary properties of this and a few similar x - ray events , particularly their enormous luminosity output , it is of utmost importance to understand what is happening in these systems .
for the first time , we have now looked at one of these flare events with high spatial and spectral resolution , in order to confirm the counterpart , follow the long - term temporal behavior , study the spectral evolution and measure the post - flare spectrum , and to use these results to test the favored outburst model : tidal disruption of a star by a supermassive black hole . among the few known x - ray
flaring galaxies , rxj12421119 was our target of choice for follow - up x - ray observations because it flared most recently , so the probability of catching the source in the declining phase was highest . also , because there were two galaxies in the _ rosat _ x - ray error box , the correct counterpart needed to be confirmed .
a hubble constant of 50 km s@xmath8 mpc@xmath8 is used throughout this paper .
we observed the field of rxj12421119 with the advanced ccd imaging spectrometer ( acis - s ) instrument on - board the _ chandra x - ray observatory _ ( weisskopf et al .
2002 ) for 4.5 ks on 2001 march 9 .
the x - ray photons from the target source were collected on the back - illuminated s3 chip of acis .
the _ chandra _ data allow us to locate precisely the counterpart of the x - ray source .
we find that the x - ray emission peaks at coordinates r.a.=@xmath9 , decl.=@xmath10 ( equinox j2000 ) . within the errors ,
this position is coincident with the center of the optically brighter of the two galaxies , rxj12421119a ( figure 1 ) .
an offset in position of @xmath11 in r.a . can be traced to residual uncertainties in the absolute pointing accuracy of the x - ray telescope .
neither _ chandra _ nor _ xmm - newton _ detected any other x - ray source within the _ rosat _ x - ray position error circle , including from the optically fainter galaxy rxj12421119b .
we do not find evidence for significant source extent .
the photons detected by _
chandra _ from rxj12421119a fall within a radius of @xmath12 , and the radial profile is consistent with a point source .
this indicates that we are still seeing late - phase flare - related emission rather than persistent , extended x - ray emission originating from the host galaxy itself .
a further result , immediately obvious upon inspection of the _ chandra _ data , is the huge drop in x - ray flux of rxj1242 - 1119 compared to its last observation by _
rosat_. only 18 source photons were detected by _
chandra _ ; the implied amplitude of variability is quantified below . given the strong indications , based on the _ chandra _ observation , that the x - ray emission from rxj12421119a is still flare - dominated , we asked for a target - of - opportunity observation with _ xmm - newton _ in order to obtain , for the first time , a good - quality x - ray spectrum of one of the flaring sources , and to follow the long - term spectral evolution before the source had declined to non - detectability . while the strength of _ chandra _ is its high spatial resolution of better than @xmath12 , _ xmm - newton _ has higher throughput ( the 18 source photons registered with _
chandra _ do not allow spectral fitting ) .
the _ xmm - newton _ ( jansen et al .
2001 ) observation of rxj12421119 was performed on 2001 june 21 - 22 , with a duration of 24.3 ks .
data from the epic - pn , mos1 , and mos2 detectors were used for analysis .
the observation was first checked for flares in the background radiation ; none were detected .
photons were then extracted from a circular area of radius @xmath13 centered on the target source . for the mos cameras
the background was determined in an annulus around the source with an inner radius of @xmath14 and an outer radius of @xmath15 .
the background counts for the pn observation were selected in a circular region close to the target source .
the total source count rate , measured in the epic - pn detector , amounts to 0.0137 counts s@xmath8 in the @xmath16 kev energy band .
no systematic decrease ( or increase ) of the count rate is present during the observation . the x - ray _ spectrum _ yields important information on the physics governing the post - flare evolution .
x - ray emission from rxj12421119a is detected in all three of the following energy intervals : 0.30.75 kev , 0.751.5 kev , 1.55 kev , with count rates of 0.0073 , 0.0040 , and 0.0024 counts s@xmath8 , respectively .
for spectral analysis , source photons were binned such that each spectral bin contains @xmath17 photons .
the resulting spectrum is best described by a power law with photon index @xmath18 ( @xmath19 ; figure 2 ) .
we do not find evidence for strong excess cold absorption .
the absorption towards rxj12421119a is consistent with the galactic value of @xmath20 @xmath21 .
thermal ( raymond - smith and mekal ) x - ray emission models yield significantly worse spectral fits if the metal abundance of the x - ray emitting gas is constrained not to fall significantly below 0.1 times the solar value . using the best - fitted power - law spectral model , the _ xmm - newton _ `` low - state '' luminosity of rxj12421119 is @xmath22 ergs s@xmath8 in the 0.12.4 kev band . compared to the peak luminosity of @xmath23 ergs s@xmath8 measured with _ rosat _ , this corresponds to a factor @xmath24 decline . finally , we used the optical monitor ( om ) data to estimate the @xmath25 and @xmath26 magnitudes of rxj12421119a ( and b ) .
the om observations were performed in standard image mode , with @xmath26 , @xmath25 , and two broad - band uv filters .
five high- and one low - resolution image per filter were collected with an exposure time of 1000 s each .
the data files were processed with _
sas 5.4.1 _ and with the task _ omichain _ to perform required corrections for tracking , bad pixels , fixed pattern , and flat - field .
both galaxies are detected in the @xmath26 , @xmath25 , and @xmath27 filter bands , and their angular separation is large enough to avoid confusion .
total source counts were measured for each galaxy in an aperture radius of 12 pixels as recommended to account for the broad wings of the om point - spread function .
corrections for coincidence losses and dead - time were applied to the total count rates , and then background was subtracted .
a mean corrected count rate was then computed for each filter band . to estimate the magnitudes ,
the most recent zero - point corrections ( as of 2003 november ) were applied .
this yields the following @xmath25 and @xmath26 magnitudes for rxj12421119a(b ) : @xmath28 mag and @xmath29 mag .
we have utilized the complimentary abilities of the x - ray observatories _ chandra _ and _ xmm - newton _ to study the peculiar _ rosat _ source rxj12421119 . with _
chandra _ , we found that the source of the bright x - ray emission detected by _
rosat _ is the galaxy rxj12421119a .
the flare emission has declined dramatically by a factor of @xmath24 , but has not yet faded away completely .
_ xmm - newton _ then provided us with the best post - flare spectrum ever taken of any of the few flaring , optically inactive galaxies .
the spectrum is of power - law shape ; there is no evidence for excess absorption , and the post - flare spectrum is harder than the flare maximum .
there is an additional argument that the detected point - like low - state x - ray emission does not arise from the ism of the host galaxy , but is related to the mechanism that produced the flare itself .
compared with the blue luminosity of rxj12421119a , inferred from the extinction - corrected blue magnitude @xmath30 mag measured with the om , the x - ray emission of this galaxy is still very high , even in its low state .
its ratio of x - ray to blue luminosity , @xmath31/@xmath32 , is above the upper end observed so far for field early - type galaxies ( irwin & sarazin 1998 ; osullivan et al .
2001 ) .
while the x - ray observations presented here demonstrate that the post - flare emission is associated with the galaxy rxj12421119a , it was established previously that the optical spectrum of rxj12421119a is that of a non - active galaxy ( komossa & greiner 1999 , gezari et al .
in addition to other arguments presented earlier , the absence of any significant excess x - ray absorption above the galactic value argues against the scenario of an extremely x - ray variable active galactic nucleus that is completely obscured optically .
we find that the galaxy rxj12421119a is not detected in the nrao vla sky survey ( nvss , condon et al .
1998 ) at 1.4 ghz , also consistent with the absence of a _ permanent _ ( hidden ) active nucleus .
the x - ray observations reported here place important constraints on the total duration of the flare maximum . the source was not detected in 1990 in a short exposure during the rass .
it was `` on '' in 1992 , and had dropped dramatically in flux during the _ chandra _ and _ xmm - newton _ observations in 2001 , limiting the `` on - state '' to @xmath33 yr .
the high outburst luminosity and amplitude of variability strongly suggest that a supermassive black hole at the center of rxj12421119a is the ultimate power source .
the new observations presented here are consistent with and corroborate the previously favored tidal disruption scenario to explain this flare event .
independent of the x - ray properties of rxj12421119a , its blue magnitude measured by the om can be used to estimate the mass of the black hole at the center of the galaxy .
the correlation between the absolute blue magnitude of the bulge of an elliptical galaxy and the mass of the central black hole ( ferrarese & merritt 2000 ) predicts a black hole mass of @xmath34 .
this estimate is uncertain by a factor of several , given the scatter in the relation between bulge luminosity and black hole mass ( e.g. , ferrarese & merritt 2000 , mclure & dunlop 2002 ) .
the observations are consistent with either partial disruption of a giant star , or complete disruption of a solar - type star ( e.g. , hills 1975 ) ; or for larger black hole masses , if the black hole is of kerr type and the star approaches from a favorable direction ( beloborodov et al .
1992 ) ] , where either only part of the debris from the solar - type star is accreted while the rest becomes unbound , or where part of the debris is accreted in a radiatively inefficient mode . to estimate ( a lower limit on ) the amount of stellar debris accreted by the black hole
, we forced the high - state and low - state luminosities to follow a @xmath35 law expected for the `` fall - back '' phase ( e.g. , rees 1990 ; li et al .
2002 ) of tidal disruption .
the integral @xmath36 then gives the total emitted energy , @xmath37 ergs , where we started the integration at the time the source was first observed in the high - state by _
rosat_. this requires accretion of @xmath38 of stellar material , where @xmath39 is the efficiency .
although the x - ray emission of rxj12421119a dropped substantially , does this necessarily mean that the _ total _ accretion luminosity declined , or is it possible that the emission just shifted out of the x - ray band ?
part of the flux may have softened into the unobservable euv part of the spectrum , as expected in some tidal disruption models ( e.g. , cannizzo et al .
1990 ) , but it is more likely that the total luminosity has declined since outburst .
otherwise , we might expect to see optical emission lines excited by the hypothesized bright euv continuum ; such emission lines do not appear in _ hst _ spectra of rxj12421119a taken several years after the outburst ( gezari et al .
2003 ) . with all observational tests possible
now performed , more detailed comparison with theory has to await refined model calculations of stellar disruption .
the results presented here demonstrate the effectiveness of combining the abilities of the two most powerful x - ray observatories in orbit to study one of the most extreme variability events ever recorded among galaxies . in our interpretation of these observations
, we are seeing the post - disruption phase of a close encounter of a star with a central supermassive black hole , into which some of the tidal debris is accreting
. continued x - ray monitoring of rxj12421119a will enable us to follow the expected further decline in luminosity .
future x - ray ( all)-sky surveys planned for missions like _ duo _ , _ rosita _ , _ lobster _ , and _ maxi _ will be very useful in finding new flare events , while their detailed study will become possible with future high - throughput x - ray missions like _ xeus _ and _ con - x_. we may then be able to probe the regime of strong gravity , since the temporal evolution of the flare x - ray emission is expected to depend on relativistic precession effects in the kerr metric . the results presented here will impact further model calculations of the tidal disruption process , which are still complex and challenging , and are expected to motivate new studies of the flare host galaxies as well as an expanded search for similar flare events in existing and future _ chandra _ and _ xmm - newton _ archives , including the deep fields .
this work is based on observations obtained with _ xmm - newton _ , an esa science mission with instruments and contributions directly funded by esa member states and nasa , and on observations obtained with the nasa mission _
chandra_. in germany , the _ xmm - newton _ project is supported by the bundesministerium fr bildung und forschung / deutsches zentrum fr luft- und raumfahrt , the max - planck society and the heidenhain - stiftung .
we thank the _ xmm - newton _ and _ chandra _ teams for their help which made these observations possible , and peter edmonds , joachim trmper and an anonymous referee for their useful comments .
gratefully remember hartmut schulz for ongoing discussions on many topics in astrophysics including the flares , for sharing his broad knowledge and deep insight into physics , for patiently listening when i needed somebody to talk to , for intense collaborations on ngc6240 ever since i started in astrophysics , and for sharing his enthusiasm for ngc4151 .
hartmut passed away in august 2003 . | in recent years , indirect evidence has emerged suggesting that many nearby non - active galaxies harbor quiescent supermassive black holes .
knowledge of the frequency of occurrence of black holes , of their masses and spins , is of broad relevance for studying black hole growth and galaxy and agn formation and evolution .
it has been suggested that an unavoidable consequence of the existence of supermassive black holes , and the best diagnostic of their presence in non - active galaxies , would be occasional tidal disruption of stars captured by the black holes .
these events manifest themselves in form of luminous flares powered by accretion of debris from the disrupted star into the black hole .
candidate events among optically non - active galaxies emerged in the past few years .
for the first time , we have looked with high spatial and spectral resolution at one of these most extreme variability events ever recorded among galaxies . here ,
we report measuring a factor @xmath0 drop in luminosity of the x - ray source rxj12421119 with the x - ray observatories _ chandra _ and _ xmm - newton _ , and perform key tests of the favored outburst scenario , tidal disruption of a star by a supermassive black hole .
we show that the detected `` low - state '' emission has properties such that it must still be related to the flare .
the power - law shaped post - flare x - ray spectrum indicates a `` hardening '' compared to outburst .
the inferred black hole mass , the amount of liberated energy , and the duration of the event favor an accretion event of the form expected from the ( partial or complete ) tidal disruption of a star . |
dissipation of stellar tides and oscillations is often considered to be mainly due to the turbulent flow in their convective zones .
usually the effects of the turbulent flow are parametrized by some sort of effective viscosity coefficient .
clearly the situation is not as simple as that and the usual `` fix '' is to allow this viscosity coefficient to depend on the perturbation being dissipated , most notably its frequency and perhaps direction of the shear it creates .
completely analytical treatments start by assuming a kolmogorov spectrum for the turbulent flow and combine it with some prescription for the effectiveness of eddies in dissipating perturbations of the given period . since kolmogorov turbulence is isotropic the direction of shear is unimportant in such prescriptions .
two such prescriptions have been used .
the first , proposed by @xcite , states that when the period of the perturbation ( @xmath0 ) is shorter than the turnover times ( @xmath1 ) of some eddies , their dissipation efficiency should be proportional to the fraction of a churn they manage to complete in half a perturbation period , in other words , the effective viscosity coefficient scales like : @xmath2.\ ] ] the second prescription is due to @xcite and @xcite .
they argue that eddies with turnover times much bigger than the period of the perturbation will not contribute appreciably to the dissipation , and hence the effective viscosity should be dominated by the largest eddies with turnover times @xmath3 .
then the kolmogorov prescription of turbulence predicts that the effective viscosity will scale as : @xmath4\ ] ] zahn s prescription has been tested against tidal circularization times for binaries containing a giant star @xcite , and is in general agreement with observations .
also zahn s prescription is in better agreement with observed tidal dissipation of binary stars in clusters @xcite and with the location of the red edge of the cepheid instability strip @xcite .
the less efficient prescription has been used successfully by @xcite , @xcite and @xcite to develop a theory for the damping of the solar @xmath5-modes . in this case
the more effective dissipation would require dramatic changes in the excitation mechanism in order to explain the observed amplitudes .
finally @xcite developed a perturbative derivation of the convective viscosity , which for a kolmogorov scaling , gives a result that is closer to the less efficient goldreich & nicholson viscosity than it is to zahn s . while providing a firmer theoretical basis for the former scaling
, this does not resolve the observational problem of insufficient tidal dissipation .
the development of 2d and 3d simulations of solar convection hint at a possible resolution of this problem .
the convective flow that these simulations predict is fundamentally very different from the assumed kolmogorov turbulence @xcite .
the first major difference is that the frequency power spectrum of the velocity field is much flatter than the kolmogorov power spectrum , and hence one might expect that the dissipation will decrease significantly slower as frequency increases relative to the kolmogorov case .
another major difference is that the velocity field is no longer isotropic and hence one would expect it to react differently to shear in different directions . that is , if we would use an effective viscosity coefficient , it should be a tensor and not a scalar quantity .
as a first step in investigating that possibility @xcite , we applied the perturbative approach developed by @xcite to a numerical model of solar convection ( robinson et al .
2003 ) to find the scaling of the components of the effective viscosity tensor with frequency .
somewhat unexpectedly , we found that the scaling closely follows zahn s prescription , even though when the same approach is applied to kolmogorov turbulence it gives results closer to those of goldreich and collaborators . that left a lot of questions unanswered .
for example , are the effects of turbulence anything at all like that of molecular viscosity , and is taking only the lowest order term in a power series expansion of the energy dissipation rate an acceptable approximation ?
the perturbative approach also assumes that the spatial scale of the external shear being dissipated is large compared to all convective scales , and hence can be assumed a linear function of position , but it would be interesting to see how quickly dissipation efficiency is lost as the spatial scale of the external shear decreases . to tackle these questions we adapted the anelastic spectral code of @xcite to simulate a convectively unstable box in which we are able to place external , time dependent shear as part of the evolution equations and observe the effects of the turbulent convective flow directly . in order to make the code suitable to simulate convection we added heat diffusion , that allows for the supply the heat that will drive the convective motions
this in turn necessitated that we allow for a more general background state slightly modified treatment of the pressure and the introduction of temperature boundary conditions .
in addition we added an external forcing term and removed the kepplerian shear and the accompanying it shearing coordinates .
the resulting code is ideal for our purposes for the following reasons : 1 .
the anelastic approximation means no shock or sound waves can exist and hence the time steps can be much larger than for a fully compressible code , and yet the anelastic approximation allows phenomena such as convection to occur unlike in completely incompressible codes .
this allows us to get much closer to the actual time scales of interest ( usually of order hours or days ) than a fully compressible code would .
the boussinesq approximation has those same properties , however it is not suitable for simulating stratified convection where the buoyancy does not depend entirely on the temperature .
spectral codes are more efficient at simulating turbulent flows than finite difference codes because their spatial accuracy is exponential rather than a power law , so they are capable of reliably reproducing the turbulent flow even at modest resolution
. 3 . the grid happens to have higher density near the top and bottom of the box where it is needed in order to resolve the boundary layers that develop there due to the boundary conditions we impose .
in this paper we present the details of the modified code as well as a perturbative estimate of the efficiency of turbulent viscosity . in section
[ sec : equations ] we derive the anelastic approximation to the navier - stokes equations modified to include a time dependent shear . in section [ sec : numerics ] we present the numerical algorithm . in section
[ sec : tests ] we show the tests we ran to confirm that our implementation actually evolves the desired equations , and verify that the numerical scheme is indeed second - order accurate in time .
finally , in section [ sec : results ] we use a perturbative calculation to get an estimate of the effective turbulent viscosity in our simulated box .
we define upfront the following quantities describing the fluid and the flow : what we pre - compute and store is a matrix @xmath7 , which we obtain by overwriting the first and last row of @xmath8 with : @xmath9 to allow for imposing the boundary conditions , then sandwiching the resulting matrix between @xmath10 and @xmath11 : @xmath12 then in order to solve equations [ eq : heat diff ] and [ eq : heat diff bc ] for each horizontal mode we construct a new matrix @xmath13 from @xmath7 as follows : @xmath14 the first new term adds the horizontal part of the @xmath15 operator for the given mode .
however , this breaks the requirements for the boundary conditions , so we repair them with the other two terms . this way the boundary conditions are directly imposed by breaking equation [ eq : heat diff ] at the physical top and bottom of the box , instead of ignoring it for the two highest chebyshev modes .
that means we do not need to perform extra greens steps like we did for the pressure equation .
then we obtain the solution to equation [ eq : heat diff ] by decomposing @xmath13 into an upper and lower triangular matrices and using back - substitution . | we have adapted the anelastic spectral code of @xcite to simulate a turbulent convective layer with the intention of studying the effectiveness of turbulent eddies in dissipating external shear ( e.g. tides ) .
we derive the anelastic equations , show the time integration scheme we use to evolve these equations and present the tests we ran to confirm that our code does what we expect .
further we apply a perturbative approach to find an approximate scaling of the effective eddy viscosity with frequency , and find that it is in general agreement with an estimate obtained by applying the same procedure to a realistic simulation of the upper layers of the solar convective zone .
hydrodynamics , anelastic approximation , stratified flows , shear flows , spectral methods , convection , turbulence , turbulent dissipation , effective viscosity |
many resonance lines of the solar spectrum show complex linear polarization patterns , especially when observed near the solar limb @xcite .
these signals are generated by scattering in the upper photosphere and chromosphere , where the plasma is very rarefied and nearly collisionless , and long integration paths are involved in the production of the observed polarized signals . under these conditions ,
the ground state of the observed transitions may harbour a significant amount of atomic polarization @xcite , and coherence effects among the atomic levels become apparent in the emergent polarization profiles ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) a general expression for the _ second - order emissivity _ describing coherent . ] resonance scattering in a two - term polarized atom was recently presented by @xcite ( @xcite ; hereafter , paper i ) . on the other hand , very few lines in the solar spectrum
can be considered strictly resonant .
the ions in the solar atmosphere are illuminated by a broadband radiation field that pumps the atomic levels through all possible transitions simultaneously , creating many different scattering channels .
these conditions are responsible for the appearance of complex coherence phenomena , where the scattered polarization signatures become coupled with the atomic polarization of the various energy levels involved in the atomic transitions .
the expression that was presented in paper i for the two - term atom is general enough that it lends itself to a straightforward extension to the treatment of more complex atomic structures .
here we generalize such expression to the description of polarized radiation effects in a @xmath0-type multi - term atom , where the lower terms of the system are all radiatively coupled to a common excited state ( see figure [ fig : model ] ) .
-type , multi - term model atom considered in this paper , which is relevant for the modeling of the polarization of the @xmath3 transition . in order to correctly describe the polarization properties of the outgoing light ,
all lower terms @xmath4 that are radiatively connected to the ( virtual ) upper term @xmath5 , also including the final term @xmath6 , must be taken into account .
our model is restricted to the case where all @xmath4 levels are sharp ( i.e. , with practically infinite lifetime ) .
