article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
the de sitter ( ds ) and anti - de sitter ( ads ) space has a constant curvature with the maximal symmetry of a given dimension .
the ( a)ds has thus attracted attention in understanding a quantum nature of spacetime and been applied to different physics . the present accelerating universe with a cosmological constant is an asymptotically pure ds space and the early inflationary universe also underwent a quasi - exponential expansion .
the pure ds space has a cosmological horizon and emits the ds radiation with the gibbons - hawking temperature.@xcite the einstein equation for the ds space can be explained by the black hole thermodynamics.@xcite an intriguing feature of the ds radiation is the solitonic nature in the global coordinates in any odd - dimensional ds space.@xcite the discrimination of dimensions for ds radiation can be explained by the stokes phenomenon , in which a pair of instanton actions interferes destructively in odd dimensions and constructively in even dimensions.@xcite the stokes phenomenon can also be explained by the coherent destructive or constructive interference of the superadiabatic particle number.@xcite the one - loop effective action in an electromagnetic field and in a curved spacetime reveals the interplay between quantum electrodynamics ( qed ) and quantum nature of curved spacetime .
a strong electromagnetic field changes the quantum vacuum structure,@xcite and an electric field with / without a parallel magnetic field pulls out virtual particles from the dirac sea and creates pairs of particles and antiparticles known as the schwinger effect.@xcite the schwinger effect in ( a)ds space entails the pair production by the electric field and the ds radiation@xcite or the suppression by the negative curvature of ads space.@xcite the one - loop effective action and the schwinger effect has been comprehensively studied in ref . .
the @xmath0-wave of a scalar field in the nariari geometry of a rotating ds black hole is equivalent to a charge in a uniform electric field in @xmath1.@xcite interestingly , the near - horizon geometry @xmath2 of an extremal or near - extremal black hole deduces that the emission of charges from the black hole is related to the schwinger effect in @xmath3.@xcite the fermionic current due to the schwinger effect@xcite and the holographic schwinger effect@xcite have been studied in the ds space . the production of particles can be explained by the ( effective ) temperature .
the hawking radiation has the bose - einstein or fermi - dirac distribution with the hawking temperature and the ds radiation has a distribution , whose leading boltzmann factor can be given by the gibbons - hawking temperature .
the schwinger effect in a constant or pulsed electric field has an effective temperature given by the inverse of the period of the instanton action in the euclidean time.@xcite narnhofer , peter and thirring have shown that the effective temperature for an accelerating observer in ds space is the geometric mean of the unruh temperature and the gibbons - hawking temperature,@xcite which is extended to the ads space by deser and levin,@xcite as shown in fig .
recently , cai and kim have shown that the schwinger effect in ( a)ds space also has a thermal interpretation in terms of an effective temperature for the unruh effect for accelerating charge by the electric field and the curvature effect,@xcite as summarized in fig .
this implies that the schwinger effect from an extremal black hole , in which the hawking radiation is suppressed due to the zero temperature , may also have such a thermal interpretation and similarly for a near - extremal black hole.@xcite in this paper we elaborate the one - loop effective action of scalar qed in ref . and
also advance a method to find the density of states for the one - loop action and the schwinger effect .
in particular , the density of states is important to obtain the one - loop qed action density from the integrated one - loop action in the in - out formalism and the rate of pair production from the mean number by the bogoliubov transformation .
the one - loop effective action is equivalent to all the one - loop diagrams with even number of photons and gravitons , as shown in fig .
[ fig2 ] , whose summation is beyond the computational practicality .
however , in the in - out formalism by schwinger and dewitt , the one - loop action is derived from the scattering amplitude between the in- and out - vacua@xcite @xmath4 in fact , the propagator or green function of the charge in the electric field and in the ( a)ds space gives the scattering amplitude .
we study the one - loop effective action of a scalar with mass @xmath6 and charge @xmath7 in a uniform electric field in the planar coordinates of @xmath1 space @xmath8 and in @xmath3 space @xmath9 the one - loop action differs from quantum electrodynamic interaction by summing feynman diagrams .
the gauge potentials are chosen to have the minkowski space limit ( @xmath10 or @xmath11 ) and the scalar curvature is @xmath12 .
then the momentum mode @xmath13 in @xmath1 has the equation @xmath14 where the dimensionless parameters are @xmath15 on the other hand , the energy mode @xmath16 in the @xmath3 space is given by the equation @xmath17 where another set of dimensionless parameters @xmath18 in the @xmath1 space , the positive and negative solutions for the in - vacuum ( @xmath19 ) are given by the whittaker function as @xmath20 with @xmath21 and those for the out - vacuum ( @xmath22 ) are @xmath23 the solutions ( [ ds in - sol ] ) and ( [ ds out - sol ] ) are chosen to satisfy the quantization rule @xmath24 = i$ ] .
similarly , the solutions for the in - vacuum ( @xmath25 ) for the tunneling boundary condition are @xmath26 with @xmath27 and those for the out - vacuum ( @xmath28 ) are @xmath29 the solutions ( [ ads in - sol ] ) and ( [ ads out - sol ] ) are chosen to satisfy the quantization rule @xmath30 = i$ ] for the tunneling boundary condition.@xcite
using ( [ ds in - sol ] ) and ( [ ds out - sol ] ) , we recalculate the bogoliubov coefficients in the @xmath1 space@xcite @xmath31 and from eqs .
( [ ads in - sol ] ) and ( [ ads out - sol ] ) , we find the bogoliubov coefficients in @xmath3 space @xmath32 hence , the mean numbers of pairs by the schwinger effect are @xmath33 @xmath34 the binding nature of @xmath3 provides an effective mass @xmath35 and thus increases the critical strength @xmath36 while the separating nature of @xmath1 gives @xmath37 and thus lowers the critical strength .
note the breitenlohler - freedman bound @xmath38 for the stability of ads against schwinger emission .
recently , cai and kim have proposed a thermal interpretation of the schwinger effect in the @xmath5 space .
note that the mean number for the ds space is determined by the actions@xcite @xmath39 here the effective temperature and the associated temperature are @xmath40 where the gibbons - hawking temperature and the unruh temperature for the accelerating charge by the electric field are , respectively , @xmath41 on the other hand , the mean number for the ads space is determined by the actions@xcite @xmath42 where @xmath43 in the minkowski limit @xmath44 , the mean number reduces to @xmath45 , as expected .
the factor of two is intrinsic nature of the schwinger effect.@xcite
the density of states is necessary to derive the qed action density and the pair - production rate in the in - out formalism . the energy - momentum integral of the off - shell mean number gives the density of states , @xmath46 , for a uniform electric field in the minkowski spacetime.@xcite in the case of a uniform magnetic field
, the wave packet of landau states gives the density of states , @xmath47 .
we now propose another method to find the density of states in the in - out formalism .
for that purpose , we note that the field has the fourier integral @xmath48 using the asymptotic form @xmath49 and changing of variable @xmath50 , we rewrite eq .
( [ fourier ] ) as @xmath51 where @xmath52 is a function independent of @xmath53 , and @xmath54 the gaussian integral from the saddle point method gives @xmath55 in the tunneling picture , the decay of the vacuum due to the schwinger effect is determined by the magnitude square of the field @xmath56 and the momentum independent part @xmath57 determines the mean number of created pairs . therefore , the density of states is given by @xmath58 in the minkowski spacetime @xmath59 the density of states becomes @xmath60 .
finally , we obtain the pair - production rate per unit two - volume @xmath61 the density of state has been proposed for spinor qed in the ds space.@xcite
from the integrated one - loop action ( [ in - out scat ] ) follows the qed action density , which is given by the bogoliubov coefficient , which is in turn expressed by gamma functions with complex arguments @xmath62 to obtain the qed action density , one uses the gamma function in the proper - time integral @xmath63 , \label{gamma fn}\end{aligned}\ ] ] and then applies the schwinger subtraction scheme to regulate divergent terms via the renormalization of the vacuum energy and charge.@xcite remarkably , the consistence relation , @xmath64 , between the imaginary part of the effective action and the pair - production rate , is the consequence of the cauchy s residue theorem applied to the gamma function ( [ gamma fn ] ) . finally , we find the qed action density in the @xmath65 space @xmath66 , \nonumber\\ { \cal l}^{(1)}_{\rm ads } & = & \frac{k^2 \gamma_{\rm ads}}{2(2 \pi ) } \int_{0}^{\infty } \frac{ds}{s } e^ { -\gamma_{\rm ads}s } \cosh(\bar{\omega}_0 s ) \bigl(\frac{1}{\sin(s/2 ) } - \frac{2}{s } - \frac{s}{12 } \bigr ) . \label{(a)ds qed}\end{aligned}\ ] ]
the scalar qed action and schwinger effect has been recalculated in the @xmath65 space .
a new method has been advanced to find the density of states and to obtain the qed action density and the schwinger pair - production rate in the @xmath65 space , which is the main result .
the author thanks christian schubert for useful discussion on the worldline instantons in de sitter space and won kim for drawing the figure .
this work was supported by ibs ( institute for basic science ) under ibs - r012-d1 . | we review the one - loop effective action in scalar qed and the schwinger effect in a uniform electric field in a two - dimensional ( anti- ) de sitter space .
the schwinger effect has a thermal interpretation in terms of the effective temperature introduced by cai and kim .
we propose a method to find the density of states for the charged scalar and obtain the qed action density and the pair - production rate in the in - out formalism . |
coherent transport in discrete low - dimensional mesoscopic quantum systems has attracted much attentions for fundamental physics as well as applications due to persistent technique progress in nanofabrication @xcite . quantum interference phenomena and quantum coherent transport properties
are extensively investigated in various quantum devices @xcite , interesting effects including fano resonance @xcite , kondo effect @xcite , aharonov - bohm ( ab ) effect @xcite , have been theoretically predicted and experimentally verified in one - dimensional nanostructures system , such as photonic nanocrystals @xcite , nanowires @xcite , quantum dots array @xcite .
much effort has been devoted to investigate coherent electron transport properties of a various types of interferometers oriented at different geometries @xcite . among these ,
quantum dots embedded two - arm electron wave interferometers are mostly studied , in particular , the ab interferometer @xcite .
the ab interferometer is composed by a two - terminal structure with single or double quantum dots embedded .
the two - terminal configuration forms a close ring shape quantum system threading by magnetic flux that is tunable via externally applied magnetic field . studying transport properties of a quantum system
is beneficial for understanding quantum processes concerning energy transfer .
the applications include quantum state transfer , quantum control and manipulation in quantum information science .
the engineered and modulated optical systems mimic a great deal of quantum systems including atomic , molecular in condensed matter physics .
for example , the dynamic localization @xcite and bloch oscillation @xcite of electron under static external field are demonstrated in periodically modulated coupled optical waveguides .
the optical systems of coupled waveguides and coupled resonators are successfully employed to investigate parity - time symmetry @xcite . in recent years , quantum optical analogue is a fruitful platform for studying coherent mechanics in quantum realm due to the simplicity and flexibility of photons .
photons , as neutral particles , although proved not directly interact with magnetic field , the photonic analogy of ab effect is proposed @xcite and verified based on dynamic modulation of material permittivity @xcite , and based on photon - phonon interaction @xcite .
the key point is to realize nonreciprocal photon tunneling .
thereafter , the study of quantum phenomena from solids to their artificial electromagnetic counterparts in optics has become a hot topic in quantum physics , quantum information and quantum optics @xcite .
the synthetic artificial gauge field is introduced to simulate and investigate phenomena usually found in electronic systems @xcite , investigations include the quantum topology in photonic system @xcite and the photonic analogue of quantum ( anomalous ) hall effect in photonic crystals @xcite . here , we study light transport in the presence of synthetic magnetic flux in a ring array of coupled resonators . in the presence of artificial gauge field ,
photons mimic electrons in magnetic field , a coupled resonators system provides an optical platform for investigating quantum phenomena .
the ring array of coupled resonators enclosed synthetic magnetic field acts as a two - arm ab interferometer . as interference is sensitive to the relative phase between light waves propagating in its arms , thus light transmission probability is relevant to the connection positions of the input and output leads , and depends on the enclosed synthetic magnetic flux .
the interference and phase sensitive feature reveal the quantum nature of photons .
we employ the bethe ansatz method to study the light transport properties .
we focus on the influence of a synthetic gauge field on light transport in a ring shaped array with arbitrary number of coupled resonators .
the transmission zeros and the resonant transmissions are further discussed .
, @xmath0 coupled together via an antiresonant auxiliary resonator ( in blue ) . the optical paths ( green and red arrows )
imbalance induces a nonreciprocal hopping phase between two primary resonators .
( b ) the ring array of coupled resonators enclosed synthetic magnetic flux .
the synthetic magnetic flux is introduced through nonreciprocal hoppings between neighbor resonators based on ( a ) .
the input lead is connected to resonator @xmath1 , the output lead is connected to resonator @xmath2 .
the system acts as a two - arm ab ring interferometer .
the output light can be modulated by the enclosed synthetic magnetic flux @xmath3.,width=16 ] we begin with two coupled passive resonators @xcite , using coupled mode theory , the dynamics of field amplitudes inside the resonators give @xmath4where @xmath5 , @xmath6 are the field amplitude in the two resonators , @xmath7 is the resonant frequency , @xmath8 is the coupling strength .
the coupled resonators system is able to be described by a tight - binding hamiltonian , @xmath9 .
similarly , an @xmath10 coupled passive resonators system with resonant frequency @xmath11 and coupling strength @xmath8 is described by @xmath12 , where @xmath13 ( @xmath14 ) is the photon creation ( annihilation ) operator for the @xmath15-th resonator with periodical boundary condition that @xmath16 , the roundtrip length of each resonator is @xmath17 , where @xmath18 is the light wavelength , @xmath19 is an integer .
the synthetic magnetic flux is induced by nonreciprocal hopping phase introduced in coupled resonators as proposed by hafezi hafezi2014 .
for any two adjacent resonators , the hopping is induced through an auxiliary resonator ( fig .
[ fig1]a ) . when optical length of the auxiliary resonator is @xmath20 , being antiresonant with two primary resonators , the auxiliary resonator results in an effective coupling strength @xmath8 in between two primary resonators describing by the hamiltonian @xmath21 . in fig .
[ fig1]a , light in the resonators is shown in yellow .
the forward going path of light from resonator @xmath22 to @xmath0 is shown in green , the backward going path of light from resonator @xmath0 to @xmath22 is shown in red . considering the path length difference between the forward and backward direction is @xmath23 , the effective coupling acquires an additional optical path dependent hopping phase factor @xmath24 with @xmath25 .
the auxiliary resonator between two primary resonators induces a nonreciprocal coupling @xmath26 . for a ring configuration shown in fig .
[ fig1]b based on the building blocks shown in fig .
[ fig1]a , photons circling around the ring array accumulate an ab type phase factor @xmath27 ( @xmath28 ) in clockwise ( counterclockwise ) direction , the total phase is @xmath29 .
the effective ab type phase introduced in the ring array equivalently realizes a synthetic magnetic flux for photons .
the synthetic magnetic field felt by photons is gauge invariant and unaffected by the distribution of nonreciprocal phase as long as the total phase @xmath3 accumulated by photons in one round is unchanged .
the ring resonator supports clockwise and counterclockwise modes , here we assume clockwise mode ( yellow circles ) as shown in fig .
the situation for counterclockwise mode is similar , where photons feel an exactly opposite synthetic magnetic flux .
to be precise , we consider a ring array with @xmath10 coupled resonators threading by synthetic magnetic flux @xmath30 .
the system is schematically illustrated in fig .
[ fig1]b , each site ( black ball ) represents a ring resonator .
the input and output leads connect to the array of coupled resonators at two different resonators ( @xmath1 and @xmath2 ) .
the connection leads split the ring array into two components : the upper arm and the lower arm . the ring array of coupled resonators functions as a two - arm ab interferometer . the light can propagate from input resonator ( @xmath1 ) to output resonator ( @xmath2 ) either in the upper or the lower arm .
photons tunneling in the upper arm acquire a difference phase comparing with tunneling in the lower arm , meet at the output resonator and interferes with each other .
the light interference depends on the different phases photons acquired in the two propagating paths ( the upper and the lower ring arms ) .
therefore , the light output is affected by the synthetic magnetic flux , which alternates the relative phase difference between two optical paths . in this sense , the wave vector of light and the connection positions of the input and output leads also influence the light transmission . in the following ,
we investigate light transport properties in a ring array of coupled resonators in the framework of interference .
the hamiltonian of a two - arm ring interferometer under synthetic magnetic flux is written as @xmath31where @xmath32 ( @xmath33 ) is the creation ( annihilation ) operator of resonator @xmath15 in the upper ring arm . corresponding creation and annihilation operators for resonator @xmath15 of the lower ring arm
are denoted by @xmath34 and @xmath35 . in order to investigate the influence of synthetic gauge field in the system ,
the coupling strength @xmath8 is set unity , the resonators are the same in resonant frequency @xmath36 .
there are no disorders in the couplings and no detunings in the array of coupled resonators , the light transport is solely controlled and affected by the synthetic magnetic flux and the ring interferometer structure .
we consider the input lead is connected to the ring array at resonators @xmath37 and @xmath38 ( @xmath39 , which is actually resonator @xmath38 ) .
the threading synthetic magnetic flux in the ring array is @xmath40 , it acts globally and is invariant under local operator transformations .
we study the dynamical process by employing the bethe ansatz method .
equivalent descriptions of a ring array of coupled resonators is guaranteed under local operator transformation . for convenience ,
we take a local transformation @xmath41 with @xmath42 $ ] in the upper ring arm , and @xmath43 with @xmath44 $ ] in the lower ring arm .
the hamiltonian is expressed as @xmath45 which describes a one - dimensional coupled ring resonators system with effective nonreciprocal tunneling phase introduced only in between resonators @xmath46 and @xmath47 .
photons acquire an additional phase @xmath30 tunneling from resonator @xmath46 to @xmath47 , and acquire an additional phase @xmath48 when tunneling inversely from resonator @xmath47 to @xmath46 .
the supporting light frequency in the ring array of coupled resonators is in the range of @xmath49 $ ] .
as @xmath8 is set unity , we consider an incident light wave with frequency @xmath50 ( @xmath51 ) incoming in the left lead and outgoing in the right lead .
the ring array of coupled resonators splits the input wave at connection resonator @xmath1 , a portion of light reflects back to the input lead and others propagate forward in the ring arms ; the forward propagating lights meet at resonator @xmath2 , interfere with each other and form the output .
meanwhile , the interference at output resonator @xmath2 also results in a portion of light reflecting back toward input resonator @xmath1 , meeting at the input resonator @xmath1 , interfere and form reflection light in the input lead . in the light propagating process , the ring array of coupled resonators acts as a two - arm ab interferometer .
the interference depends on the phase accumulated by photons at each path , which depends on enclosed synthetic magnetic flux in the ring array of coupled resonators .
the magnetic flux affects periodically , the phase factor @xmath52 is unchanged for every @xmath53 period of @xmath30 , i.e. , @xmath54 with integer @xmath55 .
thus , the light transmission probability in the system satisfies @xmath56 . in fig .
[ fig2 ] , we plot the light transmission probabilities for several different configurations with @xmath57 $ ] .
moreover , for the ring array shown in fig .
[ fig1]b , the transmission probabilities are irrelevant to the direction of the magnetic flux .
the magnetic flux is opposite when we turn over the ring array , which indicates an exchanging of the upper and the lower arms of the two - arm ab ring interferometer .
this will not change the light interference , and the light transmission remains unchanged .
thus , we have @xmath58 , which implies the counterclockwise mode supporting by the coupled resonators system has the same propagating properties as the clockwise mode we are discussing . the ring array of coupled resonators is a hermitian system , the light transmission probabilities are the same for light inputting in either side , this implies a symmetric light transport @xmath59 . .
we schematically illustrate the ring array of coupled resonator in the inserts of ( a ) for no resonator embedded in two arms , ( b ) for one resonator embedded in one arm , and ( c ) for two resonators respectively embedded in two arms .
corresponding contours of @xmath60 are plotted in the left panels as functions of the enclosed synthetic magnetic flux and the light wave vector .
the middle panels are cuts through contours of @xmath60 at @xmath61 ( red ) and @xmath62 ( blue ) .
the right panels are cuts through contours of @xmath60 at @xmath63 ( red ) , @xmath64 ( blue ) , @xmath65 ( green).,width=604 ] the dynamics in the system is governed by the schrdinger equations .
the field amplitude in each resonator has the form @xmath66 . in a rotating frame at oscillating frequency @xmath36 ,
the stationary resonator modal amplitude @xmath67 inside the ring array of coupled resonators is set as @xmath68 @xmath69 where @xmath70 and @xmath71 represent the forward and backward propagating light waves with vector @xmath72 , respectively .
the stationary resonator modal amplitude in the leads resonators of the input and output leads are set as @xmath73 and @xmath74 .
the reflection and transmission coefficients are defined as @xmath75 and @xmath76 for wave vector @xmath72 . due to the interference , light waves in the upper and the low arms of the ring array contain forward transmitted ( green arrow in fig . [ fig1 ] ) and backward reflected ( red arrow in fig . [ fig1 ] ) components . in order to calculate the reflection and transmission coefficients
, we list the schrdinger equations for all resonators in the whole coupled resonators system .
notice that the schrdinger equations for resonator @xmath77 $ ] in the upper arm and resonator @xmath78 $ ] in the lower arm give a series of equations , @xmath79the footnote is @xmath80 .
these equations all yield the eigen energy @xmath81 , which implies the dispersion relation , @xmath82 .
the remaining schrdinger equations in the ring array together with the boundary conditions ( continuity of wave functions at the connection resonators ) determine the light propagating properties .
the boundary conditions mean that the modal amplitude @xmath83 , @xmath84 should be in accordance at resonator @xmath37 and @xmath38 ( @xmath39 ) .
the boundary conditions at resonator @xmath37 are related to the reflection coefficient , @xmath85similarly , the boundary conditions at resonator @xmath38 are related to the transmission coefficient , @xmath86the schrdinger equation at resonator @xmath87 in the lower ring arm gives @xmath88,\]]this yields another boundary condition at resonator @xmath39 , @xmath89the schrdinger equation for resonator @xmath37 gives@xmath90and for resonator @xmath38 gives @xmath91 + e^{-i\phi } [ x_{\beta } e^{ik(n_{\beta } -1)}+y_{\beta } e^{-ik(n_{\beta } -1 ) } ] = te^{-ik}. \label{eq:8}\]]from equations ( [ eq:3]-[eq:8 ] ) obtained from the bethe ansatz method , we can exactly calculate the reflection and transmission coefficients @xmath75 , @xmath76 . in the following ,
we discuss light transmissions in details , in particular , the influence of synthetic magnetic flux .
the situation of @xmath92 in the ring array induces a completely destructive interference , that input light is totally reflected with zero transmission .
this is an interference effect only relevant to the structure and the input wave vector .
the physical interpretation of destructive interference is as follows .
the forward and backward going waves interfere in the ring array , forming two standing waves in the two arms .
the standing waves are sinusoidal functions that exactly being the eigenstates of uniform tight - binding chains with @xmath93 and @xmath87 coupled resonators , and the interference in the ring array results in vanishing of wave functions at the connection resonators . in this situation
, two ring arms that function as transport channels are isolated , and we have @xmath94 , @xmath95 .
the input light will be fully reflected without any transmission , while the synthetic magnetic flux will not affect the light propagation . in other aspect , the gauge invariant field introduced in open chain
is trial . in the situation of @xmath92 , the ring array of coupled resonators
will act as a cage constraining photons of wave vector @xmath72 initially localized in the ring array .
photons travel in the ring arms without escaping to the leads . for the ring array and input wave vector satisfying @xmath96 and @xmath97 , the light wave vector satisfies @xmath98 , the phase acquired by photons in the path of upper ring arm from resonator @xmath37 to resonator @xmath99 is @xmath100 .
this indicates the wave functions at two connection resonators have the same amplitude and the same ( opposite ) phase , which depends on the length of upper ring arm , @xmath101 ( @xmath102 ) .
the same amplitude of light in the connection resonators implies a none zero transmission .
otherwise , zero transmission will reduce to the situation of @xmath92 .
in the situation @xmath96 and @xmath97 , none zero light transmission can also be seen from the transmission coefficient @xmath76 , which is calculated by the bethe ansatz method . from eq .
( [ eq:3]-[eq:8 ] ) , we have @xmath103+i\mathrm{sin}(n_{\beta } k)\mathrm{sin}k\ } } , \label{eq : t_caseii}\]]where we directly notice the transmission @xmath104 because @xmath97 , and the enclosed synthetic magnetic flux modulates the light transmission probability .
the resonant light transmission @xmath105 happens at @xmath106/\mathrm{sin}k$ ] .
through tuning the enclosed synthetic magnetic flux in the system , we can always have a resonant light transmission for an input light with wave vector @xmath107 , which corresponds to the input frequency @xmath36 , i.e. , the resonator on - resonance frequency .
the on - resonance input is an optimal choice for obtaining resonant light transmission through controlling synthetic magnetic flux . for input wave vector @xmath108 , although a synthetic magnetic flux modulates the light transmission , a resonant light transmission @xmath105 may be impossible by controlling the synthetic magnetic flux , e.g. @xmath109|>|\mathrm{sin}k|$ ] .
the case for @xmath110 and @xmath111 is similar , the transmission coefficient obtained is eq .
( eq : t_caseii ) with subscript @xmath112 and @xmath113 switched ( @xmath114 ) and the transmission coefficient @xmath76 substituted by @xmath115 ( @xmath116 ) .
now we turn to discuss a general situation . after simplification eq .
( [ eq:3]-[eq:8 ] ) under @xmath117 and @xmath97 , we obtain the transmission coefficient for a two - arm ring interferometer with input and output leads connected at arbitrary positions , the light transmission coefficient of the ring array of coupled resonators threading by synthetic magnetic flux @xmath30 is in form of @xmath118}{\mathrm{sin}(n_{\alpha } k)}+\frac{\mathrm{sin}[(n_{\beta } -1)k]}{% \mathrm{sin}(n_{\beta } k)}-e^{-ik}\right ) ^{2}}.\ ] ] where the transmission zeros @xmath119 happen only when @xmath120 at wave vector @xmath72 fulfills @xmath121 . in other words ,
the transmission zeros are at wave vector @xmath122 when there is no magnetic flux @xmath123 and @xmath124 when the magnetic flux is at maximum value @xmath125 , where @xmath126 is integer number .
the wave functions in the ring arms are sinusoidal functions , waves in the upper and the lower arms interfere destructively due to the extra phase factor @xmath127 in the presents of magnetic flux .
the destructive interference effectively opens the ring array at the connection resonator @xmath99 ( @xmath47 ) coupled to the output lead , and the ring arms act like dangling chains . by tuning magnetic flux with other system parameters
set , we obtain two reflection zeros at different magnetic flux in a @xmath128 period . the transmission probability @xmath129 for several simple configurations of ring array under magnetic flux are shown in fig .
the whole dynamical process is unitary .
the reflection is @xmath130 .
the ring array with no resonator embedded ( fig .
[ fig2]a ) , one resonator embedded ( fig .
[ fig2]b ) , and two resonators embedded ( fig . [ fig2]c ) embedded are shown .
moreover , we notice that @xmath125 results in destructive interference in the ab ring interferometer with equal length ( @xmath131 ) .
this is because the phase difference of photons reaching the output lead from the upper and from the lower arm is @xmath3 for system with @xmath131 , the magnetic flux induces additional phase factor @xmath132 .
therefore , half - integer magnetic flux quantum leads to a destructive interference in the two - arm ab ring interferometer with equal length . .
( a ) either @xmath133 or @xmath134 is odd .
( b ) both @xmath135 and @xmath134 are odd .
the blue line is for system with odd @xmath136 and green line is for system with even @xmath136 .
the transmission is zero for @xmath133 and @xmath134 being both even.,width=10 ] the light with wave vector @xmath137 propagates at the fast velocity @xmath138 , the dispersion relation yields the light frequency is on - resonance with the ring resonators @xmath139 .
varying the synthetic magnetic flux enclosed in the ring interferometer , the phase difference photons accumulated in the upper and the lower arms changes , which affects the interference in the ring array .
the transmissions probabilities for @xmath137 is plotted as a function of the magnetic flux @xmath30 in fig .
[ fig3 ] . for @xmath137 ,
the destructive interference induces full reflection without any transmission in case the numbers of coupled resonators embedded in the two interferometer arms ( @xmath38 , @xmath39 ) are both even . in the situation that the number of embedded resonators being even in one arm ( @xmath38 or @xmath39 is even ) , the transmission probability equals to @xmath140 , and varies in the range of @xmath141 to @xmath142 as the synthetic magnetic flux .
the resonant light transmission happens at @xmath143 , and reaches the minimum @xmath141 at @xmath144 ( fig .
[ fig3]a ) . for ring array with @xmath38 ,
@xmath39 being both odd , the synthetic magnetic flux induces asymmetric fano lineshapes in the transmission ( fig .
[ fig3]b ) .
in this situation , we have @xmath145 , and the transmission coefficient @xmath146 ( @xmath147 is the amplitude and @xmath148 is the phase ) is closely related to the system structure .
if @xmath149 is odd , we have @xmath150 being even , therefore @xmath151}$ ] equals to @xmath152 at half - integer magnetic flux quantum ( @xmath153 , @xmath55 ) , which leads to completely destructive interference ( blue line in fig .
[ fig3]b at @xmath154 ) .
the amplitude is @xmath155^{-1}$ ] .
the phase is @xmath156 for @xmath157 and @xmath158 for @xmath159 , respectively .
if @xmath160 is even , we have @xmath150 being odd , therefore @xmath151}$ ] equals to @xmath152 at integer magnetic flux quantum ( @xmath161 , @xmath55 ) , which leads to completely destructive interference ( green line in fig .
[ fig3]b at @xmath63 ) .
the amplitude is @xmath162^{-1}$ ] .
the phase is @xmath163 for @xmath164 and @xmath165 for @xmath166 , respectively . .
other parameters are ( a ) @xmath167 ; ( b ) @xmath168 ; ( c ) @xmath169 ; ( d ) @xmath170 ; ( e ) @xmath171 ; ( f ) @xmath172.,width=16 ] in the situation that there is no resonator embedded in the upper arm of the ring array ( @xmath173 ) .
the light transmission coefficient for @xmath39 coupled resonator embedded in the lower arm is @xmath174 \mathrm{sin}k } { [ \mathrm{sin(}n_{\beta } k)/\mathrm{sin}k+2\mathrm{cos}% \phi ] + 2ie^{-in_{\beta } k}\mathrm{sin}k}. \label{t_na1}\]]in eq .
( [ t_na1 ] ) , we notice that in the region @xmath175 and @xmath176 , there is no transmission zero ( @xmath177 ) except for @xmath178 , @xmath55 ( see fig . [ fig4 ] ) .
this indicates that only half - integer or integer magnetic flux quantum could lead to complete destructive interference without light transmission .
the resonant transmissions ( @xmath179 ) happen at @xmath180 .
for @xmath181 ( @xmath55 ) , the transmission amplitude is generally larger than @xmath30 being others and the resonant transmission ( @xmath105 ) happens at @xmath111 .
the transmission and reflection are shown in fig .
[ fig4]a and fig4b for the ring array with @xmath182 and @xmath183 , where the resonant transmissions are seen at each @xmath184 , @xmath185 for @xmath186 .
we also show @xmath187 in fig .
[ fig4]c and [ fig4]d for @xmath182 and @xmath183 , where the transmission zeros ( @xmath177 ) are at @xmath188 .
the @xmath189 is a trivial case , the transmission probability is equivalent with no magnetic flux case , we show @xmath190 in fig . [ fig4]e and [ fig4]f for @xmath182 and @xmath191 as comparison , where the transmission zeros ( @xmath192 ) are at @xmath193 .
the synthetic magnetic flux modulates the light transport , significantly varies @xmath72 for the resonant transmission as well as the transmission zeros
. this might be useful in the design of quantum devices such as photon filters or optical switches . to verify our analysis on transmission probability of the ring array of coupled resonators enclosed synthetic magnetic flux
, we simulate light transmission through calculating the dynamical evolution of a gaussian profile wave packet .
the profile of the light wave packet is @xmath194 which is centered at the resonator @xmath195 , the coefficient @xmath196 represents the renormalized light intensity of resonator @xmath15 , @xmath197 is the relative phase in each resonators of the incident light wave , and the parameter @xmath198 controls the width of the gaussian profile in the array of coupled resonators .
the gaussian wave is localized at its profile center @xmath195 and keeps shape - preserving in the evolution process , in particular , a wide spreading gaussian wave packet @xcite .
the interference of a plane wave with wave vector @xmath72 can be approximately simulated by a gaussian wave packet with velocity @xmath199 in a uniform chain with nearest neighbor coupling strength @xmath8 . to simulate the dynamics
, we consider a ring array of coupled resonators threading by synthetic magnetic flux is the central of the system , connected to the input and output leads . the gaussian wave with certain velocity is initially localized in the input lead far away from the central , it moves along the input lead towards the central , interferes in the ring array of coupled resonators and outgoes from the output lead .
the dynamical process and transmission probability are shown in the upper and the lower panels of fig .
[ fig5 ] for the gaussian profile waves in a ring array of coupled resonators system with @xmath200 , @xmath201 .
figure [ fig5]a is for wave with @xmath202 when synthetic magnetic flux is @xmath144 , and fig .
[ fig5]b is for wave with @xmath137 when synthetic magnetic flux is @xmath186 .
the numerically simulated transmission probabilities are in accordance with the analytically calculated transmission probabilities , being @xmath203 ( fig .
[ fig5]a ) and @xmath204 ( fig . [ fig5]b ) .
b are simulated in the upper panels .
the transmission probabilities as a function of time are shown in the lower panels .
the system parameters are ( a ) @xmath205 , @xmath206 , @xmath207 ; ( b ) @xmath205 , @xmath208 , @xmath209.,width=604 ]
in conclusion , the influence of synthetic magnetic flux on light transport is investigated for an array of coupled resonators in ring configuration threading by synthetic magnetic flux .
the synthetic magnetic flux is introduced in the couplings through optical paths imbalance .
we discuss the transport properties of a coupled resonators system with arbitrary number of resonators embedded and arbitrary positions that the leads connected . in particular , the transmission zeros and resonant transmissions are discussed in details .
the half - integer magnetic flux quantum results in completely destructive interference and transmission zeros for the ring array of coupled resonators with two arms at equal length , and light can be confined inside without leakage .
the ring array of coupled resonators enclosed synthetic magnetic flux can be used to control light transport .
tuning the enclosed magnetic flux , resonant transmission is available , and the sensitivity to synthetic magnetic flux might be useful in the design of photon filters or optical switches .
we perform numerical simulations , the results are in accordance with our theoretical analysis .
the array of coupled resonators is a discrete quantum system , it also reflects the quantum transport of electrons in magnetic field
. our results might be useful for studying coherent dynamics of a ring aharonov - bohm interferometer and may have potential applications in quantum state transfer , quantum control and manipulation in quantum information science .
this work was supported by national basic research program ( 973 program ) of china ( grant no .
2012cb921900 ) and nankai university baiqing plan foundation ( grant no .
zb15006104 ) . | quantum transport properties are instrumental to understanding quantum coherent transport processes
. potential applications of quantum transport are widespread , in areas ranging from quantum information science to quantum engineering , and not restricted to quantum state transfer , control and manipulation . here , we study light transport in a ring array of coupled resonators enclosed synthetic magnetic flux .
the ring configuration , with an arbitrary number of resonators embedded , forms an two - arm aharonov - bohm interferometer .
the influence of magnetic flux on light transport is investigated .
tuning the magnetic flux can lead to resonant transmission , while half - integer magnetic flux quantum leads to completely destructive interference and transmission zeros in an interferometer with two equal arms .
quantum transport , coupled resonators , synthetic magnetic field , aharonov - bohm flux |
in 1893 and 1897 lodge argued that a device using electromagnetic fields should in principle be able to detect angular acceleration by the interference of light .
sixteen years later sagnac @xcite observed such an effect and realised the potential for light interferometers as accelerometers .
the continued development of such instruments and their careful exploitation by michelson and others was pivotal in the establishment of our current world view of classical physics .
sagnac recognised that his observations could open up new avenues in the development of accelerometers , an insight that has been dramatically verified . today
the ring - laser is a cornerstone of most inertial guidance systems and optoelectronics is the backbone of the communications industry ( see @xcite for historical accounts of ring - laser development ) .
the basic principles leading to the sagnac effect are well understood in broad outline . as with all interferometers , a guided wave phenomenon ( such as that experienced by electromagnetic fields in a wave guide ) is used to distinguish different propagation paths in spacetime .
the traditional electromagnetic `` passive '' type of sagnac interferometer involves splitting a light beam into two coherent components , forcing them to travel different paths in spacetime to an event where they combine to produce an interference pattern .
the nature of the interference fringes can then be correlated with the difference in the optical paths .
in such a device the frequency of the light employed is controlled externally and the circuits in spacetime are often determined by mirrors or continuous dielectric fibres . with the discovery of the laser it was found that the electromagnetic modes of a closed tubular cavity containing an active lasing medium ( excited initially by an external rf field ) could be made to produce an interference pattern that varied with the state of rotation of the whole apparatus .
such an instrument has been labelled an `` active '' device to distinguish it from the traditional passive sagnac interferometer .
certain modes of a non - rotating cavity can be made to form standing waves .
however , the co- and counter - propagating modes in a rotating lasing cavity acquire different resonant frequencies leading to a beat - frequency as observed by an on - board probe . in 1963
sperry - rand recognised that the active ring - laser could be used as an inertial - guidance device . today
such miniturised interferometers are used routinely in civil and military applications . considerable increase in sensitivity
can be achieved using larger ring - lasers .
insulated from thermal and seismic noise , sensitivities of @xmath0 are routinely obtained in the ug1 sagnac interferometer housed in the cashmere cavern in christchurch , new zealand . with an effective area of @xmath1 and perimeter
@xmath2 it operates with a he - ne laser at @xmath3 and an intracavity power of @xmath4 at each mirror .
high resolution can be achieved by pattern matching to periodic variation of the sagnac frequency induced by rotation of the earth . using matching based on geophysical modelling , detection of lunar perturbations of the earth s rotation and the amplitudes of the oppolzer modes
have recently been reported in @xcite .
the instantaneous direction of the earth s rotation axis has been measured to a precision of @xmath5 part in @xmath6 when averaged over a time interval of several hours .
the enhanced sensitivity of such large ring - lasers owes much to the significant improvement of mirror design in recent years .
it is imperative for the maintenance of a sagnac beat signal that pollution of the resonant state by competing modes is inhibited .
this requires considerable engineering skill , the use of thermally inert materials and mirrors that can maintain the dominant cavity resonant mode over long periods of time . with such advances in technology
it is natural to seek the sensitivities that are necessary to discriminate the effects of non - newtonian gravitation on the behaviour of light @xcite .
the most important of these is the gravito - magnetic field of the rotating earth , predicted by einstein s description of gravitation .
however , even crude estimates of the effects of terrestrial gravito - magnetism show that a number of competing effects that may modify the classical sagnac effect ( due to the rotation of the apparatus ) need careful consideration .
such effects are often dependent on the structure of the interferometer and so invite one to contemplate alternative designs . although , in general , one expects that the sensitivity of an active ring - laser will increase with its spatial dimensions one may also enquire about the dependence of sensitivity on its geometry and spatial shape . in order to discriminate gravito - magnetic effects
it then becomes necessary to understand in some detail how the geometry and shape of the interferometer affect the resonant mode structure of the cavity . in this way
one can hope to discriminate between a number of features that compete to modify the classical sagnac frequency .
the effect of gravitation on the sagnac beat frequency is usually discussed in terms of the high frequency ray - optic description of light . in this way
one can simply relate the interference of co- and counter - propagating electromagnetic modes to the proper time difference between events that terminate their ray histories in spacetime @xcite .
while such an analysis can give valuable insights into the leading terms in an asymptotic expansion in the lasing frequency it can not address the effects due to the vector nature of the electromagnetic modes and the effects of the infinity of competing modes possessed by the lasing cavity .
an exact analysis of the wave nature of light in an active sagnac cavity is a non - trivial mathematical problem that has to our knowledge never been fully considered .
the problem is complicated by the fact that the harmonic electromagnetic modes rarely split into simpler transverse electric ( te ) and transverse magnetic ( tm ) modes that greatly simplify the analysis in most simply - connected axially symmetric cavity geometries .
such a separation relies on a number of factors such as the shape of the entire cavity in space , the homogeneity of the lasing medium , the nature of the acceleration of the cavity and the presence of even a weak non - newtonian gravitational field .
in order to accommodate these complexities a viable approximation is mandatory . here
the problem is approached in terms of an expansion scheme determined by scales related to the geometry of the cavity .
thus slender cavities are considered where the ratio of the transverse dimension to the cavity length is small .
such cavities will be called `` wavetubes '' . in order to confidently predict near - earth general relativistic contributions to the active sagnac beat frequency a careful electromagnetic mode analysis of the cavity
must be considered .
this can be done in terms of the covariant maxwell equations @xcite for moving media @xcite on a background spacetime determined by einstein s equations for gravitation .
the procedure will exploit a local system of coordinates adapted to the geometry of a wavetube .
such a wavetube can be characterised in terms of the locus in space of the centre of each of its transverse cross - sections .
this locus is regarded as a closed space curve with possible non - zero frenet torsion and curvature .
the approximation scheme will be developed for a wavetube with central locus having local curvature departing slightly from that of a fixed plane circle . in sections 2 and 3
it is shown how to construct te and tm type solutions to the covariant maxwell equations for a particular class of coframes .
these include the fields in rotating cavities containing a homogenous , isotropic and dispersionless medium in the presence of terrestrial gravito - magnetism . the language of exterior differential forms is used throughout since it offers the most succinct formulation for the field equations and makes clear the role of the spacetime metric and local coframe that feature in the subsequent computational scheme .
the relation to the inertial 3-d formulation can be found in @xcite and references therein . in section 4 a particular coordinate system
is introduced that , together with a particular frenet frame , encodes the geometry of the wavetube and permits one to introduce dimensionless parameters that control the nature of the approximation scheme . in these coordinates differences between non - inertial motion and gravito - magnetic effects
are made explicit in sections 5.1 and 5.2 .
a complete mode analysis of the rotating wavetube is given in section 6 and the effects of multi - mode excitation and the description of sagnac beats is explored in section 7 . in particular attention
is drawn to the effects of non - planarity of the wavetube on the beat frequency by reference to a particular geometry generated by the structure of a torus - knot .
to accommodate the ideas above it is first shown how a class of spacetime metrics that admit local orthonormal coframes with particular properties are sufficient to construct te and tm type electromagnetic modes relative to such a coframe .
the language of exterior forms is used throughout range over the integers @xmath7 , indices in the middle of the latin alphabet @xmath8 range over the integers @xmath9 and greek indices @xmath10 range over @xmath11 .
expressions involving repeated indices imply use of the einstein summation convention . ]
since this greatly facilitates the discussion and leads in a direct manner to an explicit construction of the fields below .
the hollow wavetube is supposed to be constructed of a perfectly conducting material which in this section contains no lasing medium .
let @xmath12 be a spacetime and let @xmath13 be a local orthonormal coframe with respect to the metric tensor @xmath14 with dual frame @xmath15 , @xmath16 and orientation ( volume form ) defined by the hodge map @xmath17 with @xmath18 the inverse metric tensor @xmath19 is @xmath20 define the split of the cotangent space @xmath21 with @xmath22 spanned by @xmath23 and @xmath24 spanned by @xmath25 at @xmath26 .
these subspaces inherit hodge maps with @xmath27 and @xmath28 so @xmath29 now restrict to spacetime metrics @xmath30 that admit local orthonormal coframes satisfying the conditions @xmath31 these conditions permit a viable approximation scheme leading to a tractable decomposition of electromagnetic modes in a wide context to be discussed below .
a differential form @xmath32 ( vector field @xmath33 ) on @xmath34 that satisfies @xmath35 ( @xmath36 ) for all @xmath37 will be called _ _
longitudinal _ _ is the interior derivative on forms with respect to the vector field @xmath33 . ] .
a _ transverse _ differential form @xmath38 ( vector field @xmath39 ) will satisfy @xmath40 ( @xmath41 ) for all @xmath42 .
let the @xmath43-form @xmath44 be a solution to the vacuum maxwell equations @xmath45 in a spacetime region bounded by a perfectly conducting wavetube hypersurface @xmath46 and subject to the boundary condition @xmath47 the wavetube interior on spacetime is topologically @xmath48 where @xmath49 is a @xmath43-disc .
solutions to maxwell s equations are sought by adapting the above coframe to the wavetube geometry .
furthermore , we shall suppose that ( [ fresnel_bc ] ) must be satisfied with @xmath50 , for all @xmath51 , a transverse @xmath5-form . since @xmath52 the pair @xmath23 are , by frobenius theorem , normal @xmath5-forms to a local foliation @xmath53 of @xmath34 .
the leaves of @xmath53 contain the wavetube cross - sections , each of whose intersection with @xmath54 is the image of a closed curve .
locally , we seek separable propagating solutions of the form @xmath55 for some @xmath43-form @xmath56 where @xmath57 for constant components @xmath58 .
we decompose @xmath56 into longitudinal and transverse parts @xmath59 thus @xmath60 are transverse @xmath5-forms and @xmath61 are transverse @xmath43-forms .
using the identity @xmath62 where @xmath63 for any vector field @xmath33 on @xmath34 and @xmath42 is any @xmath64-form on @xmath34 it can be shown that @xmath65 where @xmath32 is a longitudinal @xmath64-form , @xmath38 is a transverse @xmath66-form and @xmath67 inserting the expression for @xmath44 into maxwell s equations ( [ maxwell_df_dstarf ] ) gives @xmath68 where ( [ hodge_as_parhodge_perphodge ] ) has been used .
the form structure indicates that the above pair splits into @xmath69 acting with @xmath17 on ( [ split_maxwell_dphi ] ) and ( [ split_maxwell_dpsi ] ) and using ( [ hodge_as_parhodge_perphodge ] ) yields @xmath70 where @xmath71 and @xmath72 . acting with @xmath73 , where @xmath74 , on equation ( [ split_maxwell_ealpha ] ) yields @xmath75 where @xmath76 .
similarly , ( [ split_maxwell_parhodgeealpha ] ) leads to @xmath77 thus , inserting ( [ phi_kkappa_related ] ) in ( [ dkkappa_phi_related ] ) and ( [ psi_kkappa_related ] ) in ( [ dkkappa_psi_related ] ) yields @xmath78 the helmholtz equations ( [ equation_for_phi ] ) and ( [ equation_for_psi ] ) determine @xmath79 and @xmath80 subject to ( [ fresnel_bc ] ) . the boundary condition ( [ fresnel_bc ] ) may be expanded @xmath81 implying @xmath82 in general there are three possibilities for the metric norm @xmath83 of @xmath84 : @xmath85 @xmath86 .
* triviality of @xmath79 * + let @xmath87 be a @xmath43-chain for which the image of @xmath88 is in @xmath54 and @xmath89 and note @xmath90 using ( [ equation_for_phi ] ) and ( [ phi_boundary_condition ] ) where @xmath91 is the complex - conjugate of @xmath79 .
since @xmath92 where @xmath93 is any transverse @xmath64-form and , by hypothesis , @xmath86 it follows from ( [ integration_phi ] ) that @xmath94 and so , using ( [ phi_boundary_condition ] ) , @xmath95 . * triviality of @xmath80 and @xmath44 * + since @xmath84 is null @xmath96 is proportional to @xmath84 and so ( [ phi_kkappa_related ] ) and ( [ psi_kkappa_related ] ) with @xmath95 imply that @xmath97 vanishes and @xmath80 is constant .
let @xmath98 be the null longitudinal @xmath5-form with constant components @xmath99 that satisfies @xmath100 .
acting with @xmath101 on ( [ split_maxwell_parhodgeealpha ] ) and ( [ split_maxwell_ealpha ] ) and noting @xmath102 where @xmath103 it is found that @xmath104 since @xmath97 vanishes @xmath105 where @xmath106 and @xmath107 .
the differential form @xmath108 is transverse and so @xmath109 follows from ( [ kkappa_boundary_condition ] ) . by integrating ( [ dkkappa_l_psi_related ] ) over the @xmath43-chain @xmath87 introduced earlier , noting that ( [ kkappa_l_boundary_condition ] ) implies @xmath110 and recalling that @xmath80 is constant , it follows that @xmath111 and @xmath112 cross - sections of the wavetube interior are simply - connected and so by the poincar lemma @xcite @xmath113 it follows from ( [ div_kkappa_l ] ) and ( [ kkappa_l_boundary_condition ] ) that @xmath114 satisfies @xmath115 subject to the dirichlet boundary condition @xmath116 where @xmath117 is a complex constant .
the solution to ( [ laplace_in_null_case ] ) and ( [ laplace_in_null_case_boundary_cond ] ) is @xmath118 and so , using ( [ exactness_of_kkappa_zeta ] ) , @xmath119 and @xmath120 @xmath121 .
* triviality of @xmath79 * + using ( [ integration_phi ] ) and ( [ positivity_parhodge ] ) it can be seen that @xmath79 vanishes since , by hypothesis , @xmath121 .
* triviality of @xmath80 and @xmath44 * + introduce the spacelike normalized @xmath5-form @xmath122 and write @xmath123 where @xmath124 and @xmath125 and @xmath126 vanishes by ( [ phi_kkappa_related ] ) because @xmath95 .
thus ( [ kkappa_boundary_condition ] ) yields @xmath127 since @xmath84 is longitudinal and @xmath128 is transverse ( [ psi_kkappa_related ] ) gives the neumann boundary condition @xmath129 the simple - connectedness of the wavetube cross - sections implies that the eigenvalues @xmath130 in ( [ equation_for_psi ] ) associated with the neumann boundary condition ( [ psi_boundary_condition ] ) are negative . to prove this an argument similar to that used in the dirichlet case
is employed .
let @xmath87 be the @xmath43-chain introduced earlier and note that ( [ psi_boundary_condition ] ) implies @xmath131 , where @xmath132 is a @xmath133-form , and @xmath134 since @xmath135 .
thus @xmath136 using ( [ equation_for_psi ] ) and ( [ psi_boundary_condition_s ] ) where @xmath137 is the complex - conjugate of @xmath80 . by hypothesis @xmath121 and using ( [ positivity_parhodge ] )
it follows that @xmath111 . therefore @xmath120
@xmath138 .
this is the only situation that leads to a non - trivial expression for @xmath44 .
introduce the timelike normalized @xmath5-form @xmath139 and write @xmath140 where @xmath141 and ( [ phi_kkappa_related ] ) and ( [ psi_kkappa_related ] ) have been used . inserting the above equation for @xmath142 into the boundary condition ( [ kkappa_boundary_condition ] ) yields
@xmath143(p ) = 0 { \quad\forall\ , p\in{\ensuremath{\mathcal{b}}}}.\ ] ] however , ( [ phi_boundary_condition ] ) means that @xmath144 on @xmath54 is proportional to @xmath145 and so again ( [ psi_boundary_condition ] ) is obtained . expressed entirely in terms of @xmath146 the maxwell 2-form @xmath44
can be written @xmath147 where @xmath148\exp(iw_{tm}),\\ \label{te_f_in_terms_of_psi } f_{te } = \biggl [ \psi{\ensuremath{\#_\perp}}1 - i\frac{dw_{te}}{|dw_{te}|^2 } \wedge { \ensuremath{\#_\perp}}d\psi\biggr]\exp(iw_{te}).\end{gathered}\ ] ] the helmholtz equations for @xmath79 and @xmath80
are @xmath149 solved subject to the boundary conditions @xmath150 where @xmath151 and @xmath152 are timelike .
the constants @xmath153 are determined by the appropriate boundary conditions and , in general , @xmath154 .
to excite electromagnetic modes in a wavetube one may fill it with a gas such as a helium - neon mixture and induce it to lase with an external rf field .
the presence of a lasing medium requires that one take into account its electrical and magnetic susceptibility .
the electromagnetic field in a material medium is described by a pair of 2-forms @xmath155 on spacetime that satisfy the equations : @xmath156 where @xmath37 is a source 3-form . for a given @xmath37 , the 2-form @xmath157 must be related to @xmath44 in order to have a closed system .
this is usually done by relating their components relative to decompositions of the form : @xmath158 where the timelike vector @xmath159 is normalised with @xmath160 and the forms @xmath161 are all annihilated by @xmath162 . with a spacetime metric tensor @xmath30 , having physical dimensions of length squared , a coherent dimensional scheme arises with the aid of the permittivity constant @xmath163 and permeability @xmath164 of free space satisfying @xmath165 where @xmath166 is the speed of light in vacuum . in terms of conventional mks components of electric and magnetic fields
one has in any @xmath30-orthonormal coframe @xmath167 adapted to @xmath159 : @xmath168 for @xmath157 one has @xmath169 and @xmath170
. for a linear isotropic homogeneous dispersionless non - conducting medium with relative permittivity @xmath171 and relative permeability @xmath172 one has the simple constitutive relations @xmath173 and @xmath174 or @xmath175 such a constitutive relation assumes that the dimensionless quantities @xmath171 and @xmath172 are frame - independent constant scalars on spacetime .
the refractive index @xmath176 of the medium is @xmath177 propagating solutions to the source - free maxwell equations @xmath178 that satisfy the boundary condition @xmath179 at the wavetube surface @xmath54 , @xmath180 have a similar structure to ( [ tm_f_in_terms_of_phi ] ) and ( [ te_f_in_terms_of_psi ] ) when @xmath159 has the form @xmath181 where @xmath182 are constant and @xmath183 belong to the frame @xmath184 dual to the coframe @xmath185 satisfying the conditions ( [ coframe_assumptions ] ) . it can be shown that @xmath186 and @xmath187 are solutions to ( [ source_free_maxwell_medium ] ) where @xmath188\exp(iw_{tm}),\\ f_{te } = \biggl [ \psi{\ensuremath{\#_\perp}}1 - i\frac{dw_{te}}{|\alpha_{te}|^2 } \wedge { \ensuremath{\#_\perp}}d\psi\biggr]\exp(iw_{te } ) \label{te_f_medium_in_terms_of_psi}\end{gathered}\ ] ] with @xmath189 the helmholtz equations for @xmath79 and @xmath80 are @xmath190 and are solved subject to the boundary conditions @xmath191 as before @xmath151 and @xmath152 where @xmath192 and @xmath193 are constant .
a moving space curve in euclidean @xmath194 may be represented by a @xmath194-valued mapping @xmath195 .
suppose that @xmath196 parameterises arc length and @xmath197 is a time variable .
the instantaneous geometry of the space curve may be described by three mutually orthogonal unit vectors @xmath198 on the image of @xmath199 at each time @xmath200 . for a frenet frame with @xmath201 tangent to @xmath202 one
has : @xmath203 where @xmath204 denotes the derivative of @xmath205 with respect to @xmath196 and @xmath206 and @xmath207 denote the curvature and torsion of @xmath202 respectively .
we shall restrict our discussion to space curves that admit continuous frenet frames along their entire length . since the triad @xmath208 forms an orthonormal basis on the curve the time derivative @xmath209 is related to @xmath208 by an anti - symmetric matrix for each @xmath210 .
we are interested in space curves that rotate rigidly with uniform angular velocity @xmath211 without deformation .
thus @xmath212 , @xmath213 and @xmath214 .
these imply @xmath215 in a fixed global cartesian triad @xmath216 one may write in cylindrical coordinates : @xmath217 if this triad is oriented so that @xmath218 then @xmath219 and @xmath220 are independent of @xmath197 and @xmath221 .
let the vector @xmath222 with cartesian components @xmath223 in the frame @xmath216 , @xmath224 locate a point at the instant @xmath200 .
if this point lies in a tubular neighbourhood about the space curve @xmath202 it may also be labelled by the frenet coordinates @xmath225 such that : @xmath226 thus for points in this region one has a transformation between the cartesian coordinates @xmath223 and the frenet coordinates @xmath227 .
when the curve moves @xmath228 and so for rigid motion the above gives : @xmath229dt\\ & + [ 1 -\kappa(s ) x_1]ds\,{{\mathbf t}}+[d x_1 - x_2 \tau(s ) ds]{{\mathbf n}}+[d x_2 + x_1 \tau(s ) ds]{{\mathbf b}}\\ = & dy_1{{\mathbf i}}+ dy_2{{\mathbf j}}+ dy_3{{\mathbf k}}\end{split}\ ] ] where @xmath230 , @xmath231 and @xmath232 .
these equations imply a @xmath197 dependent coordinate transformation between @xmath223 and @xmath225 .
the following is concerned with cavities that have small circular cross - section compared with their length and are represented by a space curve that approximates a plane circle rotating about its axis . to this end
it is convenient to introduce a number of dimensionless parameters and new coordinates .
first introduce polar variables @xmath233 and @xmath234 by @xmath235 where @xmath236 , @xmath237 and @xmath238 is the radius of the circular cross - section .
it is supposed that the absolute values of the curvature and torsion of the closed curve @xmath199 are bounded from above by @xmath239 and @xmath240 respectively and that the scale of @xmath199 is given by that of a circle of radius @xmath241 .
introduce the dimensionless parameters @xmath242 the dimensionless parameter @xmath243 indicates the order of small quantities that will be used below to approximate the metric appropriate for a cavity with small cross section .
alternative assumptions about the assignment of @xmath243 to the quantities @xmath244 will give different approximation schemes .
the above choice will be shown to give the classical sagnac beat frequency to first order in a wavetube based on a planar space curve .
the calculation will proceed for @xmath245 to lowest order in @xmath243 for a curve that approximates a circle of radius @xmath241 rotating with uniform angular speed @xmath246 about its axis .
introduce the scaled polar coordinates @xmath247 and @xmath248 where @xmath249 and the scaled quantities @xmath250 and consider the class of @xmath199 for which @xmath251(s , t ) = a_0\bigl[1 + \epsilon\gamma_1({\hat{s}},{\hat{\lambda}})\bigr],\\ \notag [ { { \mathbf n}}\cdot({{\mathbf k}}{{\mathbf \times}}{{\mathbf c}})](s , t ) = a_0\epsilon\gamma_2({\hat{s}},{\hat{\lambda}}),\\ \label{approximate_curve_conditions } [ { { \mathbf b}}\cdot({{\mathbf k}}{{\mathbf \times}}{{\mathbf c}})](s , t ) = a_0\epsilon\gamma_3({\hat{s}},{\hat{\lambda}}).\end{gathered}\ ] ] thus , to lowest ( zeroth ) order in @xmath243 @xmath252 + { { \mathbf b}}[d{\hat{x}}_2 + \mu_4\hat\tau({\hat{s } } ) { \hat{x}}_1 d{\hat{s } } ] + { { \mathbf t}}\mu_2 d{\hat{\lambda}}.\ ] ] in terms of @xmath247 and @xmath248 @xmath253{\ensuremath{\otimes}}[d\phi+\mu_4{\hat{\tau}}({\hat{s } } ) d{\hat{s } } ] + d{\hat{\rho}}{\ensuremath{\otimes}}d{\hat{\rho}}\end{split}\ ] ] and the zeroth order space curve is a circle of radius @xmath241 orthogonal to the rotation axis @xmath254 .
we shall consider only stationary axisymmetric spacetimes . thus @xmath12 admits a commuting pair of killing vector fields @xmath255 and @xmath256 : @xmath257 = 0\end{gathered}\ ] ] where @xmath258 is the lie derivative with respect to @xmath33 .
the vector field @xmath255 is timelike and future - pointing and the vector field @xmath256 is spacelike and has closed orbits . here by definition ,
material points that are `` rotating '' and @xmath256 above . ]
follow timelike worldlines whose tangents at any event are a linear combination @xmath159 of @xmath255 and @xmath256 at that event .
if the spacetime is asymptotically flat and the normalised @xmath159 is timelike asymtotically it also offers a field of stationary observers . before constructing a local coframe adapted to a twisted wavetube rotating in a spacetime with gravitation we consider the simpler case of minkowski spacetime .
the metric tensor @xmath30 on minkowski spacetime is @xmath259 in terms of the inertial coordinates @xmath260 .
it possesses the timelike future - pointing killing vector field @xmath261 and the spacelike killing vectors @xmath262 and @xmath263 where @xmath264 on the wavetube spacetime domain @xmath260 is related to the spacetime frenet coordinates @xmath265 by @xmath266,\\ y_2 = { { \mathbf j}}\cdot[{{\mathbf c}}(s , t)+x_1{{\mathbf n}}(s , t)+x_2{{\mathbf b}}(s , t)],\\ y_3 = { { \mathbf k}}\cdot[{{\mathbf c}}(s , t)+x_1{{\mathbf n}}(s , t)+x_2{{\mathbf b}}(s , t)].\end{gathered}\ ] ] the wavetube apparatus follows integral curves of @xmath267 in spacetime . by expressing the vector field @xmath267 with respect to @xmath260 : @xmath268 it can be seen that @xmath267 is a linear combination of the timelike future - pointing killing vector field @xmath269 and the spacelike killing vector field @xmath256 : @xmath270 where @xmath271=0 $ ] . at
the instant @xmath272 write @xmath199 in the form @xmath273 where @xmath274 is a plane circle in @xmath194 rotating with angular speed @xmath275 about its axis @xmath254 .
the curve image @xmath276 has tangent @xmath277 , normal @xmath278 , constant binormal @xmath279 and radius @xmath241 .
note that @xmath196 is _ not _ necessarily the arc parameter of @xmath274 .
the vector @xmath276 can be expressed @xmath280 and it is supposed that the tangent @xmath201 , normal @xmath281 and binormal @xmath282 of @xmath199 satisfy @xmath283{{\mathbf n}}_t(s ) + \sin[\beta(s , t)]{{\mathbf b}}_t(s ) + o(\epsilon),\\ \label{moving_curve_requirements } { { \mathbf b}}(s , t ) = -\sin[\beta(s , t)]{{\mathbf n}}_t(s ) + \cos[\beta(s , t)]{{\mathbf b}}_t(s ) + o(\epsilon)\end{gathered}\ ] ] for some angle @xmath284 . using @xmath285
it follows that @xmath286 which is consistent with ( [ approximate_curve_conditions ] ) .
thus , in terms of @xmath265 the metric tensor @xmath30 is @xmath287 = \rho [ d\phi + \tau(s)d s].\end{gathered}\ ] ] where @xmath13 is orthonormal to lowest order in @xmath243 .
note that the approximate minkowski metric is insensitive to @xmath288 and the curvature of @xmath199 .
the metric tensor at large distances from an isolated compact rotating body in an asymptotically flat spacetime , with newtonian gravitational mass @xmath289 and angular momentum @xmath290 , may be approximated ( for @xmath291 where @xmath292 is newton s gravitational constant ) by @xmath293 in spherical polar coordinates @xmath294 .
the use of the metric ( [ kerrm ] ) is of course an idealisation for physical applications such as terrestrial ring lasers , where one does not have two killing vectors . in particular , due to interactions with other celestial bodies the positions of both the rotation and angular momentum axes will not quite coincide with each other and their positions will vary with time . as reported in ref .
@xcite for the case of ring lasers at the earth s surface such effects are three to four orders of magnitude larger than the lense - thirring effect .
nonetheless the calculation presented here serves a useful purpose in isolating effects due to gravito - magnetism .
the implications of time variation of the rotation and angular momentum axes of the earth on terrestrial lense - thirring measurements will be discussed elsewhere @xcite .
the vector fields @xmath261 and @xmath295 are a commuting pair of killing vectors .
furthermore , @xmath269 is timelike and future - pointing and @xmath296 is spacelike and has closed orbits . consider the acceleration of matter on one of the integral curves of the 4-velocity field @xmath297 with respect to a stationary observer belonging to the 4-velocity field @xmath298 . at events where the integral curves of @xmath159 and @xmath299 coincide @xmath159 has spatial acceleration @xmath300 where @xmath301 projects the 4-acceleration to the instantaneous 3-space of @xmath299 . in the non - relativistic weak gravitational field limit one
finds that the g - orthonormal components of @xmath302 yield a centripetal acceleration with dominant magnitude @xmath303 at an event with coordinates @xmath304 .
thus each observer within @xmath299 would interpret such events within @xmath159 to have instantaneous 3-acceleration produced by rotation with angular speed @xmath275 .
the asymptotically minkowski coordinate chart @xmath260 is given in terms of @xmath294 as @xmath305 the dot @xmath306 and cross @xmath307 products of the pair of vectors @xmath308 are defined with respect to the tensor @xmath132 : @xmath309 and the @xmath132-orthonormal frame @xmath216 : @xmath310 and , as before , the vector @xmath311 , @xmath312 locates the point @xmath260 .
the metric tensor @xmath30 has the form @xmath313 \end{split}\ ] ] where @xmath260 and the spacetime coordinates @xmath314 satisfy @xmath315,\\ y_2 = { { \mathbf j}}\cdot[{{\mathbf c}}({{\underline{s}}},t)+{{\underline{x}}}_1{{\mathbf n}}({{\underline{s}}},t)+{{\underline{x}}}_2{{\mathbf b}}({{\underline{s}}},t)],\\ y_3 = { { \mathbf k}}\cdot[{{\mathbf c}}({{\underline{s}}},t)+{{\underline{x}}}_1{{\mathbf n}}({{\underline{s}}},t)+{{\underline{x}}}_2{{\mathbf b}}({{\underline{s}}},t)]\end{gathered}\ ] ] and the @xmath132-orthonormal frame @xmath208 solves the frenet - serret equations @xmath316 the functions @xmath317 are underlined as a reminder that they are normalized with respect to the tensor @xmath132 rather than the spacetime metric @xmath30 .
as in the previous section consider curves of the form @xmath318 subject to ( [ moving_curve_requirements ] ) , with @xmath196 replaced by @xmath319 , where @xmath320 is the rotating plane circle of @xmath132-radius @xmath241 centred at the point @xmath321 with the frenet frame @xmath322 and @xmath218 .
here one is considering @xmath323 perturbations about a `` circular '' space curve parallel to the `` equatorial plane '' of the kerr geometry where elements of the wavetube apparatus follow integral curves of @xmath324 in this spacetime .
as before , introduce the dimensionless functions @xmath325 since @xmath326 and the position vector @xmath327 has the form @xmath328\ ] ] and using @xmath329 + { { \mathbf b}}[d{\hat{\underline{x}}}_2 + \mu_4\hat{{\underline{\tau}}}({\hat{\underline{s } } } ) { \hat{\underline{x}}}_1 d{\hat{\underline{s } } } ] + { { \mathbf t}}\mu_2 d{\hat{\lambda}},\\ { { \mathbf b}}= { { \mathbf k}}\end{gathered}\ ] ] with ( [ moving_curve_requirements ] ) it follows that @xmath330.\ ] ] the metric tensor @xmath30 adapted to @xmath199 is @xmath331\\ & - \mu_6 \bigl[(d{\hat{\underline{s}}}+ \mu_2 d{\hat{\lambda}}){\ensuremath{\otimes}}d{\hat{\lambda}}+ d{\hat{\lambda}}{\ensuremath{\otimes}}(d{\hat{\underline{s}}}+ \mu_2 d{\hat{\lambda}})\bigr ] \end{split}\ ] ] to lowest order in @xmath243 where @xmath332 and @xmath333 a @xmath30-orthonormal coframe to lowest order in @xmath243 is @xmath334^{1/2}c dt,\\ \notag e^1 = \bigl[(1+\mu_5)^{1/2}\mu_2-\mu_6/(1+\mu_5)^{1/2}\bigr ] cdt + ds,\\ \notag e^2 = d\rho,\\ \label{lowest_order_kerr_coframe } e^3 = \rho [ d\phi + \tau(s)ds]\end{gathered}\ ] ] where @xmath335
both coframes ( [ lowest_order_minkowski_coframe ] ) and ( [ lowest_order_kerr_coframe ] ) satisfy ( [ coframe_assumptions ] ) and so single - valued propagating electromagnetic @xmath44 modes inside a wavetube with circular cross - section of radius @xmath238 can be generated from particular solutions of the helmholtz equations discussed earlier .
solutions regular in the wavetube follow from the forms @xmath336 $ ] where @xmath337 , @xmath338 and @xmath339 or @xmath340 ( see equations ( [ alpha_t_beta_t_explicit ] ) ) .
the role of the frenet torsion in solutions of the scalar helmholtz equation in coordinates adapted to non - planar curves has been noted before in a number of different contexts @xcite .
the @xmath186 and @xmath187 modes follow from @xmath341\ ] ] with @xmath342 satisfying @xmath343 and @xmath344\ ] ] with @xmath345 satisfying @xmath346 where @xmath347 is the @xmath348th regular bessel function of the first kind and @xmath349 . for both te and tm modes @xmath350 and , using @xmath351 , @xmath352 is the frequency ( hz ) , measured by the wavetube apparatus , of the field mode associated with @xmath353 or @xmath354 .
positive wavenumbers @xmath355 correspond to propagation in the direction of the tangent @xmath201 and negative wavenumbers to propagation in the opposite direction . since the wavetube is closed , i.e. the wavetube interior on spacetime is topologically @xmath48 , the maxwell @xmath43-form satisfies @xmath356 where @xmath357 is the length of the wavetube and so using ( [ xi_in_terms_omega_k ] ) and either ( [ tm_f_medium_in_terms_of_phi ] ) or ( [ te_f_medium_in_terms_of_psi ] ) @xmath358 \equiv k_\xi = \frac{2\pi n}{l } - n{\bar{\tau}},\quad\,n\in{{\mathbf z}}\ ] ] where @xmath359 is the average frenet torsion .
the mode spectrum is classified using the triple of integers @xmath360 , where the positive integer @xmath64 labels a solution of @xmath361 in the tm case or @xmath362 in the te case .
the longitudinal members of ( [ lowest_order_minkowski_coframe ] ) and ( [ lowest_order_kerr_coframe ] ) have the form @xmath363 where @xmath364 and @xmath365 are constants with the dimensions of @xmath166 . for each @xmath366 , where @xmath339 or @xmath340 ( see equations ( [ alpha_t_beta_t_explicit ] ) ) and @xmath367 is determined by the boundary conditions , the dispersion relation @xmath368 with @xmath369 yields @xmath370\ ] ] where @xmath371 it is assumed that , for terrestrial applications , @xmath372 and so the sign of the square root of the discriminant in ( [ wavetube_omega_spectrum ] ) is chosen to be positive .
in addition to @xmath373 $ ] introduce the scalars @xmath374 dependent on the mode indices @xmath360 and the type @xmath375 as follows : @xmath376,\\ \zeta_t \equiv \zeta_t[p , n]\end{gathered}\ ] ] where @xmath377)=0 $ ] and @xmath378)=0 $ ] .
the wavetube apparatus measures the frequency @xmath379 = \frac{c}{2\pi}dw_t\bigl(v)\ ] ] and so here @xmath380=\frac{\omega_t[n , p , n]c}{2\pi\sqrt{a^2-b^2}}.\ ] ] we may approximate this expression further by taking into account the relative magnitudes of @xmath381 , @xmath348 , @xmath382 , @xmath383 and @xmath384 . for a lasing medium excited by a rf field @xmath385 is expected and it is assumed that @xmath386 and @xmath387 . then ( [ wavetube_omega_spectrum ] ) has the form @xmath388 \simeq \frac{|k[n , n]|(a^2-b^2)}{a{{\ensuremath{\mathcal{n}}}}+{\text{sgn}}(k[n , n])b}\ ] ] where @xmath389 using ( [ k_in_terms_of_n_tau ] ) and @xmath390 hence @xmath391 and @xmath392 are independent of the mode type @xmath393 and the bessel root index @xmath64 in this approximation .
henceforth we write @xmath394 and @xmath395 and drop their @xmath64 dependencies . for terrestrial applications
we expect that @xmath396 ( as well as @xmath397 ) . in this case
( [ wavetube_frequency_spectrum ] ) becomes @xmath398 \simeq \frac{c|n|}{l{{\ensuremath{\mathcal{n } } } } } - { \text{sgn}}(n)\biggl[\frac{c|n|}{l}\frac{b}{a{{\ensuremath{\mathcal{n}}}}^2 } + \frac{cn{\bar{\tau}}}{2\pi{{\ensuremath{\mathcal{n}}}}}\biggr]\ ] ] where ( [ mod_k_in_terms_of_n_tau ] ) and ( [ high_freq_wavetube_omega_spectrum ] ) have been used .
if a number of electromagnetic modes of similar frequency are active in the wavetube then the time histories of the electromagnetic fields measured by the wavetube apparatus will exhibit beating .
the classical @xcite sagnac beat frequency @xmath399 induced by rotation on flat spacetime follows from ( [ approx_high_frequency_wavetube_frequency_spectrum ] ) with @xmath400 and by choosing a pair of rotationally symmetric ( @xmath401 ) modes that differ only in their directions of propagation and have @xmath402 : @xmath403 - \nu[-n,0]\\ & \simeq -\frac{2\omega n a_0}{l } \end{split}\ ] ] where @xmath364 and @xmath365 have been obtained from ( [ lowest_order_minkowski_coframe ] ) . to reveal the classical expression for the sagnac beat frequency the parameters @xmath241 and @xmath357 and the mode index @xmath381 are eliminated by introducing the area @xmath404 and perimeter @xmath405 of the plane circle approximated by the wavetube locus and the lasing wavelength @xmath406 .
it follows that @xmath407 the most general frequency difference is constructed from a pair of modes of type @xmath408 with indices @xmath409 and @xmath410 . in the approximation introduced earlier
we only need consider @xmath381 and @xmath348 dependences : @xmath411 \equiv \nu[n_1,n_1 ] - \nu[n_2,n_2].\ ] ] thus , with @xmath402 , ( [ standard_sagnac_beat_frequency ] ) has the generalization @xmath412
\simeq -\frac{2n\omega a_0}{l{{\ensuremath{\mathcal{n}}}}^2 } - \frac{cn{\bar{\tau}}}{\pi{{\ensuremath{\mathcal{n}}}}}.\ ] ] for terrestrial applications the external rf field generates counter- and co - propagating modes with indices @xmath413 and @xmath414 whose magnitudes are so large it is possible that @xmath415 . more generally , there is no reason to assume that @xmath416 and @xmath417 are identical . with @xmath418 and @xmath419 and using ( [ approx_high_frequency_wavetube_frequency_spectrum ] ) it follows that @xmath420 \simeq & \frac{c}{l{{\ensuremath{\mathcal{n}}}}}(n_1+n_2 ) - \frac{a_0\omega}{l{{\ensuremath{\mathcal{n}}}}^2}(n_1-n_2)\\ & - \frac{c{\bar{\tau}}}{2\pi{{\ensuremath{\mathcal{n}}}}}(n_1+n_2 ) .
\end{split}\ ] ] the dimensionless parameters @xmath421 , @xmath422 contain the _ euclidean _ radius @xmath241 of the multiply - wound plane circle @xmath423 that approximates the wavetube locus on kerr spacetime @xmath12 .
one may eliminate the parameter @xmath241 in favour of a length determined by the spacetime metric @xmath30 rather than the euclidean metric @xmath424 . for each constant @xmath425
the multiply - wound planar circle may be described as the image of the map @xmath426 & \rightarrow { \ensuremath{\mathcal{m}}}\\ u & \mapsto ( t = t_0,\,r = r_0,\,\theta=\theta_0,\,\varphi = u ) \end{split}\ ] ] where @xmath294 are the coordinates used in ( [ kerrm ] ) , @xmath427 and the positive integer @xmath428 is a winding number .
the arclength @xmath429 of @xmath423 is @xmath430 where @xmath431 is tangent to @xmath423 . using ( [ kerrm ] ) , ( [ spacetime_gamma_map ] ) and
( [ g_length_of_circle ] ) it follows that @xmath432 where @xmath433 introduce dimensionless parameters @xmath434 @xmath435 so that , by using ( [ a0_in_terms_of_l0 ] ) to eliminate @xmath241 : @xmath436 typically , for terrestrial applications , the magnitudes of the dimensionless parameters @xmath437 are much less than unity .
thus , it is reasonable to approximate @xmath438 by a polynomial obtained from the multivariate taylor expansion of @xmath438 with respect @xmath439 about @xmath440 .
however , when truncating the taylor expansion the numerical values of the terms in the series must be carefully scrutinized .
for example , the parameters associated with a wavetube whose locus approximates @xmath423 with `` radius '' @xmath441 fixed on the earth and centred on the poles have magnitudes @xmath442 , @xmath443 and @xmath444 . for terrestrial conditions @xmath445 is generally true .
it follows that @xmath446 is a good approximation ( the other terms in the taylor series are about 10 orders of magnitude smaller ) . using ( [ approx_high_frequency_wavetube_frequency_spectrum ] ) and ( [ approx_kerr_b_over_a ] ) it follows that @xmath447 \simeq \frac{c|n|}{l{{\ensuremath{\mathcal{n } } } } } -{\text{sgn}}(n)\biggl[\frac{l_0}{2\pi m}\frac{\omega|n|}{l{{\ensuremath{\mathcal{n}}}}^2 } \biggl(1+\frac{{g_n}m}{r_0 c^2 } -\frac{2{g_n}j}{\omega
r_0 ^ 3 c^2 } \biggl ) + \frac{cn{\bar{\tau}}}{2\pi{{\ensuremath{\mathcal{n}}}}}\biggr]\ ] ] and , more generally , for a pair of modes with indices @xmath448 and @xmath449 , where @xmath418 and @xmath419 , @xmath450 \simeq & \frac{c}{l{{\ensuremath{\mathcal{n}}}}}(n_1+n_2)\\ & -\frac{l_0}{2\pi m}\frac{\omega}{l{{\ensuremath{\mathcal{n}}}}^2 } \left[1+{{g_n}m\over r_0 c^2}-{2{g_n}j\over\omega r_0 ^ 3 c^2}\right ] ( n_1-n_2)\\ & - \frac{c{\bar{\tau}}}{2\pi{{\ensuremath{\mathcal{n}}}}}(n_1+n_2 ) . \end{split}\ ] ] by choosing a pair of rotationally symmetric ( @xmath401 ) modes that differ only in their directions of propagation and have @xmath451 we find the classical sagnac beat frequency ( [ standard_sagnac_beat_frequency ] ) is modified in this approximation to @xmath452 - \nu[-n,0]\\ & \simeq -\frac{1}{{{\ensuremath{\mathcal{n}}}}^2}\frac{4\omega{{\ensuremath{\mathcal{a}}}}}{\lambda{{\ensuremath{\mathcal{p } } } } } \left[1+{{g_n}m\over r_0 c^2}-{2g_nj\over\omega r_0 ^ 3c^2}\right ] \end{split}\ ] ] where the `` area '' @xmath453 and `` perimeter '' @xmath454 of @xmath423 are defined as @xmath455 in terms of total arclength @xmath429 and winding number @xmath428 of @xmath423 and @xmath456 . using ( [ scales_and_barred_scales ] )
it can be seen that the average torsion @xmath457 is related to the euclidean torsion @xmath458 of the wavetube locus by @xmath459 where @xmath460 . to illustrate the significance of the frenet torsion contribution to ( [ generalized_sagnac_beat_frequency ] ) we consider a wavetube based on a particular torus knot .
a type @xmath461 torus knot is a curve @xmath462 specified by a pair of integers @xmath463 and a triple of real numbers @xmath464 : @xmath465\\ { { \bm{\gamma}}}_1(\sigma ) = { { \mathbf i}}\cdot{{\bm{\gamma}}}(\sigma ) = [ v_1 + v_2\cos(u_2\sigma)]\cos(u_1\sigma),\\ { { \bm{\gamma}}}_2(\sigma ) = { { \mathbf j}}\cdot{{\bm{\gamma}}}(\sigma ) = [ v_1 + v_2\cos(u_2\sigma)]\sin(u_1\sigma),\\ { { \bm{\gamma}}}_3(\sigma ) = { { \mathbf k}}\cdot{{\bm{\gamma}}}(\sigma ) = v_3\sin(u_2\sigma),\\ { { { \mathbf c}}}({{\underline{s } } } ) = { { \bm{\gamma}}}\bigl(\sigma({{\underline{s}}})\bigr)\end{gathered}\ ] ] where the euclidean arc parameter @xmath319 satisfies @xmath466 the torus knot @xmath462 is approximated by a multiply - wound circle of euclidean radius @xmath467 .
the type @xmath468 torus knot is shown in figure [ figure : knot ] and its frenet curvature and torsion are shown in figures [ figure : curvature ] and [ figure : torsion ] . choosing mks units , so @xmath469 , this particular knot has @xmath470 and so @xmath471 . the classical sagnac beat frequency ( [ standard_sagnac_beat_frequency ] ) of a circular
ne wavetube laser ( @xmath472 ) and `` radius '' @xmath473 fixed on the earth ( @xmath474 ) and centred on the poles is @xmath475 .
clearly , for @xmath476 and @xmath400 the standard sagnac and torsion contributions to the beat frequency are quite similar and , moreover , in order to determine @xmath275 from the beat frequency both contributions would have to be taken into account .
a particular approximation scheme has been developed that permits one to determine the electromagnetic mode structure for a rotating slender wavetube containing a non - conducting isotropic homogeneous dispersionless medium .
the possible effects of a weak non - newtonian gravito - magnetic field on these modes has been included .
the approximation enables one to identify te and tm type field configurations with respect to the longitudinal axis of the wavetube .
although this need not be fixed in space , provided the dimensionless parameters @xmath477 and @xmath288 are smaller than @xmath478 the decomposition remains valid to the order prescribed . in this regime
one finds that such a wavetube with non - zero integrated frenet torsion can produce a modification to the classical sagnac beat frequency due to any rotation of the interferometer . to detect effects of gravito - magnetism in this scenario one must adjust the parameters @xmath479 so that the effects due to the dimensionless parameters @xmath480 can be distinguished experimentally from those produced by @xmath481 .
there are of course practical limitations that bound @xmath429 and @xmath238 and the possibility of increasing @xmath275 by rotating the interferometer .
furthermore the example given in section 7.1 shows that there exist slender geometrical structures based on space curves having small frenet curvature but integrated frenet torsion giving rise to interference effects commensurate with those produced by the rotation of the earth .
it is difficult to construct geometrical configurations that meet all the above competing constraints needed to reveal terrestrial gravito - magnetic effects without releasing the condition of relatively constant local frenet curvature . to overcome this limitation
it is clear that a more complete analysis of the mode spectra should take into account perturbations induced by variations of frenet curvature of the wavetube .
this is a non - trivial modification since it implies that the maxwell equations no longer separate into twisted te and tm modes .
however , for appropriate local curvature variations one may extend the approximation scheme using a perturbation approach and the basis of twisted modes found in this analysis .
this would yield a clear separation of the effects of frenet torsion and frenet curvature on the sagnac beat frequency .
such developments would provide a more complete picture of the delicate interplay between acceleration , geometry , electromagnetism and gravity that arises in any attempt to design an effective ring - laser capable of detecting terrestrial gravito - magnetism .
dab , an and rwt are most grateful for the hospitality provided by the department of physics and astronomy , university of canterbury , christchurch , new zealand and for valuable discussions with g stedman , r hurst and r reeves .
they are also grateful to m hamilton , university of adelaide , for information on optical polarising devices .
dab acknowledges financial support from the royal society , dlw from the marsden fund of the royal society of new zealand and rwt is grateful to bae systems for their support and interest in this research . | a new approximation scheme , designed to solve the covariant maxwell equations inside a rotating hollow slender conducting cavity ( modelling a ring - laser ) , is constructed .
it is shown that for well - defined conditions there exist te and tm modes with respect to the longitudinal axis of the cavity .
a twisted mode spectrum is found to depend on the integrated frenet torsion of the cavity and this in turn may affect the sagnac beat frequency induced by a non - zero rotation of the cavity .
the analysis is motivated by attempts to use ring - lasers to measure terrestrial gravito - magnetism or the lense - thirring effect produced by the rotation of the earth . |
since the recent discovery of the carbon nanotubes by iijima @xcite there has been a significant progress @xcite in the studies of the conducting properties of both single - walled @xcite and multi - walled @xcite carbon nanotubes .
conductance of a mesoscopic system connected to metallic reservoirs is well understood and is usually described by the landauer formula @xcite . for quantum point contacts in semiconductor structures and in metallic nanowires
it is well establish experimentally that the differential conductance to a good approximation is quantized in units of @xmath1 and at zero temperature given by @xmath2 where @xmath3 is the number of propagating modes . in carbon nanotubes with metallic contacts
most experiments show that the conductance is less than the conductance which one should expect for a smooth interface between tube and metal , e.g. @xmath4 for metallic single - walled tubes where the extra factor of 2 comes from the two @xmath5 bands that are crossing the fermi level @xcite .
the reasons for this lower conductance are still not fully known .
theoretically , several groups have considered the effects of vacancies @xcite , disorder @xcite , distortion @xcite , and doping @xcite on the conductance of carbon nanotubes .
the conducting properties have also been studied in e.g. the context of junctions between different metallic carbon nanotubes @xcite , aharonow
bohm effect in the presence of a magnetic field @xcite , and the luttinger liquid behavior of a one - dimensional gas of interacting electrons @xcite .
also the ideal `` hollow quantum cylinder '' , i.e. a two - dimensional electron gas on a cylinder , has been studied in context of the difference between strip - like wires and tubes @xcite . however , with the exception of the recent qualitative study of tersoff @xcite and the recent modeling - works of anantram _ et al .
_ @xcite and sanvito _ et al . _
@xcite , less attention has been focused on the conditions for a good transmission between tube and a metal contact which is an important issue for practical devices with carbon nanotubes , or other quantum tubes . in quantum point
contacts an adiabatic interface between the wire and reservoirs ensures a transmission coefficient close to unity @xcite .
the condition for adiabaticity is that the shape of the contact region varies slowly on the scale of the fermi wave length . in the opposite case with an abrupt interface ,
i.e. quasi - one - dimensional lead connected to a wide two - dimensional contact , szafer and stone @xcite found that the transmission rapidly increases to unity as the width of the confined region exceeds half of the fermi wavelength , thus giving a reflectionless contact . for the contact between a quantum tube and a three - dimensional metal it is not obvious that the assumption of an ideal reflectionless contact applies and the aim of this work is to study the contact resistance for this case .
the model we are studying is that of a hollow quantum cylinder of radius @xmath6 contacted by a three - dimensional free - electron metal which we for convenience model by a cylindrical wire with radius @xmath7 , see fig .
the system thus has full cylindrical symmetry and the angular momentum quantum number @xmath8 can therefore be used to label the scattering states . for the coupling of the quantum tube to the contact it is necessary to take a radial confinement potential for the quantum tube into account and here we model the confinement by an attractive delta - function potential .
as an example we apply this model to metal contacts of al or au ; the quantum tube parameters are chosen to mimic armchair carbon nanotubes : the strength of the confinement can be related to the work function for the material that constitutes the tube and in the case of a carbon nanotube we relate it to the work function of graphene .
it should be noted that the employed free electron model does not fully describe the actual band structure of carbon nanotubes .
nevertheless , a study of contact resistance within this idealized model should yield valuable insights which are relevant to real materials . the paper is organized as follows : in section ii the eigenstates of a quantum tube connected to a cylindrical metal contact are found . in section iii
these eigenstates are used to construct the scattering states to find the transmission coefficient , and hence the conductance of the contact . in section iv
, we apply our model to contacts between an armchair carbon nanotube and a metal . finally , in section v discussion and conclusions
are given .
essential details of analytical calculations are given in appendices a and b.
we separate the discussion into two parts : first we find the eigenstates in the tubular geometry and then the eigenstates for the cylindrical metal contact . in section
iii the matching of these eigenstates are used to construct the scattering states of the contact .
the quantum tube of radius @xmath6 with otherwise free electrons is modeled by the hamiltonian @xmath9 + v_{\scriptscriptstyle\rm t}(r),\ ] ] with a confining potential given by an attractive delta function potential @xmath10 where the confinement strength @xmath11 is taken positive .
the eigenstates of the schrdinger equation have the form @xmath12 with angular and longitudinal wave functions @xmath13 , \\
\psi _ { m}(z ) & = & \left [ k_{m}(e)\right]^{-1/2 } \exp \left [ \pm ik_{m}(e)z \right ] , \end{aligned}\ ] ] where the angular momentum quantum numbers @xmath8 are integers , @xmath14 ^{1/2}$ ] is the wave vector associated to the longitudinal free propagation , and @xmath15 is the total energy of the state . here , @xmath16 is the energy associated to the longitudinal propagation and @xmath17 is the ( binding ) energy associated to the transverse motion .
we can relate the strength of the confinement to the work function @xmath18 , which is the energy required to remove an electron at the fermi level ( disregarding surface charge effects ) .
the normalization @xmath19 ^{-1/2}$ ] is chosen such that the propagating modes carry the same amount of current .
the radial wave function @xmath20 satisfies @xmath21 \right\ } r_{m}(r)= -\gamma r_{\scriptscriptstyle\rm t } \delta(r - r_{\scriptscriptstyle\rm t})r_{m}(r),\ ] ] where @xmath22 is a dimensionless confinement strength . for the bound states ( @xmath23 ) and @xmath24 this equation has the form of bessel s modified differential equation @xcite .
the solutions are given by modified bessel functions of order @xmath8 of the first and second kind , so that the full solution is given by @xmath25 where @xmath26 ^{1/2}$ ] .
at @xmath27 , the radial wave function is continuous and the appropriate matching condition for the derivative @xmath28 at @xmath27 is found by integrating eq .
( [ schrodinger_radial ] ) from @xmath29 to @xmath30 . in this way
the matching conditions become @xmath31 and we get the following equation for the normalization coefficients @xmath32 non - trivial solutions exist if the determinant vanishes , and hereby the wave vector @xmath33 is a solution to the equation @xmath34 where the result for the wronskian @xmath35 has been used @xcite . expanding eq .
( [ kappa_m ] ) in the small-@xmath36 limit @xcite we find that a bound state with angular momentum @xmath37 exists for @xmath38 .
the number of bound states for a certain value of @xmath39 is given by @xmath40 where @xmath41 is the integer part of @xmath42 .
thus , there is always at least a single bound state corresponding to @xmath43 . from the matching conditions and the normalization of @xmath44 ( see appendix [ normalization ] )
, it follows that @xmath45 ^{-1/2}. \label{a}\end{aligned}\ ] ] a plot of the radial wave function is shown in the inset of fig .
3 . increasing the confinement strength ,
the radial wave function becomes more localized which is accompanied by an increase in the binding energy . for the metal it is convenient to assume a cylindrical geometry and consider a cylindrical wire of radius @xmath7 .
the hamiltonian is written as @xmath46 + v_{\scriptscriptstyle\rm c}(r),\ ] ] with a hard - wall confining potential @xmath47 obviously , for @xmath48 , @xmath49 and for @xmath50 the eigenstates have the form @xmath51 ,
\\ \psi _ { \nu m}(z ) & = & \left [ k_{\nu m}(e)\right ] ^{-1/2 } \exp \left [ \pm ik_{\nu m}(e)z\right ] , \end{aligned}\ ] ] where @xmath52 is a bessel function of the first kind of order @xmath8 , @xmath53 is a wave vector corresponding to the radial energy @xmath54 , @xmath55 ^{1/2}$ ] is the wave vector of the longitudinal motion , and @xmath56 is the total energy of the state . again
the normalization @xmath57 ^{-1/2}$ ] makes the propagating modes carry the same amount of current .
the boundary condition for the radial wave function leads to @xmath58 from which we find @xmath59 numerically . since @xmath60 for large @xmath42 @xcite
, we have @xmath61 with @xmath62 . the normalization @xmath63 is given by ( see appendix [ normalization ] ) @xmath64 which is a real number .
we consider an electron in the tube in mode @xmath8 incident on the contact ( see fig .
[ fig1 ] ) and compute the transmission and reflection coefficients .
we construct the scattering states in the basis of the eigenstates of the schrdinger equation ( see previous section ) . since the angular momentum is a conserved quantity , the transmitted and reflected parts of the wave function
also have the same quantum number @xmath8 . in the quantum tube ( @xmath65 )
the scattering state is given by @xmath66,\ ] ] and in the contact ( @xmath67 ) by @xmath68 here @xmath69 is the reflection amplitude for mode @xmath8 and @xmath70 is the corresponding transmission amplitude .
we assume that the effective electron mass is the same in the two materials so that the continuity of @xmath71 and @xmath72 at @xmath73 are appropriate boundary conditions . for carbon nanotubes and metals like al and au
this is a reasonable approximation ( @xmath74 @xcite ) . for general details on how to account for differences in the effective mass and the underlying symmetry of the lattice
we refer to refs . @xcite and references therein .
the boundary conditions lead to @xmath75 where @xmath76 , with the radial overlap defined as @xmath77 in addition we have the sum - rule @xmath78 , which can be used to verify the numerical convergence .
the overlap can be calculated analytically ( see appendix [ overlap ] ) and the squared overlap is given by @xmath79 ^ 2 } { \left[\left(\kappa_mr_{\scriptscriptstyle\rm t}\right)^2 + \left(\kappa_{\nu m}r_{\scriptscriptstyle\rm t}\right)^2\right]^2 \left[j_{m-1}(\kappa_{\nu m}r_{\scriptscriptstyle\rm c } ) j_{m+1}(\kappa_{\nu m}r_{\scriptscriptstyle\rm c})\right]}\\ & \quad&\times \frac{4 } { \left[k_m^2(\kappa_m r_{\scriptscriptstyle\rm t } ) i_{m-1}(\kappa_m r_{\scriptscriptstyle\rm t } ) i_{m+1}(\kappa_m r_{\scriptscriptstyle\rm t } ) -i_m^2(\kappa_m r_{\scriptscriptstyle\rm t } ) k_{m-1}(\kappa_m r_{\scriptscriptstyle\rm t } ) k_{m+1}(\kappa_m r_{\scriptscriptstyle\rm t})\right ] } \nonumber.\end{aligned}\ ] ] the total transmission from mode @xmath8 in the quantum tube into the contact is thus @xmath80 where @xmath81 projects onto the propagating modes ( @xmath82 real ) of the metal contact .
here we have assumed that the lengths of the quantum tube and the contact are semi - infinite so that tunneling through evanescent modes can be neglected .
these should be included in the case of two metal contacts connected by a quantum tube of finite length . introducing real and imaginary parts by @xmath83
we obtain @xmath84 the reflection probability @xmath85 can be calculated in a similar manner which provides us with the usual sum - rule @xmath86 , ensuring the conservation of probability current density . to summarize
, the transmission probability of mode @xmath8 can be calculated from eq .
( [ t_m ] ) with @xmath87 ^{1/2}$ ] and @xmath88 ^{1/2}$ ] , with @xmath89 being the energy of the @xmath8th transverse mode in the tube and similarly @xmath90 is the transverse energy in the contact . here
@xmath91 and @xmath92 are solutions to @xmath93 and @xmath94 , respectively . for a numerical implementation , an upper cut - off @xmath95 in the sum over modes in the contact
is needed and the sum - rule for the squared radial overlap is then a measure of the numerical convergence for a given cut - off .
when choosing values for the confinement parameters , @xmath39 and @xmath96 , we take into account that the fermi momenta of the quantum tube and the metal can be different since the relevant electrons are located at the fermi levels of the two materials . for the metal contact we use the known fermi energies for e.g. al and au to relate the confinement potential to the fermi level as @xmath97 , with
the fermi energy being defined positive . for the tube ,
the fermi energy enters as @xmath98 .
when the two materials are brought into contact the chemical potentials align , but the difference in fermi wave vectors remains .
thus for the metal contact @xmath99^{2}/2m_e$ ] , and for the tube @xmath100^{2}/2m_e$ ] . for the tube
we need to specify @xmath101 , which follows from the work function @xmath102^{2}/2m_e$ ] .
we have neglected the charge density induced at the interface by a mismatch of the work functions . for a discussion of this in the context of the screening properties of one - dimensional systems ,
see e.g. ref . @xcite .
the @xmath103 armchair single - walled carbon nanotube can be regarded as the result of rolling one sheet of graphite ( with the carbon atoms in a hexagonal lattice ) in the direction of one of the bonds @xcite .
the resulting tube has a periodicity @xmath104 along the tube axis ( @xmath105-axis ) and a radius @xmath106 with @xmath107 atoms along the perimeter , arranged in two rows that resemble a chain of armchairs , see fig .
their metallic character is caused by two @xmath5 bands crossing the fermi level at a wave vector @xmath108 . as discussed recently by tersoff
@xcite the metallic armchair carbon nanotubes have electrons at the fermi level which can be regarded as having an angular momentum quantum number @xmath43 . in order to apply our simple model to the problem of the contact resistance of @xmath103 single - walled carbon nanotubes embedded in a free - electron metal we notice that @xmath109 . in fig . [ fig3 ] the transmission probability at the fermi level is shown for several values of @xmath110 corresponding to @xmath103 armchair carbon nanotubes for various values of the dimensionless confinement strength @xmath101 . in the particular case of an al contact ,
the mismatch is given by @xmath111 and the corresponding transmission is presented in panel ( a ) of fig .
[ fig4 ] . in panel
( b ) we show similar results for an au contact for which @xmath112 . in order to estimate @xmath101
we relate it to the work function of the carbon nanotube which is of the order @xmath113 . for the quantum tube we associate a work function to the @xmath43 bound state via its binding energy , i.e. @xmath114 where @xmath115 is the solution to @xmath116 . in fig .
[ fig5 ] this work function is shown as a function of the confinement strength for quantum tubes corresponding to @xmath103 armchair carbon nanotubes . from this
we estimate that @xmath117 is a reasonable regime for the curves shown in figs .
[ fig3 ] and [ fig4 ] .
this means that a transmission close to unity ( per band at the fermi level ) can be expected if the armchair nanotube is embedded in free - electron metals like al or au . assuming a work function @xmath118 of the nanotube we have used the corresponding value of @xmath101 to calculate the transmission for the different nanotubes . for @xmath103 armchair nanotubes with al and au contacts we find @xmath119 and @xmath120 , respectively . in the case of matching fermi wave vectors ( @xmath121 ) we get @xmath122 . for the considered tubes ( @xmath123 ) ,
these transmissions are found to be almost independent of the specific value of the tube indices @xmath103 .
here we have only taken the geometry - related contact scattering into account .
physically , a lower transmission can be caused by electrons being scattered by interface imperfections / roughness , deviations from a spherical fermi surface of the metal contact , and scattering due to non - matching work functions of the nanotube and metal . also scattering due to the non - matching fermi velocities of the nanotube and the metal
could be expected . however , as shown in fig .
[ fig4 ] , a mismatch between fermi wave vectors can actually in some cases increase the transmission ( and thereby the conductance ) due to quantum interferences , even though the mismatch by itself is known to give rise to momentum relaxation and thereby resistance .
we have considered the contact resistance ( in terms of transmission ) of a quantum tube embedded in a free - electron metal . for the quantum tube
we have modeled the radial confinement of the electron motion by an attractive delta function potential which gives rise to at least one bound state in the radial direction .
the strength of the attractive potential can phenomenologically be associated to the work function of the quantum tube . within this model
we have calculated the transmission of a quantum tube contacted by a free - electron metal .
due to the cylindrical geometry of the contact , considerable analytical progress was possible and with the resulting equations the scattering problem is readily solved numerically . as an application we have considered the transparency of contacts with armchair carbon nanotubes embedded in free - electron metals .
our calculations show that in the absence of scattering mechanisms associated to e.g. interface imperfections / roughness , deviations from a spherical fermi surface of the metal contact , and scattering due to non - matching work functions of the nanotube and metal , the geometry itself allows for a high transparent contact between armchair carbon nanotubes and free - electron metal contacts .
furthermore , from this simple model we find that al would be a good candidate for such a metal as it was suggested recently by tersoff @xcite . for au however
, we find that the present 3d geometry allows for good contact in contrast to tersoff s findings for au , which were based on 1d considerations .
we would like to thank m. brandbyge , h. bruus , d.h .
cobden , and j. nygrd for useful discussions .
from the radial wave function of the quantum tube , eq .
( [ wavefunction_exact ] ) , it follows that the normalization is given by @xmath124^{-1/2},\ ] ] and since @xcite @xmath125/2 , \\
\int{\rm d}\alpha\,\alpha k_m^2(\alpha)&= & { \alpha}^2 \left[k_m^2({\alpha})-k_{m-1}({\alpha})k_{m+1}({\alpha})\right]/2,\end{aligned}\ ] ] we get the result in eq . ( [ a ] ) . similarly , from the radial wave function of the free - electron metal contact , eq . ( [ r_contact ] )
, it follows that the normalization is given by @xmath126^{-1/2},\ ] ] and since @xmath127 and @xcite @xmath128/2,\ ] ] we obtain eq .
( [ c ] ) .
the overlap of radial wave functions can be written as @xmath129 \nonumber \\ & = & \frac{a_m c_{\nu m}}{\kappa_m^2 + \kappa_{\nu m}^2 } \frac { j_m(\kappa_{\nu m}r_{\scriptscriptstyle\rm t } ) + \kappa_{\nu m}r_{\scriptscriptstyle\rm c } i_m(\kappa_mr_{\scriptscriptstyle\rm t } ) k_m(\kappa_mr_{\scriptscriptstyle\rm c } ) j_{m+1}(\kappa_{\nu m } r_{\scriptscriptstyle\rm c } ) } { k_m(\kappa_m r_{\scriptscriptstyle\rm t})},\end{aligned}\ ] ] s.j .
tans , m.h .
devoret , h. dai , a. thess , r.e .
smalley , l.j .
geerligs , and c. dekker , nature * 386 * , 379 ( 1997 ) ; m. bockrath , d.h .
cobden , p.l .
mceuen , n.g .
chopra , a. zettl , a. thess , and r.e .
smalley , science * 275 * , 1922 ( 1997 ) ; j.e .
fischer , h. dai , a. thess , r. lee , n.m .
hanjani , d.l .
dehaas , and r.e .
smalley , phys .
b * 55 * , r4921 ( 1997 ) ; a.yu .
kasumov , r. deblock , m. kociak , b. reulet , h. bouchiat , i.i .
khodos , yu.b .
gorbatov , v.t .
volkov , c. journet , and m. burghard , science * 284 * , 1508 ( 1999 ) .
ebbesen , h.j .
lezec , h. hiura , j.w .
bennett , h.f .
ghaemi , and t. thio , nature * 382 * , 54 ( 1996 ) ; a. bachtold , m. henny , c. terrier , c. strunk , c. schnenberger , j .-
salvetal , j .-
m . bonard , and l. forr , appl .
phys . lett . * 73 * , 274 ( 1998 ) ; s. frank , p. poncharal , z.l .
wang , and w.a .
de heer , science * 280 * , 1744 ( 1998 ) .
see e.g. r. egger and a.o .
gogolin , phys .
lett . * 79 * , 5082 ( 1997 ) ; eur .
j. b * 3 * , 281 ( 1998 ) ; h. yoshioka and a.a .
odintsov , phys .
lett . * 82 * , 373 ( 1999 ) ; m. bockrath , d.h .
cobden , j. lu , a.g .
rinzler , r.e .
smalley , l. balents , and p.l .
mceuen , [ cond - mat/9812233 ] . | we consider the conductance of a quantum tube connected to a metallic contact . the number of angular momentum states that the tube can support depends on the strength of the radial confinement .
we calculate the transmission coefficients which yield the conductance via the landauer formula , and discuss the relation of our results to armchair carbon nanotubes embedded in a metal . for al and au contacts and tubes with a realistic radial confinement
we find that the transmission can be close to unity corresponding to a contact resistance close to @xmath0 per band at the fermi level in the carbon nanotube .
[ firstpage ] |
let @xmath9 be the manifold obtained by @xmath10-dehn surgery on a knot @xmath11 in @xmath4 .
( here @xmath12 , @xmath13 , parametrizes the slope on the boundary of the exterior of @xmath11 that runs @xmath14 times meridionally and @xmath15 times longitudinally around @xmath11 . )
we denote the core of the attached solid torus in @xmath16 by @xmath1 .
thus @xmath1 is a knot in @xmath3 with a dehn surgery that yields @xmath4 , and it is natural to investigate what properties of @xmath1 this entails . in the present paper
we are interested in the relationship between @xmath1 and the heegaard splittings of @xmath3 ; more specifically , if @xmath17 is a heegaard surface for @xmath3 , what can we say about the bridge number @xmath18 of @xmath1 with respect to @xmath17 ?
first note that if @xmath17 is a heegaard surface for @xmath19 , the exterior of @xmath11 , i.e. @xmath17 separates @xmath20 into a handlebody and a compression body , then @xmath17 becomes a heegaard surface for @xmath3 such that @xmath21 .
in general not every heegaard surface for @xmath3 arises in this way ; for example the heegaard genus of @xmath3 may be smaller than that of @xmath20 .
nevertheless , it is shown in @xcite that , given @xmath11 , for all but finitely many slopes @xmath10 we have @xmath21 for any heegaard surface @xmath17 of @xmath3 .
recall that if @xmath22 are two isotopy classes of simple closed curves on a 2-torus , then the _ distance _ between @xmath23 and @xmath24 , denoted @xmath25 is defined to be the absolute value of the intersection number between @xmath23 and @xmath24 .
note that @xmath26 , the distance of @xmath10 from the meridian of @xmath11 on @xmath27 . since the trivial dehn surgery @xmath28 represents the maximal possible degeneration of heegaard genus , one would expect the heegaard splittings of @xmath16 to behave better as @xmath15 gets large .
indeed , it follows from @xcite that if @xmath11 is hyperbolic and @xmath17 is a heegaard surface of genus @xmath5 for @xmath16 , then @xmath29 implies @xmath21 .
here we consider the question whether there is an upper bound on @xmath30 that depends only on the genus of @xmath17 : is there a function @xmath31 such that if @xmath11 is a knot in @xmath4 and @xmath17 is a heegaard surface of genus @xmath5 for @xmath9 , where @xmath13 , then @xmath32 ?
it turns out that the answer to question 1 is `` no '' in general when @xmath33 ; see remark [ rem1 - 6 ] below .
however , corollary [ cor1 - 1 ] below says that the answer is `` yes '' if @xmath34 , provided that @xmath11 is hyperbolic , @xmath3 is not a special kind of seifert fibered space , and @xmath10 is not a boundary slope for @xmath20 .
corollary [ cor1 - 1 ] follows from our main result , theorem 8.1 , where we consider more generally the number of intersections of @xmath1 with a thick level surface in a heegaard splitting of @xmath3 with respect to which @xmath1 is in thin position . there is a linear function @xmath31 with the following property .
let @xmath11 be a hyperbolic knot in @xmath4 , @xmath9 where @xmath34 , and @xmath1 the core of the attached solid torus in @xmath3 .
suppose @xmath1 is in thin position with respect to a genus @xmath5 heegaard splitting of @xmath3 and let @xmath17 be a corresponding thick level surface . if @xmath17 is a strongly irreducible heegaard surface for @xmath3 then either * @xmath35 ; or * @xmath3 is toroidal ; or * @xmath3 is a seifert fibered space over the 2-sphere with exactly three exceptional fibers , at least one of which has order 2 or 3 .
furthermore , in cases ( 2 ) and ( 3 ) @xmath3 has a genus 2 heegaard splitting with respect to which @xmath1 has bridge number @xmath36 in case ( 2 ) and at most @xmath37 in case ( 3 ) .
in [ sec : proofofmain ] we show that @xmath38 works .
the technical assumption in theorem [ thm : main ] that @xmath17 is strongly irreducible is likely unnecessary .
it is only used in the proof of lemma [ lem : parallelesc ] .
if @xmath20 is a compact , orientable 3-manifold with torus boundary let us say that a slope @xmath39 on @xmath27 is a _
@xmath40-boundary slope for @xmath20
_ if there is a compact , connected , orientable essential surface @xmath41 with boundary slope @xmath39 , such that @xmath42 has genus at most @xmath5 and at most @xmath43 boundary components .
we say that @xmath39 is a _
@xmath5-boundary slope _ if it is a @xmath40-boundary slope for some @xmath43 .
[ cor1 - 1 ] there is a linear function @xmath31 with the following property .
let @xmath11 be a hyperbolic knot in @xmath4 , @xmath9 where @xmath34 , and @xmath1 the core of the attached solid torus in @xmath3 .
let @xmath17 be a genus @xmath5 , strongly irreducible heegaard surface in @xmath3 .
if @xmath10 is not a @xmath44-boundary slope for the exterior of @xmath11 then either @xmath32 , or conclusion ( 3 ) of theorem 8.1 holds . for terminology on thin presentations ,
see section [ sec : basics ] .
let @xmath45 be the function of theorem [ thm : main ] .
assume @xmath16 has a genus @xmath5 splitting which is strongly irreducible , and that @xmath10 is not a @xmath44-boundary slope .
in particular , @xmath3 can not be toroidal since , otherwise , by @xcite @xmath10 is a @xmath46-boundary slope ( and @xmath47 ) .
put @xmath1 in thin position with respect to this splitting and let @xmath17 be a thick level surface .
theorem [ thm : main ] says that either @xmath3 is the seifert fibered space of its conclusion @xmath48 and that @xmath1 has bridge number at most @xmath37 with respect to a genus @xmath49 splitting of @xmath3 , or @xmath50 .
this thin presentation of @xmath1 must be a bridge presentation of @xmath1 , otherwise there will be a thin level surface in the thin presentation which intersects @xmath1 fewer times than does @xmath17 , and lemma [ lem : thin = bridge ] would contradict that @xmath10 is not a @xmath44-boundary slope .
thus @xmath1 is bridge with respect to @xmath17 and @xmath32 .
[ rem1 - 2 ] by @xcite , at most finitely many slopes of the exterior of @xmath11 are boundary slopes .
note that if @xmath3 is non - haken , then @xmath10 can not be a boundary slope , by theorem 2.0.3 of @xcite ( @xmath51 and @xmath3 is irreducible by @xcite ) . if @xmath3 is a seifert fibered space over the 2-sphere with at most three exceptional fibers , then @xmath3 is non - haken .
[ rem1 - 3 ] it is conjectured that a seifert fibered space can never be obtained by non - integral dehn surgery on a hyperbolic knot in @xmath4 that is , that ( 3 ) of theorem 8.1 never occurs .
theorem 2.4 of @xcite implies that if @xmath3 is seifert fibered over the 2-sphere with three exceptional fibers and @xmath52 then @xmath1 has bridge number at most 1 with respect to some genus 2 heegaard splitting of @xmath3 , though possibly a 1-sided splitting ( which are not considered in this paper ) .
note that since a 1-sided splitting corresponds to a non - orientable surface , if @xmath14 is odd then @xmath16 does not have a 1-sided splitting .
thus when @xmath52 and @xmath14 is odd , theorem 2.4 of @xcite sharpens conclusion ( 3 ) above .
[ rem1 - 4 ] if @xmath3 is toroidal , then by @xcite ( @xmath53 and ) @xmath11 belongs to the family of knots described by eudave - muoz in @xcite . in @xcite
eudave - muoz shows that these knots have tunnel number 1 , and hence @xmath54 with respect to a minimal genus ( genus 2 ) splitting of @xmath3 .
[ rem1 - 5 ] we conjecture that the hypotheses that @xmath11 is hyperbolic in theorem 8.1 and corollary [ cor1 - 1 ] , and that @xmath10 is not a @xmath44-boundary slope in corollary [ cor1 - 1 ] , are redundant .
[ rem1 - 6 ] when @xmath33 the answer to question 1 is `` no '' in general . in @xcite teragaito
constructs infinitely many hyperbolic knots @xmath55 in @xmath4 such that 4-surgery on @xmath55 gives the same 3-manifold @xmath3 for all @xmath56 , where @xmath3 is a seifert fibered space of type @xmath57 .
in particular @xmath3 has heegaard genus 2 .
let @xmath58 be the core of the attached solid torus in @xmath59 , and define @xmath60 a genus 2 heegaard surface for @xmath61 .
it is shown in @xcite that @xmath62 is unbounded .
[ rem1 - 7 ] for small values of @xmath5 the bridge number @xmath63 is either known or conjectured to be very small .
for example , the impossibility of getting @xmath4 by non - trivial dehn surgery on a non - trivial knot @xcite can be expressed as saying that if @xmath64 and @xmath65 then @xmath66 . when @xmath67 , @xmath16 is a lens space and here the cyclic surgery theorem @xcite says that if @xmath68 then @xmath11 is a torus knot , which is easily seen to imply @xmath69 , while if @xmath70 and @xmath11 is hyperbolic the berge conjecture is equivalent to the assertion that @xmath71 . in another paper ( ( *
* theorem 2.4 ) ) , we consider the case @xmath72 and show that if @xmath52 then , generically , @xmath73 ( with respect to some genus 2 splitting , possibly 1-sided ) .
strongly irreducible heegaard surfaces and closed incompressible surfaces may exhibit similar behavior . with this theme in mind , in the appendix we adapt the proof of theorem [ thm : main ] to bound the intersection number of @xmath1 with an incompressible surface rather than a thick heegaard surface in terms of the genus of the surface : there is a linear function @xmath74 with the following property .
let @xmath11 be a hyperbolic knot in @xmath4 , @xmath9 where @xmath34 , and @xmath1 the core of the attached solid torus in @xmath3 .
let @xmath17 be an orientable , incompressible surface in @xmath3 of genus @xmath5 .
then @xmath1 can be isotoped to intersect @xmath17 at most @xmath75 times .
osoinach shows that for integral surgeries the above does not hold . in @xcite
, he gives examples of infinitely many different knots in the 3-sphere on which @xmath36-surgery gives the same manifold .
this manifold has an essential torus , and in @xcite he shows that the set of minimal intersection numbers in @xmath3 of this torus with the corresponding cores of the attached solid tori must be infinite .
our results give information on the relationship between the heegaard genus of @xmath3 and that of @xmath20 , the exterior of @xmath11 .
recall that a heegaard splitting of @xmath20 is a decomposition @xmath76 , where @xmath77 is a handlebody with @xmath78 and @xmath79 is a compression body with @xmath80 .
the _ heegaard genus _
@xmath81 of @xmath20 is the minimal genus of @xmath17 over all such decompositions . in this context one
often talks about the _ tunnel number _ @xmath82 of @xmath11 , the minimum number of arcs ( `` tunnels '' ) that need to be attached to @xmath11 so that the complement of an open regular neighborhood of the resulting 1-complex is a handlebody .
it is easy to see that @xmath83 . for any slope @xmath10
, @xmath84 is a heegaard splitting of @xmath85 ; in particular @xmath86 . the question arises as to what extent heegaard genus can decrease under dehn filling , i.e. how bad the inequality @xmath87 can be . since it is easy to see that if @xmath1 has bridge number @xmath63 with respect to a genus @xmath5 splitting of @xmath3 then @xmath88 , corollary [ cor1 - 1 ] gives a linear bound on the extent to which heegaard genus can decrease under a non - integral dehn surgery .
[ cor1 - 8 ] there is a linear function @xmath89 with the following property .
let @xmath11 be a hyperbolic knot in @xmath4 and @xmath90 . if @xmath3 has heegaard genus @xmath5 , @xmath10 is not a @xmath91-boundary slope for the exterior of @xmath11 , and @xmath92 , then the tunnel number of @xmath11 is at most @xmath93 .
we conjecture that there is always such a universal bound on the tunnel number of a knot in terms of the heegaard genus of any of its ( non - trivial dehn surgeries ) : [ conj1 - 9 ] there is a function @xmath89 such that if @xmath11 is a knot in @xmath4 and @xmath94 is a non - trivial dehn surgery on @xmath11 with heegaard genus @xmath5 , then the tunnel number of @xmath11 is at most @xmath93 .
information on the question of degeneration of heegaard genus under dehn filling in provided by rieck and sedgwick in @xcite and @xcite . as mentioned above , it is shown in @xcite that for all but finitely many slopes @xmath10 , @xmath95 with respect to any heegaard surface @xmath17 of @xmath3 .
taking @xmath17 to have minimal genus , it is easy to see that @xmath96 implies that either @xmath97 or @xmath98 .
see @xcite for details .
by @xcite , the second possibility can happen for only a finite number of lines of slopes ( where a _ line _ is a set of slopes @xmath39 such that @xmath99 for some fixed slope @xmath100 ) . regarding corollary [ cor1 - 8 ] and conjecture [ conj1 - 9 ] , in fact we know no examples where the heegaard genus of @xmath16 ( @xmath65 ) is less than @xmath82 .
so we ask is @xmath101 for all @xmath65 ? for example , the answer to question 2 is known to be `` yes '' when * @xmath102 ( @xcite ) * @xmath103 and @xmath104 ( @xcite ) * @xmath105 , @xmath52 , and @xmath16 does not contain an incompressible surface of genus 2 .
( @xcite ) finally , the bound on bridge number in corollary [ cor1 - 1 ] allows us to use a result of tomova @xcite to get a statement about the distance of splittings of exteriors of knots with genus @xmath5 dehn surgeries .
if @xmath17 is a heegaard surface for some 3-manifold , we denote by @xmath106 the _ distance _ of the corresponding splitting ; see @xcite .
[ cor1 - 10 ] there is a linear function @xmath107 with the following property .
let @xmath11 be a hyperbolic knot in @xmath4 whose exterior has a genus @xmath5 heegaard surface @xmath17 with @xmath108 . if @xmath10 is not a @xmath109-boundary slope for the exterior of @xmath11 and @xmath51 , then @xmath110 .
thus the distance of a splitting of a knot exterior is putting a limit on the degeneration of heegaard genus under dehn filling .
for instance , this applies to the examples of @xcite . recall that a knot has only finitely many boundary slopes ( @xcite ) . by @xcite , for any @xmath111 , there are knots @xmath11 in @xmath4 whose exteriors have genus @xmath5 heegaard splittings @xmath17 with @xmath108 , in fact with @xmath106 arbitrarily large ( such knots are necessarily hyperbolic ) .
( the case @xmath72 was first done in @xcite . )
corollary [ cor1 - 10 ] says that for such a knot @xmath11 , if @xmath51 and @xmath10 is not a @xmath109-boundary slope , then @xmath16 has heegaard genus @xmath5 .
let @xmath11 , @xmath10 , @xmath17 be as in the hypothesis .
set @xmath112 .
assume for contradiction that @xmath113 .
by corollary [ cor1 - 1 ] , the bridge number of @xmath1 with respect to some genus @xmath114 heegaard surface @xmath115 of @xmath16 is at most @xmath116 .
thus @xmath1 can be put in bridge position with respect to @xmath115 so that @xmath117
. then @xmath118 by assumption , and the main result of @xcite implies that , in @xmath16 , @xmath115 is isotopic to a stabilization of @xmath17 . but
@xmath115 has smaller genus than @xmath17 . throughout this article @xmath3
will be the manifold @xmath16 obtained by @xmath10-dehn surgery on a hyperbolic knot @xmath11 in @xmath4 , with @xmath34 .
it follows that * @xmath3 is irreducible @xcite ; * @xmath3 is not a lens space ( or @xmath4 ) @xcite ; * @xmath3 does not contain a klein bottle @xcite , @xcite . note
that ( 1 ) and ( 2 ) together imply that @xmath3 does not contain a projective plane .
we also assume * @xmath3 is atoroidal .
the proof of theorem [ thm : main ] occupies the remainder of this article and culminates in [ sec : proofofmain ] .
we give an overview , taking the notation from theorem [ thm : main ] .
we begin in [ sec : fatvertexedgraphs ] by briefly reviewing and setting up notation for the notions of thin position and fat - vertexed graphs of intersection between ( punctured ) heegaard spheres of @xmath4 and heegaard surfaces of @xmath3 .
in particular , @xmath119 and @xmath42 are surfaces in the exterior of @xmath11 coming from a heegaard sphere , @xmath120 , in @xmath4 and the given thick level heegaard surface , @xmath121 , in @xmath3 .
@xmath122 are the graphs of intersection between @xmath123 .
let @xmath124 be the number of components of @xmath125 .
then @xmath126 is also the number of vertices of @xmath127 and consquently the number of different labels on the vertices of @xmath128 .
our goal is to bound @xmath126 by some linear function of @xmath5 .
lemma [ lem:2 ] shows the existence of a special subgraph , @xmath129 , on @xmath128 called a great @xmath5-web . for each label @xmath130 of this graph
, @xmath131 is the subgraph of @xmath129 consisting of those edges with label @xmath130 .
we consider the set @xmath132 of all labels , @xmath130 , for which @xmath131 has a bigon or trigon face . in proposition
[ prop : webcount ] we make a key estimate showing that up to an additive linear function in @xmath5 , @xmath133 .
a bigon or trigon of @xmath131 gives rise to a bigon or trigon subgraph in @xmath129 .
we classify these subgraphs ( see @xmath134 ) as either extended scharlemann cycles or as trigons of type i , ii , iii .
the extended scharlemann cycles give rise to _ long mbius bands _ and _ long twisted @xmath135bands _ embedded in @xmath3 which intersect @xmath136 in essential curves ( or essential @xmath135-curves ) .
we choose a minimal collection , @xmath137 , of extended scharlemann cycles so that each label of @xmath132 that corresponds to an extended scharlemann cycle appears as a label in some element of @xmath137 .
let @xmath138 be the set of all labels appearing in the extended scharlemann cycles of @xmath137 .
those labels of @xmath132 that do not appear in @xmath138 then correspond to trigons of type i , ii , or iii .
section [ sec : typeiorii ] analyzes trigons of type i or ii and culminates in theorem [ thm : type12boundn ] , which bounds , in terms of a linear function in @xmath5 , those labels appearing in trigons of type i or ii but not in @xmath138 ( the collection of these labels is called @xmath139 ) .
section [ sec : typeiii ] shows that those labels appearing only in type iii trigons , up to a linear function in @xmath5 ( lemma [ lem : boundwrapping ] ) , correspond to vertices in trivial curves on @xmath136 called simple gnarls . the collection of such labels is called @xmath140 .
that is , up to an added linear function in @xmath5 , we have that @xmath141 is @xmath142 .
let @xmath143 be those labels of @xmath128 that are not in @xmath132 .
the statements of this paragraph * will all be up to an added linear function of @xmath5*. the curves of intersection , denoted @xmath144 , of @xmath136 with the long mbius bands or long twisted @xmath135-bands coming from @xmath137 , bound a collection of disjoint annuli of @xmath136 .
each curve of @xmath144 corresponds to two elements of @xmath138 , so there are @xmath145 such annuli . furthermore , half of these annuli must contain in their interiors vertices of @xmath127 which correspond to labels in @xmath128 which are in either @xmath143 or @xmath140 .
this is lemma [ lem : builda ] , whose main ingredient is lemma [ lem : triplets ] . this sub - collection of annuli is called @xmath146 , and from the above we get @xmath147 . on the other hand , it follows from lemmas [ lem : nonemptygnarl ] and [ 3ingnarl ] that each simple gnarl must contain at least three vertices of @xmath127 corresponding to labels in @xmath143 .
counting those elements of @xmath143 coming from from the interiors of the annuli in @xmath146 and the gnarls in @xmath140 we estimate ( equation sequence ( * * ) in section [ sec : proofofmain ] ) that @xmath148 .
thus @xmath149 , where @xmath150 is a linear function of @xmath5 , thereby bounding @xmath126 by @xmath151 .
the full proof of theorem [ thm : main ] is given at the beginning of [ sec : proofofmain ] .
in the course of this proof , we keep track of and explicitly determine suitable linear bounds . in the course of this work kb was partially supported by nsf grant dms-0239600 , by the university of miami 2011 provost research award , and by a grant from the simons foundation ( # 209184 to kenneth baker ) .
kb would also like to thank the department of mathematics at the university of texas at austin for its hospitality during his visits .
these visits were supported in part by nsf rtg grant dms-0636643 .
in this paper , a heegaard splitting will always be a 2-sided heegaard splitting .
given such a heegaard surface @xmath17 of a closed @xmath2-manifold @xmath152 there is a product @xmath153 so that @xmath154 and the complement of the product is the union of the cores of the two handlebodies .
this defines a height function on the complement of the the cores of the handlebodies .
consider all the circles @xmath155 embedded in the product that are morse with respect to the height function and represent the knot type of a particular knot @xmath156 .
the following terms are all understood to be taken with respect to the heegaard splitting .
following @xcite ( see also @xcite ) , the _ width _ of an embedded circle @xmath155 is the sum of the number of intersections @xmath157 where one regular value @xmath158 is chosen between each pair of consecutive critical values .
the _ width _ of a knot @xmath156 is the minimum width of all such embeddings .
an embedding realizing the width of @xmath156 is a _ thin position _ of @xmath156 , and @xmath156 is said to be _
thin_. if the critical point immediately below @xmath158 is a local minimum and the critical point immediately above @xmath158 is a local maximum , then the level @xmath159 is a _
thick level_. if the critical point immediately below @xmath158 is a local maximum and the critical point immediately above @xmath158 is a local minimum , then the level @xmath159 is a _
thin level_. the minimal number of local maxima among morse embeddings of @xmath155 is the _ bridge number _ of @xmath156 , and denoted @xmath160 , or , if @xmath17 is understood , @xmath161 .
an embedding realizing the bridge number of @xmath156 may be ambient isotoped so that all local maxima lie above all local minima , without introducing any more extrema .
the resulting embedding is a _ bridge position _ of @xmath156 , and @xmath156 is said to be _
bridge_. with @xmath156 in bridge position ,
the arcs of @xmath156 intersecting a heegaard handlebody are collectively @xmath162-parallel .
there is an embedded collection of disks in the handlebody such that the boundary of each is formed of one arc on @xmath17 and one arc on @xmath156 .
a single such disk is called a _ bridge disk _ for that arc of @xmath156 , and the arc is said to be _
bridge_. a thin position for a knot may have smaller width than that of its bridge position , with respect to the same heegaard splitting .
that is , thin position may not be bridge position .
however , this only happens when the meridian of the knot in the ambient manifold is a boundary slope of the knot exterior .
recall [ def:2bdryslope ] if @xmath163 is a compact , orientable 3-manifold with torus boundary , a slope @xmath39 on @xmath164 is a _
@xmath40-boundary slope for @xmath163
_ if there is a compact , connected , orientable essential surface @xmath165 with boundary slope @xmath39 , such that @xmath42 has genus at most @xmath5 and at most @xmath43 boundary components .
we say that @xmath39 is a @xmath5-boundary slope if it is a @xmath40-boundary slope for some @xmath43 .
[ lem : thin = bridge ] assume @xmath156 is a knot in a 3-manifold @xmath3 .
if @xmath156 has a thin position which is not a bridge position with respect to a genus @xmath5 heegaard splitting of @xmath3 , then the meridian of @xmath156 is a @xmath40-boundary slope for the exterior of @xmath156 where @xmath43 is the number of intersections of @xmath156 with any thin level surface in this thin presentation of @xmath156 .
this is proved in @xcite when @xmath64 .
the same proof works here .
we sketch it for the convenience of the reader .
let @xmath17 be the heegaard surface of a genus @xmath5 splitting of @xmath3 with respect to which @xmath156 is in thin position but not bridge position .
then this thin presentation must have a thin level , @xmath166 .
let @xmath43 the number of intersections of @xmath156 with the thin level surface @xmath167 .
there can be no bridge disks for @xmath156 to the thin level surface , else such a disk would give rise to a thinner presentation of @xmath156 .
maximally compress @xmath168 in the exterior of @xmath156 .
either some component of the result is an incompressible , @xmath169-incompressible surface of genus at most @xmath5 with at most @xmath43 boundary components each of which is a meridian of @xmath156 , or the result is a non - empty collection of boundary parallel annuli along with some closed surfaces .
but each boundary parallel annulus gives rise to a bridge disk of @xmath156 onto @xmath167 , which is not possible .
thus the meridian is a @xmath40-boundary slope for the exterior of @xmath156 .
[ lem : fandq ] assume @xmath11 is a non - trivial knot in @xmath4 with exterior @xmath170 .
let @xmath9 and @xmath1 be the core of the attached solid torus in @xmath3 .
suppose @xmath3 has a genus @xmath5 heegaard surface and that @xmath1 can not be isotoped to lie on this surface .
assume @xmath1 is in thin position with respect to this genus @xmath5 heegaard splitting of @xmath3 and let @xmath136 be a corresponding thick level surface .
then there is a punctured @xmath49-sphere @xmath119 and a punctured genus @xmath5 surface @xmath42 properly embedded and transverse in @xmath20 satisfying the following : 1 .
each component of @xmath171 is a meridian of @xmath27 and each component of @xmath125 is a @xmath10 curve on @xmath27 .
each arc of @xmath172 is essential in each of @xmath42 and @xmath119 .
3 . there are no simple closed curves of @xmath172 trivial in both @xmath42 and @xmath119 .
4 . capping off @xmath42 with disks in @xmath16 gives @xmath136 .
capping off @xmath119 with meridians of @xmath11 gives a 2-sphere , @xmath120 , in @xmath4 .
let @xmath173 be as stated .
let @xmath1 be the core of the attached solid torus in @xmath174 .
isotop @xmath1 to be in thin position with respect to the given genus @xmath5 splitting of @xmath3 so that @xmath136 is a thick level surface . in @xmath4 , put @xmath11 into thin position with respect to the genus @xmath36 heegaard splitting .
by theorem 6.2 of @xcite ( by assumption @xmath175 can not be isotoped onto their heegaard surfaces ) , there exists a thick level surface @xmath120 of @xmath4 such that each arc of @xmath172 is essential in each of @xmath176 and @xmath177 .
note that the argument of theorem 6.2 allows us to choose the fat layer with which we want to work . within the chosen fat layer ,
the intersection of any thick level surface with @xmath20 is properly isotopic in @xmath20 to the intersection of any other thick level surface with @xmath20 . as the exterior of @xmath11 is irreducible , after an isotopy in @xmath20 we may assume there are no simple closed curves of @xmath172 trivial in both @xmath42 and @xmath119 . in this section
we describe the graphs @xmath179 and the great web @xmath129 , which will be the context of the rest of the paper .
let @xmath180 be as in lemma [ lem : fandq ] .
on @xmath120 and @xmath136 form the _ fat vertexed graphs of intersection _ , @xmath128 and @xmath127 , respectively , consisting of the _ fat vertices _ that are the disks @xmath181 and @xmath182 and _ edges _ that are the arcs of @xmath172 .
choosing an orientation on @xmath183 , we may number the intersections of @xmath1 with @xmath136 , and hence the vertices of @xmath127 , from @xmath184 to @xmath185 in order around @xmath1 . similarly , if @xmath186 , by choosing an orientation on @xmath187 we may number the intersections of @xmath11 with @xmath120 and hence the vertices of @xmath128 from @xmath184 to @xmath188 in order around @xmath11 .
each component of @xmath189 intersects each component of @xmath190 a total of @xmath15 times .
thus a vertex of @xmath128 has valence @xmath191 and a vertex of @xmath127 has valence @xmath192 . since each component of @xmath193 is an endpoint of an arc of @xmath172 , each endpoint of an edge in @xmath128
may be labeled with the vertex of @xmath127 whose boundary contains the endpoint . thus around the boundary of each vertex of @xmath128 the labels @xmath194 appear in order @xmath15 times .
similarly around the boundary of each vertex of @xmath127 the labels @xmath195 appear in order @xmath15 times .
see the expository article @xcite for a more thorough discourse on such fat - vertexed graphs and standard techniques in their use .
endow each vertex of @xmath128 ( and @xmath127 ) with a sign of @xmath196 or @xmath197 according to whether or not the corresponding component of @xmath190 ( @xmath189 ) with its induced orientation is parallel or anti - parallel on @xmath198 to a chosen component of @xmath190 ( @xmath189 ) .
two vertices on the same graph are _ parallel _ if they have the same sign , _ anti - parallel _ if they have opposite signs . the orientability of @xmath42 and @xmath119 and the knot exterior gives the following useful property of these graphs * parity rule : * _ an edge connects parallel vertices on one of @xmath178 iff it connects anti - parallel vertices on the other . _ for a subgraph , @xmath129 , of @xmath128 , a _ ghost edge _ for @xmath129 is an edge of @xmath128 which does not belong to @xmath129 but is incident to a vertex of @xmath129 .
the incidence of a ghost edge for @xmath129 with a fat vertex of @xmath129 is called a _ ghost label _ of @xmath129 .
a connected subgraph , @xmath129 , of @xmath128 is called a _
@xmath5web _ if its vertices are parallel and if @xmath129 has at most @xmath199 ghost labels .
if @xmath200 is a component of @xmath201 then we say @xmath202 is a _ disk bounded by @xmath129_. a _
great @xmath5web _ is a @xmath5-web with the property that there is a disk , @xmath203 , bounded by @xmath129 such that @xmath204 .
note that as long as @xmath205 and @xmath206 then a great @xmath5-web must contain at least two vertices ( @xmath128 has no 1-sided faces ) .
[ lem:2 ] let @xmath122 be the graphs of intersection coming from @xmath123 of lemma [ lem : fandq ] as described above . if @xmath207 and @xmath206 , then @xmath128 has a great @xmath5-web , @xmath129
this is theorem 6.1 of @xcite where @xmath208 play the role of @xmath209 , where the slope @xmath210 there is the meridianal slope of @xmath11 . in @xcite ,
a great @xmath5-web is not necessarily connected , but by restricting to a connected component we may take it to be . the goal of theorem [ thm : main ] is to bound @xmath211 in terms of the genus of @xmath136 .
thus , after lemma [ lem:2 ] , we will hereafter assume the existence in @xmath128 of the great @xmath5-web , @xmath129 . for each label @xmath212 ,
the subgraph of a great @xmath5web @xmath129 consisting of all edges with an endpoint labeled @xmath130 and the vertices to which these edges are incident is denoted @xmath131 .
we think of @xmath131 as a graph in the disk bounded by @xmath129 .
a ghost edge of @xmath129 which is incident to a vertex of @xmath129 with label @xmath130 is called a _
ghost @xmath130-edge _ for @xmath129 . by the parity rule
, each ghost @xmath130-edge has at most one endpoint at a fat vertex of @xmath129 with label @xmath130 .
we refer to this endpoint as a _ ghost @xmath130-label_. let @xmath213 denote the number of ghost @xmath130labels for @xmath129 .
note that , taken over all labels , @xmath214 .
[ def : call ] a bigon ( trigon ) face of @xmath131 is referred to as an _ @xmath130-bigon _
( _ @xmath130-trigon _ , resp . ) of @xmath129 .
let @xmath132 be the set of labels @xmath130 of @xmath129 for which @xmath129 has an @xmath130bigon or @xmath130trigon .
by @xmath143 we denote the complement of @xmath132 in the set of all labels of @xmath128 .
[ lem : q2xinl0 ] if @xmath215 , and @xmath216 , then @xmath217 .
assume that @xmath215 and @xmath216 .
assume for contradiction that @xmath218 .
note that components of @xmath131 can not be nested in the disk bounded by @xmath129 .
take a connected component , @xmath219 , of @xmath131 with the fewest number of ghost @xmath130-labels .
let @xmath220 be the number of vertices , edges , and faces of @xmath219 in the disk bounded by @xmath129 .
let @xmath56 be the number of outside edges of @xmath219 , counted with multiplicity .
by assumption @xmath219 can contain no bigon or trigon face , so counting edges gives : @xmath221 .
note that @xmath222 by our assumptions , and @xmath129 has at least two vertices . as @xmath216 and
as each edge of @xmath129 has at most one label @xmath130 ( parity rule ) , we have ( @xmath223 ) @xmath224 . along with the euler characteristic equation @xmath225
, these imply that @xmath226 .
that is , @xmath227 .
this implies that @xmath228 and @xmath229 . if @xmath219 has at most two ghost @xmath130-labels then @xmath230 . using this in place of ( @xmath223 ) ,
the calculations above show that @xmath231 a contradiction .
thus we have that @xmath219 has exactly @xmath2 ghost @xmath130-labels .
this implies that @xmath232 and @xmath219 contains all the vertices of @xmath129 .
this along with the facts that @xmath233 implies that @xmath219 consists of a single edge connecting two vertices , and that one of these vertices has two ghost @xmath130-labels .
thus @xmath129 consists of two vertices connected by fewer than @xmath126 parallel edges ( @xmath128 contains no @xmath184-sided faces ) . as @xmath228 , @xmath129 has more than @xmath234 ghost labels
. thus @xmath235 a contradiction .
[ prop : webcount ] if @xmath207 and @xmath206 then @xmath236 .
assume @xmath237 and @xmath130 is a label which is not in @xmath132 .
then by lemma [ lem : q2xinl0 ] , @xmath238 .
hence @xmath239 hence @xmath240 . a _ scharlemann cycle ( of length @xmath56 )
_ is a disk face of @xmath128 or @xmath127 with @xmath56 edges , all edges having the same pair of labels and all connecting parallel vertices of the graph .
we use the same term for the set of edges defining the face .
the scharlemann cycles considered in this paper are typically on @xmath128 and of length @xmath49 or @xmath2 .
note that by the parity rule , any edge of the great @xmath5-web @xmath129 must have different labels in @xmath128 .
the subgraph of @xmath129 bounded by a bigon face in @xmath131 contains a scharlemann cycle of length @xmath49 .
this subgraph is called an _ extended scharlemann cycle of length @xmath49 _ and the scharlemann cycle it contains is referred to as its _ core _ scharlemann cycle .
[ def : trigonarms ] a trigon in @xmath131 is a _ cycle _ trigon if its edges can be oriented consistently around the trigon so that the @xmath130label is always at the tail end of an edge ( i.e. it forms an @xmath130cycle ) . see figure [ fig : xcycletrigon ] .
otherwise it is a _ non - cycle _ trigon .
see figure [ fig : noncycletrigon ] .
[ def : trigonarms2 ] if @xmath241 is a trigon face of @xmath131 then the subgraph of @xmath129 bounded by @xmath241 consists of a trigon face of @xmath129 together with three _ arms _ , each consisting of a ( possibly empty ) string of bigon faces of @xmath129 ; see for example figure [ fig : xcycletrigon ] . the subgraph of @xmath129 bounded by a trigon face , @xmath241 , in @xmath131 , either contains a single scharlemann cycle of length @xmath2 or contains one or two scharlemann cycles of length @xmath49 ( e.g. see the proof of lemma [ lem : existss22 ] ) . if it contains a scharlemann cycle of length @xmath2 , this subgraph is called an _ extended scharlemann cycle of length @xmath2 _ and the scharlemann cycle it contains is referred to as its _ core _ scharlemann cycle .
trigons of @xmath131 which are not extended scharlemann cycles will be classified later as type i , ii , or iii ( see definition [ def : trigontypes ] ) .
the arc on the boundary of a vertex between two consecutive edges of a face of a subgraph of @xmath128 or @xmath127 is a _ corner _ of that face .
if a set of edges @xmath241 is the boundary of a face @xmath242 of @xmath131 then let @xmath243 be the set of labels of @xmath129 that appear on the corners of @xmath242 .
let @xmath241 be an extended scharlemann cycle of @xmath128 .
let @xmath244 be the subgraph of @xmath127 consisting of the outermost ( in @xmath128 ) edges of @xmath241 and the vertices to which these edges are incident .
@xmath241 lies in a disk _ if @xmath244 is contained in a disk in @xmath245 .
@xmath241 lies in an essential annulus _ if @xmath244 is contained in an annulus in @xmath245 but does not lie in a disk .
[ lem : doesnotlieindisk ] no extended scharlemann cycle of @xmath128 lies in a disk .
assume @xmath241 is an extended scharlemann cycle of length @xmath14 that lies in a disk .
note that @xmath246 .
let @xmath242 be the face of @xmath241 .
let @xmath203 be a small disk in @xmath245 in which @xmath247 lies ( above ) . by disk exchanges
we may assume @xmath248 .
let @xmath249 be the @xmath184handle neighborhood of the arc of @xmath1 that forms the corners of @xmath242 .
then @xmath250 is a punctured lens space of order @xmath14 .
this can not happen since @xmath3 is neither reducible nor a lens space .
[ lem : aentscc ] no simple closed curve of @xmath251 that is trivial in @xmath119 is trivial in @xmath136 .
otherwise let @xmath252 be the disk bounded by such a simple closed curve .
let @xmath253 be @xmath127 restricted to @xmath254 . by lemma [ lem : fandq](3 ) ,
@xmath253 is non - empty . by lemma [ lem : fandq](2 )
, there are no @xmath184-sided faces in @xmath253 and no @xmath184-sided faces in the subgraph of @xmath128 corresponding to the edges of @xmath253 .
the argument of proposition 2.5.6 of @xcite , along with the assumption that @xmath206 , implies that one of @xmath253 or @xmath128 contains a scharlemann cycle .
such a scharlemann cycle along with the argument of lemma [ lem : doesnotlieindisk ] above , would imply the contradiction that either @xmath4 or @xmath3 contains a lens space summand .
lemma [ lem : fandq](2 ) along with lemma [ lem : aentscc ] imply the following .
[ cor : aentscc ] any simple closed curve of @xmath172 that lies in a disk face of @xmath119 is an essential curve in @xmath136 .
[ defn3.1 ] let @xmath257 be a collection of simple loops on a surface @xmath17 , and let @xmath258 be a non - negative integer .
we say that @xmath257 has _ property @xmath255 _
if * the elements of @xmath257 are essential and pairwise non - isotopic on @xmath17 ; * any pair of elements of @xmath257 meet transversely ; * for all @xmath259 , @xmath260 ; and * any @xmath261 meets at most @xmath258 elements of @xmath262 . for @xmath263 and @xmath264
define @xmath265 is a collection of simple loops on a closed , connected , orientable surface of genus @xmath5 with property @xmath266 . [
lem : fkg]@xmath267 , where @xmath268 and @xmath269 is : @xmath270 moreover , we have equality if @xmath72 or @xmath271 , i.e. @xmath272 in the present paper we will only make use of the formula for @xmath273 ( and @xmath274 ) .
however , we include the general case as possibly being of independent interest .
we will prove the lemma by induction on @xmath5 . since the case @xmath275 is easy and well - known , we assume @xmath276 . let @xmath257 be a collection of simple loops with property @xmath255 on a closed , connected , orientable surface @xmath17 of genus @xmath277 .
first we consider the special case where there is a loop @xmath278 that bounds a once - punctured torus @xmath279 containing a curve in @xmath257 that intersects @xmath258 other elements of @xmath257 .
let @xmath280 be the other component of @xmath17 cut along @xmath281 , and let @xmath282 be the closed surface of genus @xmath283 obtained by capping off the boundary of @xmath280 with a disk @xmath284 .
note that @xmath281 is disjoint from all the other members of @xmath257 .
it follows that either @xmath285 and there are exactly two elements of @xmath257 in @xmath286 , or @xmath287 and there are exactly three elements of @xmath257 in @xmath286 . discarding from @xmath257 these elements , together with @xmath281 , we get a collection @xmath288 of loops on @xmath280 .
regarded as a collection of loops on @xmath282 , @xmath288 satisfies all the conditions for property @xmath255 except possibly ( 1 ) : some pairs of loops in @xmath288 may become isotopic on @xmath282 .
( note that no loop in @xmath288 is inessential in @xmath282 , as such a loop would be parallel to @xmath281 in @xmath17 . )
let @xmath289 be curves that are isotopic on @xmath282 . then @xmath290 and @xmath291 are disjoint , and cobound an annulus @xmath292 in @xmath282 .
since @xmath290 and @xmath291 are not isotopic on @xmath17 , we have @xmath293 . in particular
, it follows that no triple of curves in @xmath288 is isotopic in @xmath282 .
now suppose @xmath294 is another pair of curves that are isotopic in @xmath282 .
then @xmath43 and @xmath295 cobound an annulus @xmath296 in @xmath282 with @xmath297 .
hence @xmath298 .
if @xmath299 , then either @xmath290 or @xmath291 lies in @xmath296 , or @xmath43 or @xmath295 lies in @xmath292 .
but then @xmath290 or @xmath291 would be isotopic to @xmath43 or @xmath295 in @xmath17 , a contradiction .
hence , without loss of generality , @xmath300 . therefore @xmath301 is a single transverse arc in @xmath292 .
since @xmath43 and @xmath295 are isotopic in @xmath282 , we must also have @xmath302 , and so @xmath303 is also a single transverse arc .
thus the pair @xmath304 contributes two points of intersection to each of @xmath290 and @xmath291 .
it follows that if @xmath285 then at most one pair of curves in @xmath288 are isotopic in @xmath282 .
we thus get a collection @xmath305 of loops on @xmath282 with property @xmath306 such that @xmath307 . therefore @xmath308 . by induction @xmath309 , and hence @xmath310 as claimed .
if @xmath287 and at most one pair of curves in @xmath288 are isotopic in @xmath282 then the result follows easily by induction .
so suppose that two pairs of curves @xmath311 and @xmath304 in @xmath288 become isotopic in @xmath282 .
let @xmath312 be the annuli cobounded by @xmath311 and @xmath304 respectively , extended slightly so that @xmath313 and @xmath314
. then @xmath315 is a disk containing @xmath284 , and hence @xmath316 is a once - punctured torus .
there is a simple loop @xmath317 in @xmath316 intersecting each of @xmath290 and @xmath43 transversely in a single point .
then @xmath318 is a collection of loops on @xmath282 with property @xmath319 .
hence , using the inductive hypothesis , @xmath320 we now consider the general case , and assume that @xmath257 contains no loop @xmath281 as above .
since @xmath321 is non - decreasing in @xmath258 , we may assume that @xmath257 does not have property @xmath322 , i.e. there is a loop @xmath278 that intersects @xmath258 other elements @xmath323 .
in particular , @xmath281 is non - separating on @xmath17 , and so cutting @xmath17 along @xmath281 gives a connected surface @xmath280 .
let @xmath324 be the closed surface of genus @xmath325 obtained by capping off the two boundary components of @xmath280 with disks @xmath326 and @xmath327 .
let @xmath328 .
let @xmath329 . as a collection of loops on @xmath330
, @xmath331 satisfies property @xmath255 except that some pairs of loops in @xmath331 may become isotopic in @xmath330 .
( again , no loop in @xmath331 is trivial in @xmath330 as such a loop would be parallel to @xmath281 in @xmath17 . )
let @xmath311 be such a pair .
the previous discussion applies verbatim , with @xmath288 replaced by @xmath331 and @xmath282 replaced by @xmath330 , so any other such pair @xmath304 contributes two points of intersection to each of @xmath290 and @xmath291 .
at least @xmath332 of the curves @xmath333 must intersect one of @xmath311 . hence the number of pairs of elements of @xmath331 that are isotopic in @xmath330 is at most @xmath334 . discarding one element of @xmath331 from each such pair , we get a collection of loops @xmath335 with property @xmath255 in @xmath330 .
now apply the same argument to @xmath335 , regarded as a collection of loops in @xmath336 .
we may assume that no element of @xmath335 is inessential in @xmath282 , as such a curve @xmath317 would bound a disk @xmath203 in @xmath282 with @xmath337 , and would therefore bound a once - punctured torus in @xmath17 containing @xmath281 , putting us in the special case treated earlier .
we then get a collection @xmath305 of loops in @xmath282 satisfying property @xmath255 , with @xmath338 .
hence @xmath339 now assume @xmath340 and that the theorem holds for @xmath325 .
then @xmath341 and the result follows by induction on @xmath5 . to start the induction we will show that @xmath342 .
let @xmath17 be a closed , orientable surface of genus 2 and let @xmath257 be a collection of simple loops on @xmath17 with property @xmath255 .
let @xmath343 be the hyperelliptic involution .
the quotient @xmath344 where @xmath77 consists of six points .
let @xmath345 be the quotient map .
the elements of @xmath257 can be isotoped so that @xmath346 for all @xmath261 .
if @xmath317 meets @xmath347 then @xmath348 is an arc in @xmath349 whose endpoints are distinct points in @xmath77 ; if @xmath350 then @xmath348 is a simple loop in @xmath351 . by part
( 3 ) of definition [ defn3.1 ] , if @xmath352 are distinct elements of @xmath257 then @xmath353 is either empty or a single point in @xmath77 . the arcs @xmath348 form the edges of a graph @xmath354 in @xmath349 with vertex set @xmath77 , and @xmath355 is the disjoint union @xmath356 where the components of @xmath129 are simple loops .
hence @xmath357 , where @xmath358 is the number of edges of @xmath354 .
note that no two edges of @xmath354 share the same pair of endpoints , no component of @xmath129 bounds a disk in @xmath359 , and no two components of @xmath129 are parallel in @xmath359 .
finally , by part ( 4 ) of definition [ defn3.1 ] , the sum of the valencies of the vertices at the endpoints of each edge of @xmath354 is at most @xmath360 .
it is straightforward to check that with these constraints , for @xmath361 the maximum value of @xmath362 is the @xmath269 given in the statement of the lemma .
these values are realized by the configurations shown in figure [ fig : curves2 ] .
note that by passing to the 2-fold branched covering of @xmath363 we obtain a collection @xmath257 of simple loops on @xmath17 with property @xmath255 .
for @xmath364 we have the 1-skeleton of the octahedron shown in figure [ fig : curves2 ] , with 12 edges .
an easy euler characteristic argument shows that a graph @xmath354 in @xmath349 with six vertices , no loop edges and no parallel edges , has at most 12 edges .
it follows that @xmath365 for @xmath366 .
finally we show that the inequality in the lemma is an equality when @xmath367 or 2 . for @xmath275 ,
this is easy and well known . for @xmath285 or 2 , let @xmath17 be the closed surface of genus @xmath5 obtained by attaching @xmath5 once - punctured tori @xmath368 , @xmath369 , to a @xmath5-punctured sphere @xmath370 along their boundaries . in each @xmath368 take a collection @xmath371 of 2 ( resp .
3 ) simple loops with pairwise intersection numbers equal to 1 , and in @xmath370 take a collection @xmath288 of simple loops that cut @xmath370 into pairs of pants .
define @xmath372 , a collection of loops on @xmath17 .
then @xmath257 satisfies property @xmath306 ( resp .
@xmath319 ) and @xmath373 ( resp .
@xmath374 ) .
consider the number @xmath375 .
@xcite shows @xmath376 for @xmath377 .
moreover , by essentially the same method as the base case of our induction , @xcite proves @xmath378 .
let @xmath379 be labels of @xmath128 , then @xmath380 is one of the two intervals of labels running between @xmath290 and @xmath43 on an abstract vertex of @xmath128 that contains no labels @xmath381 on its interior .
typically , one such interval is specified by the context .
if @xmath241 is an extended scharlemann cycle whose edges have labels @xmath379 then we may think of @xmath243 as an interval of labels @xmath380 .
[ lem : commonlabelsets ] let @xmath382 be extended scharlemann cycles of length @xmath49 or @xmath2
. then either 1 .
@xmath383 contains all labels ; 2 .
@xmath384 or @xmath385 ; 3 .
@xmath386 is a single interval of labels @xmath387 where @xmath130 is an extremal label of @xmath241 and @xmath166 is an extremal label of @xmath24 ; or 4 .
@xmath388 .
@xmath389 are each a single interval of labels . by convexity , either @xmath386 contains all labels or @xmath386 is convex . if nonempty , it is a single interval of intersection , @xmath387 . clearly , @xmath130 and @xmath166 must be extremal in either @xmath241 or @xmath24 .
if both are extremal in , say , @xmath241 , then @xmath243 is contained in @xmath390 .
[ def : longmobiusband ] let @xmath241 be an extended scharlemann cycle of length @xmath49 with @xmath391 .
let @xmath242 be the union of the bigons bounded by @xmath241 in @xmath128 .
construct the _ long mbius band _ @xmath392 in @xmath3 by extending the corners of @xmath242 radially to @xmath1 in a neighborhood of @xmath1 in @xmath3 .
let @xmath393 be the set of simple closed curves @xmath394 formed by pairs of edges of @xmath241 .
the _ core labels _ of @xmath241 are the two labels of the scharlemann cycle face of @xmath241 .
the _ core curve _ of @xmath241 is the element of @xmath393 on its core labels .
an _ embedded @xmath135curve _ in a surface @xmath245 is a graph @xmath135 consisting of two vertices and three edges , between these two vertices , embedded in @xmath245 so that @xmath395 is a thrice punctured sphere .
@xmath135band _ is homeomorphic to the product of an embedded @xmath135curve with an interval .
let @xmath241 be a scharlemann cycle of length @xmath2 in @xmath128 .
then , regarding as points the two vertices of @xmath127 to which the edges of @xmath241 are incident , the edges of @xmath241 form an embedded @xmath135-curve in @xmath245
. let @xmath242 be the trigon of @xmath128 bounded by @xmath241 .
let @xmath396 be the label pair on the edges of @xmath241 . with @xmath397 , @xmath210 , and @xmath23
label the corners of @xmath242 and hence label the endpoints of the edges of @xmath241 too .
these labels then appear around vertices @xmath130 and @xmath398 of @xmath127 in opposite directions .
since each edge of @xmath241 connects two distinct corners of @xmath242 , the endpoints of each edge have distinct labels .
shrinking vertices @xmath130 and @xmath398 to points , this forces @xmath241 to form an embedded @xmath135-curve in @xmath245 .
[ def : twistedthetaband ] let @xmath241 be an extended scharlemann cycle of length @xmath2 with @xmath391 .
let @xmath242 be the union of the bigons and the trigon bounded by @xmath241 in @xmath128 .
construct the _ long twisted @xmath135band _
@xmath399 in @xmath3 by extending the corners of @xmath242 radially to the core of @xmath400 .
let @xmath401 be the set of embedded @xmath135curves @xmath402 .
the _ core labels _ of @xmath241 are the two labels of the scharlemann cycle face of @xmath241 .
the _ core @xmath135curve _ of @xmath241 is the element of @xmath401 on its core labels
. the following shorthand will be convenient .
if @xmath241 is an extended scharlemann cycle of length @xmath2 , then take @xmath403 and @xmath404 .
if @xmath241 is an extended scharlemann cycle of length @xmath49 , then take @xmath405 and @xmath406 .
if @xmath241 is an extended scharlemann cycle of length @xmath49 or @xmath2 , let @xmath407 and @xmath408 .
let @xmath409 is an extended scharlemann cycle of @xmath129 of length @xmath49 or @xmath410 be the set of all labels that lie on the corners of extended scharlemann cycles of length @xmath49 or @xmath2 in @xmath129 .
[ def : sigma ] let @xmath137 be a collection of extended scharlemann cycles of length @xmath49 or @xmath2 in @xmath129 such that @xmath411 contains @xmath412 .
let @xmath413 .
assume @xmath137 is chosen so that the complexity @xmath414 is minimal in the lexicographic ordering among all such collections .
[ lem : commonlabelsetsrefined ] if @xmath415 then either 1 . @xmath383 contains all labels ; 2 .
@xmath386 is a single interval of labels @xmath387 where @xmath130 is an extremal label of @xmath241 and @xmath166 is an extremal label of @xmath24 ; or 3 .
@xmath388 .
if none of these were to occur , then by lemma [ lem : commonlabelsets ] the only other possibility is that either @xmath384 or @xmath385 .
therefore @xmath241 or @xmath24 respectively may be dropped from @xmath137 to produce @xmath416 so that @xmath417 contains @xmath418 and yet has lesser complexity than @xmath137 .
this contradicts the minimality assumption on @xmath137 .
the extended scharlemann cycle @xmath419 is obtained by _ paring down _ the extended scharlemann cycle @xmath241 if @xmath419 is contained in the disk face that @xmath241 bounds .
[ lem : atmostoneintersection ] two elements of @xmath144 intersect at most once .
furthermore , each vertex of @xmath127 belongs to at most two different elements of @xmath144 .
if there exists @xmath415 with @xmath420 and @xmath421 intersecting twice , then @xmath422 and @xmath423 have the same label pair
. therefore we may pare down either @xmath241 or @xmath24 ( or completely eliminate one ) to produce @xmath416 so that @xmath417 contains @xmath418 and yet has lesser complexity than @xmath137 .
this contradicts the minimality assumption on @xmath137 .
if label @xmath130 of @xmath128 were in @xmath424 and @xmath425 with @xmath426 , then one of these label sets must be contained in the union of the other two .
this contradicts the minimality of @xmath137 thereby proving the final statement of the lemma .
[ lem : gtsolidtorus ] let @xmath241 be an extended scharlemann cycle of length @xmath2 that lies in an essential annulus @xmath292 of @xmath136 .
let @xmath242 be the face bounded by @xmath241 and assume @xmath427 .
then @xmath428 is a solid torus and the core of @xmath292 runs @xmath2 times in the longitudinal direction of @xmath429 . let @xmath249 be the @xmath184handle neighborhood of the arc of @xmath1 that forms the corners of @xmath242
. then @xmath428 may be obtained from attaching the @xmath49handle @xmath430 to the genus @xmath49 handlebody @xmath431 .
because a meridional disk of @xmath432 intersects @xmath433 once , @xmath429 is a solid torus .
one may then observe that the core of @xmath292 runs @xmath2 times in the longitudinal direction of @xmath429 .
lemma 2.1 @xcite ; lemma 3.7 @xcite ) .
[ lem : essentialcurves ] let @xmath241 be an extended scharlemann cycle of length @xmath49 or @xmath2 in @xmath129 .
assume an element of @xmath434 lies in an annulus @xmath292 of @xmath136 .
let @xmath435 be the core of @xmath292 .
then @xmath435 can not bound a disk in @xmath3 whose framing on @xmath436 is the same as that of @xmath136 . in particular
, @xmath435 is neither trivial on @xmath136 nor a meridian curve on either side of @xmath136 .
assume some element , @xmath317 , of @xmath434 lies in an annulus @xmath292 in @xmath136 whose core is @xmath435 .
assume for contradiction that there is a disk @xmath203 in @xmath437 whose boundary has the same slope on @xmath438 as @xmath136 .
pare down @xmath241 so that @xmath317 is composed of the outermost edges of @xmath241 .
the corners of @xmath241 belong to the same sub - arc , @xmath258 , of @xmath1 .
let @xmath242 be the trigon face bounded by @xmath241 .
then @xmath439 ( whose boundary is @xmath317 ) is @xmath242 radially contracted to @xmath258 .
we view both @xmath439 and @xmath203 as properly embedded in @xmath152 . by the framing assumption on @xmath203
, we may take the boundaries of these two surfaces to be disjoint .
isotop @xmath258 rel its endpoints on @xmath440 so that @xmath258 intersects @xmath203 minimally .
then any arc components of intersection between @xmath242 and @xmath203 must be parallel on @xmath242 to the outermost edges of @xmath241 .
we may surger away any simple closed curves of intersection .
if @xmath258 is disjoint from @xmath203 , then a neighborhood of @xmath441 will be a lens space summand of @xmath3 ( using lemma [ lem : gtsolidtorus ] when the length of @xmath241 is @xmath2 ) , a contradiction .
if @xmath258 intersects @xmath203 , then the graph of intersection between @xmath203 and @xmath242 is a collection of arcs parallel to @xmath442 with the same number of endpoints on each corner of @xmath242 .
an innermost face of this graph on @xmath242 will give a length @xmath2 scharlemann cycle with respect to @xmath203 which we can use , as in lemma [ lem : doesnotlieindisk ] , to construct a lens space summand in @xmath3 .
[ lem : parallelthetas ] let @xmath443 be a set of embedded @xmath135-curves in @xmath136 such that 1 .
no @xmath444 lies in a disk in @xmath136 ; 2 . if @xmath445 then @xmath446 is either empty or a single vertex ; 3 .
any vertex belongs to at most two elements of @xmath443 .
if @xmath447 then some @xmath444 lies in an essential annulus in @xmath136 . by ( 1 ) , for each @xmath444 we may choose an essential circle @xmath448 in @xmath135 by deleting the interior of one of the edges of @xmath135 .
let @xmath449 . if @xmath450 is a subcollection such that no two elements of @xmath451 are isotopic in @xmath136 then ( after isotoping the elements of @xmath451 into general position ) @xmath451 has property @xmath319 of definition [ defn3.1 ] .
hence , by lemma [ lem : fkg ] , if @xmath452 then four elements of the @xmath453 must be isotopic in @xmath136 .
any two of these curves are either disjoint or intersect in a single vertex non - transversely . in either case
we think of two such curves as cobounding an ( essential ) annulus in @xmath136 .
one sees that one of these four curves must contain a vertex that lies strictly in the interior of the annulus cobounded by two others .
the corresponding @xmath135curve then lies within this annulus .
let @xmath241 be an extended scharlemann cycle of length @xmath49 with @xmath454 .
then we may order the elements of @xmath393 as @xmath455 such that @xmath456 is the core curve and @xmath457 bounds an annulus @xmath458 that is disjoint from the other elements of @xmath393 for @xmath459 .
the core scharlemann cycle in @xmath241 forms the core curve @xmath456 which bounds the mbius band @xmath460 .
[ def : thetaj ] let @xmath241 be an extended scharlemann cycle of length @xmath2 with @xmath461
. then we may order the elements of @xmath401 as @xmath462 such that @xmath463 is the core @xmath135curve and @xmath464 bounds a product @xmath135curve @xmath465 that is disjoint from the other elements of @xmath401 for @xmath459 .
the core scharlemann cycle in @xmath241 forms the core @xmath135curve which bounds the twisted @xmath135band @xmath466 .