[ fig : model ] ] an important example of such @xmath0-type system is contained in the ion , and underlies the formation of some notable transitions observed in the solar spectrum : the h and k lines around 395 nm , which display a complex polarization pattern spanning more than 10 nm @xcite , due to coherent scattering and quantum interference between the levels of the upper @xmath7 term of the doublet @xcite ; and the infrared ( ir ) triplet around 858 nm , in particular the 854.2 nm and 866.2 nm lines , which under solar conditions is dominated by the presence of atomic polarization in the metastable @xmath8 term @xcite .
the uv doublet and the ir triplet are connected via the common @xmath7 upper term , yet previous attemps at modeling the formation of these sets of lines have not taken into account the full complexity of this system : either coherent scattering in a multi - level system was considered , but the essential polarization of the metastable @xmath8 levels was neglected @xcite , or the atomic polarization of all the levels was included in the modeling of the ir triplet pattern , but coherence effects among the levels of the upper @xmath7 term were neglected @xcite . in the next section , we provide explicit expressions for the @xmath0-type multi - term atom with and without hyperfine structure , as well as for the multi - level atom , when the fine - structure interaction is completely neglected
. finally , in section [ sec : application ] we present the application of this formalism to the modeling of the polarized line profiles of some notable @xmath0-type atomic systems .
we consider the general form of the radiative transfer equation for polarized light in spectral lines , including the process of coherent scattering in a spectral line , which is responsible for the effects of partial redistribution of the radiation frequency when the illumination of the atom is not spectrally flat .
such transfer equation was presented in @xcite ( hereafter , paper i ) , and when the polarized radiation is expressed in terms of the stokes vector @xmath9 , it takes the form @xmath10 here @xmath11 is the absorption matrix ( corrected for radiation stimulated effects ) , and @xmath12 and @xmath13 are the source terms due to the spontaneous emission of radiation from the excited levels , and to coherent scattering in the spectral line , respectively . in the particular case of sharp lower levels , and in the absence of collisions , there is no true excitation of the upper levels , because of the infinite radiative lifetime of the lower levels .
then , the emissivity due to the spontaneous emission term @xmath12 in the transfer equation can be neglected , as well as stimulation effects of the radiation ( see paper i ) . in this case , the absorption term does not originate from true photon absorption , but corresponds instead to the atom s cross - section for the coherent scattering of radiation . accordingly , in this work we consider the scattering of radiation in resonance lines when both the initial and final states of the transition are infinitely sharp . then the only _ positive _ contribution to the line radiation comes from the coherent scattering term @xmath13 .
this is given by equation ( i.20 ) , @xmath14 here , @xmath15 represents the matrix element between the atomic states @xmath16 and @xmath17 of the @xmath18 spherical component of the electric dipole moment , @xmath19 . the geometric tensors @xmath20 were introduced by @xcite , and their algebraic expressions have been tabulated by several authors ( e.g. , * ? ? ?
* ; * ? ? ?
the radiation tensors @xmath21 are defined in terms of these geometric tensors and the incident stokes vector as follows , @xmath22 finally , the profiles @xmath23 , which describe the effects of frequency redistribution , are given by equation ( i.5 ) .
we indicate the complex conjugate of these profiles with @xmath24 .
equation ( [ eq:2emiss.1 ] ) describes the scattering of polarized light that occurs in the radiative transition from a set of lower levels @xmath4 weighted by the density matrix @xmath25 and the gas density @xmath26 for the ionic species considered to a final set of levels @xmath6 , via the virtual excitation of a set of intermediate upper levels @xmath5 .
it is important to observe that this expression generally describes the scattering polarization of a radiative transition @xmath3 that results from the combination of distinct @xmath0-type atomic systems sharing the same final branch @xmath27 , but with different initial branches @xmath28 ( see figure [ fig : model ] ) .
therefore , equation ( [ eq:2emiss.1 ] ) can be used to model the scattering of polarized radiation in multi - term atoms of the @xmath0-type , including the effects of a magnetic field and of lower - level polarization .
we consider the general case where the hyperfine structure may be present , since such a model is necessary to describe many interesting spectral lines of the solar chromosphere ( see table [ tab : lines ] ) .
we must observe that the index substitution @xmath29 in the original expression ( i.20 ) for the second - order emissivity of a two - term atom , which allowed us to write equation ( 2 ) , is not just a formal exercise .
there are supporting physical arguments showing that this new expression actually applies to the case of a @xmath0-type multi - term atom .
for example , it reproduces the spectral emission of such a system when the incident radiation field is spectrally flat , i.e. , in the limit of complete redistribution of the incident energy ( see paper i , sect . 6 ) .
this correct behavior was numerically verified in all atomic models that we tested .
cccc ion & @xmath30 ( nm ) & lower terms & upper term + & @xmath31 & @xmath32 & @xmath33 + & @xmath34 & @xmath35 & + & @xmath36 & @xmath37 & @xmath38 + & @xmath39 & @xmath40 & + & @xmath41 & @xmath42 & @xmath43 + & @xmath44 & @xmath45 & + & @xmath46 & @xmath47 & @xmath48 + & @xmath49 & @xmath50 & + & @xmath51 & @xmath52 & @xmath53 + & @xmath54 & @xmath55 & + & 2017.9 2024.6 & @xmath56 & + we indicate with @xmath57 and @xmath58 the electronic configuration of the two lower terms representing respectively the initial and final states of the transition , and correspondingly with @xmath59 the electronic configuration of the intermediate upper term ( see figure [ fig : model ] ) .
we assume the direction of the magnetic field as the quantization axis ( @xmath60-axis ) .
then the atomic states involved in equation ( [ eq : rt.1 ] ) are of the form @xmath61 where @xmath62 is the projection of the total angular momentum @xmath63 on the @xmath60-axis , @xmath64 is the quantum number of the nuclear spin , while @xmath65 is the index of the atomic hamiltonian eigenbasis spanning the subspace of all the quantum numbers @xmath66 and @xmath67 that are associated with a given value of @xmath62 . following the formalism of paper i , equation ( [ eq:2emiss.1 ] ) becomes @xmath68 where we adopted the shorthand notation @xmath69 .
the various coefficients @xmath70 represent the projection components of the eigenstates @xmath71 of the magnetic hamiltonian on the basis of atomic states of the form @xmath72 . in writing equation ( [ eq : rt.jf ] )
we also introduced the redistribution function in the atomic frame of reference , @xmath73 the transformation of equation ( [ eq : rt.jf ] ) to the laboratory frame of reference is formally attained by replacing the redistribution function of equation ( [ eq : redist ] ) with the appropriate velocity - dependent function @xmath74 , where @xmath75 and @xmath76 are the frequencies of the incoming and outgoing radiation in the laboratory frame , respectively , and @xmath77 is the scattering angle .
the task of extending the redistribution function in the laboratory frame to the case of a @xmath0-type three - term polarized atom has been undertaken in a separate work ( r. casini & r. manso sainz 2016 ; in preparation ) . in the appendix , we give the form of such redistribution function in the limit case of _ non - coherent lower term _ , which applies practically to all the examples presented in section [ sec : application ] .
it is important to note that the widths of the level @xmath78 and @xmath79 appearing in the redistribution profiles of the form @xmath80 must take into account all possible spontaneous de - excitation processes towards lower terms @xmath81 , i.e. , @xmath82 where , according to our model , the set of lower terms @xmath81 also includes the final term @xmath6 . for some applications , including those presented in the next section , it is necessary to consider atomic models without hyperfine structure , i.e. , @xmath83 .
in that case , after some straightforward racah algebra manipulations of equation ( [ eq : rt.jf ] ) , we find @xmath84 finally , in the case of the multi - level atom , there is no dependence of the line profiles on the @xmath65-indices , and so we can use the orthogonality properties of the hamiltonian eigenvectors ( see equations ( i.23 ) ) in order to perform the trivial summations over those indices .
the expression of the second - order emissivity for this model atom can be derived directly from equation ( [ eq : rt.j ] ) by imposing the additional conditions @xmath85 and @xmath86 : @xmath87 in the next section , we provide examples of the application of this formalism ( specifically , of equation ( [ eq : rt.j ] ) ) to a few notable @xmath0-type three - term atoms , namely the ly@xmath1-h@xmath2 system of hydrogen , and the h - k doublet with the ir triplet ( see table [ tab : lines ] ) .
stokes @xmath64 and @xmath88 emission profiles of the ly@xmath1 ( top panels ) and h@xmath2 ( bottom panels ) lines , respectively at 102.5 nm and 656.3 nm .
the model is for the 90-degree scattering of a collimated beam of radiation by a plasma with temperature @xmath89k , and no magnetic field .
the intensity of the incident radiation field corresponds to a planckian spectrum at @xmath90k .
the incident uv radiation ( monochromatic and unpolarized ) is nearly resonant with the ly@xmath1 transition , with a detuning of @xmath91 ( vertical dashed lines in the top panels ) .
we note the presence of a secondary contribution in the scattered radiation , which is perfectly resonant with the natural wavelength of the line .
the intensity profile amplitude is expressed as number of scattered photons per unit intervals of time , emitting volume , wavelength , and solid angle , assuming a reference gas density @xmath92 for the emitting volume . ]
as an application of equation ( [ eq : rt.j ] ) , we first consider the simplest case of the @xmath0-type system @xmath93-@xmath94-@xmath95 of , which pertains to the formation of the ly@xmath1 and h@xmath2 lines , respectively at 102.5 nm and 656.3 nm .
since both lines formed in this restricted @xmath0-type system have sharp lower levels , the corresponding three - term model atom can indeed be described through the formalism presented above .
more specifically , we model the effect of detuning of the uv radiation around the wavelength of the ly@xmath1 line on the intensity and polarization of the h@xmath2 line . for this purpose , we assume an ensemble of atoms with a pre - assigned distribution of population and atomic polarization in the ground and metastable states . this distribution is derived by solving the statistical equilibrium for the atomic system under prescribed illumination conditions , corresponding to a collimated beam of highly diluted radiation with a planckian spectrum at @xmath96 . under such conditions , the population of the excited @xmath97 term is negligible with respect to the populations of the ground and metastable states ( by about 9 and 8 orders of magnitude , respectively ) , and the prd formalism presented in this work , where the incoherent emission of photons from spontaneous de - excitation of the upper term can be neglected , thus becomes applicable .
we also remark that the atomic system so prepared will in general harbor atomic polarization , because of the condition of anisotropic illumination associated with the collimated beam of incident radiation .
contour plots of the scattered intensity of the ir line at 854.2 nm ( left ) and the k line at 393.4 nm ( right ) of , as a function of wavelength across the line spectral ranges ( @xmath98 axes ) , and of the detuning of the incident monochromatic ir radiation around the 854.2 nm resonance wavelength ( @xmath99 axis ) .
the scattering configuration is the same as for figure [ fig : lyb - ha ] , but for an incident radiation with a planckian spectrum at @xmath100k . the diagonal dash - dotted line in the left panel tracks the location of the monochromatic illumination with respect to the intensity profile of the scattered 854.2 nm line , for the various values of detuning .
the horizontal dashed lines locate the value of the detuning of the monochromatic ir incident radiation adopted for the calculation of the emission profiles shown in figure [ fig : plots.offres.8542 ] . ]
figure [ fig : lyb - ha ] shows the scattered radiation produced in this system .
the wavelength of the uv incident radiation , which again is monochromatic and unpolarized , is marked by the vertical dashed lines in the stokes @xmath64 and @xmath88 plots of the ly@xmath1 line . in order to clearly identify the various contributions to the scattered radiation in the system , we assume no incident radiation at the h@xmath2 wavelength .
hence , the radiation scattered in the spectral range of h@xmath2 is completely due to raman scattering through virtual excitation of the @xmath97 term by the uv radiation .
we note that the scattered radiation in the ly@xmath1 line is dominated by the coherent component centered around the wavelength of the monochromatic uv incident radiation .
however , because of the relatively large doppler width compared to the size of the detuning , a small contribution , resonant with the natural transition of the line and blended with the coherent component , is also present in this case .
in contrast , the emission in h@xmath2 shows these same two contributions well separated .
in fact , if we indicate with @xmath101 the detuning of the incident radiation in the @xmath0-type system , for the final branch of the scattering process @xmath102 because in figure [ fig : lyb - ha ] the detuning from the ly@xmath1 resonance is @xmath103 , the coherent component in h@xmath2 occurs with a wavelength shift from resonance given by @xmath104 . as this is larger than the doppler width corresponding to the assumed plasma temperature ( which instead scales linearly with the wavelength of the transition ) ,
the coherent and resonant contributions in the h@xmath2 line appear completely separated .
we also note that the line profiles of both transitions are 100% linearly polarized at the frequency of coherent re - emission , as it is to be expected in the wings of @xmath105-@xmath106 transitions . finally , we point out that the wavelength integrated intensity profiles of figure [ fig : lyb - ha ] , which give the total numbers of photons emitted in the ly@xmath1 and h@xmath2 lines , are in exactly the same ratio as the einstein a - coefficients of the two transitions , which is also to be expected .
next , we consider the more complex atomic system of , leading to the formation of the h and k lines around 395 nm and the ir triplet around 858 nm .
we model the effect of detuning in the two cases where the monochromatic incident radiation is located in the proximity of either one or the other of the two transitions .
we consider the same scattering configuration as in the previous example , with the exception of the temperature of the incident radiation , for which we assume instead a planckian spectrum at @xmath107 .
stokes @xmath64 and @xmath88 emission profiles of the ir 854.2 nm line ( top panels ) and the k line at 393.4 nm ( bottom panels ) of , for the same scattering model adopted for figure [ fig : plots.res.8542 ] .
these profiles correspond to the detuning of the incident radiation from the 854.2 nm wavelength resonance identified by the horizontal dashed lines in the contour plots of figure [ fig : plots.res.8542 ] .
hence , the stokes @xmath64 profiles represent the horizontal cut of those contour plots for the corresponding value of the detuning of the incident ir radiation .
the profiles shown with dashed curves correspond to the case when the atomic polarization of the metastable state of the system is neglected . ]
figure [ fig : plots.res.8542 ] shows contour plots of the scattered stokes @xmath64 emission profiles ( wavelength along the @xmath98 axis ) of the ir line at 854.2 nm ( left ) and the k line at 393.4 nm ( right ) of , as a function of the frequency of a monochromatic illumination ( wavelength along the @xmath99 axis ) varying within a spectral range of 0.18 around the 854.2 nm resonance wavelength .
each horizontal slice of those plots thus corresponds to the scattered line intensity profile for the corresponding value of the detuning from the 854.2 nm resonance .
again , we assume no direct illumination of the k line for this modeling , and so the emitted radiation in that line is purely produced by raman scattering induced via virtual excitation of the upper state of the k line by the ir illumination .
in particular , this allows us to represent the scattered radiation in both lines via the contour plots of figure [ fig : plots.res.8542 ] , as a function of only one detuning parameter .
the observed spectral spread of the re - emitted radiation along the @xmath98 axis is produced by doppler redistribution , corresponding to the plasma temperature of 1000k .
also for this model we assumed a zero magnetic field . in such case
, the wavelength dependence of the stokes @xmath88 polarization qualitatively resembles closely that of the intensity profiles ( see also figure [ fig : plots.offres.8542 ] , and therefore we omitted showing contour plots also of stokes @xmath88 . same as figure [ fig : plots.res.8542 ] , but for an inverted role of the two transitions , where the monochromatic incident radiation this time is tuned across the spectral range of the k line centered at 393.4 nm , with no radiation directly exciting the ir line at 854.2 nm .
the horizontal dashed lines locate the value of the detuning of the monochromatic uv incident radiation adopted for the calculation of the emission profiles shown in figure [ fig : plots.offres.k ] . ]
the dash - dotted diagonal line in the left plot tracks the wavelength of the monochromatic ir incident radiation as it scans across the spectral interval around the 854.2 nm resonance wavelength .
as the detuning from the 854.2 nm transition ( which is read on the @xmath99 axis ) decreases in absolute value , the signal of the scattered radiation remains initially fully coherent with the frequency of the incident radiation , while at the same time it increases in strength .
in fact , for values of the detuning larger than @xmath108 in absolute value , the virtual energy levels of the upper term attained during the scattering process lie outside the energy band corresponding to the thermal line width of the upper term .
then there is no energy overlap leading to quantum interference between these levels and the atomic hamiltonian eigen - levels of the upper term , and the scattering is purely coherent in both the k and 854.2 nm lines .
this process corresponds to rayleigh scattering of the monochromatic incident radiation , and the frequency spread observed in the scattered radiation is dominated by the doppler redistribution corresponding to the plasma temperature .
however , at about 0.06 from resonance , the signal of the non - coherent contribution at the resonance wavelength of the 854.2 nm line begins to appear , becoming the dominant term of the scattered radiation in an interval of about @xmath109 around the resonance .
the radiation emitted in the k line , which is produced by raman scattering towards the ground state , is also dominated by the resonant contribution at 393.4 nm approximately within the same interval of detuning of the ir illumination , although the profiles of this line never become double - peaked .
this can be understood if we recall equation ( [ eq : important ] ) , and it is clearly illustrated by figure [ fig : plots.offres.8542 ] , which shows the stokes @xmath64 and @xmath88 emission profiles ( solid curves ) for a detuning of @xmath110 from the 854.2 nm resonance wavelength ( horizontal slice of the contour plots of figure [ fig : plots.res.8542 ] identified by the dashed lines ) .
overplotted on these profiles , with the dashed curves , we show the same case where the contribution of the atomic polarization in the metastable state of is neglected . as we see from comparing the two sets of profiles , the linear polarization of the scattered radiation in both k and 854.2 nm lines is larger in that case .
stokes @xmath64 and @xmath88 emission profiles of the k line at 393.4 nm ( top panels ) and the ir 854.2 nm line ( bottom panels ) of , for the same scattering model adopted for figure [ fig : plots.res.8542 ] , but in the case where the monochromatic incident radiation is tuned around the k line resonance wavelength .
we note how in this case the profiles of both lines can become visibly double - peaked , as a consequence of equation ( [ eq : important ] ) .
these profiles correspond to the cut of the contour plots of figure [ fig : plots.res.k ] for the value of the detuning of the monochromatic uv incident radiation identified by the horizontal dashed lines . ]
stokes emission profiles of the k line , with the same detuning of the uv radiation around the k line resonance as for figure [ fig : plots.offres.k ] , but in the case where a magnetic field with @xmath111 g directed towards the observer is also present .
we note the manifestation of the hanle effect polarization in both stokes @xmath88 and @xmath112 , in the resonant core of the line .
the dashed curve in the panel for stokes @xmath113 represents the weak - field approximation to the circular polarization signal , computed through the first derivative of stokes @xmath64 , assuming the @xmath114-coupling value for the effective land factor of the line , @xmath115 . ] figures [ fig : plots.res.k ] and [ fig : plots.offres.k ] are analogous to figures [ fig : plots.res.8542 ] and [ fig : plots.offres.8542 ] , for the case where the monochromatic incident radiation is tuned across the resonance wavelength of the k line at 393.4 nm instead , with no radiation directly exciting the ir transition at 854.2 nm .
as evidenced by figure [ fig : plots.offres.k ] , in this case the profiles of both lines can become double - peaked ( for values of the detuning from resonance larger than @xmath116 in absolute value ) , in virtue of the separation between the coherent and resonant contributions satisfying equation ( [ eq : important ] ) . when a magnetic field is present
, the profiles evidently become more complicated , although some general conclusions can be drawn from the example presented in figure [ fig : plots.b ] .
this shows the full stokes profiles of the ir line at 854.2 nm under the same illumination conditions of figure [ fig : plots.offres.k ] , but with the addition of a magnetic field of 5 g directed towards the observer ( i.e. , normal to the direction of the incident radiation ) . because the magnetic field strength is comparable to the hanle critical field for the upper level @xmath117 of the 854.2 nm and k lines ( @xmath118
g , where @xmath119 is the land factor of the @xmath117 level , and @xmath120 is calculated via equation ( [ eq : widths ] ) ) , one expects that the stokes profiles will show evidence of hanle - effect depolarization . indeed , comparing the bottom panels of figure [ fig : plots.offres.k ] with the top panels of figure [ fig : plots.b ] , we see that the intensity profile is practically unaffected by the presence of a weak magnetic field , whereas the resonant component of stokes @xmath88 shows a depolarization with respect to the zero - field case , and accordingly a signal in stokes @xmath112 appears .
it is important to observe that the stokes @xmath112 signal manifests itself strictly in the resonant core of the line , which is also where the depolarization of stokes @xmath88 occurs .
this is to be expected , since the hanle effect is a manifestation of the relaxation of atomic coherence as the energy degeneracy of the atomic hamiltonian eigen - levels is lifted by the applied magnetic field .
stokes profiles of the line at 854.2 nm , for the same scattering configuration and magnetic model of figure [ fig : plots.b ] , and different conditions of polarization of the monochromatic uv incident radiation in the proximity of the k line : unpolarized ( solid curves ; identical to the case of figure [ fig : plots.b ] ) ; fully linearly polarized along the los ( dashed curves ) ; fully linearly polarized perpendicularly to the los ( dash - dotted curves ) . ]
we also note that the shape of stokes @xmath113 is not quite reproduced by the weak - field approximation of the circular polarization signal ( dashed curve in the bottom - right panel of figure [ fig : plots.b ] ) , despite the very small strength of the applied field , and the fact that all polarization effects due to level interference induced by the magnetic field , such as stokes-@xmath113 asymmetries associated with atomic orientation , are completely negligible in this case , because of the relatively large fine structure separation between the @xmath121 levels of the upper term .
therefore , the net circular polarization observed in figure [ fig : plots.b ] is a manifestation of the effects of partial redistribution on the circular polarization of the scattered light .
all previous examples were calculated assuming an unpolarized beam of incident radiation .
figure [ fig : ca_8542.b+q ] shows instead the stokes profiles of the 854.2 nm line for the same scattering configuration and magnetic model as in the example of figure [ fig : plots.b ] , but under different conditions of polarization of the incident uv radiation .