[ lem : annulusintheta ] let @xmath241 be an extended scharlemann cycle of length @xmath2 .
assume @xmath467 , @xmath468 , lie in an essential annulus @xmath296 in @xmath245 .
exactly two of the edges , @xmath469 of @xmath470 cobound a disk in @xmath136 .
let @xmath471 be the edges of @xmath472 that lie on the corresponding arms ( definition [ def : trigonarms2 ] ) of @xmath241 .
then @xmath471 cobound a disk on @xmath136 . since @xmath470 lies in the annulus @xmath296 , two of its edges , @xmath469 , cobound a disk , @xmath203 , in @xmath296 , and ( by lemma [ lem : essentialcurves ] ) the third creates a core curve of @xmath296 when joined with either @xmath473 or @xmath474 .
for @xmath258 between ( and including ) @xmath475 and @xmath476 , let @xmath477 be as in definition [ def : thetaj ] .
let @xmath478 be the edges of @xmath479 such that @xmath480 lies on the same arm of @xmath241 as @xmath473 , and @xmath481 on the same arm as @xmath474 .
let @xmath482 be the curve @xmath483 .
let @xmath484 be the union of the faces on each arm between @xmath485 and @xmath486 .
then @xmath487 forms an annulus @xmath292 in @xmath3 whose boundary components are @xmath488 and @xmath489 . by passing to a concentric subannulus of @xmath296 , we may assume that @xmath490 , where @xmath491 are curves in @xmath492 .
any component of @xmath493 is one of the curves @xmath482 for some @xmath258 between @xmath475 and @xmath476 .
the corresponding @xmath135-curve @xmath479 is then contained in @xmath296 , and does not lie in a disk by lemma [ lem : essentialcurves ] . also , @xmath482 is either a core curve of @xmath296 or bounds a disk in @xmath296 .
assume for contradiction that @xmath489 does not bound a disk in @xmath296 .
since @xmath488 does bound a disk in @xmath296 , there exist @xmath494 between @xmath475 and @xmath476 such that @xmath482 and @xmath495 are adjacent on @xmath296 among the components of @xmath493 , @xmath482 bounds a disk in @xmath296 and @xmath495 is a core curve of @xmath296 . by renumbering we
may therefore assume that @xmath496 and @xmath497 ; thus the annulus @xmath292 meets @xmath296 only in @xmath498 .
let @xmath163 be the disk @xmath499 . by pushing @xmath203 slightly off @xmath296 in the appropriate direction we may assume that @xmath500 .
let @xmath501 and let @xmath419 be the extended scharlemann cycle within @xmath241 that terminates at the edges of @xmath502 .
then the interior of the face bounded by @xmath419 is disjoint from @xmath296 and @xmath163 .
let @xmath429 be a regular neighborhood in @xmath3 of @xmath503 . by lemma [ lem : gtsolidtorus ] , @xmath429 is a punctured lens space of order @xmath2 , a contradiction .
the above lemma allows certain arguments for extended scharlemann cycles of length @xmath49 to apply to extended scharlemann cycles of length @xmath2 . to facilitate this
we make the following definitions .
assume @xmath241 is an extended scharlemann cycle of length @xmath2 with @xmath467 , @xmath468 , lying in an essential annulus @xmath296 in @xmath245 .
let @xmath504 be the edges of @xmath470 that cobound a disk , @xmath505 , in @xmath296 and @xmath506 be the third edge of @xmath470 .
let @xmath507 be the edges of @xmath472 on the corresponding arms of @xmath241 .
by lemma [ lem : annulusintheta ] , @xmath508 cobound a disk @xmath509 in @xmath296 .
@xmath505 ( @xmath509 ) is called the _ disk of parallelism _ for @xmath470 ( @xmath472 , resp . ) . for @xmath510 ,
let @xmath511 be the faces along an arm of @xmath241 between @xmath512 . by lemma [ lem : annulusintheta ] , @xmath513 and @xmath514 form isotopic annuli in @xmath3 whose
boundaries lie in @xmath515 and are isotopic to the core of @xmath296 .
@xmath516 are called the _ constituent annuli _ of @xmath517 of @xmath439 between @xmath492 .
[ lem : sfs ] let @xmath241 be an extended scharlemann cycle of length @xmath518 or @xmath519 .
assume @xmath520 , @xmath521 , lie on an annulus , @xmath296 , in @xmath136 such that all of the intersections of @xmath1 with @xmath296 belong to @xmath522 ( i.e. occur at the four vertices of @xmath522 ) .
then either 1 .
@xmath523 and the annulus @xmath524 between @xmath522 ( or constituent annulus when @xmath241 has length @xmath2 ) is parallel into @xmath136 ; or 2 .
@xmath3 is a seifert fibered space over @xmath349 with three exceptional fibers , one of which has order @xmath525 , and @xmath1 has bridge number @xmath36 with respect to some genus @xmath49 heegaard splitting of @xmath3 .
the conclusion above holds when @xmath241 has length @xmath2 and @xmath1 intersects @xmath296 only in @xmath522 or in the disk of parallelism of at most one of @xmath526 or @xmath527 .
we assume @xmath1 intersects @xmath296 only at the four vertices of @xmath522 .
let @xmath528 where @xmath529 when @xmath241 is of length @xmath49 , and where @xmath292 is a constituent annulus of @xmath439 between @xmath522 when @xmath241 is of length @xmath2 .
let @xmath530 .
note that @xmath292 and @xmath296 only intersect in their boundaries ( otherwise @xmath3 contains a projective plane , corollary [ cor : aentscc ] ) , thus @xmath531 is an embedded torus in @xmath3 ( @xmath3 contains no klein bottles ) . since @xmath3 is assumed to be atoroidal , @xmath531 must be compressible in @xmath3 .
let @xmath203 be a compressing disk .
_ addendum : _ in the context of the addendum , assume that @xmath1 intersects @xmath296 only in the four vertices of @xmath522 and in the disk of parallelism of , say , @xmath526
. then we may choose the constituent annulus of @xmath439 between @xmath522 and shrink @xmath296 so that @xmath292 and @xmath296 meet along their boundary and @xmath296 is disjoint from the interior of the disk of parallelism of @xmath527 .
then @xmath296 intersects @xmath1 only along @xmath522 , and we apply the arguments to this @xmath312 .
the proof of this lemma splits into three cases depending upon the relationship of @xmath532 , @xmath533 , and @xmath203 with respect to @xmath245 .
* case ( i ) : * @xmath534 and @xmath533 lie on opposite sides of @xmath245 .
the following claim shows that case i does not occur .
there is an isotopy of @xmath3 that reduces the width of @xmath1 with respect to @xmath136 .
the unfurling move from section 4.3 of @xcite gives a homeomorphism of @xmath3 that reduces the width of @xmath1 and is isotopic to the identity .
we describe this below .
let @xmath77 be a regular closed neighborhood of @xmath531 in @xmath3 .
let @xmath535 be the spanning arcs @xmath536 on the annulus @xmath292 .
then @xmath77 is an interval of tori , @xmath537 $ ] , where we coordinatize @xmath538 so that @xmath539 , @xmath540 , and @xmath541 for points @xmath542 . now , up to isotopy rel-@xmath162 in @xmath77 , we may take @xmath543 to be the pair of arcs @xmath544 which are @xmath535 extended by product arcs in @xmath545 $ ] .
let @xmath546 be the homeomorphism that rotates @xmath547 by @xmath548 in the factor @xmath549 and is the identity in the factor @xmath550 . under @xmath551 ,
@xmath552 become isotopic to spanning arcs of @xmath296 extended by product arcs across @xmath77 .
as @xmath551 is the identity on @xmath553 , it extends to @xmath3 as the identity outside @xmath77 , and the image of @xmath1 under @xmath551 is strictly thinner with respect to @xmath136 ( the arcs above and below @xmath136 after the homeomorphism are level - preserving isotopic to those of @xmath554 ) .
to finish the claim we must show that @xmath551 is isotopic to the identity on @xmath3 .
as above , let @xmath203 be a compressing disk for @xmath531 and @xmath555 .
as @xmath3 is irreducible , the 2-sphere component , @xmath17 , of @xmath556 bounds a 3-ball in @xmath3 which either * ( a ) * contains @xmath557 or * ( b ) * meets it only along @xmath17 . in case * ( a ) * , @xmath551 is the identity on @xmath3 outside the 3-ball .
so there is an isotopy on the 3-ball , keeping @xmath17 fixed , taking @xmath551 restricted to the 3-ball to the identity .
this extends to an isotopy on @xmath3 taking @xmath551 to the identity . in case
* ( b ) * , note that @xmath551 as a homeomorphism of @xmath77 is isotopic to the identity by an isotopy that keeps one boundary component of @xmath77 fixed and rotates the other component of @xmath553 by @xmath558 .
take such an isotopy that rotates the component @xmath559 of @xmath553 on the side containing @xmath203 and fixes the component , @xmath560 , on the other side .
@xmath559 bounds a solid torus , @xmath429 , in @xmath3 that intersects @xmath77 in @xmath559 .
so we can extend the isotopy of @xmath551 on @xmath77 across @xmath429 .
as this isotopy fixes @xmath560 , we may extend it across all of @xmath3 , giving the desired isotopy of @xmath551 to the identity on @xmath3 .
* case ( ii ) : * @xmath534 and @xmath533 lie on the same side of @xmath245 and @xmath203 near @xmath296 lies on the opposite side of @xmath245 .
let @xmath561 be the component of @xmath562 containing @xmath203 . by a slight isotopy supported in @xmath432 ,
push @xmath1 off of @xmath531 ( this comes from looking at the labelling around @xmath522 and the fact that @xmath531 is not a klein bottle ) .
we may assume that @xmath203 is disjoint from @xmath1 since the exterior of @xmath1 is atoroidal .
we will need to handle the cases @xmath518 and @xmath519 separately .
* case ( iia ) : * assume @xmath518 . then @xmath563 is an embedded mbius band that intersects @xmath531 in @xmath526
. now @xmath1 can be perturbed ( avoiding @xmath203 ) so that it is disjoint from @xmath531 and intersects @xmath564 in a single point .
( @xmath1 is made disjoint from @xmath531 as above . to see
it then intersects the mbius band @xmath460 once , note that when we push @xmath1 off @xmath292 it starts off on one side of @xmath532 then ends up on the other . )
we may assume that @xmath203 intersects @xmath564 in a nonempty collection of arcs which are essential in @xmath564 ( in fact parallel to @xmath565 in @xmath564 by working on the intersections before perturbing @xmath1 off @xmath564 ) else we get a projective plane or can reduce the number of intersections .
taking an outermost disk of @xmath566 and compressing @xmath564 along it constructs a compressing disk for @xmath531 in @xmath561 which intersects @xmath1 exactly once ( from the intersection of @xmath1 with @xmath564 ) .
but this contradicts that there is an essential compressing disk for @xmath531 which misses @xmath1 in @xmath561 .
* case ( iib ) : * assume @xmath519 .
then @xmath567 is an embedded twisted @xmath135band that intersects @xmath531 in @xmath568 . _
addendum : _ in the context of the addendum in this case , we expand @xmath296 to include the disks of parallelism .
if @xmath1 intersects the interior of the disk of parallelism of , say , @xmath526 we surger this disk along the faces of @xmath241 connecting the disk of parallelism of @xmath526 with that of @xmath527 as in the proof of lemma [ lem : annulusintheta ] .
we take the resulting @xmath296 for the argument below .
let @xmath569 .
then @xmath203 intersects @xmath570 away from @xmath571 .
we consider @xmath572 as a large trigon , @xmath242 , with corners along @xmath571 ( indeed it is a trigon face in @xmath131 for some label @xmath130 ) .
then we may assume @xmath203 intersects @xmath242 only in arcs parallel to the corners of @xmath242 .
furthermore , this collection of arcs is non - empty
else @xmath203 in union with @xmath570 along an annulus in @xmath531 gives rise to a punctured @xmath573 in @xmath3 by lemma [ lem : gtsolidtorus ] .
the compression of @xmath570 along an outermost disk of @xmath574 contains a mbius band @xmath564 with @xmath575 and @xmath576
. then we proceed as in case ( iia ) where @xmath518 to produce a compressing disk for @xmath531 in @xmath561 that intersects @xmath1 exactly once , contradicting that there is an essential disk disjoint from @xmath1 .
* case ( iii ) : * @xmath534 and @xmath533 lie on the same side of @xmath245 and @xmath203 near @xmath296 lies on this side of @xmath245 .
since @xmath534 and @xmath533 lie on the same side of @xmath245 , @xmath475 and @xmath476 have different parity .
let @xmath577 be the component of @xmath562 containing @xmath203 .
as in case ( ii ) , @xmath1 may be perturbed to miss @xmath577 completely . note that in the context of the addendum , the constituent annuli can be taken disjoint from the disks of parallelism of both @xmath526 and @xmath527 . compressing @xmath578 along @xmath203
gives a 2-sphere disjoint from @xmath1 .
as the exterior of @xmath1 is irreducible , we see that @xmath577 is a solid torus .
we show that either @xmath292 and @xmath296 are longitudinal annuli in @xmath577 , that is that @xmath579 on @xmath578 , thereby showing that @xmath292 is parallel into @xmath296 , or conclusion ( 2 ) holds . if @xmath580 then @xmath3 contains a punctured lens space ( coming from the elements of @xmath439 up to @xmath526 in union with @xmath203 along @xmath526 ) .
assume @xmath581 .
set @xmath582 if @xmath518 or @xmath583 if @xmath519 . in the context of the addendum
we expand @xmath296 to include disks of parallelism of @xmath522 and surger away intersects of @xmath1 with either of these disks of parallelism as in case(iib ) .
then @xmath77 is a seifert fibered space over the disk with two exceptional fibers of orders @xmath525 and @xmath56 ( note @xmath584 or @xmath585 intersect @xmath531 only along their boundary , else the core curve of @xmath296 bounds a disk outside of @xmath577 which along with @xmath577 would create a lens space summand in @xmath3 ) .
@xmath1 can be perturbed so that it is disjoint from @xmath577 and intersects @xmath584 or @xmath586 in a single point if @xmath518 or @xmath519 respectively ( this is most easily seen by checking the labellings around the vertices of @xmath526 , see figure [ fig : figure1 ] ) .
thus @xmath543 is isotopic rel boundary to the co - core of the unique essential annulus in the seifert fibered space , @xmath77 .
it follows from this that the twice - punctured torus @xmath587 is incompressible in @xmath588 .
[ clm : ontov ] @xmath589 is isotopic rel boundary in @xmath590 onto @xmath553 .
. then @xmath592 must be compressible in the exterior of @xmath1 , @xmath20 , by theorem 2.4.3 of @xcite ( where @xmath592 is the @xmath17 there and the @xmath100 there is @xmath10 ) .
as @xmath17 is incompressible in @xmath593 , @xmath592 must compress in @xmath594 .
if @xmath17 were incompressible in @xmath594 , then the handle addition lemma , lemma 2.1.1 of @xcite , would imply that @xmath553 is incompressible in @xmath590 .
but this contradicts that @xmath3 is atoroidal .
thus @xmath17 must compress in @xmath594 .
as @xmath20 is anannular , @xmath17 must compress in @xmath590 to a boundary parallel annulus .
this implies that @xmath589 is isotopic in @xmath590 to @xmath553 , as claimed .
since @xmath3 is atoroidal , and @xmath553 is incompressible in @xmath77 , @xmath595 must compress outside of @xmath77 . compressing @xmath595
then yields a @xmath49sphere that bounds a @xmath2ball @xmath596 .
if this 3-ball contained @xmath77 , then @xmath1 could be isotoped into this 3-ball , contradicting the irreducibility of the exterior of @xmath1 .
so we may assume that this 3-ball lies outside of @xmath77 , and consequently that @xmath590 is a solid torus .
then @xmath3 is a seifert - fibered space over the 2-sphere with three exceptional fibers one of which has order @xmath525 .
let @xmath597 .
one checks that @xmath598 is a genus two handlebody .
certainly , @xmath599 is a genus two handlebody .
that is , @xmath600 is a heegaard splitting of @xmath3 .
recall that @xmath589 can be isotoped rel its boundary onto @xmath553 .
thus @xmath1 can be isotoped onto @xmath601 .
this is conclusion @xmath602 of the lemma .
thus we may assume @xmath579 and @xmath577 gives an isotopy of @xmath292 to @xmath296 as desired .
if @xmath603 , then this isotopy from @xmath292 to @xmath296 gives a thinner presentation of @xmath1 ( the two sub - arcs @xmath604 of @xmath1 are flattened onto @xmath136 by this isotopy , perturb them back slightly off of @xmath136 to reinstate a morse presentation of @xmath1 with fewer extrema . ) as we started with a thin presentation of @xmath1 , it must be that @xmath523 .
these 3 cases exhaust the possibilities , thereby completing the proof of lemma [ lem : sfs ] .
[ lem : triplets ] let @xmath241 be an extended scharlemann cycle of length @xmath518 or @xmath519 .
assume that conclusion @xmath602 of lemma [ lem : sfs ] does not hold .
if @xmath605 together lie on an annulus in @xmath136 , then there must be intersections of @xmath1 with this annulus that do not belong to @xmath606 . without loss of generality
assume @xmath607 cobound annulus @xmath608 on @xmath136 and @xmath609 cobound annulus @xmath610 on @xmath136 , such that @xmath1 intersects @xmath611 only along @xmath606 and such that @xmath612 have disjoint interiors .
then by lemma [ lem : sfs ] we may assume @xmath613 and the annuli or constituent annuli , @xmath614 , of @xmath439 are parallel into @xmath615 .
but then we can use these parallelisms of the annuli to guide a thinning of @xmath1 with respect to @xmath136 .
[ lem : thetatriplets ] let @xmath241 be an extended scharlemann cycle of length @xmath519 .
assume that conclusion @xmath602 of lemma [ lem : sfs ] does not hold .
assume that @xmath135-curves @xmath605 together lie on an annulus in @xmath136 , with @xmath527 between @xmath616 on @xmath136 .
then there must be intersections of @xmath1 with this annulus that do not belong to @xmath606 or to the disks of parallelism of @xmath616 .
assume the only other intersections of @xmath1 occur within the disks of parallelism of @xmath616 .
we apply the proof of lemma [ lem : triplets ] using the addendum to lemma [ lem : sfs ] .
an extended scharlemann cycle is called _ proper _ if in its corner no label appears twice .
[ lem : parallelesc ] assume that @xmath136 is a strongly irreducible heegaard surface for @xmath3 .
let @xmath618 be proper extended scharlemann cycles of @xmath129 of length @xmath49 or @xmath2 with different core labels . if there is an annulus , @xmath292 , in @xmath136 containing an element of @xmath434 and an element of @xmath619 then either 1 .
@xmath243 and @xmath390 intersect in two non - empty label intervals , @xmath383 includes all labels , and one of @xmath241 or @xmath24 has length @xmath2 ; or 2 .
@xmath3 is a seifert fibered space over the @xmath49sphere with three exceptional fibers , and the orders of two of these exceptional fibers are the lengths of @xmath382 .
furthermore , there is a genus @xmath49 ( hence minimal genus ) heegaard splitting of @xmath3 with respect to which @xmath1 has bridge number at most @xmath620 , where @xmath621 .
after possibly paring down @xmath382 we may assume that @xmath622 are the only elements of @xmath623 ( resp . ) lying entirely in @xmath292 .
if there are intersections of @xmath624 with @xmath292 coming from the interior of faces of @xmath618 , then we may assume they are essential simple closed curves in @xmath292 ( lemma [ lem : aentscc ] ) .
take such a curve that is innermost on @xmath241 , say , bounding a disk @xmath203 .
surgery of @xmath292 along @xmath203 gives a disk @xmath564 which we may take ( after possibly further surgeries ) to intersect @xmath439 only along its boundary .
but then @xmath625 is a lens space summand in @xmath3 a contradiction .
thus we may assume that @xmath618 intersect @xmath292 only along their edges .
the argument now divides up according to @xmath386 .
see lemma [ lem : commonlabelsets ] for the possibilities .
we work through these , ending with case iv that @xmath383 contains all labels and @xmath626 . in this final case we show that at least one of @xmath382 must be of length @xmath2 , giving conclusion ( 1 ) above .
the generic argument is case i. * case i : * @xmath388 then @xmath627 and we may assume @xmath628 intersect @xmath292 only along their boundaries .
consequently @xmath629 is a seifert fibered space over the disk with two exceptional fibers whose orders are given by the lengths of @xmath618 ( e.g. lemma [ lem : gtsolidtorus ] ) .
note that as @xmath3 contains no klein bottles , the length of either @xmath241 or @xmath24 must be @xmath2 .
[ clm : stronglyirred ] @xmath3 is a seifert fibered space over the 2-sphere with three exceptional fibers , where the orders of the exceptional fibers include the lengths of @xmath241 and @xmath24 .
furthermore , @xmath10 is not a boundary slope for @xmath20 .
as @xmath3 is atoroidal , @xmath630 compresses in @xmath631 . as @xmath3 is irreducible , the resulting @xmath49-sphere bounds a 3-ball .
first assume this 3-ball contains @xmath429 .
then the core of @xmath292 is an essential curve in @xmath136 that does not bound a meridian on either side of @xmath136 ( otherwise , as argued above , it in union with @xmath439 can be surgered to produce a lens space summand in @xmath3 ) . then lemma 3.3 of @xcite ( which shows
if a heegaard surface of a @xmath2-manifold other than @xmath4 contains a non - meridional simple closed curve that lies in a @xmath2-ball , then the splitting is weakly reducible ) contradicts our assumption that @xmath136 is strongly irreducible .
thus the 3-ball is disjoint from @xmath630 .
that is @xmath631 is a solid torus .
then @xmath3 is a seifert fibered space over the two sphere with three exceptional fibers with finite first homology .
furthermore , @xmath3 contains no incompressible surface .
thus theorem 2.0.3 of @xcite implies that if @xmath10 were a boundary slope then @xmath632 would contain an incompressible surface whenever @xmath633 .
since @xmath206 and @xmath634 is @xmath4 , this is impossible .
@xmath635 consists of at most @xmath636 arcs which can be isotoped ( rel @xmath630 ) to a collection of co - cores of the essential annulus in the seifert fibered space @xmath429 .
first note that each of @xmath637 can be isotoped off of @xmath638 ( resp . ) to intersect @xmath292 in a single point .
see figure [ fig : figure1 ] .
thus we may take @xmath635 to be a collection of @xmath636 arcs of the form @xmath639 .
each of these arcs can be isotoped to a co - core of the essential annulus @xmath640 , where @xmath317 is the core of @xmath292 : in particular , say @xmath241 is a length @xmath2 extended scharlemann cycle .
let @xmath242 be the disk of parallelism in @xmath127 between two edges of @xmath641 ( as shown in figure [ fig : figure1 ] ) .
then the arc @xmath642 , where @xmath14 is a point of @xmath242 , can be isotoped , rel @xmath630 , across @xmath643 to a cocore of @xmath644 in @xmath429 .
figure [ fig : figure2 ] shows an isotopy of @xmath1 moving an intersection with @xmath242 out to an edge of @xmath433 and then through the trigon face of @xmath241 to produce an intersection with @xmath645 ; the dashed circles show the resulting new intersections @xmath211 .
@xmath646 is a 2-torus .
let @xmath647 .
maximally compress @xmath531 in @xmath20 .
claim [ clm : stronglyirred ] says that @xmath10 is not a @xmath169-slope , thus @xmath531 compresses to @xmath169-parallel annuli in @xmath20 .
note that these annuli may be nested in @xmath20 .
each @xmath169-parallel annulus defines an isotopy in @xmath3 of a subarc of @xmath1 , with endpoints in @xmath648 , onto @xmath630 keeping the endpoints of the arc fixed .
furthermore , every point of @xmath648 belongs to such an arc of @xmath1 .
let @xmath649 be the subarcs of @xmath1 lying within the outermost of these @xmath169-parallel annuli and let @xmath650 be the complementary subarcs of @xmath1 .
note that the @xmath651 lie on the same side of @xmath630 .
then the isotopies described above corresponding to the outermost annuli deform @xmath1 to the union of the arcs @xmath650 along with arcs that lie in @xmath630 . as the complement of @xmath1 in @xmath3 is hyperbolic
, it must be that the arcs @xmath652 lie in @xmath429 . by the claim above ,
@xmath652 can be isotoped to co - cores of the essential annulus in @xmath429 .
let @xmath397 be a disjoint co - core of this annulus .
then @xmath653 is a genus @xmath49 splitting of @xmath3 .
furthermore , each of @xmath652 is @xmath169-parallel in @xmath654 .
perturbing each arc of @xmath655 into @xmath631 puts @xmath1 in bridge position with respect to this genus @xmath49 splitting . noting that @xmath656
, we have conclusion @xmath602 of lemma [ lem : parallelesc ] . in the remaining cases , the argument of case i gives the desired conclusion once we show how to modify @xmath439 and @xmath657 so that they are mutually disjoint and intersect @xmath292 only along their boundaries , @xmath658 .
we write @xmath659 where @xmath660 is the union of faces of @xmath382 ( resp . )
( thought of as disks in @xmath20 ) , and where @xmath661 are rectangles in @xmath662 describing an extension of @xmath660 across @xmath662 to form the long mbius band or twisted @xmath135-band . in
what follows ( at least up to case iv ) we show how to choose @xmath661 to be disjoint , making @xmath638 disjoint and intersecting @xmath292 along @xmath498 .
let @xmath663 be the extremal labels of @xmath241 , and @xmath664 the extremal labels of @xmath24 .
let @xmath665 or @xmath666 be the corners of @xmath241 depending on the length of @xmath241 . similarly let @xmath667 or @xmath668 be the corners of @xmath24 according to its length .
* case ii : * @xmath386 is a single interval of labels @xmath387 ( including a point interval ) .
first we consider the case of a point interval , that is , when @xmath669 ( and @xmath670 ) .
then @xmath671 intersect in a single point ( at vertex @xmath669 ) .
as @xmath671 both lie in @xmath292 , they must be non - transverse around vertex @xmath130 on @xmath127 .
this means that as one reads around vertex @xmath130 on @xmath127 , labels @xmath665 ( @xmath666 ) do not separate the labels @xmath667 ( @xmath668 ) .
see figure [ endof3fig5-alt ] for @xmath382 of length @xmath49 .
we choose disjoint @xmath672 as pictured in figure [ endof3fig7 ] ( with @xmath669 ) , making @xmath624 disjoint .
the argument for case i now gives a genus @xmath49 splitting of @xmath3 with bridge number at most @xmath620 ( @xmath1 intersects @xmath292 along @xmath673 only @xmath2 times here , rather than @xmath674 ) .
thus we may assume that @xmath675 .
let @xmath676 be the component of @xmath434 through vertex @xmath166 .
then @xmath676 intersects @xmath677 in a single point ( at @xmath166 ) .
see figures [ endof3fig6 ] , [ endof3fig6-theta ] .
again @xmath676 is disjoint from @xmath678 which lies in @xmath292 with @xmath677 , so @xmath676 must intersect @xmath677 tangentially . thus
, as one transverses vertex @xmath166 in @xmath127 the @xmath679 ( @xmath680 ) labels do not separate @xmath681 ( @xmath682 ) .
we then may choose disjoint @xmath661 as pictured in figures [ endof3fig7 ] , [ endof3fig7-theta ] .
the argument of case i now gives a genus @xmath49 splitting of @xmath3 with bridge number at most @xmath636 .
* case iii : * @xmath385 or @xmath384 the argument is the same in either case .
we assume @xmath385 .
we may assume that , say , @xmath683 .
let @xmath676 be the component of @xmath434 through vertex @xmath166
connecting @xmath166 to another vertex @xmath39 .
if @xmath684 , then @xmath382 have the same core labels , contrary to our assumption .
thus we assume @xmath685 ( @xmath686 by the parity rule ) .
then @xmath676 intersects @xmath677 in a single point ( at the vertex @xmath166 ) . since @xmath676 is disjoint from @xmath678 , and @xmath673
are contained in @xmath292 , @xmath676 must intersect @xmath677 tangentially . that is , as one transverses around the ( fat ) vertex @xmath166 of @xmath127 the labels @xmath665 ( @xmath666 ) are not separated by the labels @xmath687 ( @xmath688 ) . thus in @xmath662
, we may choose disjoint @xmath661 as pictured in figure [ fig : plmb2 ] ( which illustrates the case of length @xmath49 with @xmath39 between @xmath166 and @xmath45 ) thereby making @xmath638 disjoint .
we now apply the argument from case i giving a @xmath636 bridge presentation of @xmath1 .
* case iv : * @xmath383 contains all labels of @xmath128 , and @xmath243 overlaps @xmath390 in two intervals of labels : @xmath387 and @xmath689 .
this case is conclusion ( 1 ) of the lemma once we have shown that one of @xmath382 is of length @xmath2 .
so we assume that @xmath382 are both of length @xmath49 and we are as in figure [ fig : plmb5 ] .
if @xmath690 then @xmath691 are isotopic on @xmath136 and both go through vertices @xmath692 of @xmath136 .
thus @xmath693 can be amalgamated along their boundary to create an embedded klein bottle .
next assume that @xmath669 but @xmath670 .
let @xmath676 be the component of @xmath393 through vertex @xmath45 .
as @xmath676 is disjoint from @xmath694 and intersects @xmath695 once at @xmath45 , @xmath676 and @xmath695 intersect non - transversely .
thus around vertex @xmath45 , the labels @xmath696 do not separate @xmath687 . similarly , as @xmath697 intersect in a single point at vertex @xmath130 and yet are isotopic , their intersection is non - transverse .
that is , around vertex @xmath130 , the labels @xmath696 do not separate @xmath687 .
figure [ fig : plmb7 ] ( with @xmath669 ) shows that we can choose disjoint @xmath661 .
thus we may assume @xmath675 .
let @xmath676 be the component of @xmath393 through vertex @xmath166 , and let @xmath39 be the other vertex of @xmath127 to which @xmath676 is incident
. then @xmath698 .
assume @xmath699 . then as @xmath676 intersects @xmath695 once and is disjoint from @xmath694 , it must intersect @xmath695 non - transversely . that is ,
around vertex @xmath166 the labels @xmath696 do not separate @xmath687 .
let @xmath700 be the component of @xmath393 through vertex @xmath45 .
again , @xmath700 must intersect @xmath695 non - transversely at @xmath45 . hence around @xmath45 in @xmath127
, the labels @xmath696 do not separate @xmath687 .
thus we may choose disjoint disks @xmath661 in @xmath662 as pictured in figure [ fig : plmb7 ] .
this leaves us with the case that @xmath701 , whose argument is slightly different from the preceding ones . in case iv above @xmath684 .
this is the case when the core labels of @xmath382 are `` antipodal '' labels .
let @xmath676 be the component of @xmath393 through vertices @xmath166 and @xmath45 of @xmath127
. then @xmath676 and @xmath695 intersect twice . since @xmath676 is disjoint from @xmath694 which is isotopic to @xmath695 , the algebraic intersection number of @xmath676 and @xmath695 is @xmath36 .
thus we can choose @xmath702 in @xmath662 so that they are either ( 1 ) disjoint as in figure [ fig : plmb7 ] or ( 2 ) intersect in exactly two arcs as in figure [ fig : plmb8 ] .
this follows since the labels @xmath679 must separate @xmath681 either ( 1 ) around neither vertices @xmath703 or ( 2 ) around both vertices @xmath703 .
@xmath292 is the annulus on @xmath136 between @xmath694 and @xmath695 .
we may assume @xmath704 is disjoint from the vertices of @xmath127 ( else there is another component of @xmath393 or @xmath705 in @xmath292 , which we initially assumed is not the case ) .
consider @xmath706 where @xmath660 is the union of faces of @xmath382
. then @xmath707 are either ( 1 ) disjoint or ( 2 ) intersect in two double arcs ( from @xmath130 to @xmath166 and @xmath45 to @xmath708 along @xmath1 ) .
if ( 1 ) , @xmath3 contains an embedded klein bottle .
if ( 2 ) , @xmath709 is a klein bottle that self - intersects in a single double - curve ( again , we may assume @xmath707 are disjoint from @xmath292 except along @xmath710 ) .
the two preimage curves are disjoint from the cores of each of @xmath711 , and @xmath292 , and consequently bound disjoint disks , mbius bands in the pre - image .
we may surger along the double curve to obtain an embedded projective plane or klein bottle in @xmath3 .
this completes the proof of lemma [ lem : parallelesc ] .
[ cor : noparallelesc2 ] let @xmath618 be proper extended scharlemann cycles of @xmath129 of length @xmath49 . if there is an annulus , @xmath292 , in @xmath136 containing a component of @xmath393 and a component of @xmath705 then @xmath241 and @xmath24 have the same core labels . in this context , the proof of lemma [ lem : parallelesc ] constructs within @xmath3 a seifert fibered space over the disk with two exceptional fibers each of order @xmath49 .
but then @xmath3 would contain a klein bottle , a contradiction .
[ notwo ] assume @xmath136 is a strongly irreducible heegaard surface for @xmath3 . if @xmath618 are extended scharlemann cycles of @xmath129 of length @xmath49 or @xmath2 with the property that @xmath383 includes all labels , then either 1 .
@xmath712 ; or 2 .
@xmath3 is a seifert fibered space over @xmath349 with three exceptional fibers , where at least one of the exceptional fibers has order @xmath49 or @xmath2 .
furthermore , there is a genus @xmath49 heegaard splitting of @xmath3 with respect to which @xmath1 is at most @xmath674-bridge .
we may assume , by paring down as needed , that @xmath241 and @xmath24 are proper .
we may also assume @xmath713 ; hence @xmath714 .
if @xmath715 does not hold then we have @xmath716 .
note that the elements of @xmath717 are disjoint .
since a @xmath135-curve has euler characteristic @xmath718 , the maximal number of disjoint @xmath135-curves on @xmath136 , none of which lies in an annulus , is @xmath719 .
suppose @xmath241 has length @xmath2 .
since @xmath720 , there must be more than @xmath721 elements of @xmath717 that lie in annuli , and hence at least three such elements that are isotopic on @xmath136 and contain no further elements of @xmath717 between them .
the same holds _ a fortiori _ if @xmath241 has length @xmath49 . by lemma [ lem : triplets ] , either conclusion @xmath602 of lemma [ notwo ] holds , or there must be intersections of @xmath1 with an annulus in @xmath136 between a pair of these elements , @xmath722 .
we assume the latter .
such an intersection gives rise to a vertex of @xmath127 that must correspond to an element of @xmath390 that is not in @xmath243 .
thus there is an element , @xmath43 , of @xmath619 lying in the annulus of @xmath136 cobounded by @xmath722
. then @xmath43 must be parallel to and disjoint from , say , @xmath723 .
taking @xmath43 to be the nearest such to @xmath723 , we may assume @xmath724 cobound an annulus @xmath725 such that @xmath726 ( four from the vertices of @xmath724 and at most one more in the interior of @xmath292 belonging to an element of @xmath619 whose other vertex belongs to @xmath723 ) .
now pare down @xmath618 so that @xmath727 and @xmath728 . then , as @xmath43 contains a vertex which is not in @xmath243 , @xmath243 and @xmath390 can intersect in at most one label interval .
now lemma [ lem : parallelesc ] gives conclusion ( 2 ) above .
[ lem : existss22 ] let @xmath241 be a trigon face of @xmath131 which is not an extended scharlemann cycle
. then @xmath241 must contain a length @xmath49 scharlemann cycle in its interior .
in particular , assume @xmath241 is not an @xmath130-edge cycle .
that is , one corner , @xmath210 , of @xmath241 has no @xmath130 labels , and another corner , @xmath23 , has two .
then there is a length @xmath49 extended scharlemann cycle in @xmath241 on the arm opposite @xmath210 .
furthermore , if @xmath241 has a second scharlemann cycle in its interior , it must lie on the arm of @xmath241 between @xmath210 and @xmath23 .
assume that @xmath241 is an @xmath130-edge cycle .
then we may assume it is as in figure [ fig : xcycletrigon ] , where @xmath729 and @xmath730 are not all equal .
if there is no order @xmath49 scharlemann cycle in @xmath241 then we must have @xmath731 , @xmath732 , and @xmath733 .
all three inequalities can not simultaneously be true .
now assume @xmath241 is not an @xmath130-edge cycle
. we may take it as in figure [ fig : noncycletrigon ] .
assume there is no scharlemann cycle in the arm of @xmath241 opposite to @xmath210
. then @xmath476 can not be in @xmath734 on @xmath23 .
thus @xmath735 must be in @xmath736 on @xmath397 .
hence the length of the @xmath210-corner must be at least as large as that of @xmath23 and @xmath210 would contain a label @xmath130 .
now assume that @xmath241 contains a second scharlemann cycle in its @xmath737-arm .
refer to figure [ fig : sc2intrigon ] .
looking at the @xmath738-arm , we see that the interval @xmath739 on the extension of @xmath397 must be at least as large as @xmath740 on @xmath23 , which in turn is strictly larger than @xmath741 on @xmath397 . but @xmath741 on @xmath397 must be larger than @xmath742 on @xmath210 .
looking at the @xmath743-arm , though , we see that @xmath742 on @xmath210 is larger than @xmath739 on @xmath23 a contradiction .
[ def : trigontypes ] let @xmath241 be a trigon face of @xmath131 , considered as a subgraph of @xmath129 , that does not correspond to an extended scharlemann cycle of length @xmath2 . by lemma [ lem : existss22 ] , @xmath241 must contain a scharlemann cycle of length @xmath49 . as @xmath241 is a face of @xmath131 any such scharlemann cycle must be the core of an extended scharlemann cycle in @xmath241 that abuts the central trigon of @xmath241 and is disjoint from any @xmath130-edge of @xmath241 ( i.e. the extended scharlemann cycle lies within @xmath241 abutting the central trigon ) .
there can be at most two such scharlemann cycles within @xmath241 , which must lie in different arms of @xmath241 .
we say that @xmath241 is a _
type i trigon _ if there is only one and the corners of its central trigon are on three different pairs of labels .
it is of _ type ii _ if it contains two such scharlemann cycles .
type iii trigon _ otherwise that is , it contains a single length @xmath49 scharlemann cycle and the three corners in its central trigon represent two different label pairs .
[ def : t ] let @xmath744 be one of the two subgraphs of @xmath129 pictured in figure [ fig : trigonwithlongarms ] , where @xmath745 and either 1 .
@xmath746 and @xmath531 is the subgraph on the top . in this case
we say that @xmath531 is of _ type i_. furthermore , we require that @xmath747 , which is equivalent to no label appearing twice on any corner of @xmath531 . define @xmath748 to be the extended scharlemann cycle in @xmath531 with corner @xmath749 in the @xmath738-arm . note that @xmath531 has @xmath750 bigons in the @xmath738-arm beyond the extended scharlemann cycle @xmath748 , and @xmath429 bigons in the @xmath751- and @xmath752-arms .
2 . @xmath753 and @xmath531 is the subgraph on the bottom . in this case
we say that @xmath531 is of _ type ii_. furthermore , we require that @xmath754 , which is equivalent to no label appearing twice on any corner of @xmath531
. define @xmath748 and @xmath755 to be the extended scharlemann cycles in @xmath531 with corners @xmath756 and @xmath757 in the @xmath738- and @xmath751-arms respectively .
note that @xmath531 has @xmath750 bigons in the @xmath738- and @xmath751-arms beyond the extended scharlemann cycles @xmath748 and @xmath755 , and @xmath429 bigons in the @xmath752-arm . in this figure ,
@xmath758 are corners of vertices of @xmath129 , and we refer to the @xmath397-corner of @xmath531 as its _ ideal corner _ , denoted @xmath759 . note that every corner of @xmath531 ( as a labelled interval ) is contained in @xmath759 .
note also that there are @xmath760 labels appearing in @xmath531 which are not labels in an extended scharlemann cycle in @xmath531 .
[ lem : findt ] let @xmath241 be a trigon face of @xmath131 that corresponds to a trigon of type i or ii . let @xmath56 be the shortest distance ( along a corner of @xmath241 ) from @xmath130 to an extended scharlemann cycle within @xmath241 .
after possibly relabelling @xmath241 , there is a subgraph @xmath761 of @xmath241 such that @xmath761 is either type i or type ii according to whether @xmath241 is type i or type ii , and such that @xmath762 . in particular
, @xmath130 corresponds to a label in @xmath759 ( after relabelling ) .
assume @xmath241 is of type i. then we may assume after possibly relabelling that @xmath241 is as in the left or right of figure [ fig : findt1 ] according to whether @xmath241 is an @xmath130-cycle or not ( by lemma [ lem : existss22 ] ) .
note that @xmath763 as @xmath241 is of type i. ( if @xmath764 , then @xmath241 would be of type iii . by parity , @xmath765 or @xmath766 .
if @xmath767 then there would be a second scharlemann cycle on the @xmath768-arm . if @xmath769 then , regardless of whether @xmath241 is an @xmath130-cycle or not , there is a second scharlemann cycle along the @xmath752-arm . ) letting @xmath770 , @xmath241 contains at least @xmath429 bigons along the @xmath751- and @xmath752-arms , and @xmath750 out from the extended scharlemann cycle on the @xmath771-arm ( @xmath772 ; then from the @xmath738-arm deduce that the length of the interval from @xmath130 to @xmath773 is at least @xmath750 , the length of @xmath740 ) .
furthermore , @xmath774 .
if @xmath775 , then relabel using @xmath776 and take the involution of the trigon interchanging the @xmath397 and @xmath23 corners . we then have the desired @xmath761 .
so assume @xmath241 is of type ii . then after possibly relabelling @xmath241 , we can write @xmath241 as in figure [ fig : findt2 ] where the width of the upper extended scharlemann cycle is at most that of the lower and where either @xmath777 or @xmath778 ( using lemma [ lem : existss22 ] if @xmath241 is not an @xmath130-edge cycle ) .
the first condition gives that @xmath779 , implying @xmath780 as desired .
let @xmath781 .
then there are @xmath750 bigons beyond the extended scharlemann cycle on each of the @xmath782-arms of @xmath241 . as either @xmath783 or @xmath784
, there are @xmath429 bigons on the @xmath752-arm as well .
thus @xmath241 contains @xmath761 as desired . finally , note that @xmath762 . note
that , whether @xmath241 is type i or ii , the only label that appears twice on @xmath241 is the @xmath130-label .
thus the only label that can appear twice in the constructed @xmath761 must be the extremal ones on the ideal corner of @xmath531 .
as @xmath729 have the same parity , this is not possible . in @xmath761 ,
@xmath785 have the same parity by the parity rule of section [ sec : graphcombo ] .
thus @xmath786 .
[ def : l2 ] let @xmath787 has a trigon face which is of type i or type ii , but @xmath788 .
note that @xmath789 means that @xmath130 is not a label in an extended scharlemann cycle of length @xmath49 or @xmath2 .
[ def : cornersi&ii ] a subgraph @xmath241 of @xmath129 is said to be of _ type @xmath761 _ for some @xmath790 if after relabelling it becomes @xmath761 .
for such a @xmath241 , let @xmath791 be the corner of @xmath241 ( before relabelling ) corresponding to the ideal corner of @xmath761 .
let @xmath792 .
let @xmath793 be a smallest subcollection of @xmath794 such that for any @xmath795 there is a @xmath796 with @xmath797 .
lemma [ lem : findt ] guarantees the existence of @xmath793 . for @xmath798 ,
we denote by @xmath799 , its corresponding parameters .
thus @xmath800 .
[ atmost2sigma2 ] note that if three label intervals share a common label , then one of these intervals must be contained in the union of the other two .
thus the minimality of @xmath793 guarantees that if @xmath795 then @xmath801 for at most @xmath49 elements @xmath241 of @xmath793 .
remember that @xmath791 contains all labels appearing on the corners of @xmath241 .
[ def : partitionsii ] partition @xmath802 as follows : @xmath803 and @xmath804 .
@xmath805 @xmath806 @xmath807 @xmath808 @xmath809 @xmath810 @xmath811 @xmath812 @xmath813 @xmath814 @xmath815 @xmath816 we will abbreviate @xmath817 as @xmath818 .
one checks that definition [ def : partitionsii ] does indeed partition @xmath802 .
that is , @xmath819 iff @xmath820 .
furthermore , @xmath821 . to each of the partition elements of definition [ def : partitionsii ] we associate collections of essential closed curves that will eventually allow us to bound @xmath822 in theorem [ thm : type12boundn ] at the end of this section .
[ clm : j - k ] let @xmath761 corresponding to @xmath823 be of type i. let @xmath748 be the maximal extended scharlemann cycle for @xmath241 . for each @xmath824 there is a configuration @xmath825 in @xmath136 consisting of embedded curves @xmath826 in @xmath827 connected by a path of three edges of @xmath241 .
when @xmath828 , @xmath829 and @xmath830 are disjoint , when @xmath275 , they are identical . furthermore : 1 .
if @xmath831 , @xmath832 , share any vertices or edges then either * @xmath833 , and @xmath834 with no other shared vertices or edges ; or * @xmath835 , and @xmath836 with no other shared vertices or edges .
+ in particular , @xmath837 is a collection of @xmath838 disjoint curves .
thus @xmath839 2 .
if @xmath825 shares a vertex with an element of @xmath827 then that element is either @xmath829 or @xmath830 .
no @xmath825 lies in an annulus on @xmath136 .
it is convenient to work with @xmath241 after it is relabelled as @xmath761 ; however @xmath825 is taken to be under the original labelling . for @xmath824 ,
let @xmath829 be the essential curve of @xmath827 formed by the two @xmath840-edges in @xmath748 and let @xmath830 be the essential curve of @xmath827 formed by the two @xmath841-edges . @xmath829 and @xmath830 share no edges or vertices as long as @xmath828 .
otherwise , @xmath829 and @xmath830 are the same .
the distinct edges @xmath842 of figure [ fig : johntwo ] now join @xmath829 and @xmath830 on @xmath245 .
define @xmath825 to be the configuration on @xmath136 given by @xmath843 along with the vertices to which they are incident as shown in figure [ fig : johnthree ] . when @xmath275 , @xmath829 and @xmath830 are the same and the edges @xmath844 and @xmath845 are incident to different vertices of @xmath829 and on opposite sides of @xmath829 .
thus when @xmath275 , @xmath825 can not lie in an annulus on @xmath136 .
one checks that @xmath831 share vertices or edges only as described in item ( 1 ) above .
in particular , @xmath846 are disjoint . to check ( 2 ) , note that for @xmath825 the labels of @xmath847 are either those of @xmath829 or @xmath830 or lie outside the labels of @xmath748 .
now fix @xmath848 and assume that @xmath825 lies in an annulus on @xmath136 . as noted above , @xmath849 .
then @xmath830 and @xmath829 are different and cobound an annulus @xmath296 on @xmath136 .
see figure [ fig : johnthree ] . as the edges @xmath842 lie in @xmath296 , no vertex from @xmath766 to @xmath476 can lie in the interior of @xmath296 .
let @xmath292 be the annulus contained in @xmath850 with boundary @xmath851 obtained from the union of the appropriate bigons in @xmath748 .
note the annuli @xmath292 and @xmath296 will have disjoint interiors ( by ( 2 ) ) .
the orientations of @xmath830 and @xmath829 given by moving along an edge from labels @xmath397 to @xmath23 must be in the same direction along @xmath296 since otherwise @xmath852 would form a klein bottle .
this forces the labeling of the edges in figure [ fig : johnthree ] to be as pictured .
( this labeling is also forced because @xmath853 and @xmath854 have opposite parity . )
let @xmath443 be the torus @xmath852 which is necessarily separating .
we will arrive at a contradiction by finding an embedded curve that intersects @xmath443 once .
let @xmath855 be the arc which is the corner of figure [ fig : johntwo ] along @xmath397 that runs between the labels @xmath854 and @xmath476 within the extended scharlemann cycle @xmath748 .
noting that the curves @xmath856 must lie outside of @xmath296 ( since they are essential on @xmath245 and disjoint from @xmath857 ) , we see that @xmath855 lies on one side of @xmath443 . in particular @xmath855 intersects @xmath443 only at the vertex @xmath854 .
let @xmath858 be the arc that is the corner in figure [ fig : johntwo ] along @xmath397 running between labels @xmath859 and @xmath476 .
when @xmath860 the vertices @xmath861 must lie outside of @xmath296 , hence so do the edges of @xmath744 which are incident to @xmath858 in figure [ fig : johntwo ] . therefore @xmath858 lies entirely on one side of @xmath443 .
in particular , @xmath858 intersects @xmath443 only at the vertex @xmath859 .
when @xmath862 , then the vertices @xmath861 may share vertex @xmath863 with @xmath296 ( this happens when @xmath864 ) .
but the edge in @xmath531 connecting @xmath863 to @xmath865 in @xmath858 must lie outside of @xmath296 on @xmath127 as it is incident to vertex @xmath863 at label @xmath210 which from the labelling in figure [ fig : johnthree ] directs this edge outside of @xmath296 . again
, we conclude that @xmath858 intersects @xmath443 only at vertex @xmath859 . the arcs @xmath855 and
@xmath858 approach their endpoints at the vertices @xmath854 and @xmath859 , respectively , in opposite directions along @xmath1 . since the vertices @xmath854 and
@xmath859 have the same parity and lie on @xmath866 , @xmath855 and @xmath858 lie on opposite sides of @xmath443 .
hence @xmath867 forms a loop that may be perturbed to transversally intersect @xmath443 just once .
see figure [ fig : johnthreefive ] .
[ clm : k - j ] let @xmath761 corresponding to @xmath823 be of type ii .
let @xmath868 be the extended scharlemann cycles for @xmath241 with corners @xmath869 ( resp . ) in @xmath761 .
for each @xmath870 there is a configuration @xmath825 in @xmath136 consisting of disjoint curves @xmath826 in @xmath871 ( resp . )
connected by a path of three edges of @xmath241 . furthermore : 1 . @xmath831 share no vertices or edges when @xmath872 . in particular
, @xmath873 is a collection of @xmath874 disjoint curves on @xmath136 .
thus @xmath875 2 .
if @xmath825 shares a vertex with an element of @xmath827 or @xmath876 then that element is either @xmath829 or @xmath830 .
no @xmath825 lies in an annulus on @xmath136 .
we work with @xmath241 after it is relabelled as @xmath761 ; however @xmath825 is taken to be under the original labelling . for @xmath877 ,
let @xmath829 be the essential curve of @xmath827 formed by the two @xmath878edges in @xmath748 , and let @xmath830 be the essential curve of @xmath876 formed by the two @xmath879edges in @xmath755 .
define @xmath825 to be the connected subgraph of @xmath136 which is the union of @xmath829,@xmath830 , the edges @xmath857 of figure [ fig : johnfour ] and the vertices they connect as shown in figure [ fig : johnfive ] .
note that each of these edges is in @xmath761 since @xmath880 and that they are all distinct .
one now checks that @xmath831 share no vertices or edges if @xmath872 .
also the two vertices of @xmath825 not in @xmath881 lie outside @xmath827 and @xmath876 as they are labeled @xmath882 and @xmath854 , which implies conclusion ( 2 ) .
assume some @xmath825 lies in an annulus on @xmath136 .
then @xmath826 are parallel on @xmath136 , and corollary [ cor : noparallelesc2 ] implies that @xmath748 and @xmath755 have the same core labels , a contradiction . [ lem : differencekj ] @xmath883 let @xmath884 , with @xmath885 as in claims [ clm : j - k ] , [ clm : k - j ] . by claim [ clm : j - k ] , claim [ clm : k - j ] , and remark [ atmost2sigma2 ] , any vertex belongs to at most two elements of @xmath453 .
partition @xmath453 as @xmath886 where @xmath887 is disjoint from all other elements of @xmath888 , @xmath889 there exists @xmath890 such that @xmath155 and @xmath891 meet at a single vertex } , and @xmath892 there exists @xmath893 such that @xmath155 and @xmath891 meet at two vertices } .
note that for @xmath894 , the elements of @xmath895 are disjoint from the elements of @xmath896 .
let @xmath897 be a subcollection of disjoint curves in @xmath898 with @xmath899 .
let @xmath900 be the collection of simple closed curves on @xmath136 obtained from the elements of @xmath901 by removing all non - transverse points of intersection by small perturbations .
thus any two elements of @xmath900 either intersect transversely at a single point or are disjoint
. we may abstractly represent @xmath900 by a graph @xmath354 whose nodes correspond to the elements of @xmath900 and where two nodes are joined by an edge if and only if the corresponding elements of @xmath900 intersect . partition @xmath900 as @xmath902 where @xmath903 consists of the elements of @xmath900 that correspond to isolated nodes of @xmath354 .
the elements of @xmath904 then correspond to components of @xmath354 that are either paths with @xmath258 nodes , @xmath905 , or cycles with @xmath258 nodes , @xmath906 . in each path ( resp .
cycle ) component of @xmath354 with @xmath258 nodes , we can choose @xmath907 $ ] ( resp .
@xmath908 $ ] ) nodes such that the corresponding elements of @xmath904 are disjoint .
let the resulting subcollection of @xmath904 be @xmath909 .
let @xmath910 .
note that since any element of @xmath909 has algebraic intersection number @xmath184 with some element of @xmath904 , no element of @xmath909 is parallel on @xmath136 to any element of @xmath911 .
furthermore , if two elements of @xmath909 are parallel on @xmath136 then the corresponding nodes of @xmath354 belong to either a path component of @xmath354 with three nodes or a cycle component with four nodes .
removing one element from each parallel pair gives a subcollection @xmath912 of @xmath909 with @xmath913 , such that no element of @xmath912 is parallel to either another element of @xmath912 or an element of @xmath911 .
let @xmath914 .
then there exists @xmath915 such that @xmath155 and @xmath891 intersect , non - transversely , at a single vertex .
if @xmath916 were contained in an annulus in @xmath136 then by corollary [ cor : noparallelesc2 ] the corresponding core scharlemann cycles would have to have the same label pair and hence @xmath155 and @xmath891 would share two vertices , a contradiction .
it follows that in any parallelism class of elements of @xmath911 the corresponding elements of @xmath453 ( i.e. before perturbing to remove non - transverse intersections ) are disjoint .
we claim that no more than six elements of @xmath911 can be parallel on @xmath136 . for suppose
we have seven elements of @xmath911 that are parallel .
let @xmath917 be the outermost elements of this group .
let @xmath918 be elements of @xmath453 between these .
let @xmath919 and @xmath920 . in the notation of claims [ clm : j - k ] , [ clm : k - j ]
: if @xmath921 is of type i , then @xmath922 ; if @xmath921 is of type ii , then @xmath923 is @xmath924 or @xmath925 . then for each @xmath475
, @xmath926 must share a vertex with either @xmath927 or @xmath928 , else @xmath926 lies in an annulus , contradicting claims [ clm : j - k ] , [ clm : k - j ] .
as @xmath923 is disjoint from @xmath929 , we can assign to each @xmath930 a vertex @xmath931 which lies in @xmath927 or @xmath928 and is the closest such to @xmath923 in the path metric of @xmath926 . as @xmath927 and
@xmath928 each contain only two vertices , it must be that , without loss of generality , @xmath932 .
let @xmath933 with @xmath934 . by claims [ clm : j - k ] , [ clm : k - j ] , @xmath935 .
( say @xmath936 .
then @xmath937 , implying that @xmath938 are parallel .
this and the definition of @xmath939 imply that @xmath940 lies in an annulus . )
then remark [ atmost2sigma2 ] implies that @xmath941 . by claims [ clm : j - k ] ,
[ clm : k - j ] , it must be that either @xmath942 is of type i and @xmath943 ; or that @xmath942 is of type ii and @xmath944 . in either case , it must be that one of @xmath940 or @xmath945 , say @xmath940 , intersects @xmath928 in a vertex closer to @xmath946 in @xmath940 than @xmath939 , a contradiction .
hence there exists a subcollection @xmath947 of @xmath911 with @xmath948 such that no two elements of @xmath947 are parallel .
then @xmath949 is a collection of disjoint , essential , non - parallel simple closed curves on @xmath136 .
hence @xmath950 .
therefore @xmath951 as claimed .
[ clm:2k - j ] let @xmath761 corresponding to @xmath823 be of type i such that @xmath952 and @xmath953 .
let @xmath748 be the maximal extended scharlemann cycle for @xmath241 .
for each @xmath954 there is a configuration @xmath825 in @xmath136 consisting of disjoint curves @xmath826 in @xmath827 connected by a path of four edges of @xmath241 . 1 .
if @xmath831 , @xmath832 , share any vertices or edges then @xmath955 . in this case @xmath834 and
the edge @xmath844 of @xmath940 is the edge @xmath956 of @xmath945 . besides these edges and the vertices
to which they are incident , there are no other shared vertices or edges between @xmath831 . in particular
, @xmath957 is a collection of @xmath958 disjoint curves .
thus @xmath959 2 .
each vertex of @xmath127 lies in at most one element of @xmath960 .
if @xmath825 shares a vertex with an element of @xmath827 then that element is either @xmath829 or @xmath961 .
no @xmath825 lies in an annulus on @xmath136 .
we work with @xmath241 after it is relabelled as @xmath761 ; however @xmath825 is taken to be under the original labelling .
note that as @xmath241 is type i , @xmath962 . for @xmath954 let @xmath829 be the essential curve of @xmath827 on @xmath245 formed by the two @xmath963edges in @xmath748 ( note that @xmath964 ) .
let @xmath830 be the essential curve of @xmath827 formed by the two @xmath965edges in @xmath748 ( note that @xmath966 ) .
the four edges @xmath967 shown in figure [ fig : trigonwithlongarmsdetail ] , along with @xmath826 , give the configuration , @xmath825 , in the graph @xmath127 depicted in figure [ fig : nicelyconnected ] .
the vertices these edges connect are included in @xmath825 .
assume @xmath968 share vertices .
this can only happen if @xmath955 and the shared vertices and edges are as claimed .
in particular @xmath969 shares no vertices ( hence no edges ) with @xmath970 .
this verifies conclusions @xmath715 and @xmath602 .
as the edges @xmath967 lie outside of @xmath748 , conclusion @xmath48 holds .
assume for contradiction that @xmath825 lies in an annulus in @xmath136 .
let @xmath296 be the annulus in @xmath136 cobounded by @xmath826 .
let @xmath292 be the annulus contained in @xmath850 with boundary @xmath881 obtained from the union of the bigons in @xmath748 with labels @xmath971 and @xmath972 .
the labeling of figure [ fig : makeklein ] shows that the induced orientation on @xmath498 does not cancel an orientation on @xmath973 . note that in figure [ fig : makeklein ] we use that @xmath974 is even , and hence that vertices @xmath975 and @xmath853 are parallel .
thus @xmath852 is a klein bottle .
this is a contradiction .
[ clm : type1j ] let @xmath744 corresponding to @xmath823 be of type ii such that @xmath976 and @xmath977 .
let @xmath978 be the extended scharlemann cycles for @xmath241 with corners @xmath869 ( resp . ) in @xmath761 .
for each @xmath979 there is a configuration @xmath825 in @xmath136 consisting of disjoint curves @xmath826 in @xmath827 connected by a path of four edges of @xmath241 .
furthermore : 1 .
if @xmath970 shares vertices with @xmath945 , @xmath872 , then @xmath834 and @xmath980 share no other vertices or edges .
in particular , @xmath981 is a collection of @xmath982 disjoint curves .
thus @xmath983 2 .
each vertex of @xmath127 lies in at most one element of @xmath960 .
if @xmath825 shares a vertex with an element of @xmath827 or @xmath876 then that element is either @xmath829 or @xmath961 .
no @xmath825 lies in an annulus on @xmath136 .
we work with @xmath241 after it is relabelled as @xmath761 ; however @xmath825 is taken to be under the original labelling .
note that as @xmath241 is type ii , @xmath984 .
see figure [ fig : trigonwithlongarmsdetailii ] .
for @xmath979 let @xmath829 be the essential curve of @xmath827 on @xmath245 formed by the two @xmath878edges in @xmath748 ( note that @xmath985 ) .
let @xmath830 be the essential curve of @xmath827 formed by the two @xmath986edges in @xmath748 ( note that @xmath987 ) .
the four edges @xmath967 shown in figure [ fig : trigonwithlongarmsdetailii ] , along with @xmath826 , give a configuration , @xmath825 , in the graph @xmath127 .
the vertices these edges connect are included in @xmath825 .
note that these seven vertices of @xmath825 are distinct ( @xmath988 is even ) .
assume @xmath970 shares vertices with @xmath940 , @xmath872 . as the vertices of @xmath970 are labelled between @xmath766 and @xmath476 and the vertices @xmath989
are outside this range , @xmath970 can only share vertices with @xmath969 or @xmath990 . as these are curves of @xmath827 , either @xmath991 or @xmath834 .
as @xmath992 for each @xmath848 , it must be that @xmath834 ( corresponding to @xmath993 or @xmath994 ) .
this proves parts @xmath715 and @xmath602 of the claim . for parts @xmath48 and @xmath995 , we argue as for @xmath48 and @xmath995 of claim [ clm:2k - j ] . [
clm : boundcb ] for @xmath996 define @xmath997 .
then : @xmath998 consider the collection @xmath999 of all elements of @xmath960 taken ( without equating elements ) over all @xmath823 ( and hence over all @xmath1000 ) . by remark [ atmost2sigma2
] , each vertex of @xmath127 is a label in at most two elements of @xmath793 . on the other hand , for a fixed @xmath241 a vertex of @xmath127
lies in at most one element of @xmath960 .
thus any vertex belongs to at most two elements of @xmath999 .
partition @xmath999 as @xmath1001 as in the proof of lemma [ lem : differencekj ] .
again as in the proof of that lemma , we get disjoint collections of curves @xmath912 and @xmath911 , where no element of @xmath912 is parallel on @xmath136 to either another element of @xmath912 or an element of @xmath911 , and any pair of parallel elements of @xmath911 are ( before perturbation ) disjoint .
we claim that at most four elements of @xmath911 can be parallel on @xmath136 . for suppose we had five such parallel curves .
they all come from extended scharlemann cycles of length @xmath49 within elements of @xmath793 .
consequently , these five parallel , disjoint curves must arise from at most two different elements of @xmath1002 .
otherwise , by corollary [ cor : noparallelesc2 ] , the maximal extended scharlemann cycles in each of three different elements @xmath1003 of @xmath1002 must have the same core scharlemann cycles .
this would imply that the ideal corners of these elements would overlap , contradicting the minimality of @xmath793 ( see remark [ atmost2sigma2 ] ) .
thus three of these five parallel curves , @xmath1004 , all belong to @xmath960 for a single @xmath1005 .
they all belong to the same extended scharlemann cycle in @xmath241 .
take @xmath1006 to be between @xmath1007 .
then @xmath1008 coming from @xmath241 as described in claims [ clm:2k - j ] and [ clm : type1j ] .
if @xmath825 is disjoint from both @xmath723 and @xmath1009 , then @xmath825 lies in the annulus between @xmath1007 , contradicting these claims .
so we may assume that @xmath825 intersects , say , @xmath723 .
but then , again by these claims , @xmath722 must form the curves @xmath826 of @xmath825 .
( in these claims , if @xmath825 intersects an element of an extended scharlemann cycle in @xmath241 , then that element is @xmath829 or @xmath830 . )
then @xmath825 lies in the annulus between @xmath722 , contradicting these claims .
( if @xmath825 intersected @xmath1009 the same argument applied to @xmath1009 would equate it with @xmath723 . )
we thus get a subcollection @xmath947 of @xmath911 such that no two elements of @xmath947 are parallel and @xmath1010 .
therefore @xmath1011 .
hence @xmath1012 as claimed .
lemmas [ lem : differencekj ] and [ clm : boundcb ] linearly bound @xmath1013 where this sum is taken over @xmath1014 .
these bounds come from the collections @xmath885 and @xmath1015 . in a similar way ,
the following collections @xmath1016 will be used to linearly bound @xmath1017 where the sum is taken over the remaining @xmath1018 .
[ clm : typeiiesscurves ] let @xmath761 corresponding to @xmath823 be of type i such that @xmath1019 ( so that @xmath1020 ) . for each @xmath1021
there is an essential curve @xmath1022 in @xmath136 consisting of 6 distinct edges and vertices . furthermore , @xmath1023 intersects @xmath1024 , @xmath1025 , if and only if @xmath1026 is an element of @xmath1027 .
indeed , assume @xmath1026 is an element of @xmath1027 .
either 1 .
@xmath849 and @xmath1028 share exactly two vertices and a single edge spanning them .
after perturbing to be transverse along their common edge , @xmath1023 and @xmath1024 intersect transversely once . or 2 .
@xmath275 and @xmath1029 intersect in exactly two edges and the four vertices they connect . in this case
@xmath1028 can be perturbed along these edges so that they are transverse and intersect algebraically , geometrically twice .
as @xmath1030 , for each @xmath848 there is an @xmath1031 such that @xmath1032 . in particular ,
each @xmath1022 is essential in @xmath136 .
define @xmath1033 .
then the elements of @xmath1015 are disjoint from the elements of @xmath1016 .
furthemore , each vertex of @xmath127 belongs to at most two elements of @xmath1016 .
we work with @xmath241 after it is relabelled as @xmath744 ; however @xmath1022 is taken to be under the original labelling .
@xmath531 is shown in figure [ fig : smalltypeitrigon ] .
let @xmath1034 , @xmath281 , @xmath1035 , @xmath1036 , @xmath1037 , and @xmath1038 be the edges pictured in figure [ fig : smalltypeitrigon ] .
let @xmath1039 be the curve obtained by taking their union .
for @xmath1021 we may take the corresponding edges @xmath1040 , @xmath1041 , @xmath1042 , @xmath1043 , @xmath1044 , and @xmath1045 as we move out the arms of @xmath531 and form the curve @xmath1022 by taking their union .
thus we obtain a sequence of curves @xmath1046 on @xmath245 .
note that @xmath1047 . if @xmath1023 and @xmath1024 intersect on the heegaard surface then they intersect in vertices and possibly coincide along edges .
the labels of the vertices of @xmath1022 are @xmath1048 , @xmath1049 , @xmath1050 , @xmath1051 , @xmath1052 , and @xmath1053 in decreasing order .
each of these vertices is different , so @xmath1022 is embedded .
recall @xmath1054 . as no label appears twice along a corner of @xmath531 , the vertices of @xmath1023 and @xmath1024 overlap if and only if : 1 .
@xmath1055 ; @xmath1056 and @xmath1057 2 .
@xmath1058 ; @xmath1059 and @xmath1060 3 .
@xmath1061 ; @xmath1062 and @xmath1063 .
figure [ fig : smalltypeigf ] shows @xmath1022 as it lies on @xmath245 .
if @xmath1029 intersect then they agree exactly along a single edge ( note in figure [ fig : smalltypeitrigon ] that the sum of the labels at an edge determines the arm in which the edge lies , and no label appears twice along a corner of @xmath531 ) .
if @xmath828 , then @xmath1023 and @xmath1024 can be perturbed so that they intersect exactly once ( the @xmath758 labels that are not labels of edges of @xmath1023 , say , alternately appear on one side or the other of @xmath1023 ) . if @xmath275 , then @xmath715 and @xmath48 both hold ( @xmath1064 and @xmath1065 ) and one similarly checks that @xmath1028 intersect algebraically and geometrically twice after perturbation . in any case , each vertex of @xmath127 belongs to at most two elements of @xmath1016 .
the edges constituting elements of @xmath1015 come from the extended scharlemann cycle of @xmath241 , while the edges of an element of @xmath1016 lie outside the extended scharlemann cycle .
( note @xmath1066 . )
this verifies the last sentence of the claim .
[ clm : typeiesscurves ] let @xmath761 corresponding to @xmath823 be of type ii such that @xmath1067 ( so that @xmath1068 ) . for each @xmath1069
there is an essential curve @xmath1022 in @xmath136 consisting of 6 distinct edges and vertices . furthermore , @xmath1023 intersects @xmath1024 , @xmath1025 , if and only if @xmath1026 is an element of @xmath1070 .
indeed , assume @xmath1026 is an element of @xmath1070 .
either 1 .
@xmath1071 and @xmath1028 share exactly two vertices and a single edge spanning them . furthermore ,
after perturbing to be transverse along their common edge , @xmath1023 and @xmath1024 intersect transversely once . or 2 .
@xmath1072 and @xmath1029 intersect in exactly two edges and the four vertices they connect . in this case
@xmath1028 can be perturbed along these edges so that they are transverse and intersect algebraically , geometrically twice . as @xmath1073
, for each @xmath848 there is an @xmath1031 such that @xmath1074 . in particular ,
each @xmath1022 is essential in @xmath136 .
define @xmath1075 . then the elements of @xmath1016 are disjoint from those of @xmath1015 . furthermore
, a vertex of @xmath127 belongs to at most two elements of @xmath1076 .
we work with @xmath241 after it is relabelled as @xmath744 ; however @xmath1022 is taken to be under the original labelling .
@xmath531 is shown in figure [ fig : smalltypeiitrigon ] .
let @xmath1034 , @xmath281 , @xmath1035 , @xmath1036 , @xmath1037 , and @xmath1038 be the edges pictured in figure [ fig : smalltypeiitrigon ] .
let @xmath1039 be the curve obtained by taking their union .
for @xmath1069 we may take the corresponding edges @xmath1040 , @xmath1041 , @xmath1042 , @xmath1043 , @xmath1044 , and @xmath1045 as we move out the arms of @xmath531 and form the curve @xmath1022 in @xmath136 by taking their union .
thus we obtain a sequence of curves @xmath1077 on @xmath245 .
note that @xmath1078 . if @xmath1023 and @xmath1024 intersect on the heegaard surface then they intersect in vertices and possibly coincide along edges .
the vertices of @xmath1022 are @xmath1079 , @xmath1080 , @xmath854 , @xmath882 , @xmath1081 , and @xmath1082 .
these vertices are distinct and @xmath1022 is embedded .
furthermore , the vertices of @xmath1023 and @xmath1024 overlap if and only if : 1 .
@xmath1083 ; @xmath1084 and @xmath1085 2 .
@xmath1058 ; @xmath1086 and @xmath1087 3 .
@xmath1088 ; @xmath1089 and @xmath1090 .
figure [ fig : smalltypeiigf ] shows @xmath1022 as it lies on @xmath245 . as before one sees that if @xmath1028 intersect , then they must coincide along an edge .
they can then be perturbed so that they intersect exactly once except when @xmath1072 when ( 1 ) and ( 2 ) both hold . if @xmath1072 one checks that @xmath1023 and @xmath1024 intersect algebraically and geometrically twice after perturbation . in any case
, a vertex of @xmath127 belongs to at most two elements of @xmath1076 .
the edges constituting elements of @xmath1015 come from the extended scharlemann cycles of @xmath241 , while the edges of an element of @xmath1016 lie outside the extended scharlemann cycles .
this verifies the last sentence of the claim .
[ lem : curvestogether ] for @xmath1091 define @xmath997 . for @xmath1092
define @xmath1093 .
then : @xmath1094 consider the collection @xmath1095 of all elements of @xmath1096 taken ( without equating elements ) over all @xmath1097 .
let @xmath1098 be an element of @xmath1076 .
then @xmath1098 is disjoint from any element of @xmath960 , meets at most three other elements of @xmath1076 , and each vertex of @xmath1098 belongs to at most one other element of @xmath1076 .
also , any two elements of @xmath960 are disjoint . by remark [ atmost2sigma2 ] , each vertex of @xmath127 is a label in at most two elements of @xmath793 .
it follows that each element of @xmath1099 meets at most @xmath1100 other elements of @xmath1099 , the maximum occurring when an element @xmath1101 meets three other elements of @xmath1076 , and for each vertex @xmath1102 of @xmath1098 there exists @xmath1103 such that @xmath1102 belongs to two elements of @xmath1104 .
that is we find at least @xmath1105 disjoint curves coming from @xmath1095 .
we need to show that there are at most @xmath1106 such curves .
assume for contradiction that @xmath1107 .
then there must be a collection @xmath412 of six disjoint curves of @xmath1095 that are isotopic on @xmath136 .
[ clm : noa ] no element of @xmath412 belongs to @xmath1076 for @xmath1108 .
assume there were an @xmath1109 for some @xmath241 and @xmath848 as in claims [ clm : typeiiesscurves ] , [ clm : typeiesscurves ] . by the same claims
, we can find another element @xmath1110 coming from the same @xmath1076 that shares at least one edge with @xmath1111 and which can be perturbed to intersect @xmath1111 algebraically once or twice . as @xmath1110 must share at least one vertex with any element of @xmath412 that it intersects , one of the five curves of @xmath1112 must be disjoint from @xmath1110 . but all of these isotopic curves must have algebraic intersection number at least @xmath184 with @xmath1110 .
[ clm : aeps ] for some @xmath823 there are three elements of @xmath412 that belong to @xmath960 .
furthermore there is an extended scharlemann cycle @xmath748 in @xmath241 such that each of these three belong @xmath827 . by claim [ clm : noa ] ,
all the elements of @xmath412 come from @xmath1113 .
consequently they all come from extended scharlemann cycles of length @xmath49 within elements of @xmath793 . by corollary [ cor : noparallelesc2 ]
, these extended scharlemann cycles must have the same core labels . as each of these core labels will lie in the ideal corners of the corresponding elements of @xmath793 , remark [ atmost2sigma2 ] implies that these six parallel , disjoint curves in @xmath412 belong to at most two different elements of @xmath793 .
hence three belong to the same @xmath823 .
again by corollary [ cor : noparallelesc2 ] they must belong to the same extended scharlemann cycle within @xmath241 .
let @xmath1004 be three parallel curves of claim [ clm : aeps ] with @xmath1006 between @xmath1007 .
then @xmath1008 coming from @xmath241 as described in claims [ clm:2k - j ] and [ clm : type1j ] .
if @xmath825 is disjoint from both @xmath723 and @xmath1009 , then @xmath825 lies in the annulus between @xmath1007 , contradicting these claims .
so we may assume that @xmath825 intersects , say , @xmath723 .
but then , again by these claims , @xmath722 must form the curves @xmath826 of @xmath825 .
( in these claims , if @xmath825 intersects an element of an extended scharlemann cycle in @xmath241 , then that element is @xmath829 or @xmath830 . )
then @xmath825 lies in the annulus between @xmath722 , contradicting these claims .
( if @xmath825 intersected @xmath1009 the same argument applied to @xmath1009 would equate it with @xmath723 . ) after lemmas [ lem : differencekj ] and [ clm : boundcb ] , to get a linear bound on @xmath1114 we need only bound @xmath1115 .
this can be done as in the proof of lemma [ lem : curvestogether ] using only the collections @xmath1076 . by including the @xmath960 above
, we get a slightly improved bound .
[ thm : type12boundn ] @xmath1116 that is , there are at most 960(g-1 ) labels @xmath130 in @xmath128 such that @xmath131 has a trigon face which is of type i or type ii but @xmath789 .
the first inequality comes from definition [ def : cornersi&ii ] , so we focus on the second which follows from : @xmath1117 inequality @xmath1118 comes from lemma [ lem : curvestogether ] , lemma [ lem : differencekj ] , and lemma [ clm : boundcb ] . for inequality @xmath1119 we take the definitions of @xmath885 , @xmath960 , and @xmath1076 coming from claims [ clm : j - k ] , [ clm : k - j ] , [ clm:2k - j ] , [ clm : type1j ] , [ clm : typeiiesscurves ] , and [ clm : typeiesscurves ] along with the stipulations that if @xmath996 then @xmath997 and if @xmath1120 then @xmath1093 . the estimates of figure [ fig : sumtable ] verify inequality @xmath1119 .
these estimates arise from the partition of @xmath802 in definition [ def : partitionsii ] along with the counts listed in the above claims . in figure
[ fig : sumtable ] , each quantity in the left hand side of the second column is at least @xmath1121 . to see this , note that @xmath1122 for @xmath1123 and @xmath1124 for @xmath1125 , and then further note the upper bounds on @xmath1126 for @xmath1127 given in definition [ def : partitionsii ]
this section is primarily dedicated to elucidating the structure of type iii trigons which are cycles .
we construct a collection @xmath1128 of such trigons that captures all the labels @xmath130 of @xmath129 that give rise to type iii trigons ( whether @xmath130-cycles or not ) that do not belong already to @xmath138 .
recall that a type iii trigon on the label @xmath130 is a trigon face of @xmath131 which , regarded as a subgraph of @xmath129 , has a single scharlemann cycle ( of length @xmath49 ) and whose central trigon face ( of @xmath129 ) has only two corner types .
figure [ fig : figuretypeiii ] shows a generic ( as we shall see below ) type iii @xmath130-cycle trigon where the central trigon face is in the two corner types @xmath1129 . we now define the terms _ ideal corner _ , _ trunk _ , and _ gnarl _ associated to a type iii trigon .
let @xmath241 be a type iii trigon on @xmath130 which is a cycle ( i.e. the edges of @xmath241 form a cycle in @xmath131 so that , when oriented , the tails have label @xmath130 ) .
let @xmath1130 be the labels of its core scharlemann cycle .
let @xmath748 be the extended scharlemann cycle containing the core scharlemann cycle of @xmath241 which abuts its central trigon face .
there is a label @xmath166 such that @xmath241 has exactly two edges , @xmath1131 with labels @xmath1132 .
one can see that @xmath1133 @xmath1134 .
that is , around a vertex of @xmath129 , the labels @xmath1132 are equidistant from labels @xmath1130 .
define the _ ideal corner of @xmath241 _ , denoted @xmath791 , to be the label interval on a vertex of @xmath129 between labels @xmath130 and @xmath166 that contains labels @xmath1130 .
these edges @xmath1131 lie on different arms , @xmath1135 of @xmath241 .
the ideal corner @xmath791 appears twice as an actual corner of @xmath241 in @xmath129 once between @xmath1131 . for any pair of labels @xmath1136 in @xmath791 such that
@xmath1137 @xmath1134 , there are exactly two edges of @xmath129 in @xmath241 with endpoints @xmath1136 .
these edges lie in the arms @xmath1138 of @xmath241 .
the subgraph of @xmath127 formed by the vertices @xmath1136 and these two edges is a circle in @xmath115 . if it is an element of @xmath827 we call it a _ trunk _ of @xmath241 ; otherwise we refer to it as a _ gnarl _ of @xmath241 ( see also definition [ def : simpleandwrapping ] )
. to sum up , the ideal corner @xmath791 has the following properties : * every label in @xmath791 belongs to a unique gnarl or trunk of @xmath241 .
* @xmath130 is a label in @xmath791 .
* @xmath791 is symmetric about the consecutive labels @xmath1130 of the core of @xmath241 . as a subgraph of @xmath129 ,
two of the actual corners of @xmath241 are the same as @xmath791 .
the other either properly contains @xmath791 or is properly contained in it . after paring down @xmath241
, we may assume this corner is properly contained in @xmath791 .
note that the type iii trigon resulting from this paring down has the same ideal corner but is now a cycle trigon in @xmath166 rather than @xmath130 .
[ def : sigmaiii ] let @xmath1139 be the collection of labels of @xmath128 that are in the ideal corners of cycle trigons but are not the labels of extended scharlemann cycles of length @xmath1140 .
that is , @xmath1141 .
let @xmath1142 be a collection of smallest cardinality of type iii cycle trigons with the property that @xmath1143 .
we may assume that each element , @xmath241 , of @xmath1128 has been pared down so that each actual corner of @xmath241 is a subset of its ideal corner .
[ sigma2remarks1 ] 1 . for every @xmath1144 such that @xmath131 has a trigon which is a type iii non - cycle trigon
, there is a @xmath1145 with @xmath801 , by lemma [ lem : noncycleiii ] below . 2 .
item ( 1 ) and the definition of @xmath1142 imply : the ideal corners of elements of @xmath1128 include all labels , @xmath130 , of @xmath1146 such that @xmath131 has a type iii trigon .
the elements of @xmath1142 have different core scharlemann cycles ( by the minimality of @xmath1142 , because the ideal corners are symmetric about the core labels ) . as different core scharlemann cycles can not lie on a common annulus in @xmath136 ( else
there is a klein bottle ) , we can construct a family of @xmath1147 different curves on @xmath136 that satisfy property @xmath319 .
hence @xmath1148 , by lemma [ lem : fkg ] .