more precisely , the atomic system is still prepared assuming the same flat and unpolarized radiation field as in all previous examples , and only the incident beam of radiation used in equation ( [ eq : rt.j ] ) is now assumed to be linearly polarized .
the solid curves represent the case of unpolarized incident radiation as a reference , which are identical to those shown in figure [ fig : plots.b ] .
the other curves represent instead the cases of fully linearly polarized radiation with @xmath122 ( polarization along the los ; dashed curves ) and @xmath123 ( polarization perpendicular to the los ; dash - dotted curves ) .
when the incident beam of radiation is linearly polarized along the los , the linear polarization of the scattered radiation is practically completely suppressed ( in fact , by approximately six orders of magnitude compared to the unpolarized case ) , as it can be expected also on the basis of simple classical arguments @xcite . as a concluding remark
, we want to point out that , while the atomic polarization produced by the anisotropy of the incident radiation is fully accounted for in the modeling examples presented in this work , all those examples are realized under physical conditions where the atomic coherence in the lower terms is practically negligible .
this is either because the magnetic field is absent ( figures [ fig : lyb - ha ] to [ fig : plots.offres.k ] ) , or because the magnetic strength is above the critical value for the hanle effect of the lower term ( figures [ fig : plots.b ] and [ fig : ca_8542.b+q ] ) , yet small enough not to induce any @xmath66-@xmath124 level interference via the paschen - back effect .
the realization of this range of physical conditions allows us to adopt a reduced form of the redistribution function for the polarized atom introduced in @xcite ( @xcite ) , which corresponds to a direct generalization to the three - term atom of the @xmath125 function adopted elsewhere in the prd literature ( e.g. , @xcite ) .
we thank tanaus del pino alemn ( instituto de astrofsica de canarias ; iac ) for helpful discussions during the initial testing of the code with which the results of this paper were produced .
we also thank rebecca centeno elliot ( hao ) and javier trujillo bueno ( iac ) for reading the manuscript and for providing helpful comments and suggestions .
rms acknowledges partial support by the spanish ministry of economy and competitiveness ( project aya2010 - 18029 ) , and the support of the hao visitor program , which made this work possible .
we dedicate this work to the memory of david g. hummer ( 1934 - 2015 ) , who greatly contributed to the field of frequency redistribution .
david , who for many years of his active career was a vital collaborator of hao and ncar , as witnessed by his many works published together with hao scientists , passed away on december 17 , while this paper was receiving the final strokes .
in the case of infinitely sharp lower levels ( s.l.l . ) , and assuming that the initial term of the @xmath0-type transition is _ non - coherent _ (
i.e. , @xmath126 ) , the redistribution function for the @xmath0-type three - term polarized atom , expressed in the laboratory reference frame , is given by @xmath127 \\ & & \kern -3 cm { } \times \biggl [ w\biggl(\frac{\kappa^+ v_{ul}+\kappa^- w_{uf}}{s\xi_l\xi_f } , \frac{a_u}{s\xi_l\xi_f } \biggr ) + \overline{w}\biggl(\frac{\kappa^+ v_{u'l}+\kappa^- w_{u'f}}{s\xi_l\xi_f } , \frac{a_{u'}}{s\xi_l\xi_f } \biggr ) \biggr]\ ; , \nonumber\end{aligned}\ ] ] where @xmath128 with @xmath129 and @xmath130 being respectively the voigt and faraday - voigt functions . the full derivation of the above result is contained in an upcoming paper ( r. casini & r. manso sainz 2016 ; in preparation ) . for each scattering event , @xmath77 is the angle between the propagation directions of the incoming and outgoing photons .
we introduced accordingly the associated quantities @xmath131 next we defined @xmath132 @xmath133 where @xmath134 is the doppler width of the transition between the atomic states @xmath135 and @xmath136 , with bohr frequency @xmath137 .
we introduced the normalized frequency variables @xmath138 where the incoming and outgoing radiation frequencies , @xmath139 and @xmath140 , are expressed in the laboratory frame of reference .
we also introduced normalized damping parameters associated with the inverse lifetimes of the transition levels , using the same `` reduced '' doppler width of equation ( [ eq : avg_doppler ] ) , @xmath141 finally , we introduced the transitions `` weights '' @xmath142\;.\ ] ] casini , r. , landi deglinnocenti , m. , manso sainz , r. , landi deglinnocenti , e. , & landolfi , m .. 2014 , , 791 , 94 , a. 2000 , the second solar spectrum : a high spectral resolution polarimetric survey of scattering polarization at the solar limb in graphical representation .
volume i : 4625 to 6995(zurich : vdf hochschulverlag ) , a. 2002 , the second solar spectrum : a high spectral resolution polarimetric survey of scattering polarization at the solar limb in graphical representation .
volume ii : 3910 to 4630(zurich : vdf hochschulverlag ) , a. 2005 , the second solar spectrum : a high spectral resolution polarimetric survey of scattering polarization at the solar limb in graphical representation .
volume iii : 3160 to 3915(zurich : vdf hochschulverlag ) landi deglinnocenti , e. 1984 , , 91 , 1 landi deglinnocenti , e. 1998 , , 392 , 256 landi deglinnocenti , e. , & landolfi , m. 2004 , polarization in spectral lines ( dordrecht : springer ) manso sainz , r. , & trujillo bueno , j. 2001 , in advanced solar polarimetry , theory , observation , and instrumentation,20th nso / sacramento peak summer workshop , asp conference proceedings , 236 , ed .
m. sigwarth ( san francisco : asp ) , 213 manso sainz , r. , & trujillo bueno , j. 2003 , , 91 , 111102 mitchell , a. c. g. , & zemansky , m. w. 1934 , resonance radiation and excited atoms ( cambridge : cambridge ) sampoorna , m. , nagendra , k. n. , & stenflo , j. o. 2013 , , 770 , 92 smitha , h. n. , nagendra , k. n. , stenflo , j. o. , bianda , m. , sampoorna , m. , ramelli , r. , & anusha , l. s. 2013 , a&a , 541 , 24 stenflo , j. o. 1980 , a&a , 84 , 68 stenflo , j. o. 1997 , a&a , 324 , 344 stenflo , j. o. , & keller , c. u. 1996 , , 382 , 588 stenflo , j. o. , & keller , c. u. 1997 , a&a , 321 , 927 stenflo , j. o. , baur , t. g. , & elmore , d. f. 1980 , a&a , 84 , 60 , j. o. , twerenbold , d. , harvey , j. w. 1983 , , 52 , 161 , j. o. , twerenbold , d. , harvey , j. w. , brault , j. w. 1983 , , 54 , 505 stenflo , j. o. , keller , c. u. , & gandorfer , a. 2000 , a&a , 355 , 789 trujillo bueno , j. , & landi deglinnocenti , e. 1997 , , 482 , l183 trujillo bueno , j. , landi deglinnocenti , e. , collados , m. , merenda , l. , & manso sainz , r. 2002 , , 415 , 403 wiehr , e. 1975 , a&a , 38 , 303 | we study the effects of rayleigh and raman scattering on the formation of polarized spectral lines in a @xmath0-type multi - term atom .
we fully take into account the partial redistribution of frequency and the presence of atomic polarization in the lower states of the atomic model .
problems that can be modeled with this formalism include , for example , the formation of the h - k and ir triplet , the analogous system of , and the ly@xmath1-h@xmath2 system of hydrogenic ions . |
we performed simulations of fracturing ranked surfaces on square lattices .
it is worth noting that , despite the similarities with random percolation , the suppressing of connectivity poses a statistically different problem .
for example , while for @xmath13 there is only a single configuration in random percolation ( all bonds occupied ) , in ranked percolation there are @xmath14 , evenly weighted , possible configurations , where @xmath15 is the total number of bonds . in classical percolation
the total number of configurations is @xmath16 .
figure [ fig::mbbsize ] shows the dependence of the number of bridge bonds @xmath17 on system size , for different fractions of occupied bonds , namely , @xmath18 , @xmath19 , and @xmath20 .
as expected , at the percolation threshold of classical percolation ( @xmath7 ) , the number of bridge bonds diverges with system size as @xmath21 , where @xmath10 is the correlation length exponent , with @xmath22 in @xmath23 ; while for @xmath20 , @xmath24 , with @xmath25 .
the latter is in fact observed at any @xmath26 .
we found the same result ( see _ appendix _ ) for site percolation and on other lattices ( star , triangular , and honeycomb ) , which provides strong evidence for the universality of this exponent .
for @xmath27 , like @xmath19 we can observe , as depicted in fig .
[ fig::mbbsize ] , a crossover between the two different regimes .
the inset of fig .
[ fig::bbscale ] shows @xmath17 , rescaled by @xmath28 , as a function of @xmath2 , for different system sizes .
the number of bridge bonds grows with @xmath2 , such that , @xmath29 , where @xmath30 is a novel exponent , which we call bridge - growth exponent .
the overlap of the different curves confirms that the fractal dimension of the bridge bonds above @xmath11 is @xmath12 , for all @xmath26 .
this result differs from classical percolation where fractality is solely observed at criticality @xcite while , above @xmath11 , bridge bonds are only observed for finite systems @xcite . , for different @xmath2 , namely , @xmath18 ( circles ) , @xmath19 ( stars ) , and @xmath20 ( triangles ) .
the solid lines stand for the best fit . at @xmath7 , as conjectured by coniglio @xcite , @xmath21 , where @xmath22 is the critical exponent of the correlation length in @xmath23 . for @xmath26 ,
the number of bridge bonds scales with @xmath28 .
a crossover between the two regimes in system size is observed ( stars ) for @xmath2 in the neighborhood of @xmath11 .
systems of size @xmath31 have been considered , with @xmath32 ranging from @xmath33 to @xmath34 .
all results have been averaged over @xmath35 samples .
error bars are smaller than the symbol size .
[ fig::mbbsize ] ] , as a function of the fraction of occupied bonds , @xmath2 , for @xmath23 with different system sizes @xmath36 .
the scaling function given by equation ( [ eq::scaling ] ) is applied , with @xmath37 , obtaining @xmath30 . in the inset
, @xmath17 has been rescaled by @xmath28 , where @xmath38 .
all results have been averaged over @xmath35 samples .
[ fig::bbscale ] ] for polymer chains , at high temperatures , the excluded volume prevails over attractive forces and the chain can be described as a self - avoiding walk .
when the temperature is reduced , the attractive forces become relevant leading , at a theta - temperature , to a new exponent at the crossover between two dimensions @xcite .
analogously , in ranked percolation the fractal dimension of the bridge bonds is @xmath8 , at @xmath7 , between @xmath12 above @xmath11 and zero below @xmath11 . for the tricritical
scaling we verify the following _ ansatz _ , @xmath39,\ ] ] where @xmath40\sim x^{\zeta}$ ] for @xmath41 , and is nonzero at @xmath42 ; and the exponent @xmath43 is the crossover exponent .
therefore , the following relation is obtained , @xmath44 where @xmath45 . in the main plot of fig .
[ fig::bbscale ] we see , for @xmath23 , the scaling given by equation ( [ eq::scaling ] ) , with @xmath37 . the results above disclose a tricritical @xmath11 below which the fraction of bridges in the bridge line vanishes in the thermodynamic limit .
this is identical to a minimum height in the bridge line , @xmath46 , in the context of landscapes ( figs .
[ fig::rank](a)-(b ) ) . for a cumulative distribution function of heights @xmath47 , @xmath48 ,
the minimum is given by @xmath49 .
note that , for uncorrelated landscapes , regardless the distribution of heights , the set of bridges only depends on their position in the rank .
to observe the new set of exponents on discretized landscapes ( fig .
[ fig::rank](c ) ) , @xmath17 is the number of sites in the bridge line with both neighbors ( one at each side ) having height lower than @xmath50 , where @xmath51 .
to study the dependence of exponents @xmath52 and @xmath53 on the spatial dimension , we analyze the same problem up to dimension six . on a simple - cubic lattice ( @xmath54 ) , above @xmath11 , the set of all bridge bonds has a fractal dimension @xmath55 and grows with @xmath56 ( see _ appendix _ ) .
figure [ fig::nbbdim ] shows the size dependence of @xmath17 , in the limit @xmath57 , for lattices with size @xmath58 , where @xmath59 is the spatial dimension . in the inset , we plot the exponent @xmath53 as a function of @xmath60 . since the set of bridge bonds blocks connectivity from one side to the other , its fractal dimension must follow @xmath61 , i.e. , @xmath62 . with increasing dimension
@xmath53 decreases . at the upper - critical dimension of percolation , @xmath63 , @xmath64 and
the set of bridge bonds becomes dense having the spatial dimension @xmath60 .
table [ tab::exponents ] summarizes the exponents for dimensions @xmath65 to @xmath66 .
the bridge - growth exponent grows with @xmath60 and converges to @xmath67 at the upper - critical dimension . for @xmath68 , above the critical dimension of percolation , the exponent @xmath53 remains zero and the dimension of the set of bridges is equal to @xmath60 .
this can be understood from the fact that above @xmath69 one has an infinity of spanning clusters @xcite and thus many more possible bridges .
since the dimension of the set of bridges at @xmath7 is @xmath8 and above @xmath11 is @xmath60 , the crossover exponent increases with @xmath60 , with the relation given by equation ( [ eq::scaling.relation ] ) , where @xmath70 . .
table of exponents .
values of the exponents @xmath52 and @xmath53 up to dimension six . with increasing dimension
, the @xmath52 exponent converges to @xmath67 and @xmath53 goes to zero , revealing that the set of bridge bonds is dense .
[ tab::exponents ] [ cols="^,^,^",options="header " , ] , in the limit @xmath57 , up to dimension six .
the solid lines stand for the best fit .
the mass of the shortest path in loopless percolation ( sp ) in @xmath23 is also included for comparison .
results have been averaged over @xmath35 samples for @xmath23 and @xmath54 , and @xmath71 samples for higher dimensions .
error bars are smaller than the symbol size .
the inset shows the dependence on the spatial dimension of the exponent @xmath53 . at the upper - critical dimension of percolation , @xmath72 ,
the set of bridge bonds becomes dense having the spatial dimension .
[ fig::nbbdim ] ] since one can interchange occupied and empty bonds , there exists a symmetry between bridges and cutting bonds ( red bonds ) , so one can raise the question of what happens when bonds are removed from a percolating system with the constraint that connectivity can not be broken . initially all bonds are occupied and at the end , since cutting bonds are never removed , a line of cutting bonds is obtained , which we denote here as the _ cutting line_. above @xmath11 , as in the classical case , the percolation cluster is compact and there are no cutting bonds . for @xmath73 ,
the set of cutting bonds is fractal with the same fractal dimension as the bridge - bond set , @xmath74 , whereas at @xmath11 it is @xmath8 .
for the crossover scaling a similar _ ansatz _ to the one given by equation ( [ eq::scaling ] ) is verified , where the argument of the scaling function is then @xmath75 .
the same hyperscaling of equation ( [ eq::scaling.relation ] ) is obtained with @xmath76 . at @xmath77 ,
our numerical results corroborate the hypothesis of the same @xmath52 and @xmath53 , for cutting and bridge bonds ( see _ appendix _ ) . for @xmath78 ,
the set of cutting bonds is a line with @xmath79 and the one of bridge bonds has a dimension above @xmath80 , so that the fractal dimensions differ . above the critical dimension ,
the set of cutting bonds has dimension two , like the shortest path at @xmath11 @xcite , for any @xmath81 .
fisher @xcite proposed a bond - site transformation to map bond percolation on a lattice onto site percolation on a covering graph .
for example , the covering graph of the square lattice is obtained by adding all diagonal edges to every other face @xcite .
considering such a mapping for ranked percolation and , following the analogy with a random landscape , where sites are sequentially removed from the lowest number ( height ) and suppressing global disconnection , the obtained cutting line is identical to the bridge one with the constraint applied in the perpendicular direction .
the same is also observed for site - ranked percolation on a square lattice . given the relation between connectivity and disconnection , in both cases , the cutting version needs to be defined on different topologies , namely , the star lattice for the square ( on sites ) and the same lattice for its covering graph , but with swapped cells of diagonal bonds . for the triangular lattice ,
the relation between cutting and bridges sites is straightforward without requiring additional connections . , for different fractions of occupied bonds @xmath2 , namely @xmath18 ( circles ) , @xmath82 ( stars ) , and @xmath83 ( triangles ) .
the solid lines are guides to the eye . at @xmath7 , @xmath84 , where @xmath85 is the critical exponent related with the correlation length . for @xmath73 , the number of cutting bonds scales with @xmath86 .
a crossover between the two regimes with the system size is observed ( stars ) for @xmath2 in the neighborhood of @xmath11 .
systems of size @xmath31 have been considered , with @xmath32 ranging from @xmath87 to @xmath88 .
results have been averaged over @xmath89 samples for most system sizes and over @xmath90 samples for the largest one .
error bars are smaller than the symbol size .
the inset shows the number of cutting bonds , @xmath91 , as a function of occupied bonds , @xmath2 , for two - dimensional lattices with @xmath32 ranging from @xmath92 to @xmath93 ( averaged over @xmath71 samples ) .
the scaling is given by eq .
( @xmath94 ) in the article , where the argument is then @xmath75 , with @xmath37 .
we obtain @xmath95 , which is compatible with the one for bridge bonds .
[ fig::cb ] ]
for several different models in 2d the same fractal dimension as the one of bridge bonds has been reported . here
we discuss how some of them can be related with the process of fracturing ranked surfaces .
let us construct ranked percolation configurations on a random landscape starting with the bond with the largest number ( height ) and then occupying bonds sequentially in order of decreasing number . each time a chosen bond closes a loop , it is not occupied ( loopless ) .
one stops the procedure when , for the first time , a path of connected sites spans from one side to the other ( percolation threshold ) . since all clusters are trees , the backbone and the shortest path are the same and their fractal dimension is identical to the bridge - bond line , as verified in fig .
[ fig::nbbdim ] .
in fact , given the connectivity / disconnection relation discussed in the previous section , this line corresponds to the bridge line when bonds are occupied from the lowest to the largest number in the covering graph .
this line also corresponds to the line of cutting bonds when bonds are removed from the smallest to the largest number .
a similar procedure was also proposed to obtain the minimum spanning tree ( mst ) , for which the same fractal dimension is found for the paths between all pairs of sites @xcite .
there bonds are sequentially removed ( from the highest to the lowest value ) under the constraint that all sites remain connected .
if the constraint is relaxed such that only connectivity between two opposite borders is imposed , the cutting - bond line is obtained .
therefore , the cutting - bond line is the path between borders on the mst for which the largest value ( height ) is minimum , a valid relation in any spatial dimension @xmath60 .
, on several different lattices .
the solid lines are guides to the eye , all with slope @xmath96 .
the best fit to each data set gives the same fractal dimension of @xmath97 .
systems of linear size @xmath32 have been considered , with @xmath32 ranging from @xmath33 to @xmath34 .
results have been averaged over @xmath98 samples .
[ fig::difnet ] ] , as a function of the fraction of occupied bonds , @xmath2 , for @xmath54 ( simple - cubic lattices ) with different sizes @xmath99 .
the scaling function proposed in this article is applied , with @xmath100 , obtaining @xmath56 . in the inset
, @xmath17 has been rescaled by @xmath28 , where @xmath101 .
all results have been averaged over @xmath35 samples .
[ fig::bbscale3d ] ] the fractal dimension @xmath12 was also observed for the backbone of the optimal path crack ( opc ) @xcite . for a random landscape ,
the sequence of optimal paths between opposite borders is obtained and their highest site removed .
every such path crosses the bridge line .
the highest site on successive paths is either on the bridge line itself or is higher than the lowest bridge - line site .
therefore , as in the loopless percolation described before , the removed sites percolate when all sites on the bridge line are removed , which is the backbone of the crack , giving the same fractal dimension @xmath12 .
this is in fact the case for the crack of any sequence of self - avoiding paths between opposite borders .
since in this case the duality between cutting and bridges is not used ( only valid in 2d ) , the relation between opc and bridge bonds is still true in higher dimensions .
let us now define the optimal minimax path ( omp ) in the following way .
one starts selecting the set of paths on the landscape for which the highest - ranked site is minimum ( minimax paths ) .
such set is then reduced to include only the paths for which the second highest site is also minimum and one proceeds iteratively to the following sites , until a unique path is obtained .
this path is the optimal path in strong disorder @xcite , since under such disorder strength each site is higher than the sum over the height of all sites with lower rank .
this path is also identical to the backbone of the loopless - ranked - percolation cluster when occupied from the lowest to the largest number and , therefore , the cutting line in any dimension . in summary
, suppressing connectivity ( disconnection ) between opposite borders on ranked surfaces leads to a fractal set of bridge ( cutting ) bonds , with a universal fractal dimension , even far away from the critical point of classical percolation . in @xmath23
, there is an equivalence between bridges and cutting bonds and @xmath102 , whereas for @xmath78 cutting bonds are still a fractal line but bridge bonds form a surface .
the discussed models are then split into two groups .
the ones suppressing connectivity ( e.g. , watershed @xcite and optimal path crack @xcite ) are in the bridges universality class , while the ones keeping connectivity ( e.g. , optimal path on the mst or in strong disorder media @xcite ) are in the universality class of cutting bonds . for @xmath68 , @xmath103 and @xmath104 , so we conjecture that the upper - critical dimension of bridges and cutting bonds is also @xmath63 . finally , we show that , at the percolation threshold of classical percolation , ranked percolation displays a theta - point - like crossover
. , rescaled by @xmath28 , as a function of the fraction of occupied bonds @xmath2 .
the inset shows @xmath105 defined by eq .
( [ eq::def.chi ] ) .
results have been obtained on a square lattice with @xmath32 ranging from @xmath93 to @xmath34 .
all results have been averaged over @xmath35 samples .