[ lem : noncycleiii ] let @xmath1144 . if there is a type iii @xmath130trigon of @xmath129 that is not an @xmath130cycle , then there is a trigon @xmath1145 such that @xmath801 .
the @xmath130-trigon has one corner , @xmath397 , with both endpoints labelled @xmath130 .
it has a corner , @xmath210 , with no label @xmath130 , and a corner , @xmath23 , with exactly one endpoint labelled @xmath130 .
let @xmath1149 be the labels of the central trigon face at @xmath397 .
let @xmath1150 be the other labels of this central trigon , such that @xmath1151 have opposite parity .
two of the edges of the @xmath130-trigon will have labels @xmath1132 where @xmath1152 @xmath1134 .
using the fact that no other labels @xmath130 appear on @xmath758 , one sees that , up to whether the labels increase or decrease as you move clockwise along these corners , the @xmath130-trigon appears as in figure [ figurenoncyclev2 ] . inside this @xmath130-trigon
is the pictured @xmath166-cycle , forming a type iii cycle trigon .
the ideal corner of this type iii @xmath166-cycle trigon contains @xmath130 ( it is given by @xmath23 in the figure ) . by definition
, there is a @xmath1145 such that @xmath791 includes all elements in this corner . let @xmath241 be a type iii cycle trigon . if @xmath1153 are labels in @xmath791 , then @xmath1154 is the interval of labels between them contained in @xmath791 .
let @xmath1145 be as above .
then up to the ordering of the labels around a corner ( i.e. whether they increase as you move clockwise or anti - clockwise around a corner ) and rechoosing which label is called @xmath130 , we may assume @xmath241 appears as in figure [ fig : figuretypeiii ] .
the core scharlemann cycle of @xmath241 has labels @xmath1130 where @xmath1155 .
there @xmath791 occurs along the corners labelled @xmath1156 .
the shorter corner is labelled @xmath210 .
the lower arms in the figure are the arms @xmath1135 .
pair the vertices in the disjoint intervals @xmath1157 and @xmath736 .
each such pair of vertices has two edges in @xmath241 that forms a gnarl on @xmath136 .
see figure [ fig : figuregnarl ] .
[ property:1 ] the @xmath1158 labels at the ends of edges of @xmath1145 ( figure [ fig : figuretypeiii ] ) incident to a gnarl of @xmath241 ( but not belonging to a gnarl of @xmath241 ) lie on the same side of the gnarl in @xmath136 . see figure [ fig : figuregnarl ] .
specifically , let @xmath1159 be the labels in @xmath241 of two edges @xmath1131 forming a gnarl , with @xmath39 the label at corner @xmath397 . there is another edge , @xmath845 , in @xmath241 incident to the @xmath397-corner with label @xmath848 .
assume there is another edge , @xmath956 , in @xmath241 with label @xmath39 at the @xmath210-corner .
consider these edges in @xmath136 .
then @xmath1160 lie on the same side in @xmath136 of the gnarl @xmath1161 .
( in figure [ fig : figuregnarl ] , @xmath1162 is @xmath1163 and @xmath1164 . ) set @xmath1165 .
we separate the edges in the upper arm of @xmath241 into mod @xmath45 classes : let @xmath56 be such that @xmath1166 [ def : simpleandwrapping ] let @xmath39 be such that @xmath1167 . note that @xmath1168 .
we may now label @xmath241 as in figure [ figuretypeiiirefined ] . for each
@xmath39 we see in @xmath245 the _ component of @xmath241 _ pictured in figure [ figurecomponentofsigma ] .
note that a gnarl in this component may or may not be essential on @xmath245 if essential we call it a _ wrapping gnarl _ , if inessential we call it a _
simple gnarl_. the _ interior _ of a simple gnarl is the interior of the disk it bounds on @xmath245 .
recall that @xmath748 is the extended scharlemann cycle in @xmath241 .
the component of @xmath241 ( a subset of @xmath245 ) corresponding to @xmath39 consists of a curve of @xmath827 between the vertices with labels @xmath1169 , called the _ trunk _ of the component , along with two _ branches _ : @xmath1170 and @xmath1171 .
we think of a branch of the component of @xmath241 as a path of gnarls in @xmath136 .
we orient the branch by labelling its vertices 0,1,2 , as pictured on the left of figure [ figurebranchorder ] ( which shows this labeling for the two branches of a component ) .
the right side of that figure pictures the corresponding labelling on the trigon @xmath241 .
if @xmath1172 is a branch of @xmath241 and @xmath130 is a vertex of this branch , let @xmath1173 be the label coming from the oriented @xmath1174 ( e.g. @xmath1175 for @xmath1176 the first branch described above . ) observe : * each branch terminates in a gnarl ( @xmath241 has been pared down , preserving @xmath791 , so that its third corner is properly contained in @xmath791 ) . * every label in @xmath791 lies in a unique gnarl or unique trunk ( curve of @xmath827 ) , and in a unique component of @xmath241 .
note that if a label corresponds to the vertex of a trunk curve then it is in @xmath138 . by examining the oscillation of the labels \{0,1 , } along @xmath791 in figure [ figurebranchorder ] , we get the following .
[ property:2 ] let @xmath1177 be vertices on a branch @xmath1172 of @xmath241 .
then @xmath1132 are of the same sign ( as vertices of @xmath127 ) if and only if @xmath1178 @xmath1179 .
furthermore , assume @xmath1180 and @xmath1181 @xmath1179 . if either * @xmath1182 ; or * @xmath1183 and @xmath1184 @xmath1179 then @xmath1185 . in particular , if @xmath1132 form a gnarl of @xmath1172 and @xmath708 immediately precedes this gnarl along @xmath1172 , then @xmath1186 .
every vertex in @xmath127 corresponding to a label in @xmath241 lies in a unique component of @xmath241 that includes an essential curve from @xmath827 , its trunk , and is connected to that curve by edges in its branch .
such a vertex in the interior of a simple gnarl of @xmath241 must lie on a component of @xmath241 that , by property [ property:1 ] , would lie entirely within the interior of the gnarl .
in particular , the essential trunk curve of this component lies within the simple gnarl , a contradiction .
thus we have the following : [ property:3 ] let @xmath1145 .
a simple gnarl of @xmath241 contains no vertices in its interior which are labels of @xmath241 .
[ lem : entersonce ] thinking of a branch as a directed path of gnarls in @xmath136 , if a branch of @xmath1128 enters the interior of a simple gnarl of @xmath1128 , it never leaves .
that is , let @xmath1187 .
let @xmath1188 be a simple gnarl of @xmath942 and @xmath1189 be a branch of @xmath1190 .
if a vertex ( of @xmath127 ) , @xmath130 , of @xmath1189 lies in the interior of @xmath1188 , then at least one of the vertices in @xmath1188 is in @xmath1189 .
furthermore , if @xmath166 is a vertex in both @xmath1188 and @xmath1189 then @xmath1191 .
the vertex @xmath130 of @xmath1189 in the interior of @xmath1188 must belong to a simple gnarl @xmath1192 of @xmath1189 ( otherwise the vertex belongs to an essential simple closed curve intersecting @xmath1188 at most once ) . let @xmath1193 be the vertices of @xmath1188 and @xmath1194 the vertices of @xmath1192 .
see figure [ figureinandout ] .
as @xmath1189 ( as a directed path in @xmath115 ) must pass through @xmath1188 to reach @xmath130 from its trunk , we may take @xmath1195 so that it is in @xmath1189 such that @xmath1196 .
we take @xmath1197 to be the same sign as @xmath158 .
possibly @xmath1198 and @xmath1199 or @xmath1200 and @xmath1201 .
we assume for contradiction that @xmath1202 is also in @xmath1189 and that @xmath1203 ( @xmath1194 are consecutive along @xmath1189 ) .
note that together @xmath1204 does not account for all labels .
( otherwise lemma [ lem : nonemptygnarl ] and property [ property:3 ] applied to @xmath1192 implies that @xmath1192 must be the only gnarl of @xmath1189 .
lemma [ 3ingnarl ] says that there must be a vertex in the interior of @xmath1188 that does not belong to @xmath1192 . by property [ property:3 ] , this vertex must belong then to a trunk of @xmath1189 .
but the trunk gives rise to an essential curve on @xmath115 intersecting @xmath1188 at most once . )
thus @xmath1205 overlap on a label interval .
let @xmath1206 be the interval between @xmath1193 included in both @xmath1207 and @xmath1208 .
* case ( i ) : * @xmath1209 applying property [ property:2](b ) to the sequence @xmath1210 on the branch @xmath1189 ( with @xmath1211 ) , @xmath1212 .
but then @xmath1213 and in the interior of @xmath1188 , contradicting property [ property:3 ] .
* case ( ii ) : * @xmath1214 and @xmath1201 .
then @xmath1215 is in the interior of @xmath1188 . as @xmath1216 @xmath1179 and @xmath1217
there must be another vertex @xmath1218 in @xmath1189 in the interior of @xmath1188 such that @xmath1219 and such that @xmath1220 @xmath1179 .
applying property [ property:2](b ) to the sequence @xmath1221 we conclude that @xmath1222 .
that is , @xmath1218 is a label in @xmath942 , contradicting property [ property:3 ] . by taking a union of faces in the lower arms of figure [ figuretypeiiirefined ] ,
we get : [ property:4 ] let @xmath1145 .
two gnarls in @xmath241 cobound an embedded annulus constructed from the faces of @xmath241 .
[ lem : nonemptygnarl ] let @xmath1223 .
if any simple gnarl of @xmath241 contains no vertices of @xmath127 in its interior , then this gnarl is the only gnarl of @xmath241 .
( i.e. all other labels in @xmath241 are parts of trunks of components of @xmath241 ) .
there are at most @xmath273 such gnarls .
the proof of this lemma divides into two cases according to whether or not the extended scharlemann cycle within @xmath241 is actually just a scharlemann cycle . * case ( i ) : *
the extended scharlemann cycle within @xmath241 is a scharlemann cycle .
let @xmath1224 be the bigon bounded by this scharlemann cycle .
see figure [ figureemptygnarl1 ] for the labeling we will use .
the _ initial gnarl _ is the gnarl on vertices @xmath1225 .
we assume that @xmath241 has more than the initial gnarl , in particular , that there is a face @xmath1226 as pictured .
assume some simple gnarl of @xmath241 bounds a disk @xmath1227 in @xmath245 whose interior is disjoint from @xmath1 .
let @xmath296 be the annulus between this gnarl and the initial gnarl given by property [ property:4 ] .
( if the initial gnarl bounds @xmath1227 then @xmath1228 . ) set @xmath1229 . then @xmath203 is a disk bounded by the initial gnarl and can be taken to have interior disjoint from both @xmath1 and the faces of @xmath241 .
let @xmath1230 be the rectangle on @xmath198 bounded by the curves @xmath1231 , @xmath476 , @xmath1232 and @xmath1233 that is disjoint from @xmath210 . let @xmath1234 be the rectangle on @xmath198 bounded by the curves @xmath1235 , @xmath1236 , @xmath1237 and @xmath1238 that is disjoint from @xmath397 .
( by the curves @xmath1239 on @xmath198 , we mean the curves that are the boundaries of those vertices . ) in fact , let @xmath1240 be two push - offs of the face @xmath1224 so that @xmath1224 and @xmath1241 agree along their @xmath1242-edge and so that @xmath1243 agrees with @xmath1244 along its @xmath1242-edge .
modify @xmath1245 along @xmath27 between @xmath1246 by deforming in slightly .
see figure [ figurerectangles - alt ] .
form the disk @xmath1247 of figure [ figurethindisk1-alt ] .
this gives a thinning long disk @xmath1248 taking the subarc @xmath1249 of @xmath1 ( view @xmath1249 as part of @xmath1250 ) to the arc on @xmath245 pictured in figure [ figuresurface1 ] .
this contradicts the thinness of @xmath1 and finishes the proof of case ( i ) .
* case ( ii ) : * the extended scharlemann cycle in @xmath241 is not a scharlemann cycle ( has length at least @xmath49 ) .
see figure [ figureemptydisk2 ] .
we assume that @xmath241 has more than one gnarl , giving rise to face @xmath1251 in the figure .
let @xmath1234 and @xmath1230 be the rectangles on @xmath198 bounded by @xmath1252 and @xmath1253 , respectively .
see figure [ figurerectangles2 ] . as in case ( i )
, the assumption of the empty simple gnarl allows us to construct a disk @xmath203 in @xmath20 which is disjoint from the faces of @xmath241 and from @xmath1 .
construct the thinning disk @xmath1254 depicted in figure [ figurethindisk2 ] .
@xmath1248 gives an isotopy of @xmath1249 onto the arc in @xmath245 indicated in figure [ figurethindisk2 ] thereby thinning @xmath1 .
this contradicts the thinness of @xmath1 and finishes the proof of case ( ii ) . for the last statement in the lemma ,
recall from remark [ sigma2remarks1](3 ) that @xmath1255 . by the above
, the gnarls containing no interior vertices belong to different elements of @xmath1128 .
[ 3ingnarl ] let @xmath1223 .
if a gnarl in @xmath241 has vertices of @xmath127 in its interior , it must have at least three .
in particular , if @xmath241 has more than one gnarl , then any simple gnarl of @xmath241 must contain at least three vertices in the interior of the disk it bounds on @xmath245 .
if a simple gnarl only has one vertex in its interior then the graph within would have a @xmath184sided face ( a contradiction ) , because the number of labels on the interior of the gnarl is less than @xmath191 ( see figure [ figurecomponentofsigma ] for the labeling around a gnarl ) , the total number of labels ( recall that @xmath15 is the denominator of the dehn surgery coefficient ) .
now assume the simple gnarl on vertices @xmath1132 has exactly two vertices @xmath939 and @xmath1256 in its interior .
there are fewer than @xmath126 edges connecting @xmath1257 .
otherwise they are all parallel and describe a cable space in the exterior of @xmath1 ( section 5 of @xcite ) contradicting that @xmath1 is hyperbolic .
thus there are more than @xmath1258 labels on @xmath1132 that are endpoints of edges connecting @xmath1257 to @xmath1132 ( no 1-sided faces ) .
but there are fewer than @xmath191 such labels on @xmath1132 in the interior of the gnarl .
this is a contradiction since @xmath15 is at least @xmath49 . by lemma [ lem : nonemptygnarl ] ,
if @xmath241 has more than one gnarl , any of its simple gnarls must contain vertices in its interior .
[ lem : parallelwrapping ] let @xmath1223 .
let @xmath1172 be a branch of ( a component of ) @xmath241 .
let @xmath1259 be the trunk of @xmath1172 and @xmath1260 be a gnarl of @xmath1172 . if @xmath1259 and @xmath1260 cobound an annulus @xmath292 on @xmath115 , then either @xmath1172 is disjoint from @xmath704 or @xmath3 is a seifert fibered space over the 2-sphere with three exceptional fibers , one having order @xmath49 and another order @xmath2 . in this latter case , if @xmath136 is from a genus @xmath49 splitting of @xmath3 with respect to which @xmath1 has smallest bridge number , then @xmath1261 .
let @xmath292 be the annulus in @xmath245 cobounded by @xmath1260 and @xmath1259 .
if @xmath241 contains a simple gnarl @xmath1262 , then let @xmath1263 be the disk in @xmath245 that @xmath1262 bounds .
property [ property:4 ] gives an annulus @xmath296 from @xmath1262 to @xmath1260 ( and transverse to @xmath136 there ) arising from @xmath241 .
therefore , @xmath1260 bounds the embedded disk @xmath1264 transverse to @xmath115 along @xmath1260 ( by property 3 , @xmath296 is disjoint from the interior of @xmath1263 ) . on the other hand
, @xmath1259 bounds a long mbius band , @xmath163 , coming from the extended scharlemann cycle of @xmath241 which is transverse to @xmath115 at @xmath1259 .
then @xmath1265 is an immersed projective plane that can be surgered to produce an embedded projective plane in @xmath3 a contradiction .
therefore @xmath241 contains no simple gnarls .
we assume @xmath1172 intersects @xmath704 .
then property [ property:1 ] along with the fact that @xmath1172 contains no simple gnarls guarantees that @xmath1260 is the only gnarl in @xmath1172 .
that is , @xmath1172 consists of @xmath1259 , @xmath1260 , and a single edge @xmath1224 connecting them that lies in @xmath292 . without loss of generality ,
we may assume we are as in figure [ fig : figureparallelwrapping ] .
note that @xmath704 contains no vertices that appear as labels in @xmath241 as any such would be part of an essential curve in @xmath704 disjoint from @xmath1224 .
similarly , @xmath1224 is the only edge of @xmath241 contained in @xmath704 .
let @xmath1266 , @xmath1267 , and @xmath1268 be the consecutive handles in @xmath662 running between vertices @xmath1269 .
let @xmath1270 where the @xmath1271 are the disks in @xmath241 from figure [ fig : figureparallelwrapping ] .
then @xmath1272 has presentation @xmath1273 where @xmath1132 correspond to @xmath1274 , respectively , @xmath1275 are retracted to a base point , and @xmath317 represents the core of @xmath292 .
this can be rewritten @xmath1276 ( using @xmath1277 ) .
thus @xmath429 is a trefoil knot exterior and furthermore the core of @xmath292 is its meridian ( attaching a disk along @xmath317 kills the group ) .
let @xmath1278 be a thickening of the long mbius band coming from the extended scharlemann cycle , and consider the the submanifold @xmath1279 . as @xmath3 is atoroidal
, @xmath1280 must compress in @xmath1281 .
as @xmath3 is irreducible , compressing @xmath1280 gives a 2-sphere which bounds a 3-ball in @xmath3 . if this 3-ball contained @xmath1282
, then @xmath3 would contain a projective plane constructed from a disk in the @xmath2-ball bounded by the meridian of @xmath429 and the long mbius band .
thus the 3-ball is disjoint from @xmath1282 , and @xmath1281 is a solid torus .
this implies that @xmath3 is a seifert fibered space over the 2-sphere with three exceptional fibers , one of which has order @xmath49 and another of order @xmath2 .
so we now assume that @xmath136 is a genus @xmath49 splitting with respect to which @xmath1 has smallest bridge number .
let @xmath1283 .
let @xmath203 be the disk in @xmath292 gotten by removing from @xmath292 an open neighborhood of @xmath1284 .
then @xmath1 can be perturbed so that it intersects @xmath1282 in @xmath1285 arcs , each of the form @xmath1286 in @xmath1287 ( e.g. perturb @xmath1 from label @xmath45 to label @xmath708 between @xmath397 and @xmath23 in @xmath662 ) .
now @xmath1288 is a genus @xmath49 handlebody as neighborhoods of an arc in @xmath1289 and an arc in @xmath1290 give a pair of meridians .
furthermore , the first of these meridians shows that the annulus along which @xmath1291 and @xmath1292 meet is primitive in @xmath1291 .
thus @xmath1293 is a genus @xmath49 handlebody .
note that the surgery curve @xmath10 is not a boundary slope for @xmath20 , the exterior of @xmath1 .
for as @xmath3 contains no incompressible surface , this would contradict theorem 2.0.3 of @xcite .
consider the punctured torus @xmath1294 .
maximally compress @xmath531 in @xmath20
. then @xmath531 compresses to @xmath169-parallel annuli .
each @xmath169-parallel annulus defines an isotopy in @xmath3 of a subarc of @xmath1 , with endpoints in @xmath1295 , onto @xmath1280 and keeping the endpoints of the arc fixed .
furthermore , every point of @xmath1295 belongs to such an arc of @xmath1 .
let @xmath649 be the subarcs of @xmath1 lying within the outermost of these @xmath169-parallel annuli and let @xmath650 be the complementary subarcs of @xmath1 .
note that the @xmath651 lie on the same side of @xmath1280 .
then the isotopies described above corresponding to the outermost annuli deform @xmath1 to the union of the arcs @xmath650 along with arcs that lie in @xmath630 . as the complement of @xmath1 in @xmath3 is hyperbolic
, it must be that the arcs @xmath652 lie in @xmath1282 .
that is , @xmath652 are of the form @xmath1286 for @xmath1296 .
let @xmath24 be an arc @xmath1297 for @xmath1298 that is disjoint from @xmath1 . then @xmath1299 \cup [ ( m - n ' ) \cup \operatorname{nbhd}(\tau)]$ ] is a genus @xmath49 heegaard splitting of @xmath3 .
furthermore , each of @xmath652 is @xmath169-parallel in the handlebody @xmath1300 .
perturbing each arc of @xmath655 into @xmath1281 puts @xmath1 in bridge position with respect to this genus @xmath49 splitting .
noting that @xmath1301 and that @xmath126 is at most twice the minimal bridge number that @xmath1 has with respect to a genus @xmath49 splitting of @xmath3 , we get that @xmath1302 as desired .
[ lem : samelabels ] a gnarl or trunk of @xmath1142 shares at most a single vertex ( of @xmath127 ) with another gnarl or trunk of @xmath1142 .
furthermore , a vertex of @xmath127 belongs to at most two elements among the collection of gnarls and trunks of @xmath1142 .
let @xmath1132 be shared vertices of one gnarl or trunk with another gnarl or trunk .
the two must belong to different trigons @xmath1303 in @xmath1142 .
let @xmath1130 be the labels of the core scharlemann cycle of one of these trigons .
see figure [ figurexyedge ] .
then @xmath1304 @xmath1134 implying that @xmath1305 or @xmath1306 .
if the cores of @xmath942 and @xmath1190 are different , then all labels of @xmath128 are contained in @xmath1204 and the core scharlemann cycles are ` antipodal ' . by the minimality of @xmath1142 , @xmath1303
are the only elements of @xmath1128 . because their cores are antipodal , we may pare one of them down so that their corners are disjoint but their union contains all labels .
we may then have assumed that this was @xmath1128 and the resulting gnarls and trunks share no labels at all .
if a vertex of @xmath127 belongs to three different gnarls or trunks of @xmath1142 , then it must belong to @xmath1307 for three different trigons in @xmath1142 .
but this contradicts the minimality of @xmath1142 .
lemma [ lem : samelabels ] says that two simple gnarls of @xmath1142 whose interiors overlap must be nested .
thus we define a simple gnarl of @xmath1142 to have depth @xmath36 if it is innermost among simple gnarls of @xmath1142 .
a simple gnarl of @xmath1142 has depth @xmath1308 if it contains a gnarl of depth @xmath56 in its interior ( and no higher ) .
[ lem : depth<2 ] any two simple gnarls of @xmath1142 whose interiors overlap ( hence are nested ) must share a vertex .
a simple gnarl of @xmath1142 has at most two gnarls of @xmath1142 nested within it .
every simple gnarl of @xmath1142 has depth at most @xmath184 .
let @xmath1309 be simple gnarls on branches @xmath1310 of @xmath1311 in @xmath1128 .
assume @xmath1312 is nested within @xmath1188 . by property @xmath2 , @xmath942 and @xmath1313 must be different elements of @xmath1128 .
let @xmath1314 be the vertices of the gnarl @xmath1315 for @xmath1316 .
assume that @xmath1317 .
as the trunk curve of @xmath1312 is essential in @xmath136 , at least one of @xmath1318 must precede @xmath1312 on @xmath1319 , say @xmath1215 . by property 2(b ) , @xmath1320 . as property 3 implies that neither @xmath1321 nor @xmath1322 can be in @xmath1323
, it must be that @xmath1324 . by assumption , and the minimality of @xmath1142 , @xmath1325 .
but this contradicts the minimality of @xmath1142 .
this proves that two nested simple gnarls of @xmath1142 must share a vertex . as a vertex can belong to at most two gnarls in @xmath1142
, this implies that a simple gnarl has at most two nested within it .
let @xmath1326 be simple gnarls of @xmath1142 with @xmath1038 nested within @xmath1188 which is nested with @xmath1192 .
let @xmath1327 be the vertices of @xmath1315 , @xmath1328 .
by the paragraph above we may assume @xmath1329 .
by lemma [ lem : samelabels ] , @xmath1322 must lie strictly in the interior of @xmath1188 . again by the paragraph above @xmath1312 and @xmath1192 must share a vertex .
it can not be @xmath1322 as it lies in the interior of @xmath1312 .
thus @xmath1321 lies in all three gnarls , contradicting lemma [ lem : samelabels ] .
thus a simple gnarl of @xmath1142 has depth at most @xmath184 .
[ lem : boundwrapping ] there are at most @xmath1330 wrapping gnarls in @xmath1128 unless @xmath3 is a seifert fibered space over the 2-sphere with three exceptional fibers , one having order @xmath49 and another order @xmath2 . in the latter case ,
if @xmath1331 is a genus @xmath49 heegaard splitting of @xmath3 with respect to which @xmath1 has smallest bridge number ( among genus @xmath49 splittings ) , then there are at most @xmath1332 wrapping gnarls in @xmath1128 .
first assume that @xmath3 is not a seifert fibered space as described .
let @xmath453 be the collection of wrapping gnarls coming from the trigons in @xmath1128 , and assume for contradiction that @xmath1333 .
lemma [ lem : samelabels ] shows that @xmath453 satisfies properties @xmath1334 of definition [ defn3.1 ] . if two elements of @xmath453 intersect non - transversely , perturb them so they are disjoint so that property @xmath602 of definition [ defn3.1 ] is also satisfied
. then lemma [ lem : fkg ] guarantees that there are at least seven elements of @xmath453 that are isotopic on @xmath136 .
any two such elements much be disjoint .
let @xmath1335 be the outermost gnarls of these seven cobounding an annulus @xmath292 .
let @xmath1260 be a gnarl in @xmath292 between @xmath1335 .
then @xmath1260 belongs to a branch of @xmath1128 .
call the part of the branch between @xmath1260 and its trunk , the _ short branch _ of @xmath1260 .
if the short branch of @xmath1260 is disjoint from @xmath1335 , then the trunk of @xmath1260 is parallel to @xmath1260 in @xmath292 ; furthermore , as a branch always locally lies on the same side of a constituent gnarl ( property 1 ) , the entire branch will lie in @xmath292 between @xmath1260 and its trunk , contradicting lemma [ lem : parallelwrapping ] .
thus the short branch of @xmath1260 must intersect either @xmath1188 or @xmath1192 . as each vertex of a branch either belongs to a gnarl or trunk curve , we may associate to @xmath1260 , @xmath1336 , the first gnarl or trunk curve on the short branch that shares a vertex with either @xmath1188 or @xmath1192 ( orienting the short branch from @xmath1260 to its trunk ) .
if @xmath1337 are different gnarls between @xmath1335 , then @xmath1336 and @xmath1338 must be different .
otherwise , @xmath1337 belong to the same branch ( @xmath1339 can not be the trunk curve of two different branches by lemma [ lem : samelabels ] ) and are therefore disjoint before perturbation . then @xmath1340 , say , must occur on the short branch of @xmath1260 .
as above , property 1 shows that the entire branch of @xmath1260 lies between @xmath1337 , contradicting lemma [ lem : parallelwrapping ] .
similarly , one sees that @xmath1336 is neither @xmath1188 nor @xmath1192 .
lemma [ lem : samelabels ] says that there are at most four trunks or gnarls of @xmath1128 that share a vertex with @xmath1188 or @xmath1192 , other than @xmath1188 or @xmath1192 themselves .
this contradicts that there are five gnarls between @xmath1335 .
now assume @xmath3 is a seifert fibered space over the 2-sphere with three exceptional fibers including orders @xmath1140 and @xmath136 comes from a genus @xmath49 splitting for which @xmath1 has smallest bridge number .
let @xmath453 be the collection of wrapping gnarls coming from the trigons in @xmath1128 , and assume for contradiction that @xmath1341 .
now at least eight elements of @xmath453 are isotopic on @xmath136 .
the above shows that there must be two gnarls @xmath1337 of @xmath453 whose branches lie entirely in an annulus @xmath292 on @xmath136 .
in particular the trunk curves of these gnarls are parallel ( or equal ) in @xmath292 .
then by corollary [ cor : noparallelesc2 ] and remark [ sigma2remarks1](3 ) , there is a @xmath1145 containing both of these branches .
these branches then are either disjoint or share the same trunk curve .
let @xmath516 be the sub - annuli of @xmath292 that lie between these gnarls and their trunk curves .
then the interiors of these annuli are disjoint and each intersects the branch of the corresponding gnarl .
lemma [ lem : parallelwrapping ] shows that @xmath1342 vertices of @xmath127 lie in the union of the interiors of these annuli .
but there are at least six more vertices coming from the gnarls and their trunk curve(s ) , contradicting that @xmath127 has @xmath126 vertices .
in this section we prove our main theorem , which is the following . [ thm : main ] there is a linear function @xmath31 with the following property .
let @xmath11 be a hyperbolic knot in @xmath4 , @xmath9 where @xmath34 , and @xmath1 the core of the attached solid torus in @xmath3 .
suppose @xmath1 is in thin position with respect to a genus @xmath5 heegaard splitting of @xmath3 and let @xmath17 be a corresponding thick level surface . if @xmath17 is a strongly irreducible heegaard surface for @xmath3 then either * @xmath35 ; or * @xmath3 is toroidal ; or * @xmath3 is a seifert fibered space over the 2-sphere with exactly three exceptional fibers , at least one of which has order 2 or 3 .
furthermore , in cases ( 2 ) and ( 3 ) @xmath3 has a genus 2 heegaard splitting with respect to which @xmath1 has bridge number @xmath36 in case ( 2 ) and at most @xmath37 in case ( 3 ) .
theorem [ thm : main ] is proved by establishing the existence of a collection of annuli in @xmath1343 whose boundary components are , roughly speaking , elements of @xmath144 , and which capture , in their interiors , vertices of @xmath127 that do not correspond to labels in @xmath132 .
this is stated precisely in lemma [ lem : builda ] , which is proved in subsection 8.1 below . before giving the proof of the lemma
, we show how it leads to the proof of theorem [ thm : main ] .
let @xmath11 be a knot in @xmath4 and @xmath174 where @xmath206 .
let @xmath1 be the core of the attached solid torus in @xmath16 .
we are given a genus @xmath5 , strongly irreducible heegaard surface @xmath17 for @xmath3 , and we may assume @xmath1 can not be isotoped onto this surface , else we may take @xmath715 to hold .
in the special case that @xmath3 is a seifert fibered space as in @xmath48 , we further assume that @xmath17 comes from a genus @xmath49 heegaard splitting of @xmath3 with respect to which @xmath1 has the smallest bridge number among all genus @xmath49 splittings of @xmath3 . note that as @xmath3 is neither a lens space , @xmath4 , @xmath1344 , nor a connected sum , any genus @xmath49 splitting of @xmath3 is strongly irreducible .
then there are surfaces @xmath180 in the exterior of @xmath11 as described by lemma [ lem : fandq ] , where @xmath121 is a thick level surface in a thin presentation of @xmath1 with respect to the given splitting .
let @xmath179 be the corresponding graphs of intersection .
let latexmath:[$t=|\partial f|= is atoroidal so that @xmath602 of theorem [ thm : main ] does not hold .
we then show that either conclusion @xmath715 ( for @xmath38 ) or @xmath48 ( with the bridge number bound on minimal bridge number , genus @xmath49 splitting ) of theorem [ thm : main ] holds .
we may assume @xmath1346 , so lemma [ lem:2 ] guarantees that @xmath128 has a great web @xmath129 .
let @xmath132 be the collection of labels of @xmath129 given in definition [ def : call ] , and @xmath143 the complement of @xmath132 among all labels of @xmath128 .
let @xmath137 be the minimal collection of extended scharlemann cycles of @xmath129 given by definition [ def : sigma ] .
let @xmath793 and @xmath1128 be as in definitions [ def : cornersi&ii ] and [ def : sigmaiii ] .
then for every @xmath217 , either @xmath1347 or there is a @xmath1348 such that @xmath801 .
let @xmath140 be the set of labels of @xmath128 that correspond to vertices of simple gnarls of @xmath1128 that are not in @xmath138 ( section [ sec : typeiii ] ) .
a gnarl hereafter will always be a gnarl of @xmath1128 .
every element of @xmath132 belongs to either @xmath138 , @xmath139 ( definition [ def : l2 ] ) , or corresponds to a vertex of a gnarl ( definition [ def : simpleandwrapping ] ) .
a gnarl may be either wrapping or simple . by theorem [ thm : type12boundn ] and
lemma [ lem : boundwrapping ] , @xmath1349 let @xmath146 be the collection of annuli given by lemma [ lem : builda ] .
we can write @xmath1350 , where @xmath1351 are the elements of @xmath140 corresponding to vertices that lie outside @xmath1352 , and @xmath1353 are those corresponding to vertices in the interiors of the annuli in @xmath146 . by part @xmath995 of lemma [ lem : builda ]
, we have @xmath1354 .
using lemma [ lem : nonemptygnarl ] to remove from consideration those gnarls with no vertices in their interior , corollary [ 3ingnarl ] and lemma [ lem : depth<2 ] imply that there are at least @xmath1355 vertices of @xmath127 outside @xmath1356 that lie in the interior of simple gnarls but are not vertices of gnarls .
such vertices can not belong to @xmath138 , so they are either in @xmath139 or in @xmath143 .
first assume @xmath146 is non - empty
. then @xmath1357 where @xmath1358 is the function defined in lemma [ lem : builda ] , and where the last line uses @xmath1359 .
thus by theorem [ thm : type12boundn ] @xmath1360 by proposition [ prop : webcount ] , @xmath1361 .
therefore @xmath1362 thus @xmath1363 where @xmath1364 proving theorem [ thm : main ] when @xmath146 is non - empty .
now assume @xmath146 is empty
. then lemma [ lem : builda](1 ) implies @xmath1365 .
furthermore , @xmath1366 , and @xmath1367 along with theorem [ thm : type12boundn ] yields @xmath1368 .
then @xmath1359 gives @xmath1369 .
using proposition [ prop : webcount ] , we get @xmath1370 this gives @xmath1371 where @xmath7 is as above .
if @xmath135 is a @xmath135-curve , a _ curve _ in @xmath135 is a circle @xmath23 obtained by removing from @xmath135 the interior of an edge . if @xmath241 is an extended scharlemann cycle of length @xmath2 , @xmath1372 will denote the set of curves in @xmath135-curves belonging to @xmath401 . if @xmath241 is an extended scharlemann cycle of length @xmath49 or @xmath2 , @xmath1373 will denote @xmath1374 ( where @xmath1375 if @xmath241 has length @xmath49 ) .
[ lem : builda ] there is a ( possibly empty ) collection @xmath146 of annuli in @xmath136 such that 1 .
@xmath1376 where @xmath1377 ; 2 . for any @xmath1378
there is a @xmath1379 such that each component of @xmath973 belongs to @xmath1373 ; 3 .
the interiors of any two distinct annuli in @xmath146 are disjoint ; 4 .
no vertex in the interior of an annulus in @xmath146 belongs to @xmath138 , and any vertex in the interior of an annulus in @xmath146 that belongs to @xmath132 is either a vertex of a simple gnarl or belongs to @xmath139 ; 5 .
if @xmath1378 and @xmath1380 is the number of vertices in the interior of @xmath296 that belong to simple gnarls then there are at least @xmath1381 vertices in @xmath296 that belong to @xmath1382 .
if @xmath146 is empty ( @xmath1383 ) , then ( 1 ) says that @xmath1384 . by lemma [ lem : parallelthetas ]
there are at least @xmath1385 @xmath135curves of @xmath1386 that lie in essential annuli . since @xmath1387 and each element of @xmath1388 lies in an essential annulus
, at least @xmath1389 of the elements of @xmath144 lie in essential annuli .
let @xmath453 be this subcollection of @xmath144 .
call two elements of @xmath453 _ isotopic _ if the cores of the corresponding annuli are isotopic in @xmath136 .
then any two elements of @xmath453 which are isotopic are either disjoint or intersect non - transversely in a single vertex .
furthermore , the elements of @xmath453 form at most @xmath273 isotopy classes ( using ` core ' curves from the @xmath135-curves and perturbing non - transverse intersections , lemma [ lem : atmostoneintersection ] shows that a subcollection of non - isotopic elements of @xmath453 have property @xmath319 apply lemma [ lem : fkg ] ) . thus all but at most @xmath1390 of these curves lie in isotopy classes with at least four elements of @xmath453 .
let @xmath1391 be the distinct isotopy classes of @xmath453 , each of which contains at least four elements of @xmath453 .
a _ pinched annulus _
@xmath725 is a disk @xmath203 with two points on its boundary identified to a single point @xmath1392 .
the _ interior _ of @xmath292 is the interior of @xmath203 .
note that under the identification @xmath1393 becomes the union of the two simple closed curves @xmath23 and @xmath1394 that intersect non - transversely at @xmath1392 .
we write @xmath1395 .
consider @xmath1396 , say , where @xmath1397 , and the @xmath1398 s are ordered sequentially as they lie on @xmath136 .
then we have 1 .
@xmath1398 and @xmath1041 are disjoint unless @xmath1399 and @xmath1398 and @xmath1041 share a single vertex ( lemma [ lem : atmostoneintersection ] ) ; 2 .
for @xmath1400 there exists @xmath1401 such that 1 .
@xmath1402 is an annulus or pinched annulus according as @xmath1403 is empty or non - empty ; 2 .
@xmath1404 where @xmath1405 is a curve in @xmath1406 ; 3 .
@xmath1402 contains @xmath1407 ; 4 .
there are no vertices of any @xmath1041 in the interior of @xmath1402 .
we have @xmath1408 .
note that if @xmath1402 is pinched then @xmath1409 .
[ clm : intar ] if @xmath1410 then the number of vertices of @xmath127 in the interior of @xmath1402 is at least @xmath1411 , and at least @xmath1412 if @xmath1402 is pinched .
we may assume @xmath1413 , hence lemmas [ lem : parallelesc ] and [ notwo ] imply that @xmath3 is a seifert fibered space over @xmath349 with three exceptional fibers at least one of which has order @xmath49 or @xmath2 .
this puts us in the context of theorem [ thm : main](3 ) .
thus @xmath136 is a genus @xmath49 splitting surface for @xmath3 with respect to which @xmath1 has the smallest bridge number .
again by lemmas [ lem : parallelesc ] and [ notwo ] , @xmath1 has bridge number at most @xmath620 , where @xmath1414 .
thus @xmath1415 ( twice the bridge number is an upper bound for the number of intersections with a thick surface ) .
let @xmath1416 be the number of vertices in the interior of @xmath1402 .
then @xmath1417 or @xmath1418 according as @xmath1402 is an annulus or a pinched annulus , giving @xmath1419 or @xmath1412 , respectively .
let @xmath1420 be the collection @xmath1421 coming from @xmath895 .
[ clm : disjointannuli ] for any @xmath894 , any element of @xmath895 is disjoint from some element of @xmath896 .
thus any element of @xmath895 is either disjoint from a given element of @xmath896 or intersects it non - transversely in a single vertex .
no vertex in the interior of an annulus of @xmath1420 lies in the interior of an annulus of @xmath1422 , @xmath894 . by claim [ clm : intar ]
the number of vertices in a pinched annulus is at least @xmath1423 .
it follows that there are at most two pinched annuli in @xmath1420 , and if there are two , then they must share two vertices and hence must be adjacent . since @xmath1424 , there are at least three disjoint elements of @xmath895 .
the middle of these is then disjoint from any element of @xmath896 as proposed , and any element of @xmath896 can intersect an element of @xmath895 only non - transversely .
let @xmath1425 . by the above
, the boundary components of @xmath312 can be perturbed to be disjoint . if their interiors intersected , then some component of @xmath498 would be isotopic to one of @xmath973 , a contradiction .
thus @xmath312 share no interior vertices .
let @xmath1426 .
it is easy to see from claim [ clm : intar ] that there are at most two elements @xmath1427 such that @xmath1428 , where @xmath1429 .
moreover , if there are two then they are adjacent ( and hence belong to the same @xmath1420 ) .
let @xmath1430 be @xmath1420 with any such elements removed .
thus for any annulus @xmath1431 , there is a @xmath1379 such that @xmath1432 . in particular ,
there are no pinched annuli in @xmath1430 . if @xmath1433 , their interiors will be disjoint unless they share a @xmath135-curve @xmath135 , in which case two of the edges of @xmath135 will cobound a disk , @xmath203 , of parallelism in @xmath136 with @xmath1434 .
since the cores of the annuli in @xmath1430 are all isotopic in @xmath136 , assigning an orientation to this isotopy class allows us to talk about one element of @xmath1430 being to the _ right _ or _ left _ of another .
if @xmath1435 share a disk of parallelism @xmath203 , with @xmath1436 on the left of @xmath1437 , we define @xmath1438 .
doing this for all pairs in @xmath1430 that share a disk of parallelism , we get a collection @xmath1430 of annuli in @xmath136 whose interiors are disjoint .
let @xmath1439 .
then @xmath1440 satisfies conditions @xmath602 and @xmath48 of the lemma . also , recall from the first paragraph of the proof that @xmath1441 and @xmath1442 .
hence @xmath1443 [ clm : intofa ] there are at most @xmath1444 annuli in @xmath1440 with no vertices of @xmath127 in their interior .
the elements of @xmath1430 appear sequentially along @xmath136 .
since @xmath1445 is obtained from @xmath1446 by removing either @xmath1447 or @xmath49 adjacent elements from some @xmath1420 , the annuli in @xmath1440 appear in at most @xmath1448 sequential groups , with , say @xmath1449 annuli in the @xmath39th group , @xmath1450 .
we show that no two consecutive annuli in @xmath1430 have empty interiors .
so assume @xmath312 are consecutive annuli in @xmath1430 , neither of which contains vertices in its interior .
by @xmath1451 and the construction of @xmath1430 , @xmath498 and @xmath973 belong to @xmath1373 for some @xmath1379 . but
this contradicts lemma [ lem : triplets ] or lemma [ lem : thetatriplets ] ( or theorem [ thm : main]@xmath48 holds with the bridge number bound ) .
note that if @xmath241 were of length @xmath2 , letting @xmath1452 be the @xmath135-curve which meets both @xmath292 and @xmath296 , then , by the way we defined the elements of @xmath1440 , the disk of parallelism of @xmath135 is contained in either @xmath292 or @xmath296 .
so lemma [ lem : thetatriplets ] applies .
it follows that the number of annuli in @xmath1440 with no vertices in their interior is at most @xmath1453 as claimed .
since any label in @xmath138 corresponds to a vertex in an element of @xmath144 , it is clear from the definition of the original collection @xmath1454 that no annulus in @xmath1440 contains vertices in its interior corresponding to elements of @xmath138 .
finally , we define @xmath146 to be the collection of annuli obtained by discarding from @xmath1440 any annulus whose interior either * has no vertices of @xmath127 ; or * contains the vertex of a wrapping gnarl ; or * contains the vertex of a simple gnarl with no vertices in its interior .
then @xmath146 satisfies condition @xmath1455 and @xmath995 of lemma [ lem : builda ] . by claim [ clm : intofa ] ,
lemma [ lem : boundwrapping ] , and lemma [ lem : nonemptygnarl ] @xmath1456 hence , using inequality ( [ eq : atilde ] ) above , we have @xmath1457 by claim [ clm : disjointannuli ] and lemma [ lem : fkg ] , @xmath1458 .
setting @xmath1459 , we get @xmath1460 where @xmath1377 .
thus @xmath146 satisfies condition @xmath715 .
it remains to show that @xmath146 satisfies condition @xmath1461 .
[ clm : atleast2 ] each element of @xmath146 contains at least two vertices in its interior .
assume the annulus @xmath1378 contains a single vertex in its interior . by ( 2 ) ,
the components of @xmath973 belong to @xmath1373 for the same element , @xmath241 , of @xmath137 .
first assume that @xmath241 is an extended scharlemann cycle of length @xmath49 .
then there is an annulus @xmath1462 that runs between the components of @xmath973 .
@xmath292 intersects @xmath296 only along their mutual boundary , forming a 2-torus , @xmath531 .
furthermore , @xmath1 may be perturbed off @xmath292 so that it intersects @xmath531 either once or three times .
but then @xmath531 is a non - separating surface in @xmath3 , which is impossible .
when @xmath241 has length @xmath2 , we arrive at an analogous contradiction .
@xmath296 contains a @xmath135-curve @xmath1463 . as in the proof of corollary [ 3ingnarl ] ,
if the unique vertex of @xmath127 in the interior of @xmath296 lay between the two parallel edges of @xmath135 in @xmath296 , then @xmath127 would have a 1-sided face ( the number of edges interior to the parallelism is less than the valence of the interior vertex ) .
hence this vertex can not lie in this disk of parallelism .
the proof of lemma [ lem : annulusintheta ] shows how to construct an annulus @xmath292 within @xmath399 that runs between components of @xmath973 such that @xmath1 can be perturbed off of @xmath292 to intersect @xmath296 transversely in either one or three points .
but then @xmath852 is a non - separating torus in @xmath3 , a contradiction .
let @xmath1378 , let @xmath1380 be the number of vertices in the interior of @xmath296 that belong to simple gnarls , and @xmath1464 the number that correspond to labels in @xmath1382 .
we must show that @xmath1465 .
note that if a vertex in the interior of @xmath296 belongs to @xmath132 , it must in fact belong to @xmath139 by condition @xmath995 of lemma [ lem : builda ] .
if @xmath1466 then @xmath1467 by claim [ clm : atleast2 ] . if @xmath1468 then the corresponding vertices belong to simple gnarls where each such gnarl contains vertices in its interior . by lemma [ lem : depth<2 ] , simple gnarls are of depth @xmath36 or @xmath184 . by corollary [ 3ingnarl ] , each depth @xmath36 gnarl contains in its interior at least three vertices , which must belong to @xmath1382 and hence contribute @xmath2 to @xmath1464 . by lemma [ lem : depth<2 ]
, a depth @xmath184 gnarl must share a vertex with an internal depth @xmath36 gnarl . by lemma [ lem : samelabels ]
, a vertex of @xmath127 belongs to at most two different gnarls .
let @xmath1469 be the vertices that belong to gnarls lying in @xmath296 ( possibly with vertices on @xmath973 ) .
then @xmath1470 . let @xmath56 be number of depth @xmath36 gnarls contained in @xmath296 , @xmath525 be the number of vertices in depth @xmath184 gnarls in @xmath296 that are not in depth @xmath36 gnarls , and @xmath39 the number of vertices that are contained in two depth @xmath36 gnarls in @xmath296 . from the observations in the preceding paragraph ,
@xmath1471 . furthermore , there are at least @xmath1472 vertices in @xmath1382 in the interior of @xmath296
. thus @xmath1473 $ ] . noting that @xmath1474 , the bracketed quantity is non - negative . therefore , @xmath1475 it follows that @xmath1465 .
[ thm : appendix ] there is a linear function @xmath1476 with the following property .
let @xmath11 be a hyperbolic knot in @xmath4 , @xmath9 where @xmath34 , and @xmath1 the core of the attached solid torus in @xmath3 .
let @xmath17 be an orientable , incompressible surface in @xmath3 of genus @xmath5 .
then @xmath1 can be isotoped to intersect @xmath17 at most @xmath75 times .
compare the above with theorem 4 of @xcite which says the following .
let @xmath11 be an hyperbolic knot in @xmath4 that does not contain a genus @xmath5 , closed incompressible surface in its exterior and such that @xmath174 contains an incompressible surface of genus @xmath5 .
let @xmath1 be the core of the attached solid torus in @xmath3 .
if @xmath1477 , then there is an incompressible surface of genus @xmath5 in @xmath3 which @xmath1 intersects at most @xmath1478 times .
the proof is just that of theorem [ thm : main ] given for strongly irreducible heegaard surfaces , except basically that the proof of claim [ clm : stronglyirred ] must be altered to replace the assumption of strong irreducibility with incompressibility .
in fact the arguments for an incompressible surface tend to be easier than in the context of an heegaard surface , but as those arguments have been already been made , we will make use of them . as @xmath3 is a non - integral surgery on a knot in @xmath4 , it is irreducible ( @xcite ) .
we may assume that @xmath3 is atoroidal .
otherwise by @xcite @xmath11 is an eudave - munoz knot and , consequently , @xmath3 is the union of two seifert fibered spaces over the disk , each with two exceptional fibers , along a torus @xmath17 .
the seifert fibers of the two halves intersect once along @xmath17 .
hence @xmath17 is the unique orientable incompressible surface .
( any other connected incompressible surface would have to be the union of horizontal surfaces in the two seifert fibered spaces .
consequently such a surface would be non - separating , contradicting that @xmath3 is a rational homology sphere . )
again @xcite shows that @xmath1 can be isotoped to intersect @xmath17 twice . note that this also shows that @xmath3 can not be a seifert fibered space as any such which is atoroidal contains no incompressible surface . finally , recall that @xmath3 can not contain a klein bottle by theorem 1.3 of @xcite ( see also @xcite ) .
so let @xmath136 be an incompressible surface in @xmath3 of genus at least @xmath49 and isotop @xmath1 to intersect @xmath136 minimally .
let @xmath42 be the punctured surface , @xmath1479 , where @xmath20 is the exterior of @xmath1 in @xmath3 ( which is the exterior of @xmath11 in @xmath4 ) .
the thin position argument given in lemma 4.4 of @xcite ( in place of , and simpler than , that of @xcite ) shows that there is a level surface @xmath120 in a thin presentation of @xmath11 in @xmath4 such that the corresponding punctured surface @xmath1480 intersects @xmath42 in arcs which are essential in both @xmath119 and @xmath42 .
that is , no arcs of @xmath172 are boundary parallel in either @xmath42 or @xmath119 .
the combinatorial arguments of section [ sec : graphcombo ] go through in this context .
in particular , proposition [ prop : webcount ] still holds for @xmath132 as defined in that section . for each label @xmath217
, there is a bigon or trigon in @xmath131 , and the classification of such as either an extended scharlemann cycle or as a trigon of type i , ii , or iii is the same .
one checks that all of the results of section [ sec : thinning ] hold ; however , with the helpful addition now that @xmath3 can not be a seifert fibered space . note that an isotopy which gave a thinning in this section now reduces the intersection number of @xmath1 and @xmath136 .
indeed this is overkill .
for example , neither conclusion of lemma [ lem : sfs ] can hold in this context : conclusion @xmath715 allows us to reduce the intersection of @xmath1 with @xmath136 , and conclusion @xmath602 says that @xmath3 is a seifert fibered space .
there is one place in section [ sec : thinning ] where the fact that @xmath136 is a strongly irreducible heegaard surface is used .
that is in the proof of claim [ clm : stronglyirred ] .
strong irreducibility is used there to show that the seifert fibered space @xmath429 does not lie in a 3-ball . in the current context , this does not occur as it would imply that a curve that is homotopically essential in @xmath136 lies in this 3-ball , and consequently is homotopically trivial in @xmath3 thereby violating the incompressibity of @xmath136 .
to3em , _ non - hyperbolic manifolds obtained by dehn surgery on hyperbolic knots _ , geometric topology ( athens , ga , 1993 ) , ams / ip stud
, vol . 2 , amer .
soc . , providence , ri , 1997 , pp .
( 98i:57007 ) cameron mca .
gordon , _ combinatorial methods in dehn surgery _
, lectures at knots 96 ( tokyo ) , ser .
knots everything , vol .
15 , world sci .
publishing , river edge , nj , 1997 , pp .
mr mr1474525 ( 98m:57020 ) | we show there exists a linear function @xmath0 with the following property .
let @xmath1 be a hyperbolic knot in a hyperbolic @xmath2manifold @xmath3 admitting a non - longitudinal @xmath4 surgery . if @xmath1 is put into thin position with respect to a strongly irreducible , genus @xmath5 heegaard splitting of @xmath3 then @xmath1 intersects a thick level at most @xmath6 times .
typically , this shows that the bridge number of @xmath1 with respect to this heegaard splitting is at most @xmath7 , and the tunnel number of @xmath1 is at most @xmath8 . |
disks are an integral part of star formation ; they are the birthplace of planets .
how they form is a long - standing , unresolved problem .
a major difficulty is that their formation is greatly affected by magnetic braking , which has been hard to quantify until recently .
there is now increasing theoretical evidence that magnetic braking may suppress the formation of rotationally supported disks ( rsds hereafter ) in dense cores magnetized to a realistic level , with dimensionless mass - to - flux ratios @xmath0 of a few to several @xcite .
@xcite first demonstrated through 2d ( axisymmetric ) simulations that rsds are suppressed by a moderately strong magnetic field in the ideal mhd limit .
@xcite showed analytically that the disk suppression is due to the formation of a split magnetic monopole , which is an unavoidable consequence of flux freezing .
the efficient disk braking was confirmed numerically by @xcite and @xcite ( see , however , @xcite and @xcite for a different view , and @xcite for a more detailed discussion ) .
the ideal mhd approximation must break down in order for rsd to form .
two non - ideal mhd effects have already been explored in the context of disk formation : ambipolar diffusion and ohmic dissipation .
@xcite found that ambipolar diffusion tends to make disk braking more efficient , because it enables the magnetic flux that would have gone into the central protostellar object ( and form a split monopole ) in the ideal mhd limit to pile up at small radii outside of the object instead , making the region strongly magnetized ( @xcite , @xcite , @xcite ) .
@xcite showed that the enhanced braking is strong enough to suppress the formation of rsds for realistic core conditions .
the effect of ohmic dissipation was examined by several groups , starting with @xcite .
they suggested that enhanced resistivity ( well above the classical value ) is needed for ohmic dissipation to weaken the magnetic braking enough to form a relatively large rsd of tens of aus or more .
the suggestion was confirmed by @xcite ( @xcite ; kls10 hereafter ) , although @xcite were apparently able to form rsds using the classical resistivity computed by @xcite ( see also @xcite ) . in any case , a third non - ideal mhd effect , the hall effect , has not been explored in the context of disk formation ( see , however , the independent work of @xcite in her unpublished phd thesis ) ; it is the focus of this paper .
our goal is to determine whether the hall effect by itself can enable an rsd to form in the presence of a relatively strong magnetic field and , if yes , to estimate the magnitude of hall coefficient needed for rsd formation .
we find that the hall effect can actively spin up the inner part of the protostellar collapsing flow , potentially to keplerian speed , unlike the other two non - ideal mhd effects .
the combined effect of all these three non - ideal mhd terms on disk formation will be explored in another investigation ( @xcite ; lks11 hereafter ) .
we adopt the same problem setup as in kls10 , where we consider the collapse onto a star of @xmath1 of a rotating envelope that is uniformly magnetized initially . in the absence of any non - ideal mhd effect
, the magnetic field would prevent an rsd from forming through magnetic braking .
the hall effect changes the evolution of the magnetic field ( and thus the braking efficiency ) through the induction equation : @xmath2\right\ } \label{induction}\ ] ] where @xmath3 is a coefficient of the hall effect which , for simplicity , we will assume to be spatially constant .
the mhd equations are solved using zeustw , a 3d non - ideal mhd code based on zeus3d @xcite .
we treat the hall term in the induction equation using an explicit method based on @xcite and @xcite , which includes subcycling to speed up the computation . as in kls10
, we adopt a spherical polar coordinate system ( @xmath4 , @xmath5 , @xmath6 ) , and fill the computation domain with an isothermal gas of sound speed @xmath7 between @xmath8 and @xmath9 .
a uniform density @xmath10 is assumed , so that the total envelope mass is @xmath11 .
for the initial rotation , we adopt the following prescription : @xmath12 where @xmath13 is the cylindrical radius , and @xmath14 and @xmath15 are chosen . with this setup ,
an rsd of @xmath16 in radius is produced at a representative time of @xmath17 in the absence of any magnetic field ( see fig . 1 of kls10 ) .
it is completely destroyed by a moderately strong initial magnetic field of @xmath18 in the ideal mhd limit ( see fig . 2 of kls10 ) . whether the rsd can be restored by the hall effect
is the question that we seek to address .
we consider first a model where the initial magnetic field is parallel to the rotation axis ( model para in table [ table : first ] ) and the coefficient @xmath19 in the lorentz - heaviside units convenient to use with the zeus family of codes , corresponding to @xmath20 in gaussian cgs units ( @xmath21 ) .
an rsd is able to form in this case , as illustrated fig .
[ hall_19 ] , which plots a color map of density distribution , velocity field and magnetic field lines in the meridian plane ( left panel ) , and the infall and rotation speeds on the equator ( right panel ) , at a representative time @xmath22 .
the rsd shows up clearly on the color map as the flattened high density region of about @xmath23 in radius .
it has an infall speed that is close to zero and a rotation speed that is keplerian .
the hall effect has clearly enabled an rsd to form , although it is not by weakening the magnetic braking that would have suppressed the rsd formation in the ideal mhd limit , unlike the case of ohmic dissipation ( kls10 ) .
the rsd is formed because the inner part of the collapsing flow is actively spun up by the hall effect . a strong support for hall spin - up
as the cause for the rsd formation comes from a second simulation where we flip the sign of the initial magnetic field direction ( model anti ) , so that the field is anti - parallel , rather than parallel , to the rotation axis . in this case ,
an rsd disk is again formed , but it rotates in a direction opposite to the initial rotation , as shown in fig .
[ hall_22 ] .
the reversal of the disk rotation direction means that the hall effect does not merely weaken the magnetic braking so that enough of the original angular momentum is retained for rsd formation .
lllll para & @xmath24 & 35.4 & yes & yes + anti & @xmath24 & - 35.4 & yes & yes + norot & @xmath24 & 35.4 & no & yes + lob & @xmath24 & 10.6 & no & yes + loq & @xmath25 & 35.4 & no & yes + loq1 & @xmath26 & 35.4 & no & yes + loq2 & @xmath27 & 35.4 & no & no + the reason that the disk rotation direction is reversed when the magnetic field direction is flipped is the following ( see @xcite for a similar discussion in the context of disk - wind launching ) . from the induction equation ( [ induction ] ) , it is easy to see that the hall effect forces the field lines to move with a velocity @xmath28 where @xmath29 is the current density .
the pinching of the poloidal magnetic field near the equatorial region produces a strong toroidal current , @xmath30 , which changes sign as the poloidal field direction is flipped .
the toroidal current of different signs causes the field lines to wind up toroidally in opposite directions , giving rise to magnetic torques of opposite sense . in the case where the initial field line is parallel to the rotation axis , @xmath30 is positive , giving rise to a toroidal field - line speed @xmath31 that is negative ( assuming that the hall coefficient is positive , which is not necessarily true , see , e.g. , @xcite and lks11 ; @xmath31 changes sign if @xmath3 is negative ) . since the toroidal current is the strongest on the equator , the field line is twisted backward most on the equator and less so above and below the equatorial plane .
the differential twist forces the field line to bend in the positive azimuthal direction , creating a torque that spins up the material near the equatorial plane in the same direction as the initial rotation .
this explains why the rsd in the field - rotation aligned case rotates in the same sense as the initial rotation . when the initial field direction is flipped , the toroidal current @xmath30 ( and thus @xmath31 ) changes sign , forcing the field lines to bend in the negative azimuthal direction .
the resulting negative torque not only overcame the initial positive rotation but also produced a counter - rotating rsd in the anti - aligned case .
the hall spin - up is even more unambiguous in model norot where the envelope is initially non - rotating ; any subsequent rotation must be due to the hall effect .
the right panel of fig .
[ hall_24 ] shows that an rsd is again formed in this case , with the rotation speed approaching the keplerian speed and the infall speed dropping close to zero inside a radius @xmath32 .
since the envelope is initially non - rotating , the spin of the rsd must be balanced by material rotating in the opposite direction .
the left panel of fig .
[ hall_24 ] demonstrates that this is indeed the case .
it shows that , outside the equatorial plane , materials of positive and negative angular momentum occupy alternating shells . inside a radius of @xmath33 ,
the angular momentum is positive , dominated by the ( small ) rsd and highly flattened pseudodisk outside rsd ( see the isodensity contours ) . at larger distances ,
the angular momentum becomes negative , dominated by the outermost counter - rotating shell shown in the fig .
[ hall_24 ] .
the total positive angular momentum ( @xmath34 ) does not exactly balance out the total negative angular momentum ( @xmath35 ) , however , indicating some angular momentum is transported out of the simulation box , presumably through torsional alfvn waves .
note that the hall effect modifies the envelope rotation well beyond the @xmath36-scale rsd .
we have verified that when the initial magnetic field direction is flipped , the direction of the hall - induced rotation is reversed .
besides torquing up the envelope in the toroidal direction , the hall effect also enables the magnetic field lines to diffuse out radially . in the absence of any radial diffusion
, the field lines would be pulled by the accreting material into a split magnetic monopole , which is not present in any of the simulations discussed so far .
indeed , at small radii where the hall spin - up is most efficient , the inward advection of poloidal magnetic field is almost exactly balanced by the outward magnetic diffusion enabled by the hall effect .
this balance is illustrated in the right panel of fig .
[ hall_24 ] , which shows that , inside a radius @xmath37 , the radial component of the hall - induced field line drift velocity on the equator has almost the same magnitude as the infall speed but with an opposite sign .
the radial magnetic diffusion allows material to fall inward without dragging field lines along with it .
the situation is similar in the azimuthal direction , where the hall - induced field line drift speed and the fluid rotation speed are nearly identical in magnitude but opposite in sign .
the azimuthal hall - induced drift prevents the field lines from being wound up continuously by rotation .
the required toroidal current is provided by the bending of the field lines inside the disk ( see left panel of fig . 1 ; contrast with the unbent field lines in fig
. 7 of kls10 )
. an interesting feature of the envelope collapse is that the infall speed of the equatorial material remains well below the free - fall speed .
the equatorial collapse is slowed down three times ( see right panel of fig .
[ hall_24 ] ) , for different reasons .
the first slowdown near @xmath38 corresponds to the edge of the magnetic bubble inflated by field twisting ( outside the region shown in the right panel of fig .
[ hall_24 ] ) , where a magnetic barrier forces the collapsing material over a large solid angle into a narrow equatorial channel ( see fig . 2 of @xcite and associated discussion ) .
the second slowdown occurs near @xmath39 ( or about @xmath40 ) , inside which the accreted magnetic flux is left behind due to hall diffusion .
it is similar in origin to the ambipolar diffusion - induced accretion shock first discussed in @xcite ( @xcite ; a similar behavior is also present in the ohmic dissipation - only case , lks11 ) .
the increase in the poloidal magnetic field strength enabled by hall diffusion in the radial direction makes it easier to spin up the equatorial material through the hall - induced magnetic torque .
the third slowdown near @xmath32 is obviously due to centrifugal effect , since an rsd is formed interior to it . to check how robust the above results from model norot are , we did a number of variants of the model .
first , we reduced the initial field strength to @xmath41 of its original value , from @xmath42 to @xmath43 ( model lob ) .
the result is shown in the left panel of fig .
[ hall_compare ] ( see dashed lines ) .
an rsd is again formed at the representative time @xmath22 , despite the reduced initial field strength .
it is somewhat smaller than that in the original case , which is to be expected since the hall spin - up depends on current density , which increases with field strength .
another difference is that the infall speed outside the rsd is higher in the weaker field case , because the magnetic forces that oppose gravity are weaker .
second , we reduced the coefficient @xmath3 by a factor of 10 , from @xmath44 to @xmath45 in gaussian cgs units ( model loq ) .
an rsd is still formed at @xmath22 . with a radius of @xmath46 or @xmath47 ,
it is much smaller than the rsd in the original model norot .
if we reduce the value of @xmath3 by another factor of 3 ( to @xmath48 , model loq1 ) , an rsd still forms , but it extends barely outside the inner boundary of our computation domain , which has a radius of @xmath49 . when the value of q is reduced further to @xmath50 ( model loq2 ) , the rsd disappears .
we therefore take @xmath51 as a rough estimate for the value of @xmath3 needed for the hall - enabled formation of a relatively large rsd of tens of aus or more in size .
we have verified that the rsd formed in model norot is little affected when the classical ohmic resistivity used in kls10 ( based on @xcite ) is included .
an enhanced resistivity can , however , reduce the size of the rsd significantly .
experimentation shows that it suppresses the hall - enabled rsd formation in model norot altogether when the resistivity is on the order of @xmath52 or larger .
the negative effect of the resistivity on the hall spin - up is not surprising , since it reduces the electric current density in general , and the toroidal current density in particular ; the latter is the driver of the spin - up through the hall - induced magnetic torque .
how does the critical value of @xmath53 estimated for rsd formation compare with the microscopic values of expected in realistic dense cores ? in a separate study ( lks11 ) , we have computed the values of @xmath3 as a function of hydrogen number density @xmath54 ) , based on a simple prescription for the magnetic field strength as a function of density @xcite and a simplified chemical network for charge densities @xcite , including two extreme grain size distributions : ( 1 ) the standard mrn power - law distribution ( @xcite ; appropriate for diffuse interstellar clouds ) , and ( 2 ) grains of single , large size of @xmath55 .
these two cases should bracket the situation in dense molecular cloud cores , where some grain growth is expected . as can be seen from the right panel of fig .
[ hall_compare ] , although the microscopic values of @xmath3 can be larger than the critical value @xmath56 at relatively low densities , they are smaller than @xmath56 by about an order of magnitude at densities of order @xmath57 ( or higher ) that are crucial for rsd formation ( see the left panel of fig .
[ hall_19 ] ) .
we have studied the collapse of a non - self - gravitating rotating , magnetized envelope onto a central stellar object of fixed mass to illustrate the influence of the hall effect on disk formation . in this idealized problem ,
the formation of a rotationally supported disk ( rsd ) is completely suppressed by magnetic braking in the ideal mhd limit . including a hall term with a coefficient @xmath58 in the induction equation enables an rsd of radius @xmath59 to form .
the rsd is formed not because the hall effect has reduced the efficiency of the magnetic braking .
rather , it is produced because the hall effect actively spins up the inner part of the equatorial material to keplerian speed .
the spin - up comes about because the collapsing envelope drags the magnetic field into a highly pinched configuration near the equator , producing a large toroidal electric current density , which forces the field lines to rotate differentially due to the hall effect .
the resulting twist of field lines yields the magnetic torque that spins up the equatorial material , even in the absence of any initial rotation .
the spin - up is most effective in the inner part of the accretion flow where radial hall diffusion enables the magnetic flux that would have been dragged into the central object by the accreted material in the ideal mhd limit to stay behind ; the resulting pileup of magnetic flux at small radii is similar to the cases with either ambipolar diffusion or ohmic dissipation . when the field direction is flipped , the hall - induced magnetic torque changes direction as well .
an implication is that the direction of the angular momentum of the material close to a protostar , including rsd , may depend more on the magnetic field orientation than on the initial rotation of the dense core that the star is formed out of , at least in principle . in practice , it is uncertain whether the hall - induced magnetic torque can produce a sizable rsd or not .
first , the value of the coefficient @xmath3 required to produce an rsd of tens of aus in size in our idealized model appears to be larger than the microscopic values expected in dense cores , by roughly an order of magnitude . whether the coefficient can be enhanced somehow , perhaps through anomalous processes
, is unclear .
more importantly , the hall effect is only one of the non - ideal mhd effects that are present in the lightly ionized , magnetized , dense cores .
it can be dominated by ambipolar diffusion at low densities and by ohmic dissipation at high densities .
it will be interesting to explore the interplay of all these three non - ideal mhd effects and how they affect the collapse of dense cores and the formation of protostellar disks .
nevertheless , we have demonstrated that the hall effect is unique among the non - ideal effects in its ability to actively spin up a magnetized collapsing flow , potentially to keplerian speed , providing a new mechanism for disk formation .
we acknowledge support by the theoretical institute for advanced research in astrophysics ( tiara ) , by the national science council of taiwan through grant nsc97 - 2112-m-001 - 018-my3 , and by nasa through nng06gj33 g and nnx10ah30 g .
allen , a. , li , z .- y . , & shu , f. h. 2003 , , 599 , 363 braiding , c. 2011 , unpublished phd thesis , macquarie university ciolek , g. e. , & knigl , a. 1998 , , 504 , 257 clarke , d. a. , norman , m. l. , & fiedler , r. a. 1994 , zeus-3d user manual ( tech . rep .
015 ; urbana - champaign : national center for supercomputing applications ) contopoulos , i. , ciolek , g. e. , & knigl , a. 1998 , , 504 , 247 dapp , w. b. , & basu , s. 2010 , , 521 , 56 galli , d. , lizano , s. , shu , f. h. , & allen , a. 2006 , , 647 , 374 hennebelle , p. , & fromang , s. 2008 , , 477 , 9 hennebelle , p. , & ciardi a. 2009 , , 506 , l29 huba , j. d. 2003 , in space plasma simulation , ed .
j. bchner , c. dum , & m. scholer , lecture notes in physics , vol . 615 ( berlin : springer ) , 166 krasnopolsky , r. , & knigl , a. 2002 , , 580 , 987 krasnopolsky , r. , li , z .- y . , & shang , h. 2010 , , 716 , 1541 ( kls10 ) li , z .- y . ,
krasnopolsky , r. , & shang , h. 2011 , , to be submitted ( lks11 ) li , z .- y . , & mckee , c. f. 1996 , , 464 , 373 machida , m. n. , inutsuka , s. , & matsumoto , t. 2010 , arxiv:1009.2140 mathis , j. s. , rumpl , w. , & nordsieck , k. h. 1977 , , 217 , 425 mellon , r. r. , & li , z .- y .
2008 , , 681 , 1356 mellon , r. r. , & li , z .- y .
2009 , , 698 , 922 nakano , t. , nishi , r. , & umebayashi , t. 2002 , , 573 , 199 nishi , r. , nakano , t. , & umebayashi , t. 1991 , , 368 , 181 sano , t. , & stone , j. m. 2002 , , 570 , 314 shu , f. h. , galli , d. , lizano , s. , & cai , m. 2006 , , 647 , 382 troland , t. h. , & crutcher , r. m. 2008 , , 680 , 457 wardle , m. , & knigl , a. 1993 , , 410 , 218 wardle , m. , & ng , c. 1999 , , 303 , 246 | stars form in dense cores of molecular clouds that are observed to be significantly magnetized .
a dynamically important magnetic field presents a significant obstacle to the formation of protostellar disks .
recent studies have shown that magnetic braking is strong enough to suppress the formation of rotationally supported disks in the ideal mhd limit .
whether non - ideal mhd effects can enable disk formation remains unsettled .
we carry out a first study on how disk formation in magnetic clouds is modified by the hall effect , the least explored of the three non - ideal mhd effects in star formation ( the other two being ambipolar diffusion and ohmic dissipation ) . for illustrative purposes
, we consider a simplified problem of a non - self - gravitating , magnetized envelope collapsing onto a central protostar of fixed mass .
we find that the hall effect can spin up the inner part of the collapsing flow to keplerian speed , producing a rotationally supported disk .
the disk is generated through a hall - induced magnetic torque .
disk formation occurs even when the envelope is initially non - rotating , provided that the hall coefficient is large enough .
when the magnetic field orientation is flipped , the direction of disk rotation is reversed as well .
the implication is that the hall effect can in principle produce both regularly rotating and counter - rotating disks around protostars .
we conclude that the hall effect is an important factor to consider in studying the angular momentum evolution of magnetized star formation in general and disk formation in particular . |
`` hard diffraction '' is an experimental phenomenon which lies at the borderline between `` hard '' and `` soft '' interactions .
it appears as a scattering process initiated by a hard probe ( e.g. a virtual photon at hera @xcite or a forward jet at the tevatron @xcite ) , but in which the proton target is not destroyed , in a similar way to conventional soft diffractive processes .
it is associated with large rapidity gaps between the hadronic remnants of the target and of the projectile as was discovered at hera , but was first identified by high-@xmath2 jets in diffraction at the isr @xcite .
there has been a lot of debate about the theoretical explanation of hard diffraction . indeed , assuming a partonic content of the pomeron led to a nice prediction @xcite of the phenomenon and , supplemented by the qcd evolution of diffractive structure functions , to a quantitative description of hera data @xcite . as satisfactory as can be this phenomenological analysis ,
the main unsolved problem in this approach is the lack of relation between diffractive and non diffractive hard scattering leading to a profusion of input parameters ( e.g. the non - perturbative parton distributions in the pomeron ) .
if one wants to enter more deeply in the study of hard diffraction and in the still mysterious nature of the pomeron interaction , one has to look for theoretical links with qcd .
this is our goal . in the present paper
, we shall focus on three existing different theoretical approaches of hard diffraction , for which we propose a new , unifying , formulation .
the first one we will refer to is the `` partonic pomeron '' approach .
the hard photon is here supposed to probe the parton distributions of the pomeron regge pole considered as a hadronic particle .
in fact we can also consider it as an extension of the regge theory of soft diffraction @xcite to incorporate the effect of the hardness of the probe . a second approach is the soft colour interaction one , where hard diffraction is described by a two - step process : during a relatively short `` interaction time '' , the probe initiates a typical hard deep - inelastic interaction .
then , at large times / distances , a soft colour interaction is assumed which will decide of the separation between diffractive and non - diffractive events according to a simple probabilistic rule : it gives rise to colour neutralization of the final state with probability of order @xmath3 ( where @xmath4 is the number of colours ) and thus to rapidity gaps and diffraction .
the assumed extreme softness of soft colour interaction ensures that the dynamics of hard partons remains unchanged .
various models based on soft colour interaction @xcite lead to satisfactory phenomenological descriptions .
a third approach is based on the small @xmath0 regime of perturbative qcd , where the resummation of leading @xmath5 contributions allows one to obtain some theoretical information on high energy hard interaction processes following the balitsky , fadin , kuraev , lipatov ( bfkl ) approach @xcite .
calculations of hard diffraction using the related qcd dipole approach @xcite or in the original bfkl framework @xcite have been performed .
in fact , in the line of qcd dipole models for proton structure functions @xcite , models @xcite for diffractive proton structure functions have been derived and give a convenient description of hera data . for further abbreviation
we shall call it the `` qcd dipole '' approach . keeping a general point of view , one may notice that each of these approaches has advantages and disadvantages with respect to the others .
the partonic pomeron approach synthetises what we know about the effect of the hard probe and the factorization properties of the hard interaction but gives no prediction about the soft pomeron dynamics .
the soft colour interaction approach give a nice relation ( even in normalization ) between hard and soft deep - inelastic scattering but the nature of the colour rearrangement at long distances is unknown .
finally the qcd dipole approach gives detailed prescriptions for the form of the amplitudes , but the problem of the perturbative / non - perturbative qcd interface and the crucial point of the relative normalization of diffractive _ vs _ non - diffractive contributions , remain obscure .
in the present paper we will show that the three aproaches may find a common formulation and intrinsic mutual equivalence through s - matrix relations on different analytic discontinuities of @xmath6 forward elastic amplitudes in the so - called triple - regge regime . within this unifying s - matrix framework and using the hard probe as a hint for the application of perturbative qcd to the calculation of these discontinuities whenever it is reasonably justified
, we will end with a constrained formulation of the diffractive structure functions , transcending the limitations of each of them taken separately .
the plan of the paper is as follows : in the next section we will give the general framework , using both the s - matrix and perturbative qcd properties , allowing one to mutually relate the three abovementionned approaches to hard diffraction .
in the next section 3 , concentrating on the _ inelastic _ component , we show the compatibility of the qcd dipole and partonic pomeron approaches . in section 4 the same is done for the qcd dipole and soft colour interaction approaches . as a result , a determination of the longitudinal and transverse diffractive structure functions emerges incorporating both perturbative and non perturbative ingredients in an overall consistent qcd picture of hard diffraction . in the last section 5 ,
we give a summary of our results , and an outlook on phenomenological tests and theoretical consequences of our unifying approach .
the usual kinematical variables for the diffractive amplitudes and cross - sections are @xmath7 the photon virtuality , @xmath8 the total rapidity available for the final hadronic state , @xmath9 the rapidity gap , @xmath10 where @xmath2 is the momentum transferred at the proton vertex .
we use also @xmath11 where @xmath12 is the invariant mass of the diffractively produced state but due to the sharp cut - off in t , one can often neglect t in the denominator . ] . in the following
we shall make use of the important s - matrix connection , also called mueller - regge relation @xcite , between semi - inclusive amplitudes and specific discontinuity contributions of forward elastic @xmath6 amplitudes .
it naturally applies to hard diffraction initiated by a virtual photon , as sketched in fig.1 , namely ^ * + pp + x disc_1\{^ * |p p ^ * |p p } .
[ 3to3 ] in fact , this relation , derived in the context of soft physics @xcite , can also be applied to the definition of hard diffraction when @xmath13 and @xmath14 are large enough ( typically , @xmath15 in this case , the mueller - regge relation applies in a region where the pomeron regge poles are supposed to take into account the colour singlet exchange responsible for the gap .
if moreover , the diffractive mass @xmath12 is large enough , one considers a pomeron - like singularity describing the large @xmath12 behaviour .
however the resulting triple - pomeron contribution is to be interpreted here in a loose sense since this third regge singularity may differ from the other ones due to the hard probe , as mentionned in the previous section .
we will thus keep the word pomeron for those responsible for the gaps . in the triple - regge kinematical region ,
one thus writes @xcite : ~disc_1 a(3 3 ) g_p(t ) _ ^2 e^2(-1 ) y ^tot_^*-p , [ regge ] where @xmath16 is the `` unintegrated '' ( t - dependent ) diffractive structure functions ( related to the measurable hard diffractive cross - sections ) , _ ( t ) ( -(t ) ) \{1+e^-i(t ) } [ phase ] is the ( partonic ) pomeron regge phase factor in the amplitude and @xmath17 is the ( partonic ) pomeron - pole regge trajectory with intercept @xmath18 and slope @xmath19 .
@xmath20 is a function , not constrained in the regge formalism , describing the coupling of incident pomerons to the proton .
the pomeron - photon cross - section @xmath21 describes the interaction of the hard probe with the incident pomeron .
for instance , in the partonic pomeron model @xcite , it is expressed in terms of pomeron structure functions , in much the same way as proton - photon cross - sections are expressed in terms of the proton structure functions , but with a quite different partonic content @xcite .
quite interestingly , the existence of the regge phase factors ( [ phase ] ) allows one to relate other discontinuities of @xmath22 to @xmath23 as sketched in fig.2 , one may consider a double discontinuity @xmath24 taking into account also the analytic discontinuity in the subenergy of one of the incident pomeron exchanges and then a triple discontinuity @xmath25 including the discontinuity over the two pomeron exchanges .
the expression of the discontinuities , through generalized unitarity relations , is obtainedthrough the imaginary part of the relevant regge phase factors .
one writes eliminates already all terms but one in the rather involved expression of the uncut vertex . ]
@xcite : disc_1 a = disc_2 a = disc_3 a = v_p|p , disc where @xmath26 ia a real function describing the @xmath27 vertex in @xmath28 the notation @xmath29 indicates that the third regge exchange , corresponding to the summation over diffractive final states , may have ( and indeed _ has _ in our calculations ) a different effective trajectory avoiding a definition problem in ( [ disc ] ) . ] . beyond these general relations of the s - matrix framework in the triple - regge regime , we will now take advantage of the hard probe , allowing one to introduce in the game the ( resummed ) perturbative qcd expansion at high energy ( small @xmath0 ) .
let us start with general considerations . in a generic s - matrix approach , the analytic discontinuities of scattering amplitudes
are related to a summation over a complete set of asymptotic _ hadronic _ final states . if however , the underlying microscopic field theory is at work with small renormalized coupling constant due to the hard probe , it is possible in some cases to approximate the same discontinuity using a complete set of _ partonic _ states .
in particular , at high energy and within the approximation of leading logs ( and also large @xmath4 ) , qcd dipoles can be identified as providing such a basis .
the interest is that the wave function of incoming states at the hard vertex is completely determined at the leading log level . using this wave - function and @xmath30 dipole transitions
are needed @xcite . ] together with the appropriate @xmath31 factorization properties at the proton vertices @xcite , one is able to estimate the contributions to the various discontinuities of @xmath32 depicted in fig.2 within the ( resummed ) perturbative qcd framework .
hence , the s - matrix relations ( [ disc ] ) , which express a simple phase relation in a pure regge framework , acquire a non - trivial meaning if one considers the various perturbative qcd ingredients and non - perturbative interfaces for the different discontinuities . in some sense , at it is shown in fig.3 , the discontinuities can be associated , whenever it is possible , with the hard qcd interaction at some time @xmath33 supplemented by a perturbative resummation of qcd radiation around this interaction time .
the perturbative / non - perturbative interface has to be taken into account in order to put a reasonable limit to this derivation .
depending on the discontinuity considered , one has different interaction pictures with a hard interaction at short time , and soft interaction diluted in space - time . instead of considering these pictures as different , we will use the s - matrix relations as a link between these pictures and thus will get non - trivial relations , as we will see later on .
let us comment in turn on the pictures corresponding to @xmath34 * i ) * @xmath35 as we have seen from relation ( [ regge ] ) , is parametrized in terms of soft pomeron trajectories and of @xmath36 the unknown quantity parametrizing the interaction of the incident pomeron with the hard probe .
the same @xmath21 is the one described by the partonic structure functions of the pomeron in the original scheme of ref.@xcite .
@xmath37 is thus a natural framework for the partonic pomeron approach , since the analytical discontinuity over the diffractive final states is effectively expressed in terms of the hard interaction of the photon with the partons inside the incident pomeron .
this is schematically depicted in the `` time - dependent '' scheme of fig.3 , where the soft incident pomerons are represented as living during a `` long '' time , while the partonic ( or qcd dipole ) process is initiated by the hard interaction .
* ii ) * @xmath38 is quite interesting since it appears as a good candidate for a description of the soft colour interaction approach , see fig.3 .
it appears as a partonic interaction very similar to the one describing ordinary deep - inelastic processes , in parallel with a `` soft '' correction evolving during a long time , corresponding to the uncut pomeron singularity in fig.3 . in section 4
, we will see how this can be quantified using the interrelation with the qcd dipole picture .
* iii ) * @xmath39 in this case the discontinuity is the `` full '' one , cutting the two pomerons together with the @xmath40 interaction .
indeed , this full cutting is required by consistency of the qcd dipole picture based on a qcd calculation of the partonic states at the interaction time @xcite . in the description of hard diffraction ( see , e.g. , @xcite , the amplitude is deduced from a @xmath41 dipole transition , starting either from the primordial @xmath42 state of the photon ( `` elastic component '' ) or from a dipole excited in the photon wave function for higher diffractive masses ( `` inelastic component '' ) . in the following we shall focus on the main `` inelastic component '' , leaving the `` elastic component '' for a further study .
this last analysis requires some more care about the perturbative / non - perturbative interface , which can not be restricted to usual factorisation properties . as studied in ref .
@xcite , one has to distinghuish the cases whether we consider strictly the forward @xmath43 case or not .
indeed , whenever @xmath44 the qcd radiation accompanying the hard probe extends far into the infrared region at the triple pomeron vertex , making the perturbative treatment difficult . at @xmath45
due to an interesting conservation law of conformal dimensions in the bfkl framework , the perturbative resummation acquires more credibility , at least in the same approximation as for the total structure functions , .
we will thus stick to using the perturbative calculations at @xmath43 , more precisely inspiring ourselves from those performed in paper ref .
@xcite in the framework of the qcd dipole model limit of the qcd dipole model .
the inclusion of such a state will be welcome , especially for large diffractive masses , but lies beyond the scope and the dipole framework of our present paper . ] . due to the relations ( [ disc ] )
, we expect that all three descriptions of hard diffraction , even if having seemingly distinct `` time evolutions '' , indeed correspond to an unique one in a complete qcd theory satisfying both macroscopic ( s - matrix ) and microscopic ( with interacting quarks and gluons ) requirements . in the absence
yet of such a complete realization of the strong interaction theory , we shall adopt the following point of view : looking for a synthesis , we will first check whether all three approaches are consistent between themselves and then formulate hard diffractive structure functions realizing the synthesis .
our starting point is the formula for the differential structure functions ( inelastic component ) at @xmath46 for longitudinal and transverse photon given in @xcite : ( q^2,y , y;p_t=0 ) = = _ c - i^c+i ( 1-_1-_2- ) pf_t , l ( qq_0)^2 \{y [ ( _ 1)+(_2 ) ] + ( y - y ) [ ( ) ] } , biczy where ( ) ( ) = \{2(1)-2()-2(1- ) } chi is the bfkl evolution kernel @xcite , @xmath47 the non - perturbative scale of primordial dipoles in the proton @xcite , ^diff n^2_eff , nordif the normalization with @xmath48 the phenomenological number of primordial dipoles in the proton at scale @xmath47 and @xmath49 the sum over squared quark charges .
the other prefactors of perturbative qcd origin are : pf_t , lhg(_1 ) hg(_2 ) i_t , l ( ) g^-1 ( ) , pf with h()=1 4 ^ 2(1-)^2 ; g()= , hg and finally , ( i_l + i_t ) 1 ( 1- ) ( + ( 1+)(2- ) ) , [ itl ] which are the ` ` impact factors '' @xcite of the longitudinal and transverse photon in the bfkl formalism .
@xmath50 is a cut - off function taking into account the non - perturbative description of the end of the rapidity spectrum of emitted gluons @xcite . in practice
, it mainly affects the overall normalization ( also the form at edges of the @xmath51-dependence ) .
it is one aspect of the lack of predictivity of the qcd dipole approach about the relative normalization of diffractive vs. non diffractive contributions
. one important ingredient of formula ( [ biczy ] ) is the delta function @xmath52 it expresses the conservation law of conformal dimensions at the triple pomeron vertex , which is a special feature of the bfkl property in the forward direction .
as discussed in the previous section , we have choosen the qcd dipole formulation for the forward diffraction instead of the non - forward one @xcite , since it is expected to be less affected by non -perturbative contributions . on the other hand , it is phenomenologically not difficult to insert the non - perturbative phenomenological information on the @xmath53-dependence coming from the proton vertices in the non - forwardregion .
indeed , it is convenient to parametrize the integrated diffractive structure function assuming an exponential behaviour @xmath54 well verified experimentally is presently known @xcite only on a global average to be of order @xmath55 it could depend on other variables but it happens to be very similar to the diffractive photoproduction slope , confirming the stability of this non perturbative parameter in different kinematical regions .
we shall assume it constant in the following .
we will thus in the following assume the relations : f_t , l^diff(q^2,y , y ) dp_t^2
< p_t^2>(p_t=0 ) .
fdiff the first step of the computation of formula ( [ biczy ] ) is @xcite to use the saddle - point approximation at large @xmath56 to integrate over the difference @xmath57 one easily gets ( p_t=0)= n^diff1x_p q_0 ^ 2 _ c - i^c+i pf_t , l \{2y [ ( 2 ) ] + ( y - y ) [ ( ) ] + 2 qq_0 } . biczy1 for comparison the qcd dipole formula for the total proton structure functions @xcite is : f_t , l = n^tot _ c - i^c+i e^y ( ) ( qq_0)^2 h ( ) i_t , l ( ) , tot with ^tot = n_eff e_f^2 .
normtot in order to proceed further in our theoretical analysis , it is suitable to look for an analytical solution of ( [ biczy1 ] ) . for this sake we shall now use another saddle - point approximation near @xmath58 in previous analyses , the integral
was either approximated in some regions of the variables @xcite or numerically evaluated @xcite .
for sake of a theoretical discussion in parallel with the formula ( [ tot ] ) for the total structure functions , we better choose a gaussian approximation valid in a reasonably large physical region .
the gaussian approximation makes use of the following expansions ( 2 ) & & + 8 ^ 2 + o(^4 ) ( ) & & + 2 ( 12-)^2 + o((12-)^4 ) , expan where by definition ( = 12 ) = 42 ; ^
^(=12 ) = 28 ( 3 ) .
zeta note that @xmath59 which allows one to use the gaussian saddle - point for diffraction in a region avoiding the edges of integration valid when @xmath60 @xcite and with numerical estimates @xcite .
more precisely , @xmath61 giving @xmath62 of order 1 .
hence the approximations ( [ expan ] ) will be good for @xmath63 less for @xmath64 but enough for our theoretical purpose . ]
@xmath65 .
it is important to realize here a different expected result with the gaussian approximation for the total structure function ( [ tot ] ) in the bfkl formalism where one uses a saddle - point approximation near @xmath66 in particular , the dynamical exponent @xmath67 will be moved away from its canonical value at @xmath68 even without the virtuality factor @xmath69 after using the gaussian approximation , and relation ( [ fdiff ] ) , the resulting structure functions read : f_t , l^diff(q^2,y , y ) = n^diff pf_t , l \ { ( y - y ) * * _ s } ( qq_0)^2*_s * ( - ) , biczy2 where the saddle - point determines both the new `` effective dimension '' and `` effective intercept '' : = ; * * _ s = + epsilon for the pomeron - photon cross - section . by definition
eta the connection between the qcd dipole model and the partonic pomeron is thus well illustrated by the identification of our resulting formula ( [ biczy2 ] ) with the triple regge prediction ( [ regge ] ) considered at @xmath70 indeed , we identify the partonic pomeron intercept as @xmath71 which is natural in a bfkl framework at @xmath72 see the following discussion .
we get now the following new relation : ^tot_^*-p ~ \ { ( y - y ) * * _ s } ( qq_0)^2*_s * ( - ) , biczy3 qcdregge with ( [ epsilon ] ) defining the `` effective '' anomalous dimension and intercept of a `` deformed '' bfkl formula for the photon pomeron ( with zero mass @xmath73 ) total cross - section .
it is interesting to compare expression ( [ biczy2 ] ) and @xmath74 with the canonical bfkl formula ( [ tot ] ) in the saddle - point approximation namely : f_l , t ^tot h i_t , l ( qq_0 ) ( - ) .
tot2 the `` effective bfkl '' parameters for @xmath21 differ from the bfkl ones by an amount depending on @xmath75 also note that the `` diffusion '' term , which quantifies the amount of transverse momentum drift of gluons ( or dipoles ) along the bfkl evolution is renormalized by a factor @xmath76 the same as for the anomalous dimension @xmath77 while at large @xmath78 the formulae become similar , since for instance @xmath79 they sizeably differ from them provided @xmath80 in fact , this is realized in practice , since in present experiments @xmath81 on contrary , the features of the pomeron exchanges responsible for the gaps are those of the `` bare '' bfkl values taken at @xmath82 which can be interpreted as the partonic pomerons in the bfkl formalism since the scale @xmath47 is characteristic of the starting point of the qcd evolutions .
thus the `` deformed '' bfkl parameters ( [ epsilon ] ) lead to a situation which can be called intermediate between a `` hard '' ( with effective dimension @xmath83 ) and a `` soft '' ( with @xmath84 ) pomeron , depending on the ratio @xmath85 also , the energy dependence is faster for the `` deformed '' bfkl singularity than for the partonic pomeron .
this original situation is confirmed by numerical estimates @xcite .
in order to confront and unify the qcd dipole and soft colour interaction approaches , we will estimate the overall contribution of hard diffraction to the total structure function at fixed value of @xmath86 as previously done , we will stick here to the `` inelastic '' ( leading - twist ) diffractive component leaving for further study the `` elastic '' component with small diffractive masses .
let us then consider the following integral : f_t , l^diff / tot & = & _ x_bj^x_gap dx_pdp_t^2 ( q^2,y , y;p_t ) & & _
x_bj^x_gap dx_pdp_t^2 < p_t^2 > ( q^2,y , y;p_t=0 ) & = & _ c - i^c+i ( 1-_1-_2- ) pf_t , l & & ( qq_0)^2_y_gap^ydy e^ , biczy4 where @xmath87 is the minimal value retained for the rapidity gap .
indeed , the integral over @xmath56 is expected to be dominated by the behaviour of the integrand for large enough gaps .
we also keep the approximation @xmath88 technically , the integrals in ( [ biczy4 ] ) can be performed using a saddle - point method in both variables @xmath89 and @xmath56 provided the conditions y 1 ; 1 conditions are realized .
the saddle - point equations , see appendix * a2 * , lead to remarkably simple solutions y_c & = & ( y+ ) ( 1+)^-1 ( _ c)&=&2 ( ) solution resulting in a value of @xmath90 which is `` universal '' , i.e. independent of the kinematics of the reaction .
the saddle - point calculation leads to the following expression : f_t , l^diff / tot = ^diff ( qq_0)^2_c .
biczy7 note that @xmath91 when @xmath92 let us now come to the main outcome of this calculation .
let us compare the result ( [ biczy7 ] ) to a _
total _ hard contribution of bfkl type , see equation ( [ tot ] ) , which can be rewritten in a saddle - point approximation : f_l , t ^tot h(_s.p . ) i_t , l(_s.p . )
( qq_0)^2_s.p . , tot1 with the bfkl saddle - point value _ s.p .
= 12 ( 1- 4 ) .
value hence , ( [ biczy7 ] ) is strikingly similar to a bfkl formula ( [ tot ] ) with a shift in the value of the saddle - point namely : _
shift it is remarkable that the total diffractive contribution to the structure function has the same analytical form as the non - diffractive one , up to the substitution ( [ shift ] ) and the prefactors , which we will study next .
this qcd derivation is thus compatible with the postulate of the soft colour interaction approach , namely the relation between the diffractive and non diffractive deep - inelastic processes stating that the diffractive part of the total hard cross - section is not intrinsically different . in the sci framework , this is explicitely realized as a whole in the first paper of ref .
@xcite while it is realized graph by graph in the approach of paper @xcite . in the framework of our qcd calculations , it is however realized in a somewhat modified way , through a shift ( [ shift ] ) in the `` effective '' parameters of the hard interaction
we will comment later upon possible phenomenological consequences of this effect . using our identification of the hard interaction of bfkl type present in @xmath93 we are now able to take advantage of the second postulate of the soft colour interaction approach , i.e. the probabilistic evaluation of the soft rearrangement of colour at large distances , leading to the famous factor @xmath3 in the ratio of diffractive vs. non diffractive cross - sections .
this beautifully simple proposal expected from non perturbative qcd properties can be incorporated in the unified approach in an easy way .
we are thus led to a proportionnality between the prefactors of the bfkl expressions ( [ tot2 ] ) and ( [ biczy7 ] ) , namely n^tot h ( ) i_t , l ( ) .
ident using our result ( [ biczy7 ] ) and assuming thus a relation between the prefactors compatible with ( [ ident ] ) , the total diffractive contribution can be rewritten f_t , l^diff / tot n^tot
h(_c ) i_t , l(_c ) ( qq_0)^2_c .biczy9 coming back now to our saddle - point result ( [ biczy2 ] ) for the differential hard diffractive structure function ( [ biczy ] ) , one rewrites f_t ,
l^diff(q^2,y , y ) \ { ( y - y ) * * _ s } ( qq_0)^2*_s * , biczy10 where @xmath94 and @xmath95 are defined in ( [ epsilon ] ) and a _ known _ normalization ^(2)+^(_c ) h(_c ) i_t , l(_c ) ^tot , norm with @xmath96 as in ( [ biczy9 ] ) .
note that we could also have considered @xmath97 instead of the constant @xmath96 in formula ( [ norm ] ) since the prefactors are expected to be slowly varying functions of @xmath98 in fact , care must be taken about poles at @xmath99 contained in the functions @xmath100 ( the same poles appear consistently in both sides of ( [ ident ] ) , as explicit in the definition ( [ pf ] ) of @xmath101 ) .
a normalization correction could appear in ( [ biczy9],[biczy10 ] ) by a known rescaling factor @xcite .
we leave a detailed phenomenological study for futher work .
the relations ( [ biczy9],[biczy10 ] ) call for some comments . on a theoretical point of view
, these relations are out of range of the present knowledge of qcd , for its large distance , confining regime .
however , some arguments are in favor of such kind of relations . for instance
, there exists a complementarity between the perturbative results and the non perturbative ansatze present here . the value of the strong coupling constant @xmath102 the number of primordial dipoles @xmath48 and , in the derivation @xcite of hard diffraction , the expression of the cut - off @xmath50 are not determined in the bfkl approach which essentially means that the relative normalization of diffractive vs. non - diffractive contributions remain unknown .
it is just this ratio which is determined by the soft colour interaction anstz . on the other hand , the non - trivial form of the structure functions inherited from the qcd dipole calculations
could not be guessed just from what we know about soft colour processes , which are limited by our present lack of knowledge of the theory of strong interactions at long distances .
summarizing our results , let us quote : * 1 ) * using s - matrix properties of triple - regge contributions , a relation is found between three popular approaches to hard diffraction : the partonic pomeron , soft colour interaction and qcd dipole formulations . * 3 ) * the `` effective '' parameters of the @xmath96-pomeron total cross - section corresponding to the partonic pomeron are determined from leading log perturbative qcd resummation .
they are found to depend not only on on @xmath13 but also on rapidities through the parameter @xmath103 * 4 ) * the diffractive structure functions @xmath104 are found to be determined both in form and normalization , using their perturbative ( through qcd dipoles ) and non perturbative ( through soft color interaction ) relationship with the total structure functions @xmath105 concerning the _ partonic pomeron _ approach , we find that the `` effective '' parameters of @xmath106 depend on a parameter @xmath78 external to the @xmath107 reaction .
looking to the deep reason of this fact , we find the constraint @xmath108 which prevents the bfkl pomeron singularities of the involved bfkl pomerons from taking their ordinary saddle - point positions . hence those pomerons responsible for the gaps become @xmath13-independent while the one in @xmath106 becomes both @xmath13 and @xmath78-dependent .
this kind of `` off - shellness '' is quite remarkable , showing that a universal pomeron kernel @xmath67 can lead to non - universal effective behaviours when put in a hard interaction context
. it would be interesting to look for an experimental study of this intringuing feature of the partonic pomeron .
the compatibility with dglap evolution equations is also to be studied in detail .
concerning the _ soft colour interaction _ prediction , relating the diffractive and non - diffractive components of deep - inelastic scattering , it seems quite compatible with the qcd predictions on the form of the amplitudes .
however , an unexpected difference appears , since the typical values of the `` effective '' dimensions are different in both cases .
it is clear for the integrated diffractive contribution @xmath109 , see ( [ biczy4 ] ) , comparing @xmath110 and @xmath111 it is also explicit from the analysis of @xmath112 itself that the relevant regions in the inverse mellin transforms are always for values of @xmath89 lower than for @xmath113 it will be interesting to test this prediction with data . for developping this study beyond the first theoretical step
, we would like to stress the feasability of phenomenological tests and applications .
however this requires a similar analysis for the `` elastic '' qcd dipole component .
one knows that the mueller regge formulae can be extended to generalized pomeron regge amplitudes ( e.g. @xcite ) .
we find interesting to elaborate on the s - matrix discontinuity properties in this context and their possible relations with `` higher - twist '' qcd contributions which are known to be relevant in this case .
this deserves a specific study which is beyond the scope of this paper . on the theoretical ground ,
some interesting problems arise .
for instance , the known problem of the application of a ( resummed ) perturbative scheme at @xmath114 require a better understanding of the matching between perturbative and non - perturbative features . also are interesting to develop the more precise extension of this scheme to regge cuts , next - leading bfkl contributions and , last but not least , unitarity corrections , which are required to identify properly the soft pomeron contributions .
finally , the study of tevatron results on diffraction will be welcome , since the soft colour interaction approach seems to give some interesting clues , while factorization of the parton structure functions of the pomeron appears to be violated .
thanks are due to jochen bartels , who draw the attention of one of us ( r.p . ) to the relations betwen regge discontinuities cf .
the papers @xcite , to andrzej bialas for fruitful discussions and the inspiring paper in collaboration with wieczlaw czyz @xcite , and to christophe royon , for his help and advice .
+ + starting with eq . ( [ biczy4 ] ) , the integration over @xmath115 leads to f_t , l^diff=
n^diff _ c - i^c+i pf_t , l ( y , ) , biczy11 where the function ( y,)\{2y ( 2 ) + ( y - y ) ( ) + 2 qq_0 } debut determines the saddle - point equations , namely & & -y ^ ( 2 ) + ( y - y ) ^ ( ) + 2 qq_0=0 & & 2()-()=0 , cols whose solution @xmath116 is given in ( [ solution ] ) , see the text .
the prefactors appearing in the integrated formula are coming from the partial second derivatives of @xmath117 a first one comes from the partial second derivative over @xmath89 1= 1 .
pref1 the second one is slightly more involved .
it reads = .
pref2 the simplification appearing in the second prefactor comes from an identity + = 0 identity at the saddle - point @xmath118 which came from the calculation .
f. abe et al . , cdf coll . , _ phys .
* 78 * ( 1997 ) 2698 ; * 79 * ( 1997 ) 2636 ; * 80 * ( 1998 ) 1156 ; s. abachi et al . ,
lett . _ * b440 * ( 1998 ) 5278 .
t. affolder et al . , cdf coll . , _ phys .
_ * 84 * ( 2000 ) 5043 ; * 85 * ( 2000 ) 4215 .
a. capella , a. kaidalov , c. merino , j. tran thanh van , [ hep - ph/9408212 ] , _ phys .
_ * b343 * ( 1995 ) 403 ; a. capella , a. kaidalov , c. merino , d. pertermann , j. tran thanh van [ hep - ph/9506454 ] , _ phys .
rev . _ * d53 ( 1996 ) 2309 * ( 1996 ) 2309 .
w. buchmuller , a. hebecker , [ hep - ph/9504374 ] , _ phys .
* b355 * ( 1995 ) 573 ; [ hep - ph/9512329 ] , _ nucl .
* b476 * ( 1996 ) 203 .
an extensive review of this and other approaches can be found in a. hebecker , [ hep - ph/9905226 ] , _ phys .
* 331 * ( 2000 ) 1 .
a. edin , g. ingelman , j. rathsman , [ hep - ph/9605281 ] , _ z. phys .
_ * c75 * ( 1997 ) 57 . for a review and discussion including tevatron results ,
see : g. ingelman , [ hep - ph/9912534 ] , lectures at st .
croix summer school 1998 : _ techniques and concepts of high energy physics x _ , ed .
t. ferbel , kluwer academic publishers 1999 , p. 597 .
l.n.lipatov , _ sov.j . nucl.phys._
* 23 * ( 1976 ) 642 ; v.s.fadin , e.a.kuraev and l.n.lipatov , _ phys .
lett . _ * b60 * ( 1975 ) 50 ; e.a .
kuraev , l.n .
lipatov and v.s .
fadin , _ sov.phys
* 44 * ( 1976 ) 45 , * 45 * ( 1977 ) 199 ; i.i.balitsky and l.n.lipatov , _ sov.j . nucl.phys._
* 28 * ( 1978 ) 822 ; l.n .
zh.eksp.teor.fiz . _
* 90 * ( 1986 ) 1536 ( eng .
. jetp _ * 63 * ( 1986 ) 904 ) .
a. h. mueller , b. patel , _ nucl .
_ * b425 * ( 1994 ) 471 ; + a.bialas , h.navelet , r.peschanski , _ phys .
_ * b427 * ( 1998 ) 147 .
+ for a related approach , mixing perturbative and non - perturbative considerations , see : n. n. nikolaev , b. g. zakharov , _ z. phys . _ * c49 * ( 1991 ) 607 ; _ z. phys .
_ * c53 * ( 1992 ) 331 .
a. bialas , r. peschanski , _ phys .
_ * b378 * ( 1996 ) 302 ; * b387 * ( 1996 ) 405 .
+ a.bialas , r.peschanski and c.royon , _ phys .
_ * d57 * ( 1998 ) 6899 ; s.munier , r.peschanski and c.royon , _ nucl .
phys . _ * b354 * ( 1998 ) 297 .
detar , c.e .
jones , f.e .
low , j.h .
weis , j.e .
young , c - i tan , _ phys . rev
_ * 26 * ( 1971 ) 675 . for a detailed account of the origin of the soft pomeron formula , see w.r .
frazer , _ rev
* 44 * ( 1972 ) 284 .
s. catani , m. ciafaloni , f. hautmann , _ nucl .
phys_. * b366 * ( 1991 ) 135 .
j. c. collins , r. k. ellis , _ nucl .
* b360 * ( 1991 ) 3 .
e. m. levin , m. g. ryskin , yu . m. shabelskii , a. g. shuvaev , _ sov .
_ * 53 * ( 1991 ) 657 . | using a combination of s - matrix and perturbative qcd properties in the small @xmath0 regime , we propose a formulation of hard diffraction unifying the partonic ( ingelman - schlein ) pomeron , soft colour interaction and qcd dipole descriptions .
in particular , we show that all three approaches give an unique and mutually compatible formula for the proton diffractive structure functions @xmath1 incorporating perturbative and non perturbative qcd features . |
in this paper we will pursue some technical details concerning the computation and the physical understanding of the recent analysis by stupakov @xcite on the effect of coulomb diffusion on bunching coefficients in echo enabled harmonic generation ( eehg ) free electron laser ( fel ) seeded devices @xcite .
the considerations developed in this paper should be understood as a complement to refs .
@xcite , with the aim of providing a more general computational framework , benefitting from the formalism of beam transport employing algebraic means @xcite . in this treatment
the transport through magnetic elements is treated in terms of exponential operator acting on an initial phase space distribution and usually do not contain transport elements provided by a heat type diffusion . here
we will show that diffusion mechanisms can be included in such a framework in a fairly straightforward way @xcite , preserving the spirit of exponential operator concatenation .
the effect of the coulomb diffusion on the e - beam longitudinal phase space distribution is ruled , according to @xcite , by the equation @xmath0 where @xmath1 is the coordinate of propagation , @xmath2 is the longitudinal coordinate in the beam , @xmath3 is associated to the beam energy , @xmath4 is the coulomb diffusion coefficient and @xmath5 is the initial distribution .
we will discuss the physical meaning of the previous quantities and we will provide their explicit expression in the following , for the moment we note that , having assumed that @xmath4 is independent of @xmath3 , we can solve via the following gauss - weierstrass ( g - w ) transform @xcite @xmath6 the bunching coefficients can be defined as those of the fourier components of the expansion @xcite @xmath7 and are , therefore , specified by @xmath8 the dilution of the bunching coefficients induced by the coulomb diffusion can therefore be mathematically expressed as @xmath9 following ref .
@xcite , we assume the initial distribution @xmath10^{2}}{2}}\label{initdistr}\ ] ] whose physical meaning will be discussed in the forthcoming section .
the solution of eq . can be therefore easily computed , even though can not be obtained in analytical form . in fig .
1 we have reported the effect of the dispersion on the initial distribution ( fig .
1a ) , after a drift @xmath11 with @xmath12 ( fig .
@xmath13 undergoing the coulombian diffusion .
a ) @xmath14 b ) @xmath15.,title="fig : " ] undergoing the coulombian diffusion .
a ) @xmath14 b ) @xmath15.,title="fig : " ] it is evident that the coulomb term smears out the oscillations , associated with the bunching coefficients , which tend to disappear or to be strongly suppressed , as we will further discuss in the following .
if we neglect in eq .
the contributions @xmath16 we can expand @xmath17 in series of bessel functions , namely @xmath18 where @xmath19 are cylindrical bessel functions of the first kind .
the associated bunching coefficients are , neglecting contributions in @xmath20 , provided by @xmath21 the g - w transform can be accordingly written as ( see appendix ) @xmath22 and bunching coefficients now read @xmath23 @xmath24 the effect of the coulombian diffusion is therefore twofold , it induces \a ) a dispersion due to the term @xmath25
\b ) a suppression of the higher orders harmonics occurring through @xmath26 . in these introductory remarks
we have provided a few remarks on the type of formalism we are going to use to treat the bunching mechanism in eehg fel seeded devices , the forthcoming sections will cover more physical details .
we have quoted the initial distribution given in eq . without any comment about its physical meaning .
its origin can be traced back to the following liouville equation @xmath27 @xmath28 @xmath29 which rules the evolution of an ensemble of non - interacting particles , driven by the hamiltonian @xmath30 the solution of eq . can be cast in the form @xmath31
@xmath32 we have denoted by @xmath33 and @xmath34 the evolution and liouville operators , respectively , of our dynamical problem @xcite .
we note that the liouville operator breaks into two non - commuting parts , therefore any treatment of the associated evolution operator demands for an approximate disentanglement of the exponential , which , for simplicity will be assumed to be provided by @xcite @xmath35 which is accurate to @xmath36 $ ] .
the use of the rule @xmath37 allows to cast the solution of eq . in the form
@xmath38 therefore by setting @xmath39 and keep as initial distribution a gaussian in @xmath3 , we recover eq . . the physical scheme
, we are dealing with , is reported in ref .
@xcite ) , to which the reader is addressed for further details .
the e - beam undergoes two successive modulations induced by two different lasers in two different sections .
the beam initial distribution is that occurring after the first chicane at the entrance of the second modulator .
the physical content of the previous variables is the following @xmath40 @xmath41 where @xmath42 is the e - beam energy spread , @xmath43 is the induced energy modulation , @xmath44 is the laser wave vector .
the dynamics of the bunching coefficients can be obtained from eqs . and as ( @xmath45 ) @xmath46.\label{ziter}\ ] ]
@xmath47 eq . yields an idea of the coupling between the various bunching coefficients and has already been derived in a closely similar fom in ref .
@xcite .
the inclusion of the coulomb diffusion in this scheme can be modeled by transforming the liouville into a vlasov equation , namely @xmath48 the diffusion coefficient , expressed in practical units , reads @xcite @xmath49}{\varepsilon_{x}\left[\mu m\right]\sigma_{x}\left[100\mu m\right]\left(\sigma_{e}\left[kev\right]\right)^{2}}\ ] ] with @xmath50 and @xmath51 being the emittance and beam section respectively the solution of eq .
can be obtained as @xmath52 according to this approximation the effect of the diffusion is calculated separately from that induced by the liouvillian contribution .
higher order disentanglements @xcite can be used to get more accurate results , but the present approximation is adequate for our purposes . in fig .
2 we report the contour plots of the liouville distribution with and without the effect of the coulombian diffusion for the same s value . it is evident that the presence of a non - zero @xmath4 value provides a significant reduction of the distribution harmonic content . a global view is provided by fig .
3 where we have reported the 3d plot of the liouville distribution under the action of diffusion for different s - values . .,title="fig:"].,title="fig : " ] ) for different s values : a ) @xmath53 , b ) @xmath54 , c)@xmath11.,title="fig : " ] ) for different s values : a ) @xmath53 , b ) @xmath54 , c)@xmath11.,title="fig : " ] @xmath13@xmath55 ) for different s values : a ) @xmath53 , b ) @xmath54 , c)@xmath11.,title="fig : " ] the nice feature of the procedure we are adopting is its modularity , which will be further appreciated in the forthcoming section .
in the introductory section we have expanded the initial distribution , by neglecting the contributions in @xmath56 .
such an approximation implies that the induced energy modulation is not large compared to the natural e - beam energy spread .
however if we relax these assumptions and employ the formalism of generalized bessel functions @xcite , we can obtain a more general view on the problem under study .
it has indeed been shown that two variable bessel functions can be expressed through the generating function @xmath57 where @xmath60 are two variable hermite - kamp d frit polynomials @xcite .
they are actually understood as belonging to the family of hermite based functions and are widely exploited to deal with radiation scattering problems @xcite , going beyond the dipole approximation . which holds if the laser , in the second modulator , is the same of the first .
we have removed the @xmath1 variable because absorbed into the @xmath67 and @xmath68 terms .
if we use the bessel function expansion @xmath69 we can therefore cast the distribution at the end of the second modulator as @xmath70 the effect of the coulombian diffusion on the distribution can be easily computed , either numerically or analytically . by keeping the lowest order contribution @xmath74 and by approximating the bessel function as from the above equation we easily deduce that the coulombian contribution yields a suppression factor proportional to @xmath77 .
more in general as also pointed in ref .
@xcite the larger is the order of bunching the more significant is the effect of the reduction .
the formalism we have presented in this paper provides a fairly detailed analysis of how the coulombian diffusion affect the bunching coefficients in echo enabled fel devices .
our treatment is complementary to the seminal contributions of refs .
@xcite and confirm , within a more general framework , their conclusions .
we believe that the modularity of the methods we have described in this paper offer further opportunities as e. g. that of including in the treatment more complicated arrangements of transport lines and of external fields .
the possibility of embedding these procedures with diffusion and damping in storage rings , will be discussed elsewhere . | coulombian diffusion determines a dilution of bunching coefficients in free electron laser seeded devices . from the mathematical point of view
the effect can be modeled through a heat type equation , which can be merged with the ordinary liouville equation , ruling the evolution of the longitudinal phase space beam distribution .
we will show that the use of analytical tools like the generalized bessel functions and algebraic techniques for the solution of evolution problems may provide a useful method of analysis and shine further light on the physical aspects of the underlying mechanisms . |
standard big - step semantics and hoare logics do not support reasoning about nonterminating runs of programs .
essentially , they ignore them .
but of course nonterminating runs are important .
not only need we often program a partial function whose domain of definedness we can not decide or is undecidable , e.g. , an interpreter , but we also have to program functions that are inherently partial . in programming with interactive input / output , for example , diverging runs are often what we really want . in search for a foundational framework for reasoning about possibly nonterminating programs
constructively ( intuitionistically ) and intrigued by attempts in this direction in the literature , we have previously devised a big - step semantics for while based on traces @xcite . in this
semantics , traces are possibly infinite sequences of states that a program run goes through .
they are defined coinductively , as is the evaluation relation , relating initial states of program runs to traces they produce . on terminating runs ,
this nonstandard semantics agrees with the standard , inductive state - based big - step semantics . in this paper , we put forward a hoare logic to match this big - step semantics . in this new trace - based logic ,
program runs are reasoned about in terms of assertions on states and traces .
more precisely , our hoare triple @xmath0 is given by a statement @xmath1 , a state assertion @xmath2 ( a condition on the initial state of a run of @xmath1 ) and a trace assertion @xmath3 ( a condition on the trace produced by the run ) .
in the presentation we have chosen for this paper , assertions are nothing but predicates expressible in the meta - logic , i.e. , we do not confine ourselves to a particular language of state and trace assertions .
nonetheless , we do not want to downplay the question of what makes a good assertion language for traces .
we are after a set of connectives that allows for a concise formulation of a sound and complete hoare logic over state and trace predicates and logical entailment as given by the constructive meta - logic .
we adopt a solution that is reminiscent of interval temporal logic @xcite , with a chop - connective .
this gives us a set of connectives that is spartan in terms of convenience of expression , but suffices for our meta - theoretical study .
our logic is intended foundational framework into which more specialized and more applied logics with more limited assertion languages can be embedded . besides being deterministic , the while language is also total as soon as we accept that traces of program runs
can be infinite .
this allows our logic to conservatively extend both the standard , state - based partial - correctness hoare logic as well as the state - based total - correctness hoare logic . on the level of derivability alone
this can be proved semantically by going through the soundness and completeness results .
but we go one step further : we show that derivations in these two state - based logics are directly transformable into derivations in our logic , yielding embeddings on the level of derivations , not just mere derivability .
the transformations are relatively straightforward and do not require invention of new invariants or variants , demonstrating that our logic incurs no undue proof burden in comparison to the standard hoare logics . in the converse direction , we can project derivations in our trace - based logic into derivations in the state - based logics : a derivation of a hoare triple in the trace - based logic is translated into a derivation in the state - based logics with a translated , weaker postcondition .
however , the power of our logic goes beyond that of the state - based partial - correctness and total - correctness hoare logics .
the assertions have access to traces . as suggested by the similarity of our ( open ) assertion language to the that of interval temporal logic , this allows us to specify liveness properties of diverging runs .
we will demonstrate this extra expressiveness of our logic by a series of examples . also , interpreted into a constructive underlying logic , our assertion language becomes quite discerning .
in particular we can distinguish between termination and nondivergence , e.g. , unbounded classically total ( constructively nonpartial ) search fails to be terminating , but is nonetheless nondivergent .
we do not discuss this in the paper , but our logic can be adjusted to deal with exceptions and nondeterminism . the paper is organized as follows . in section [ sec : sem ] , we present our trace - based big - step semantics . in section
[ sec : hoare ] , we proceed to the question of a corresponding hoare logic .
we explain our design considerations and then present our hoare logic and the soundness and completeness proofs . in section [ sec : hoare ] , we show the embeddings of the state - based partial - correctness and total - correctness hoare logics into our logic and the projections back . in section
[ sec : examples ] , we consider examples . in section [ sec : related ] , we discuss the related work , to conclude in section [ sec : concl ] .
we have formalized the development fully constructively in coq version 8.1pl3 using the syntax extension library .
the coq development is available at http://cs.ioc.ee/~keiko/code/abyss.tgz . both the paper and the accompanying coq code use coinductive types , corecursion and coinduction extensively . for an introduction
, we can refer the reader to the exposition of bertot and castran ( * ? ? ?
13 ) . in the paper , we have sought to abstract over the more bureaucratic aspects involved in the coq formalization ( e.g. , working with coq s restricted guardedness condition on cofix definitions , i.e. , definitions by corecursion , proofs by coinduction ) .
we start with our big - step semantics .
this is defined in terms of states and traces .
the notion of a state is standard .
a state @xmath4 is an assignment of integer values to the variables .
traces @xmath5 are defined coinductively by the rules @xmath6 so a trace is a non - empty colist ( possibly infinite sequence ) of states .
we also define ( strong ) bisimilarity of two traces , @xmath7 , coinductively by @xmath8 bisimilarity is straightforwardly seen to be an equivalence .
we think of bisimilar traces as equal , i.e. , type - theoretically we treat traces as a setoid with bisimilarity as the equivalence relation .
is bisimilar to either a trace of the form of @xmath9 or one of the form @xmath10 . ]
accordingly , we have to make sure that all functions and predicates we define on traces are setoid functions and predicates ( i.e. , insensitive to bisimilarity ) .
we define the initial state @xmath11 of a trace @xmath12 by case distinction by @xmath13 .
the function @xmath14 is a setoid function .
we also define finiteness of a trace ( with a particular final state ) and infiniteness of a trace inductively resp .
coinductively by @xmath15 finiteness and infiniteness are setoid predicates .
it should be noticed that infiniteness is defined positively , not as negation of finiteness .
constructively , it is not the case that @xmath16 , which amounts to asserting that finiteness is decidable . in particular
, @xmath17 is constructively provable , but @xmath18 is not . the statements of the while
language are given by the following grammar where @xmath19 ranges over ( integer ) variables and @xmath20 over ( arithmetic ) expressions built over variables .
@xmath21 the integer value of an expression @xmath20 in a state @xmath22 is denoted @xmath23 .
we also interpret expressions as booleans ; @xmath24 stands for @xmath20 being true in @xmath22 .
evaluation @xmath25 , expressing that running a statement @xmath1 from a state @xmath22 produces a trace @xmath12 , is defined coinductively by the rules in figure [ fig : sem ] .
the rules for sequence and while implement the necessary sequencing with the help of extended evaluation @xmath26 , also defined coinductively , as the coinductive prefix closure of evaluation : @xmath26 expresses that running a statement @xmath1 from the last state ( if it exists ) of an already accumulated trace @xmath12 results in a total trace @xmath27 .
a remarkable feature of the definition of @xmath26 is that it does not hinge on deciding whether the trace @xmath12 is finite or not , which is constructively impossible .
a proof of @xmath26 simply traverses the already accumulated trace @xmath12 : if the last element is hit , which is the case when @xmath12 is finite , then the statement is run , otherwise the traversal goes on forever .
we look closer at the sequence rule .
we want to conclude that @xmath28 from the premise @xmath29 .
classically , either the run of @xmath30 terminates , i.e. , @xmath31 for some @xmath32 , or it diverges , i.e. , @xmath33 . in the first case , we would like to additionally use that @xmath12 is a finite prefix of @xmath27 and that @xmath34 , where @xmath35 is the rest of @xmath27 .
in the second case , it should be case that @xmath7 . in both cases ,
the desirable condition is equivalent to @xmath36 , which is the second premise of our rule .
the use of extended evaluation , defined as the coinductive ( rather than inductive ) prefix closure of evaluation , allows us to avoid the need to decide whether the run of @xmath30 terminates or not .
@xmath37}}\rangle } } } } } { } \quad \infer= { { ( { \ensuremath{\mathsf{skip}}},{\sigma } ) \rightarrow { \langle{\sigma}\rangle } } } { } \quad \infer= { { ( { \ensuremath{s_0;s_1}},{\sigma } ) \rightarrow { \ensuremath{\tau } } ' } } { { ( s_0,{\sigma } ) \rightarrow { \ensuremath{\tau } } } & { ( s_1,{\ensuremath{\tau } } ) \stackrel{*}{\rightarrow } { \ensuremath{\tau } } ' } } \\[1ex ] \infer= { { ( { \ensuremath{\mathsf{if~ } e \mathsf{~then~ } s_t \mathsf{~else~ } s_f}},{\sigma } ) \rightarrow { \ensuremath{\tau } } } } { { \ensuremath{{\sigma}\models e } } & { ( s_t,{\ensuremath{{\sigma}::{\langle{\sigma}\rangle } } } ) \stackrel{*}{\rightarrow } { \ensuremath{\tau } } } } \quad \infer= { { ( { \ensuremath{\mathsf{if~ } e \mathsf{~then~ } s_t \mathsf{~else~ } s_f}},{\sigma } ) \rightarrow { \ensuremath{\tau } } } } { { \ensuremath{{\sigma}\not\models e } } & { ( s_f,{\ensuremath{{\sigma}::{\langle{\sigma}\rangle } } } ) \stackrel{*}{\rightarrow } { \ensuremath{\tau } } } } \\[1ex ] \infer= { { ( { \ensuremath{\mathsf{while~ } e \mathsf{~do~ } s_t}},{\sigma } ) \rightarrow { \ensuremath{\tau } } ' } } { { \ensuremath{{\sigma}\models e } } & { ( s_t,{\ensuremath{{\sigma}::{\langle{\sigma}\rangle } } } ) \stackrel{*}{\rightarrow } { \ensuremath{\tau } } } & { ( { \ensuremath{\mathsf{while~ } e \mathsf{~do~ } s_t}},{\ensuremath{\tau } } ) \stackrel{*}{\rightarrow } { \ensuremath{\tau } } ' } } \quad \infer= { { ( { \ensuremath{\mathsf{while~ } e \mathsf{~do~ } s_t}},{\sigma } ) \rightarrow { \ensuremath{{\sigma}::{\langle{\sigma}\rangle } } } } } { { \ensuremath{{\sigma}\not\models e } } } \\[2ex ] \infer= { { ( s,{\langle{\sigma}\rangle } ) \stackrel{*}{\rightarrow } { \ensuremath{\tau } } } } { { ( s,{\sigma } ) \rightarrow { \ensuremath{\tau } } } } \quad \infer= { { ( s,{\ensuremath{{\sigma}::{\ensuremath{\tau } } } } ) \stackrel{*}{\rightarrow } { \ensuremath{{\sigma}::{\ensuremath{\tau } } ' } } } } { { ( s,{\ensuremath{\tau } } ) \stackrel{*}{\rightarrow } { \ensuremath{\tau } } ' } } \end{array}\ ] ] [ fig : sem ] evaluation is a setoid predicate . for any @xmath1 , @xmath22 , @xmath12 and @xmath27 , if @xmath25 and @xmath7 , then @xmath38 .
the trace produced from a state begins with this state .
for any @xmath1 , @xmath22 and @xmath12 , if @xmath25 , then @xmath39 .
moreover , for while , evaluation is deterministic ( up to bisimilarity , as is appropriate for our notion of trace equality ) .
[ prop : determ ] for any @xmath1 , @xmath22 , @xmath12 and @xmath27 , if @xmath25 and @xmath38 , then @xmath7 . and
it is also total .
[ prop : total ] for any @xmath1 and @xmath22 , there exists @xmath12 such that @xmath25 . in our definition ,
we have made a choice as regards to what grows the trace of a run .
we have decided that assignments and testing of guards of if- and while - statements augment the trace by a state ( but @xmath40 does not ) , e.g. , we have @xmath41}}\rangle}}}}$ ] , @xmath42 and @xmath43 . this is good for several reasons .
first , @xmath40 becomes the unit of sequential composition , i.e. , the semantics does not distinguish @xmath1 , @xmath44 and @xmath45 .
second , we get a notion of small steps that fully agrees with a very natural coinductive trace - based small - step semantics arising as a straightforward variation of the textbook inductive state - based small - step semantics .
the third and most important outcome is that any while - loop always progresses , because testing of the guard is a small step .
for instance , in our semantics @xmath46 can only derive @xmath43 ( up to bisimilarity ) . as we discuss below , giving up insisting on progress in terms of growing the trace would introduce some semantic anomalies .
it also ensures that evaluation is total as we should expect .
given that it is also deterministic , we can thus equivalently turn our relational big - step semantics into a functional one : the unique trace for a given statement and initial state is definable by corecursion .
( for details , see our previous paper @xcite . )
the coinductive trace - based semantics agrees with the inductive state - based semantics . for any @xmath1 , @xmath22 , @xmath32 , existence of @xmath12 such that @xmath25 and @xmath31 is equivalent to @xmath47 .
we notice that the inductive state - based semantics can not be made total constructively .
it is unproblematic to complement the inductively defined terminating evaluation relation with a coinductively defined diverging evaluation relation , but this does not help , as we can not decide the halting problem .
we look at several seemingly not so different but problematic alternatives that we reject , thereby revealing some subtleties in designing coinductive big - step semantics and motivating our design choices .
since progress of loops is not required for wellformedness of the definitions of @xmath48 and @xmath49 , one might be tempted to regard guard testing to be instantaneous and modify the rules for the while - loop to take the form @xmath50 this leads to undesirable outcomes .
we can derive @xmath51 , which means that the non - terminating @xmath46 is considered semantically equivalent to the terminal ( immediately terminating ) @xmath40 .
worse , we can also derive @xmath52}}\rangle}}}}$ ] , which is even more inadequate : a sequence can continue to run after the non - termination of the first statement .
yet worse , inspecting the rules closer we discover we are also able to derive @xmath53 for any @xmath12 .
mathematically , giving up insisting on progress in terms of growing the trace has also the consequence that the relational semantics can not be turned into a functional one , although while should intuitively be total and deterministic .
in a functional semantics , evaluation must be a trace - valued function and in a constructive setting such a function must be productive .
another option , where assignments and test of guards are properly taken to constitute steps , could be to define @xmath49 by case distinction on the statement by rules such as @xmath54 here , @xmath55 , defined corecursively , traverses @xmath12 and duplicates its last state , if it is finite .
similarly , @xmath56 and @xmath57 traverse @xmath12 and evaluate @xmath20 in the last state , if it is finite : @xmath58 ( the rules for @xmath40 and sequence are very simple and appealing in this design . )
the relation @xmath48 would then be defined uniformly by the rule @xmath59 it turns out that we can still derive @xmath53 for any @xmath12 .
we can even derive @xmath60 for any @xmath12 .
the third alternative ( leroy and grall use this technique in @xcite ) is most close to ours .
it introduces , instead of our @xmath49 relation , an auxiliary relation @xmath61 , defined coinductively by @xmath62 so that @xmath63 expresses that the trace @xmath27 can be split into a concatenation of traces @xmath64 and @xmath65 glued together at a mid - state @xmath32 .
then the evaluation relation is defined by replacing the uses of @xmath49 with @xmath61 , e.g. , the rule for the sequence statement would be : @xmath66 this third alternative does not cause any outright anomalies for while .
but alarmingly @xmath67 has to be run from some ( underdetermined ) state within a run of @xmath68 even if the run of @xmath30 does not terminate . in a richer language with abnormal terminations ,
we get a serious problem : no evaluation is derived for @xmath69 although the @xmath70 statement should not be reached .
we now proceed to the hoare logic and its soundness and completeness proof .
we base our consequence rule on semantic entailment rather than derivability in some fixed proof system .
this allows us to sidestep the problem of its unavoidable incompleteness due to the impossibility of complete axiomatization of any theory containing arithmetic . by identifying assertions with state and trace predicates ( more precisely , predicates expressible in the meta - logic )
, we also avoid the risk of possible incompleteness due to a chosen narrower assertion language not being closed under weakest preconditions / strongest postconditions . to formulate the rules of the hoare logic
, we introduce a small set of assertion connectives , i.e. , operations on predicates . to be able to express the strongest postcondition of any precondition
, we need a few additional connectives .
our assertions are predicates over states and traces .
a state predicate
@xmath2 is any predicate on states ( in particular , it need not be decidable ) . from a trace predicate @xmath3 , we require additionally that it is a setoid predicate , i.e.
, it must be unable to distinguish bisimilar traces .
although we refrain from introducing a language of assertions , we introduce a number of connectives for our assertions , which are operations on predicates .
all trace predicate connectives yield setoid predicates
. the inference rules of the hoare logic make use of these connectives .
indeed , it was an intriguing exercise for us to come up with connectives that would be small but expressive enough for practical specification purposes and at the same time allow us to prove the hoare logic sound and complete in our constructive setting .
the definitions of these connectives are given in figure [ fig : assert ] . to highlight application of a predicate to a state or trace .
we are not defining a single satisfaction relation @xmath71 for some assertion language , but a number of individual state / trace predicates and operations on such predicates .
some of these operations are defined inductively , some coinductively , some definitions are not recursive at all . ]
@xmath72 \infer { { \ensuremath{\tau}}\models { \ensuremath{\mathsf{true } } } } { } \qquad \infer { { \ensuremath{\tau}}\models \neg p } { \neg ( { \ensuremath{\tau}}\models p ) } \qquad \infer { { \ensuremath{\tau}}\models p \wedge q } { { \ensuremath{\tau}}\models p & { \ensuremath{\tau}}\models q } \qquad \ldots \\[2ex ] \infer { { { \langle{\sigma}\rangle } \models { \langleu\rangle } } } { { { \sigma}\models u } } \qquad \infer { { { \ensuremath{\tau } } ' \models { p \mathbin{\ast\ast } q } } } { { { \ensuremath{\tau}}\models p } & { { \ensuremath{\tau } } ' \models_{{\ensuremath{\tau } } } q } } \qquad \infer={{{\ensuremath{\tau}}\models { p^\dagger } } } { { { \ensuremath{\tau}}\models { \langle{\ensuremath{\mathsf{true}}}\rangle } } } \quad \infer={{{\ensuremath{\tau } } ' \models { p^\dagger } } } { { { \ensuremath{\tau}}\models p } & { { \ensuremath{\tau } } ' \models_{{\ensuremath{\tau } } } { p^\dagger } } } \\[2ex ] \infer { { { \ensuremath{{\sigma}::({\ensuremath{{\sigma}[x\mapsto { \ensuremath{\llbracket e \rrbracket { \sigma } } } ] } } ) } } \models { \ensuremath{u[x\mapsto e ] } } } } { { { \sigma}\models u } } \qquad \infer { { { \ensuremath{{\sigma}::{\langle{\sigma}\rangle } } } \models { \langle u \rangle^2 } } } { { { \sigma}\models u } } \\[2ex ] \infer { { { \sigma}\models { \mathit{last } ~p } } } { { { \ensuremath{\tau}}\models p } & { { \ensuremath{\tau}}\mathbin{\downarrow } { \sigma } } } \qquad \infer{{{\ensuremath{\tau}}\models { \mathit{finite } } } } { { { \ensuremath{\tau}}\mathbin{\downarrow } { \sigma } } } \qquad \infer{{{\ensuremath{\tau}}\models { \mathit{infinite } } } } { { { \ensuremath{\tau}}^{\shortuparrow } } } \\[2.5ex ] \infer=[\mathsf{[flw\mbox{-}nil ] } ] { { { \ensuremath{\tau}}\models_{{\langle{\sigma}\rangle } } q } } { { \mathit{hd}}~{\ensuremath{\tau}}= { \sigma}&{{\ensuremath{\tau}}\models q } } \quad \infer=[\mathsf{[flw\mbox{-}delay ] } ] { { { \ensuremath{{\sigma}::{\ensuremath{\tau } } ' } } \models_{{\ensuremath{{\sigma}::{\ensuremath{\tau } } } } } q } } { { { \ensuremath{\tau } } ' \models_{{\ensuremath{\tau } } } q } } \\[2.5ex ] \infer{u \models v } { \forall { \sigma}.\ , { { \sigma}\models u } \to { { \sigma}\models v } } \qquad \infer{p \models q } { \forall { \ensuremath{\tau}}.\ , { { \ensuremath{\tau}}\models p } \to { { \ensuremath{\tau}}\models q } } \end{array}\ ] ] the two most primitive state ( resp . trace ) predicates are @xmath73 and @xmath74 , which are respectively true and false for any state ( resp.trace ) .
we can also use the standard connectives @xmath75 and quantifiers @xmath76 to build state and trace predicates .
the context disambiguates the overloaded notations for these state and trace predicates . for a state predicate @xmath2 ,
the singleton @xmath77 is a trace predicate that is true of singleton traces given by a state satisfying @xmath2 .
in particular @xmath78 is true of any singleton trace . for a state predicate @xmath2 ,
the doubleton @xmath79 is true of a doubleton trace whose two states are identical and satisfy @xmath2 . for a state predicate @xmath2 ,
the update @xmath80}}$ ] is the strongest postcondition of the statement @xmath81 for the precondition @xmath2 .
it is true of a doubleton trace whose first state @xmath22 satisfies @xmath2 and second state is obtained from the first by modifying the value of @xmath19 to become @xmath82 . for trace
predicates @xmath3 and @xmath83 , the chop @xmath84 is a trace predicate that is true , roughly speaking , of a trace @xmath27 that has a prefix @xmath12 satisfying @xmath3 , with the rest of @xmath27 satisfying @xmath83 .
( to be more precise , the prefix and the rest overlap on a mid - state which is the last state of the prefix and the first state of the suffix . ) but its definition is carefully crafted , so that @xmath83 is not checked , if @xmath12 is infinite ( in which case necessarily @xmath85 , and this happens without case distinction on whether @xmath12 is finite .
this effect is achieved with the premise @xmath86 .
the relation @xmath86 is defined coinductively .
it traverses all of @xmath12 , making sure that it is a prefix of @xmath27 ( rule @xmath87 ) , and , upon possible exhaustion of @xmath12 in a finite number of steps , checks @xmath83 against the rest of @xmath27 ( rule @xmath88 ) . this way the problem of deciding whether @xmath12 is finite is avoided , basically by postponing it , possibly infinitely .
our chop operator is classically equivalent to the chop operator from interval temporal logic @xcite ( cf .
also the separating conjunction of separating logic ) .
indeed , classically , @xmath89 holds iff * either , for some finite prefix @xmath12 of @xmath27 , we have @xmath90 and @xmath91 , where @xmath35 is the rest of @xmath27 , * or @xmath27 is infinite and @xmath92 .
this is how the semantics of chop is defined in interval temporal logic .
but it involves an upfront decision of whether @xmath3 will be satisfied by a finite or an infinite prefix of @xmath27 .
our definition is fine - tuned for constructive reasoning . for a trace predicate @xmath3 ,
its iteration @xmath93 is a trace predicate that is true of a trace which is a concatenation of a possibly infinite sequence of traces , each of which satisfies @xmath3 .
( this is modulo the overlap of the last and first states of consecutive traces in the sequence and the empty concatenation being a singleton trace . )
it is reminiscent of the kleene star operator .
it is defined by coinduction and takes into account possibilities of both infiniteness of some single iteration and infinite repetition .
for a trace predicate @xmath3 , @xmath94 is a state predicate that is true of states that can be the last state of a finite trace satisfying @xmath3 .
trace predicates @xmath95 and @xmath96 are true of finite and infinite traces , respectively . for state predicates @xmath2 and @xmath97 ( resp .
trace predicates @xmath3 and @xmath83 ) , @xmath98 ( resp .
@xmath99 ) denotes entailment .
[ prop : asserts_setoid ] for any @xmath2 , @xmath77 , @xmath80}}$ ] and @xmath79 are setoid predicates .
for any setoid predicates @xmath3 , @xmath83 , @xmath84 is a setoid predicate . for any setoid predicate @xmath3 ,
@xmath93 is a setoid predicates .
moreover , @xmath95 and @xmath96 are setoid predicates . that @xmath77 , @xmath80}}$ ] and @xmath79 are setoid predicates follows from the definition .
we prove that @xmath84 and @xmath93 are setoid predicates when @xmath3 and @xmath83 are by coinduction . that @xmath95 and @xmath96 are setoid predicates is proved by induction and by coinduction respectively .
[ prop : asserts_monotone ] for any @xmath2 and @xmath97 , if @xmath98 , then @xmath100 , @xmath80 } } \models { \ensuremath{v[x\mapsto e]}}$ ] and @xmath101 .
for any setoid predicates @xmath102 and
@xmath83 , if @xmath103 , then @xmath104 and @xmath105 . for any setoid predicates @xmath3 and @xmath83 , if @xmath99 , then @xmath106 and @xmath107 . that @xmath77 , @xmath80}}$ ] , @xmath79 are monotone follows from the definition
we prove that @xmath84 and @xmath93 are monotone and that @xmath94 is monotone by coinduction and by induction respectively .
a number of logical consequences and equivalences hold about these connectives , to be proved in lemma [ lemma : misc_asserts ] .
we have the trivial equivalence : @xmath108 .
the chop operator is associative : @xmath109 .
the iterator operator @xmath93 repeats @xmath3 either zero times or once followed by further repetitions : @xmath110 .
a trace is infinite if and only if @xmath74 holds for any last state : @xmath111 . if every trace satisfying @xmath3 is infinite , i.e. , if @xmath112 , then any trace satisfying @xmath3 has no last state , i.e. , @xmath113 .
we have @xmath114 , so that if a trace satisfies @xmath3 , then its last state , if exists , satisfies @xmath94 .
the last state of a singleton trace @xmath9 is @xmath22 , therefore we have @xmath115 .
we also have @xmath116 , but the converse does not hold . e.g. , @xmath117 . instead , we have @xmath118
. finally we have @xmath119 .
namely , a state satisfies @xmath2 and can be the last state of a finite trace satisfying @xmath3 if and only if it can be the last state of a finite trace satisfying @xmath120 .
we define the concatenation of traces @xmath12 and @xmath27 , @xmath121 , by replacing the last state of @xmath12 by @xmath27 .
formally , it is defined by corecursion by @xmath122 we first observe three results , which are useful for later proofs .
[ lemma : follows_true ] for any @xmath12 , @xmath123 . by coinduction and case analysis on @xmath12 .
the case of @xmath124 follows from @xmath125 .
the case of @xmath126 : we get @xmath127 from the coinduction hypothesis , from which we obtain @xmath123 .
[ lemma : follows_singleton ] for any @xmath2 and @xmath128 , if @xmath129 , then @xmath7 . by coinduction and inversion on @xmath129 .
[ lemma : conv_last ] for any @xmath2 and @xmath12 , if for any @xmath22 , @xmath130 implies @xmath131 , then @xmath132 . by coinduction with case analysis on @xmath12 . the case of @xmath133 :
we have @xmath130 , hence @xmath134 by our hypothesis .
we conclude @xmath132 .
the case of @xmath135 : since @xmath136 implies @xmath130 , we have that , for any @xmath22 , @xmath136 implies @xmath131 .
we get @xmath137 by the coinduction hypothesis , from where we conclude @xmath132 .
[ lemma : misc_asserts ] for any @xmath138 and setoid predicates @xmath139 and @xmath140 , we have 1 . [ lemma : sglt_dup ] @xmath141 2 . [ lemma : sglt_chop]@xmath142 3 . [ lemma : sglt_true_chop]@xmath108 4 . [ lemma : chop_assoc ] @xmath109 5 . [ lemma : rep_unfold ] @xmath110 6 .
[ lemma : rep_idem]@xmath143 7 .
@xmath111 8 .
if @xmath144 , then @xmath113 . 9 .
@xmath145 1 .
[ lemma : last_sglt]@xmath115 2 .
[ lemma : last_chop]@xmath116 3 .
[ lemma : last_last]@xmath118 4 .
[ lemma : last_chop_sglt ] @xmath146 5 .
[ lemma : last_rep ] @xmath147 1
. follows from the definition .
2 . follows from the definition .
@xmath148 follows from the definition .
@xmath149 holds by lemma [ lemma : follows_true ] .
suppose @xmath150 .
there exists @xmath12 such that @xmath90 and @xmath151 . by lemma [ lemma : follows_singleton ] @xmath7 holds , so we must have @xmath92 since @xmath3 is a setoid predicate .
this proves @xmath152 .
4 . suppose @xmath153 .
there exist @xmath12 and @xmath27 such that @xmath90 and @xmath86 and @xmath154 .
we prove , for any @xmath155 and @xmath156 , @xmath157 and @xmath158 imply @xmath159 by coinduction and inversion on @xmath157 . this yields @xmath160 therefore @xmath161 .
+ the converse is more subtle . given @xmath161 , we have to find a prefix @xmath27 of @xmath35 that satisfies @xmath84 while @xmath154 . to do so
, we define a function @xmath162 by corecursion ( we take @xmath64 and @xmath65 to be implicit parameters of @xmath163 , inferred from the proof argument ) . below ; @xmath164 is for an anonymous dummy argument .
@xmath165 is the constructor of sigma - types in coq . ]
@xmath166 { \mathit{midp}}~(\mathsf{flw\mbox{-}delay } ~{\sigma}~{\ensuremath{\tau}}_0~{\ensuremath{\tau}}_1~(h:{{\ensuremath{\tau}}_1 \models_{{\ensuremath{\tau}}_0 } p_0{\mathbin{\ast\ast}}q_0 } ) ) = { \ensuremath{{\sigma}::{\mathit{midp}}~h } } \end{array}\ ] ] we then prove that , for any @xmath155 and @xmath167 , @xmath168 and @xmath169 hold by coinduction and inversion on @xmath170 . + now assume @xmath161 .
there exists @xmath12 such that @xmath90 and @xmath171 .
we have @xmath172 , since @xmath173 .
this together with @xmath174 proves @xmath175 , as required .
5 . follows from the definition .
suppose @xmath176 .
we have to prove @xmath177 . from lemma [ lemma : follows_true ] and ( 5 ) , we deduce @xmath178 , which gives us @xmath177 .
conversely , suppose @xmath177 .
there exists @xmath27 such that @xmath179 and @xmath180 .
we close the case by proving the following two conditions by mutual coinduction : 1 .
@xmath181 2 .
+ ( a ) : we perform inversion on @xmath176 . the case of @xmath183 , i.e. , @xmath124 : from @xmath184 , we conclude @xmath179 .
the case of @xmath185 and @xmath186 : we get @xmath187 by ( b ) , from where we conclude @xmath179 .
+ ( b ) : we perform inversion on @xmath184 . the case of @xmath124 and @xmath188 and @xmath179 follows from ( a ) .
the case of @xmath189 and @xmath190 and @xmath191 : we must have @xmath192 and @xmath193 . the coinduction hypothesis ( b ) gives us @xmath194 , from which we conclude @xmath186
we prove an auxiliary condition : for any @xmath12 , @xmath195 iff @xmath196 by coinduction .
@xmath197 follows from the condition .
@xmath198 follows from the condition and lemma [ lemma : follows_singleton ] .
. follows from @xmath199 .
suppose @xmath90 . by the definition of @xmath94
, we have for any @xmath22 , @xmath130 implies @xmath200 .
we then deduce @xmath201 by lemma [ lemma : conv_last ] , thus conclude @xmath202 .
+ conversely , suppose that @xmath203 , i.e. , @xmath90 and @xmath204 for some @xmath12 .
by lemma [ lemma : follows_singleton ] , @xmath7 holds , so we must have @xmath92 since @xmath3 is a setoid predicate . 1
. follows from the definition .
2 . suppose @xmath205 .
there exist @xmath12 and @xmath27 such that @xmath136 and @xmath90 and @xmath86 .
we have to prove @xmath206 .
we do so by proving an auxiliary condition : for any @xmath207 and @xmath64 , if @xmath208 , then for any @xmath65 , @xmath209 implies @xmath210 by induction on the derivation of @xmath208 .
3 . suppose @xmath211 .
there exist @xmath12 and @xmath27 such that @xmath90 , @xmath212 , @xmath213 and @xmath136 .
we then have that the concatenation of @xmath12 and @xmath27 has the desired properties .
namely , @xmath214 and @xmath215 .
this proves @xmath205 as we wanted .
+ conversely , suppose @xmath205 .
there exist @xmath12 and @xmath27 such that @xmath90 , @xmath86 and @xmath136 .
the finiteness of @xmath27 implies that of @xmath12 , i.e. , we have @xmath31 for some @xmath32 .
we can therefore find the suffix @xmath35 of @xmath27 such that @xmath216 and @xmath217 .
( basically , we drop the first @xmath218 elements from @xmath27 to obtain @xmath35 , where @xmath218 is the length of @xmath12 . since @xmath12 is finite ,
its length is defined . ) together @xmath86 and @xmath31 proves @xmath91 .
this concludes @xmath211 , as we wanted .
4 . by the monotonicity of the last and chop operators , it suffices to prove @xmath219 .
suppose @xmath220 .
there exists @xmath12 such that @xmath130 and @xmath221 .
we now prove @xmath222 by induction on the proof of @xmath130 , from where @xmath134 follows .
5 . by ( 13 ) and the monotonicity of the chop and iterator operators , it suffices to prove @xmath223 .
suppose @xmath224 .
there exists @xmath27 such that @xmath225 and @xmath226 , which give us @xmath227 and @xmath228 .
we want to prove @xmath132 .
we do so by proving the following conditions by mutual coinduction
@xmath229 2 .
+ ( a ) : we perform inversion on @xmath227 . the case of @xmath183 : we have @xmath124 .
this and @xmath228 prove @xmath132 .
the case of @xmath231 and @xmath226 : the former gives us @xmath232 for some @xmath35 .
we therefore conclude @xmath132 by ( b ) .
+ ( b ) : we perform inversion on @xmath233 : the case of @xmath124 and @xmath188 and @xmath234 : we have @xmath235 , therefore @xmath192 for some @xmath236 with @xmath237 , and @xmath238 . from ( a ) , we obtain @xmath239 , which yields @xmath240 , as required . the case of @xmath190 and @xmath189 and
@xmath241 : we have @xmath192 and @xmath242 . by ( b ) , we get @xmath239 , which yields @xmath240 .
the derivable judgements of the hoare logic are given by the inductively interpreted inference rules in figure [ fig : hoare ] .
the proposition @xmath0 states derivability of the judgement .
the intent is that @xmath0 should be derivable precisely when running a statement @xmath1 from an initial state satisfying @xmath2 is guaranteed to produce a trace satisfying @xmath3 .
the rules for assignment and @xmath40 are self - explanatory .
the rule for sequence is defined in terms of the chop operator .
the precondition @xmath97 for the second statement @xmath67 is given by those states in which a run of the first statement @xmath30 may terminate . in particular ,
if @xmath243 and @xmath144 , i.e. , @xmath30 is necessarily diverging for the precondition @xmath2 , then we have @xmath244 . in this case , from the derivability of @xmath245 for any @xmath83 , we get @xmath246 for any @xmath83 . but
this makes sense , since @xmath247 as soon as @xmath144 .
the rule for if - statement uses the doubleton operator in accordance with the operational semantics where we have chosen that testing the boolean guard grows the trace .
the rule for while - statement is inspired by the corresponding rule of the standard , state - based partial - correctness hoare logic .
it uses a loop invariant @xmath248 .
this is a state predicate that has to be true each time the boolean guard is about to be ( re-)tested in a run of the loop . accordingly , the precondition @xmath2 should be stronger then @xmath248 . also , @xmath248 must hold each time an iteration of @xmath249 has finished , as enforced by having @xmath250 as the postcondition of @xmath249 .
the postcondition @xmath251 of the loop consists of three parts .
@xmath79 accounts for the first test of the guard ; @xmath252 accounts for iterations of the loop body in alternation with re - tests of the guard ( notice that that we are again using the doubleton operator ) ; @xmath253 accounts for the state in which the last test of the guard is finished .
we have chosen to introduce a separate rule for instantiating auxiliary variables .
alternatively , we might have stated the consequence rule in a more general form , as suggested by kleymann @xcite ; yet the separation facilitates formalization in coq .
the various logical consequences and equivalences about the connectives suggest also further alternative and equivalent formulations .
for instance , we could replace the rule for the while - statement by @xmath254 if we strengthened the consequence rule to @xmath255 with our chosen rule for while , this strengthened version of consequence is admissible : [ lemma : semax_hd ] for any @xmath2 , @xmath1 and @xmath3 , if @xmath0 , then @xmath256 .
we prove the following more general statement by induction on the derivation of @xmath0 : for any @xmath2 , @xmath1 and @xmath3 , if @xmath0 , then for any @xmath97 , @xmath257 .
we do not attempt to argue that our formulation is the best choice ; yet we found that the present formulation is viable from the points - of - view of both the meta - theory and applicability of the logic .
@xmath258}}\ } } } { } \quad \infer { { \{u\}~{\ensuremath{\mathsf{skip}}}~\{{\langleu\rangle}\ } } } { } \quad \infer { { \{u\}~{\ensuremath{s_0;s_1}}~\{{p \mathbin{\ast\ast } q}\ } } } { { \{u\}~s_0~\{{p \mathbin{\ast\ast } { \langlev\rangle}}\ } } & { \{v\}~s_1~\{q\ } } } \\[2ex ] \infer { { \{u\}~{\ensuremath{\mathsf{if~ } e \mathsf{~then~ } s_t \mathsf{~else~ } s_f}}~\{{{\langle u \rangle^2 } \mathbin{\ast\ast } p}\ } } } { { \{{e \wedge u}\}~s_t~\{p\ } } & { \{{\neg e \wedge u}\}~s_f~\{p\ } } } \\[2ex ] \infer { { \{u\}~{\ensuremath{\mathsf{while~ } e \mathsf{~do~ } s_t}}~\ { { \langle u \rangle^2 } { \mathbin{\ast\ast } } { ( { } p { \mathbin{\ast\ast}}{\langle i \rangle^2})^\dagger } { \mathbin{\ast\ast}}{\langle { } \neg e\rangle}\ } } } { u \models i & { \{{e \wedge i}\}~s_t~\{{p \mathbin{\ast\ast } { \langlei\rangle}}\ } } } \\[2ex ] \infer { { \{u\}~s~\{p\ } } } { u \models u ' & { \{u'\}~s~\{p'\ } } & p ' \models p } \quad \infer { { \{\exists z.\ , u\}~s~\{\exists z.\ , p\ } } } { \forall z.\ , { \{u\}~s~\{p\ } } } \end{array}\ ] ] the soundness result states that any derivable hoare triple is semantically valid in the sense that , if the precondition holds of the initial state of an evaluation , then the postcondition is true of the trace produced . [
prop : sound ] for any @xmath1 , @xmath2 and @xmath3 , if @xmath0 , then , for all @xmath22 and @xmath12 , @xmath131 and @xmath25 imply @xmath259 . by induction on the derivation of @xmath0 .
we show the main cases of sequence and while .
* @xmath260 : we are given as the induction hypothesis that , for any @xmath261 , @xmath29 and @xmath131 imply @xmath262 , and that , for any @xmath261 , @xmath263 and @xmath264 imply @xmath265 .
we have to prove @xmath266 , given @xmath131 and @xmath267 and @xmath268 . the induction hypothesis for @xmath30
gives us @xmath269 .
by lemma [ lemma : follows_singleton ] and that @xmath3 is a setoid predicate , we derive @xmath270 and @xmath271 .
we prove by coinduction an auxiliary lemma : for any @xmath12 , @xmath27 , @xmath272 and @xmath36 give @xmath86 , using the induction hypothesis for @xmath67 .
the lemma gives us @xmath273 .
we can now close the case by @xmath274 and @xmath275 .
* @xmath276 : we are given as the induction hypothesis that for any @xmath22 and @xmath12 , @xmath277 and @xmath278 imply @xmath279 .
we also have @xmath280 .
we have to prove @xmath281 , given @xmath131 and @xmath282 .
we prove that , for any @xmath22 and @xmath12 , @xmath282 implies @xmath283 by coinduction .
it remains to prove @xmath284 . by inversion on @xmath282
, we learn that @xmath135 for some @xmath32 and @xmath27 such that @xmath285 .
so , we close the case by proving the following conditions by mutual coinduction : * * for any @xmath22 and @xmath12 , if @xmath286 and @xmath287 , then @xmath288 * * for any @xmath12 and @xmath27 , if @xmath289 and @xmath290 , then @xmath291 .
thanks to proposition [ prop : total ] ( totality of evaluation ) , as an immediate corollary of proposition [ prop : sound ] ( soundness ) we learn that , if a state satisfies the precondition of a derivable hoare triple , then there exists an evaluation producing a trace satisfying the postcondition .
[ coro : semax_satisfiable ] for any @xmath1 , @xmath2 and @xmath3 , if @xmath0 , then , for any @xmath22 such that @xmath134 , there exists @xmath12 such that @xmath25 and @xmath90 .
the completeness result states that any semantically valid hoare triple is derivable . following the standard approach ( see , e.g. , @xcite ) we define , for a given statement @xmath1 and a given precondition @xmath2 , a trace predicate @xmath292the candidate strongest postcondition .
then we prove that @xmath292 is a postcondition according to the logic ( i.e. , @xmath293 is derivable ) and that @xmath292 is semantically stronger than any other trace predicate that is a postcondition semantically .
completeness follows .
the trace predicate @xmath292 is defined by recursion on @xmath1 in figure [ fig : sp ] .
the definition is mostly self - explanatory , as it mimics the inference rules of the logic , except that we need the loop - invariant @xmath294 .
@xmath294 characterizes the set of states that are reachable by some run of @xmath295 from a state satisfying @xmath2 and where the boolean guard is tested in that run . and
@xmath292 , we have used impredicativity in our coq development . ] @xmath296}}\\ { \mathit{sp}({\ensuremath{\mathsf{skip } } } , u ) } = { \langleu\rangle}\\ { \mathit{sp}({\ensuremath{s_0;s_1 } } , u ) } = p { \mathbin{\ast\ast}}{\mathit{sp}(s_1 , { \mathit{last } ~p } ) } \mathit{~where~ } p = { \mathit{sp}(s_0 , u ) } \\ { \mathit{sp}({\ensuremath{\mathsf{if~ } e \mathsf{~then~ } s_t \mathsf{~else~ } s_f } } , u ) } = { \langle u \rangle^2 } { \mathbin{\ast\ast}}({{\mathit{sp}(s_t , { e \wedge u } ) } \vee { \mathit{sp}(s_f , { \neg e \wedge u } ) } } ) \\ { \mathit{sp}({\ensuremath{\mathsf{while~ } e
\mathsf{~do~ } s_t } } , u ) } = { \langle u \rangle^2 } { \mathbin{\ast\ast}}{({\mathit{sp}(s_t , { e \wedge i } ) } { \mathbin{\ast\ast}}{\langle i \rangle^2})^\dagger } { \mathbin{\ast\ast}}{\langle\neg e\rangle}\\ \hspace*{2 cm } \mathit{~where~ } i = { \mathit{inv}(e , s_t , u ) } \\ \\[2ex ] \infer { { { \sigma}\models { \mathit{inv}(e , s , u ) } } } { { { \sigma}\models u } } \quad \infer { { { \sigma}\models { \mathit{inv}(e , s , u ) } } } { v \models { \mathit{inv}(e , s , u ) } & { { \sigma}\models { \mathit{last } ~({\langle{\mathit{inv}(e , s , u ) } \wedge e\rangle } { \mathbin{\ast\ast}}{\mathit{sp}(s , v ) } ) } } } \end{array}\ ] ] for any @xmath1 and @xmath2 , the predicate @xmath292 is a monotone setoid predicate .
[ lemma : sp_setoid ] for any @xmath1 , @xmath2 , @xmath12 , @xmath27 , if @xmath297 and @xmath7 , then @xmath298 . by induction on the structure of @xmath1 .
[ lemma : sp_cont ] for any @xmath1 , @xmath2 , @xmath299 , if @xmath300 , then @xmath301 . by induction on the structure of @xmath1 .
the following lemma states that any trace which satisfies @xmath292 has its first state satisfying @xmath2 .
[ lemma : sp_hd ] for any @xmath1 , @xmath2 , @xmath12 , if @xmath297 , then @xmath302 . by induction on the structure of @xmath1 .
the next lemma states a crucial property of @xmath294 .
[ lemma : inv_correct ] for any @xmath1 , @xmath20 , @xmath2 , @xmath303 .
( @xmath304 ) : suppose @xmath305 .
it suffices to prove @xmath306 .
we have @xmath307 by lemma [ lemma : sp_hd ] , and @xmath308 by lemma [ lemma : sp_cont ] and @xmath309 .
these give us @xmath310 . by the definition of _ inv
_ , we have for any @xmath22 , @xmath130 implies @xmath311 .
therefore we conclude @xmath306 by lemma [ lemma : conv_last ] .
( @xmath312 ) : suppose @xmath313 .
we then have some @xmath27 such that @xmath314 and @xmath315 .
the latter proves @xmath7 by lemma [ lemma : follows_singleton ] .
we conclude @xmath305 by lemma [ lemma : sp_setoid ] .
we are now ready to establish that @xmath292 is a postcondition according to the hoare logic .
[ lemma : a ] for any @xmath1 , @xmath2 , we have @xmath293 . by induction on @xmath1 .
we show the main cases of sequence and while .
* @xmath260 : we are given as the induction hypotheses that , for any @xmath316 , @xmath317 and @xmath318 .
we have to prove @xmath319 where @xmath320 . by the induction hypothesis , we have @xmath243 , thus @xmath321 by ( 9 ) of lemma [ lemma : misc_asserts ] and the consequence rule .
we therefore close the case with @xmath322 given by the induction hypothesis . *
@xmath276 : we are given as the induction hypothesis that @xmath323 , for any @xmath316 .
we have to prove @xmath324 where @xmath325 .
it is sufficient to prove @xmath326 , which follows from the induction hypothesis and lemma [ lemma : inv_correct ] .
following the standard route , it remains to prove the following condition : for any @xmath1 , @xmath2 , @xmath3 , if for all @xmath261 , @xmath131 and @xmath25 imply @xmath259 , then @xmath327 .
this will be an immediate corollary from lemma [ lemma : sp_hd ] and the following lemma , stating that any trace satisfying @xmath292 is in fact produced by a run of @xmath1 .
[ lemma : b ] for any @xmath1 , @xmath2 , @xmath12 , if @xmath297 then @xmath328 . by induction on @xmath1 .
we show the main cases of sequence and while .
* @xmath260 : we are given as the induction hypotheses that , for any @xmath329 , @xmath330 ( resp .
@xmath331 ) implies @xmath332 ( resp .
@xmath333 ) . we have to prove @xmath334 , given @xmath335 , which unfolds into @xmath336 and @xmath337 . by the induction hypothesis for @xmath30
, we have @xmath338 . using the induction hypothesis for @xmath67 , we prove by coinduction that , for any @xmath339 , @xmath340 implies @xmath341 , thereby we close the case .
* @xmath276 : we are given as the induction hypothesis that , for any @xmath329 , @xmath342 implies @xmath343 .
we have to prove @xmath344 , given @xmath345 where @xmath325 .
we do so by proving the following two conditions simultaneously by mutual coinduction : * * for any @xmath12 , @xmath346 implies @xmath347 , * * for any @xmath12 and @xmath27 , @xmath348 implies @xmath290 .
[ cor ] for any @xmath1 , @xmath2 and @xmath3 , if for all @xmath22 and @xmath12 , @xmath131 and @xmath25 imply @xmath259 , then @xmath327 .
completeness is proved as a corollary of lemma [ lemma : a ] and corollary [ cor ] .
[ prop : complete ] for any @xmath1 , @xmath2 and @xmath3 , if for all @xmath22 and @xmath12 , @xmath131 and @xmath25 imply @xmath259 , then @xmath0 .
assume that for all @xmath261 , @xmath131 and @xmath25 imply @xmath259 .
by corollary [ cor ] , we have that @xmath327 . by lemma
[ lemma : a ] , we have @xmath293 .
applying consequence , we get @xmath0 . combining propositions [ prop : determ ] ( determinacy of evaluation ) and [ prop : complete ] ( completeness ) , we immediately get completeness for total correctness . for any @xmath1 , @xmath2 and @xmath3 , if , for all @xmath22 such that @xmath131 , there is @xmath12 such that @xmath25 and @xmath259 , then @xmath0 .
it is easy to see , by going through the soundness and completeness results , that our trace - based hoare logic is a conservative extension of the standard , state - based partial - correctness and total - correctness hoare logics .
but more can be said .
the derivations in these two logics are directly transformable into derivations in our logic , preserving their structure , without invention of new invariants or variants . and in the converse direction , derivations in our logic are transformable into derivations into the standard logics in a way that removes from postconditions information about intermediate states .
in this direction , the variant for a while - loop is obtained by bounding the length of traces satisfying the trace invariant of the loop .
concerning total correctness , we use two variations of the while - rule . in the forward transformation ,
we use a version of the while - rule with a dedicated variant ( a natural - valued _ function _ on states ) whereas , in the backward transformation , we work with a version where the invariant ( a state predicate ) is made dependent on a natural number ( i.e. , becomes a _ relation _ between states and naturals ; crucially , there is no functionality requirement : in the same state , the invariant can be satisfied by zero or one or several naturals ) . the two alternative while - rules for total correctness are : @xmath349 { { \{i \wedge t = m\}~{\ensuremath{\mathsf{while~ } e \mathsf{~do~ } s_t}}~\{i \wedge t \leq m \wedge \neg e\ } } } { \forall n:\mathit{nat}. ~{\{e \wedge
i \wedge t = n\}~s_t~\{i \wedge t < n\ } } } \ ] ] and @xmath350 { { \{j\ , m\}~{\ensuremath{\mathsf{while~ } e \mathsf{~do~ } s_t}}~\{\exists k.\ , k \leq m \wedge j\ , k \wedge \neg e\ } } } { \forall n:\mathit{nat}. ~{\{e \wedge j\ , n\}~s_t~\{\exists k.\ , k < n \wedge j\ , k\ } } } \ ] ] there is a reason for this discrepancy , which reflects our compromise between pursuing a constructive approach and striving for purely syntactic translations .
we will discuss it in section [ sec : hoare : discussion ] after having presented the transformations .
we have tried to fine - tune the inference rules in the different hoare logics and the transformations between them for smoothness .
there is some room for variations in them .
the transformations are quite sensitive to the exact division of labor in the source and target hoare logics between the rules for the statement constructors and the consequence rule , but the effects of the possible variations are mostly inessential . for reference , the inference rules of the state - based logics appear in the appendix .
notice that also here we use predicates as assertions and entailment as consequence , so there is no dedicated assertion language or proof system for assertions .
we formalize our claim of embeddability of the standard hoare logics in the following two propositions , whose direct proofs are algorithms for the transformations .
proposition [ prop : partial_hoare ] states that , if @xmath351 is a derivable partial - correctness judgement , then @xmath352 is derivable in our logic .
the trace predicate @xmath353 indicates that @xmath354 holds of any state that is reachable by traversing , in a finite number of steps , the whole trace @xmath12 produced by running @xmath1 .
classically , this amounts to @xmath354 being true of the last state of @xmath12 , if @xmath12 is finite and hence has one ; if @xmath12 is infinite , then nothing is required .
proposition [ lemma : total_hoare ] states that , if @xmath351 is a derivable total - correctness judgement , then @xmath355 is derivable in our logic ( in fact , it states a little more ) .
the trace predicate @xmath356 expresses that the trace @xmath12 produced by running @xmath1 is finite and @xmath354 holds of the last state of @xmath12 ; the finiteness of @xmath12 guarantees the existence of this last state .