[ fig::bbp ] ] this work opens up several challenges . besides the need for a more precise numerical estimation of the bridge - growth exponent
, it would be interesting to formulate a renormalization group scheme and to obtain the new set of exponents from analytic treatments such as , e.g. , exact results in the mean - field limit and a schramm - loewner evolution in two dimensions .
another interesting possibility is to find the corresponding exponents in other related problems with different universality classes like , e.g. , the kasteleyn - fortuin clusters in the @xmath106-state potts model @xcite with or without magnetic field . regarding the fractal dimension of the surface of discontinuous percolation clusters @xcite
, it would also be interesting to understand how it relates with the herein introduced bridge - bonds universality class . for the cutting bonds
, the study of the crossover for higher dimensions represents another computational challenge .
finally , it would also be interesting to try to identify the third scaling field of our theta - like point .
all numerical results have been obtained with monte carlo simulations .
results have been averaged over @xmath35 samples for @xmath23 and @xmath54 , and @xmath71 samples for higher dimensions .
k.j.s . and n.a.m.a .
carried out the numerical experiments . all authors conceived and designed the research , analyzed the data , worked out the theory and wrote the manuscript .
we acknowledge financial support from the eth risk center .
we also acknowledge the brazilian agencies cnpq , capes and funcap , and the pronex grant cnpq / funcap , for financial support .
in fig . [ fig::cb ]
we see the size dependence of the number of cutting bonds in @xmath23 when bonds are removed with the constraint that cutting bonds are never removed . as discussed in the article for bridge bonds ,
a crossover is observed at @xmath7 , between two fractal dimensions . in @xmath23 , for @xmath73 the fractal dimension of the cutting bonds @xmath107 is the same as observed for the bridge bonds for @xmath26 , whereas at @xmath11 it is @xmath8 .
in fig . [ fig::difnet ] one sees results for the bridge - bond line in several different lattices , namely , square , star , triangular , and the honeycomb lattices .
the data points for bridge sites on the square lattice are also included .
for all considered lattices the obtained fractal dimension of the bridge line is @xmath97 , a strong evidence for the universality of this fractal dimension .
figure [ fig::bbscale3d ] shows the crossover scaling for the number of bridge bonds @xmath17 on a simple - cubic lattice with different sizes .
bonds are occupied except for the bridge bonds .
we applied the scaling function proposed in eq .
( 1 ) of the article with @xmath100 , giving @xmath56 .
as discussed in the article , the set of bridge bonds merges towards a single connected line . in fig .
[ fig::bbp ] we see the size of the largest cluster of sites on the edge of bridge bonds @xmath108 , rescaled by @xmath28 , as a function of @xmath2 , for different system sizes .
the inset shows @xmath105 , defined as , @xmath109,\ ] ] where the sum runs over all clusters of sites on the edge of bridge bonds and @xmath110 is their size .
alike percolation on a line , in the thermodynamic limit , the bridge - bond line only percolates at @xmath13 .
10 stark , c. p. an invasion percolation model of drainage network evolution .
_ nature _ * 352 , * 423425 ( 1991 ) .
maritan , a. , colaiori , f. , flammini , a. , cieplak , m. & banavar , j. r. universality classes of optimal channel networks .
_ science _ * 272 , * 984986 ( 1996 ) .
manna , s. s. & subramanian , b. quasirandom spanning tree model for the early river network .
lett . _ * 76 , * 34603463 ( 1996 ) .
knecht , c. l. , trump , w. , ben - avraham , d. & ziff , r. m. retention capacity of random surfaces .
lett . _ * 108 , * 045703 ( 2012 ) .
baek , s. k. & kim , b. j. critical condition of the water - retention model .
e _ * 85 , * 032103 ( 2012 ) .
yan , j. , zhao , b. , wang , l. , zelenetz , a. & schwartz , l. h. marker - controlled watershed for lymphoma segmentation in sequential ct images .
phys . _ * 33 , * 24522460 ( 2006 ) .
ikedo , y. _ et al .
_ development of a fully automatic scheme for detection of masses in whole breast ultrasound images .
phys . _ * 34 , * 43784388 ( 2007 ) .
kerr , b. , neuhauser , c. , bohannan , b. j. m. & dean , a. m. local migration promotes competitive restraint in a host - pathogen tragedy of the commons. _ nature _ * 442 , * 7578 ( 2006 ) .
mathiesen , j. , mitarai , n. , sneppen , k. & trusina , a. ecosystems with mutually exclusive interactions self - organize to a state of high diversity .
lett . _ * 107 , * 188101 ( 2011 ) .
cieplak , m. , maritan , a. & banavar , j. r. optimal paths and domain walls in the strong disorder limit .
lett . _ * 72 , * 23202323 ( 1994 ) .
cieplak , m. , maritan , a. & banavar , j. r. invasion percolation and eden growth : geometry and universality .
lett . _ * 76 , * 37543757 ( 1996 ) .
fehr , e. _ et al .
_ new efficient methods for calculating watersheds .
_ j. stat .
_ p09007 ( 2009 ) .
fehr , e. , kadau , d. , andrade jr .
, j. s. & herrmann , h. j. impact of perturbations on watersheds . _ phys .
lett . _ * 106 , * 048501 ( 2011 ) .
andrade jr . , j. s. , oliveira , e. a. , moreira , a. a. & herrmann , h. j. fracturing the optimal paths . _
lett . _ * 103 , * 225503 ( 2009 ) .
oliveira , e. a. , schrenk , k. j. , arajo , n. a. m. , herrmann , h. j. & andrade jr . ,
j. s. optimal - path cracks in correlated and uncorrelated lattices .
e _ * 83 , * 046113 ( 2011 ) .
porto , m. , havlin , s. , schwarzer , s. & bunde , a. optimal path in strong disorder and shortest path in invasion percolation with trapping .
lett . _ * 79 , * 40604062 ( 1997 ) .
porto , m. , schwartz , n. , havlin , s. & bunde , a. optimal paths in disordered media : scaling of the crossover from self - similar to self - affine behavior .
e _ * 60 , * r2448r2451 ( 1999 ) .
barabsi , a .-
invasion percolation and global optimization .
_ * 76 , * 37503753 ( 1996 ) .
dobrin , r. & duxbury , p. m. minimum spanning trees on random networks .
* 86 , * 50765079 ( 2001 ) .
jackson , t. s. & read , n. theory of minimum spanning trees .
i. mean - field theory and strongly disordered spin - glass model . _ phys .
* 81 , * 021130 ( 2010 ) .
ifti , k. minimum spanning tree reflects the alterations of the default mode network during alzheimer s disease .
* 39 , * 14931504 ( 2011 ) .
goyal , s. & puri , r. k. formation of fragments in heavy - ion collisions using a modified clusterization method .
c _ * 83 , * 047601 ( 2011 ) .
hubbe , m. , harvati , k. & neves , w. paleoamerican morphology in the context of european and east asian late pleistocene variation : implications for human dispersion into the new world .
j. phys . anthropol . _ * 144 , * 442453 ( 2011 ) .
stauffer , d. & aharony , a. _ introduction to percolation theory _
( taylor and francis , 1994 ) .
broadbent , s. r. & hammersley , j. m. percolation processes : i. crystals and mazes .
_ cambridge philos .
* 53 , * 629641 ( 1957 ) .
coniglio , a. fractal structure of ising and potts clusters : exact results .
lett . _ * 62 , * 30543057 ( 1989 ) .
scholder , o. anti - red bond calculation algorithm in percolation .
c _ * 20 , * 267272 ( 2009 ) .
de gennes , p .-
_ scaling concepts in polymer physics _ ( cornell univ
. press , 1979 ) .
chang , i. & aharony , a. flory approximation for self - avoiding walks near the theta - point on fractal structures . _
i _ * 1 , * 313316 ( 1991 ) .
poole , p. h. , coniglio , a. , jan , n. & stanley , h. e. universality classes of the @xmath43 and @xmath111 points .
b _ * 39 , * 495504 ( 1989 ) . newman , c. m. & schulman , l. s. infinite clusters in percolation models .
_ j. stat .
_ * 26 , * 613628 ( 1981 )
. havlin , s. & nossal , r. topological properties of percolation clusters .
_ j. phys . a : math .
* 17 , * l427l432 ( 1984 ) .
fisher , m. e. critical probabilities for cluster size and percolation problems .
_ j. math .
* 2 , * 620627 ( 1961 ) .
wierman , j. c. substitution method critical probability bounds for the square lattice site percolation model . _
comput . _ * 4 , * 181188 ( 1995 ) .
fortuin , c. m. & kasteleyn , p. w. on the random - cluster model : i. introduction and relation to other models .
_ physica _ * 57 , * 536564 ( 1972 ) .
wu , f. y. percolation and the potts model .
_ j. stat .
phys . _ * 18 , * 115123 ( 1978 ) .
coniglio , a. & klein , w. clusters and ising critical droplets : a renormalisation group approach .
_ j. phys . a : math .
* 13 , * 27752780 ( 1980 ) .
deng , y. , blte , h. w. j. & nienhuis , b. geometric properties of two - dimensional critical and tricritical potts models .
e _ * 69 , * 026123 ( 2004 ) .
arajo , n. a. m. & herrmann , h. j. explosive percolation via control of the largest cluster .
lett . _ * 105 , * 035701 ( 2010 ) .
schrenk , k. j. , arajo , n. a. m. & herrmann , h. j. gaussian model of explosive percolation in three and higher dimensions .
* 84 , * 041136 ( 2011 ) .
prakash , s. , havlin , s. , schwartz , m. & stanley , h. e. structural and dynamical properties of long - range correlated percolation .
a _ * 46 , * r1724r1727 ( 1992 ) .
sahimi , m. long - range correlated percolation and flow and transport in heterogeneous porous media .
_ j. phys .
i france _ * 4 , * 12631268 ( 1994 ) .
sahimi , m. & mukhopadhyay , s. scaling properties of a percolation model with long - range correlations .
e _ * 54 , * 38703880 ( 1996 ) .
makse , h. a. , havlin , s. , schwartz , m. & stanley , h. e. method for generating long - range correlations for large systems .
e _ * 53 , * 54455449 ( 1996 ) .
, edited by peitgen , h. & saupe , d. ( springer , 1988 ) .
mandelbrot , b. b. & van ness , j. w. fractional brownian motions , fractional noises and applications .
_ siam rev . _
* 10 , * 422437 ( 1968 ) . | discretized landscapes can be mapped onto ranked surfaces , where every element ( site or bond ) has a unique rank associated with its corresponding relative height .
by sequentially allocating these elements according to their ranks and systematically preventing the occupation of bridges , namely elements that , if occupied , would provide global connectivity , we disclose that bridges hide a new tricritical point at an occupation fraction @xmath0 , where @xmath1 is the percolation threshold of random percolation . for any value of @xmath2 in the interval @xmath3 ,
our results show that the set of bridges has a fractal dimension @xmath4 in two dimensions . in the limit @xmath5
, a self - similar fracture is revealed as a singly connected line that divides the system in two domains .
we then unveil how several seemingly unrelated physical models tumble into the same universality class and also present results for higher dimensions .
any real landscape can be duly coarse - grained and represented as a two - dimensional _ discretized map _ of regular cells ( e.g. , a square lattice of sites or bonds ) to which average heights can be associated .
this process is exemplarily shown in figs .
[ fig::rank](a)-(c ) . as such
, the concept of discretized maps has been considered as a way to delimit spatial boundaries in a wide range of seemingly unrelated problems , ranging from tracing water basins and river networks in landscapes @xcite to the identification of cancerous cells in human tissues @xcite , and the study of spatial competition in multispecies ecosystems @xcite . moreover , previous studies have shown that cracks or surviving paths through discretized maps possess a universal fractal dimension which can be physically realized in terms of optimal paths under strong disorder @xcite , optimal path cracks @xcite , loopless percolation @xcite , or minimum spanning trees @xcite . here
we show that all these problems can be understood in terms of the same universal concept of _ fracturing a ranked surface_. we start by defining a ranked surface . given a two - dimensional discretized map of size @xmath6
, we generate a list containing the heights of its elements ( sites or bonds ) in crescent order , and then replace the numerical values in the original map by their corresponding ranks . as depicted in fig .
[ fig::rank](d ) , the result is a ranked surface .
the process of fracture generation is rather simple .
once the ranked surface is obtained , we sequentially occupy the elements of an empty lattice with the same size following the crescent rank order of the corresponding elements ( i.e. , in the same position ) on the ranked surface . during each step of the allocation process , only bridges , identified as those lattice elements which , once occupied , would create a spanning cluster ( i.e. , a globally connecting cluster ) @xcite , are never occupied .
these elements will eventually form a macroscopic fracture . in fig .
[ fig::snap ] we show the evolution of the fracture line on a large ranked surface with the fraction of occupied bonds @xmath2 .
as displayed , the lattice is initially seeded by a set of disconnected bridge elements at low values of @xmath2 , while for @xmath5 the fracture finally emerges towards a singly connected line that divides the system in two . as we show later in this article
, our results reveal that this line is fractal with dimension @xmath4 .
interestingly , this value is statistically identical to the dimensions of fractures generated from different models previously investigated @xcite .
however , at the percolation threshold value of the classical random percolation model @xcite , @xmath7 , the set of bridge bonds appearing in any configuration of our model should be identical to the set of the so - called _ anti - red bonds _ in random percolation @xcite
. as first proposed by coniglio @xcite and numerically verified by scholder @xcite , at @xmath7 , this set is also fractal , but with dimension @xmath8 ( @xmath9 in 2d ) , where @xmath10 is the correlation length exponent .
motivated by this substantial change in the fractal behavior at distinct stages of our fracturing process , in what follows we address how the set of bridges scales with the fraction of occupied bonds and the system size at @xmath7 analogously to a theta point @xcite , while for all values of @xmath2 above @xmath11 , it has a fractal dimension @xmath12 .
moreover , we introduce a new tricritical crossover exponent , which we study up to dimension six , the upper - critical dimension of percolation . |
the rationale for the _ xmm_-large scale structure ( _ xmm_-lss ) survey was presented in @xcite . a first catalogue for the 5.5 @xmath0 surveyed until year 2003 was presented in ( * ? ? ?
* hereafter paper i ) . in the present paper
, we supersede the first release with a new version which covers the entire 11.1 @xmath0area of the survey .
all the data were processed or re - processed afresh with the latest version of our pipeline ( see section [ secxamin ] ) .
we release two families of x - ray database tables ( see section [ seconline ] ) , a standard catalogue ( termed 2xlss ) for event files truncated to a common uniform exposure of 10 ks , and a _ deeper _ catalogue ( termed 2xlssd ) using the full exposure time . the _ xmm_-lss survey area , located around @xmath1 @xmath2 , was covered in the optical band by the canada france hawaii telescope legacy survey wide and deep synoptic fields ( cfhtls - w1 and d1 ) ; in the nir band partially by the ukirt infrared deep sky survey ( ukidss ; @xcite ) ; in the ir by the _
spitzer _ wide - area infrared extragalactic survey ( swire ; @xcite ) ; and in the uv by the galaxy evolution explorer ( _ galex _ ; @xcite ) all - sky survey .
we release also a multiwavelength database table ( using data from the sources just described ) in conjunction with each of the x - ray table families .
data from the present catalogue have already been used in other works , published , submitted or in preparation , e.g. , , @xcite , @xcite , @xcite .
the plan of the paper is as follows : in section [ secnewcat ] we describe the layout and content of our catalogue , in particular section [ seconline ] presents our database system by which users can have public access to the entire data tables and associated data products ( a reduced summary will be available via the centre de donnes de strasbourg [ cds ] ) ; section [ secnewproc ] describes the x - ray data processing and section [ secopt ] the generation of the multiwavelength catalogue .
finally some statistics are presented in section [ secstat ] , and concluding remarks in section [ secconclusion ] .
[ cols="<,<,<,>,<,<,<,>,<,<,<,>",options="header " , ] probabilities of chance association between a counterpart and an x - ray source are computed using formula ( [ eq1 ] ) in section [ secoptprob ] , where the density @xmath3 is computed from simple linear fits as reported in table [ tabprob ] .
the same table indicates also the magnitudes or fluxes used to look - up at the density for the appropriate band .
we did only a rough estimate of the parametrization of the density for the entire _
xmm_-lss area , neglecting any possible spatial variation .
x - ray to cfhtls probability , called @xmath4 , is computed for sources with a cfhtls counterpart in the order : d1 if present , else w1 . in the case of undefined cfhtls magnitudes ,
the pointing tile limiting magnitude was used ( read directly from the w1 table , or fixed to @xmath5 for d1 ) .
the other probabilities , x - ray to swire probability @xmath6 , x - ray to ukidss probability @xmath7 , and x - ray to galex probability @xmath8 , are computed as given in the notes in table [ tabprob ] .
a probability of 99 ( `` undefined '' ) is assigned whenever it could not be computed .
a statistics of the probability ranking defined in section [ secoptprob ] is shown in fig .
[ fighisto ] .
this figure indicates an acceptable tuning with the cfhtls , swire and ukidss catalogues , while _ galex _ data are perhaps _ overtuned _ in the sense that there is an excess of good probabilities .
this may indicate that the probability computation has to be revised .
we provide here details on the comparisons : ( a ) between 2xlssd and version i xlss ( section [ secxstatnewold ] ) ; ( b ) between 2xlssopt and 2xlssoptd ( section [ secappcompoct ] ) ; ( c ) between our catalogues and the xmds one ( section [ secappxmds ] ) ; ( d ) between our catalogues and the 2xmm one ( section [ secappwatson ] ) .
we first checked that the new pipeline version provides results consistent with the previous idl version by performing detailed tests on simulated and real _ xmm _
the detection parts of both pipelines give nearly identical results for point - like and extended sources .
the characterization parts ( maximum likelihood fitting ) are in excellent agreement for point - like sources . regarding extended sources , comparison of fluxes , sizes and likelihoods estimates from the two pipelines are in very good agreement .
some differences do however show up for individual faint sources or sources close to the detector borders and gaps .
these differences can be attributed to statistical fluctuations , and comparing these values to the input characteristics of simulated sources in a statistical sense , we conclude that both pipelines perform equivalently .
finally we compared directly the results .
the association between sources in the version 2 catalogues ( 2xlss and 2xlssd ) with the earlier xlss release ( in the smaller area covered by the latter ) is possible using the database column @xmath9 , as explained in section [ secnaming ] .
2824 sources are common between xlss and 2xlssd ( out of 3385 in xlss ) .
there are 452 objects appearing only in 2xlssd in the pointings covered by xlss , and 561 xlss objects not confirmed in 2xlssd . a majority ( respectively 95% of the former and 88% of the latter ) have a poor detection likelihood ( @xmath10 corresponding conventionally to @xmath11 ) . as an example of the rather good agreement between the two catalogue versions ( mainly between the older and newer xamin version , and
also between the 6 vs 10 band merging ) we compare the fluxes between xlss and 2xlssd ( which makes sense since xlss was also using full exposures ) , via the plots given in fig .
[ figcompa3 ] .
there are 20837 potential counterpart sets for 6721 x - ray sources in input for 2xlssoptd , and 16813 for 5572 sources for 2xlssopt ( the comparison includes negative rank _ rejected _ counterparts since rejection may act differently because of the displacement ) . of these , 1188 entries ( for 428 x - ray sources )
in 2xlssopt have no obvious correspondent in the deep 2xlssoptd . in 38 additional cases ( only 0.2% of the common ones )
the counterpart set is the same between a couple of 2xlssopt and 2xlssoptd entries , but the latter are not associated by column @xmath12 .
i.e. two sources have the same counterpart set but are not the closest . in 10 cases
this is due to ambiguous band merging , but in the rest ( which is anyhow an extremely small number ) this probably means that there are two 10 ks sources both displaced from but close to a given full exposure one ( or v.v . ) .
the 2171 remaining 2xlssopt counterpart sets ( covering 1479 individual 2xlss x - ray sources ) can be _ altogether different _ from all counterpart sets in 2xlssoptd for the associated x - ray source ( the x - ray source position moved so much that entire counterpart sets are farther than 6 from either position ) , or may partially match ( from 1 to 4 out of the 5 d1 , w1 , swire , ukidss or galex catalogues ) .
a different approach for the comparison is to consider only the _ best counterparts _ i.e. those ranked 0 - 1 ( by definition one per x - ray source ) .
let us exclude the 1607 2xlssoptd ( 24% of 6721 ) x - ray sources not confirmed in the 10ks catalogue , and the 528 2xlssopt ( 8% of 5572 ) new in the 10ks catalogue , and let us concentrate on the common sources . of their best counterparts : a subset of the xmds ( _ xmm _ medium deep survey ) data was published as the xmds / vvds @xmath14 catalogue (; the same paper contained also the logn - logs of the entire catalogue ) .
contextually to the release of the entire xmds catalogue through our database contextually with version 2 _ xmm_-lss ones ( see table [ tabtables ] ) , we report here some details of a comparison between the catalogues presented in this paper ( mainly 2xlssd ) and xmds . 1 .
the xmds covers the fields g01 to g19 in table [ tabpointing ] 2 .
xmds data were analysed with a _ fully independent _ ( and more traditional ) pipeline based on the one by @xcite .
the xmds pipeline uses the sas to detect candidates in 5 energy bands simultaneously ( and not in 2 independent bands with later merging ) operating on event files merged from all 3 _ xmm _ cameras and from the entire _ xmm _ field of view ( not just the central 13 ) , 4 .
it then applies the @xcite characterization , which is robust but oriented to point sources only ( unlike the wavelet method in xamin which handles better extended sources ) .
the event pattern selection is different ( non - standard and broader in xmds ) .
the removal of redundant sources is handled differently , in particular the primary detection is chosen differently , the position is inherited from the primary detection , but the flux is obtained stacking data from all overlapping pointings .
the astrometric correction offsets are different . 8 .
also the xmds catalogue does not include spurious objects ( but only those above a probability threshold ) , so the difference between the raw database table and the catalogue is only due to the overlap removal procedure . as anticipated in section [ secxmds ] ,
the xmds catalogue includes 1168 sources , by definition all in the g - labelled fields .
comparison with the 2xlssd or 2xlss catalogues may also involve adjacent b fields if the _ xmm_-lss overlap removal procedure preferred those .
the association between xmds and _ xmm_-lss sources is done within a radius of 10 .
the comparison occurs naturally with the deep catalogue , because it involves the _ same _ input ( odf ) event data ( for the full exposures ) with different pipelines and procedures . of those 1168 objects : similarly 43 are potential ultrasoft sources ( the band with the highest s / n ratio in xmds is the a band [ 0.3 - 0.5 kev ] , which is not processed by the current release of xamin ) , so they are legitimately excluded ; concerning the distance between the ( astrometrically corrected ) x - ray positions in the xmds and 2xlssd catalogues , 58% of the sources are closer than 2 , 88% closer than 4 and only 4% more distant than 6 , in general concentrated among the sources with lesser significance , and the few extended ones .
the agreement between xmds and 2xlssd positions , peaking around 1 , is better than the typical inter - band distance between 2xlssd detections in the two energy bands , which peaks around 2 . compare panels ( d ) and ( c ) of fig .