[ prop : partial_hoare ] for any @xmath1 , @xmath2 and @xmath354 , if @xmath351 is derivable in the partial - correctness hoare logic , then @xmath352 is derivable in the trace - based hoare logic . by induction on the derivation of @xmath351 .
we show the main cases of sequence and while .
* @xmath260 : we are given as the induction hypotheses @xmath357 and @xmath358 .
we have to prove @xmath359 , which is derived by @xmath360 { { \{u\}~{\ensuremath{s_0;s_1}}~\{{\ensuremath{\mathsf{true}}}{\mathbin{\ast\ast}}{\langlez\rangle}\ } } } { \infer { { \{u\}~{\ensuremath{s_0;s_1}}~\{{\ensuremath{\mathsf{true}}}{\mathbin{\ast\ast}}{\ensuremath{\mathsf{true}}}{\mathbin{\ast\ast}}{\langlez\rangle}\ } } } { \infer*[{\mathrm{ih}}_0]{{\{u\}~s_0~\{{\ensuremath{\mathsf{true}}}{\mathbin{\ast\ast}}{\langlev\rangle}\ } } } { } & \infer*[{\mathrm{ih}}_1]{{\{v\}~s_1~\{{\ensuremath{\mathsf{true}}}{\mathbin{\ast\ast}}{\langlez\rangle}\ } } } { } } } \ ] ] 1 .
we have @xmath361 * @xmath276 : we are given as the induction hypothesis @xmath362 .
we have to prove @xmath363 , which is derived by @xmath360 { { \{i\}~{\ensuremath{\mathsf{while~ } e \mathsf{~do~ } s_t}}~\{{\ensuremath{\mathsf{true}}}{\mathbin{\ast\ast}}{\langlei \wedge \neg e\rangle}\ } } } { \infer { { \{i\}~{\ensuremath{\mathsf{while~ } e \mathsf{~do~ } s_t}}~\{{\langle i \rangle^2 } { \mathbin{\ast\ast } } { ( { } { \ensuremath{\mathsf{true}}}{\mathbin{\ast\ast}}{\langle i \rangle^2})^\dagger } { \mathbin{\ast\ast}}{\langle { } \neg e\rangle}\ } } } { \infer*[{\mathrm{ih}}_t]{{\{e \wedge i\}~s_t~\{{\ensuremath{\mathsf{true}}}{\mathbin{\ast\ast}}{\langlei\rangle}\ } } } { } } } \ ] ] 1 .
we have @xmath364 for the embedding of total - correctness derivations , we prove a slightly stronger statement to have the induction go through .
[ lemma : total_hoare ] for any @xmath1 , @xmath2 and @xmath354 , if @xmath351 is derivable in the total - correctness hoare logic with @xmath365 , then for any @xmath366 , @xmath367 is derivable in the trace - based hoare logic . by induction on the derivation of @xmath351 .
we show the main cases of sequence and while .
* @xmath260 : we are given as the induction hypotheses that , for any @xmath368 , @xmath369 and , for any @xmath370 , @xmath371 .
we have to prove that , for any @xmath366 , @xmath372 .
this is done by the derivation @xmath360 { { \{u \wedge w\}~{\ensuremath{s_0;s_1}}~\{{\langlew\rangle } { \mathbin{\ast\ast}}{\mathit{finite}}{\mathbin{\ast\ast}}{\langlez\rangle}\ } } } { \infer{{\{u \wedge w\}~{\ensuremath{s_0;s_1}}~\{{\langlew\rangle } { \mathbin{\ast\ast}}{\mathit{finite}}{\mathbin{\ast\ast}}{\langlev\rangle } { \mathbin{\ast\ast}}{\mathit{finite}}{\mathbin{\ast\ast}}{\langlez\rangle}\ } } } { \infer*[{\mathrm{ih}}_0\ , w]{{\{u \wedge
w\}~s_0~\{{\langlew\rangle } { \mathbin{\ast\ast}}{\mathit{finite}}{\mathbin{\ast\ast}}{\langlev\rangle}\ } } } { } & \infer{{\{v\}~s_1~\{{\langlev\rangle } { \mathbin{\ast\ast}}{\mathit{finite}}{\mathbin{\ast\ast}}{\langlez\rangle}\ } } } { \infer*[{\mathrm{ih}}_1\ , v]{{\{v \wedge v\}~s_1~\{{\langlev\rangle } { \mathbin{\ast\ast}}{\mathit{finite}}{\mathbin{\ast\ast}}{\langlez\rangle}\ } } } { } } } } \ ] ] 1 .
we have @xmath373 * @xmath276 : we are given as the induction hypothesis that , for all @xmath374 and @xmath375 , @xmath376 .
we have to prove that , for any @xmath366 , @xmath377 .
this is accomplished by the derivation @xmath360 { { \{i \wedge t = m \wedge w\}~{\ensuremath{\mathsf{while~ } e \mathsf{~do~ } s_t}}~\{{\langlew\rangle } { \mathbin{\ast\ast}}{\mathit{finite}}{\mathbin{\ast\ast}}{\langlei \wedge \neg e \wedge t \leq m\rangle}\ } } } { \infer { { \begin{array}{l}\{i \wedge t = m \wedge w\ } ~{\ensuremath{\mathsf{while~ } e \mathsf{~do~ } s_t } } \\
\quad \{{\langle i \wedge t = m \wedge w \rangle^2 } { \mathbin{\ast\ast}}{((\exists n. { \langlet = n\rangle } { \mathbin{\ast\ast}}{\mathit{finite}}{\mathbin{\ast\ast}}{\langlet < n\rangle } ) { \mathbin{\ast\ast}}{\langle i \wedge t \leq m \rangle^2})^\dagger } { \mathbin{\ast\ast}}{\langle\neg e\rangle}\ } \end{array } } } { \infer { { \{e \wedge
i \wedge t \leq m\}~s_t~\{(\exists n. { \langlet = n\rangle } { \mathbin{\ast\ast}}{\mathit{finite}}{\mathbin{\ast\ast}}{\langlet < n\rangle } ) { \mathbin{\ast\ast}}{\langlei \wedge t \leq m\rangle}\ } } } { \infer { { \{\exists n.\ , e \wedge i \wedge t = n \wedge t \leq m\}~s_t~\{\exists n.\ , { \langlet = n\rangle } { \mathbin{\ast\ast}}{\mathit{finite}}{\mathbin{\ast\ast}}{\langlei \wedge t <
n \wedge t \leq m\rangle}\ } } } { \infer[(2 ) ] { \forall n.\ , { \{e \wedge
i \wedge t = n \wedge t \leq m\}~s_t~\{{\langlet = n\rangle } { \mathbin{\ast\ast}}{\mathit{finite}}{\mathbin{\ast\ast}}{\langlei \wedge
t < n \wedge t \leq m\rangle}\ } } } { \infer*[\forall n.\ , { \mathrm{ih}}_t\ , n\ , ( t = n \wedge t \leq m)]{\forall n.\ , { \{e \wedge i \wedge t = n \wedge t = n \wedge t \leq m\}~s_t~\{{\langlet =
n \wedge t \leq m\rangle } { \mathbin{\ast\ast}}{\mathit{finite}}{\mathbin{\ast\ast}}{\langlei \wedge t < n \rangle}\ } } } { } } } } } } \ ] ] 1 .
the repetition gives rise to a non - empty colist of values of @xmath378 , which is strictly decreasing and must hence be of finite length .
the concatenation of a finite colist of finite traces is finite .
2 . if @xmath379 and @xmath380 in the first state of a trace , then @xmath381 ( everywhere ) , so in the last state @xmath382 gives @xmath383 , which can be weakened to @xmath380 .
[ prop : total_hoare ] for any @xmath1 , @xmath2 and @xmath354 , if @xmath351 is derivable in the total - correctness hoare logic with @xmath365 , then @xmath355 is derivable in the trace - based hoare logic .
immediate from proposition [ lemma : total_hoare ] by choosing @xmath384 .
given that derivations in the standard hoare logics can be transformed into derivations in the trace - based hoare logic , it is natural to wonder , if it is also possible to translate derivations in the converse direction .
in this direction we would expect some loss ( or displacement ) of information . reducing a condition on traces into a condition on those last states that happen to exist or into a condition that also requires their existence
must lose or displace the constraints on the intermediate states .
we now proceed to demonstrating that meaningful transformations ( `` projections '' ) from the trace - based logic to the standard logics are indeed possible .
we will show that , if @xmath0 is derivable in the trace - based hoare logic , then so are @xmath385 in the partial - correctness hoare logic and @xmath386 in the total - correctness hoare logic .
we will shortly define @xmath387 formally , but intuitively , the total - correctness judgement states that @xmath1 terminates in a state satisfying @xmath388 from any initial state @xmath22 such that @xmath2 holds in @xmath22 and any @xmath22-headed trace satisfying @xmath3 has length at most @xmath389 . since @xmath389 is universally quantified on the top level , we can actually derive @xmath390 , stating that @xmath1 terminates in a state satisfying @xmath388 from any initial state @xmath22 such that @xmath2 holds and the length of @xmath22-headed traces satisfying @xmath3 is bounded .
[ lemma : back_hoare ] for any @xmath1 , @xmath2 and @xmath3 , if @xmath0 in the trace - based hoare logic , then for any @xmath366 , @xmath391 is derivable in the partial - correctness hoare logic .
by induction on the derivation of @xmath0 .
we show the main cases of sequence and while .
* @xmath260 : we are given as the induction hypotheses that @xmath392 , for any @xmath368 , and @xmath393 , for any @xmath370 .
we have to show that , for any @xmath366 , @xmath394 .
let @xmath395 .
we have the derivation @xmath360{{\{u \wedge w\}~{\ensuremath{s_0;s_1}}~\{{\mathit{last } ~({\langlew\rangle } { \mathbin{\ast\ast}}p { \mathbin{\ast\ast}}q)}\ } } } { \infer{{\{u \wedge w\}~{\ensuremath{s_0;s_1}}~\{{\mathit{last } ~({\langlev'\rangle } { \mathbin{\ast\ast}}q)}\ } } } { \infer*[{\mathrm{ih}}_0\ , w]{{\{u \wedge w\}~s_0~\{v'\ } } } { } & \infer[(2)]{{\{v'\}~s_1~\{{\mathit{last } ~({\langlev'\rangle } { \mathbin{\ast\ast}}q)}\ } } } { \infer*[{\mathrm{ih}}_1\ , v']{{\{v \wedge v'\}~s_1~\{{\mathit{last } ~({\langlev'\rangle } { \mathbin{\ast\ast}}q)}\ } } } { } } } } \ ] ] 1 . recalling that @xmath396 is monotone
, we have @xmath397 2 . by lemma [ lemma : misc_asserts ] ( 10 ) and ( 11 ) , @xmath398 , therefore @xmath399 . * @xmath276 : we are given as the induction hypothesis that , for any @xmath375 , @xmath400 .
we now have to prove that , for any @xmath366 , @xmath401 , given @xmath280 .
let @xmath402 .
we close the case by the derivation @xmath360 { { \{u \wedge w\}~{\ensuremath{\mathsf{while~ } e \mathsf{~do~ } s_t}}~\{{\mathit{last } ~({\langlew\rangle } { \mathbin{\ast\ast}}{\langle u \rangle^2 } { \mathbin{\ast\ast}}{(p { \mathbin{\ast\ast}}{\langle i \rangle^2})^\dagger } { \mathbin{\ast\ast}}{\langle\neg e\rangle})}\ } } } { \infer { { \{i'\}~{\ensuremath{\mathsf{while~ } e \mathsf{~do~ } s_t}}~\{i ' \wedge \neg e\ } } } { \infer[(2 ) ] { { \{e \wedge i'\}~s_t~\{i'\ } } } { \infer*[{\mathrm{ih}}_t\ , i']{{\{e \wedge
i \wedge i'\}~s_t~\{{\mathit{last } ~({\langlei'\rangle } { \mathbin{\ast\ast}}p { \mathbin{\ast\ast}}{\langlei\rangle})}\ } } } { } } } } \ ] ] 1 .
we have @xmath403 2 .
we have @xmath404 [ prop : back_hoare ] for any @xmath1 , @xmath2 and @xmath3 , if @xmath0 , then @xmath405 is derivable in the partial correctness hoare logic .
immediate from proposition [ lemma : back_hoare ] and lemma [ lemma : misc_asserts ] ( 3 ) by instantiating @xmath384 . to define the assertion translation for the projection of the trace - based hoare logic into the total - correctness hoare logic , we introduce two new connectives .
the inductively defined trace predicate @xmath406 is true of finite traces with length at most @xmath218 ( we take the singleton trace to have length 0 ) . given a trace predicate @xmath3 , the state predicate @xmath407 is defined to be true of a state @xmath22 , if every trace headed by @xmath22 and satisfying @xmath3 also satisfies @xmath406 .
@xmath408 \infer{{\sigma}\models { \lceil p \rceil
n } } { \forall { \ensuremath{\tau}}.\ , { \mathit{hd}}~{\ensuremath{\tau}}= { \sigma}\wedge { \ensuremath{\tau}}\models p \to { \ensuremath{\tau}}\models { \mathit{len}\ , n } } \end{array}\ ] ] [ lemma : ceil_antimonotone ] for any @xmath218 , @xmath3 and @xmath83 , if @xmath409 then @xmath410 . for setoid predicates @xmath3 and @xmath83 , we say @xmath83 _ extends _
@xmath3 , written @xmath411 , if , whenever @xmath90 , there exists @xmath27 such that @xmath86 , i.e. , @xmath412 .
[ lemma : after_monotone ] for any @xmath413 and @xmath414 , if @xmath415 and @xmath416 then @xmath417
. follows from the definition .
[ lemma : after_chop ] for any @xmath139 and @xmath140 , @xmath418 .
suppose @xmath419 .
there exists @xmath65 such that @xmath420 and @xmath421 .
we then have @xmath422 .
the hypothesis gives us that there exists @xmath156 such that @xmath423 .
we then build a trace @xmath424 such that @xmath425 from @xmath156 by dropping the first @xmath218 elements , where @xmath218 is the length of @xmath65 .
( @xmath218 is welldefined because @xmath65 is finite . )
[ lemma : after_last ] for any @xmath3 and @xmath83 , @xmath426 .
suppose @xmath90 . from the hypothesis
, we know that , for any @xmath22 , if @xmath130 , then there exists @xmath27 such that @xmath427 and @xmath213 . using this , we build a trace @xmath27 such that @xmath86 , by traversing @xmath12 and invoking the hypothesis when the last state of @xmath12 is hit .
[ prop : back_hoaretotal ] for any @xmath1 , @xmath2 and @xmath3 , if @xmath0 in the trace - based hoare logic , then for any @xmath428 and @xmath140 such that @xmath429 , @xmath430 is derivable in the total - correctness hoare logic with @xmath431 .
by induction on the derivation of @xmath0 .
we show the main cases of sequence and while .
* @xmath260 : we are given as the induction hypotheses that , for any @xmath432 and @xmath433 such that @xmath434 , we have that @xmath435 and , for any @xmath436 and @xmath437 such that @xmath438 , we have that @xmath439 we have to prove that , for any @xmath440 and @xmath140 such that @xmath441 we have that @xmath442 + let @xmath443 .
we close the case by the derivation @xmath444{{\{u \wedge w \wedge { \lceil p { \mathbin{\ast\ast}}q { \mathbin{\ast\ast}}r \rceil m}\}~{\ensuremath{s_0;s_1}}~\{{\mathit{last } ~({\langlew\rangle } { \mathbin{\ast\ast}}p { \mathbin{\ast\ast}}q ) } \wedge { \lceil r \rceil m}\ } } } { \infer{{\{u \wedge w \wedge { \lceil p { \mathbin{\ast\ast}}q { \mathbin{\ast\ast}}r \rceil m}\}~{\ensuremath{s_0;s_1}}~\{{\mathit{last } ~({\langlev'\rangle } { \mathbin{\ast\ast}}q ) } \wedge { \lceil r \rceil m}\ } } } { \infer[(3)]{{\begin{array}{l}\{u \wedge w \wedge { \lceil p { \mathbin{\ast\ast}}q { \mathbin{\ast\ast}}r \rceil m}\ } s_0 \\ \quad \{v ' \wedge { \lceil q { \mathbin{\ast\ast}}r \rceil m}\ } \end{array } } } { \infer*[(1)~{\mathrm{ih}}_0\ , m , w , q { \mathbin{\ast\ast}}r]{{\begin{array}{l}\{u \wedge w \wedge { \lceil p { \mathbin{\ast\ast}}{\langlev\rangle } { \mathbin{\ast\ast}}q { \mathbin{\ast\ast}}r \rceil m}\ } s_0 \\ \quad \{v ' \wedge { \lceil q { \mathbin{\ast\ast}}r \rceil m}\ } \end{array } } } { } } & \infer[(4)]{{\begin{array}{l}\ { v ' \wedge { \lceil q { \mathbin{\ast\ast}}r \rceil m}\}s_1 \\
\quad \{{\mathit{last } ~({\langlev'\rangle } { \mathbin{\ast\ast}}q ) } \wedge { \lceil r \rceil m}\ } \end{array } } } { \infer*[(2)~{\mathrm{ih}}_1\ , m , v ' , r]{{\begin{array}{l}\{v \wedge v ' \wedge { \lceil q { \mathbin{\ast\ast}}r \rceil m}\}s_1 \\ \quad \{{\mathit{last } ~({\langlev'\rangle } { \mathbin{\ast\ast}}q ) } \wedge { \lceil r \rceil m}\ } \end{array } } } { } } } } \ ] ] 1 . to invoke the induction hypothesis
, we have to prove @xmath445 . by corollary [ coro : semax_satisfiable ] and the hypothesis @xmath446 , we have that , for any @xmath22 such that @xmath447 , there exists @xmath12 such that @xmath448 and @xmath449 .
this gives us @xmath450 . by lemmata [ lemma : misc_asserts ] ( 11 ) and
[ lemma : after_monotone ] , we have @xmath451 , therefore @xmath452 by lemma [ lemma : after_last ] . now , for any @xmath12 , suppose @xmath453 .
there exists @xmath27 such that @xmath86 .
we deduce @xmath454 . from the hypothesis @xmath455 , we know that there exists @xmath35 such that @xmath154 , therefore @xmath160 , as required .
2 . to invoke the induction hypothesis
, we have to prove @xmath456 , namely @xmath457 . by lemma
[ lemma : after_chop ] , it suffices to show @xmath458 , which follows from the hypothesis ( @xmath455 ) , @xmath459 and lemma [ lemma : after_monotone ] .
3 . @xmath460 follows from lemma [ lemma : ceil_antimonotone ] and @xmath461 .
4 . @xmath399 holds because @xmath462 from lemmata [ lemma : misc_asserts ] ( 10 ) and ( 11 ) .
we have @xmath463 * @xmath276 .
we are given as the induction hypothesis that , for any @xmath464 , @xmath375 and @xmath465 such that @xmath466 , @xmath467 we also know @xmath280 and @xmath468 , the latter of which gives us @xmath469 we have to prove that , for any @xmath440 and @xmath140 such that @xmath470 we have that @xmath471 let @xmath472 and @xmath473 and @xmath474 we first prove that , for all @xmath218 , @xmath475 holds by case analysis on @xmath218 .
* * @xmath476 .
we prove @xmath477 .
then , by the derivability of @xmath478 for any @xmath1 and @xmath2 in the total correctness hoare logic , we obtain ( [ eq : back_hoare : while_body ] ) .
suppose @xmath479 holds . from @xmath480
, there exists a trace @xmath64 such that @xmath481 and @xmath208 . by lemma [ lemma : misc_asserts ]
( 14 ) , @xmath480 and @xmath482 give us @xmath483 .
applying corollary [ coro : semax_satisfiable ] to ( [ eq : back_hoare : whilei ] ) , we know that there exists a trace @xmath65 such that @xmath484 and @xmath485 . hence , we have @xmath486 . we deduce @xmath487 where @xmath488 was defined by corecursion by @xmath489
the assumption @xmath455 gives us a trace @xmath156 such that @xmath490 .
since @xmath64 is finite , we find the suffix @xmath424 of @xmath156 such that @xmath491 by dropping the first @xmath218 elements from @xmath156 , where @xmath218 is the length of @xmath64 .
we also have @xmath492 by construction .
the trace @xmath424 can not be a singleton @xmath493 , which would imply @xmath494 and @xmath495 .
contradiction by @xmath496 . * * @xmath497 .
we invoke the induction hypothesis @xmath498 on @xmath218 , @xmath499 and @xmath83 .
to do so , we have to prove @xmath500 by lemma [ lemma : after_chop ] , it suffices to prove @xmath501 . from corollary [ coro : semax_satisfiable ] and
( [ eq : back_hoare : whilei ] ) , we deduce @xmath502 , hence @xmath503 by lemma [ lemma : after_last ] .
this together with the hypothesis ( @xmath455 ) proves ( [ eq : back_hoare : ih_cond ] ) .
+ we have now proved @xmath504 we close the case by invoking the consequence rule with @xmath505 and @xmath506 + \(1 ) given a trace @xmath507 and a state @xmath22 satisfying @xmath508 and @xmath509 , we have @xmath510 and @xmath511 .
+ \(2 ) from lemma [ lemma : misc_asserts ] ( 14 ) and @xmath280 , we know , for any @xmath22 , if @xmath512 , namely @xmath513 , then @xmath514 .
suppose @xmath515 .
it suffices to prove that , for any @xmath12 , if @xmath448 and @xmath516 , then @xmath517 . we have @xmath518 , which together with @xmath519 yields @xmath520 .
this gives us @xmath517 .
+ by the rule for the while - statement by taking @xmath521 to be @xmath522 , we deduce @xmath523 we have @xmath524 and @xmath525 so , by consequence , we have @xmath526 as we wanted .
[ cor : back_hoaretotal ] for any @xmath1 , @xmath2 and @xmath3 , if @xmath0 in the trace - based hoare logic , then @xmath527 is derivable in the total - correctness hoare logic with @xmath431 . from proposition [ prop : back_hoaretotal ] by taking @xmath384 , @xmath528 and then invoking the consequence rule and the admissible rule @xmath529 of the total - correctness hoare logic .
let us now discuss the transformations from and to the total - correctness hoare logic .
we used two different rules for while . in the forward direction ( the embedding ) , we used @xmath365 , whereas in the backward direction ( the projection ) , we used @xmath431 . with @xmath431 ,
the total - correctness hoare logic is complete , constructively . in particular
, @xmath365 is weaker and straightforwardly derivable from @xmath431 by instantiating @xmath530 . in the converse direction , attempts to derive @xmath365 from @xmath431 constructively meet several problems .
it is natural to define the invariant @xmath248 as @xmath531 .
the variant @xmath378 should then be defined as @xmath532 this is a sensible classical definition .
but constructively , there are two issues with it .
first , in the first case , we must be able to find the least @xmath533 such that @xmath534 .
we have a bound for this search , which is given by the witness @xmath535 of @xmath536 .
but in order to search , we must be able to decide where @xmath534 or @xmath537 for each @xmath538 , which requires @xmath539 to be decidable .
second , in order to choose between the two cases , we must be able to decide whether @xmath536 or @xmath540 , which is generally impossible .
these issues show that @xmath365 is not fine - tuned for constructive reasoning . to formulate a more sensitive rule
, we may notice that we only ever need @xmath378 in contexts where @xmath531 can be assumed to hold .
so instead of @xmath541 we could use @xmath542 making the variant @xmath378 depend on the proof of the invariant .
the corresponding modified rule is @xmath543 { { \{\sigma p : i.\ , t\ , p = m\}~{\ensuremath{\mathsf{while~ } e \mathsf{~do~ } s_t}}~\{\sigma p : i.\ , t\ , p \leq m \wedge \neg e\ } } } { \forall n:\mathit{nat } ~{\{e \wedge \sigma p : i.\ , t\ , p = n\}~s_t~\{\sigma p : i.\ , t\ , p < n\ } } } \ ] ] ideally , we would like to use @xmath431 ( or some interderivable rule ) both in the transformations to and from the trace - based hoare logic .
but in the forward direction we have been unable to do it .
we conjecture the reason to be that the state - based total - correctness triples and trace - based ( partial - correctness ) triples express different types of propositions about the underlying evaluation relations : total - correctness triples are about existence of a final state satisfying the postcondition , whereas partial - correctness triples are about all traces produced satisfying the postcondition .
this discrepancy shows itself also in our backward transformation : we need to invoke the totality of the trace - based evaluation relation .
propositions [ prop : partial_hoare ] and [ lemma : total_hoare ] show that our trace - based logic is expressive enough to perform the same type of analyses that the state - based partial or total correctness hoare logics can perform .
however , the expressiveness of our logic goes beyond that of the partial and the total correctness hoare logics . in this section
, we demonstrate this by a series of examples .
we adopt the usual notational convention that any occurrence of a variable in a state predicate represents the value of the variable in the state , e.g. , a state predicate @xmath544 abbreviates @xmath545 .
since we work in a constructive underlying logic , we can distinguish between termination of a run , @xmath95 , and nondivergence , @xmath546 .
for instance , any unbounded nonpartial search fails to be terminating but is nonetheless nondivergent .
this example is inspired by markov s principle : @xmath547 for any decidable predicate @xmath548 on natural numbers , i.e. , a predicate satisfying @xmath549 .
markov s principle is a classical tautology , but is not valid constructively .
this implies we can not constructively prove a statement @xmath1 that searches a natural number @xmath218 satisfying @xmath548 by successively checking whether @xmath550 to be terminating . in other words , we can not constructively derive a total correctness judgement for @xmath1 .
the assumption @xmath551 only guarantees that @xmath548 is not false everywhere , therefore the search can not diverge ; indeed , we can constructively prove that @xmath1 is nondivergent in our logic .
we assume given a decidable predicate @xmath548 on natural numbers and an axiom : @xmath552 stating that @xmath548 is not false everywhere . therefore running the statement @xmath553 can not diverge
: this would contradict _ b_noncontradictory_. in proposition [ prop : markov ] we prove that any trace produced by running @xmath554 is nondivergent and @xmath555 holds of the last state .
we define a family of trace predicates @xmath556 coinductively as follows : @xmath557 @xmath558 is a setoid predicate .
a crucial observation is that , in the presence of _
b_noncontradictory _ , @xmath559 is stronger than nondivergent : [ lemma : cofinally_not_infinite ] @xmath560 .
it is sufficient to prove that , for any @xmath12 , @xmath561 and @xmath562 are contradictory .
suppose there is a trace @xmath12 such that @xmath561 and @xmath562 .
then by induction on @xmath218 we can show that , for any @xmath218 , there is a trace @xmath27 such that @xmath563 and @xmath564 .
but whenever this holds for some @xmath27 and @xmath218 , then @xmath565 .
hence we also have @xmath566 . but
this contradicts @xmath567 .
it is straightforward to prove that , if @xmath12 satisfies @xmath559 , then its last state ( if exists ) satisfies @xmath548 at @xmath19 .
( we refer to the program variable @xmath19 in the statement @xmath554 . )
[ lemma : cofinally_colast ] @xmath568 .
we prove the following condition by coinduction , from which the lemma follows : for any @xmath569 , if @xmath570 , then @xmath571 . [ prop : markov ] @xmath572 .
a sketch of the derivation is given in figure [ fig : markov ] .
( at several places we have applied the consequence rule silently . ) 1 .
we use lemmata [ lemma : cofinally_not_infinite ] and [ lemma : cofinally_colast ] .
@xmath19 is incremented by one in every iteration until @xmath548 holds at @xmath19 .
@xmath573}}{\mathbin{\ast\ast}}{\langle { \ensuremath{\mathsf{true}}}\rangle^2})^\dagger } { \mathbin{\ast\ast}}{\langleb~x\rangle } \models { { \it cofinally}}~0 $ ] follows from the definition of @xmath574 .
( it is proved by coinduction . )
we take @xmath73 as the loop invariant . as the similarity of our assertion language to the interval temporal logic suggests
, we can specify and prove liveness properties . in proposition
[ prop : liveness ] , we prove that the statement @xmath576 eventually sets the value of @xmath19 to @xmath218 for any @xmath577 at some point .
the example is simple but sufficient to demonstrate core techniques used to prove liveness properties of more practical examples .
for instance , imagine that assignment to @xmath19 involves a system call , with the assigned value as the argument .
it is straightforward to enrich traces to record such special events , and we can then apply the same proof technique to prove the statement eventually performs the system call with @xmath218 as the argument for any @xmath218 . for every @xmath218 ,
we define inductively a trace predicate @xmath578 stating that a state @xmath22 in which the value of @xmath19 is @xmath218 is eventually reached by finitely traversing @xmath12 : @xmath579 if @xmath12 satisfies @xmath580 , then it has a finite prefix followed by a state that maps @xmath19 to @xmath218 .
[ lemma : eventually_finite ] for any @xmath569 , if @xmath581 then @xmath582 . by induction on the derivation of @xmath580 .
[ prop : liveness ] for any @xmath577 , @xmath583 .
a sketch of the derivation is given in figure [ fig : liveness ] . 1
. we use lemma [ lemma : eventually_finite ] .
@xmath19 is incremented by one in every iteration , starting from zero .
each iteration is finite : the length of the trace of each iteration is invariably two .
@xmath19 must eventually become @xmath218 after a finite number of iterations for any @xmath218 .
we take @xmath73 as the invariant .
@xmath575 { { \{{\ensuremath{\mathsf{true}}}\}~{\ensuremath{{\ensuremath{x : = 0}};{\ensuremath{\mathsf{while~ } { \ensuremath{\mathsf{true}}}\mathsf{~do~ } { \ensuremath{x : = x+1}}}}}}~\{{\mathit{finite}}~{\mathbin{\ast\ast}}{\langlex = n\rangle } { \mathbin{\ast\ast}}{\ensuremath{\mathsf{true}}}\ } } } { \infer { { \{{\ensuremath{\mathsf{true}}}\}~{\ensuremath{{\ensuremath{x : = 0}};{\ensuremath{\mathsf{while~ } { \ensuremath{\mathsf{true}}}\mathsf{~do~ } { \ensuremath{x : = x+1}}}}}}~\{{\ensuremath{{\ensuremath{\mathsf{true}}}[x\mapsto 0 ] } } { \mathbin{\ast\ast}}{\mathit{eventually}}~n\ } } } { { \{{\ensuremath{\mathsf{true}}}\}~{\ensuremath{x : = 0}}~\{{\ensuremath{{\ensuremath{\mathsf{true}}}[x\mapsto 0]}}\ } } & \infer[(2)]{{\{x=0\}~{\ensuremath{\mathsf{while~ } { \ensuremath{\mathsf{true}}}\mathsf{~do~ } { \ensuremath{x : = x+1}}}}~\{{\mathit{eventually}}~n\ } } } { { \hspace*{2 mm } \deduce{\line(-1,6){2}}{\makebox[0 cm ] { \hspace{-4 cm } \infer[(3 ) ] { { \begin{array}{l}\{x=0\ } ~{\ensuremath{\mathsf{while~ } { \ensuremath{\mathsf{true}}}\mathsf{~do~ } { \ensuremath{x : = x+1 } } } } \\
\quad \{{\langle x=0 \rangle^2 } { \mathbin{\ast\ast}}{({\ensuremath{{\ensuremath{\mathsf{true}}}[x\mapsto x+1 ] } } { \mathbin{\ast\ast}}{\langle { \ensuremath{\mathsf{true}}}\rangle^2})^\dagger } { \mathbin{\ast\ast}}{\langle{\ensuremath{\mathsf{false}}}\rangle}\ } \end{array } } } { { \{{\ensuremath{\mathsf{true}}}\}~{\ensuremath{x : = x+1}}~\{{\ensuremath{{\ensuremath{\mathsf{true}}}[x\mapsto x+1]}}\ } } } } } \hspace*{2 mm } } } } } \end{array}\ ] ] the last example is inspired by a notion of weak trace equivalence : two traces are weakly equivalent if they are bisimilar by identifying a finite number of consecutive identical states with a single state .
it is conceivable that ( strong ) bisimilarity is too strong for some applications and one needs weak bisimilarity .
for instance , we may want to prove that the observable behavior , such as the colist of i / o events of a potentially diverging run , is bisimilar to a particular colist of i / o events
. then we must be able to collapse a finite number of non - observable internal steps .
we definitely should not collapse an infinite number of internal steps , otherwise we would end up concluding that a statement performing an i / o operation after a diverging run , e.g. , @xmath584 , is observably equivalent to a statement immediately performing the same i / o operation , e.g. , @xmath585 . in this subsection , we prove that the trace produced by running the statement @xmath586 is weakly bisimilar to the ascending sequence of natural numbers @xmath587 , by projecting the value of @xmath19 , assuming that @xmath19 is initially 0 .
the statement differs from that of the previous subsection in that it `` stutters '' for a finite but unbounded number of steps , i.e. , @xmath588 , before the next assignment to @xmath19 happens .
this exercise is instructive in that we need to formalize weak trace equivalence in our constructive underlying logic .
we do so by supplying an inductive predicate @xmath589 stating that @xmath27 is obtained from @xmath12 by dropping finitely many elements from the beginning , until the first state with a different value of @xmath19 is encountered , and a coinductive trace predicate @xmath590 stating that the given trace is weakly bisimilar to the ascending sequence of natural numbers starting at @xmath218 , by projecting the value of @xmath19 .
formally : @xmath591 @xmath592 these definitions are tailored to our example . but a more general weak trace equivalence can be defined similarly @xcite .
we note that our formulation is not the only one possible nor the most elegant .
in particular , with a logic supporting nesting of induction into coinduction as primitive @xcite , there is no need to separate the definition into an inductive part , @xmath589 , and a coinductive part , @xmath590 . yet
our formulation is amenable for formalization in coq .
we also use an auxiliary trace predicate @xmath593 that is true of a finite trace in which the value of @xmath19 does not change and is non - negative at the end and hence everywhere .
it is defined inductively as follows : @xmath594 [ prop : wkbism ] @xmath595 where @xmath596 .
a sketch of the derivation is given in figure [ fig : wstream ] .
@xmath19 is incremented by one in every iteration , starting from zero , and the loop is iterated forever .
each iteration is finite , although globally unbounded .
so @xmath19 must increase forever .
we take @xmath597 as the invariant .
@xmath598 is decremented by one in every iteration , starting from a non - negative number .
it must become zero after a finite number of iterations .
@xmath19 does not change .
we take @xmath599 as the invariant .
coinductive big - step semantics for nontermination have been considered by cousot and cousot @xcite and by leroy and grall @xcite ( in the context of the compcert project , which is a major demonstration of feasibility of certified compilation ) .
cousot and cousot @xcite study fixpoints on bi - semantic domains , partitioned into domains of terminating and diverging behaviors ; they prove a specific fixed - point theorem for such domains ( bi - induction ) , and then produce a bi - inductive big - step semantics for lambda - calculus ( avoiding some duplication of rules between what would otherwise be distinct inductive and coinductive definitions of terminating resp . diverging evaluation ) .
leroy and grall @xcite investigate two big - step semantics approaches for lambda - calculus .
the first , based on cousot and cousot @xcite , has different evaluation relations for terminating and diverging runs , one inductive ( with finite traces ) , the other coinductive ( with infinite traces ) . to conclude that any program either terminates or diverges
, they need the law of excluded middle ( amounting to decidability of the halting problem ) , and , as a result , can not prove the standard small - step semantics sound wrt .
the big - step semantics constructively .
the second approach , applied in @xcite , uses a coinductively defined evaluation relation with possibly infinite traces .
while - loops are not ensured to be progressive in terms of growing traces ( an infinite number of consecutive silent small steps may be collapsed ) and this leads to problems .
some other works on coinductive big - step semantics include glesner @xcite and nestra @xcite .
in these it is accepted that a program evaluation can somehow continue after an infinite number of small steps . with glesner
, this seems to have been a curious unintended side - effect of the design , which she was experimenting with just for the interest of it .
nestra developed a nonstandard semantics with transfinite traces on purpose in order to obtain a soundness result for a widely used slicing transformation that is unsound standardly ( can turn nonterminating runs into terminating runs ) .
our trace - based coinductive big - step semantics @xcite was heavily inspired by capretta s
@xcite modelling of nontermination in a constructive setting similar to ours . rather than using coinductive
possibly infinite traces , he works with a coinductive notion of a possibly infinitely delayed value ( for statements , this corresponds to delaying the final state ) .
the categorical basis appears in rutten s work @xcite .
but rutten only studied the classical setting ( any program terminates or not ) , where a delayed state collapses to a choice of between a state or a designated token signifying nontermination .
a general categorical account of small - step trace - based semantics has been given by hasuo et al .
@xcite .
while hoare logics for big - step semantics based on inductive , finite traces have been considered earlier ( to reason about traces of terminating runs ) , hoare or vdm - style logics for reasoning about properties of nonterminating runs do not seem to have been studied before , with one exception , see below .
neither do we in fact know about dynamic logic or kat ( kleene algebra with tests ) approaches that would have assertions about possibly infinite traces .
rather , nonterminating runs have been typically reasoned about in temporal logics like ltl and ctl@xmath601 or in interval temporal logic @xcite .
these are however essentially different in spirit by their `` endogeneity '' : assertions are made about traces in a fixed transition system rather than traces of runs of different programs .
notably , however , interval temporal logic has connectives similar to ours in fact they were a source of inspiration for our design .
hofmann and pavlova @xcite consider a vdm - style logic with finite trace assertions that are applied to all finite prefixes of the trace of a possibly nonterminating run of a program .
this logic allows reasoning about safety , but not liveness .
we expect that we should be able to embed a logic like this in ours .
we have presented a sound and complete hoare logic for the coinductive trace - based big - step semantics of while .
the logic naturally extends both the standard state - based partial and total correctness hoare logics .
its design may be exploratory at this stage in the sense that one might wish to consider alternative choices of primitive connectives .
we see our logic as a viable unifying foundational framework facilitating translations from more applied logics .
we are grateful to martin hofmann , thierry coquand and adam chlipala for discussions .
our anonymous reviewers provided very detailed and useful feedback .
20 y. bertot , p. castran .
_ interactive theorem proving and program development : coqart : the calculus of inductive constructions _ , _ texts in theoretical computer science : an eatcs series_. springer , 2004 .
s. blazy , x. leroy .
mechanized semantics for the clight subset of the c language . _ j. of autom . reasoning _ , 43(3):263288 , 2009
. v. capretta .
general recursion via coinductive types .
methods in comput .
_ , 1(2:1 ) , 2005 . p.
cousot , r. cousot .
inductive definitions , semantics and abstract interpretation . in _ conf .
record of 19th acm sigplan - sigact symp . on principles of programming languages , popl 92 ( albuquerque , nm , jan.1992 )
_ , pp . 8394 .
acm press , 1992 .
p. cousot , r. cousot .
bi - inductive structural semantics . _ inform . and comput .
_ , 207(2):258283 , 2009 .
n. a. danielsson , t. altenkirch .
subtyping , declaratively . in c. bolduc , j. desharnais , b. ktari , eds . , _ proc . of 10th int .
conf . on mathematics of program construction , mpc 2000 ( qubec city , june 2010 )
_ , vol .
6120 of _ lect .
notes in comput .
_ , pp . 100118 .
springer , 2010 .
s. glesner . a proof calculus for natural semantics based on greatest fixed point semantics . in j.
knoop , g. c. necula , w. zimmermann , eds . , _ proc . of 3rd int .
wksh . on compiler
optimization meets compiler verification , cocv 2004 ( barcelona , apr .
2004 ) _ , vol .
132(1 ) of _ electron .
notes in theor .
_ , pp . 7393 .
elsevier , 2004 .
j. halpern , z. manna , b. moszkowski . a hardware semantics based on temporal intervals . in j.
daz , ed .
, _ proc . of 10th int .
on automata , languages and programming , icalp 83 ( barcelona , july 1983 ) _ , vol .
154 of _ lect .
notes in comput .
_ , pp . 278191 .
springer , 1983 .
i. hasuo , b. jacobs , a. sokolova .
generic trace semantics via coinduction .
methods in comput .
_ , 3(4:11 ) , 2007 .
m. hofmann , m. pavlova .
elimination of ghost variables in program logics . in g.
barthe , c. fournet , eds . , _ revised selected papers from 3rd symp . on trustworthy global computing , tgc 2007 ( sophia antipolis , nov .
2007 ) _ , vol .
4912 of _ lect .
notes in comput .
_ , pp . 120 .
springer , 2008 .
x. leroy .
coinductive big - step operational semantics . in p.
sestoft , ed . , _ proc . of 15th europ .
symp . on programming ,
esop 2006 ( vienna , march 2006 ) _ , vol .
3924 of _ lect .
notes in comput .
_ , pp . 5468 .
springer , 2006 .
x. leroy , h. grall .
coinductive big - step operational semantics .
_ inform . and comput .
_ , 207(2):285305 , 2009 .
t. kleymann .
hoare logic and auxiliary variables .
_ formal asp . in comput .
_ 11(5):541566 , 1999
. b. moszkowski . a temporal logic for reasoning about hardware .
_ computer _ , 18(2):1019 , 1985 .
k. nakata , t. uustalu . a hoare logic for the coinductive trace - based big - step semantics of while . in a.
d. gordon , ed .
, _ proc . of 19th european symp . on programming , esop 2010 ( paphos , march 2010 )
_ , vol .
6012 of _ lect .
notes in comput .
_ , pp . 488506 .
springer , 2010 .
k. nakata , t. uustalu .
resumptions , weak bisimilarity and big - step semantics for while with interactive i / o : an exercise in mixed induction - coinduction . in l.
aceto , p. sobocinski , eds .
, _ proc . of 7th wksh . on structural operational semantics , sos 2010 ( paris , aug .
2010 ) _ , vol .
32 of _ electron .
proc . in theor .
_ , pp . 5775 .
open publishing assoc . , 2010 .
k. nakata , t. uustalu .
trace - based coinductive operational semantics for while : big - step and small - step , relational and functional styles . in s.
berghofer , t. nipkow , c. urban , m. wenzel , eds .
, _ proc . of 22nd int .
conf . on theorem proving in higher order logics , tphols 2009 ( munich , aug .
2009 ) _ , vol .
5674 of _ lect .
notes in comput .
_ , pp . 375390 .
springer , 2009 .
h. nestra .
fractional semantics . in m.
johnson , v. vene , eds .
, _ proc . of 11th int .
conf . on algebraic methodology and software technology ,
amast 2006 ( kuressaare , july 2006 ) _ , vol . 4019 of _ lect .
notes in comput .
_ , pp . 278292 .
springer , 2006 . h. nestra .
transfinite semantics in the form of greatest fixpoint .
_ j. of log . and algebr
. program .
_ , 78(7):574593 , 2009 .
h. riis nielson , f. nielson .
_ semantics with applications : a formal introduction_. wiley , 1992 .
j. rutten . a note on coinduction and weak bisimilarity for while programs .
_ theor .
inform . and appl .
_ , 33(45):393400 , 1999 .
the figures below give the rules of the standard , state - based partial correctness and total correctness hoare logics in the form used in section [ sec : hoare ] .
@xmath602\}~{\ensuremath{x : = e}}~\{u\ } } } { } \quad \infer { { \{u\}~{\ensuremath{\mathsf{skip}}}~\{u\ } } } { } \quad \infer{{\{u\}~s_0;s_1~\{z\ } } } { { \{u\}~s_0~\{v\ } } & { \{v\}~s_1~\{z\ } } } \\[2ex ] \infer { { \{u\}~{\ensuremath{\mathsf{if~ } e \mathsf{~then~ } s_t \mathsf{~else~ } s_f}}~\{z\ } } } { { \{e \wedge u\}~s_t~\{z\ } } & { \{\neg e \wedge u\}~s_f~\{z\ } } } \quad \infer { { \{i\}~{\ensuremath{\mathsf{while~ } e \mathsf{~do~ } s_t}}~\{i \wedge \neg e\ } } } { { \{e \wedge i\}~s_t~\{i\ } } } \\[2ex ] \infer { { \{\exists z.\ , u\}~s~\{\exists z.\ , v\ } } } { \forall z.\ , { \{u\}~s~\{v\ } } } \quad \infer { { \{u\}~s~\{z\ } } } { u \models u ' & { \{u'\}~s~\{z'\ } } & z ' \models z } \end{array}\ ] ] @xmath602\}~{\ensuremath{x : = e}}~\{u\ } } } { } \quad \infer { { \{u\}~{\ensuremath{\mathsf{skip}}}~\{u\ } } } { } \quad \infer{{\{u\}~s_0;s_1~\{z\ } } } { { \{u\}~s_0~\{v\ } } & { \{v\}~s_1~\{z\ } } } \\[2ex ] \quad \infer { { \{u\}~{\ensuremath{\mathsf{if~ } e \mathsf{~then~ } s_t \mathsf{~else~ } s_f}}~\{z\ } } } { { \{e \wedge u\}~s_t~\{z\ } } & { \{\neg e \wedge u\}~s_f~\{z\ } } } \\[2ex ] \infer[{\textrm{while - fun } } ] { { \{i \wedge t = m\}~{\ensuremath{\mathsf{while~ } e \mathsf{~do~ } s_t}}~\{i \wedge t \leq m \wedge \neg e\ } } } { \forall n:\mathit{nat } ~{\{e \wedge i \wedge t = n\}~s_t~\{i \wedge t < n\ } } } \\[2ex ] \mathrm{alt . }
\\[2ex ] \infer[{\textrm{while - rel } } ] { { \{j\ , m\}~{\ensuremath{\mathsf{while~ } e \mathsf{~do~ } s_t}}~\{\exists k.\ , k \leq m \wedge j\ , k
\wedge \neg e\ } } } { \forall n:\mathit{nat } ~{\{e \wedge j\ , n\}~s_t~\{\exists k.\ , k < n \wedge j\ , k\ } } } \\[2ex ] \infer { { \{\exists z.\ , u\}~s~\{\exists z.\ , v\ } } } { \forall z.\ , { \{u\}~s~\{v\ } } } \quad \infer { { \{u\}~s~\{z\ } } } { u \models u ' & { \{u'\}~s~\{z'\ } } & z ' \models z } \end{array}\ ] ] | in search for a foundational framework for reasoning about observable behavior of programs that may not terminate , we have previously devised a trace - based big - step semantics for while . in this
semantics , both traces and evaluation ( relating initial states of program runs to traces they produce ) are defined coinductively . on terminating runs
, this semantics agrees with the standard inductive state - based semantics . here
we present a hoare logic counterpart of our coinductive trace - based semantics and prove it sound and complete .
our logic subsumes the standard partial - correctness state - based hoare logic as well as the total - correctness variation : they are embeddable . in the converse direction
, projections can be constructed : a derivation of a hoare triple in our trace - based logic can be translated into a derivation in the state - based logic of a translated , weaker hoare triple .
since we work with a constructive underlying logic , the range of program properties we can reason about has a fine structure ; in particular , we can distinguish between termination and nondivergence , e.g. , unbounded classically total search fails to be terminating , but is nonetheless nondivergent .
our metatheory is entirely constructive as well , and we have formalized it in coq . |
methane ( ) is an organic molecule in earth s atmosphere primarily produced by living organisms @xcite .
it has also been measured in mars s atmosphere by telescopic and spacecraft remote sensing @xcite . however , these measurements produced inconsistent results and some of these measurements have been called into question ( e.g. * ? ? ? * ) .
recently , an _ in situ _ measurement of methane on mars has been made : msl s tunable laser spectrometer ( tls ) as a part of the sample analysis at mars ( sam ) instrument determined a background mixing ratio of @xmath1 ppbv and a pulse of @xmath2 ppbv observed over two months @xcite .
these measurements suggest strong temporal variability of the methane abundance on the surface of mars .
methane s atmospheric existence requires a geologically recent or continually replenishing source , because methane has an lifetime of @xmath3 years in mars oxidizing atmosphere @xcite .
this source can be photo - degradation of organic matter in the meteorites fallen onto mars @xcite .
alternatively , methane has to come from mars itself , which challenges the conventional understanding of a geologically and biologically dead mars ( e.g. , * ? ? ?
* ; * ? ? ?
an indigenous source of methane is corroborated by a recent discovery of methane evolved from martian meteorites @xcite .
furthermore , atmospheric processes alone can not produce the variability of methane detected by the curiosity , because the atmospheric mixing time of methane is much shorter than its chemical lifetime @xcite .
the discovery of methane thus compels a new chapter of mars research to explain the existence of methane and its variability in the martian atmosphere .
extrapolating our knowledge of terrestrial biotic sources to mars , many consider methanogens ( a type of -producing archaean microbe ) as a probable analog to martian life forms @xcite .
some methanogens are able to utilize inorganic compounds ( and ) as their only source of energy and produce methane
. being independent of photosynthesis for subsistence , methanogens can thrive in the deep subsurface where is the predominant oxidant and ( aq ) is abundant from water - rock interactions ( e.g. , ferrous - iron reduction of to during serpentinization @xcite ) .
may also come from photolysis of in the atmosphere @xcite .
in fact , methanogens thrive in some of the harshest environments on earth , including extremely acidic environments and inside greenland glacial ice 3-km deep , which is analogous to martian subsurface ice environments @xcite .
alternatively , fischer - tropsch - type ( ftt ) reactions can be a potential methane source @xcite .
ftt is the most widely posited abiotic source of methane on earth . catalyzed by transition metals and related oxides ,
these reactions have the same overall chemistry as the methanogenesis , take place in hydrothermal environments , and probably source also from serpentinization .
abundant evidence indicates that volcanism and hydrothermal environments existed , and might still exist , on mars @xcite .
these environments provide heat and liquid water to support ftt and/or microbial production .
thus , atmospheric may point to either serpentinization or the existence of life itself , both of which are associated with a warm , habitable backdrop .
in this paper we focus on the singularly important discovery of the episodically enhanced emission on mars .
the elevated methane levels occurred in the southern - hemisphere winter , except for a single and statistically marginal measurement on sol 306 @xcite .
a first - order question about the curiosity s methane spikes is whether they represent new source to the atmosphere , or some cyclic processes that conserve methane .
if the methane spikes represent new methane produced from the deep subsurface @xcite or from a meteoritic source @xcite , then this methane would be oxidized and remain in the martian atmosphere in the form of . as a lower estimate , if 7 ppb of methane is produced per year across the entire martian surface ( representing the spike from @xmath4 ) , then over 3 billion years , the total amount of produced is 20 times the present atmospheric content .
this is much higher than the amazonian outgassing rates determined from photogeologic constraints @xcite or what would be allowed by atmospheric evolution models @xcite . on the other hand ,
if the methane is converted to organics and stored in the regolith , then the molar density of organics originating from methane would be ( 8.4/@xmath5 ) mol per gram of soil , assuming as the organic molecule , with @xmath5 as the depth in the soil in units of cm to which the organics are stored .
this is far greater than the observed molar density for any reasonable values of @xmath5 @xcite .
therefore , if the spikes are new methane , the source must be local ; alternatively , the spikes are signatures of cyclic processes that produce fast source and sink .
interestingly , the methane spikes were coincident with surface relative humidities greater than 60% measured by the rover environmental monitoring station ( rems ) @xcite .
this apparent correlation motivates us to consider the surface - atmosphere exchange as a major modulator for the atmospheric methane abundances in gale crater , the landing site of msl .
a high surface relative humidity may have a strong impact on the atmosphere - surface exchange because perchlorate salts in the regolith may deliquesce to form liquid solutions @xcite .
martian regolith contains @xmath60.5% perchlorate salts by weight , measured at both the phoenix and the curiosity landing sites @xcite .
they can form liquid solutions under martian conditions due to their low eutectic temperatures and low deliquescence relative humidities ( drh ) @xcite .
recent laboratory measurements have determined that the drh of calcium perchlorate is only @xmath7 at 198 k , and that the wet salt does not lose moisture until the relative humidity drops to @xmath8 @xcite . by comparing the eutectic temperature and the drh to the measured surface temperature and relative humidity , we confirm that perchlorate salts in the subsurface of gale crater may deliquesce in the southern winter , in which the methane spikes were measured .
we therefore formulate three hypotheses in an attempt to explain the variability of the atmospheric methane abundance at gale crater .
the first hypothesis ( hypothesis i ) is that the regolith in gale crater adsorbs methane when dry and releases this methane to the atmosphere when the relative humidity in the regolith is high enough for perchlorate salts to deliquesce .
the second hypothesis ( hypothesis ii ) is that microorganisms convert organic matter in the soil to methane when they are in liquid solutions .
this scenario does not require regolith adsorption .
the third hypothesis ( hypothesis iii ) is that deep subsurface aquifers produce the bursts of methane .
the first two hypotheses explain the methane spikes as signatures of cyclic processes that produce fast source and sink , whereas the third hypothesis explains the methane spikes as local emission of new methane .
we model the subsurface temperature and humidity of dusty terrains at gale crater by solving diffusion equations of heat and moisture with the hourly - averaged rems data as the boundary conditions @xcite .
the absolute humidity of each measurement is computed from the measured relative humidity and the sensor temperature archived in nasa s planetary data system ( pds ; https://pds.nasa.gov/ ) , with the saturation vapor pressure from @xcite .
the uncertainty of each surface temperature measurement is @xmath9k at low temperatures @xcite , and the uncertainty of each relative humidity measurement is @xmath10% @xcite .
the combined uncertainty of the hourly - averaged values is one - order - of - magnitude smaller for gaussian statistics .
the data clearly shows the effect of the rover traversing diverse terrains , as the magnitudes of the diurnal variations in both temperature and humidity change significantly @xcite . to remove this effect , we use daily means for a mars year ( from sol 39 to sol 707 ) and
add the hourly means from sol 70 to sol 89 ( when the rover is at rocknest , a dusty terrain ) to the daily means .
we use the diurnal variation measured over a dusty terrain as a proxy for regolith diurnal temperature and moisture variability .
we neglect the seasonal change in the magnitude of the diurnal variation in temperature and humidity , which is valid given the equatorial location of gale crater .
m.,scaledwidth=80.0% ] the rems relative humidity data in the daytime are thought to be unreliable , because the daytime relative humidity is often lower than 2% @xcite .
multiplied by the saturation vapor concentration , the rems daytime relative humidity measurements imply atmospheric water concentrations up to @xmath11 kg m@xmath12 .
this water concentration appears to be too large compared with the column water abundances in the equatorial region .
the mini - tes instrument on spirit and opportunity rovers measured daytime water columns seasonally variable from a few to 20 pr .
@xmath13 m @xcite .
the tes instrument onboard mars global surveyor also indicated a peak daytime water column of 20 pr .
@xmath13 m @xcite .
using a boundary layer thickness of 5 km @xcite , this column corresponds to a concentration of @xmath14 kg m@xmath12 at the surface .
if the boundary layer has a thickness of 2 km , the surface concentration can be up to @xmath15 kg m@xmath12 .
however , what if the rems daytime data are actually right ?
a detailed inspection reveals that the several hundred measurements in each hour in the daytime are not consistent with a random noise around zero .
the standard deviation of each group of measurements is small enough to statistically declare a non - zero mean value at @xmath16 or higher .
of course , the standard deviation may be artificially small as a result of the data processing pipeline .
we therefore consider all these possibilities of daytime water concentration in the analysis : for case l ( low water concentration ) we impose an upper limit of @xmath14 kg m@xmath12 on the rems data ; for case m ( intermediate water concentration ) we impose an upper limit of @xmath15 kg m@xmath12 ; and for case h ( high water concentration ) we use the rems data as reported in the pds .
the thermal and moisture diffusion equations are solved on a 41-level grid from zero to 5.2 m ( i.e. , four thermal skin depths ) .
the size of layers increases with depth , and the smallest size is 0.01 cm at the shallow subsurface
. we use the thermal inertia derived for rocknest @xcite , and the volumetric heat capacity for the martian aeolian dunes @xcite .
figure [ temp ] shows the modeled subsurface temperature .
the moisture diffusion equation ( i.e. , eq . 13 in @xcite ) includes physical adsorption and desorption by regolith , condensation , and corrections due to soil porosity and tortuosity @xcite .
we assume a soil porosity of 0.5 , and a porosity / tortuosity ratio of 0.3 @xcite . physical adsorption and desorption is assumed to be in equilibrium , and follows a langmuir - freundlich isotherm measured for palagonite , a terrestrial analog of martian basaltic soil @xcite .
the effect of physical adsorption and desorption is to retard water vapor diffusion by several orders of magnitude @xcite .
figure [ moist ] shows the model subsurface relative humidity for three assumptions of the daytime water concentration .
the diurnal variation of relative humidity extends to a depth of @xmath17 cm , on the same order of magnitude as the diurnal temperature variation . below the depth of diurnal variation ,
the relative humidity is constantly below 20% , except for the southern winter during which the relative humidty can be as high as 50% within the depth of 1 m for the high water concentration scenario ( figure [ moist ] ) .
further inspection of the results reveals that the absolute humidty below the depth of diurnal variation actually has little seasonal variation , consistent with the picture that water vapor diffusion is significantly retarded by physical adsorption .
therefore , the seasonally high subsurface relative humdity is driven by the seasonally low subsurface temperature as shown in figure [ temp ] .
furthermore , a zoom - in view of the relative humidity variation indicates that condensation can occur near the surface early morning during the southern winter ( figure [ moist ] ) , a finding consistent with @xcite .
the modeled temperature and relative humidity are then compared with the experimentally measured drh for perchlorate @xcite to determine whether deliquescence occurs ( figure [ deli ] ) .
we assume calcium perchlorate because both phoenix and curiosity identify calcium perchlorate as the likely form of the parent salt @xcite .
we find that the surface temperature and humidity conditions allow deliquescence to occur in the top @xmath18 cm of soil , each sol before sunrise and after sunset ( figure [ deli ] ) .
the liquid solution does not persist over diurnal cycles close to the surface .
the depth range of the transient liquid solution varies with the surface daily mean temperature and is largest in the southern winter . when sam s methane measurements were taken , with the blue arrows indicating background measurements and the red arrows indicating the spike measurements .
, scaledwidth=50.0% ] for a conservative daytime water concentration of @xmath14 kg m@xmath12 or lower ( case l ) , deliquescence only occurs in the shallow subsurface and results in transient solutions .
this is consistent with the findings of @xcite , in which they omitted all daytime relative humidity measurements .
nonetheless , if the daytime water concentration is indeed higher , our calculation indicates that persisting liquid solutions could occur below the shallow subsurface in the southern winter , starting from @xmath19 is the mars - sun angle , measured from the northern hemisphere spring equinox where @xmath20 . for the southern hemisphere , @xmath21 corresponds to winter solstice , and @xmath22 corresponds to summer solstice . ]
( figure [ deli ] , cases m and h ) . the range of the liquid solution is larger for a higher water concentration . because of a small diffusion coefficient , the absolute humidity of the soil at depths is roughly constant .
the deep liquid solution is formed by slow cooling of the deep soil and the increase of the relative humidity .
the liquid solution persists to @xmath23 , longer than the high relative humidity season , manifesting hysteresis .
this scenario assumes that methane is normally adsorbed by the regolith . however , when deliquescence occurs , the liquid solutions of perchlorate that form can conceivably coat the soil particles and deactivate most active sites , releasing methane into the atmosphere .
wetted particles have much smaller surface area than dry particles .
10-@xmath13 m round - shape particles of basaltic composition only have surface area of @xmath24 m@xmath25 g@xmath0 , and this is even smaller for larger particles .
dry regolith on mars , however , has surface area @xmath26 m@xmath25 g@xmath0 @xcite .
it is therefore plausible to hypothesize that wetting could effectively remove surface area and displace adsorbed methane .
the methane is then mixed in the planetary boundary layer , and may be transported and re - adsorbed by the regolith elsewhere . taking the boundary layer as a box ( e.g. , 33% of the column mass for a boundary layer thickness of 5 km ) , the column mass of methane in the box ( @xmath27 ) is modeled by the following equation , @xmath28 @xmath29 is the local atmospheric residence time of methane that brackets removal of methane by diurnal thermal tides , horizontal advection and diffusion , and soil adsorption .
the horizontal advection and diffusion can be a sink in this model because gale crater is not fully covered with thick regolith , so spreading methane emitted from regions of thick regolith across gale crater can greatly reduce the local methane concentration .
@xmath29 is taken as a free parameter , because the footprint of the methane emission is unknown .
@xmath30 is the flux of methane to maintain the 0.7-ppb background level of methane , and is equal to @xmath31 .
@xmath30 is invoked to complete the equation and it represents the partitioning of methane between the atmosphere and the regolith .
equation ( [ ad ] ) can be rewritten using the departure from the background level @xmath32 as the variable and without the @xmath30 term .
once deliquescence occurs , the displaced methane is assumed to be released to the atmosphere immediately .
this is probably close enough to reality because the deliquescence starts from the shallow subsurface and progresses deeper , because solubility of methane in salty water is negligible , and because the diffusivity of methane is fast enough such that it should quickly reach the surface @xcite .
the flux of methane produced by the onset of deliquescence is @xmath33 where @xmath34 kg m@xmath12 is soil density , @xmath35 m@xmath36 is the monolayer coverage of methane per unit surface area @xcite , @xmath37 is the mass of methane molecule , @xmath38 is the coverage ratio of the methane , and @xmath39 m@xmath25 g@xmath0 is the specific surface area .
the specific surface area adopted in the study is estimated based on in situ measurements of viking @xcite , or measurements of jsc mars-1 analog @xcite .
the specific surface area could be as high as 1000 m@xmath25 g@xmath0 if the soil was mainly made of clay minerals @xcite , which is not seen in gale crater .
@xmath38 is calculated from the langmuir isotherm , @xmath40 where @xmath41 is the number density of methane in the gas phase , and @xmath42 where @xmath43 is the thermal velocity of methane , @xmath44 is the planck constant , @xmath45 is the boltzmann constant , @xmath46 is the gas constant , and @xmath47 is the adsorption energy .
@xmath48 is the number density of and @xmath49 is the equilibrium constant for adsorption .
the equation ( [ langmuir ] ) is applicable with being the sole competing substance for adsorption , because ice can not form in gale crater , ice may form but only at the top @xmath50 cm of soil during nighttime , and inhomogeneous grains , like perchlorate grains , only make less than 1% of soil mass . given that in any case @xmath51 , equation ( [ langmuir ] ) is reduced to @xmath52 where @xmath53 is the coverage ratio of , @xmath54 we adopt the @xmath53 measurement for palagonite under martian conditions fitted to the format of equation ( [ langmuirc ] ) by @xcite .
@xmath47 has been measured over the martian soil analog jsc - mars-1 and the measured value is @xmath55 kj mol@xmath0 @xcite .
we however postulate that the detail surface properties could matter for the amount of methane adsorption , specified by @xmath47 in our formulation . to account for model uncertainties , including surface properties , advection , diffusion and the effect of topography on diurnal accumulation , we treat @xmath47 and @xmath29 as free parameters .
we solve equations ( [ ad]-[dflux ] ) to determine whether methane adsorption and desorption would contribute to the variability seen in gale crater .
figure [ fit1 ] shows the best - fit models for case l ( low water concentration ) and case h ( high water concentration ) .
we only include the day - to - day change in the range of deliquescence when evaluating equation ( [ dflux ] ) .
this is because adsorption of methane is kinetically slow @xcite and a diurnal deliquescence - efflorescence cycle would not allow the soil to be loaded by methane when dry .
we hypothesize that the seasonal variation in the maximum depth of deliquescent soil ( figure [ deli ] ) could be the main driver of methane variability in the atmosphere .
figure [ fit1 ] shows that deliquescence - induced desorption of methane can account for the methane spikes .
the season in which deliquescence progresses to deeper subsurface , in both low and high water cases , is consistent with the season in which the methane spikes are measured ( i.e. , the beginning of southern winter ) . the methane released from the regolith can quickly dissipate , consistent with the rapid drop - off at @xmath56 , for a residence time of a few days .
a small residence time indicates a small footprint of the methane source in gale crater , and fast re - adsorption of methane to the regolith . this model does not predict a methane spike at @xmath57
however , given the low significance of the measurement , the overall goodness of fit achieved by this hypothesis ( @xmath58 ) is satisfactory . , and a residence time of 12 days .
the goodness of fit is @xmath59 .
the best - fit model of case h ( high water concentration ) requires an adsorption energy of 36.4 kj mol@xmath0 , and a residence time of 4 days .
the goodness of fit is @xmath60 . for hypothesis
ii , none of the models produce a good fit because of the high flux at @xmath56 .
we instead show two examples for illustration : both cases assume a baseline residence time of soil carbon of 30 days ; the case l assumes @xmath29 of 100 sols and the case h assumes @xmath29 of 10 sols . ,
scaledwidth=80.0% ] the main challenge of this model is that the adsorption energy required is 2-fold greater than what is measured by @xcite .
this is because a sufficient amount of methane needs to be temporarily stored in the upper 10 cm ( case l ) or 60 cm ( case h ) of the soil layer to account for the 7-ppb spike .
increasing the residence time ( e.g. , enlarging the footprint ) could help reduce the adsorption energy , but the residence time is limited by the firm non - detection at @xmath56 .
how is the required adsorption energy compared with that of methane for other materials ?
existing measurements , often performed under room or higher temperatures , suggest a fairly wide range of adsorption energy : 15 - 25 kj mol@xmath0 for various kinds of zeolites @xcite , 16 - 21 kj mol@xmath0 for activated carbons @xcite , and 39 kj mol@xmath0 for a synthetic nano - porous titanium silicate @xcite .
further laboratory studies are warrented to determine whether the high adsorption energy required by this scenario is possible for martian regolith .
this scenario postulates extant microorganisms in the regolith , and assumes that in the presence of liquid water and organic compounds , these microorganisms are able to produce methane . like hypothesis i , this hypothesis is based on the seasonal deliquescence - efflorescence cycle , but it does not require a large adsorption energy .
this hypothesis is motivated by the recent discovery of organic matter in drill samples of a mudstone in the yellowknife bay region of gale crater @xcite .
the detected organic matter includes trace - level chlorobenzene ( ) and dichloroalkanes ( ) in the evolved gases , whose precursors are suggested to be benzenecarboxylates and aliphatic hydrocarbons .
benzene rings are hard to break , but one may imagine microorganisms instead use the aliphatic hydrocarbons by performing the following hydrogenation reaction which is exothermic ( @xmath61 kj mol@xmath0 ) .
alternatively , methanogens could directly source carbon from the atmosphere the source of hydrogen for this metabolism is photochemical dissociation of water vapor in the atmosphere @xcite .
we imagine yeast - like microorganisms performing the biochemical reactions above in liquid solutions formed by deliquescence of perchlorate salts . on earth
, methanogenesis occurs as a result of oxygen - depleted decomposition of organic c compounds , and consume acetate ( ) to produce and , or and to produce and @xcite .
the present - day production of predominantly occurs in water - logged , oxygen - depleted ( anaerobic ) environments .
temperature is a limiting factor to methanogenesis : emissions from soils have been found to vary exponentially as a function of temperature @xcite .
other factors that influence the production of in oxygen - depleted environments include soil ph and redox potential @xcite .
life is known to exist in highly salty and highly saline environments on earth @xcite .
however , it is doubtful that any terrestrial microorganisms , even those living in highly saline environments , can live in an eutectic brine of ca perchlorate on mars , because the water activity on mars would be too low @xcite . phenomenologically , and as a bold assumption , we consider the possibility that martian life has developed biological mechanism that allows functioning in such a harsh environment .
assuming methanogenesis occurs in gale crater , we model @xmath62 here , @xmath29 includes not only horizontal transport and physical adsorption , but also conversion from methane back to organic matter , probably by methanotrophs ( see figure [ bio_schematic ] ) , given that oxidation in aerobic and anaerobic environments is a source of c and energy for terrestrial methanotrophs @xcite .
@xmath63 is the methane flux produced by methanogens , which is modeled as @xmath64 where @xmath65 kg / kg is the soil content of aliphatic hydrocarbons , @xmath66 is the baseline residence time of organic carbon at 0 @xmath67c , and @xmath68 is an exponential temperature dependence constant . equation ( [ t_analog ] ) implies that @xmath66 , together with the temperature and the liquid water availability in the soil , determines the residence time of organic carbon in the soil .
the measured and modeled biological residence times of carbon in anaerobic environments span across several orders of magnitude ( from 30 days for root exudates to @xmath69 days for soil carbon @xcite ) .
we choose @xmath70 , broadly consistent with terrestrial biological methane emission studies @xcite .
the deliquescence - enabled methanogenesis can produce the methane spikes during the southern winter as measured by msl , but can not produce the rapid drop - off at @xmath56 ( figure [ fit1 ] ) .
compared with hypothesis i , the methane spikes predicted by the methanogenesis model are later in the season .
this is simply because the methane emission flux here is proportional to the thickness of the deliquescent soil , while the methane flux due to desorption is proportional to the derivative of that .
for the same reason , this hypothesis is not directly consistent with the firm non - detection of elevated methane at @xmath56 or @xmath71 , during which the subsurface soil remains wet .
it is however possible that msl was not downstream of the methane source during this period . ) and the baseline residence time of soil carbon ( @xmath66 ) that produce a @xmath72 ppb methane spike at @xmath56 as measured by msl .
, scaledwidth=60.0% ] a high surface water concentration , and then a large thickness of deliquescent soil during winter is probably necessary for methanogenesis to produce the 7-ppb methane spikes .
figure [ fit2 ] shows the parameter ranges that produce the strength of the spikes .
additionally , one needs to consider replenishment of organic carbon in the soil .
the methane production rate is controlled by the baseline residence time of soil carbon ( @xmath66 ) , which is a free parameter in the model .
the atmospheric residence time of methane ( @xmath29 ) needs to be shorter than the lifetime of the organic carbon in the soil to ensure replenishment , if the main source of carbon of the methanogens is the soil organic carbon ( reaction [ r2 ] ) . in the model
we find this condition corresponds to @xmath73 , shown as a dashed line in figure [ fit2 ] .
using this condition , we find that it is only marginally possible to produce the 7-ppb methane spikes in case l ( low water concentration ) , while producing the methane spikes in case h ( high water concentration ) is much easier .
if the methanogens directly source carbon from the co@xmath74 atmosphere ( reaction [ r1 ] ) , the atmospheric residence time of methane can be longer and the methane spikes are more readily fit by this scenario .
another possibility is that the elevated methane measured by tls represents sources of `` new '' methane emerging from a deep subsurface aquifer .
the sources of methane may include subsurface gas - water - rock chemistry and microbial methanogenesis @xcite , with the difference being that biological methanogenesis is much faster than gas - water - rock reactions @xcite .
such a source of methane , along with cometary and meteoric sources , may be necessary to balance the loss of atmospheric methane due to uv photolysis over a time scale of a few hundred years .
a subsurface aquifer may exist at a depth of 5 km assuming a geothermal gradient of 10 k / km @xcite .
the aquifer can be partially sealed by an ice or clathrate layer to produce bursts of methane , in similar ways as the terrestrial arctic tundra .
the arctic tundra is one of the major sources of methane for earth s atmosphere .
concentrated bursts of methane have been observed at tundra sites in late fall as the seasonally thawed active layer refreezes , forcing sub - surface methane into the atmosphere @xcite , and during the spring freeze - thaw transition when subsurface methane trapped by the frozen surface escapes @xcite .
the fall and spring bursts are transitory ( occurring only for a @xmath75-day window immediately surrounding the freeze - thaw transition ) and episodic ( they do not occur every season ) , and the magnitude of the emissions is highly variable . on mars , however , if the methane was released deep from the subsurface , this release should be sporadic and have no seasonality .
this is because the seasonal variations of temperature and relative humidity damp out at a depth of a few meters ( figure [ temp ] ) , and the condition of the permanent aquifer is not controlled by surface conditions .
other geological processes may trigger the subsurface methane to be vented out , such as seepage from mud volcanoes ( e.g. , * ? ? ?
* ; * ? ? ?
the source of methane emission in this scenario must be localized , because on the mass balance calculation in 1 .
a source of methane close to msl that emits @xmath2 ppb of methane over a small area could be the cause of the methane spike , while the subsequent decrease back to the background level would be easily explained by the dispersal of the methane into the rest of the atmosphere .
the small amount of methane released would have very little effect on the carbon budget of the rest of the atmosphere .
it is an open question what areas of mars would be amenable to such localized methane releases .
we propose the first working theories that explain the sub - annual variability of atmospheric methane abundance in gale crater , in the form of three testable hypotheses .
first , adsorption of methane in the regolith up to a few tens cm depth , and its release driven by deliquescence during the winter , can explain the apparent methane variability .
this scenario requires an adsorption energy 2-fold higher than laboratory measurements on mars regolith analogs .
second , biological conversion from organic matter to methane in the regolith by microorganisms in perchlorate solutions can produce the magnitude of the methane spikes .
however , this model can not produce the rapid drop - off of the methane abundance at @xmath56 , and likely requires a daytime water concentration higher than standard mars atmospheric models .
third , the methane spikes may come from a deep subsurface aquifer , representing new methane to the atmosphere .
emitted intermittently , this methane may be the methane that maintains the 0.7-ppb background on a time scale of several hundred years , if the footprint of the emission is small .
any of the three hypotheses , if confirmed , leads to profound ramification in our understanding of mars as an active and potentially habitable world .
if adsorption and desorption by regolith modulates atmospheric methane in gale crater , greater variability can be expected for higher latitudes , where perchlorate also exists @xcite and deliquescence may occur for a larger fraction of the surface @xcite .
the methane variability suggested on the basis of telescopic observations @xcite , though debated @xcite , could be explained in this way .
both the first and the second hypothesis work better when the daytime water concentration at the surface is high . if true , these scenarios warrant further investigation of the boundary layer dynamics and water vapor distribution on mars .
lastly , if the methane spikes are from the deep subsurface , one might ask how lucky msl has to be to observe the spikes and whether the geographic low of gale crater helps . continued monitoring of methane by msl will test these hypotheses .
the first two hypotheses predict the methane variability is seasonal and should repeat annually .
the peak of the methane abundance occurs in different seasons between the two hypotheses , with the adsorption - desorption model in the early winter and the methanogenesis model in the later winter .
we caution that observing the repeated signals , even if either of the hypotheses is true , can be hard because of the changing wind patterns and the rover location .
the last hypothesis predicts the methane spikes are sporadic .
meanwhile , our investigations of the three hypotheses call for laboratory studies of gas adsorption and desorption during deliquescence , as well as further laboratory studies of the adsorption energy of methane for fine , porous silicate materials . finally , future _ in situ _ mars exploration may conduct improved characterization of the daytime water vapor concentration at the surface and measurement of the water s isotopic signatures will further constrain the rate of the atmosphere - regolith exchange , and reveal the smoking gun of a near - surface methane reservoir .
support for this work was partially provided by nasa through hubble fellowship grant # 51332.01 awarded by the space telescope science institute , which is operated by the association of universities for research in astronomy , inc .
, for nasa , under contract nas 5 - 26555 .
yly was supported in part by an nai virtual planetary laboratory grant from the university of washington to the jet propulsion laboratory and california institute of technology .
the research was carried out at the jet propulsion laboratory , california institute of technology , under a contract with the national aeronautics and space administration .
84 natexlab#1#1url # 1`#1`urlprefix blamey , n. j. , parnell , j. , mcmahon , s. , mark , d. f. , tomkinson , t. , lee , m. , shivak , j. , izawa , m. r. , banerjee , n. r. , flemming , r. l. , 2015 .
evidence for methane in martian meteorites .
nature communications 6 .
cavenati , s. , grande , c. a. , rodrigues , a. e. , 2004 .
adsorption equilibrium of methane , carbon dioxide , and nitrogen on zeolite 13x at high pressures .
journal of chemical & engineering data 49 ( 4 ) , 10951101 . ,
f. h. , oneill , k. , bradley , p. m. , meth , b. a. , ciufo , s. a. , knobel , l. l. , lovley , d. r. , jan . 2002 . a hydrogen - based subsurface microbial community dominated by methanogens .
nature 415 , 312315 .
clifford , s. m. , lasue , j. , heggy , e. , boisson , j. , mcgovern , p. , max , m. d. , 2010 .
depth of the martian cryosphere : revised estimates and implications for the existence and detection of subpermafrost groundwater . journal of geophysical research : planets 115 ( e7 ) , n / a
n / a , e07001 .
etiope , g. , oehler , d. , allen , c. , 2011 .
methane emissions from earths degassing : implications for mars .
planetary and space science 59 ( 23 ) , 182 195 , methane on mars : current observations , interpretation and future plans .
freissinet , c. , glavin , d. p. , mahaffy , p. r. , miller , k. e. , eigenbrode , j. l. , summons , r. e. , brunner , a. e. , buch , a. , szopa , c. , archer , p. d. , franz , h. b. , atreya , s. k. , brinckerhoff , w. b. , cabane , m. , coll , p. , conrad , p. g. , des marais , d. j. , dworkin , j. p. , fairn , a. g. , franois , p. , grotzinger , j. p. , kashyap , s. , ten kate , i. l. , leshin , l. a. , malespin , c. a. , martin , m. g. , martin - torres , f. j. , mcadam , a. c. , ming , d. w. , navarro - gonzlez , r. , pavlov , a. a. , prats , b. d. , squyres , s. w. , steele , a. , stern , j. c. , sumner , d. y. , sutter , b. , zorzano , m .-
p . , the msl science team , 2015 . organic molecules in the sheepbed mudstone , gale crater , mars .
journal of geophysical research : planets 120 ( 3 ) , 495514 . fries , m. , christou , a. , archer , d. , conrad , p. , cooke , w. , eigenbrode , j. , ten kate , i. l. , matney , m. , niles , p. , sykes , m. , steele , a. , treiman , a. , aug 2015 . a meteor shower origin for martian methane .
meteoritics & planetary science 50 ( 1 , si ) , 78th annual meeting of the meteoritical - society , berkeley , ca , jul 27 - 31 , 2015 .
glavin , d. p. , freissinet , c. , miller , k. e. , eigenbrode , j. l. , brunner , a. e. , buch , a. , sutter , b. , archer , p. d. , atreya , s. k. , brinckerhoff , w. b. , cabane , m. , coll , p. , conrad , p. g. , coscia , d. , dworkin , j. p. , franz , h. b. , grotzinger , j. p. , leshin , l. a. , martin , m. g. , mckay , c. , ming , d. w. , navarro - gonzlez , r. , pavlov , a. , steele , a. , summons , r. e. , szopa , c. , teinturier , s. , mahaffy , p. r. , 2013 .
evidence for perchlorates and the origin of chlorinated hydrocarbons detected by sam at the rocknest aeolian deposit in gale crater .
journal of geophysical research : planets 118 ( 10 ) , 19551973 .
gomez - elvira , j. , armiens , c. , carrasco , i. , genzer , m. , gmez , f. , haberle , r. , hamilton , v. e. , harri , a .- m . , kahanp , h. , kemppinen , o. , lepinette , a. , martn soler , j. , martn - torres , j. , martnez - fras , j. , mischna , m. , mora , l. , navarro , s. , newman , c. , de pablo , m. a. , peinado , v. , polkko , j. , rafkin , s. c. r. , ramos , m. , renn , n. o. , richardson , m. , rodrguez - manfredi , j. a. , romeral planell , j. j. , sebastin , e. , de la torre jurez , m. , torres , j. , urqu , r. , vasavada , a. r. , verdasca , j. , zorzano , m .-
p . , 2014 .
rover environmental monitoring station : overview of the first 100 sols .
journal of geophysical research : planets 119 ( 7 ) , 16801688 .
gomez - elvira , j. , armiens , c. , castaner , l. , domnguez , m. , genzer , m. , gmez , f. , haberle , r. , harri , a .- m .