[ figdisthisto ] .
we have cross calibrated graphically the detection likelihood of xamin with the chance probability of the xmds ( for definition see @xcite ) and the detection likelihood of xamin with the significance in terms of number of @xmath17 of the xmds ( see and references therein ) .
we only show the latter in fig .
[ xmdssn ] .
one can see that a likelihood of 75 corresponds more or less to the @xmath14 level , and one of 40 to the @xmath16 level .
the xmds by construction includes measurements in all 5 energy bands even if the source is above the probability threshold only in one .
if we consider as good detections for xmds only those with @xmath18 in the band , we have that , of the sources in common with 2xlssd : the fluxes , computed for xmds according to the prescriptions of @xcite and for _ xmm_-lss as explained in section [ secflux ] , are compared in fig .
[ xmdsflux ] .
extended sources classified c1 are excluded as their 2xlssd flux is set to undefined .
the fluxes match qualitatively , although there is a systematic difference : namely 2xlssd fluxes are 0.895 lower than the xmds fluxes in the b band , while they are only 1.040 higher in the cd band .
it can be seen that a larger scatter in fluxes occurs for the sources which have poorer significance in either catalogue , while outliers are generally due to sources presumably falling near a chip gap on one detector ( and as such characterized by a fluxflag of 2 ) , or exceptionally by residual c2 extended sources ( for which the _ xmm_-lss flux is computed from the point - like rate ) .
it is also possible , similarly to what was done in section [ seccompoct ] , to compare the counterparts in optical ( and other ) bands between the xmds catalogue and one of our _ xmm_-lss catalogues .
in doing this one should consider : 1 . that the identification procedure for xmds is historically different from the one used here , in particular it was done in several incremental steps , and uses distances capped to 2 in the computation of probabilities , 2 . that the catalogues used for xmds were in larger number ( for a total of 27 ) and included many other , including older , data sources ( e.g. vvds , radio data , simbad and ned , cfhtls t003 , etc . )
therefore we present a comparison limited to the 2xlssoptd catalogue ( the one which matches better xmds in exposures ) , and , for xmds , to _
reduced counterpart sets _ considering only cfhtls t004 d1 and w1 , swire dr6 and galex .
ukidss was not included since 2xlssoptd uses release dr5 , while xmds used release dr3 .
more specifically we consider only the 1057 x - ray sources in common between xmds and 2xlssd , as described in [ secappxmdsx ] .
they correspond to 4316 counterpart sets ( of any rank ) in 2xlssoptd and 4916 in xmds .
we found that a large fraction ( 3620 ) of the possible counterpart sets are identical ( i.e. they have the same counterparts , irrespective of ranking , in both catalogues ) .
we quickly compared our 2xlssd catalogue with the 2xmm catalogue .
namely we used the 2xmmi - dr3 `` slim '' reduced catalogue , which contains exposure - merged sources ( not individual detections ) , and thus somewhat compare with our post overlap - removal catalogues .
we restricted a comparison to a rectangular area fully encompassing 2xlssd and noticed that such an area includes precisely just our pointings , with the only exception of a single additional pointing centered on the bright star mira ceti ( obsid 014850 0201 ) .
the rectangular area contains 6181 2xmm sources . since the slim catalogue does not contain indication on the pointings , we can tentatively flag sources in the mira ceti field as the 60 within 13 from the respective pointing centre .
we checked the association with our 6721 2xlssd sources within the customary radius of 10 .
we find : we note that 59% of the unassociated 2xlssd sources have a rather poor likelihood ( @xmath19 ) , and 93% are below @xmath20 ( i.e. @xmath16 ) . despite the different definition of the 2xmm likelihood ( parameter @xmath21 )
, we note that of 1141 unassociated 2xmm sources , 67% have @xmath22 . of the associated cases ,
all but 69 have a single association .
the ambiguous ones are all plain couples , and usually well separated ( one 2xmm source at 1 - 2 from the 2xlssd position , and the other at 8 - 9 ) .
the distance between the astrometrically corrected position ( 2xmm used the usno catalogue for this purpose ) for the associated primary ( closest ) cases is within 2 in 55% of the cases , within 4 in 86% and within 6 in 95% ( while 90% of the 69 secondaries are above 6 ) .
the histogram of distances however peaks at a 1 offset ( not unlike xmds vs 2xlssd ) .
the common subset contains 100 of our extended sources , and 88 2xmm extended sources ( i.e. those with their parameter @xmath23 ) .
56 are considered extended in both catalogues .
110 of our extended sources do not appear at all in 2xmm , and 113 2xmm extended sources do not appear in 2xlssd .
the 2xmm slim catalogue does not contain count rates .
it contains fluxes in several energy bands , but these are different from ours .
we can directly compare the sum of 2xmm bands 2 and 3 with our b band ( 0.5 - 2 kev ) , while at harder energies we can compare the sum of 2xmm bands 4 and 5 ( 2 - 12 kev ) with our cd band ( 2 - 10 kev ) . the comparison of the fluxes is reported in fig .
[ fig2xmm ] .
the average 2xmm/2xlssd flux ratio for common point - like sources is 0.92 in the soft band ( same energy range ) , and 1.22 in the hard bands ( where 2xmm extends 2 kev further ) .
again , considering the differences in the pipelines , the agreement is acceptable . | we present the final release of the multi - wavelength _ xmm_-lss data set , covering the full survey area of 11.1 square degrees , with x - ray data processed with the latest _
xmm_-lss pipeline version .
the present publication supersedes the @xcite catalogue pertaining to the initial 5 @xmath0 .
we provide x - ray source lists in the customary energy bands ( 0.5 - 2 and 2 - 10 kev ) for a total of 6721 objects in the deep full - exposure catalogue and 5572 in the 10ks - limited one , above a detection likelihood of 15 in at least one band .
we also provide a multiwavelength catalogue , cross - correlating our list with ir , nir , optical and uv catalogues .
customary data products ( x - ray fits images , cfhtls and swire thumbnail images ) are made available together with our interactively queriable database in milan , while a static snapshot of the catalogues has been supplied to cds .
[ firstpage ] catalogues , surveys , x - rays : general |
this paper derives an exact deterministic nonlinear observer to compute the continuous state of an inertial navigation system based on partial discrete measurements .
the main analysis tool is nonlinear contraction theory @xcite .
recent work on nonlinear observer design for mechanical systems based on nonlinear contraction theory can be found in @xcite .
specifically , we consider the classical strap - down problem in inertial navigation @xcite , where angular position ( euler angles ) @xmath0 and inertial position * r * are computed from the body turn rate @xmath1 and inertial acceleration @xmath2 , measured continuously in intrinsic ( body - fixed ) coordinates , @xmath3 with @xmath4\ ] ] and @xmath5\ ] ] as made precise in @xcite such a system lies at the boundary between convergence and divergence , much like a triple integrator . in this paper ,
the continuous measurements of @xmath1 and @xmath2 are augmented by _
discrete _ measurements of @xmath6 and @xmath7 , leading to a globally exponentially convergent nonlinear observer design .
such combinations of measurements are typical in inertial navigation , whether for vehicles or robots ( see e.g. @xcite for a recent discussion ) .
the human vestibular system also features a similar structure , with otolithic organs measuring linear acceleration and semi - circular canals estimating angular velocity through heavily damped angular acceleration signals , an information then combined with visual data at much slower update rate .
section [ sec : general ] introduces the basic observer designs .
we build simple observers to compute @xmath8 based on partial discrete measurements @xmath9 and @xmath10 . in section [ sec : extension ] we discuss extensions , such as the use of nonlinear measurements , and the effects of system disturbance and measurement disturbance @xcite .
we also study the case where the inertial navigation system is expressed in quaternion form @xcite .
section [ sec : simulation ] presents simulation results on a 3-dimensional system .
brief concluding remarks are offered in section [ sec : conclusion ] .
a very brief review of basic results in contraction theory is presented in the appendix .
in this section , we construct a discrete observer for system ( [ eq : system ] ) , which consists of a hierarchy of three sub - systems , mirroring the hierarchical nature of systems physics ( [ eq : system ] ) . the observer is based on the partial - measurements of the state * x * and * r * at a series of instants @xmath11 .
+ + * first * , based on the discrete measurement @xmath12 , compute * x * with the observer @xmath13 where the first equation describes a continuous update between measurements , and the second equation a discrete measurement incorporation .
virtual displacements , which are systematically used in mathematical physics and optimization theory , also represent basic tools in contraction theory ( see appendix ) .
computing virtual displacements in ( [ eq : observer_ep1 ] ) leads to @xmath14 based on @xcite , define @xmath15 with @xmath16 .
this implies that @xmath17 + from ( [ eq : delta_x_hat ] ) , we have @xmath18 from hybrid contraction condition ( [ eq : hybrid_condition ] ) in the appendix , if @xmath19 where @xmath20 , then both @xmath21 and @xmath22 tend to zero exponentially .
so @xmath23 tends to * x * exponentially .
+ + * second * , based on the discrete measurement of @xmath24 , compute * v * with the observer @xmath25 from ( [ eq : observer_ep2 ] ) and the first step , we get @xmath26 since @xmath27 tends exponentially to a constant , we have @xmath28 using ( [ eq : dp_delta_v ] ) , this implies that @xmath29 , which by continuity implies that the constant which @xmath27 tends to must be zero .
we thus have , exponentially , @xmath30 since by design @xmath31 is a particular solution of ( [ eq : observer_ep2 ] ) , this implies that @xmath32 tends to @xmath33 exponentially .
+ + * third * , based on the discrete measurement @xmath34 , use the observer @xmath35 since we know @xmath27 tends to zero exponentially , we have @xmath36 if @xmath37 , i.e. @xmath38 where @xmath39 is the largest eigenvalue of @xmath40 .
so @xmath41 tends to * r * exponentially .
+ * extension 1 * : when we compute * v * and * r * , we only use the discrete - time measurement @xmath34 without @xmath42 .
this allows @xmath42 and @xmath34 to be measured at different instants , with the same computation .
+ + * extension 2 * : the metric can also be written @xmath43 since * a * is orthogonal .
so we can simply use @xmath44 .
+ + * extension 3 * : assume that in ( [ eq : observer_ep1 ] ) we replace the discrete update law by the more general @xmath45 where @xmath46 and @xmath47 commute .
then @xmath48 the hybrid contraction condition ( [ eq : condition_ep1 ] ) becomes @xmath49 where @xmath50 is the largest eigenvalue of @xmath51 .
note that because the generalized jacobians are zero at each step of the hierarchy , the hybrid contraction conditions simply define the _ metrics _ in which the discrete measurement incorporation steps should be contracting . as we shall see later
, the flexibility offered within this constraint will allow us to trade - off model error vs measurement error , similarly in spirit to a standard kalman filter .
discussions about full discrete measurements , use of a linear observer , nonlinear measurement , disturbance effects , and quaternion representation are offered in this section . an observer based on full measurement
is described in section [ sec : full ] .
effects of system disturbance and measurement disturbance are discussed in section [ sec : error ] .
section [ sec : nonlinear ] we develop a more general discrete observer applicable to nonlinear measurements .
use of quaternions is studied in section [ sec : quaternion ] .
assume that _ all _ states * x * , * v * , and * r * are actually measured , at a series of discrete instants @xmath52 .
then steps 1 and 3 are unchanged , but we can replace step 2 ( the estimation of @xmath33 ) by the observer @xmath53 since we know @xmath22 tends to zero exponentially , we have @xmath54 with @xmath55 , we have @xmath56 where @xmath57 is the largest eigenvalue of @xmath58
. so @xmath32 tends to * v * exponentially .
note that in some cases one only needs to estimate orientation @xmath6 and velocity @xmath33 , and that the discrete measurement of @xmath33 may be obtained from optical flow , which can be computationally `` expensive '' and thus infrequent .
effects of bounded inputs and measurement disturbances can be quantified and obeserver gains chosen accordingly .
consider input disturbance @xmath59 and measurement disturbance @xmath60 , with @xmath61 and @xmath62 , leading to the modified system @xmath63 using the basic robustness result in @xcite , we can quantify the corresponding quadratic bounds @xmath64 on the estimation error @xmath65 where @xmath66 is the largest eigenvalue of the symmetric part of @xmath67 .
define the objective function ( @xmath68 ) @xmath69 then , @xmath70 , where @xmath71 and @xmath72 .
we know @xmath73 should also satisfy @xmath74 define @xmath75 as an upper bound of @xmath73 .
therefore , @xmath76 where @xmath77 .
finally , we obtain the minimum of @xmath78 @xmath79 where @xmath71 and @xmath72 .
when different measurements are available , the above formulas can also be used to select _ a priori _ the most informative measurement .
this can be the case for instance for selecting the direction of gaze of the eyes in hopping robot @xcite .
this can also be the case when the measurements are `` expensive '' , for instance computationally . + * extension * : the discussions above will still work when the bounds of input disturbance and measurement disturbance are time - varying . if @xmath80 and @xmath81 when @xmath82 .
similar to the above , we have @xmath83 where @xmath71 and @xmath84 .
for the system @xmath85 , consider the observer @xmath86 where @xmath87 we have @xmath88 defining @xmath89 , we have @xmath90 . using equation ( [ eq : nonlinear_delta ] ) yields @xmath91 where @xmath92 and @xmath93 .
the sufficient contraction condition on hybrid systems can be written @xmath94 where @xmath95 and @xmath66 is the largest eigenvalue of the symmetric matrix @xmath96 .
if condition ( [ eq : condition_gn ] ) is satisfied by an appropriate choice of @xmath97 , then @xmath98 will tend to * x * exponentially . a a simple illustration , consider using distance measurements instead of direct cartesian position measurements . in the 3-dimensional space ,
measure the distances from one point @xmath99 to four time - varying reference points @xmath100^t$ ] , @xmath101^t$ ] , @xmath102^t$ ] , and @xmath103^t$ ] , @xmath104 the discrete - update part of observer ( [ eq : nonlinear_observer ] ) can be built up as below , @xmath105
= \left [ \begin{array}{l } \hat{x}_{1,i}^- \\ \hat{x}_{2,i}^- \\ \hat{x}_{3,i}^- \end{array } \right ] -\frac{1}{2 } { \bf k}_i \left [ \begin{array}{l } ( \hat{y}_{1,i}^-)^2-(\hat{y}_{2,i}^-)^2-(y_{1,i}^2-y_{2,i}^2 ) \\ ( \hat{y}_{2,i}^-)^2-(\hat{y}_{3,i}^-)^2-(y_{2,i}^2-y_{3,i}^2 ) \\ ( \hat{y}_{3,i}^-)^2-(\hat{y}_{4,i}^-)^2-(y_{3,i}^2-y_{4,i}^2 ) \end{array } \right]\ ] ] where @xmath106 is a 3 by 3 time - varying gain matrix . using equation ( [ eq : nonlinear_y ] ) yields @xmath107 @xmath108\ ] ] where subscript @xmath109 refers to the value at time @xmath110 .
+ assume @xmath111 is non - singular .
then we can choose @xmath112 with equation ( [ eq : nonlinear_dvarying ] ) , we have @xmath113 by choosing @xmath114 , we can make @xmath115 satisfy the following contraction contidion that makes @xmath21 tends to zero exponentially.@xmath116 where @xmath117 and @xmath66 is the largest eigenvalue of the symmetric matrix @xmath96 .
therefore , @xmath22 will tend to zero , and @xmath98 will tend to * x * exponentially .
+ * remark * when @xmath111 is singular , one has @xmath118 + equation ( [ eq : nonlinear_singular1 ] ) is equivalent to @xmath119 \cdot \left| \begin{array}{ccc } { \bf i } & { \bf j } & { \bf k } \\ ( c_{1i}-b_{1i } ) & ( c_{2i}-b_{2i } ) & ( c_{3i}-b_{3i } ) \\ ( d_{1i}-c_{1i } ) & ( d_{2i}-c_{2i } ) & ( d_{3i}-c_{3i } ) \end{array } \right|=0\ ] ] + which we can write @xmath120 this means that points a , b , c , and d are in the same plane , and therefore that the geometry does not contain enough information to infer position .
+ to compute velocity , one can rewrite observer ( [ eq : observer_ep2 ] ) as @xmath121 - { \bf k}_i \left [ \begin{array}{c } ( \hat{y}_{1,i}^+)^2-(\hat{y}_{2,i}^+)^2-(y_{1,i}^2-y_{2,i}^2 ) \\ ( \hat{y}_{2,i}^+)^2-(\hat{y}_{3,i}^+)^2-(y_{2,i}^2-y_{3,i}^2 ) \\ ( \hat{y}_{3,i}^+)^2-(\hat{y}_{4,i}^+)^2-(y_{3,i}^2-y_{4,i}^2 ) \end{array } \right ] \ } \end{array } \right.\ ] ] where @xmath122 and @xmath123^{-1 } ~~\textrm{and}~~ { \bf k}_{i+1}= \left [ \begin{array}{ccc } ( b_{1i+1}-a_{1i+1 } ) & ( b_{2i+1}-a_{2i+1 } ) & ( b_{3i+1}-a_{3i+1 } ) \\ ( c_{1i+1}-b_{1i+1 } ) & ( c_{2i+1}-b_{2i+1 } ) & ( c_{3i+1}-b_{3i+1 } ) \\ ( d_{1i+1}-c_{1i+1 } ) & ( d_{2i+1}-c_{2i+1 } ) & ( d_{3i+1}-c_{3i+1 } ) \end{array } \right]^{-1}\ ] ]
we then have @xmath124 which is the same as equation ( [ eq : dp_delta_v ] ) . similarly to the second step of section [ sec : general ] , this shows that @xmath125 tends to * v * exponentially .
+ note that the geometry problem of going from distances to positions is solved by a dynamic system , the observer , rather than explicitly at each instant .
in general , one may also use linear measurements at some instants and nonlinear ones at others . note that
if a measurement is delayed , the algorithms work similarly but the actual information is available after the delay ( i.e. the measurement is incorporated at some past time and the forward simulation runs instantly to the current time ) .
consider now , extending section [ sec : error ] , the effect of model and measurement errors .
for the system @xmath126 , with the following nonlinear observer , @xmath127 where@xmath128 and @xmath129 + + so we have @xmath130 we know the quadratic bounds @xmath64 on the estimation error @xmath131 where @xmath132 , @xmath133 , and @xmath134 is the largest eigenvalue of the symmetric part of @xmath67 .
+ + we can choose the most relevant discrete update function @xmath135 which will best contribute to improving the estimate @xmath98 ( i.e. , to minimize @xmath136 ) .
angular position can be expressed in quaternion form , avoiding representation singularities @xcite .
quaternions express a rotation of angle @xmath137 about the unit vector @xmath138 as @xmath139 . with @xmath140
the quaternion vector , this leads to @xmath141 where @xmath142\ ] ] and @xmath143\ ] ] in this representation , the fact that the dynamics of * q * is indifferent is obvious , since @xmath144 is skew - symmetric .
the observers can be derived as earlier , simply by replacing ( [ eq : observer_ep1 ] ) by @xmath145 based on the discrete measurements @xmath146 .
computing virtual displacements @xmath147 and because the dynamics of * q * is indifferent , we only need @xmath148 where @xmath50 is the largest eigenvalue of @xmath149 . under condition ( [ eq :
condition_qp1 ] ) , @xmath150 tends to zero exponentially , and @xmath151 tends to * q * exponentially .
the other two steps are unchanged , with @xmath152 being replaced by @xmath153 .
all the above variations and extensions can of course be combined .
in this section , we will do a 3-dimentional simulation about system ( [ eq : system ] ) based on the discrete measurement @xmath42 and the nonlinear distance measurements @xmath154 , @xmath155 , @xmath156 , and @xmath157 , as in section [ sec : nonlinear ] .
consider system ( [ eq : system ] ) in the 3-dimensional case . where @xmath158~~\textrm{and}~~ { \bf \gamma}=~ \left [ \begin{array}{c } \cos(2 t ) \\
\frac{1 + 2\sin t}{3 } \end{array } \right]\ ] ] four time - varying reference points are chosen as below ( all move on circular trajectories ) , @xmath159 @xmath160 observer ( [ eq : observer_ep1 ] ) with @xmath161 is used to compute * x*. observer ( [ eq : observer_ep2_distance ] ) with gain ( [ eq : nonlinear_gain ] ) is used to compute * v*. using observer ( [ eq : nonlinear_observer],[eq : nonlinear_od1 ] ) and gain ( [ eq : nonlinear_k_i ] ) , we choose @xmath162 to satisfy condition ( [ eq : condition_gn ] ) , thus we can compute * r*. figure [ fig : example_ep ] shows @xmath163 tends to @xmath164 exponentially .
observers similar to those developed in this paper can in principle be applied to other continuous nonlinear systems besides inertial navigation systems , although much simplification was afforded by exploiting the hierarchical structure of the system physics .
an animation of the basic observer as applied to head stabilization @xcite in a simulated robot hopper @xcite can also be found in _ http://web.mit.edu / nsl / www / hopping_robot.mpg_. * acknowledgement * this paper benefited from stimulating discussions with dr .
agostino martinelli .
0 aghannan , n. , rouchon , p. , an intrinsic observer for a class of lagrangian systems , _
i.e.e.e . trans .
control _ ,
* 48(6 ) * ( 2003 ) .
berthoz , a. , the brain s sense of movement , _ harvard university press , cambridge , massachusetts _ ( 2000 ) .
egeland , o. , kristiansen , e. , and nguyen , t.d . , observers for euler - bernoulli beam with hydraulic drive , _
i.e.e.e . conf .
( 2001 ) .
goldstein , h. , classical mechanics , _ addison - wesley _ ( 1980 ) .
grassia , f.s . , practical parameterization of rotations using the exponential map , _ the journal of graphics tools _ , vol . 3.3
( 1998 ) .
hestenes , d. , new foundations for classical mechanics , _ kulwer academic publishers , dordrecht , the netherlands _ ( 1986 ) .
jouffroy , j. , and j. opderbecke , underwater vehicle trajectory estimation using contracting pde - based observers , _ american control conference , boston , massachusetts _ ( 2004 ) .
kristiansen , d. , and egeland , o. , time and spatial discretization methods that preserve passivity properties for systems described by partial differantial equations , _ in proceedings of the 2000 american control conference , chiago , illinois _ ( 2000 ) lohmiller , w. , and slotine , j.j.e .
, on contraction analysis for nonlinear systems , _ automatica _ , * 34(6 ) * ( 1998 ) .
lohmiller , w. , contraction analysis of nonlinear systems , _ ph.d .
thesis , department of mechanical engineering , mit _ ( 1999 ) .
lohmiller , w. , and slotine j.j.e . , nonlinear process control using contraction theory , _ a.i.ch.e
. journal _ ( 2000 ) .
lohmiller , w. , and slotine j.j.e . , control system design for mechanical systems using contraction theory , _
i.e.e.e . trans .
control _ ,
* 45(5 ) * ( 2000 ) .
slotine , j.j.e . , and li , w. , applied nonlinear control , _ prentice - hall _ ( 1991 ) .
slotine , j.j.e . , and lohmiller , w. , modularity , evolution , and the binding problem : a view from stability theory , _ neural networks _ , * 14(2 ) * ( 2001 ) .
raibert , m.h .
, legged robots that balance , _ the m.i.t . press , cambridge , massachusetts _ ( 1986 ) .
ryu , j. , rossetter e.j . , and gerdes , j.c . , vehicle sideslip and roll parameter estimation using gps . _
the proceedings of avec 2002 , hiroshima , japan _ ( 2002 ) .
varshalovich , d.a . ,
moskalev , a.n . , and khersonskii , v.k . ,
description of rotation in terms of the euler angles , _ singapore : world scientific _ ( 1988 ) .
the main theorem of contraction analysis @xcite can be stated as is uniformly negative definite , then all system trajectories then converge exponentially to a single trajectory , with convergence rate @xmath169 , where @xmath170 is the largest eigenvalue of the symmetric part of @xmath171 .
the system is said to be contracting .
[ th : theoremf ] it can be shown conversely that the existence of a uniformly positive definite metric with respect to which the system is contracting is also a necessary condition for global exponential convergence of trajectories . in the linear time - invariant case , a system is globally contracting if and only if it is strictly stable , with @xmath171 simply being a normal jordan form of the system and @xmath172 the coordinate transformation to that form . furthermore , since where is @xmath174 the symmetric part of @xmath171 , all transformations @xmath175 corresponding to the same @xmath176 lead to the same eigenvalues for @xmath177 , and therefore to the same contraction rate @xmath169 .
consider now a hybrid case @xcite , consisting of a continuous system @xmath178 which is switched to a discrete system @xmath179 every @xmath180 for one discrete step .
letting , in the _ same coordinate system @xmath172 _
, @xmath66 be the largest eigenvalue of the symmetric matrix @xmath96 , and @xmath181 be the largest eigenvalue of @xmath182 ( the corresponding discrete - time quantity , where @xmath183 , see @xcite ) , a sufficient condition for the overall system to be contracting is @xmath184 contraction theory proofs and this paper make extensive use of _ virtual displacements _ , which are differential displacements at fixed time borrowed from mathematical physics and optimization theory .
formally , if we view the position of the system at time @xmath185 as a smooth function of the initial condition @xmath186 and of time , @xmath187 , then @xmath188 . | we derive an exact deterministic nonlinear observer to compute the continuous state of an inertial navigation system based on partial discrete measurements , the so - called strapdown problem .
nonlinear contraction is used as the main analysis tool , and the hierarchical structure of the system physics is sytematically exploited .
the paper also discusses the use of nonlinear measurements , such as distances to time - varying reference points . |
the search for galaxy alignments has a long history , beginning with searches for alignments in `` spiral and elliptical nebulae '' during the late 19th century .
recent scrutiny of the problem has been motivated by the understanding that establishing the level of galaxy spin vector ( @xmath0 ) alignments could offer an additional constraint on various theories of galaxy formation and evolution .
for example , `` top - down '' scenarios of large - scale structure formation can lead to ordered distributions of angular momentum on cluster and supercluster scales through a variety of mechanisms ( @xcite , @xcite , @xcite , @xcite ) .
in addition to galaxy @xmath0 alignments resulting from various formation mechanisms , galaxy @xmath0 alignments may also be the evolutionary result of anisotropic merger histories ( @xcite ) , galaxy - galaxy interactions ( @xcite ) , or strong gravitational gradients ( @xcite , @xcite ) . for a summary of the history of the field , see djorgovski ( 1987 ) and cabanela & aldering ( 1998 ) [ hereafter paper i ] .
observational support exists for some forms of galaxy @xmath0 alignments with surrounding large - scale structure .
for example , @xcite discovered that the major axes of cd galaxies tended to be aligned with the axes of their parent cluster .
however , most previous searches for galaxy alignments have had results that one could describe as negative or statistically significant but not strongly so .
one complication in earlier efforts has been that most have not truly determined @xmath0 , but rather simply used the position angle ( and sometimes ellipticity ) of the galaxies in an attempt to determine the possible distribution of @xmath0 .
however , for each combination of galaxy position angle and ellipticity , there are four solutions for the true orientation of the galactic angular momentum axis ( @xmath0 ) .
this degeneracy in @xmath0 can only be removed by establishing both which side of the major axis is moving toward the observer and whether we are viewing the north or south side of the galaxy , where `` north '' is in the direction the galaxy angular momentum vector .
therefore previous studies have either restricted themselves to using only position angles of galaxies , or they have often taken all four possible solutions of @xmath0 with equal weight ( @xcite , @xcite ) .
several studies have been published regarding searches for alignments using completely determined galaxy angular momentum axes .
@xcite used and optical observations of 20 galaxies in the virgo cluster to show that no very strong @xmath0 alignments exist . however , a followup to this study by @xcite found evidence for anti - alignments of spin vectors for binary pairs of galaxies in a sample of 31 such pairs .
@xcite briefly investigated the possibility of galaxy alignments by plotting up the @xmath0 orientations for @xmath1 galaxies with fully determined spin vectors from their virgo cluster sample and found no obvious alignments .
most recently , @xcite used a sample of 60 galaxies from the third reference catalogue of bright galaxies ( @xcite , hereafter referred to as the rc3 ) in the `` ursa major filament '' and found no evidence of galaxy alignments .
there are several criticisms one can level against these earlier studies .
all the studies attempted to use relatively small samples to map out orientation preferences over the entire sky .
thus only very strong @xmath0 alignment signatures could have been discovered via this method .
the samples were selected using source catalogs with `` visual '' criteria which may have led to a biased sample .
for example , as noted in paper i , the source catalog for the `` ursa major filament '' study , the rc3 , suffers from the `` diameter - inclination effect , '' which leads to a strong bias for preferentially including face - on galaxies over edge - on galaxies of the same diameter ( @xcite ) .
finally , no attempt was made to consider the positions of the galaxies within the local large - scale structure before looking for alignments .
considering that the local mass density is critical for determining which alignment mechanism may be dominant , an attempt should be made to look for @xmath0 alignments relative to local large - scale structures .
this study is an attempt to avoid some of the issues citied above and obtain a sample of galaxies with well determined @xmath0 in various environments in a supercluster using a mechanically - selected sample of galaxies . for this study , we selected a subsample of the minnesota automated plate scanner pisces - perseus galaxy catalog ( hereafter maps - pp ) , which is a true major - axis diameter - limited catalog built using automated , mechanical methods and does not exhibit the `` diameter - inclination '' effect ( see paper i ) .
we determined the @xmath0 orientation for the galaxies in this subsample using observations .
the sample selection criteria are outlined in section [ hiobservations ] .
the analysis methods are discussed in section [ hianalysis ] .
section [ hiresults ] discusses the results of the data analysis .
our interpretation of these results is provided in section [ hiconclusions ] .
the galaxy sample for this study was selected from the maps - pp .
the maps - pp catalog was designed to avoid several of the pitfalls of previous attempts to measure galaxy orientations .
the maps - pp contains @xmath2 galaxies in the pisces - perseus supercluster field with ( roughly ) isophotal diameter @xmath330 constructed from digitized scans of the blue and red plates of the palomar observatory sky survey ( poss i ) . by using a mechanical measure of the diameter
, this catalog avoids the `` diameter - inclination '' effect seen in both the uppsala general catalog ( @xcite , hereafter ugc ) and the rc3 . the maps - pp also uses a two - dimensional , two - component fit of the galaxy light profile in order to obtain a more accurate position angle and ellipticity measurement for the component of the galaxy with most of the angular momentum ( e.g. - the disk in spirals ) .
such a full two - dimensional fit has been shown ( @xcite ) to be very effective at recovering the image parameters in situations were a simple ellipse fit fails ( e.g. - edge - on spirals with a large bulge ) .
more details as to the construction of the maps - pp are available in paper i. for this study , we selected a subsample of the maps - pp that could have their @xmath0 determined through observations and at the same time could probe the galaxy @xmath0 orientations relative to the large - scale structure of the pisces - perseus supercluster ( hereafter pps ) .
observations can determine which side of the major - axis is approaching us , reducing the four - fold degeneracy in the @xmath0 to two solutions .
however , because of the great distance to the pps ( @xmath4 ) , the poss i images do nt generally have enough detail to make out spiral arm structure , so determining if we were viewing the north or south side of a galaxy would be difficult without re - imaging the galaxies .
instead , we choose to constrain the inclinations of the galaxies in our subsample to be edge - on .
this means we effectively reduce the two - fold degeneracy in @xmath0 solution to a single solution and simultaneously we reduce the galaxy alignments analysis problem from a full three - dimensional problem to a much simpler one - dimensional problem . and
because the pps plane itself is viewed very close to edge - on ( @xcite ) , we are simplifying the problem without losing the ability to probe the angular momentum distribution in relationship to the pps plane . the primary requirement for including a maps - pp galaxy in this study was therefore an ellipticity greater than 0.66 .
other criteria for selecting a maps - pp galaxy for our program were based on observational considerations . to ensure the galaxy could be observed from arecibo , the declination was required to be less than 36 .
an o ( blue ) major - axis diameter between 44 and 100 was needed so that the disk of the galaxy was not too small to be targeted on both sides by the arecibo beam and not too large to be fully sampled .
the galaxy was required to be within 2.25 of the pps midplane ( as determined in paper i ) and if the redshift was known , it needed to be between @xmath5 in order to increase the chances it was a true pps member . finally to reduce the sample size , we selected galaxies with o magnitude brighter than 17 .
this maps - pp subsample consisted of 105 galaxies .
the maps - pp subsample was cross - identified with the nasa / ipac extragalactic database ( ned ) in order to obtain previous radio flux measurements and redshifts .
we also examined the field around each subsample galaxy and eliminated those in crowded fields , which led to a final maps - pp subsample of 96 galaxies ( which will hereafter be referred to as the arecibo sample ) , listed in table [ arecibo_subset ] .
we obtained 21 cm line spectra with the 305 m arecibo telescope of the national astronomy and ionosphere center over 14 nights between august 6 and august 20 , 1998 .
the new gregorian feed was used with the narrow l band receiver using a 25 mhz bandpass centered on 1394 mhz ( 1024 channels ) .
one observation was performed using a 50 mhz bandpass centered at 1400.5 mhz .
the beamsize of the 305 m arecibo dish is approximately 3.3 fwhm . for each of our arecibo sample galaxies , we made two sets of on - off observations , one 90 to the east of the central position along the major - axis , and a corresponding observation to the west of the galaxy center .
typically , 5 minute integrations were used for each observation , although some galaxies were re - observed to allow better measurement of their weak flux and others known to be bright in were observed with shorter integrations .
preliminary data reduction was performed using analyz at the arecibo facility . for each observation
, the two polarizations were averaged together .
for each galaxy we then archived both the sum of the east and west ( @xmath6 ) spectra and the difference ( in the sense east minus west ) .
it is the difference ( @xmath7 ) spectra that can be used to determine the spin vector , by allowing us to determine which side of the major - axis is moving toward us relative to the galaxy center . of the 96 galaxies in the original sample ,
6 were not observed , 16 were not detected in , 3 suffered from strong radio frequency interference ( rfi ) , and one suffered from a distorted baseline .
we therefore had a total of 70 galaxies for which there were good detections .
subsequent data reduction was performed on the 70 galaxies for which good @xmath6 detections existed .
the spectra were doppler corrected and the fluxes corrected for gain differences with zenith angle and changes in system temperature . a visual estimate of each galaxy s redshift was made and then radio frequency interference ( rfi ) within @xmath8 of the line was removed from the spectra .
rfi removal was performed interactively and the rfi was replaced with a linear interpolation between the two endpoints of the spectra .
noise was added to the linear interpolation , using the surrounding spectral channels to determine the noise level .
both the @xmath6 and @xmath7 spectra were baseline corrected using a linear fit to non- line channels within @xmath9 .
we determined the line properties of the galaxy using the @xmath6 spectra .
all velocities follow the optical convention , @xmath10 , and are adjusted to be in the heliocentric frame .
the flux - weighted mean velocity , @xmath11 , of the galaxy as well as the line flux is computed .
the line width used the mean of the line widths at a threshold of 50% of the boxcar equivalent flux and at a threshold of 20% the maximum flux determined by using a outward searching algorithm ( @xcite ) .
the reported line width has been corrected for noise and channel width using the method outlined in lavezzi and dickey ( 1997 ) .
the direction of the galaxy s spin vector was determined by taking the first moment of the @xmath7 spectra , @xmath12 , where @xmath13 ^ 2 dv } , \ ] ] where @xmath14 and @xmath15 are the minimum and maximum velocity of the line respectively , @xmath11 is the flux - weighted mean velocity of the galaxy , and @xmath16 and @xmath17 are the fluxes of the @xmath6 and @xmath7 spectra respectively .
negative @xmath12 implies that the eastern side of the galaxy is approaching us relative to the galaxy center , meaning the galaxy s @xmath0 points northward .
positive @xmath12 implies @xmath0 points to the south .
the uncertainty in @xmath12 due to bad baseline and spectral noise was measured using two variants of the normal first moment . to determine the effect of spectral noise on the first moment
, we computed @xmath18 , where we measure the first moment of the flux outside the line by conserving @xmath19 , but offset the @xmath11 , @xmath14 , and @xmath15 in equation [ ewmom ] to lie outside the line ( see figure [ mudef ] ) .
this gave us a measure of the contribution of spectral noise ( presumably similar outside the line as inside ) to the value of @xmath20 .
to determine the effect of uncertainty in the baseline fit to the first moment determination , we also computed @xmath21 , where we find the 1st moment about @xmath11 of the flux outside the line .
we then scaled this by @xmath22 to determine the amount of @xmath12 uncertainty due to uncertainty in the baseline fit . both @xmath18 and @xmath21
are illustrated in figure [ mudef ] .
@xmath18 and @xmath21 measurements suggest that galaxies with @xmath23 should be considered to have undetermined spin ( see figure [ fig3 - 1 ] ) . to confirm
that the @xmath7 spectra are the result of gas being observed on both sides of the major axis , we also computed the cross - correlation , @xmath24 , of the @xmath6 and @xmath7 spectra , @xmath25 ^ 2 dv } , \ ] ] since we would expect that the @xmath6 and @xmath7 spectra would be orthogonal in those cases where the flux is from both the eastern and western positions .
we have empirically found that if @xmath26 the @xmath7 flux was likely to be entirely from only one position and thus the spin measurement should be considered undetermined .
it should be noted that this process will not eliminate observations of galaxies with an asymmetric distribution if there is significant flux in both the eastern and western positions .
such a asymmetric distribution would affect the mean velocity , @xmath11 , and thus may affect the amplitude of @xmath12 , but it should not change the sign of @xmath12 , which is the observable we use later .
the final dataset had 54 galaxies with well determined spin vectors out of the 70 galaxies with good detections ( see table [ hi_data ] ) , 16 galaxies having been rejected from the sample due to either large @xmath24 or small @xmath12 .
for these galaxies , we computed @xmath27 which is the projection of @xmath0 on the plane of the sky .
since the arecibo sample is chosen to be nearly edge - on , @xmath28 is essentially a complete description of @xmath0 , allowing simple one - dimensional statistical analysis to be used for what is normally a three - dimensional problem .
identification of anisotropies in the observed @xmath28 and @xmath29 distributions was initially done by using the kuiper v statistic , which is a two - sided variant of the kolmogorov - smirnov ( k s ) d statistic ( @xcite ) .
we use the kuiper v statistic because the k s d statistic can systematically underestimate the significance of differences between the observations and the models , especially if the differences are near the ends of the distribution ( @xcite ) . for this test ,
we compare the cumulative distributions of a variable , @xmath30 ( such as @xmath28 , @xmath31 , etc . ) , in the observed sample , @xmath32 , with that for a model of 100000 randomly - oriented galaxies , @xmath33 .
the kuiper statistic , @xmath34 , is then defined as @xmath35 + max [ s_m(x)-s(x ) ] , \ ] ] the sum of the absolute values of the maximum positive ( @xmath36 ) and negative ( @xmath37 ) differences between @xmath32 and @xmath33 .. it does nt distinguish between differences above or below the @xmath33 the curve . ]
@xmath34 is essentially a measure of the difference between two distributions ( see figure [ vdef ] ) .
if the number of degrees of freedom is known a priori , a simple functional form exists for the probability , @xmath38 , that the two samples whose cumulative distributions differ by @xmath34 were drawn from the same parent distribution ( see press 1992 , for example ) . therefore ,
if we are comparing the distribution of @xmath30 for the observed sample to that of a modeled , randomly - oriented sample , we have a way of estimating the probability that the observed sample is drawn from an isotropic distribution . in this study
, we considered a distribution s anisotropy significant if the probability , @xmath38 , that the arecibo sample could have been drawn from the randomly - oriented sample was less than 5% . in those cases where
the number of degrees of freedom is not well determined a priori , we used monte carlo comparisons of the observations with 1000 model samples of equal size .
this was necessary in order to avoid overestimating the significance of an observed anisotropy .
we model a randomly oriented distribution of galaxies by taking the observed sample , randomly reassigning the observed @xmath24 and @xmath20 values to various galaxies ( @xmath20 is determined by randomly reversing the sign of @xmath39 ) , and then randomly generating the major - axis position angle , @xmath29 .
this model kept the spatial distribution of the original sample and the observational selection effects while otherwise being a completely randomly oriented model .
comparison of the real distribution of a variable versus its distribution in the 1000 monte - carlo samples is used to determine the significance of an anisotropy in some of the more complicated distributions discussed in section [ hiresults ] .
as a followup to the work done in paper i , we initially examined some of the distributions similar in nature to the ones investigated that study .
we divided the entire maps - pp and arecibo samples into 3 subsets each : the high density subset , the low density subset , and the complete sample .
the high and low density subsets were created using surface density estimates , @xmath40 , from the maps - pp catalog to compute the median surface density .
the high and low density subsets include all galaxies with @xmath40 greater than and less than this median value , respectively . for the maps - pp subsets we tested the @xmath29-based distributions , whereas for the arecibo subsets
, we tested the @xmath28-based distributions . examinations of the @xmath28 and @xmath29 distributions show no significant anisotropy in any of the arecibo or maps - pp subsets .
similar results were seen for distributions of @xmath28 and @xmath29 relative to other critical angles including the following : * @xmath41 and @xmath42 : the difference of @xmath28 and @xmath29 , respectively , between nearest neighbor galaxies in that sample . note that @xmath42 is used in the arecibo sample only to separate the significance of any @xmath41 alignments from any @xmath42 alignments .
* @xmath43 : the difference of @xmath28 from the geodesic to the nearest neighbor galaxy .