, jimnez , v. , kahanp , h. , kowalski , l. , lepinette , a. , martn , j. , martnez - fras , j. , mcewan , i. , mora , l. , moreno , j. , navarro , s. , de pablo , m. , peinado , v. , pea , a. , polkko , j. , ramos , m. , renno , n. , ricart , j. , richardson , m. , rodrguez - manfredi , j. , romeral , j. , sebastin , e. , serrano , j. , dela torre jurez , m. , torres , j. , torrero , f. , urqu , r. , vzquez , l. , velasco , t. , verdasca , j. , zorzano , m .-
, martn - torres , j. , 2012 .
rems : the environmental sensor suite for the mars science laboratory rover .
space science reviews 170 ( 1 - 4 ) , 583640 .
gough , r. v. , tolbert , m. a. , mckay , c. p. , toon , o. b. , 2010 .
methane adsorption on a martian soil analog : an abiogenic explanation for methane variability in the martian atmosphere .
icarus 207 ( 1 ) , 165 174 .
harri , a .- m . ,
genzer , m. , kemppinen , o. , gomez - elvira , j. , haberle , r. , polkko , j. , savijrvi , h. , renn , n. , rodriguez - manfredi , j. a. , schmidt , w. , richardson , m. , siili , t. , paton , m. , torre - juarez , m. d. , mkinen , t. , newman , c. , rafkin , s. , mischna , m. , merikallio , s. , haukka , h. , martin - torres , j. , komu , m. , zorzano , m .-
p . , peinado , v. , vazquez , l. , urqui , r. , 2014 . mars science laboratory relative humidity observations : initial results .
journal of geophysical research : planets 119 ( 9 ) , 21322147 .
hecht , m. h. , kounaves , s. p. , quinn , r. c. , west , s. j. , young , s. m. m. , ming , d. w. , catling , d. c. , clark , b. c. , boynton , w. v. , hoffman , j. , deflores , l. p. , gospodinova , k. , kapit , j. , smith , p. h. , 2009 .
detection of perchlorate and the soluble chemistry of martian soil at the phoenix lander site .
science 325 ( 5936 ) , 6467 .
himeno , s. , tomita , t. , suzuki , k. , yoshida , s. , 2007 . characterization and selectivity for methane and carbon dioxide adsorption on the all - silica \{dd3r } zeolite .
microporous and mesoporous materials 98 ( 13 ) , 62 69 .
komatsu , g. , ori , g. g. , cardinale , m. , dohm , j. m. , baker , v. r. , vaz , d. a. , ishimaru , r. , namiki , n. , matsui , t. , 2011 .
roles of methane and carbon dioxide in geological processes on mars .
planetary and space science 59 ( 23 ) , 169 181 , methane on mars : current observations , interpretation and future plans .
kounaves , s. p. , chaniotakis , n. a. , chevrier , v. f. , carrier , b. l. , folds , k. e. , hansen , v. m. , mcelhoney , k. m. , oneil , g. d. , weber , a. w. , 2014 .
identification of the perchlorate parent salts at the phoenix mars landing site and possible implications .
icarus 232 ( 0 ) , 226 231 .
leshin , l. a. , mahaffy , p. r. , webster , c. r. , cabane , m. , coll , p. , conrad , p. g. , archer , p. d. , atreya , s. k. , brunner , a. e. , buch , a. , eigenbrode , j. l. , flesch , g. j. , franz , h. b. , freissinet , c. , glavin , d. p. , mcadam , a. c. , miller , k. e. , ming , d. w. , morris , r. v. , navarro - gonzlez , r. , niles , p. b. , owen , t. , pepin , r. o. , squyres , s. , steele , a. , stern , j. c. , summons , r. e. , sumner , d. y. , sutter , b. , szopa , c. , teinturier , s. , trainer , m. g. , wray , j. j. , grotzinger , j. p. , team , m. s. , 2013 .
volatile , isotope , and organic analysis of martian fines with the mars curiosity rover .
science 341 ( 6153 ) .
martinez , g. m. , renn , n. , fischer , e. , borlina , c. s. , hallet , b. , de la torre jurez , m. , vasavada , a. r. , ramos , m. , hamilton , v. , gomez - elvira , j. , haberle , r. m. , 2014 .
surface energy budget and thermal inertia at gale crater : calculations from ground - based measurements .
journal of geophysical research : planets 119 ( 8) , 18221838 .
martn - torres , f. j. , zorzano , m .- p .
, valentn - serrano , p. , harri , a .- m . , genzer , m. , kemppinen , o. , rivera - valentin , e. g. , jun , i. , wray , j. , 2015 .
transient liquid water and water activity at gale crater on mars .
nature geoscience 8 ( 5 ) , 361 .
ming , d. w. , archer , p. d. , glavin , d. p. , eigenbrode , j. l. , franz , h. b. , sutter , b. , brunner , a. e. , stern , j. c. , freissinet , c. , mcadam , a. c. , mahaffy , p. r. , cabane , m. , coll , p. , campbell , j. l. , atreya , s. k. , niles , p. b. , bell , j. f. , bish , d. l. , brinckerhoff , w. b. , buch , a. , conrad , p. g. , des marais , d. j. , ehlmann , b. l. , fairn , a. g. , farley , k. , flesch , g. j. , francois , p. , gellert , r. , grant , j. a. , grotzinger , j. p. , gupta , s. , herkenhoff , k. e. , hurowitz , j. a. , leshin , l. a. , lewis , k. w. , mclennan , s. m. , miller , k. e. , moersch , j. , morris , r. v. , navarro - gonzlez , r. , pavlov , a. a. , perrett , g. m. , pradler , i. , squyres , s. w. , summons , r. e. , steele , a. , stolper , e. m. , sumner , d. y. , szopa , c. , teinturier , s. , trainer , m. g. , treiman , a. h. , vaniman , d. t. , vasavada , a. r. , webster , c. r. , wray , j. j. , yingst , r. a. , team , m. s. , 2014 .
volatile and organic compositions of sedimentary rocks in yellowknife bay , gale crater , mars .
science 343 ( 6169 ) .
miyajima , t. , wada , e. , hanba , y. t. , vijarnsorn , p. , 1997 .
anaerobic mineralization of indigenous organic matters and methanogenesis in tropical wetland soils .
geochimica et cosmochimica acta 61 ( 17 ) , 37393751 . , m. j. , villanueva , g. l. , novak , r. e. , hewagama , t. , bonev , b. p. , disanti , m. a. , mandell , a. m. , smith , m. d. , feb .
2009 . strong release of methane on mars in northern summer 2003 .
science 323 , 1041. nuding , d. , rivera - valentin , e. , davis , r. , gough , r. , chevrier , v. , tolbert , m. , 2014 .
deliquescence and efflorescence of calcium perchlorate : an investigation of stable aqueous solutions relevant to mars .
icarus 243 ( 0 ) , 420 428 . , j. d. , beaty , d. w. , jones , m. a. , bakermans , c. , barlow , n. g. , boston , p. j. , chevrier , v. f. , clark , b. c. , de vera , j .-
p. , gough , r. v. , hallsworth , j. e. , head , j. w. , hipkin , v. j. , kieft , t. l. , mcewen , a. s. , mellon , m. t. , mikucki , j. a. , nicholson , w. l. , omelon , c. r. , peterson , r. , roden , e. e. , sherwood lollar , b. , tanaka , k. l. , viola , d. , wray , j. j. , nov .
2014 . a new analysis of mars `` special regions '' : findings of the second mepag special regions science analysis group ( sr - sag2 ) .
astrobiology 14 , 887968 .
smith , m. d. , wolff , m. j. , spanovich , n. , ghosh , a. , banfield , d. , christensen , p. r. , landis , g. a. , squyres , s. w. , 2006 .
one martian year of atmospheric observations using mer mini - tes .
journal of geophysical research : planets 111 ( e12 ) , n / a
n / a , e12s13 .
solomon , s. c. , head , j. w. , 1990 .
heterogeneities in the thickness of the elastic lithosphere of mars : constraints on heat flow and internal dynamics .
journal of geophysical research : solid earth 95 ( b7 ) , 1107311083 .
song , c. , xu , x. , sun , x. , tian , h. , sun , l. , miao , y. , wang , x. , guo , y. , 2012 . large methane emission upon spring thaw from natural wetlands in the northern permafrost region .
environmental research letters 7 ( 3 ) , 034009 .
squyres , s. w. , arvidson , r. e. , ruff , s. , gellert , r. , morris , r. v. , ming , d. w. , crumpler , l. , farmer , j. d. , marais , d. j. d. , yen , a. , mclennan , s. m. , calvin , w. , bell , j. f. , clark , b. c. , wang , a. , mccoy , t. j. , schmidt , m. e. , de souza , p. a. , 2008 .
detection of silica - rich deposits on mars .
science 320 ( 5879 ) , 10631067 .
treude , t. , orphan , v. , knittel , k. , gieseke , a. , boetius , a. , 2007 .
consumption of methane and co2 by methanotrophic microbial mats from gas seeps of the anoxic black sea . applied and environmental microbiology 73 ( 7 ) , 22712283 .
tung , h. c. , bramall , n. e. , price , p. b. , 2005 . microbial origin of excess methane in glacial ice and implications for life on mars .
proceedings of the national academy of sciences of the united states of america 102 ( 51 ) , 1829218296 .
van hulzen , j. , segers , r. , van bodegom , p. , leffelaar , p. , 1999 .
temperature effects on soil methane production : an explanation for observed variability . soil biology and biochemistry 31 ( 14 ) , 19191929 .
wania , r. , melton , j. , hodson , e. , poulter , b. , ringeval , b. , spahni , r. , bohn , t. , avis , c. , chen , g. , eliseev , a. , et al . , 2013 .
present state of global wetland extent and wetland methane modelling : methodology of a model inter - comparison project ( wetchimp ) .
geoscientific model development 6 ( 3 ) , 617641 .
wania , r. , ross , i. , prentice , i. , 2010 .
implementation and evaluation of a new methane model within a dynamic global vegetation model : lpj - whyme v1 .
geoscientific model development 3 ( 2 ) , 565584 . webster , c. r. , mahaffy , p. r. , atreya , s. k. , flesch , g. j. , mischna , m. a. , meslin , p .- y . ,
farley , k. a. , conrad , p. g. , christensen , l. e. , pavlov , a. a. , martn - torres , j. , zorzano , m .-
mcconnochie , t. h. , owen , t. , eigenbrode , j. l. , glavin , d. p. , steele , a. , malespin , c. a. , archer , p. d. , sutter , b. , coll , p. , freissinet , c. , mckay , c. p. , moores , j. e. , schwenzer , s. p. , bridges , j. c. , navarro - gonzalez , r. , gellert , r. , lemmon , m. t. , the msl science team , 2015 . mars methane detection and variability at gale crater .
science 347 ( 6220 ) , 415417 .
yvon - durocher , g. , allen , a. p. , bastviken , d. , conrad , r. , gudasz , c. , st - pierre , a. , thanh - duc , n. , del giorgio , p. a. , 2014 .
methane fluxes show consistent temperature dependence across microbial to ecosystem scales .
nature 507 ( 7493 ) , 488491 .
zent , a. p. , quinn , r. c. , 1995 .
simultaneous adsorption of co2 and h2o under mars - like conditions and application to the evolution of the martian climate .
journal of geophysical research : planets 100 ( e3 ) , 53415349 . | the curiosity rover recently detected a background of 0.7 ppb and spikes of 7 ppb of methane on mars .
this in situ measurement reorients our understanding of the martian environment and its potential for life , as the current theories do not entail any geological source or sink of methane that varies sub - annually . in particular , the 10-fold elevation during the southern winter
indicates episodic sources of methane that are yet to be discovered .
here we suggest a near - surface reservoir could explain this variability . using the temperature and humidity measurements from the rover , we find that perchlorate salts in the regolith deliquesce to form liquid solutions , and deliquescence progresses to deeper subsurface in the season of the methane spikes .
we therefore formulate the following three testable hypotheses .
the first scenario is that the regolith in gale crater adsorbs methane when dry and releases this methane to the atmosphere upon deliquescence .
the adsorption energy needs to be 36 kj mol@xmath0 to explain the magnitude of the methane spikes , higher than existing laboratory measurements .
the second scenario is that microorganisms convert organic matter in the soil to methane when they are in liquid solutions .
this scenario does not require regolith adsorption , but entails extant life on mars .
the third scenario is that deep subsurface aquifers produce the bursts of methane
. continued _ in situ _ measurements of methane and water , as well as laboratory studies of adsorption and deliquescence , will test these hypotheses and inform the existence of the near - surface reservoir and its exchange with the atmosphere .
mars , methane , astrobiology , regolith |
according to bell s theorem [ 1 ] , one can not construct a local realistic hidden - variable theory that can reproduce all the statistical predictions of quantum mechanics for a two - particle singlet - state . here , by the term locality
we mean bell s locality condition which in a local stochastic hidden - variable theory is equivalent to the statistical independence of the values of the spin components of the two particles [ 2 ] .
bell s locality condition at the level of hidden variables also implies a more general tenet named the `` common cause principle '' [ 3 ] . according to this principle
, one can explain how two spatially separated particles can be correlated , when there is no direct causal relationship between them . despite the fact that one can not decisively close the door to other potential alternatives which may provide a better understanding of the so - called bell - type correlations in a causal fashion [ 4 ] , the common cause description is still a very attractive and comprehensible way of understanding the nature of these correlations [ 5 ] .
but , thanks to bell s theorem , we do know that this description can not be compatible with the predictions of quantum mechanics .
so , we are still facing the crucial problem of whether correlations resulting from the entanglement of two spatially separated particles are an essential characteristic of microphysical reality .
what is then the nature of these correlations ? could they still be explained by invoking a field of common cause in the past light cones of the two separated particles ? the main purpose of this article is to give a positive answer to the last question . here , we want to provide a stochastic hidden - variable model which honors the common cause principle but can not be ruled out by bell s theorem .
the key point here is that it is a local _ contextual _ model . by the term contextual we mean that each probability function is defined in a measurement setup where a definite set of hidden variables is used .
this approach is very similar to what has been recently suggested by de muynck [ 6 ] . here
, we are going to provide a more concrete realization of what de muynck has proposed as contextual reality which might be accomplished by some hidden variables in a subquantum theory . in section 2 ,
we first review the common cause principle in the domain of probabilistic causality .
then , in section 3 , we construct a local contextual hidden - variable model which confirms the quantum mechanical predictions . in section 4 , we explain why our proposed model is local and contextual .
subsequently , we talk about an alternative classical example in section 5 which can make the subtleties of our argument more transparent . we have summarized our results in the last section .
it seems better to explain common cause principle through the term spurious correlation .
eells defines the usual understanding of this term in the following way [ 7 ] : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [ t]wo factors are spuriously correlated when ( roughly ) neither causes the other and the correlation disappears when a third variable is introduced and held fixed- that is , the correlation disappears both in the presence and in the absence of the third factor_. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ if a correlation between factors @xmath0 and @xmath1 disappears both in the presence and in the absence of third factor @xmath2 , then the explanation may be either that the correlation results from the joint causal effect of @xmath2 on @xmath0 and @xmath3 ( @xmath2 is a common cause of @xmath0 and @xmath1 ) or that @xmath2 is an intermediate causal factor between @xmath0 and @xmath1 ( @xmath0 operates on @xmath1 through @xmath2 or @xmath1 operates on @xmath0 through @xmath2 ) . since one of @xmath0 and @xmath1 may in fact be a genuine positive causal factor for the other in the second case , we do not consider it as a case of spurious correlation .
so , as it is indicated by eells , two factors @xmath0 and @xmath1 are spuriously correlated if neither causes the other and the correlation of @xmath1 with @xmath0 disappears when @xmath2 is held fixed .
now , suppose that , in the simplest kind of a common cause pattern , @xmath2 is a common cause of both @xmath0 and @xmath4 then , the following relations hold for the various kinds of probabilities : @xmath5 @xmath6 @xmath7 @xmath8 @xmath9 @xmath10 where @xmath11 denotes the negation .
relations ( 1 ) and ( 2 ) correspond to the assumption that @xmath2 is a common cause of @xmath0 and @xmath1 ; ( 3 ) and ( 4 ) say that the correlation between @xmath0 and @xmath1 disappears when @xmath2 is held fixed ( positively or negatively ) and ( 5 ) as well as ( 6 ) mean that when @xmath2 is held fixed ( positively or negatively ) , the joint probability of @xmath0 and @xmath1 will be factorized .
let us consider an ideal version of epr - bohm experiment in which a spin zero source emits two entangled spin @xmath12 particles in the singlet - state [ 8 ] .
alice ( bob ) measures the spin component of particle @xmath13 ( @xmath14 ) along the direction @xmath15 ( @xmath16 ) , denoted by @xmath17 @xmath18 which takes the values @xmath19 @xmath20 in units of @xmath21 .
the directions @xmath15 and @xmath16 lie on the @xmath1-@xmath2 plane .
now , we consider a hidden - variable model in which we have complete information about the spin component of each particle along any direction , given that the corresponding hidden variables along the same direction are known .
moreover , this model enables us to make some statistical predictions along the other directions .
the framework of our proposed hidden - variable model consists of the following suppositions : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ * s1*. the set of hidden variables @xmath22 ( @xmath23 ; @xmath24 in units of @xmath25 ) constitute the field of common cause , where @xmath26 @xmath27 @xmath28 is a unit vector in the @xmath1-@xmath2 plane .
taking into account the conservation principle of spin angular momentum , @xmath22 uniquely determine the values @xmath29 and @xmath30 for the spin components of particles @xmath13 and @xmath14 , respectively , along the direction @xmath31 .
the outcomes of any spin measurement along @xmath32 are also assumed to be _ predetermined _ by @xmath33 .
* s2*. the spin component @xmath34 is assumed to contain complete information about the spin value of particle @xmath35 ( @xmath36 or @xmath14 ) along the direction @xmath37 subsequently , we define a spin vector along @xmath31 which is a unit vector , i.e. , @xmath38 * s3*. there is an equal chance for the hidden variables @xmath39 and @xmath40 to occur in a given run of the experiment .
so , we are assuming that : @xmath41 where @xmath42 stands for probability .
* s4*. regarding the stochastic character of the model , if @xmath33 are given , we shall have only information about the _
mean _ value of the spin component of particle @xmath35 along another direction @xmath43 , which is denoted by @xmath44 .
the mean value is postulated to be obtained by projecting the spin vector @xmath45 on the purposed direction .
that is , @xmath46 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ applying * s4 * for @xmath47 , one gets @xmath48 which means that we have complete information along the direction @xmath49 , given that @xmath22 is known .
furthermore , using * s2 * and * s4 * , one can also show that , e.g. , @xmath50 where @xmath51 .
similarly , @xmath52 . considering * s1 * and * s2 * , when alice measures the spin component of the first particle along @xmath15 and obtains a result @xmath53 , it will be obvious for her that the hidden variables @xmath54 have caused such a result .
she can also infer that if bob measures along the same direction , he will obtain the result @xmath55 .
however , alice can not predict with certainty that if bob measures along a different direction @xmath56 , what result he will obtain . nevertheless , using the assumptions * s1 * , * s2 * and * s4 * , after her measurement
, alice can refer to @xmath54 to have some statistical predictions at a subquantum level .
for example , according to the relation ( 11 ) she can predict that the mean value of bob s results will be @xmath57 , if he measures along @xmath16 too . ] .
albeit this result is not observable , one should average over the hidden variables @xmath54 to obtain the empirical results ( see below ) .
the same situation holds for bob : after his measurement along @xmath16 and getting the result @xmath58 , he concludes that the mean value of alice s results will be definitely @xmath59 .
( for @xmath60 , @xmath61 , according to * s1*. ) since the choice of measuring direction can be arbitrarily made by each local observer , different sets of hidden variables may be used by alice and bob to make predictions about the results of the other one .
these predictions might be different at a subquantum level .
the hidden variables , however , constitute the same field of common cause .
so , each observer can refer to a definite set of hidden variables to reproduce the _ observational _ predictions .
their predictions should be the same , experimentally .
hence , to reproduce the bell correlations , alice can use the local hidden variables @xmath54 , when she reveals the result @xmath53 for @xmath17 , which we call hereafter the alice description .
bob can also invoke @xmath62 , when he reveals the result @xmath58 for @xmath63 , which we call the bob description ( fig 1 . ) as we shall show in the following , the observational results in both descriptions are the same and they are supposed to be equal to the predictions of quantum mechanics . in the following , we first use the alice description . to see some of the predictions of this model at the level of hidden variables
, it is apparent from * s1 * and * s2 * that : @xmath64 then , due to the factorizability of the joint probabilities in ( 5 ) , in the presence of the common cause ( supposing in * s1 * ) , one can show that @xmath65 here , @xmath66 and @xmath67 = \frac{1}{2}\left [ 1-rq\cos \varphi _ { ab}\right ] \tag{14}\ ] ] where the last equality is obtained by using the relation ( 11 ) .
we can also calculate the correlation function both at the level of hidden - variables and quantum mechanics .
the correlation function at the level of hidden variables is equal to : @xmath68 where @xmath69 is the average value of the spin components of the two particles along @xmath15 and @xmath70 , and it is defined as : @xmath71 using the relations ( 13 ) and ( 14 ) in ( 16 ) , one gets : @xmath72 accordingly , we reach the following result : @xmath73 which means that there is no correlation at the subquantum level , as was expected . in quantum mechanics ,
the correlation function is equal to : @xmath74 where @xmath75 stands for the quantum mechanical expectation values . using * s3 *
, one can show that @xmath76 where we have used the relation ( 11 ) for @xmath77 and @xmath78 . similarly one can show that @xmath79 .
we also have @xmath80 in which the relation ( 17 ) is used .
relations ( 20 ) and ( 21 ) lead us to the following result : @xmath81 which is the correlation function predicted by quantum mechanics**. * * the same results could be obtained , if we used the bob description . for example , one could replace the left hand side of relation ( 14 ) with @xmath82 and get the same result as @xmath83 .$ ] here , @xmath84 and the relations ( 17 ) to ( 22 ) could be derived again , if one substitutes @xmath62 in place of @xmath85 this shows that each local observer by herself ( himself ) can simulate bell correlations without any communication with her ( his ) distant partner .
so , there is no communication cost for simulating bell correlations in our approach [ 4 ] .
this demonstrates the correct application of a common cause description in an appropriate manner .
our proposed model is an instance of a local contextual hidden - variable theory .
contextulaity in our definition means that we should always refer to a definite set of hidden variables , in a given measuring setup , to be able to have some observational predictions .
hence , the joint probability of @xmath86 and @xmath87 being @xmath53 and @xmath58 respectively , in a context which is characterized by the hidden variables @xmath54 and the measuring directions @xmath15 and @xmath16 in alice description ( as was denoted by @xmath88 in ( 13 ) ) , is essentially different from @xmath89 and/or @xmath90 where in the latter case , alice measures the spin component of the first particle along a different direction @xmath91 .
we are not allowed to sum up conditional probabilities in different contexts where distinctive hidden variables characterizes different conditionals .
consequently , bell s inequality can not be derived because different sets of hidden variables ( including different pieces of information ) should be used in different contexts . however , the model is obviously local .
alice and bob refer to their own local hidden variables @xmath54 and @xmath62 , respectively .
there is no causal link between them .
each local observer refers to her ( his ) context dependent hidden variables to make a proper prediction .
notwithstanding , the predictions of distant observers are not necessarily the same , whenever a hidden - variable description is taken into account .
they have no correlation at this level .
for example , alice can attribute the probability @xmath92 to the outcome @xmath93 , whereas bob predicts @xmath94 $ ] for the same outcome at the same time .
nevertheless , their observational predictions are always the same .
the spurious correlation appears when the local hidden variables are statistically averaged .
it is also important to stress here that the statistical description of this model is intrinsically realistic .
while in quantum mechanics probabilities do not describe the current situation of the system , here we can describe the statistical behavior of the particles in a definite context , if the corresponding hidden variables are known .
for example , we can simultaneously attribute probabilities to the spin components of particle @xmath14 along different directions .
that is , all spin components are real and the relation ( 14 ) holds true for every @xmath16 , given that @xmath54 are known at the same time .
but in practice , since one can not _ measure _
@xmath87 for different directions simultaneously , we should use the counterfactual statements for what could be described for different directions in a given measuring setup .
this means that the observational predictions of this model for incompatible observables should be interpreted counterfactually .
this is different from the standard interpretation of quantum mechanics in which the randomness of events is usually believed to be irreducible . instead
, the quantum results are always reproduced in this model , because we are _ ignorant _ of which of the hidden variables , e.g. , @xmath95 or @xmath96 is applicable in different runs of a given experiment .
we are now going to present a classical counterpart for a bell - type experiment which obeys similar rules as what was mentioned for a local - contextual model in section 3 .
let us assume that alice and bob are two observers who are located at different positions with a space - like separation .
there is no communication between them .
intermediately , there is a place where a sender ( sam ) stores balls with different colors . on each ball ,
one of the two signs @xmath97 or @xmath98 is inscribed .
sam sends balls with different colors to each of the two observers in different stages of an experiment successively .
each observer has a color - recognizer device which is able to detect the passage of any colored ball but can be adjusted to identify only _ one _ type of color during each step of an experiment . choosing from sam s colors set ,
let us assume here that alice decides to collect balls with color @xmath99 and bob collects balls with color @xmath100 , standing for , e.g. , amber and blue , respectively .
they are not aware of which color their partner chooses , nor sam needs having knowledge of their decision .
subsequently , alice and bob record the signs on their balls .
then , they will upload their outcomes to a remote computer which can calculate the correlation of data . as a matter of fact
, this experiment is composed of different stages . without loss of generality
, one can suppose that sam owns a set of three colored balls including amber ( @xmath99 ) , blue ( @xmath100 ) and cherry ( @xmath101 ) .
in three different stages , sam sends both members of the pairs ( @xmath102 ) , ( @xmath103 ) and ( @xmath104 ) to alice and bob . for each pair of colored balls including ( @xmath102 ) , ( @xmath105 ) and ( @xmath106 ) ,
sam exploits specific executive algorithms which cause definite correlations between the signs inscribed on balls . in each stage of the experiment ,
alice s ( bob s ) device is capable of recognizing only one of the two colors @xmath99 or @xmath101 ( @xmath100 or @xmath101 ) .
the reader can find a full description of our proposed experiment in the following appendix .
nevertheless , for our present purposes , let us consider a more restricted situation where alice and bob choose @xmath99 and @xmath100 colors , respectively .
so , it is sufficient henceforth to focus on those events in which only @xmath99 and @xmath107 colored balls are detected simultaneously , and leave out the @xmath101 events . in this way
, sam is committed to send these colors according to two complementary executive algorithms which we call @xmath108 and @xmath109 .
each algorithm includes some important features which are described in the following : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 1- sam either sends @xmath99 colored balls with engraved symbols @xmath97 ( hereafter called as @xmath110 ) or the same colored balls with engraved symbols @xmath98 ( @xmath111 ) along with @xmath100 colored balls ( denoted by @xmath112 ) to alice .
the first situation is controlled by @xmath108 algorithm and the second one by @xmath113 .
sam uses the two algorithms with equal chances , i.e. , @xmath114 .
2- suppose , for example , that sam sends an @xmath110 and a @xmath115 ball to alice .
then , at the same time , he will surely send an @xmath116 and a @xmath117 ball to bob .
this means that sam takes into account the anticorrelation rule in choosing @xmath118 symbols for the same colored balls which are sent to different observers .
the anticorrelation rule is fulfilled by either of the @xmath119 and @xmath113 algorithms .
strictly speaking , e.g. , under the @xmath108 algorithm , @xmath110 and @xmath116 balls are always sent to alice and bob , respectively .
the @xmath100 colored balls should also satisfy the anticorrelation rule at the same time .
3- when sam executes the @xmath108 algorithm , the probability that bob records @xmath117 results is chosen to be 0.15 .
this means that sam sends @xmath120 balls to bob ( joined with @xmath121 balls to alice ) , with a frequency of 15 per 100 @xmath100 colored balls total results for which our supposed outcome occures 0.15@xmath122 times . ] .
when sam uses the @xmath113 algorithm , the probability that bob records the same outcome is 0.85 ( i.e. , a frequency of 85 per 100 @xmath100 colored balls ) .
accordingly , when alice records @xmath110 ( @xmath123 ) results , bob will receive @xmath117 ( @xmath124 ) balls with a probability of 0.15 .
4- when sam uses the @xmath108 algorithm , each @xmath115 ball is also sent to alice with a probability of 0.15 .
this is a consequence of the anticorrelation rule in 2 .
hence , according to @xmath113 algorithm , the probability that the same ball is sent to alice is 0.85 .
this accomplishes the stochastic characters of two executive algorithms used by sam .
5- this is an ideal experiment , in the sense that the color - recognizer device is capable of registering any colored ball passing through it , even if it can not recognize its color .
the experiment is ended when alice and bob both have full registered data ( i.e. , the number of all color - registered balls is compatible with all passing balls ) for @xmath99 and @xmath100 colored balls , respectively .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ table 1 summarizes the executive routines of @xmath108 and @xmath125 in terms of some definite frequencies . here ,
e.g. , the results ( @xmath126 , @xmath115 ; @xmath117 , @xmath116 ) denote that sam sends simultaneously a pair of @xmath99 colored balls with signs @xmath97 and @xmath98 as well as a pair of @xmath100 colored balls with opposite signs to alice and bob , respectively .
( here , we are neglecting the @xmath101 colored balls , as was said before . for a full description of this experiment see the following appendix . ) under the @xmath119 , the probability ( frequency ) of occurrence of this event is equal to : @xmath127 the frequencies of other results can be defined similarly .
@xmath128 table 1 .
the executive routines of @xmath108 and @xmath109 in terms of some given probabilities .
the results indicated in table 1 are not exactly what alice and bob can observe . as was said before ,
alice only detects one of the two results @xmath126 or @xmath123 .
correspondingly , bob can only record @xmath117 and @xmath129 outcomes .
so , what both observers can really observe is that they are receiving their desired colors with an equal distribution of @xmath97 and @xmath98 signs . at the end of experiment , alice and
bob upload their data to a remote computer which is able to compare their results and calculate the correlation of them .
table 2 demonstrates what the remote computer calculates for the frequency of occurrence of the observed results and their correlation .
@xmath130 table 2 .
the remote computer calculates the frequency of occurrence of detected results and their correlation . here
, for instance , the joint probability of getting an @xmath110 result by alice in conjunction with a @xmath117 result by bob is equal to : @xmath131 considering the outcomes of alice and bob collectively , the above result , e.g. , can be calculated directly by the remote computer .
however , one can also reach the same result , using sam s information about the statistical features of two executive algorithms @xmath108 and @xmath113 described above : @xmath132 now let us see what sam obtains for the correlations of the twofold results in table 2 in the _ presence _ of the determining factors @xmath133 and @xmath113 which should be viewed as two complementary calculational contexts .
hence , one can define , e.g. : @xmath134 where @xmath135 is the correlation function under the @xmath108 algorithm and the other functions are mean values under the same program . from table 1
, it is apparent that @xmath136 , @xmath137 and @xmath138 .
so , @xmath139 .
one can similarly show that @xmath140 is also equal to zero under the @xmath113 algorithm .
consequently , _ _ the correlation disappears when a third variable ( _ _ @xmath108 _ or _
@xmath113 _ _ ) is introduced and held fixed . _ _ accordingly , the executive algorithms @xmath141 and @xmath113 constitute a field of common cause . comparing this with a bell - type experiment , if one replaces the colored balls @xmath99 and @xmath100 with the spin components of two particles along the directions @xmath15 and @xmath16 respectively , it will be apparent that the two executive algorithms @xmath108 and @xmath113 in the present example stand for the hidden variables @xmath95 and @xmath142 in our bell story . here , however , we restricted ourselves to a single measuring setup which includes only @xmath99 and @xmath100 colored balls .
so , @xmath108 and @xmath113 are solely defined in such a measuring context . yet ,
in such a context too , the two algorithms require complementary calculational conditions under which one is not allowed to sum up the probabilities directly .
the expressions like @xmath143 are meaningless . alternatively ,
if one applies the probability rules correctly , one can hide the role of different conditions .
correlation appears as an emergent result of such a disguisement .
bell s theorem states that there are no local hidden variables which can screen off the mysterious correlations of two entangled particles .
however , it is still important to know what we can infer about the nature of these correlations . is there any non - local action ? or , do alice and bob only update their state of knowledge , when they receive new information via the classical channels of information ? here , the correlation appears when the whole information is considered .
a positive response to either of the above questions introduces other problems which are regularly discussed in the literature [ 9 , 10 , 11 ]
. there is , however , another possibility when one introduces the notion of contextuality .
the consistency of the local contextual hidden - variable theories is not necessarily ruled out by bell s theorem , although this really depends on what one means by the term contextual ( for example , see references [ 6 , 12 ] ) . here ,
we have tried to construct a local contextual hidden - variable model which can reproduce the predictions of quantum mechanics .
the contextual character of this model requires that the type of hidden variables used by two observers be held fixed according to their corresponding measuring setup .
so , it is feasible to explain the two - particle quantum correlations according to the common cause principle in a contextual manner .
let us assume that in three different stages of an experiment- hereafter called first , second and third stages- sam sends three pairs of colored balls ( @xmath102 ) , ( @xmath103 ) and ( @xmath106 ) to both alice and bob .
then , after many trials , alice and bob upload their registered results in distinct categories ( indicating the time and duration of the corresponding experiment ) to a remote computer which can perform statistical calculations .
each stage of experiment is controlled by two complementary executive algorithms which are chosen by sam with equal chance .
hence , there are three equiprobable couples of such executive algorithms : ( @xmath108 , @xmath113 ) , ( @xmath144 , @xmath145 ) , and ( @xmath146 , @xmath147 ) .
either of the two first algorithms @xmath108 or @xmath113 is used when sam decides to send @xmath148 joined with @xmath149 balls or @xmath150 joined with @xmath151 balls to alice and bob respectively , as was described in section 5 .
the second and the third pairs of algorithms are characterized by replacing @xmath100 or @xmath99 colored balls with @xmath101s in the above settings , correspondingly .
table * * a**@xmath152 shows the probabilistic features of two algorithms @xmath153 and @xmath154 .
it is assumed that for both situations , when sam executes any of the two algorithms @xmath153 or @xmath154 , the probability that bob records a positive result is equal to 0.04 . for simplicity
, we have not mentioned the executive routines of @xmath145 and @xmath155 which are simply the complements of two earlier programs , respectively .
when sam executes , say , @xmath158 algorithm , it may happen that bob decides to collect @xmath100 colored balls instead of @xmath101s in a given course of experiment .
for such a case , he does not record any result .
in other words , @xmath159 . a similar situation may happen for alice .
so , in some given courses , it may occur that alice and/or bob , because of their unsuitable choices , register no result .
nevertheless , the experiment is repeated for many times and one can assume that there are always situations where alice and bob register simultaneously all transmitted colored balls in a given course .
moreover , because in all executive algorithms , the probability of occurring an event is supposed to be equivalent to the frequency of occurrence of the same event ( see footnote in section 5 ) , a remote computer can neglect the stages in which any of the two observers detects no result and calculate the frequency of occurrence of the joint outcomes in those stages where all joint data are registered .
this can be done by comparing the total sum of color - registered data with all that the color - recognizer device detects as passing balls .
the experiment is declared to be terminated when alice and bob both have full registered data , i.e. , when all color - registered balls match the passing balls for all triple stages ( see assumption 5 in section 5 ) .
for the second and third stages of this trilogy , the statistical calculations of the remote computer are summarized in table * * a**@xmath162 . here ,
all the complementary algorithm pairs ( @xmath108 , @xmath113 ) , ( @xmath158 , @xmath163 ) and ( @xmath154 , @xmath164 ) constitute the field of common cause .
this means that the corresponding correlations disappear when any of the two algorithms in each pair is held fixed . on the other hand ,
each of the above pairs is defined in a specific context which is characterized by a definite pair of colored balls sent to the observers . disregarding the role of conditional contexts in the application of the probability rules
, one can derive some bell - type inequalities , such as ( see , e.g. , reference [ 13 ] ) the inconsistency , however , is removed , when one remembers that the probability functions in relation ( a-1 ) are derived under different contexts .
one such derivation is shown in the relation ( 25 ) . indeed ,
if we apply the probability rules correctly , the bell - type inequality ( a-1 ) will be never obtained . | considering the common cause principle , we construct a local - contextual hidden - variable model for the bohm version of epr experiment .
our proposed model can reproduce the predictions of quantum mechanics .
it can be also extended to classical examples in which similar correlations may be revealed . |
the level of the sun s magnetic activity is observed to vary on an 11-year time scale and we are currently in the minimum between cycles 23 and 24 .
the current solar minimum is attracting a great deal of attention as it is proving to be quite unusual .
observations of surface and atmospheric effects , such as the number of visible sunspots , the rate of occurrence of solar flares and the strength of the solar wind , highlight just how quiet the sun is .
the sun s activity cycle influences everyday life on the earth .
the rate of occurrence of solar flares is dependent on the number of spots on the sun s surface , and large solar flares can disrupt satellite communications and cause power outages .
coronal mass ejections ( cmes ) , which are another source of radiation that can disrupt life on the earth , are still being observed regularly on the sun , even in this unusual solar minimum ( see the stereo cor1 cme catalog ) .
cosmic rays , which are a significant space radiation hazard , are anticorrelated with the solar cycle .
the level of solar activity is possibly correlated to the earth s climate ( see , for example , * ? ? ? * and references therein ) .
space - weather groups use predictions of solar cycles to anticipate the amount of orbital drag experienced by satellites .
the next solar cycle has already proven difficult to predict as the current solar minimum is lasting significantly longer than expected .
@xcite reviews 50 predictions for the size and timing of cycle 24 and finds a wide range of results , especially in comparison to predictions made before the previous solar cycle @xcite .
for example , predictions of when cycle 24 will reach its maximum range from 2009 december @xcite to 2014 december @xcite .
in fact , the official noaa , nasa and ises solar cycle 24 prediction panel , which studied the predictions collated by @xcite , failed to reach a consensus on when the peak of cycle 24 will occur and how active the upcoming cycle will be . meanwhile , the number of predictions for cycle 24 is ever increasing .
surface measures of the sun s activity , such as the number of sunspots , which are used to aid cycle predictions , indicate that we are still in an extended solar - cycle minimum .
we ask the question : can we learn anything about this unusual solar minimum from the sun s interior ? to answer this question
we investigate the variation with the solar cycle of the frequencies of the sun s natural resonant oscillations , which are known as @xmath0 modes .
solar @xmath0 modes are trapped in cavities below the surface of the sun and their frequencies are sensitive to properties , such as temperature and mean molecular weight , of the solar material .
it has been known since the mid 1980s @xcite that @xmath0-mode frequencies vary throughout the solar cycle with the frequencies being at their largest when the solar activity is at its maximum . by examining the changes in the observed @xmath0-mode frequencies throughout the solar cycle , we can learn about solar - cycle - related processes that occur beneath the sun s surface .
the birmingham solar - oscillations network ( bison ; * ? ? ?
* ) makes sun - as - a - star ( unresolved ) doppler velocity observations , which are sensitive to the @xmath0 modes with the largest horizontal scales ( or the lowest angular degrees , @xmath1 ) .
consequently , the frequencies measured by bison are of the truly global modes of the sun .
these modes travel to the sun s core but their dwell time at the surface is longer than at the solar core because the sound speed inside the sun increases with depth . consequently , the low-@xmath1 modes are most sensitive to variations in regions of the interior that are close to the surface and so are able to give a global picture of the influence of near - surface activity .
bison is a network of autonomous ground - based observatories that are strategically positioned at various longitudes in order to provide as continuous coverage as possible of the sun .
bison is in a unique position to study the changes in oscillation frequencies that accompany the solar cycle as it has now been collecting data for over 30 years .
however , when the network was first established the quality of the data was relatively poor , in comparison to recent years , because of the sporadic coverage of the observations . here
, we have been able to analyze the mode frequencies observed during the last two solar cycles in their entirety .
the precision with which @xmath0-mode frequencies can be determined is directly related to the length of time series under consideration .
consequently , @xmath0-mode frequencies are often determined from time series whose lengths are of the order of years . however
, a compromise must be made here regarding the appropriate length of time series for study so that changes in the solar cycle can be resolved . before 1985
april the observed data are sparse because the early bison data were collected from just one or two sites and , initially , just in summer months in the northern hemisphere . therefore
, when considering the data collected before 1985 april the @xmath0-mode frequencies were obtained from seven time series of different lengths , which reflected the availability of data .
these data are included here for completeness but are not used in the later analysis as the data are too sparse to provide reliable results . after the third bison site , at carnarvon , western australia , was established in 1985 the duty cycle of the bison data increased significantly .
therefore , after 1985 april 22 the time series were truncated to 54 days in length , which corresponds to approximately twice the rotation period of the solar surface .
the analysis performed here concentrates on the data obtained after this date .
a standard likelihood maximization method was used to fit the power spectra of these time series ( e.g. , * ? ? ?
* ) , enabling the observed mode frequencies to be determined .
we have concentrated on the strongest modes of oscillation , which are in the frequency band from 2100 to 3500@xmath2 . to maintain consistency with previous work ( e.g. , * ? ? ?
* ) a minimum activity reference set was determined by averaging the frequencies from time series observed during the minimum activity epoch at the boundary between cycle 22 and cycle 23 .
we then defined the solar - cycle frequency shifts , @xmath3 , as the differences between frequencies given in the minimum activity reference set and the frequencies of the corresponding modes observed at different epochs and , consequently , different levels of activity .
the size of a frequency shift has a well - known dependency on frequency and mode inertia and these dependencies were removed in the manner described in @xcite .
this allowed a weighted average of the frequency shifts observed for each time series to be determined , which provided a mean frequency shift for each epoch , @xmath4 .
figure [ figure[flux shifts ] ] shows the frequency - shift data that were observed during cycles 21 , 22 , and 23 .
the data are shown up to 2009 april .
after 1985 april the data plotted in figure [ figure[flux shifts ] ] have been smoothed over five points , as the mean frequency shifts were obtained from time series of the same length .
one of the most obvious manifestations of the sun s solar cycle is the variation in the number of sunspots on the solar surface , formally recorded as the international sunspot number ( isn ) .
however , the sunspot number is not the only variable that can be used as a proxy for the solar activity .
for example , the radio flux emitted from the sun at a wavelength of 10.7 cm ( hereafter @xmath5 ) also responds to changes in the solar cycle .
many authors ( e.g. , * ? ? ?
* ; * ? ? ?
* and references therein ) have determined and commented upon the good correlations between shifts in the @xmath0-mode frequencies and activity proxies such as the @xmath5 and the isn .
therefore , for comparison purposes , the average 10.7 cm flux has been plotted on top of the observed frequency shifts in figure [ figure[flux shifts ] ] .
after the first seven points plotted in figure [ figure[flux shifts ] ] the @xmath6 has been smoothed as per the frequency shifts . to determine the scale on which the proxy axis was plotted , a linear least - squares fit between the observed frequency shifts and the activity proxy was performed .
the rise and decline of cycle 23 appears to be slow in comparison to cycle 22 .
the oscillation frequencies observed in the current unusual solar minimum are significantly lower than the previous minimum in 1996 and the most recent data indicate that we are still on a downward trend .
notice that the shapes of the previous and current solar minima are very similar , with both appearing to show a double minimum .
this implies that we may still have to wait some time before the rising phase of cycle 24 begins .
it is also interesting to note that , if account is taken of the lower quality of the data observed before 1990 , the minimum between cycles 21 and 22 is also deeper than the last solar minimum .
this could be because both the minimum between cycles 21 and 22 and the current solar minimum ( between cycles 23 and 24 ) correspond to the same phase of 22-year magnetic hale cycles .
it is clearly apparent from figure [ figure[flux shifts ] ] that there is an unusually large difference between the activity proxy and the frequency shifts on the declining phase of cycle 23 .
notably , the frequency shifts are still observed to be changing throughout the solar minimum , which is in contrast to the activity proxy data that show little structure in its variation with time since 2006
. + + [ figure[flux proxies ] ] figure [ figure[flux proxies ] ] allows a comparison between the bison frequency shifts and two proxies of the sun s activity , other than the @xmath6 . as noted above the isn
is a measure of the number of active sunspot groups and individual spots on the visible disk .
the mgii core - to - wing ratio @xcite is determined from space - based uv spectral irradiance measurements .
the mgii h and k line cores originate in the chromosphere and the ratio of the core size to that of the more stable background gives a robust indication of the chromospheric activity .
the @xmath6 and the mgii core - to - wing ratio appear to show better agreement with the frequency shifts than the isn .
this behavior has been observed before @xcite and occurs because the frequency shifts are sensitive to both the strong _ and _ weak components of the sun s magnetic flux .
the @xmath5 and the mgii core - to - wing ratio show similar relative sensitivity to the different components of the sun s magnetic flux , whereas , the isn is predominantly sensitive to the strong component of the sun s magnetic field . for the purposes of this letter it is most important to notice that there are large discrepancies between , respectively , all three of the proxies plotted in figures [ figure[flux shifts ] ] and [ figure[flux proxies ] ] and the frequency shifts in the declining phase of cycle 23 .
furthermore , the structure that is observed in the frequency shifts in the current solar minimum is not replicated by any of the proxies .
it is also important to note that the seismic minimum is decidedly lower than the minimum predicted by the three proxies .
the oscillations are sensitive to the conditions beneath the solar surface , whereas the activity proxy is a measure of the magnetic field that is present on the surface .
we therefore assume that there are processes happening in the solar interior , which have yet to manifest themselves at the surface .
we now turn our attention to the shorter - term variable structure in the frequency shifts visible on top of the general 11-year trend , which has a period of around two years .
similar `` quasi - biennial '' variation of activity proxies has been noted before in data from the green coronal emission line at 530.3 nm at high solar activity @xcite .
these authors used variations near the solar equator and poles and evidence was found of quasi - biennial variability at both low and high latitudes .
an explanation of such quasi - biennial behavior has been put forward in terms of two different types of dynamo operating at different depths @xcite . elsewhere
@xcite , it has been possible to identify shorter secondary periods in the activity cycles of some stars and it is conceivable that asteroseismology may be able to detect such effects in data sets which will become available , for example , from the recently launched @xmath7 satellite @xcite . in spite of the known behavior of the proxies ,
this is the first time that the quasi - biennial variability has been noted in seismic data .
furthermore , the signature of the two - year signal in the activity index was previously restricted to times of moderate to high solar activity .
for the first time , we are seeing this short - term variability at low solar activity . to investigate this further
, a sine wave was fitted to the frequency shifts observed after 1985 april ( see figure [ figure[flux shifts ] ] ) .
the sine wave took the form @xmath8 } s(\delta t)=a_0\sin\left(2\pi a_1\delta t+a_2\right)+a_3\delta t+a_4,\ ] ] where @xmath9 is the time in years after 1985 april 22 , which is the start date of the first 54d time series .
the @xmath10 coefficients are given in table [ table[a coefficients ] ] .
the sine wave was calculated using a weighted least - squares fit to the unsmoothed frequency shifts . in equation
[ equation[sine wave ] ] , the first collection of terms accounts for the sine wave structure of the frequency shifts .
the second and third terms account for an offset which linearly decreases with time .
@xmath10 coefficients used in calculating equation [ equation[sine wave ] ] .
[ cols="^,^",options="header " , ] [ table[a coefficients ] ] we have determined the residuals between the best - fitting sine wave and the observed frequency shifts and the results are plotted in figure [ figure[residuals ] ] .
flux residuals were also determined between a sine wave , which was scaled in amplitude using the linear fit between the flux and the frequency shifts , and the @xmath5 . clearly , there is a lot of structure in both sets of observed residuals and since @xmath11 the frequency - shift residuals are quite substantial in size
. large discrepancies between the two sets of residuals are clearly evident .
a periodogram of the data from the last two cycles shows significant periods of about two and three years with greater than 99% confidence , but the phase of the signal is not locked to that of the underlying 11-year cycle and several incidences of phase jumps are evident , meaning that periods are not easy to determine .
the three - year periodicity is not visible in a periodogram of the activity proxy residuals .
the solar - cycle shifts that are observed in @xmath0-mode frequencies are usually well correlated with proxies of the sun s activity , such as the 10.7 cm radio flux . however , in the declining phase of cycle 23 and the current solar minimum we find unusually large differences between the frequencies observed in bison data and the activity levels .
the current cycle minimum indicated by the helioseismic data is significantly deeper than the minima observed by the activity proxies and the structure that is clearly evident in the frequency shifts is not replicated in the proxy data .
we also observe a quasi - biennial signal in the @xmath0-mode frequencies at _ all _ activity levels .
interestingly , this signal is _ only _ visible at _ high_-activity levels in the proxy data . as the frequency shifts respond to conditions beneath the surface of the sun whereas the proxies respond to changes at or above the surface , we suggest that these differences may be caused by changes in the magnetic flux that have yet to manifest at the surface .
it is also possible that the magnetic flux responsible for the discrepancies between the frequency shifts and the activity proxies will never reach the sun s surface .
the analysis presented here was based on averages made over groups of modes ( @xmath12 , @xmath13 ) .
further work on individual modes may allow one to isolate the location of the variability because each mode shows a different sensitivity to the latitudinal distribution of the surface activity .
such an analysis is currently in progress . in the seismic data , the previous solar minimum exhibited a double minimum and it appears that the current solar minimum is showing a similar structure .
furthermore , the most recent @xmath0-mode frequencies indicate that the current minimum is still declining .
therefore , it may still be some time before we observe the rising phase of cycle 24 .
there have been suggestions that this recent strange behavior of the sun is indicative that the current grand maximum is about to end @xcite .
that would indeed be an occurrence of great significance .
although our results can not predict whether this is true they do indicate that the next solar cycle should be observed with great interest .
this letter utilizes data collected by the birmingham solar - oscillations network ( bison ) .
we thank the members of the bison team , both past and present , for their technical and analytical support .
we also thank p. whitelock and p. fourie at the south african astronomical observatory ( saao ) , the carnegie institution of washington , the australia telescope national facility ( australian commonwealth scientific and research organization , csiro ) , e.j .
rhodes ( mt .
wilson , californa ) and members ( past and present ) of the instituto de astrofisica de canarias ( iac ) , tenerife .
bison is funded by the science and technology facilities council ( stfc ) .
the authors also acknowledge the financial support of stfc .
abreu , j.a . , beer , j. , steinhiber , f. , tobias , s.m . &
weiss , n.o .
, 2008 , , 35 , 20109 benevolonskaya , e. e. , 1998a , , 509l , 49 benevolonskaya , e. e. , 1998b , , 181 , 479 chaplin , w. j. , elsworth , y. , isaak , g. r. , miller , b. a. , & new , r. , 1999 , , 308 , 424 chaplin , w. j. , elsworth , y. , miller , b. a. , verner , g.a .
& new , r. , 2007 , , 659 , 1749 chaplin , w. j. et al . ,
1996 , , 168 , 1 christensen - dalsgaard et al . , 2008 , commun . in asteroseismol . , 157 , 266 elsworth , y. et al . , 1990 , , 345
, 536 howe , r. , 2008 , adv .
space res . , 41 ,
846 joselyn , j. a. et al . , 1997 , eos trans . , 78 , 205 lockwood , m. & frhlich , c. , 2007 , proc .
r. s. , 463 , 2447 maris , g. & oncica , a. , 2006 , sun geosphere , 1 , 1 mccomas , d. j. et al . , 2008 , , 35 , 18103 pesnell , w. d. , 2008 , , 252 , 209 saar , s. h. , & brandenburg , a. , 2002 , astron .
nachr . , 323 , 357 tsirulnik , l. b. , kuznetsova , t. v. & oraevsky , v. n. , 1997 , adv . s. res . , 20 ,
2369 vecchio , a. & carbone , v. 2008 , , 683 , 536 viereck , r. et al .
, 2001 , , 28 , 1343 woodard , m. f. & noyes , r. w. 1985 , , 318 , 449 | the sun is a variable star whose magnetic activity and total irradiance vary on a timescale of approximately 11 years .
the current activity minimum has attracted considerable interest because of its unusual duration and depth .
this raises the question : what might be happening beneath the surface where the magnetic activity ultimately originates ?
the surface activity can be linked to the conditions in the solar interior by the observation and analysis of the frequencies of the sun s natural seismic modes of oscillation - the @xmath0 modes .
these seismic frequencies respond to changes in activity and are probes of conditions within the sun .
the birmingham solar - oscillations network ( bison ) has made measurements of @xmath0-mode frequencies over the last three solar activity cycles , and so is in a unique position to explore the current unusual and extended solar minimum .
we show that the bison data reveal significant variations of the @xmath0-mode frequencies during the current minimum .
this is in marked contrast to the surface activity observations , which show little variation over the same period .
the level of the minimum is significantly deeper in the @xmath0-mode frequencies than in the surface observations .
we observe a quasi - biennial signal in the @xmath0-mode frequencies , which has not previously been observed at mid- and low - activity levels .
the stark differences in the behavior of the frequencies and the surface activity measures point to activity - related processes occurring in the solar interior , which are yet to reach the surface , where they may be attenuated . |
parity - violating nucleon - nucleon interactions cause the spin of transversely polarized neutrons to undergo a rotation when traveling through a target medium , even in the absence of magnetic fields . in this paper
we report the results of a calculation of neutron spin rotation from neutron - proton and neutron - deuteron forward scattering using pionless effective field theory , eft(@xmath7 ) .
this work is part of an effort to provide one consistent eft(@xmath7)framework with reliable theoretical uncertainties to parity - violating ( pv ) interactions in few - nucleon systems .
we hope that the results presented here , along with the results from refs .
@xcite , will assist in the planning , analysis , and interpretation of related pv experiments .
the pv component of the force between nucleons stems from the weak interactions between the standard model constituents of the nucleons . compared to the parity - conserving ( pc ) part
, it is typically suppressed by a factor of @xmath8 to @xmath9 ; for reviews see e.g. refs .
parity - violating neutron spin rotation observables were first discussed by michel in 1964 @xcite and studied further in refs .
recently , an upper bound on the effect in helium-4 was obtained at nist @xcite . at present ,
the effects of hadronic parity violation on the nuclear level can not be predicted from first principles .
early approaches to pv nucleon - nucleon ( @xmath10 ) interactions include the parameterization in terms of s - p wave transitions @xcite and the more widely used phenomenological meson - exchange models , particularly the framework developed by desplanques , donoghue and holstein ( ddh ) in ref .
the ddh approach provides estimated ranges for pv meson - nucleon couplings based on a number of model assumptions . using the `` best values , ''
it has been applied to the study of neutron spin rotation on various targets @xcite .
calculations have also been performed in the so - called hybrid formalism @xcite , where phenomenological wave functions in the strong sector are combined with a pv effective field theory ( eft ) treatment .
we discuss this further in sec .
[ sec : comp ] . in order to avoid model assumptions and to treat all interactions within a unified framework
, we apply effective field theory techniques consistently to the neutron - hydrogen and neutron - deuterium systems .
the identification of a small parameter allows a systematic expansion of our results and a reliable estimate of the size of theoretical errors .
in particular , since typical neutron energies in parity - violating spin - rotation experiments on the lightest nuclei are low enough that pion exchange can not be resolved , we use the pionless eft with only nucleons as dynamical degrees of freedom .
this theory has proven highly successful in the parity - conserving sector ; for an overview see e.g. refs . @xcite . for early eft descriptions of hadronic parity violation
see refs .
a comprehensive formulation of pv effects in efts with and without pions was given in ref .
@xcite . in the pv sector of eft(@xmath7 ) , five independent operators appear at leading order ( lo ) in eft(@xmath7 ) .
they correspond to the five transition amplitudes from ref .
@xcite expressed in a field theory language . the five accompanying parameters , or low - energy constants ( lecs ) , encode the unresolved short - distance physics .
at present , only experimental input can determine these couplings without introducing additional model dependence .
note , however , that for the pionful sector a first study using lattice simulations to determine the pv @xmath11 coupling has been performed @xcite . our final , next - to - leading order ( nlo )
results for @xmath0 and @xmath2 spin rotation provide these processes in terms of the pv lecs , along with estimates of the associated theoretical errors .
measurements of these observables can determine two independent combinations of the pv low - energy constants .
this article is organized as follows : we first review the general formalism of neutron spin rotation in sec .
[ sec : rot ] and present the necessary pc and pv pieces of the lagrangian in secs .
[ sec : pc ] and [ sec : pvlag ] , respectively .
the results for neutron spin - rotation on the proton up to nlo , along with error estimates , are given in sec .
[ sec : npresult ] .
the results for deuterium up to nlo are derived with a detailed discussion of the expected theoretical uncertainties in sec .
[ sec : ndspinrot ] . in sec .
[ sec : comp ] we estimate numerical predictions and compare with earlier work .
conclusions and outlook are given in sec .
[ sec : outlook ] .
appendices contain the general construction principle for the partial - wave projectors of the three - nucleon system and its results for the @xmath12- and @xmath13-waves as well as details of the numerical calculations .
in this section , we define the spin rotation angle , its relation to the scattering amplitude , and the associated conventions we will use .
important resources are refs.@xcite , but note that conventions vary .
a beam of very - low energy neutrons passing through a medium picks up a phase factor from scattering in the target .
this phase factor is related to the index of refraction @xmath14 of the medium . in the simplest case of very low - energy scattering with plane waves describing the incoming state ,
the phase accumulated after traversing a target of thickness @xmath15 is given by @xmath16 where @xmath17 is the magnitude of the wave vector of the incoming particle .
the index of refraction is in turn related to the scattering length @xmath18 by @xmath19 where @xmath20 is the density of scattering centers in the target . in our convention ,
the scattering length and scattering amplitude at zero energy are related by @xmath21 with @xmath22 the reduced mass of the beam - target system .
the phase @xmath23 can therefore be written as @xmath24 for a beam chosen in the @xmath25 direction , a perpendicular polarization in the @xmath26 direction can be written as @xmath27 where the states @xmath28 represent states with positive / negative helicity along @xmath29 .
when traversing a medium , each helicity state evolves with a phase factor : @xmath30 as long as @xmath31 , which is the case for parity - conserving interactions , the polarization of the beam is unchanged ; the state simply picks up an overall phase factor . in the case of parity violation , however , @xmath32 and the neutron spin is rotated by an amount @xmath33 a positive value of the spin rotation angle @xmath34 corresponds to a rotation about the neutron momentum in the sense of a right - handed screw . using eq . , the spin rotation angle per unit length @xmath15 is @xmath35 where @xmath36 is the scattering amplitude for @xmath37-helicity neutrons including the statistical mixture of available target spins . for further details
see refs . @xcite .
a description of eft(@xmath7)and its power counting can be found in refs .
@xcite , for example .
pionless eft is applicable to energies @xmath38 , where @xmath39 and @xmath40 are the pion and nucleon mass , respectively .
the short - distance details of the interactions are encoded in the low - energy couplings ( lecs ) of @xmath10 contact interactions .
operators and observables are organized in terms of the small dimensionless expansion parameter @xmath41 , where @xmath42 is the typical external momentum or momentum transfer and @xmath43 is the breakdown scale of the theory ( the scale at which pion exchange can be resolved ) .
the expansion parameter is typically 1/5 to 1/3 .
the following leading - order ( lo ) and next - to - leading order ( nlo ) calculations only require s - wave interactions in the parity - conserving @xmath10 sector ; higher partial waves are suppressed according to the power counting .
it is convenient ( in particular for three - body calculations ) to introduce spin - triplet and spin - singlet dibaryon fields @xmath44 and @xmath45 with the quantum numbers of the corresponding two - nucleon @xmath46-wave states @xcite .
the dibaryon ( auxiliary ) field @xmath44 also serves as the deuteron interpolating field since both have identical quantum numbers .
the relevant terms of the lagrangian up to nlo are @xmath47-y\left [ d_s^{a \dagger } ( n^t p^a_s n ) + \mathrm{h.c.}\right ] \notag\\ & + d_t^{i\dagger}\left[\delta_t -c_{0t}\left({\mathrm{i}}\partial_0+\frac{\vec{\partial}^2}{4m}+\frac{\gamma_t^2}{m}\right ) \right ] d_t^i + d_s^{a\dagger}\left[\delta_s -c_{0s}\left({\mathrm{i}}\partial_0 + \frac{\vec{\partial}^2}{4m}+\frac{\gamma_s^2}{m}\right)\right ] d_s^a \notag \\ & + \ldots,\end{aligned}\ ] ] using the conventions of ref .
@xmath48 is the isospin doublet of nucleon weyl spinors @xmath49 ( proton ) and @xmath14 ( neutron ) . with @xmath50 ( @xmath51 ) the pauli matrices in spin ( isospin )
space , the matrices @xmath52 and @xmath53 project two - nucleon states onto the @xmath54 and @xmath55 partial waves @xcite .
the parameters of the lagrangian are fixed using z - parameterization @xcite .
choosing @xmath56 the lo parameters @xmath57 are determined from the poles of the @xmath10 @xmath46-wave scattering amplitudes at @xmath58 . in the @xmath54-wave , this reproduces the experimental binding energy of the deuteron @xmath59 .
the leading - order dibaryon propagators are then given by @xmath60 at nlo , only the additional parameters @xmath61 enter .
in @xmath62-parameterization , they are chosen such that the residues of the poles in the dibaryon propagators do not receive any corrections beyond nlo ; see again ref .
@xcite for details : @xmath63 where @xmath64 is related to the effective range @xmath65 . up to nlo , the dibaryon propagators in eq .
are modified by an insertion of the effective - range term @xmath61 of the lagrangian of eq .
( [ eq : pclag ] ) : @xmath66 with the nlo correction ( z - parameterization variant in the second line ) @xmath67 calculations with a deuteron as an external state require wave function renormalization , @xmath68 . in z - parameterization ,
the lo expression is @xmath69 and up to nlo @xmath70 with @xmath71 , the nlo correction results in a @xmath72-shift from the lo value .
while this contribution is much larger than expected from nave power counting , there are no further corrections to @xmath68 at higher orders .
this has the important advantage that the correct asymptotic normalization of the deuteron wave function at large distances @xmath73 is exactly reproduced at nlo , with no corrections at higher orders : @xmath74 taking into account the unusually large nlo term therefore significantly increases overall convergence of the expansion of eft(@xmath7 ) , as demonstrated e.g. in refs . @xcite .
the consistency requirement of including a three - nucleon interaction ( 3ni ) in the @xmath75channel even at lo is discussed in reviews ; see e.g. refs .
the corresponding lagrangian is given by @xmath76^\dagger \left[d^i_t(\sigma_in)-d_s^a(\tau_an)\right],\ ] ] where @xmath77 denotes the three - nucleon coupling , which depends on the uv regulator @xmath78 .
this is the only parameter of @xmath79 scattering up to nlo not determined from @xmath10 experiments .
the 3ni strength @xmath77 can be chosen to reproduce the triton binding energy , or the @xmath75-scattering length .
choosing different low - energy data to fix it provides one method to estimate the theoretical uncertainties in sec . [
sec : ndresult ] .
the parity - conserving @xmath2 scattering amplitude is found by solving a faddeev equation , see e.g. ref .
its pictorial representation in fig . [ fig:3nampt ] specifies the center - of - mass kinematics : the total non - relativistic energy is @xmath80 , and the momentum of the incoming ( outgoing ) deuteron is @xmath81 ( @xmath82 ) .
the on - shell point is at @xmath83 and @xmath84 . ;
thick line : lo dibaryon propagator @xmath85 of eq . ; thin line ( @xmath86 ) : propagator of the exchanged nucleon ; disc : pc 3ni ( @xmath87 ) . ] in the dibaryon framework , the three - nucleon system receives contributions from @xmath88 and @xmath89 configurations , which are conveniently taken into account by decomposing all operators and amplitudes into the so - called cluster - decomposition basis ( see ref .
a.1 ) and our app .
[ sec : projectors ] ) , @xmath90 where the spinor and isospinor indices @xmath91 and @xmath92 , respectively , are often suppressed in the following .
the vector indices @xmath93 label the spin pauli matrices while @xmath94 label the isospin pauli matrices . as an example , the lo dibaryon propagator in the cluster - decomposition basis is defined by a diagonal matrix in terms of the dibaryon propagators eq .
: @xmath95 an analogous expression holds for the nlo correction of eq . .
the spin - quartet channel only receives contributions from @xmath96 , and the corresponding amplitude @xmath97 is the solution to an integral equation that only involves the @xmath98-element of the cluster matrix ( the argument ( @xmath99 applies to each entry in the matrix ) : @xmath100 with @xmath78 the uv regulator .
the projection of the exchange - nucleon propagator onto a specific orbital angular momentum @xmath101 is @xmath102 with @xmath103 and @xmath104 the @xmath101th legendre polynomials of the first and second kind with complex argument , respectively @xcite , and @xmath105 . in the spin - doublet channel , amplitudes with different cluster - decompositions mix , so that with @xmath106 denoting the amplitude for the process @xmath107 and @xmath108 : @xmath109 \notag\\ & -\frac{1}{\pi } \int_0^\lambda { \mathrm{d}}q \ , q^2 \left[{\mathcal{k}}^{(l)}(e;q , p ) \begin{pmatrix } 1 & -3 \\ -3 & 1 \end{pmatrix } + \delta_0^l \frac{2h_0(\lambda)}{\lambda^2 } \begin{pmatrix } 1 & -1 \\ -1 & 1 \end{pmatrix } \right ] \notag\\ & \quad\quad\quad\quad\quad\quad \times{\mathcal{d}}^\text{lo}(e;q ) \begin{pmatrix } t_{d , tt}^{(l ) } & t_{d , st}^{(l)}\\ t_{d , ts}^{(l ) } & t_{d , ss}^{(l ) } \end{pmatrix}(e;k , q)\;\;.\end{aligned}\ ] ] since there is no partial - wave mixing in the pc sector even at nlo , the quartet- and doublet scattering amplitudes can be combined into one cluster matrix @xmath110= \left\ { \begin{array}{ll}\begin{pmatrix } t_q^{(l ) } & 0 \\0 & 0 \end{pmatrix}(e;k , q ) & \mbox { for the spin - quartet , $ s=\frac{3}{2}$}\\[5ex ] \begin{pmatrix } t_{d , tt}^{(l ) } & t_{d , st}^{(l)}\\ t_{d , ts}^{(l ) } & t_{d , ss}^{(l ) } \end{pmatrix}(e;k , q ) & \mbox { for the spin - doublet , $ s=\frac{1}{2}$ } \end{array}\right . \;\;,\ ] ] where @xmath111 is the spin , @xmath101 the orbital angular momentum and @xmath112 the total angular momentum of the @xmath113 partial wave considered . for the nlo pc amplitudes , we use the so - called `` partially - resummed '' formalism , in which the kernel and inhomogeneous part of the integral equations are expanded to nlo and
then iterated , see fig .
[ fig:3nscatteringnlo ] and ref .
@xcite . as defined in the text ; cross : insertion of the effective - range correction @xmath61 to the dibaryon propagator ; diamond : pc 3ni @xmath114 . ]
this modifies the dibaryon propagators in eqs .
( [ eq : quartetfaddeev]/[eq : doubletfaddeev ] ) by an insertion of the effective - range contribution , i.e. by replacing @xmath115 with @xmath116 , see eq . .
no new pc 3ni enters at nlo , but in order to reproduce the three - nucleon observable , @xmath117 has to be adjusted at nlo .
this is most conveniently achieved by dividing @xmath117 into a lo piece @xmath87 and a term at nlo , @xmath114 . in this approach , in addition to _ all _ lo and nlo contributions , some higher - order contributions are also included in the amplitude referred to as @xmath118 .
this does not increase the accuracy of the calculations , which is still set by the order to which the kernel is expanded .
figure [ fig : h0dependence ] shows the cutoff dependence of the 3ni @xmath117 at lo and lo+nlo in the partially - resummed formulation .
-wave pc 3ni @xmath77 , determined to reproduce the @xmath75scattering length .
@xmath119(dashed lines ) : lo ; @xmath120(solid lines ) : lo+nlo in the partially - resummed formulation .
the linear extrapolations are only meant to guide the eye . ]
while @xmath117 varies considerably , observables are cutoff - independent , see e.g. @xcite .
this point will be of importance in the discussion of renormalization of the pv amplitudes in secs .
[ sec : ndnlo ] , [ sec : cutoffind ] and app .
[ app : insertions ] .
the following parameters are used @xcite : @xmath121 , the isospin - averaged nucleon mass @xmath122 ; @xmath123 , @xmath124 , @xmath71 , and @xmath125 are the effective - range parameters of the @xmath10 system ; and the @xmath2 @xmath75scattering length @xmath126 @xcite or triton binding energy @xmath127 determines the 3ni @xmath77 .
the leading - order pv lagrangian relevant to our calculation is given in terms of five s@xmath128p wave transitions @xcite , @xmath129 + \mathrm{h.c . } + \ldots,\end{aligned}\ ] ] where @xmath130 , @xmath131 is a spin - isospin operator , and @xmath132 is a diagonal matrix in isospin space . for equivalent
lagrangians in different bases see refs .
@xcite .
we note as an aside that the relation between these pv dibaryon couplings @xmath4 and the non - dibaryon couplings @xmath133 of ref .
@xcite differs from that given in ref .
@xcite due to different conventions used in the pc sector .
the general expression remains @xmath134 with @xmath135 for @xmath136 and @xmath137 for @xmath138 , but the values for @xmath57 and @xmath139 in the z - parameterization used here differ from the conventions of ref .
@xcite .
higher - order contributions to the lagrangian of eq . are suppressed by additional powers of @xmath140 .
corrections to s@xmath128p wave operators are expected to be suppressed by @xmath141 since they contain the same spin - space structure as the terms in eq . , but with two additional derivatives .
the effects of different partial waves , such as p@xmath128d wave mixing , are suppressed even further . as in the parity - conserving case , a simplistic application of the power counting suggests that parity - violating 3nis first start to appear at n@xmath142lo . unlike the pc case
, this simplistic power counting is valid for pv @xmath2 scattering ; parity - violating corrections to s@xmath128p wave transitions from pv 3nis do not contribute at lo or nlo @xcite .
the lagrangian of eq .
( [ eq : pvlag ] ) is therefore sufficient to determine pv @xmath2 scattering up to and including nlo .
in the dibaryon formalism , the only diagrams contributing to @xmath0 spin rotation at lo are the tree - level diagrams shown in fig . [ fig : pv2n ] .
the corresponding non - zero amplitudes are obtained by using the lo dibaryon propagator of eq .
( [ eq : dibprop ] ) : @xmath143 = & { \mathrm{i}}8 \sqrt{2 } k y \ \frac{{g^{(^3 \ !
p_1 ) } } } { \gamma_t+{\mathrm{i}}k}\ ; , \\ { \mathrm{i}}{\cal m}[{\text{n}\!\ ! \downarrow}{\text{p}\!\ ! \downarrow}\rightarrow { \text{n}\!\ ! \downarrow}{\text{p}\!\ ! \downarrow } ] = & - { \mathrm{i}}8 \sqrt{2 } k y \ \frac{{g^{(^3 \ !
p_1)}}}{\gamma_t+{\mathrm{i}}k } \ ; , \\ { \mathrm{i}}{\cal m}[{\text{n}\!\ ! \uparrow}{\text{p}\!\ !
\downarrow}\rightarrow { \text{n}\!\ ! \uparrow}{\text{p}\!\ ! \downarrow } ] = & { \mathrm{i}}4 \sqrt{2 } k y \ \left ( \frac{{g^{(^3 \ ! s_1-^1 \ !
p_1)}}}{\gamma_t+{\mathrm{i}}k } + \frac{{g^{(^1 \
! s_0-^3 \ !
p_0)}_{(\delta i=0)}}-2{g^{(^1 \ !
p_0)}_{(\delta i=2)}}}{\gamma_s+{\mathrm{i}}k}\right ) \ ; , \\ { \mathrm{i}}{\cal m}[{\text{n}\!\ ! \downarrow}{\text{p}\!\ ! \uparrow}\rightarrow { \text{n}\!\ ! \downarrow}{\text{p}\!\ ! \uparrow } ] = & - { \mathrm{i}}4 \sqrt{2 } k y \
\left ( \frac{{g^{(^3 \ ! s_1-^1 \ !
p_1)}}}{\gamma_t+{\mathrm{i}}k } + \frac{{g^{(^1 \
! s_0-^3 \ !
p_0)}_{(\delta i=0)}}-2{g^{(^1 \ !
! p_0)}_{(\delta i=2)}}}{\gamma_s+{\mathrm{i}}k}\right)\ ; , \end{split}\ ] ] with @xmath144 the magnitude of the center - of - mass momentum .
applying eq .
( [ eq : spinrotation ] ) with @xmath145 for the reduced mass and @xmath146 , the spin rotation angle for polarized neutrons on a hydrogen target is given by @xmath147,\ ] ] where @xmath148 is the ( initial and final ) proton polarization .
since the ( thermal ) external nucleon momentum @xmath144 appears only in the dibaryon propagators , it is negligible compared to the parameters @xmath149 . with eq .
( [ eq : npamplitudes ] ) , the pv rotation angle for a hydrogen target at lo is given by @xmath150 { \rm rad \ mev^{-{\frac{1}{2 } } } } , \nonumber\end{aligned}\ ] ] where the @xmath4 carry units of mev@xmath151 .
since no new pv operators enter at nlo , as discussed in sec .
[ sec : pvlag ] , the scattering amplitudes at nlo are given by the diagrams in fig .
[ fig : pv2nnlo ] . using the dibaryon propagator of eq .
( [ eq : caldnlo ] ) , the lo+nlo result is obtained by multiplying eq . by @xmath152 for each term
@xmath153 if the neutron momentum is neglected compared to @xmath149 .
this correction is 35% for the p@xmath154 waves and
5% for the p@xmath155 waves , in agreement with the nave power counting estimate of the eft(@xmath7)expansion .
our expression up to nlo becomes @xmath156 \
, \left(2{g^{(^3 \ !
p_1)}}+ { g^{(^3 \ ! s_1-^1 \ !
p_1)}}\right ) \\ & - [ 18.5 \pm 1.9 ] \ , \left({g^{(^1 \
! s_0-^3 \ !
p_0)}_{(\delta i=0)}}-2 { g^{(^1 \ ! s_0-^3 \ !
p_0)}_{(\delta i=2)}}\right ) \bigg ) { \rm rad \ mev^{-{\frac{1}{2}}}}\ ; , \end{split}\ ] ] where we have conservatively assigned errors of @xmath157 .
the parity - violating @xmath2 scattering amplitude receives contributions from tree - level ( fig .
[ fig : tree ] ) , `` one - loop '' ( fig .
[ fig : oneloop ] ) , and `` two - loop '' diagrams ( fig .
[ fig : twoloop ] ) .
this nomenclature refers to the number of loops that contain a parity - violating interaction ; the strong amplitudes receive contributions from an infinite series of multi - loop diagrams .
all graphs use the same interaction kernel , namely the tree - level pv diagrams in the off - shell kinematics specified in fig .
[ fig : tree ] . and [ fig : twoloop ] . ] ) contributions not displayed . ]
these are in turn computed using the lagrangians of eqs .
( [ eq : pclag ] ) and ( [ eq : pvlag ] ) , and with the propagator @xmath86 of the exchanged nucleon in eq .
( [ eq : nucprop ] ) as well as the cluster - decomposition version of the dibaryon propagator in eq . .
amplitudes between two states with angular momenta @xmath101 and @xmath158 scale as @xmath159 for small nucleon momenta , see e.g. ( * ? ? ?
* eq . ( 157 ) ) .
since spin - rotation measurements are performed with neutrons with momenta @xmath160 , only transitions between @xmath161 and @xmath162 waves contribute at this order , i.e. @xmath163 .
the projectors necessary to filter out specific three - nucleon partial waves are constructed and listed in app .
[ sec : projectors ] .
while the pv lagrangian contains five independent couplings , forward @xmath79 scattering at low energies only depends on three independent ( isospin - dependent ) linear combinations : @xmath164 since the @xmath79 system is an iso - doublet , the pv coupling @xmath165 can not contribute . here
we are interested in scattering on a neutron , so only the @xmath166 component of the isospin matrix is needed , effectively replacing @xmath167 with @xmath168 .
the projected scattering amplitudes of fig .
[ fig : tree ] ( a ) are : @xmath169 & = { \mathrm{i}}\frac{\sqrt{2}y m}{3pq}\left[2p q_0(-{\mathcal{e}})+qq_1(-{\mathcal{e}})\right ] \begin{pmatrix}{\mathcal{s}}_1&-{\mathcal{t}}\\{\mathcal{s}}_1&-{\mathcal{t}}\end{pmatrix } , \\[1ex ] i{\mathcal{m}}^{(a)}\left[{{\ensuremath{{}^{2}\mathrm{p}_{{\frac{1}{2}}}}\xspace}}\to { { \ensuremath{{}^{2}\mathrm{s}_{{\frac{1}{2}}}}\xspace}};q , p\right ] & = { \mathrm{i}}\frac{\sqrt{2}y m}{3pq}\left[2p q_1(-{\mathcal{e}})+qq_0(-{\mathcal{e}})\right ] \begin{pmatrix}{\mathcal{s}}_1&-{\mathcal{t}}\\{\mathcal{s}}_1&-{\mathcal{t}}\end{pmatrix } , \\[1ex ] { \mathrm{i}}{\mathcal{m}}^{(a)}\left[{{\ensuremath{{}^{2}\mathrm{s}_{{\frac{1}{2}}}}\xspace}}\to { { \ensuremath{{}^{4}\mathrm{p}_{{\frac{1}{2}}}}\xspace}};q , p\right ] & = { \mathrm{i}}\frac{4ym}{3pq } \left[2p q_0(-{\mathcal{e}})+qq_1(-{\mathcal{e}})\right ] \begin{pmatrix}{\displaystyle}\frac{{\mathcal{s}}_1-{\mathcal{s}}_2}{3}&{\mathcal{t}}\\0&0\end{pmatrix } , \\[1ex ] { \mathrm{i}}{\mathcal{m}}^{(a)}\left[{{\ensuremath{{}^{4}\mathrm{p}_{{\frac{1}{2}}}}\xspace}}\to { { \ensuremath{{}^{2}\mathrm{s}_{{\frac{1}{2}}}}\xspace}};q , p\right ] & = { \mathrm{i}}\frac{2y m}{3pq}\left[2p q_1(-{\mathcal{e}})+qq_0(-{\mathcal{e}})\right ] \begin{pmatrix}{\mathcal{s}}_2&0\\{\mathcal{s}}_2&0\end{pmatrix } , \\[1ex ] { \mathrm{i}}{\mathcal{m}}^{(a)}\left[{{\ensuremath{{}^{4}\mathrm{s}_{\frac{3}{2}}}\xspace}}\to { { \ensuremath{{}^{2}\mathrm{p}_{\frac{3}{2}}}\xspace}};q , p\right ] & = { \mathrm{i}}\frac{\sqrt{2}y m}{3pq}\left[2p q_0(-{\mathcal{e}})+qq_1(-{\mathcal{e}})\right ] \begin{pmatrix}{\mathcal{s}}_2&0\\{\mathcal{s}}_2&0\end{pmatrix } q^r{}_s , \\[1ex ] { \mathrm{i}}{\mathcal{m}}^{(a)}\left[{{\ensuremath{{}^{2}\mathrm{p}_{\frac{3}{2}}}\xspace}}\to { { \ensuremath{{}^{4}\mathrm{s}_{\frac{3}{2}}}\xspace}};q , p\right ] & = { \mathrm{i}}\frac{2\sqrt{2}y m}{3pq}\left[2p q_1(-{\mathcal{e}})+qq_0(-{\mathcal{e}})\right ] \begin{pmatrix}{\displaystyle}\frac{{\mathcal{s}}_1-{\mathcal{s}}_2}{3}&{\mathcal{t}}\\0&0\end{pmatrix } q^r{}_s,\\[1ex ] { \mathrm{i}}{\mathcal{m}}^{(a)}\left[{{\ensuremath{{}^{4}\mathrm{s}_{\frac{3}{2}}}\xspace}}\to { { \ensuremath{{}^{4}\mathrm{p}_{\frac{3}{2}}}\xspace}};q , p\right ] & = -{\mathrm{i}}\frac{2\sqrt{10}y m}{3pq}\left[2p q_0(-{\mathcal{e}})+qq_1(-{\mathcal{e}})\right ] \begin{pmatrix}{\displaystyle}\frac{{\mathcal{s}}_1-{\mathcal{s}}_2}{3}&0\\0&0\end{pmatrix } q^r{}_s , \\[1ex ] { \mathrm{i}}{\mathcal{m}}^{(a)}\left[{{\ensuremath{{}^{4}\mathrm{p}_{\frac{3}{2}}}\xspace}}\to { { \ensuremath{{}^{4}\mathrm{s}_{\frac{3}{2}}}\xspace}};q , p\right ] & = -{\mathrm{i}}\frac{2\sqrt{10}y m}{3pq}\left[2p q_1(-{\mathcal{e}})+qq_0(-{\mathcal{e}})\right ] \begin{pmatrix}{\displaystyle}\frac{{\mathcal{s}}_1-{\mathcal{s}}_2}{3}&0\\0&0\end{pmatrix } q^r{}_s , \end{split}\ ] ] with @xmath83 as before and @xmath170 the structure of these matrices in the cluster space of eq .
already appeared in ref .
the projector @xmath171 onto the spin quartet is defined in eq . .
the index @xmath172 ( @xmath173 ) denotes the spin of the outgoing ( incoming ) nucleon , and @xmath73 ( @xmath174 ) is the spin component of the outgoing ( incoming ) @xmath54dibaryon .
the diagrams of fig .