* @xmath44 : the difference of @xmath28 from angle of the pisces - perseus supercluster ridgeline at its nearest point ( as determined in paper i ) . *
@xmath45 : the difference of @xmath28 from the galaxy concentration position angle built using a percolation length of @xmath46 arcminutes ( galaxy concentrations groupings of galaxies identified using a 2 dimensional friends - of - friends algorithm ( redshift is ignored ) , see paper i for details ) . * @xmath47 : the difference of @xmath28 from the radial line to the center of the galaxy concentration built using a percolation length of @xmath46 arcminutes .
these results , shown in table [ lvectests ] , support the observations in paper i in that no simple @xmath29 or @xmath28 alignments appear to be present .
examination of the @xmath45 distribution does not support the tentative anti - alignments seen in paper i. we looked for twisting of @xmath44 versus distance from the pps ridgeline , and could not corroborate this signal seen in the @xmath48 distribution of the maps - pp in paper i. we note that the arecibo sample is considerably smaller that the maps - pp , so we can not rule out the trends seen in paper i , but we simply can not support them .
an initial visual inspection of the plot of the distribution of @xmath28 on the sky ( figure [ lvecmap ] ) appears to show some @xmath28 alignments .
specifically , in many cases if you pick a galaxy at random and then compare its @xmath28 with that of nearby galaxies , the difference is often less than 90 .
it appeared to the authors that there was a visual impression of the pps being divided up into `` spin vector domains , '' regions with preferred @xmath0 orientations .
because visual impressions are subjective , we devised tests to look for possible spin vector domains as well as looking for the alignments of the sort reported in paper i for the galaxy major - axes .
we attempted to confirm visual impression of @xmath0 domains seen in figure [ lvecmap ] by examining the orientations of several nearest neighbors , instead of just the nearest neighbor . to this end , we computed the @xmath49 distribution , which is the summed distribution of @xmath31 ( respectively ) for the n closest galaxies within 3 of each galaxy
. if @xmath0 domains exist , the @xmath49 distribution should be peaked toward the lower values of @xmath49 . because the @xmath49 distribution about one galaxy is not independent of the distribution about that galaxy s nearest neighbors ,
the number of degrees of freedom is uncertain _ a priori_. this means that the standard function to determine the probability , @xmath38 , of two distributions being identical does nt work .
instead , we gauge @xmath38 by generating 1000 monte carlo samples and computing the kuiper v statistic of their @xmath49 distributions . by comparing the value of v for the observed sample with the distribution of v in the 1000 monte carlo samples , we have an estimate of the likelihood that a greater value of v is obtained , @xmath38 .
we therefore use @xmath50 in leu of the @xmath38 used in cases where we know the number of degrees of freedom .
we examined the @xmath49 distributions for the n closest galaxies of the arecibo samples , for n ranging from 3 to 10 .
these samples show no significant anisotropy when compared to monte carlo generated datasets , indicating that the visual impression of @xmath0 domains is either incorrect , or the @xmath0 domains are too weakly aligned to confirm with this test . because in paper i only a simple nearest neighbor test was performed
, we also examined the @xmath51 distribution for the maps - pp samples , in order to see if @xmath0 domains might be visible in the larger maps - pp dataset .
we found that for n ranging from 3 to 10 , the @xmath51 distributions showed no evidence of significant anisotropies .
this appears to indicate that it is unlikely that @xmath0 domains exist in the pisces - perseus supercluster . in order to quantify the largest anisotropic signature that could remain `` hidden '' from our statistical techniques
, we performed a simple simulation .
we generated samples drawn from random ` sinusoidal ' distributions described by the probability distributions @xmath52 d\theta , \mbox{where}~\theta \in \left[0,\lambda\right],\ ] ] and @xmath53 d\theta , \mbox{where}~\theta \in \left[0,\frac{\lambda}{2}\right],\ ] ] where @xmath54 is the amplitude of the sinusoidal component of the probability in percent .
in these two distributions , @xmath55 represents either the expected @xmath29 or @xmath28 distributions in the cases of large - scale alignments ( equation [ cos_prob_2 ] ) , or the @xmath56 and @xmath31 distributions in the cases of alignments ( equation [ cos_prob_2 ] ) , anti - alignments ( equation [ cos_prob_2 ] ) or both ( equation [ cos_prob ] ) between nearby galaxies . using these two distributions
, we can generate samples with a predetermined amplitude , @xmath54 , for the alignments present and then compute the value of @xmath38 , the probability of the sample having been drawn from a random sample . by repeatedly doing this , we can determine the distribution of @xmath38 for a given @xmath54 and sample size . for samples of 30 , 54 , 100 , 615 , and 1230 galaxies ( the sizes of our subsets as noted in table [ lvectests ] ) , we computed the @xmath38 distribution for 100 generated @xmath55 samples with amplitudes , @xmath54 , ranging from 0% to 100% in steps of two percent ( see figure [ ampvprobs ] ) .
we then examined at which point 95% of the @xmath38 distribution dropped below 0.05 , our criterion for calling a distribution significantly anisotropic .
this gave us an estimate of the largest amplitude sinusoidal anisotropy that could have been missed , which we call @xmath57 .
@xmath57 is therefore the smallest amplitude of a sinusoidal anisotropy for which there is a 95% chance of detection given the criteria of @xmath58 . for our arecibo sample
, we find that with 54 galaxies @xmath59 , therefore we can only eliminate global spin vector alignments with sinusoidal amplitudes greater than 75% .
this sample does not place very strong limits on level of any spin vector alignments present . with the 1230 galaxies in the maps - pp catalog , we find @xmath60 .
therefore we can eliminate the possibility of galaxy major - axis alignments at amplitudes greater than 15% .
major - axis alignments place very weak limits on the level of spin vector alignments due to the fact that the orientation of the major - axis of the galaxy , with no additional information , only restricts the spin vector to a plane .
however , if there is a spin vector alignment , it must be reflected in the major - axis distribution of the edge - on galaxies .
we find that in a subsample of 729 maps - pp galaxies restricted to @xmath61 , there is no significant major - axis anisotropy of any sort . for a sample of 729 galaxies ,
we find @xmath62 , therefore , we can confidently state that there are no spin vector alignments with sinusoidal amplitude greater than 20% ( within the uncertainty due to the two - fold degeneracy in mapping major - axis position angle to spin vector )
. we would like to have computed @xmath57 for the spin vector domain tests in order to gauge their sensitivity but it was computationally too expensive .
we have constructed the only catalog of well determined spin vectors for galaxies in the pisces - perseus supercluster .
our study is the first radio study that explicitly looks at the spin vector distribution of galaxies in a supercluster and was optimized toward that end .
we developed a simple technique for obtaining spin vector determinations and accessing the level of uncertainty in the spin vector determinations due to both spectral noise and uncertainty in fitting the continuum .
we were intentionally rather conservative in our data selection criteria , possibly rejecting several well measured spin vectors .
there are several problems currently hampering the determination of the angular momentum distribution of galaxies relative to each other and to the surrounding large - scale structure .
one major problem is that we do not have a very clear understanding of the internal extinction in galaxies and its effect on the appearance of the galaxy with changing inclination .
therefore , it is very difficult to accurately determine the inclination of a galaxy based solely on its ellipticity and position angle .
this also makes it more difficult to construct a proper volume - limited sample for a large - scale angular momentum study .
one could obtain redshifts for all the galaxies in a diameter - limited or magnitude - limited galaxy catalog and select a volume - limited subsample , but without a clear understanding of internal extinction , we can not correct magnitudes and diameters for inclination .
we compensated for these uncertainties of the effects of internal galaxy extinction by restricting our sample to highly edge - on galaxies .
this had the added benefit of making the spectra of the galaxies as broad as possible , and thus making it easier to determine the @xmath0 orientation .
we note that this restriction to edge - ons could make reduction of alignments relative to large - scale structure difficult , since we would be restricting analysis to galaxies with @xmath0 in the plane of the sky .
however , in this study , the edge - on orientation of the pisces - perseus supercluster means our sample galaxies @xmath0 lie in the plane perpendicular to the supercluster plane , which is advantageous for reducing the complexity of the analysis .
this does reduce our sensitivity to any @xmath0 alignments that lie outside the plane of the sky .
for example , if galaxies @xmath0 are preferentially oriented in a given direction within the plane of the pisces - perseus supercluster ( e.g. toward a cluster in the supercluster plane ) rather than simply being restricted to that plane , we may not detect such an alignment in our sample , since we restrict @xmath0 of sample galaxies to the plane perpendicular to the supercluster plane .
it would be interesting to perform similar observations of a `` face - on '' version of pisces - perseus , where we would then be restricting @xmath0 to the supercluster plane and possibly investigating a new class of @xmath0 alignments .
the technique we outline for obtaining spin vector measurements could be applied to quickly obtain @xmath0 measurements for many galaxies in superclusters other than pisces - perseus .
it is also notable that this technique could be transferred to multi - fiber spectroscopy . by assigning two fibers to each galaxy
, one could simultaneously determine the @xmath0 directions of many galaxies much more quickly than a comparable line slit spectrograph observations .
no rotation curve information would be available , but it would allow quick collection of a large sample of well determined galaxy @xmath0 . our examination of the @xmath0 distribution of galaxies in pisces - perseus provides no support for any form of anisotropic @xmath0 distribution .
we do not provide confirmation of the possible @xmath0 alignments noted in paper i for the major - axis distributions of pisces - perseus galaxies . given the relatively small size of the arecibo sample , rather large anisotropies in the spin vector distribution of the arecibo sample
( see section [ limits ] ) could remain undetected with our technique .
we do note that by using a sample of 729 nearly edge - on galaxies from the original maps - pp catalog , we feel we can restrict the sinusoidal amplitude of any spin vector anisotropy present to be less than approximately 20% the background ` random ' distribution , at least in the plane perpendicular to the pisces - perseus supercluster ridge .
it is unclear at what level galaxy @xmath0 alignments might be expected as no recent simulations have been designed with the goal of estimating galaxy alignments .
we expect that if galaxy alignments are produced by large - scale structure formation , the alignments would be strongest in areas of low density , where the relative scarcity of subsequent galaxy - galaxy interactions suggests the initial @xmath0 distribution would be better preserved .
however , as noted in the introduction , galaxy alignments can arise from a variety of evolutionary processes , in both high and low density environments
. it would be interesting if in modern computer simulations of galaxy evolution , the angular momentum of the resulting galaxies was investigated for @xmath0 alignments and predictions as to the amplitude ( and type ) of any anisotropies in the @xmath0 distribution were made .
as we showed in section [ limits ] , sample sizes need to be large ( on the order of at least 500 galaxies ) in order to unambiguously detect weak alignments .
there are two paths toward increasing the sample size .
we could examine a denser cluster with a greater number of targets satisfying our edge - on criteria such as the coma cluster .
it would be interesting to investigate the possibility of tidally induced galaxy alignments in denser environments as predicted by ciotti and dutta ( 1994 ) and ciotti and giampieri ( 1998 ) .
the only previous study looking for galaxy alignments in coma was plagued by stretched imaging ( @xcite ) , so alignment results for this cluster are still unclear .
our other option for increasing sample size is to develop a better understanding of the internal extinction in galaxies so that we could use galaxies of all inclinations .
the first author is currently investigating using image parameters of a large number of galaxies obtained using the aps database in order to better determine the internal extinction properties of galaxies .
we would like to thank telescope operators miguel boggiano , willie portalatin , pedro torres , and norberto despiau for their good humor and help with observing ( and especially norberto for his `` lucky coffee '' ) .
jec would like to thank chris salter , tapasi ghosh , jo ann etter , and phil perillat for helping make his first radio observing experience excellent , both professionally and personally .
travel was sponsored by the national astronomy and ionosphere center ( naic ) and the university of minnesota graduate school .
this research has made use of the aps catalog of the poss i , which is supported by the national aeronautics and space administration and the university of minnesota .
the aps databases can be accessed at _ http://aps.umn.edu/ _ on the world wide web .
some data reduction was performed at the laboratory for computational science and engineering ( lcse ) at the university of minnesota .
information about the lcse can be found online at _
llccccccc o_778_873376 & ugc 11993 & 22:18:25.5 & 34:58:14.8 & 15.67 & 1.49 & 61.7 & 0.89 & 5.60 o_778_731211 & 22208 + 3548 & 22:20:49.5 & 35:48:04.8 & 14.66 & 1.42 & 52.0 & 0.25 & 25.35 o_778_700353 & 22233 + 3556 & 22:23:19.2 & 35:55:31.0 & 15.32 & 1.52 & 57.4 & 0.58 & 10.40 o_778_847676 & none & 22:24:45.4 & 35:10:40.3 & 15.41 & 1.39 & 44.0 & 0.27 & 11.23 o_778_849054 & none & 22:25:53.3 & 35:04:31.8 & 15.39 & 1.66 & 46.7 & 0.12 & 9.07 o_778_1040599 & ngc 7320b & 22:35:10.3 & 33:39:46.9 & 15.54 & 1.47 & 49.2 & 0.34 & 7.17 o_778_923367 & none & 22:39:36.3 & 34:39:30.7 & 14.03 & 1.39 & 94.8 & 1.12 & 6.39 o_778_755586 & none & 22:41:32.8 & 35:38:55.2 & 15.39 & 1.52 & 52.1 & 2.19 & 3.20 o_1184_28270 & none & 22:50:24.2 & 33:07:53.6 & 16.26 & 1.64 & 44.3 & 0.86 & 3.11 o_1184_66567 & 22508 + 3230 & 22:50:47.0 & 32:29:50.0 & 14.99 & 1.44 & 49.5 & 0.32 & 11.28 o_1184_128370 & ugc 12231 & 22:51:11.0 & 31:21:11.7 & 14.84 & -0.35 & 62.4 & 0.84 & 10.07 o_1184_81567 & none & 22:55:40.5 & 32:11:55.6 & 15.90 & 1.34 & 49.2 & 0.12 & 6.26 o_1184_196807 & ugc 12320 & 22:59:40.4 & 30:29:42.0 & 15.83 & 1.45 & 66.8 & 1.50 & 6.00 o_1184_121956 & ugc 12362 & 23:03:51.4 & 31:36:49.0 & 14.44 & 1.00 & 67.4 & 0.17 & 2.27 o_1184_275037 & mcg + 05 - 54 - 039 & 23:07:04.4 & 29:12:43.0 & 15.38 & 1.16 & 53.0 & 1.43 & 2.35 o_1184_313727 & ugc 12427 & 23:10:57.6 & 28:40:55.9 & 14.94 & 0.19 & 49.4 & 0.97 & 5.94 o_1184_347214 & none & 23:12:38.9 & 28:00:36.0 & 15.60 & 0.46 & 52.7 & 0.99 & 6.24 o_1184_189398 & ugc 12458 & 23:12:43.0 & 30:40:27.0 & 16.22 & 1.51 & 59.9 & 0.52 & 6.82 o_843_144830 & none & 23:18:38.9 & 25:22:01.9 & 15.86 & 1.50 & 44.0 & 2.07 & 4.80 o_843_65466 & none & 23:18:59.9 & 26:12:15.6 & 15.95 & 0.96 & 44.9 & 1.31 & 4.95 o_914_404205 & ugc 12557 &
23:20:01.1 & 28:54:24.1 & 14.72 & 1.49 & 75.0 & 0.86 & 5.05 o_914_406137 & none & 23:23:49.0 & 28:58:48.0 & 16.70 & 1.46 & 44.8 & 1.60 & 5.66 o_914_344933 & ugc 12625 & 23:26:38.5 & 29:29:58.3 & 14.73 & 1.57 & 84.2 & 1.72 & 4.36 o_914_409641 & ugc 12644 & 23:28:58.8 & 28:54:47.7 & 14.86 & 0.56 & 59.1 & 0.95 & 3.69 o_914_371708 & none & 23:32:07.5 & 29:26:30.5 & 16.45 & 1.45 & 48.7 & 0.89 & 3.28 o_914_413061 & none & 23:32:17.4 & 29:02:11.2 & 16.07 & 1.47 & 45.5 & 0.59 & 3.05 o_914_511814 & none & 23:33:55.7 & 27:39:32.0 & 16.07 & 1.24 & 55.7 & 0.72 & 3.32 o_914_416774 & ugc 12730 & 23:38:03.3 & 28:54:39.3 & 14.29 & 1.56 & 98.6 & 0.32 & 2.84 o_914_286281 & ugc 12741 & 23:39:25.1 & 30:18:15.5 & 14.54 & 0.94 & 57.3 & 0.74 & 1.49 o_914_286479 & none & 23:39:59.4 & 30:19:12.6 & 15.63 & 0.68 & 44.0 & 0.72 & 1.50 o_914_514191 & none & 23:40:00.1 & 27:46:03.4 & 16.26 & 1.17 & 52.8 & 1.53 & 6.24 o_914_437214 & cgcg 498 - 006 & 23:42:35.0 & 28:47:11.1 & 15.55 & 1.38 & 46.1 & 0.93 & 4.75 o_1257_181710 & mcg + 05 - 01 - 003 & 23:52:48.7 & 30:06:24.8 & 15.11 & 1.51 & 54.2 & 0.27 & 3.38 o_1257_106449 & ugc 12845 & 23:53:09.1 & 31:37:15.8 & 15.11 & 1.99 & 52.4 & 1.25 & 2.20 o_1257_149828 & ugc 12864 & 23:54:50.8 & 30:42:49.4 & 14.49 & 1.11 & 85.0 & 0.34 & 3.85 o_1257_140283 & ngc 7799 & 23:56:46.8 & 31:00:22.1 & 15.96 & 1.19 & 66.9 & 0.64 & 2.90 o_1257_307025 & ugc 124 & 00:10:48.1 & 28:05:27.8 & 15.45 & 1.85 & 55.1 & 2.12 & 5.95 o_1257_224112 & ugc 147 & 00:13:07.5 & 29:23:21.2 & 15.39 & 1.83 & 45.6 & 0.79 & 7.56 o_1257_212633 & 00139 + 2939 & 00:13:52.2 & 29:38:45.5 & 15.37 & 0.78 & 67.6 & 0.53 & 18.31 o_1244_265500 & ugc 238 & 00:22:25.5 & 31:03:58.5 & 13.70 & 1.15 & 99.0 & 0.99 & 3.82 o_1244_376417 & 00254 + 3029 & 00:25:21.3 & 30:29:18.9 & 15.20 & 1.34 & 55.9 & 0.48 & 5.08 o_1244_340721 & ugc 279 & 00:25:36.7 & 30:31:37.5 & 13.90 & 1.09 & 92.3 & 0.51 & 4.61 o_1244_270335 & 00267 + 3106 & 00:26:45.2 & 31:06:17.2 & 15.57 & 1.07 & 83.8 & 1.12 & 6.73 o_1244_679996 & ugc 310 & 00:28:39.0 & 28:42:58.5 & 14.84 & 0.62 & 68.2 & 1.23 & 2.29 o_1244_578706 & none & 00:29:14.4 & 29:25:41.8 & 15.88 & 0.74 & 45.7 & 0.54 & 3.48 o_1244_241809 & 00313 + 3110 & 00:31:15.4 & 31:10:30.6 & 15.40 & 0.00 & 55.4 & 1.30 & 4.67 o_1244_767827 & ugc 345 & 00:32:09.7 & 28:08:02.5 & 15.22 & 0.09 & 76.1 & 1.75 & 1.76 o_1244_185805 & 00333 + 3136 & 00:33:15.9 & 31:35:41.4 & 14.23 & 1.24 & 68.0 & 1.75 & 3.54 o_1244_655424 & 00347 + 2853 & 00:34:41.6 & 28:52:27.1 & 14.85 & 0.84 & 98.4 & 0.95 & 7.96 o_1244_554203 & ugc 412 & 00:36:49.5 & 29:29:16.8 & 15.25 & 1.29 & 55.9 & 0.34 & 10.90 o_1244_595079 & ugc 449 & 00:39:41.7 & 29:25:26.2 & 14.96 & 0.93 & 63.5 & 0.40 & 7.98 o_1244_521380 & none & 00:40:48.3 & 29:46:45.6 & 15.32 & 0.75 & 51.2 & 0.04 & 12.24 o_1244_442897 & none & 00:41:53.2 & 30:04:51.5 & 16.59 & 0.50 & 47.1 & 0.42 & 11.42 o_1244_484644 & ugc 478 & 00:43:44.5 & 29:57:54.9 & 14.45 & 1.13 & 80.4 & 0.17 & 13.26 o_601_2598615 & ugc 501 & 00:46:21.1 & 27:56:44.1 & 15.47 & 1.55 & 91.9 & 1.95 & 2.21 o_601_927741 & ugc 511 & 00:47:27.3 & 31:27:32.8 & 15.46 & 0.96 & 69.8 & 1.47 & 5.16 o_601_2448395 & cgcg 501 - 024 & 00:48:09.5 & 28:25:40.3 & 15.69 & 1.19 & 44.6 & 1.50 & 4.85 o_601_1985337 & ugc 525 & 00:48:52.4 & 29:26:42.9 & 15.69 & 1.42 & 73.9 & 0.64 & 4.96 o_601_1986315 & 00494 + 2924 & 00:49:30.1 & 29:24:19.4 & 14.42 & 1.67 & 60.4 & 0.66 & 6.53 o_601_2601958 & none & 00:50:49.0 & 28:00:21.8 & 16.80 & 1.60 & 53.5 & 2.08 & 3.56 o_601_1152906 & ugc 557 & 00:52:03.8 & 31:05:40.0 & 14.56 & 0.73 & 59.5 & 0.76 & 20.54 o_601_2454111 & ugc 554 & 00:52:04.6 & 28:26:44.9 & 15.22 & 1.79 & 50.9 & 1.71 & 5.65 o_601_2363374 & 00521 + 2835 & 00:52:06.5 & 28:35:46.1 & 14.78 & 1.79 & 58.7 & 1.57 & 4.48 o_601_1044227 & ugc 565 & 00:52:38.6 & 31:24:14.1 & 15.13 & 0.98 & 48.1 & 0.92 & 13.23 o_601_1267760 & cgcg 501 - 048 & 00:53:26.0 & 30:48:15.4 & 15.38 & 1.21 & 55.5 & 0.47 & 9.02 o_601_1050939 & ugc 598 & 00:55:06.2 & 31:12:52.1 & 14.19 & 1.44 & 71.3 & 0.49 & 6.20 o_601_1498038 & ugc 624 & 00:57:52.7 & 30:23:58.1 & 13.00 & 0.96 & 99.9 & 0.60 & 6.63 o_601_1063679 & ugc 633 & 00:58:37.0 & 31:14:23.6 & 14.16 & 0.95 & 90.7 & 0.01 & 6.20 o_601_863271 & none & 01:00:55.7 & 31:46:46.0 & 15.29 & 1.04 & 60.7 & 0.15 & 8.98 o_601_1294945 & 01011 + 3056 & 01:01:04.3 & 30:55:47.2 & 15.71 & 1.11 & 57.4 & 0.61 & 4.65 o_601_978154 & ugc 669 & 01:02:34.0 & 31:24:52.0 & 15.19 & 1.32 & 72.1 & 0.39 & 6.37 o_601_1197694 & ugc 673 & 01:03:24.6 & 31:08:18.2 & 15.15 & 0.81 & 57.1 & 0.72 & 4.65 o_601_657010 & ugc 679 & 01:04:17.9 & 32:07:21.1 & 16.17 & 0.63 & 51.3 & 0.10 & 22.35 o_601_992015 & cgcg 501 - 092 & 01:05:17.4 & 31:24:29.8 & 15.25 & 1.34 & 45.1 & 0.80 & 9.63 o_601_175820 & a82 - 91 & 01:05:18.4 & 33:11:11.4 & 14.49 & 1.25 & 57.8 & 0.82 & 17.03 o_601_264028 & ngc 407 & 01:07:49.8 & 32:51:38.8 & 14.09 & 1.44 & 96.5 & 0.34 & 18.31 o_601_359427 & none & 01:07:57.9 & 32:45:57.2 & 16.27 & 0.63 & 47.3 & 0.25 & 19.90 o_406_436424 & 01104 + 3443 & 01:10:24.4 & 34:42:01.4 & 16.52 & 0.66 & 49.0 & 1.94 & 2.14 o_406_502375 & ugc 809 & 01:13:04.0 & 33:32:50.3 & 14.84 & 0.74 & 74.2 & 0.68 & 2.66 o_1189_285025 & none & 01:19:39.9 & 33:46:42.9 & 15.99 & 0.83 & 45.8 & 0.69 & 8.12 o_1189_293769 & ngc 512 & 01:21:10.7 & 33:38:47.5 & 13.74 & 1.05 & 69.5 & 0.49 & 6.95 o_1189_251296 & 01287 + 3432 & 01:28:43.2 & 34:31:32.6 & 14.81 & 0.55 & 56.1 & 0.76 & 3.02 o_1189_224928 & ngc 634 & 01:35:25.4 & 35:06:38.8 & 13.70 & 1.13 & 84.6 & 0.57 & 3.98 o_1189_244311 & ugc 1166 & 01:35:42.0 & 34:44:18.4 & 14.17 & 1.23 & 71.1 & 0.39 & 3.38 o_1189_234487 & 01366 + 3455 & 01:36:39.7 & 34:54:18.6 & 15.20 & 0.61 & 45.6 & 0.25 & 3.32 o_1189_216659 & ngc 653 & 01:39:31.7 & 35:23:12.3 & 14.17 & 1.42 & 83.3 & 0.36 & 2.96 o_1189_199608 & 01446 + 3547 & 01:44:34.6 & 35:47:04.8 & 14.76 & 1.08 & 44.8 & 0.39 & 5.38 o_1225_610652 & none & 01:45:28.6 & 33:35:47.0 & 16.99 & 0.52 & 48.1 & 1.81 & 2.49 o_1225_389102 & ugc 1307 & 01:47:51.9 & 35:41:06.0 & 14.32 & 1.34 & 76.9 & 0.26 & 18.09 o_1225_412197 & ugc 1339 & 01:49:28.8 & 35:36:36.5 & 14.39 & 1.55 & 48.8 & 0.29 & 34.59 o_1225_369028 & ngc 714 & 01:50:33.0 & 35:58:31.7 & 14.06 & 1.62 & 77.5 & 0.20 & 42.61 o_1225_369332 & ugc 1363 & 01:50:58.4 & 35:59:02.1 & 14.71 & 1.49 & 70.0 & 0.29 & 47.93 o_1225_394977 & 01561 + 3549 & 01:56:09.1 & 35:49:14.2 & 14.93 & 1.48 & 79.2 & 0.12 & 9.32 o_1225_484388 & ugc 1569 & 02:01:55.5 & 34:54:35.3 & 15.47 & 1.46 & 47.3 & 0.53 & 2.50 o_1225_531870 & 02023 + 3434 & 02:02:19.7 & 34:33:02.9 & 15.01 & 1.15 & 51.8 & 0.76 & 2.67 o_1225_606718 & 02087 + 3349 & 02:08:40.7 & 33:48:35.6 & 14.64 & 1.13 & 63.9 & 1.58 & 2.54 llccccccr o_778_873376 & 11 & 5604 & 334 & 1.03(0.08 ) & n & @xmath63 & @xmath64 & o_778_731211 & 08 & & & & u & & & non - detection o_778_700353 & 18 & & & & u & & & non - detection o_778_847676 & 16 & & & & u & & & non - detection o_778_849054 & 13 & & & & u & & & non - detection o_778_1040599 & 07 & 6379 & 328 & 1.12(0.08 ) & u & @xmath65 & @xmath66 & o_778_923367 & 14 & 6576 & 547 & 2.85(0.16 ) & u & @xmath67 & @xmath68 & o_778_755586 & 12 & 6661 & 269 & 1.23(0.10 ) & u & @xmath69 & @xmath70 & o_1184_28270 & 12 + 17 & 6705 & 443 & 1.30(0.10 ) & s & @xmath71 & @xmath72 & o_1184_66567 & 13 & 6791 & 267 & 0.85(0.09 ) & u & @xmath73 & @xmath74 & o_1184_128370 & 06 & 3869 & 200 & 4.55(0.06 ) & u & @xmath75 & @xmath76 & o_1184_81567 & 09 & 6588 & 283 & 1.50(0.07 ) & s & @xmath77 & @xmath78 & o_1184_196807 & 06 & 6597 & 342 & 3.10(0.09 ) & s & @xmath79 & @xmath80 & o_1184_121956 & 07 & 6436 & 358 & 9.36(0.10 ) & s & @xmath81 & @xmath82 & o_1184_275037 & 06 & 3691 & 275 & 5.25(0.09 ) & n & @xmath83 & @xmath84 & o_1184_313727 & 06 & 3674 & 224 & 7.24(0.08 ) & u & @xmath85 & @xmath86 & o_1184_347214 & 14 & 5836 & 235 & 2.34(0.07 ) & u & @xmath87 & @xmath88 & o_1184_189398 & 09 & 6838 & 406 & 1.58(0.10 ) & s & @xmath89 & @xmath90 & o_843_144830 & 15 & & & & u & & & non - detection o_843_65466 & 16 & 5879 & 235 & 3.24(0.09 ) & u & @xmath91 & @xmath92 & o_914_404205 & 07 & 5903 & 555 & 5.24(0.10 ) & n & @xmath93 & @xmath94 & o_914_406137 & 13 + 17 & 5534 & 187 & 0.28(0.09 ) & u & @xmath95 & @xmath96 & o_914_344933 & 12 & 5225 & 435 & 1.95(0.09 ) & s & @xmath97 & @xmath98 & o_914_409641 & 11 & 5466 & 213 & 4.38(0.08 ) & u & @xmath99 & @xmath100 & o_914_371708 & 16 & & & & u & & & non - detection o_914_413061 & 15 & & & & u & & & non - detection o_914_511814 & 20 & 8802 & 372 & 4.56(0.09 ) & s & @xmath101 & @xmath102 & 50 mhz bandpass o_914_416774 & 14 & 5191 & 507 & 3.87(0.10 ) & n & @xmath103 & @xmath104 & o_914_286281 & 14 & 5208 & 395 & 6.34(0.09 ) & n & @xmath105 & @xmath106 & o_914_286479 & 17 & 5057 & 224 & 1.82(0.09 ) & u & @xmath107 & @xmath108 & o_914_514191 & 15 & 7134 & 261 & 3.10(0.07 ) & u & @xmath109 & @xmath110 & o_914_437214 & 15 & 6895 & 318 & 1.43(0.08 ) & n & @xmath111 & @xmath112 & o_1257_181710 & 16 & 5111 & 272 & 1.34(0.08 ) & u & @xmath113 & @xmath114 & o_1257_106449 & 07 & 4844 & 245 & 11.67(0.06 ) & s & @xmath115 & @xmath116 & o_1257_149828 & 12 & 4647 & 302 & 12.82(0.17 ) & n & @xmath117 & @xmath118 & o_1257_140283 & 13 & 4970 & 245 & 5.07(0.07 ) & s & @xmath119 & @xmath120 & o_1257_307025 & 17 & & & & u & & & non - detection o_1257_224112 & 18 & & & & u & & & non - detection o_1257_212633 & 13 & 4835 & 245 & 4.47(0.09 ) & n & @xmath121 & @xmath122 & o_1244_265500 & 15 & 6765 & 411 & 8.75(0.12 ) & n & @xmath123 & @xmath124 & o_1244_376417 & 15 & 6287 & 390 & 3.90(0.09 ) & u & @xmath125 & @xmath126 & o_1244_340721 & 11 & 6269 & 446 & 10.75(0.10 ) & n & @xmath127 & @xmath128 & o_1244_270335 & 14 & 6296 & 275 & 1.99(0.09 ) & n & @xmath129 & @xmath130 & o_1244_679996 & 06 & 4635 & 179 & 3.15(0.14 ) & u & @xmath131 & @xmath132 & o_1244_578706 & 17 & 6966 & 288 & 3.08(0.10 ) & u & @xmath133 & @xmath134 & o_1244_241809 & 12 & 4593 & 138 & 5.55(0.10 ) & u & @xmath135 & @xmath136 & o_1244_767827 & 06 & 4148 & 200 & 4.71(0.09 ) & s & @xmath137 & @xmath138 & o_1244_185805 & 18 & 6292 & 419 & 2.05(0.10 ) & n & @xmath139 & @xmath140 & o_1244_655424 & 14 & 5238 & 342 & 4.31(0.08 ) & n & @xmath141 & @xmath142 & o_1244_554203 & 16 & & & & u & & & non - detection o_1244_595079 & 13 & 5249 & 248 & 5.04(0.06 ) & n & @xmath143 & @xmath144 & o_1244_521380 & 20 & 4891 & 283 & 1.43(0.08 ) & u & @xmath145 & @xmath146 & o_1244_442897 & 20 & & & & u & & & strong rfi o_1244_484644 & 18 & 4901 & 459 & 1.34(0.09 ) & n & @xmath147 & @xmath148 & o_601_2598615 & 06 & 5059 & 382 & 4.71(0.19 ) & n & @xmath149 & @xmath150 & o_601_927741 & 13 & 4575 & 323 & 3.60(0.08 ) & n & @xmath83 & @xmath76 & o_601_2448395 & 17 & 4972 & 219 & 1.69(0.08 ) & u & @xmath151 & @xmath152 & o_601_1985337 & 07 & 4892 & 227 & 2.97(0.07 ) & n & @xmath153 & @xmath154 & o_601_1986315 & 18 & & & & u & & & non - detection o_601_2601958 & none & & & & u & & & o_601_1152906 & 11 & 6267 & 446 & 10.77(0.10 ) & n & @xmath127 & @xmath155 & o_601_2454111 & none & & & & u & & & o_601_2363374 & 18 & & & & u & & & baseline distorted o_601_1044227 & 17 & 5635 & 232 & 1.13(0.09 ) & u & @xmath156 & @xmath157 & o_601_1267760 & 18 & 4623 & 318 & 2.08(0.10 ) & n & @xmath158 & @xmath159 & o_601_1050939 & 18 & 5087 & 146 & 1.42(0.07 ) & n & @xmath160 & @xmath161 & o_601_1498038 & 12 & 4758 & 523 & 9.16(0.12 ) & n & @xmath162 & @xmath163 & o_601_1063679 & 14 & 5535 & 416 & 7.15(0.10 ) & n & @xmath164 & @xmath165 & o_601_863271 & none & & & & u & & & o_601_1294945 & 16 & 6218 & 267 & 2.85(0.08 ) & n & @xmath166 & @xmath167 & o_601_978154 & 20 & 5823 & 347 & 1.15(0.13 ) & u & @xmath168 & @xmath169 & o_601_1197694 & 16 & 6213 & 277 & 3.72(0.08 ) & u & @xmath170 & @xmath171 & o_601_657010 & 16 & 5069 & 219 & 2.51(0.09 ) & s & @xmath172 & @xmath173 & o_601_992015 & none & & & & u & & & o_601_175820 & 20 & & & & u & & & non - detection o_601_264028 & none & & & & u & & & o_601_359427 & 20 & 4643 & 267 & 1.22(0.09 ) & u & @xmath174 & @xmath175 & o_406_436424 & 20 & 4708 & 221 & 1.84(0.08 ) & s & @xmath176 & @xmath177 & o_406_502375 & 16 & 4160 & 299 & 3.45(0.10 ) & u & @xmath178 & @xmath179 & o_1189_285025 & 17 & 4851 & 277 & 1.00(0.11 ) & u & @xmath180 & @xmath181 & o_1189_293769 & 06 & 4839 & 534 & 5.36(0.10 ) & n & @xmath182 & @xmath183 & o_1189_251296 & 12 & 4127 & 213 & 3.28(0.07 ) & u & @xmath131 & @xmath184 & o_1189_224928 & 13 & 4884 & 497 & 6.83(0.10 ) & s & @xmath185 & @xmath186 & o_1189_244311 & 14 & & & & u & & & strong rfi o_1189_234487 & 15 & 5126 & 176 & 1.71(0.08 ) & u & @xmath187 & @xmath136 & o_1189_216659 & 16 & & & & u & & & strong rfi o_1189_199608 & 17 & 4786 & 272 & 1.11(0.08 ) & n & @xmath63 & @xmath188 & o_1225_610652 & 09 & 5728 & 221 & 1.11(0.07 ) & u & @xmath189 & @xmath190 & o_1225_389102 & 18 & & & & u & & & non - detection o_1225_412197 & 18 & & & & u & & & non - detection o_1225_369028 & none & & & & u & & & o_1225_369332 & 20 & & & & u & & & non - detection o_1225_394977 & 12 & 5406 & 366 & 1.75(0.08 ) & s & @xmath191 & @xmath192 & o_1225_484388 & 13 & & & & u & & & non - detection o_1225_531870 & 10 & 4334 & 288 & 1.14(0.09 ) & u & @xmath170 & @xmath193 & o_1225_606718 & 08 & 6128 & 435 & 3.33(0.10 ) & n & @xmath194 & @xmath195 & lrrrrrr @xmath196 & 54 & 23 & 30 & 1230 & 616 & 615 @xmath28 & 0.220 & 0.061 & 0.064 & & & @xmath29 & 0.449 & 0.448 & 0.807 & 0.388 & 0.270 & 0.341 @xmath41 & 0.096 & 0.462 & 0.120 & & & @xmath42 & 0.121 & 0.119 & 0.361 & 0.667 & 0.197 & 0.352 @xmath197 & & & & 0.606 & 0.730 & 0.680 @xmath43 & 0.384 & 0.894 & 0.374 & & & @xmath48 & & & & 0.799 & 0.126 & 0.900 @xmath44 & 0.257 & 0.142 & 0.113 & & & + + + @xmath31/@xmath198 & ( 000 ) & ( 000 ) & ( 000 ) & 0.651 ( 013 ) & 0.651 ( 013 ) & ( 000 ) @xmath31/@xmath199 & 0.279 ( 001 ) & 0.279 ( 001 ) & ( 000 ) & 0.700 ( 093 ) & 0.700 ( 093 ) & ( 000 ) @xmath31/@xmath200 & 0.699 ( 009 ) & 0.884 ( 008 ) & 0.279 ( 001 ) & 0.823 ( 224 ) & 0.846 ( 220 ) & 0.731 ( 004 ) @xmath31/@xmath201 & 0.489 ( 018 ) & 0.954 ( 014 ) & 0.108 ( 003 ) & 0.113 ( 431 ) & 0.062 ( 389 ) & 0.797 ( 043 ) @xmath31/@xmath202 & 0.086 ( 030 ) & 0.410 ( 021 ) & 0.252 ( 008 ) & 0.774 ( 581 ) & 0.727 ( 455 ) & 0.596 ( 127 ) @xmath31/@xmath203 & 0.046 ( 042 ) & 0.065 ( 023 ) & 0.160 ( 018 ) & 0.590 ( 722 ) & 0.881 ( 497 ) & 0.639 ( 226 ) @xmath31/@xmath204 & 0.124 ( 048 ) & 0.244 ( 023 ) & 0.624 ( 024 ) & 0.095 ( 817 ) & 0.113 ( 508 ) & 0.946 ( 310 ) @xmath31/@xmath205 & 0.214 ( 052 ) & 0.016 ( 023 ) & 0.815 ( 028 ) & 0.307 ( 902 ) & 0.380 ( 509 ) & 0.692 ( 394 ) @xmath31/@xmath206 & 0.494 ( 052 ) & 0.106 ( 023 ) & 0.917 ( 028 ) & 0.616 ( 948 ) & 0.742 ( 509 ) & 0.905 ( 440 ) @xmath31/@xmath207 & ( 000 ) & ( 000 ) & ( 000 ) & 0.858 ( 013 ) & 0.858 ( 013 ) & ( 000 ) @xmath31/@xmath208 & 0.279 ( 001 ) & 0.279 ( 001 ) & ( 000 ) & 0.839 ( 093 ) & 0.839 ( 093 ) & ( 000 ) @xmath31/@xmath209 & 0.931 ( 009 ) & 0.683 ( 008 ) & 0.279 ( 001 ) & 0.741 ( 224 ) & 0.720 ( 220 ) & 0.232 ( 004 ) @xmath31/@xmath210 & 0.764 ( 018 ) & 0.365 ( 014 ) & 0.850 ( 003 ) & 0.779 ( 431 ) & 0.665 ( 389 ) & 0.862 ( 043 ) @xmath31/@xmath211 & 0.767 ( 030 ) & 0.523 ( 021 ) & 0.871 ( 008 ) & 0.346 ( 581 ) & 0.605 ( 455 ) & 0.765 ( 127 ) @xmath31/@xmath212 & 0.359 ( 042 ) & 0.473 ( 023 ) & 0.707 ( 018 ) & 0.040 ( 722 ) & 0.247 ( 497 ) & 0.546 ( 226 ) @xmath31/@xmath213 & 0.917 ( 048 ) & 0.711 ( 023 ) & 0.475 ( 024 ) & 0.670 ( 817 ) & 0.264 ( 508 ) & 0.271 ( 310 ) @xmath31/@xmath214 & 0.281 ( 052 ) & 0.253 ( 023 ) & 0.428 ( 028 ) & 0.457 ( 902 ) & 0.849 ( 509 ) & 0.127 ( 394 ) @xmath31/@xmath215 & 0.148 ( 052 ) & 0.568 ( 023 ) & 0.039 ( 028 ) & 0.274 ( 948 ) & 0.196 ( 509 ) & 0.973 ( 440 ) binggeli , b. 1982 , , * 107 * , 338 byun , y.i . and freeman , k.c .
1995 , , * 448 * , 563 cabanela , j.e . and aldering , g.l .
1998 , , * 106 * , 1094 ( paper i ) ciotti , l. and dutta , s.n .
1994 , , * 270 * , 390 ciotti , l. and giampieri , g. 1998 , _ astro - ph/9801261 _ colberg , j.m . , white , s.d.m . ,
jenkins , a. , and pearce , f.r .
1998 , _ astro - ph/9711040 _ de vaucouleurs , g. , de vaucouleurs , a. , corwin , h. g. jr . ,
buta , r. j. , paturel , g. , fouqu , p. 1991 , _ third reference catalogue of bright galaxies _
( new york , springer - verlag ) ( rc3 ) doroshkevich , a. and shandarin , s. 1978 , , * 184 * , 643 djorgovski , s. 1987 , in _ nearly normal galaxies _ , ed .
faber , s.m .
( new york , springer - verlag ) , 227 flin , p. 1988
, , * 235 * , 857 giovanelli , r. and haynes , m.p . 1988 , in _ large - scale structures of the universe _ , eds .
audouze , j. , pelletan , m. , & szalay , a. ( dordrecht , kluwer academic publishers ) , 113 han , c. , gould , a. , and sackett , p.d .
1995 , , * 445 * , 46 helou , g. and salpeter , e. 1982 , , * 252 * , 75 helou , g. 1984 , , * 284 * , 471 hoffman , g.l . ,
lewis , b.m . ,
helou , g. , salpeter , e.e . , and williams , h.l .
1989 , , * 69 * , 95 huizinga , j. e. 1994 , ph.d .
thesis , groningen kashikawa , n. and okamura , s. 1992 , , * 44 * , 493 lavezzi , t.e . and dickey , j.m .
1997 , , * 114 * , 2437 nilson , p. 1974
, uppsala astron .
annals , no . 6 ( ugc )
press , w.h . ,
teukolsky , s.a . ,
vetterlin , w.t . , and flannery , b.p .
1992 , _ numerical recipes in c ( second edition ) _
( new york , cambridge university press ) sofue , y. 1992 , , * 44 * , l1 west , m.j .
1994 , , * 268 * , 79 white , s.d.m . 1984 , , * 286 * , 38 zeldovich , y.b . 1970 , , * 5 * , 84 | we use observations made with the upgraded arecibo 305 m telescope in august 1998 to obtain accurate spin vector determinations for 54 nearly edge - on galaxies in the minnesota automated plate scanner pisces - perseus survey ( maps - pp ) .
we introduce a simple observational technique of determining the sense of rotation for galaxies , even when their disks are not fully resolved .
we examined the spin vector distribution of these 54 galaxies for evidence of preferential galaxy alignments .
we use the kuiper statistic , a variant of the kolmogorov
smirnov statistic , to determine the significance of any anisotropies in the distribution of galaxy spin vectors .
the possibility of `` spin vector domains '' is also investigated .
we find no significant evidence of preferential galaxy alignments in this sample .
however , we show that the small sample size places weak limits on the level of galaxy alignments . |
Subsets and Splits
Data Mining Abstracts
This query searches for and retrieves 2000 documents containing the phrase 'data mining' in the abstract, providing limited but specific document samples.