[ fig : tree](b ) with the pv interactions on the lower line can be obtained from those of fig .
[ fig : tree](a ) by hermitian conjugation in cluster , spin , and isospin space : @xmath175^{r\alpha}{}_{s\beta}=\left({\mathcal{m}}^{(a)}[y \to x;p , q]^{s\beta}{}_{r\alpha}\right)^\dagger\;\;,\ ] ] where @xmath176 ( @xmath177 ) denotes the partial wave of the incoming ( outgoing ) state and the spin indices are made explicit . to find the contributions of fig .
[ fig : tree ] to pv scattering , choose the on - shell point @xmath178 : @xmath179={\mathrm{i}}{\mathcal{m}}^{(a)}[x \to y;k , k]+ { \mathrm{i}}{\mathcal{m}}^{(b)}[x \to y;k , k]\ ] ] the `` one - loop '' and `` two - loop '' contributions of figs . [ fig : oneloop ] and [ fig : twoloop ] are generated by convoluting the pv tree - level results with the pc amplitudes of sec .
[ sec : pc ] , following the calculation of higher - order corrections in the pc sector in ref .
@xcite . when the pc amplitude is attached to the left of the pv kernel in a `` one - loop '' diagram , as in fig .
[ fig : oneloop ] , the @xmath180 integration picks the nucleon pole , @xmath181 , and the angular integration is trivial , @xmath182= \sum\limits_{j = a , b}\int\limits_0^\lambda \frac{{{\mathrm{d}}^{}\ ! q\;}q^2}{2\pi^2}\ ; { \mathrm{i}}{\mathcal{m}}^{(j)}\left[x \to y;q , k\right ] { \mathrm{i}}{\mathcal{d}}^\text{lo}(e;q)\ ; { \mathrm{i}}t^\text{lo}[x;k , q]\;.\ ] ] note that we choose the same uv regulator @xmath78 as in the integral equations for the pc amplitudes , eqs .
( [ eq : quartetfaddeev]/[eq : doubletfaddeev ] ) .
when the pc amplitude is attached to the right , the amplitudes are obtained by reading fig .
[ fig : oneloop ] as if the time direction were reversed : @xmath183= \sum\limits_{j = a , b}\int\limits_0^\lambda \frac{{{\mathrm{d}}^{}\ ! q\;}q^2}{2\pi^2}\ ; { \mathrm{i}}t^\text{lo}[y;k , q]\ ; { \mathrm{i}}{\mathcal{d}}^\text{lo}(e;q)\ ; { \mathrm{i}}{\mathcal{m}}^{(j)}\left[x \to y;k , q\right]\;,\ ] ] where we used that the pc amplitudes are time - reversal invariant , i.e. identical when exchanging incoming and outgoing nucleon momenta , @xmath184=t[x;q , p]$ ] . the `` two - loop '' convolutions of fig .
[ fig : twoloop ] are @xmath185=\\ & \sum\limits_{j = a , b}\int\limits_0^\lambda \frac{{{\mathrm{d}}^{}\ ! q\;}q^2}{2\pi^2
} \int\limits_0^\lambda \frac{{{\mathrm{d}}^{}\ ! p\;}p^2}{2\pi^2}\ ; { \mathrm{i}}t^\text{lo}[y;k , p]\ ; { \mathrm{i}}{\mathcal{d}}^\text{lo}(e;p)\ ; { \mathrm{i}}{\mathcal{m}}^{(j)}\left[x \to y;q , p\right ] { \mathrm{i}}{\mathcal{d}}^\text{lo}(e;q)\ ; { \mathrm{i}}t^\text{lo}[x;k , q]\;\;.\notag\end{aligned}\ ] ] we numerically solve the integral equations for the pc amplitudes @xmath186 , eqs .
( [ eq : quartetfaddeev]/[eq : doubletfaddeev ] ) , by the hetherington - schick method @xcite in a mathematica code , with @xmath78 a hard cutoff corresponding to a step function , as detailed in ref . @xcite .
the convolutions for the pv amplitudes eqs .
( [ eq:1loopleft]/[eq:1loopright]/[eq:2loop ] ) are performed numerically , using again a mathematica code .
the numerical error of @xmath187 of these procedures is negligible relative to the systematic uncertainties discussed in sec .
[ sec : ndresult ] .
nucleon - deuteron scattering corresponds to the ( 11 ) element of the cluster matrix .
multiplication with the lo wave - function normalization factor @xmath188 , eq .
, for each external deuteron leg results in the renormalized , physical scattering amplitude @xmath189 between partial waves @xmath176 and @xmath177 at nucleon momentum @xmath144 : @xmath190=&\left(\sqrt{{\mathcal{z}}_t^\text{lo}},0\right)\big ( { \mathrm{i}}{\mathcal{m}}_{\text{tree}}[x \to y;k]+ { \mathrm{i}}{\mathcal{m}}_{1\text{-loop , pc right}}^\text{lo}[x \to y;k]\\ & + { \mathrm{i}}{\mathcal{m}}_{1\text{-loop , pc left}}^\text{lo}[x \to y;k]+ { \mathrm{i}}{\mathcal{m}}_{2\text{-loop}}^\text{lo } [ x \to y;k]\big)\left.\binom{\sqrt{{\mathcal{z}}_t^\text{lo}}}{0}\right|_{\tau_3\to -1}\;. \end{split}\ ] ] no new pv operators contribute to the nlo amplitudes since higher - order pv interactions are suppressed by at least @xmath141 as described in sec . [
sec : pvlag ] .
there are three types of nlo corrections .
the first type is given by diagrams with one insertion of an effective - range correction to the dibaryon propagators proportional to @xmath61 , see eq . .
at this order , the momentum - independent pc 3ni parameter @xmath77 requires an additive adjustment , called @xmath191 , to recapture the low - energy 3n observable that initially fixed @xmath77 at leading order .
this results in a second type of nlo correction to the @xmath2 system : diagrams with @xmath191 inserted once between lo amplitudes .
the third type of corrections comes from the change in the dibaryon wave - function renormalization and is taken into account by replacing @xmath192 in the lo pv amplitude of eq . by @xmath193
, see eq . . in a strictly perturbative approach
, there are two classes of diagrams at nlo .
the class - i diagrams of fig .
[ fig : pv3nstrictnlodifferentloop ] are generated when the pc lo amplitudes @xmath194 $ ] are replaced with the partially - resummed amplitudes @xmath195 $ ] in the topologies of the lo pv `` one- '' and `` two - loop '' diagrams of figs .
[ fig : oneloop ] and [ fig : twoloop ] , i.e. @xmath194\to t^\text{lo+nlo}[x;k , q]$ ] at each occurrence in eqs .
( [ eq:1loopleft]/[eq:1loopright]/[eq:2loop ] ) , see fig .
[ fig : pv3n - nlodifferentloop ] . using the amplitudes @xmath118 of fig . [ fig:3nscatteringnlo ] .
this also replaces the lo pv contributions of figs .
[ fig : oneloop ] and [ fig : twoloop ] .
diagrams with the pv vertex on the lower line as well as `` time - reversed '' contributions not displayed . ]
in addition to the nlo class - i diagrams , the amplitudes obtained from fig . [ fig : pv3n - nlodifferentloop ] also contain all lo contributions as well as a partial resummation of higher - order effects . in analogy to the notation at lo ,
the resulting contributions are called @xmath196 $ ] , @xmath197 $ ] , and @xmath198 $ ] .
this approach has computational advantages since it avoids diagrams with up to four numerical integrations and convolutions with parity - conserving lo full - off - shell amplitudes @xmath199 $ ] , @xmath200 .
in addition , the diagrams in class i contain divergences , studied by refs .
@xcite , that in a strictly perturbative calculation are cancelled by corresponding contributions in class - ii diagrams , to be discussed shortly .
these divergences are challenging to treat numerically , but are avoided in the approach used here .
the amplitudes corresponding to the diagrams of fig .
[ fig : pv3n - nlodifferentloop ] are individually renormalized ; each diagram approaches a unique and finite limit as the cutoff is removed , @xmath201 .
diagrams that are not generated by the replacement @xmath194\to t^\text{lo+nlo}[x;k , q]$ ] in figs .
[ fig : oneloop ] and [ fig : twoloop ] are referred to as `` class - ii '' diagrams and are shown in fig .
[ fig : pv3nstrictnlosameloop ] . for @xmath75and @xmath202 , and as @xmath203 for @xmath204 and @xmath205 .
diagrams with the pv vertex on the lower line as well as `` time - reversed '' contributions not displayed . ] the corresponding amplitudes are given by @xmath206= \sum\limits_{j = a , b}\int\limits_0^\lambda \frac{{{\mathrm{d}}^{}\ !
q\;}q^2}{2\pi^2}\ ; { \mathrm{i}}{\mathcal{m}}^{(j)}\left[x \to y;q , k\right]\ ; { \mathrm{i}}{\mathcal{d}}^\text{nlo}(e;q)\ ; { \mathrm{i}}t[x;k , q]\;,\\ \label{eq:1loopinsertionright } & { \mathrm{i}}{\mathcal{m}}_{1\text{-loop , pc right}}^\text{nlo , class - ii}[x \to y;k]= \sum\limits_{j = a , b}\int\limits_0^\lambda \frac{{{\mathrm{d}}^{}\ ! q\;}q^2}{2\pi^2}\ ; { \mathrm{i}}t[y;k , q]\ ; { \mathrm{i}}{\mathcal{d}}^\text{nlo}(e;q)\ ; { \mathrm{i}}{\mathcal{m}}^{(j)}\left[x \to y;k , q\right]\;,\\ \label{eq:2loopinsertion } & { \mathrm{i}}{\mathcal{m}}_{2\text{-loop}}^\text{nlo , class - ii}[x \to y;k]=\\ & \quad\quad\quad\quad \sum\limits_{j = a , b}\int\limits_0^\lambda \frac{{{\mathrm{d}}^{}\ !
\int\limits_0^\lambda \frac{{{\mathrm{d}}^{}\ ! p\;}p^2}{2\pi^2}\ ; { \mathrm{i}}t[y;k , p]\big[{\mathrm{i}}{\mathcal{d}}^\text{lo}(e;p)\ ; { \mathrm{i}}{\mathcal{m}}^{(j)}\left[x \to y;q , p\right ] { \mathrm{i}}{\mathcal{d}}^\text{nlo}(e;q)\nonumber\\ & \quad\quad\quad\quad\quad\quad\quad\quad\quad\quad+{\mathrm{i}}{\mathcal{d}}^\text{nlo}(e;p)\ ; { \mathrm{i}}{\mathcal{m}}^{(j)}\left[x \to y;q , p\right ] { \mathrm{i}}{\mathcal{d}}^\text{lo}(e;q)\big ] { \mathrm{i}}t[x;k , q]\;\;.\notag\end{aligned}\ ] ] since the kernel of these convolutions is now already nlo , it appears that the lo pc amplitudes @xmath194 $ ] can be used directly .
this is the strategy implemented for the pc @xmath13-wave amplitudes . for a convolution of the class - ii diagrams involving the lo pc @xmath75-wave amplitude
the situation is more complicated .
it was demonstrated on general grounds in ref .
@xcite , and confirmed numerically here that when using @xmath203 the `` one - loop '' amplitudes diverge logarithmically , and the `` two - loop '' amplitudes diverge as @xmath207 ; see app .
[ app : insertions ] .
these are _ not _ divergences that are removed by additional parity - violating 3nis at nlo @xcite . in a strictly perturbative calculation in the pc sector ,
the 3ni @xmath114 absorbs a linear divergence generated by insertions of the effective - range term @xcite .
it is this linear divergence in @xmath114 which , when inserted next to a pv interaction ( see class - i diagrams in fig . [
fig : pv3nstrictnlodifferentloop ] ) , would renormalize the divergence in the class - ii contributions . however , in the partially - resummed formalism used here for the class - i diagrams @xmath208 does not diverge linearly as @xmath201 , see fig .
[ fig : h0dependence ] , since the high - off - shell momentum part of the scattering equation , and hence of the amplitude , becomes softer .
therefore , the treatment of class - i diagrams using the partial resummation technique , see fig .
[ fig : pv3n - nlodifferentloop ] , removes those divergences from the class - i diagrams .
this in turn means that the class - i diagrams do not renormalize the divergent class - ii contributions , so that the divergences in the class - ii diagrams require separate renormalization . by using the renormalized ( partially - resummed ) pc amplitude @xmath209
$ ] for the @xmath75channel the uv behavior of the class - ii diagrams is changed and no divergences appear .
as demonstrated below in fig .
[ fig : partialwaveresults - nd-2s ] , this leads to renormalized , cutoff - independent pv amplitudes .
we therefore choose the renormalized pc amplitude @xmath209 $ ] for this channel .
we also choose the lo+nlo version for the pc @xmath202-wave amplitude in the class - ii diagrams .
this is not required by renormalization , as the loop - integrations in eqs .
( [ eq:1loopinsertionleft]/[eq:1loopinsertionright]/[eq:2loopinsertion ] ) converge well in this channel , see again ref .
however , neutron spin - rotation in deuterium is particularly sensitive to the scattering lengths of the @xmath2 system since it is a process essentially at zero kinetic energy .
it is therefore important to reproduce the experimental values of the scattering lengths .
in particular , the @xmath202scattering length is a factor of @xmath210 larger than the @xmath75one , so it will most likely dominate spin - rotation observables . at lo in the z - parameterization
used here , the @xmath202scattering length of @xmath211 would require a 25 percent correction to achieve the experimental value of @xmath212\;{\ensuremath{\mathrm{fm}}}$ ] @xcite , while the nlo amplitude leads to @xmath213 , in close agreement with the experimental value , see ref .
we therefore use the nlo pc amplitude @xmath214 $ ] for this channel .
this choice does not violate the power counting of the theory and yet improves the utility of our result . to summarize , the pc amplitudes used in eqs .
( [ eq:1loopinsertionleft]/[eq:1loopinsertionright]/[eq:2loopinsertion ] ) are : @xmath215=\left\{\begin{array}{ll } t^\text{lo+nlo}[x;k , q ] & \mbox { for } x\in\{{{\ensuremath{{}^{2}\mathrm{s}_{{\frac{1}{2}}}}\xspace}};{{\ensuremath{{}^{4}\mathrm{s}_{\frac{3}{2}}}\xspace}}\}\;,\\ t^\text{lo}[x;k , q ] & \mbox { for } x\in\{{\ensuremath{{}^{2}\mathrm{p}_{}}\xspace};{\ensuremath{{}^{4}\mathrm{p}_{}}\xspace}\}\;. \end{array}\right.\ ] ] further details and comparisons to other choices are discussed in app .
[ app : insertions ] .
the renormalized lo+nlo pv amplitudes are obtained by summing and multiplying with the lo+nlo wave - function renormalization : @xmath216=\left(\sqrt{{\mathcal{z}}_t^\text{lo+nlo}},0\right)\big ( { \mathrm{i}}{\mathcal{m}}_{\text{tree}}[x \to y;k]\\&+ { \mathrm{i}}{\mathcal{m}}_{1\text{-loop , pc left}}^\text{lo+nlo , class - i}[x \to y;k]+ { \mathrm{i}}{\mathcal{m}}_{1\text{-loop , pc left}}^\text{nlo , class - ii}[x \to y;k]\nonumber\\&+ { \mathrm{i}}{\mathcal{m}}_{1\text{-loop , pc right}}^\text{lo+nlo , class - i}[x \to y;k]+ { \mathrm{i}}{\mathcal{m}}_{1\text{-loop , pc right}}^\text{nlo , class - ii}[x \to y;k]\nonumber\\&+ { \mathrm{i}}{\mathcal{m}}_{2\text{-loop}}^\text{lo+nlo , class - i}[x \to y;k]+ { \mathrm{i}}{\mathcal{m}}_{2\text{-loop}}^\text{nlo , class - ii}[x \to y;k ] \big)\binom{\sqrt{{\mathcal{z}}_t^\text{lo+nlo}}}{0}\nonumber\end{aligned}\ ] ] finally , we reiterate that these amplitudes are complete up to nlo , but also contain some higher - order contributions .
the pc amplitudes @xmath118 used are partially resummed .
in addition , we choose to multiply the entire lo+nlo amplitude by the lo+nlo wave - function renormalization , rather than the lo pv amplitude @xmath189 by @xmath217 and the nlo correction @xmath218 only by @xmath219 . it has been demonstrated repeatedly that this speeds up convergence since the larger - than - usual wave - function renormalization correction at nlo is expected to be the dominant correction to insertions of higher - dimension operators ; see e.g. refs .
@xcite .
the computational effort can be halved by taking advantage of the fact that the individual tree - level amplitudes @xmath220 and @xmath221 are hermitian conjugates , eq . , and that the pc amplitudes are time - reversal invariant , @xmath184=t[x;q , p]$ ] .
the amplitudes of partial waves @xmath222 and @xmath223 are therefore related : @xmath224&=\left({\mathcal{m}}_{1\text{-loop , pc left}}[y \to x;k]\right)^\dagger\;,\\ { \mathcal{m}}_{1\text{-loop}}[x \to y;k]&=\left({\mathcal{m}}_{1\text{-loop}}[y \to x;k]\right)^\dagger\;,\\ { \mathcal{m}}_{2\text{-loop}}[x \to y;k]&=\left({\mathcal{m}}_{2\text{-loop}}[y \to x;k]\right)^\dagger\;,\\ \label{eq : conjugateamps } { \mathcal{m}}_\text{r}[x \to y;k]&={\mathcal{m}}_\text{r}[y \to x;k]\;.\end{aligned}\ ] ] numerically , it is difficult to perform some of the integrations at very low energies .
however , between @xmath225 and @xmath226 the amplitudes of the @xmath2 system deviate from the linear relation only at the percent - level . for the spin rotation angle ,
the relevant quantity is @xmath227 , see eq . .
the difference in @xmath227 at @xmath228 and @xmath229 is less than @xmath230 .
all results quoted are for @xmath228 . for the analysis of the cutoff dependence of the partial - wave amplitudes , we decompose them in terms of the pv interactions @xmath231 , @xmath232 , and @xmath233 of eq .
[ eq : calt ] : @xmath234}{k } = & d[x \to y;{\mathcal{s}}_1](\lambda)\;{\mathcal{s}}_1\\ & + d[x \to y;{\mathcal{s}}_2](\lambda)\;{\mathcal{s}}_2+d[x \to y;{\mathcal{t}}](\lambda)\;{\mathcal{t}}\;\;.\nonumber\end{aligned}\ ] ] in figs .
[ fig : partialwaveresults - nd-2s ] and [ fig : partialwaveresults - nd-4s ] , the dependence of the functions @xmath235 on the cutoff @xmath78 used in the pc 3n integral equations eqs .
( [ eq : quartetfaddeev]/[eq : doubletfaddeev ] ) and in the convolutions of secs .
[ sec : ndlo ] and [ sec : ndnlo ] is shown for each partial - wave and order .
( \lambda)$ ] , eq . .
columns ( left to right ) : partial waves @xmath236 , @xmath237 ; rows ( top to bottom ) : coefficients of @xmath231 , @xmath232 , @xmath233 .
@xmath119 ( dashed lines ) : lo , 3ni @xmath117 determined from @xmath2 @xmath75-scatt . length ; @xmath238(dashed lines ) : lo , 3ni from triton binding energy ; @xmath120 ( solid lines ) : nlo , 3ni from @xmath75-scatt .
length ; @xmath239 ( dotted lines ) : nlo , 3ni from triton binding energy .
the linear extrapolations are only meant to guide the eye .
notice the different scales on the vertical axes .
amplitudes which are identically zero are not displayed . ]
( \lambda)$ ] , eq . .
columns ( left to right ) : partial waves @xmath240 , @xmath241 ; rows ( top to bottom ) : coefficients of @xmath231 , @xmath232 , @xmath233 .
@xmath119 ( dashed lines ) : lo ; @xmath120 ( solid lines ) : nlo
. comments as in fig .
[ fig : partialwaveresults - nd-2s ] .
] since a unique , finite limit exists as @xmath201 , the result is properly renormalized in each partial wave .
these results confirm the theoretical findings of ref .
@xcite that no pv 3ni is required at lo and nlo .
comparison with fig .
[ fig : h0dependence ] also shows that at both lo and nlo the physical amplitudes are smooth even where the 3ni @xmath117 diverges .
a small kink in the cutoff dependence of the partial wave amplitudes is seen at nlo for @xmath242 .
it is evident in all partial waves , and so is not related to the renormalization of the 3ni @xmath117 of the @xmath75-wave .
the phenomenon appears only at cutoffs well beyond the breakdown scale of 200 mev and hence does not have any impact on our final result . for computational convenience , the calculations
so far have been performed in a partial - wave basis . however , in order to obtain the spin rotation angle , we need to determine the forward scattering amplitude for specific neutron helicity states .
recall that in the kinematics of sec .
[ sec : pc ] , the incident and outgoing nucleons have momentum @xmath243 , so that an incoming neutron @xmath244 with positive helicity corresponds to choosing the spin and isospin components @xmath245 .
the relation to the results in the partial - wave basis is obtained by inserting a complete set of projection operators : @xmath246 where @xmath247 $ ] are the partial - wave amplitudes calculated above .
the sum over @xmath248 stands for all partial waves including spin and isospin polarizations . using
the three - body projectors @xmath249 $ ] constructed in app .
[ sec : projectors ] from eq .
, the partial - wave projected matrix elements of the neutron - deuteron state are @xmath250 \right)^{b\beta}{}_{a\alpha}\right]_{11}\;\;,\ ] ] with appropriate vector and spinor indices @xmath251 . the outer bracket @xmath252_{11}$ ] indicates that only the @xmath98-entry of the cluster - decomposition matrix is physically allowed . as an example
, the projection onto the @xmath253-wave is @xmath254 where the indices @xmath255 are needed to specify the spinor and isospinor magnetic quantum numbers of the @xmath253 state , and @xmath256 for forward - scattering .
the positive - helicity neutron - deuteron amplitude is then : @xmath257\;\frac{1}{3}\sum_{i\in\{0;\pm1\ } } \left[\left(\left({\mathcal{p}}_{\{m\}}{}^{ia}[y]\right)^\dagger{\mathcal{p}}^{\{m\}}{}_{ia}[x ] \right)_{\beta=2,\alpha=2;a=2,b=2}\right]_{11}\;\;.\ ] ] note that the iso - vector index @xmath258 is irrelevant for the final @xmath98-component of the cluster - decomposition matrix .
this formula can also be understood as a decomposition into the reduced matrix elements , represented by the amplitudes @xmath259 $ ] calculated above , and the equivalent of clebsch - gordan coefficients weighting the various partial waves .
the negative - helicity amplitude is @xmath260 . using the partial - wave projectors of app .
[ sec : projectors ] leads to our final result for the neutron spin - rotation in deuterium from eq .
with @xmath261 for the reduced mass : @xmath262 -2\sqrt{2}{\mathcal{m}}_\text{r}[{{\ensuremath{{}^{2}\mathrm{s}_{{\frac{1}{2}}}}\xspace}}\to{\ensuremath{{}^{4}\mathrm{p}_{{\frac{1}{2}}}}\xspace};k]\\ & -4{\mathcal{m}}_\text{r}[{{\ensuremath{{}^{4}\mathrm{s}_{\frac{3}{2}}}\xspace}}\to{\ensuremath{{}^{2}\mathrm{p}_{\frac{3}{2}}}\xspace};k ] -2\sqrt{5}{\mathcal{m}}_\text{r}[{{\ensuremath{{}^{4}\mathrm{s}_{\frac{3}{2}}}\xspace}}\to{\ensuremath{{}^{4}\mathrm{p}_{\frac{3}{2}}}\xspace};k]\bigg]\;,\nonumber\end{aligned}\ ] ] where the amplitudes for p to s wave transitions are taken into account by using the relations of eq . . for the detailed discussion of theoretical errors below
, we decompose the spin - rotation result of eq . in terms of functions
@xmath263(\lambda)$ ] which multiply the pv interactions : @xmath264(\lambda)\;{g^{(^3 \ !
! p_1)}}+\tau_3\;c[(^3 \
! p_1)](\lambda)\;{g^{(^3 \
! s_1-^3 \ !
p_1)}}+c[{\mathcal{t}}](\lambda)\;{\mathcal{t}}\;\ ] ] the isospin matrix @xmath167 is replaced by @xmath168 for neutron spin - rotation , and @xmath265 , see eq . .
figure [ fig : spinrotresults - nd ] shows the cutoff dependence of the functions @xmath263(\lambda)$ ] .
( \lambda)$ ] for the neutron spin - rotation angle in deuterium , .
notation as in fig .
[ fig : partialwaveresults - nd-2s ] .
grey band : estimated theoretical uncertainties as described in eq . .
the linear extrapolations are only meant to guide the eye .
notice the different scales on the vertical axes.,title="fig : " ] + ( \lambda)$ ] for the neutron spin - rotation angle in deuterium , .
notation as in fig .
[ fig : partialwaveresults - nd-2s ] .
grey band : estimated theoretical uncertainties as described in eq . .
the linear extrapolations are only meant to guide the eye .
notice the different scales on the vertical axes.,title="fig : " ] + ( \lambda)$ ] for the neutron spin - rotation angle in deuterium , .
notation as in fig .
[ fig : partialwaveresults - nd-2s ] .
grey band : estimated theoretical uncertainties as described in eq . .
the linear extrapolations are only meant to guide the eye .
notice the different scales on the vertical axes.,title="fig : " ] as expected from the discussion of the cutoff dependence of the partial - wave amplitudes , the result for the spin - rotation angle is also properly renormalized .
the final result for neutron spin - rotation in deuterium is : @xmath266\;{g^{(^3 \ !
s_1-^1 \ ! p_1)}}\;+\;[17.0\pm1.7]\;{g^{(^3 \
! s_1-^3 \ !
p_1)}}\\ & + \;[2.3\pm0.5]\ ; \left(3{g^{(^1 \ ! s_0-^3 \ !
p_0)}_{(\delta i=0)}}- 2 { g^{(^1 \ ! s_0-^3 \ !
p_0)}_{(\delta i=1)}}\right ) \big)\text{rad}\;{\ensuremath{\mathrm{mev}}}^{-{\frac{1}{2}}}\;,\nonumber\end{aligned}\ ] ] where the estimate of the residual theoretical uncertainties is justified below .
it will turn out that this estimate is rather conservative ; drawing from the experience in calculations of pc 3n observables using the @xmath62-parameterization , theoretical systematic errors might be estimated to be on the order of @xmath267% , see e.g. @xcite . for now , we note that the theoretical errors of @xmath268 to @xmath269 are comparable to the statistical and systematic errors expected of the most ambitious planned experiments .
we use three methods to estimate theoretical uncertainties , with the eft(@xmath7)parameter @xmath270 as a conservative value for typical momenta in the @xmath79 system on the order of @xmath271 ( see ref .
@xcite ) . * at nlo , higher orders should contribute corrections of order @xmath272 to the total result . * since eft(@xmath7)is not valid at high momentum ,
a residual dependence of observables on momentum modes above the breakdown scale @xmath273 is an indicator of the size of higher - order corrections .
varying the momentum cutoff @xmath78 in the faddeev equations and convolutions of the 3n system from the breakdown scale to higher values provides an estimate of n@xmath142lo effects .
the parameters @xmath4 of the pv lagrangian eq .
are independent of the regularization scale @xmath78 .
therefore , the @xmath263(\lambda)$ ] of the spin - rotation results of eq .
must also be cutoff - independent up to higher - order effects .
in particular , this implies that these coefficients should approach a unique , finite limit as the cutoff is removed . as in a variety of previous calculations in the 3n system
, we consider cutoff - variations from @xmath274 to @xmath275 , see e.g. refs .
* finally , the size of higher - order effects can be assessed by using different low - energy data to determine the pc parameters of eft(@xmath7 ) . since we chose z - parameterization precisely because of its well - established improved convergence , we do not vary parameters of the @xmath10 system , e.g. by replacing @xmath276 at nlo by the effective ranges . on the other hand ,
the strength of the pc 3ni @xmath117 can for example be determined from the @xmath2 scattering length of the @xmath75-wave , or from the triton binding energy .
the difference between both approaches is again a measure of n@xmath142lo effects , i.e. expected to be on the order of @xmath272 .
note that methods ( ii ) and ( iii ) do not apply to the 2n system .
we will base our error - estimate for the @xmath2 spin - rotation coefficients on the most conservative of the above methods .
the nlo corrections are as large as @xmath277% relative to the lo result .
the size of this correction stems from the unnaturally large residue of the deuteron pole , @xmath278 , in z - parameterization .
however , once this effect is taken into account , convergence at n@xmath142loand higher is actually improved @xcite .
different inputs to determine the pc 3ni affect only the @xmath279 partial - wave amplitudes . at nlo
, this can lead to a change of up to @xmath280 in the amplitudes @xmath281 $ ] , in line with the error - estimate criterion ( iii ) . in the functions @xmath263(\lambda)$ ] multiplying the pv couplings @xmath282 and @xmath283 , fig .
[ fig : spinrotresults - nd ] , contributions involving the @xmath75-wave are doubly suppressed relative to those which contain the @xmath202-wave . not only are their relative weights in eq .
small , they are also small in absolute size , as seen from the fact that the @xmath75-wave scattering length itself is a factor of @xmath284 smaller than the @xmath202one . therefore , applying criterion ( iii ) leads to very small variations . varying the cutoff @xmath285\;{\ensuremath{\mathrm{mev}}}$ ] ,
criterion ( ii ) , produces a range at nlo of @xmath286(\lambda)=[7.8\dots8.1]\;\text{rad}\;{\ensuremath{\mathrm{mev}}}^{-{\frac{1}{2}}}$ ] and @xmath287(\lambda)=-[16.5\dots17.4]\;\text{rad}\;{\ensuremath{\mathrm{mev}}}^{-{\frac{1}{2}}}$ ] . for error - estimate criterion ( i ) , i.e. @xmath272 of the total result , corrections are @xmath288 and @xmath289 , significantly larger than the estimate from varying @xmath78 .
we therefore adopt the range from criterion ( i ) as a conservative estimate of the theoretical uncertainties in these coefficients .
overall , the amplitude @xmath290 $ ] dominates @xmath287(\lambda)$ ] , providing more than @xmath291% of its total value . due to symmetries , the function @xmath292(\lambda)$ ]
does not receive any contributions from the @xmath293 channel , and the amplitude @xmath294 $ ] is very small .
the @xmath295 amplitude thus dominates @xmath292(\lambda)$ ] . since this amplitude does not depend significantly on the input used to determine the pc 3ni , criterion ( iii ) significantly underestimates the theoretical uncertainty of the nlo calculation .
cutoff - variation , criterion ( ii ) , is significant in this channel , mapping out the range @xmath292(\lambda)=[1.8\dots2.8]\;\text{rad}\;{\ensuremath{\mathrm{mev}}}^{-{\frac{1}{2}}}$ ] , while criterion ( i ) provides an error - estimate of only @xmath296 .
we therefore adopt the variation from the cutoff changes as a conservative estimate of the theoretical uncertainties . while this error is @xmath297 , the magnitude of @xmath292(\lambda)$ ] is at most a third of the other two @xmath263(\lambda)$ ] .
if there are no accidental cancellations or significant differences in the magnitudes of the pv couplings themselves , the overall contribution of @xmath292(\lambda)$ ] to the spin rotation angle is expected to be small .
our results for @xmath0 and @xmath2 spin rotation are given in terms of the lecs @xmath4 of the lagrangian ( [ eq : pvlag ] ) .
as discussed in the introduction , these couplings are not predicted by the eft , but can be estimated on dimensional grounds . since on the microscopic level
the dominant pv process stems from a single weak gauge boson exchange , the pv @xmath10 interactions are suppressed relative to the pc ones by a factor @xmath298 , where @xmath271 is taken as a typical low - momentum scale . as in the pc interactions between two nucleons and the auxiliary dibaryon field in eq .
, an overall factor of @xmath299 should be included . in order to match the dimensionality @xmath300 $ ] of the couplings @xmath4 , an additional inverse mass dimension enters .
since the pv couplings used here are renormalization - group invariant and encode short - distance physics , the mass scale can only be set by the breakdown scale @xmath301 of eft(@xmath7 ) .
overall , therefore , we expect the magnitude of the pv couplings to be of order @xmath302 an estimate which is compatible with this number may be obtained by appealing to ref .
@xcite , where a value for the combination of couplings involved in pv @xmath303 scattering at 13.6 mev @xcite was extracted .
( asymmetries measured at higher energies are not within the realm of validity of eft(@xmath7 ) . ) using this value as an estimate of all lecs , @xmath304 in our conventions .
note that these are only rough estimates . with these values for the magnitude of the pv couplings and a target density of @xmath305 , the magnitude of the spin - rotation signal in hydrogen of eq .
is @xmath306\ ; \frac{\text{rad}}{\text{m } } \;\;.\ ] ] analogously , the magnitude of the spin - rotation signal in deuterium of eq .
is roughly @xmath307\;\frac{\text{rad}}{\text{m } } \;\;.\ ] ] we stress again that without reliable values of the pv parameters
these results are dimensional , order - of - magnitude estimates and may well be off by factors of @xmath210 or more .
another order - of - magnitude estimate can be obtained by using the parameter set of ref .
@xcite extracted from the ddh `` best estimates '' ; using these values ( eq .
( 31 ) of ref .
@xcite ) and adjusting for different sign conventions yields spin rotation values of @xmath308 for @xmath0 spin rotation and @xmath309 in the @xmath2 case .
note that there is a large spread in the ddh `` reasonable ranges '' surrounding these ddh `` best estimates . ''
we now compare our estimated result with the results from other calculations . without measured pv parameters ,
none of these can be considered as more than order - of - magnitude estimates .
hence we do not normalize with respect to less than order - of - magnitude changes in choice of target density , for example .
our estimate for the @xmath0 spin rotation angle agrees well with earlier results obtained in the ddh and hybrid formalisms @xcite , which found rotation angles between @xmath310 and @xmath311 , depending on model choices made for the couplings and the strong interaction potentials .
the rotation angle in neutron deuteron spin rotation was determined in ref .
@xcite using the ddh model and the argonne @xmath312 interaction , supplemented with the urbana - ix three - nucleon potential .
the result is @xmath313 while a subsequent calculation also employing the a@xmath312+uix potentials found @xcite , @xmath314 this agrees with our findings that the coefficients of the pv couplings in @xmath0 and @xmath2 rotations are of the same order of magnitude , and also agrees with our rough estimate of eq . .
note that result is just one of several @xmath2 spin rotation predictions provided by ref .
@xcite , who consider two other parameter set estimations in addition to a set with the `` best values '' from the wide ranges provided by ddh .
these two parameter sets , collected by bowman @xcite , yield values of @xmath315 and @xmath316 .
in addition , refs .
@xcite and @xcite performed a hybrid calculation in which the phenomenological wave functions of the pc sector are combined with a pv potential derived from eft(@xmath7 ) . in this approach the consistency of treating interactions , wave functions , and currents within a unified framework
in particular , different degrees of freedom appear in the pc and pv interactions . in both references ,
results for two different values of the regularization parameter @xmath22 used in the pv eft(@xmath7)potential are given , while in the pc sector a particular pionful potential with a fixed parameter set is used .
for the choices @xmath317 and @xmath318 , the results for the coefficients multiplying the pv parameters differ by up to two orders of magnitude .
while this regularization dependence can in principle be removed by running the pv couplings to absorb cutoff dependence , these large differences are an indication of the resolution mismatch between eft(@xmath7)and the phenomenological potentials . also note that different regularization schemes and degrees of freedom are employed for pc and pv interactions .
since the renormalization - scale dependence of the pv and pc couplings in the hybrid formalism is unknown , one can not determine whether the calculation is indeed cutoff - independent , and varying the cutoff can not be used to assess residual theoretical uncertainties . on the other hand , as shown above , our results are independent of any choices related to regularization and renormalization within the errors estimated in the eft approach .
at nlo in parity - violating eft(@xmath7 ) , two- and three - body low - energy pv observables depend upon five unknown lecs , the @xmath4 of eq . .
pv eft(@xmath7)is based on the symmetries of qcd and the weak interaction , and the power counting of the theory . in order to verify that qcd is appropriately encoded into this eft , it must be demonstrated that the lecs extracted from independent observables are consistent and are of `` natural '' size
@xmath319 in dimensionless units . to overconstrain
the system requires more than five model - independent calculations involving linearly independent combinations of the lecs , and the availability of the corresponding measurements .
this paper provides two of these calculations : using eft(@xmath7)consistently in both pv and pc sectors of the nucleon interactions , we obtain model - independent results for neutron spin rotation in hydrogen and deuterium targets . at nlo
they are given by @xmath320 \ , ( 2{g^{(^3 \ ! s_1-^3 \ !
! s_1-^1 \ !
p_1 ) } } ) \right.\\ & \left . -
[ 18.5 \pm 1.9 ] \ , ( { g^{(^1 \ ! s_0-^3 \ !
p_0)}_{(\delta i=0)}}- 2 { g^{(^1 \ ! s_0-^3 \ !
p_0)}_{(\delta i=2 ) } } ) \right ) { \rm rad \ mev^{-{\frac{1}{2}}}}\end{aligned}\ ] ] for @xmath0 , and @xmath321\;{g^{(^3 \ ! s_1-^1 \
! p_1)}}\;+\;[17.0\pm1.7]\;{g^{(^3 \
! s_1-^3 \ !
p_1)}}\right.\\ & \left .
+ [ 2.3\pm0.5]\;(3{g^{(^1 \ ! s_0-^3 \ !
p_0)}_{(\delta i=0)}}-2{g^{(^1 \ !
p_0)}_{(\delta i=1 ) } } ) \right){\rm rad \ mev^{-{\frac{1}{2}}}}\end{aligned}\ ] ] for @xmath2 .
they yield two independent constraints on the five lecs .
absent any cancellation between different pv parameters we estimate that the rotation angles for both hydrogen and deuterium targets are of the same size , @xmath322\;\frac{\text{rad}}{\text{m } } \;\;.\ ] ] there is no indication that the spin rotation observable is enhanced for a deuteron target , in agreement with refs .
numerical stability analyses verify the theoretical findings of ref .
@xcite that no parity - violating 3ni is necessary at leading or next - to - leading order .
the two calculations presented here join three others published in the same framework ; the longitudinal asymmetry in @xmath323 scattering @xcite and two pv observables from the @xmath324 system @xcite .
note that other two - nucleon pv calculations using pionless eft can be found in refs .
@xcite . to go beyond consistency to the realm of potential prediction
, we will next consider pv observables in @xmath79 scattering , @xmath325 and @xmath326 before considering heavier systems .
while this paper was under review , j. vanasse posted ref .
@xcite with a result on spin rotation off deuterium . we thank him for discussions clarifying some differences between his and our results that led to the correction of an error in our calculations .
we thank d. r. phillips for important suggestions and encouragement ; m. snow for extensive discussions on possible future pv experiments and his detailed write - up of spin - rotation definitions ; as well as v. gudkov , a. micherdzinska , a. k. opper , m. paris , and y. -h .
song for insightful discussions and helpful suggestions .
we are particularly indebted to the organizers and participants of the department of energy s institute for nuclear theory ( int ) program 10 - 01 : `` simulations and symmetries '' at the university of washington , and for the department of energy s financial support of our visit .
hwg is grateful for the kind hospitality of the nuclear experiment group of the institut laue - langevin ( grenoble , france ) .
mrs thanks the lattice and effective field theory group at duke university for their hospitality .
this work was carried out in part under national science foundation career award phy-0645498 ( hwg , mrs ) , and us - department of energy grants de - fg02 - 95er-40907 ( hwg , mrs ) and de - fg02 - 05er41368 ( rps ) .
we also acknowledge support by university facilitating funds of the george washington university ( hwg ) , and by the center for nuclear studies of the george washington university ( hwg , mrs ) .
in this appendix we construct the projectors used to extract the desired partial wave state(s ) from a given @xmath327 state .
the partial wave state is labelled @xmath328 $ ] , where @xmath111 is the total spin , @xmath101 the orbital angular momentum , @xmath112 the total angular momentum , and @xmath329 the isospin . the projectors given here do not exhaust what is needed to do a partial wave decomposition for higher order or inelastic calculations ; that would require constructing projectors acting on @xmath330 states as well .
we employ the cluster - configuration basis introduced in ref .
@xcite , which also presents the s - wave projectors .
the @xmath331 field has two free indices : the su(2 ) isospin index @xmath18 and the su(2 ) spin index @xmath172 .
the dibaryon field @xmath332 is an isosinglet and spin - triplet ( @xmath54 ) with a free vector index @xmath333 , while @xmath334 is an isovector and spin - singlet ( @xmath55 ) with a free isovector index @xmath258 . to obtain a total @xmath335 from an @xmath336 combination ( indices suppressed )
, we can consider @xmath88 , which is already purely @xmath335 from the su(2 ) decomposition @xmath337 .
an @xmath335 term is also available from @xmath89 via @xmath338 .
the @xmath335 piece is isolated by contracting with a pauli matrix in spin space : @xmath339 .
summation of repeated indices is implied .
the normalization chosen will be explained at the end of this section .
an @xmath340 term can only be obtained from @xmath341 .
it has four degrees of freedom in spin space : one free vector index ( the @xmath333 in @xmath332 ) with one free su(2 ) spin index ( the @xmath172 in @xmath331 ) gives six degrees of freedom , but the two that contribute to @xmath335 are removed by additional constraints on the projector .
the most general form is given by @xmath342 requiring that the @xmath335 and @xmath343 projectors are orthonormal @xmath344 yields @xmath345,\ ] ] where the second expression uses @xmath346 .
orthogonality to the @xmath335 projector results in the required constraints that reduce the number of degrees of freedom to four .
the projector onto @xmath343 satisfies @xmath347 for @xmath348-waves , the projector is obtained by combining @xmath349 powers of the unit vector @xmath350 in the direction of the @xmath351 center - of - mass momentum with the auxiliary field @xmath351 and the nucleon . in the case of the spin-1 field
@xmath44 , this leads to total angular momentum @xmath352 .
the projector onto @xmath353 is : @xmath354\;,\ ] ] which is symmetric and traceless in @xmath355 and @xmath356 separately , @xmath357 @xmath358 and orthonormal to the @xmath335 and @xmath343 projectors above : @xmath359 projectors onto definite states are obtained by contracting @xmath360 and @xmath361 with @xmath362 to extract the desired spin - state and multiplying by @xmath101 powers of momentum .
any remaining free indices are then contracted with the appropriate projector onto the desired total angular momentum , which is again given by @xmath360 , @xmath361 , or @xmath363 , and with kronecker - deltas or @xmath364 as needed .
for example , @xmath365 has the quantum numbers @xmath366 , i.e. @xmath367 . or , as a more complex example : @xmath368 mandates first coupling @xmath369 to @xmath343 using @xmath370 , resulting in @xmath371 , and then multiplying by @xmath372 .
the two free vector indices have to be contracted with another projector @xmath373 , multiplied from the left , such that a @xmath374 state results .
this can only be achieved by multiplying with @xmath375 .
the projector is thus proportional to @xmath376 .
the same construction principle holds for isospin projections . states with @xmath377 can be generated by combining either @xmath44 ( @xmath378 ) or @xmath45 ( @xmath379 ) with the @xmath377 field @xmath380 , since as before @xmath381 and @xmath382 . to obtain an @xmath383 state , first contract all spin - vector indices of @xmath384 using @xmath385 , or all isospin - vector indices of @xmath386 using @xmath51 , leading to two equivalent terms @xmath387n^{b\beta}\equiv \left[\sigma_i\ , d_t^i+\tau_a\,d_s^a \right]n.\ ] ] here
, all spin and isospin indices are listed explicitly in the first expression , while obvious index contractions are suppressed in the second . as two cluster - configurations exist , namely @xmath388 and @xmath389 , it is convenient to follow ref .
a.1 ) in decomposing all operators as @xmath390 each operator is represented in the cluster - configuration basis by a 2x2-matrix which carries spin and isospin indices , and all operators act in the direct tensor product space @xmath391 .
the operator in the cluster - configuration basis that projects onto the @xmath392 component is @xmath393 on the other hand , only @xmath388 contains an @xmath394 component , so that it is useful to define @xmath395 here , @xmath396 is the matrix projecting onto the only physical component of the @xmath343 cluster - configuration matrix .
the following relations hold : @xmath397 the unit matrix in spin and isospin space is @xmath398 . below are the projectors as sources of the fields @xmath380 and @xmath351 with the desired quantum numbers .
first , in analogy with the auxiliary two - nucleon fields @xmath351 , we introduce three - nucleon interpolating fields with the quantum numbers of the @xmath336 state . in symbolic notation
, they are the cluster - configuration vectors representing sources @xmath399^\dagger$ ] with total spin @xmath111 , orbital angular momentum @xmath101 , total angular momentum @xmath112 , isospin @xmath329 , and appropriate vector and spinor indices @xmath251 .
the projection onto a definite partial wave in cluster - configuration space is then @xmath400^\dagger\right)_{\{m\}}\ ; { \mathcal{p}}^{\{m\}}{}_{la}[{}^{2s+1}{l}_{j},i]\;\binom{d_t^l}{d_s^a}\;n\;\;.\ ] ] the complete set of s- or p - wave projectors using auxiliary fields as source states is finally : @xmath401_{la}&=&{\displaystyle}{\mathcal{p}}_{d , la}\ , \\[1ex ] { \mathcal{p}}[{\ensuremath{{}^{2}\mathrm{p}_{\frac{1}{2}}}\xspace},i={\frac{1}{2}}]_{la}&= & { \displaystyle}\left(\vec{\sigma}\cdot{{\vec{e}}}\right){\mathcal{p}}_{d , la}\ , \\[1ex ] { \mathcal{p}}[{\ensuremath{{}^{2}\mathrm{p}_{\frac{3}{2}}}\xspace},i={\frac{1}{2}}]^i{}_{la}&= & { \displaystyle}\sqrt{3}\;q^i{}_k\,e^k\,{\mathcal{p}}_{d , la}\ , \\[1ex ] { \mathcal{p}}[{\ensuremath{{}^{4}\mathrm{s}_{\frac{3}{2}}}\xspace},i={\frac{1}{2}}]^i{}_{l}&=&{\displaystyle}{\mathcal{p}}_{q , l}^i\equiv q^i{}_l{\mathcal{p}}_q\ , \\[1ex ] { \mathcal{p}}[{\ensuremath{{}^{4}\mathrm{p}_{\frac{1}{2}}}\xspace},i={\frac{1}{2}}]{}_{l}&= & { \displaystyle}\sqrt{\frac{3}{2}}\,e_k\,q^k{}_l\,{\mathcal{p}}_q \ , \\[2ex ] { \mathcal{p}}[{\ensuremath{{}^{4}\mathrm{p}_{\frac{3}{2}}}\xspace},i={\frac{1}{2}}]^i{}_{l}&= & { \displaystyle}\frac{3{\mathrm{i}}}{\sqrt{5}}\,q^i{}_n\,\epsilon^{nk}{}_m\,e_k \,q^m{}_l\,{\mathcal{p}}_q= { \displaystyle}\frac{1}{3\sqrt{5}}\left[\sigma^i\,\delta^k_l+\sigma_l\,\delta^{ik } -4\sigma^k\,\delta^i_l-5{\mathrm{i}}\,\epsilon^{ik}{}_{l}\right]e_k\ , { \mathcal{p}}_q \ , \\[2ex ] { \mathcal{p}}[{\ensuremath{{}^{4}\mathrm{p}_{\frac{5}{2}}}\xspace},i={\frac{1}{2}}]^{ij}{}_{l}&= & { \displaystyle}q^{ij}{}_{kl } e^k\,{\mathcal{p}}_q\ . \end{array}\ ] ] the spin - quartet projectors carry no isovector index ; the normalization is discussed below .
the two expressions for @xmath402 $ ] are equivalent since @xmath403 contains two pauli matrices and is multiplied with one more from the left , and products of pauli matrices can be reduced to a sum of terms containing at most one pauli matrix .
the first form contains the spin - quartet projector @xmath404 explicitly and hence is manifestly orthogonal to the spin - doublet projectors .
the second form contains the minimal number of linearly independent structures , derived by building the most general matrix @xmath405 out of @xmath406 , @xmath407 and @xmath364 .
the orthonormalization condition was imposed as follows : projectors to different partial waves are orthogonal after contraction over the auxiliary - field variable @xmath408 and integration over the solid angle element @xmath409 of the auxiliary field cm - momentum direction @xmath350 : @xmath410^{\{m\}}{}_{la}\ ; \left({\mathcal{p}}[{}^{2s^\prime+1}l^\prime_{j^\prime},i^\prime]^\dagger \right)_{\{n\}}^{la}=0 \;\;\forall\;\ ; l\not = l^\prime \vee s\not = s^\prime\vee j\not = j^\prime \vee i\not = i^\prime.\ ] ] when the states are identical , integration and contraction must yield the identity element in the space of total angular momentum states . in the cluster - configuration basis notation : @xmath411_{la}\;{\mathcal{p}}^\dagger[{}^{2s+1}l_{\frac{1}{2}},i]^{la}=\mathbbm{1 } \ , \\
\text{for $ j=\frac{3}{2}$ : } & \quad \frac{1}{4\pi}\int{{\mathrm{d}}^{}\ ! \omega_e\ ; } { \mathcal{p}}[{}^{2s+1}l_\frac{3}{2},i]^i{}_{la}\ ; { \mathcal{p}}^\dagger[{}^{2s+1}l_\frac{3}{2},i]^{la}{}_j=\begin{cases } q^i{}_j&\text{for } s={\frac{1}{2}}\ , \\ q^i{}_j\,{\mathcal{p}}_{q}&\text{for } s=\frac{3}{2}\ , \end{cases } \\
\text{for $ j=\frac{5}{2}$ : } & \quad \frac{1}{4\pi}\int{{\mathrm{d}}^{}\ ! \omega_e\ ; } { \mathcal{p}}[{}^{2s+1}l_\frac{5}{2},i]^{ij}{}_{mn}\ ; { \mathcal{p}}^\dagger[{}^{2s+1}l_\frac{5}{2},i]^{mn}{}_{kl}= q^{ij}{}_{kl}\,{\mathcal{p}}_{q } \;.\end{aligned}\ ] ]
in this appendix , we discuss in more detail the choices of the pc amplitudes @xmath412 $ ] of eq . in the class - ii contributions at nlo appearing in eqs .
( [ eq:1loopinsertionleft]/[eq:1loopinsertionright]/[eq:2loopinsertion ] ) and shown in fig . [ fig : pv3nstrictnlosameloop ] .
as discussed in the main text , the convolution kernel itself is already nlo , so that in principle a lo pc amplitude suffices . figures [ fig : partialwaveresults - nd - compareinsertions-2s ] , [ fig : partialwaveresults - nd - compareinsertions-4s ] and [ fig : spinrotresults - nd - compareinsertions ] compare the dependence of the total amplitudes on the different treatments of @xmath412 $ ] , benchmarked against the leading - order result and the choice of eq . to include the @xmath413-wave amplitudes as nlo and
the @xmath414-wave amplitudes as lo , see eq . .
$ ] , eq . , for different choices of the order of the pc amplitudes in the class - ii contributions of fig .
[ fig : pv3nstrictnlosameloop ] and eqs .
( [ eq:1loopinsertionleft]/[eq:1loopinsertionright]/[eq:2loopinsertion ] ) .
@xmath119 ( dashed lines ) : lo result , 3ni @xmath117 from @xmath2 @xmath75-scatt .
length ; all other : nlo results with different treatments of pc amplitudes , with 3ni from @xmath75-scatt . length .
@xmath120(solid lines ) : pc amplitudes @xmath75as nlo , @xmath414 as lo ( eq . ) ; @xmath415(dotted lines ) : all as nlo ; @xmath416 ( dotted lines ) : all as lo .
linear extrapolations used to guide the eye .
notice the different scales on the vertical axes .
amplitudes which are identically zero are not displayed .
columns ( left to right ) : partial waves @xmath236 , @xmath237 ; rows ( top to bottom ) : coefficients of @xmath231 , @xmath232 , @xmath233 . ] $ ] , eq . , for different choices of the order of the pc amplitudes in the class - ii contributions of fig .
[ fig : pv3nstrictnlosameloop ] and eqs .
( [ eq:1loopinsertionleft]/[eq:1loopinsertionright]/[eq:2loopinsertion ] ) .
@xmath119 ( dashed lines ) : lo result ; all other : nlo results with different treatments of pc amplitudes .
@xmath120(solid lines ) : pc amplitudes @xmath202as nlo , @xmath414 as lo ( eq . ) ; @xmath415(dotted lines ) : all as nlo ; @xmath417(dotted lines ) : all as lo .
linear extrapolations used to guide the eye . notice the different scales on the vertical axes .
amplitudes which are identically zero are not displayed .
columns ( left to right ) : partial waves @xmath240 , @xmath241 ; rows ( top to bottom ) : coefficients of @xmath231 , @xmath232 , @xmath233 . ]
$ ] of the neutron spin - rotation angle in deuterium , eq . , for different choices of the order of the pc amplitudes in the class - ii contributions of fig .
[ fig : pv3nstrictnlosameloop ] and eqs .
( [ eq:1loopinsertionleft]/[eq:1loopinsertionright]/[eq:2loopinsertion ] ) . notation as in figs .
[ fig : partialwaveresults - nd - compareinsertions-2s ] and [ fig : partialwaveresults - nd - compareinsertions-4s ] . linear extrapolations to guide the eye ; different scales on the vertical axes.,title="fig : " ] + $ ] of the neutron spin - rotation angle in deuterium , eq . , for different choices of the order of the pc amplitudes in the class - ii contributions of fig .
[ fig : pv3nstrictnlosameloop ] and eqs .
( [ eq:1loopinsertionleft]/[eq:1loopinsertionright]/[eq:2loopinsertion ] ) . notation as in figs .
[ fig : partialwaveresults - nd - compareinsertions-2s ] and [ fig : partialwaveresults - nd - compareinsertions-4s ] . linear extrapolations to guide the eye ; different scales on the vertical axes.,title="fig : " ] + $ ] of the neutron spin - rotation angle in deuterium , eq . , for different choices of the order of the pc amplitudes in the class - ii contributions of fig .
[ fig : pv3nstrictnlosameloop ] and eqs .
( [ eq:1loopinsertionleft]/[eq:1loopinsertionright]/[eq:2loopinsertion ] ) . notation as in figs .
[ fig : partialwaveresults - nd - compareinsertions-2s ] and [ fig : partialwaveresults - nd - compareinsertions-4s ] . linear extrapolations to guide the eye ; different scales on the vertical axes.,title="fig : " ] consider first the inclusion of the pc @xmath75-wave amplitude at lo instead of nlo ( @xmath416 in fig .
[ fig : partialwaveresults - nd - compareinsertions-2s ] ) , while treating all other amplitudes according to eq . .
this leads to individual contributions diverging logarithmically and as @xmath207 , as demonstrated on general grounds in ref .
@xcite and numerically confirmed here .
the problem is evident in the @xmath78 dependence of the functions @xmath235 and @xmath418 of eqs . and , as shown in figs .
[ fig : partialwaveresults - nd - compareinsertions-2s ] and [ fig : spinrotresults - nd - compareinsertions ] .
this choice must thus be discarded as not properly renormalized .
next , it is consistent to include the pc @xmath202-wave amplitude at lo instead of nlo ( @xmath417 in fig .
[ fig : partialwaveresults - nd - compareinsertions-4s ] ) , while treating all other amplitudes according to eq . .
this still determines the pv amplitudes at nlo since the kernel of the convolution already counts as nlo . with the exception of @xmath419 $ ] , however , the corresponding results are very close to those obtained in a lo calculation of the pv amplitudes ( @xmath119 ) . using the pc @xmath202-wave amplitude at nlo ( @xmath120and @xmath415 )
induces a considerable shift of all functions @xmath420 $ ] . as discussed in sec .
[ sec : ndnlo ] , pc lo and nlo @xmath202amplitudes predict values of the @xmath202scattering length that differ by about @xmath421 , with the nlo result very close to the experimental value .
the pv amplitudes , and therefore the @xmath420 $ ] , are expected to be highly sensitive to the @xmath202scattering length .
this is supported by the plots in fig .
[ fig : partialwaveresults - nd - compareinsertions-4s ] .
we therefore choose the nlo amplitude for the pc @xmath202wave .
higher - order corrections from the @xmath202-wave amplitude are expected to be small , since the nlo expression for the scattering length is already in good agreement with experiment .
finally , it is consistent to include the pc @xmath414-wave amplitudes at nlo ( @xmath415 in figures ) instead of lo ( @xmath120 in figures ) , while treating all other amplitudes according to eq . .
one might speculate that this leads to significant changes since the zero - energy effective - range parameter of the amplitudes , the scattering volume , is up to a factor of 2 bigger at nlo than at lo ; see ref .
@xcite . instead
, the amplitudes appear largely insensitive to the effective - range parameters of the @xmath162-waves .
however , the spurious cutoff dependence already seen when the p - wave amplitudes are included at lo is widened and increased to a pole in the window @xmath422\;{\ensuremath{\mathrm{mev}}}$ ] in the total amplitudes .
its origin seems to be that the partially - resummed approach includes some contributions beyond nlo which need to be renormalized by @xmath414-wave 3nis which are not present at nlo . outside that window , the functions @xmath418 and @xmath235 agree with the choice made in eq . within the error - estimate of sec .
[ sec : ndresult ] .
r. schiavilla , m. viviani , l. girlanda , a. kievsky and l. e. marcucci , phys .
c * 78 * , 014002 ( 2008 ) [ erratum - ibid .
c * 83 * , 029902 ( 2011 ) ] [ arxiv:0805.3599 [ nucl - th ] ] . c. -p .
liu , phys .
rev . * c75 * , 065501 ( 2007 ) . [ nucl - th/0609078 ] .
j. h. hetherington , l. h. schick , phys .
137 * , b935-b948 ( 1965 ) ; r. t. cahill , i. h. sloan , nucl .
* a165 * , 161 - 179 ( 1971 ) ; r. aaron , r. d. amado , phys . rev . * 150 * , 857 - 866 ( 1966 ) ; e. w. schmid , h. ziegelmann , the quantum mechanical three - body problem , vieweg tracts in pure and applied physics vol . 2 , pergamon press ( 1974 ) .
d. eversheim , w. schmitt , s. e. kuhn , f. hinterberger , p. von rossen , j. chlebek , r. gebel and u. lahr _ et al .
_ , phys .
b * 256 * , 11 ( 1991 ) . j. d. bowman , `` hadronic weak interaction '' , int workshop on electric dipole moments and cp violations , march 19 - 23 , 2007 , http://www.int.washington.edu / talks / workshops / int_07_1/. | we calculate the ( parity - violating ) spin rotation angle of a polarized neutron beam through hydrogen and deuterium targets , using pionless effective field theory up to next - to - leading order .
our result is part of a program to obtain the five leading independent low - energy parameters that characterize hadronic parity - violation from few - body observables in one systematic and consistent framework .
the two spin - rotation angles provide independent constraints on these parameters .
our result for @xmath0 spin rotation is @xmath1\ , { \rm rad}\ , { \ensuremath{\mathrm{mev}}}^{-{\frac{1}{2 } } } \ , \left(2{g^{(^3 \ !
s_1-^3 \ ! p_1)}}+{g^{(^3 \
! s_1-^1 \ !
p_1)}}\right ) - [ 18.5 \pm 1.9]\ , { \rm rad } \ , { \ensuremath{\mathrm{mev}}}^{-{\frac{1}{2 } } } \ , \left({g^{(^1 \
! s_0-^3 \ !
p_0)}_{(\delta i=0)}}- 2 { g^{(^1 \ ! s_0-^3 \
! p_0)}_{(\delta i=2)}}\right)$ ] , while for @xmath2 spin rotation we obtain @xmath3\ , { \rm rad } \ , { \ensuremath{\mathrm{mev}}}^{-{\frac{1}{2}}}\;{g^{(^3 \ ! s_1-^1 \ !
p_1)}}\;+\;[17.0\pm1.7]\ , { \rm rad } \ , { \ensuremath{\mathrm{mev}}}^{-{\frac{1}{2}}}\;{g^{(^3 \
! s_1-^3 \ !
p_1)}}+[2.3\pm0.5]\ , { \rm rad } \ , { \ensuremath{\mathrm{mev}}}^{-{\frac{1}{2}}}\;\left(3{g^{(^1 \ ! s_0-^3 \ !
p_0)}_{(\delta i=0)}}-2{g^{(^1 \ !
s_0-^3 \
! p_0)}_{(\delta i=1)}}\right)$ ] , where the @xmath4 , in units of @xmath5 , are the presently unknown parameters in the leading - order parity - violating lagrangian . using nave dimensional analysis to estimate the typical size of the couplings
, we expect the signal for standard target densities to be @xmath6\ ; \frac{\text{rad}}{\text{m } } $ ] for both hydrogen and deuterium targets .
we find no indication that the @xmath2 observable is enhanced compared to the @xmath0 one .
all results are properly renormalized .
an estimate of the numerical and systematic uncertainties of our calculations indicates excellent convergence .
an appendix contains the relevant partial - wave projectors of the three - nucleon system . |
relativistic jets in agn and microquasars carry energy from very small to very large scales .
these jets form in the surroundings of compact objects , such as neutron stars or stellar black - holes in the case of microquasar jets , and supermassive black - holes ( smbh ) in the case of agn jets .
the forming scales are of the order of a few times to about 30 times the radius of the central object , depending on the forming mechanism.@xcite from these regions , jets propagate up to nine powers of ten in distance .
this implies formidable stability and collimation taking into account that different instabilities can affect the evolution of jets independently of their nature .
moreover , their propagation is possibly changing the properties and evolution of the host galaxies and their environments through heating by shocks and/or mixing , and removing of gas by shocks and transfer of momentum .
thus , jet stability has been a matter of study during the last four decades.@xcite depending on the nature of jets , i.e. , whether they are magnetically or particle dominated , different types of instabilities can arise . in the case of magnetically dominated jets , both current - driven ( cdi ) and
kelvin - helmholtz ( khi ) instabilities can grow , whereas in the case of particle dominated jets , only the latter may be present.@xcite different stabilising mechanisms for both kinds of instabilities have been proposed , and some of them could explain the remarkable collimation and long distances covered by extragalactic jets , as opposed to the view that the reason for the stability of jets remains unknown.@xcite in this text i review the state - of - the - art of relativistic jet stability .
section [ cd ] is focused on the cdi .
section [ kh ] is devoted to the khi . in section [ nl ]
i overview several non - linear processes that can also dramatically affect the evolution of jets and their collimation .
finally , in section [ pv ] i discuss the possible reasons why relativistic jets manage to carry large amounts of energy from the central regions of galaxies or binary stars to very large distances .
relativistic jets are thought to be formed by magnetohydrodynamical processes in the surroundings of compact objects.@xcite thus , it is reasonable to expect that jets are magnetically dominated close to these regions.@xcite in this regime , cdi is triggered by differences in the magnetic forces when a toroidal field is present . in the case that a pinch is produced , the larger force produced by the compressed lines allows the pinch to grow . in the case
that a kink is produced , the magnetic lines get closer in the inner part of the kink , increasing the magnetic force and enlarging the curvature of the jet .
both modes are stabilised by the presence of a poloidal field component , as the magnetic tension of these lines acts against the compression or distortion of the jet .
cdi in magnetized flows has been studied from an analytical perspective in the non - relativistic regime by a number of authors.@xcite results show that this instability becomes dominant at large values of helicity ( ratio between the toroidal and poloidal field components ) and it is stabilised by a strong poloidal magnetic field . under the physical conditions typical of jets close to the forming regions
, cdi is expected to be dominant , though with small growth rates.@xcite in the relativistic regime , numerical simulations of static columns have shown that growth rates decrease with decreasing alfvn speed and , regarding the non - linear regime , an increasing helicity of the magnetic field with increasing radius in the jet , stabilises the flow with respect to kink cdi.@xcite the introduction of a sub - alfvnic velocity shear generates important differences in the growth of this instability : in the case that the shear is inside the characteristic radius of the static column , the plasma flows through a temporally growing kink , whereas if the shear is outside that radius , the kink is advected with the flow and grows in distance.@xcite it has been suggested that the conversion from poynting to kinetic flux should occur in the first hundreds of gravitational radii , on the basis of observations and modelling of the jet in m87.@xcite magnetic acceleration has been proposed@xcite as an efficient mechanism .
the growth of cdi to non - linear amplitudes can result in mass - load and acceleration of the entrained particles , so this process could also be related to a change from magnetically to particle dominated jets.@xcite however , there is no direct evidence for any of both .
the role of cdi could be significantly reduced by jet expansion and rotation , which can have a stabilising effect .
further simulations with realistic conditions are needed to solve this question . nevertheless , extragalactic jets embed stars and are necessarily entrained by gas clouds and stars rotating around the agn , so the mass - load and conversion of jet main energy channel from poynting flux to particles is difficult to avoid .
it has been shown that sub - alfvnic flows are khi stable,@xcite but once the jets are super - alfvnic and particle dominated , khi may take over cdi and start to play an important role in the long - term stability of jets .
the khi develops at the boundary between two fluids with relative velocity .
it couples to any perturbation with a certain periodicity at which the system presents unstable modes .
the sound - wave generated by the perturbation propagates through the body of the jet and grows in amplitude due to over - reflection at the boundaries.@xcite in the non - linear regime , the jet can be disrupted and decelerated by the entrainment of external matter .
assuming cylindrical symmetry in jets , solutions to the linearised equations of relativistic hydrodynamics have the form @xmath0,@xcite @xmath1 being the wave - number , @xmath2 the azimuthal wave - number ( an integer giving the number of oscillations around the jet s circular cross - section ) , and @xmath3 the frequency .
these solutions may have complex values of wave - number and/or frequency , the imaginary part giving the exponential growth of the amplitude .
unstable modes are separated into surface modes and body modes .
their distinctive property is the number of zeros that the the radial component of the wave - number ( @xmath4 ) of the unstable wave ( e.g. , the pressure wave ) has between the jet axis and its boundary .
the surface mode shows no zeros between the axis and the boundary , whereas the body modes show as many zeros as their order indicates ( e.g. , one zero in the case of the first body mode ) .
figure 1 shows some extracted modes from the solution to the linear problem of a sheared jet .
the upper lines represent the real part of the solution ( frequency ) versus the wave - number , whereas the lower lines represent the imaginary part ( growth rate ) versus wave - number .
[ f1 ] [ [ the - linear - regime ] ] the linear regime + + + + + + + + + + + + + + + + + the growth of khi has been shown to depend on the jet velocity ( unstable modes grow faster for slower jets ) , temperature ( faster growth for hotter jets ) and density ( faster growth for more dilute jets).@xcite there is a general correlation between this dependence and the long - term stability properties of jets , which is more remarkable in colder , faster and denser ones.@xcite the analytical , and numerical in some cases , knowledge acquired about the effect that the growth of khi modes could have on jets@xcite has been widely used in analysing the structures observed in different extragalactic jets and has allowed to derive estimates of the physical parameters that govern them ( e.g. , 3c 345 , 3c 120 , 3c 273 , m87 , s5
0836 + 710 ) .
@xcite@xcite in particular , it was shown that expansion and acceleration can provide the jet in 3c 120 with long - term stability@xcite or that the jet in m87 could be decelerating and heating , leading to destabilization in the kpc scales.@xcite numerical plus analytical work has also led to the conclusion that the observed structures in the jet in 3c 273 could be coupling to khi modes , which may show up at different observing frequencies , depending on the section of the jet where they develop.@xcite this result also has the implicit conclusion that different observing frequencies may be showing different regions of a transversally structured jet .
@xcite in this respect , several works have also shown that shear - layers are easily generated in jets by different mechanisms , including the growth of khi modes.@xcite in the classical regime , there is no limit to the growth of khi , other than disruption of the jet or the saturation of short wavelengths ( the latter will be discussed in the section about the non - linear phase ) , both involving an increase of the width of the velocity shear with a decrease of the radial gradient of velocity.@xcite for the case of relativistic jets , there is an analytical prediction of the saturation of the linear growth when the amplitude of the velocity perturbation reached the speed of light in the reference frame of the jet.@xcite figure 2 shows the saturation of the perturbation in velocity close to the speed of light .
numerical simulations have confirmed this result and shown that the limiting amplitude is smaller for faster jets , thus making them more stable.@xcite overall , this shows that relativistic jets are intrinsically more stable than classical ones with respect to the growth of any kind of instability . on top of this
, semi - analytic plus numerical studies of sheared jets have also shown that the growth of small - wavelength , fast - growing modes , as those shown on the right of fig .
1,@xcite can lead to long - term collimation of fast jets with moderately thick shear - layers ( @xmath5 , with @xmath6 the jet radius ) .
different analytical and numerical works have certified that the presence of winds surrounding the jet@xcite or thick shear - layers@xcite reduce the growth rates of the khi modes .
it has also been shown analytically that in the presence of a poloidal magnetic field , sub - alfvnic and even super - alfvnic jets can be stabilised by the presence of a magnetised surrounding sheath.@xcite [ f2 ] [ [ the - non - linear - regime ] ] the non - linear regime + + + + + + + + + + + + + + + + + + + + + starts typically after saturation.@xcite if the dominating mode is a low - order one with a relatively long wavelength , the jet shows strong deformations and can be disrupted by the generation of shocks at the boundary with the ambient medium.@xcite in this context , the jet develops a wide mixing layer and undergoes significant deceleration to mildly relativistic speeds . otherwise , if small - wavelength modes like resonant modes ( see fig .
1 ) dominate the growth of the instability , the jet develops a hot shear - layer and keeps collimation.@xcite figure 3 shows the schlieren plots for the last snapshot of two simulations of the temporal growth of khi in jets with different lorentz factors ( @xmath7 for the left image and @xmath8 for the right one ) .
otherwise , both simulations are equal .
the left panel , corresponding to a slower jet developing long - wavelength modes , shows turbulent mixing in a wide region , whereas the right panel , corresponding to the faster jet , which develops resonant modes , shows a significant degree of collimation .
this result has been recently confirmed by 3d simulations and its implications extensively discussed@xcite ( see figure 4 ) . a similar , although not so efficient , mechanism for long - term jet stabilisation was also found for classical jets when the shortest modes dominate the growth of the instability up to the non - linear regime,@xcite which is also true for relativistic jets.@xcite the distribution of a number of simulated jets in a relativistic mach - number versus lorentz factor plane@xcite shows that there is a trend of larger stability in the non - linear regime in the case of faster and colder jets , which fall in the upper - right corner of the plane .
[ f3 ] [ f4 ]
in the previous section , the growth of instabilities has been discussed as starting in the linear regime , i.e. , from small perturbations .
however , jets are destabilised by perturbations with large ( non - linear ) amplitudes and may undergo reconfinement shocks or meet irregularities in the ambient media .
these processes are non - linear and can not be accounted for by linear theory of jet stability .
they are of interest because generally imply strong shocks and thus represent candidate locations for very - high energy emission .
[ [ extragalactic - jets ] ] extragalactic jets + + + + + + + + + + + + + + + + + + most probably jets are not in pressure equilibrium with the ambient medium because the pressure in the surrounding cocoon changes with time , as it expands .
over - pressured jets expand and recollimate after becoming under - pressured with respect to the environment,@xcite or meet high density regions in an inhomogeneous environment ( e.g. , supernova remnants ( snr ) or massive stars with powerful winds ) . depending on the initial pressure difference
, these shocks can trigger small amplitude pinching or helical motion that may couple to khi,@xcite or they can generate a large amplitude pinch and destroy the jet@xcite ( see figure 5 ) . only in the case that jets are close to pressure equilibrium with respect to the ambient medium , the generation and influence of reconfinement shocks can be neglected .
when crossing the galaxy , the jet flow may encounter stars and clouds of gas that can be embedded in the jet or entrain it due to their proper motions.@xcite in the case of stars , the stellar wind has been claimed to efficiently mix with the jet flow and decrease the mean jet velocity,@xcite mainly in the case of massive stellar winds and very light jets .
also , irregularities in the intergalactic medium have been claimed to play a role in the hybrid morphology of a number of sources that show both fri and frii structure in their two jets.@xcite large amplitude initial perturbations , such as those triggered by changes in the injection angle of the flow could generate non - linear distortions of jets and lead to their disruption.@xcite the injection of overdense plasma in jets has been invoked to explain the observation of components travelling through parsec - scale jets and the generation of trailing components.@xcite these have been shown to generate significant pinching that could couple to khi modes.@xcite in the kiloparsec scales , the arcs observed in some fri jets * * ? ? ?
* * have also been suggested to be due to changes in the injection rates from the source.@xcite [ f5 ] [ [ microquasar - jets ] ] microquasar jets + + + + + + + + + + + + + + + + inhomogeneities in the ambient medium can also have a strong effect on microquasar jets in high - mass x - ray binaries ( hmxb ) with typical powers ( @xmath9).@xcite such inhomogeneities are encountered by the jet in the transition between the unshocked and shocked stellar wind from the massive companion .
another change of ambient medium occurs for a young hmxb still embedded in the snr , at the transition between the shocked wind and the shocked snr , and finally from the latter to the interstellar medium ( ism ) through the shocked ism . for the case of older hmxb ,
the system has moved away from the snr or this has become very dilute and the region of interaction between the shocked stellar wind and the shocked ism also implies a change in the conditions encountered by jets . in these density jumps jets
are decelerated and need several thousand years to drill the shocked ism before being able to propagate into the ism .
proper motion of the hmxb provides a further destabilising mechanism , as the jet is impacted by the otherwise quasi - steady shocked stellar wind , propagating in the direction opposite to that of the proper motion . within the binary system , the stellar wind impacting
the jet can also generate jet entrainment , deceleration and loss of collimation for jets with powers @xmath10 , mainly if a strong reconfinement shocks that decelerate the flow occur in this region , as has been shown to be the possible.@xcite see also figure 6 .
in the previous sections , the main properties of instabilities that can affect the large - scale morphology of relativistic jets , namely cdi and khi , have been summarised .
these have been proven to be disruptive enough to trigger efficient entrainment and decelerate jets .
however , several mechanisms that may damp their growth and thus make relativistic jets stable have been given : * cdi : poloidal magnetic field , mildly relativistic sheaths and jet expansion .
* khi : thick shear - layers or surrounding winds , the decrease of the cocoon density with time and jet expansion .
most importantly , it has been shown that relativistic jets are less sensitive to the growth of instabilities than classical jets , due to the saturation of their growth at the speed of light .
in addition , resonant modes that can show up in fast jets surrounded by thin shear - layers , or simply the development of short wavelength modes as compared to the jet radius can result in little mixing , restricted to the boundaries of jets , and thus little loss of jet thrust and collimation .
if one or several of these mechanisms play a role , jets can remain collimated along large distances , as it has been shown by simulations .
other mechanisms such as jet rotation or different configurations of the magnetic field are still to be explored in the relativistic regime . the properties of jets and the media outside the leading bow - shock change with distance to the source .
the cocoons surrounding jets show a homogeneous pressure in long - term simulations of jets due to their sound - speeds being larger than the expansion velocities . as the cocoon expands , its density and pressure drop and , as long as the properties of the jet do not change
, it will become denser with respect to its environment , gaining extra stability .
the only caveat is that when the cocoon is very under - pressured , strong reconfinement shocks may form .
nevertheless , even when jets are entrained and decollimated , they seem to preserve a large velocity that can be sub - relativistic , but could still be enough to keep the propagation of the jet to large scales . in fri jets , as compared to frii s , it is more likely that non - linear processes like mass - loading from stars@xcite or strong reconfinement shocks@xcite are the cause of their disrupted morphologies .
these jets have apparently similar velocities in the parsec - scales as the friis,@xcite so other intrinsic properties such as density , temperature or composition should play a role to explain their lower powers .
they are observed to have an opening angle in the inner regions , but there is not any hint of growth of instabilities before the jet deceleration and loss of collimation . in hybrid sources
in which the jet and counterjet appear to have different morphologies,@xcite it is possible that irregularities in the ambient medium are the cause of the difference , for jets with powers between those typical of fris and friis . only in one case an frii classified jet with an apparently irregular structure in the largest scales,@xcite has been reported to develop instabilities that could cause the jet disruption.@xcite this implies that , under regular conditions , i.e. , without strong non - linear perturbations , the stabilising mechanisms proposed by different authors in the last years and summarised here play a significant role .
it has also been suggested that frii jets can turn into fri jets in the long term@xcite or after the end of an active phase , as numerical simulations show.@xcite however , this possibility should be matched to the relativistic velocities measured in fri jets at the parsec scales .
the long - term stability of relativistic jets allows them to carry energy from very small to very large scales .
the bow - shocks generated by extragalactic jets need @xmath11 years to reach distances of the order of hundreds of kpc . in the most powerful radio - sources ,
strong shocks are triggered in the ism and icm , which can change the evolution of the galaxies themselves by removal of ism .
this process can suppress star formation significantly and heating the surrounding gas , as shown by recent long - term simulations of relativistic jets.@xcite thus , the importance of the study of this interaction and the evolution of the jet itself . the research on jet stability in the next years should be addressed to investigate the effect of magnetic fields on the long - term evolution of jets .
this should help to study , with input from observations and modelling , to which distance and extent the magnetic field is dynamically important and the role of instabilities in the eventual transfer of poynting flux into kinetic energy of particles .
in addition , more realistic situations should be explored , such as jet rotation and expansion .
these works will certify or falsify the role of the stabilising mechanisms summarised in this contribution and could even add further ones , suggesting that there may be more than one answer to the problem of jet stability .
finally , it would be important to study , via detailed numerical simulations , the process of mixing and entrainment that jets undergo as a consequence not only of the growth of instabilities , but also of the interaction with stellar winds and clouds of cold gas .
i acknowledge j.m . mart , m. hanasz and p.e .
hardee for sharing their understanding of jet physics with me during the last years
. financial support by the spanish `` ministerio de ciencia e innovacin '' ( micinn ) grants aya2010 - 21322-c03 - 01 , aya2010 - 21097-c03 - 01 and consolider2007 - 00050 is also acknowledged .
0 r. d. blandford and r. l. znajek , _ mon . not .
r. astron .
soc . _ , * 179 * , 433 ( 1977 ) .
r. d. blandford and d. g. payne , _ mon . not .
r. astron .
_ , * 199 * , 883 ( 1982 ) . b. d. turland and p. e. g. scheuer , _ mon . not .
r. astron .
_ , * 176 * , 421 ( 1976 ) .
r. d. blandford and j. e. pringle , _ mon . not .
r. astron .
_ , * 176 * , 443 ( 1976 )
. a. ferrari , e. trussoni , l. zaninetti , _ astron .
_ , * 64 * , 43 ( 1978 ) .
e. hardee , _ astrophys .
j. _ , * 234 * , 47 ( 1979 ) .
e. hardee , the stability of astrophysical jets , in _ jets at all scales , proc .
of the 275 iau symposium _
, eds . g. e. romero , r. a. sunyaev , t. belloni ( cambridge university press , cambridge , 2011 ) , p. 41 .
j. c. mckinney and r. d. blandford , _ mon . not .
r. astron .
_ , * 394l * , 126 ( 2009 ) . y. mizuno , y. lyubarsky , k .-
nishikawa and p. e. hardee , _ astrophys .
* 700 * , 684 ( 2009 ) .
y. mizuno , p. e. hardee and k .-
nishikawa , _ astrophys .
j. _ , * 734 * , 19 ( 2011 ) .
m. sikora , m. c. begelman , g. m. madejski and j .- p .
lasota , _ astrophis .
j. _ , * 625 * , 72 ( 2005 ) .
p. e. hardee , and j. a. eilek , _ astrophys .
j. _ , * 735 * , 61 ( 2011 ) .
s. s. komissarov , n. vlahakis , a. knigl and m. v. barkov , _ mon . not .
r. astron .
_ , * 394 * , 1182 ( 2009 ) . y. mizuno , p. hardee and k .-
nishikawa , _ astrophys .
j. _ , * 662 * , 835 ( 2007 ) .
d. g. payne and h. cohn , _ astrophys .
j. _ , * 291 * , 655 ( 1985 ) .
e. hardee , _ astrophys .
j. _ , * 533 * , 176 ( 2000 ) .
m. birkinshaw , the stability of jets , in _ beams and jets in astrophysics _ ,
a. hughes , ( cambridge university press , cambridge , 1991 ) , p. 278 .
m. perucho , j. m. mart and m. hanasz , _ astron .
_ , * 443 * , 863 ( 2005 ) .
p. e. hardee , _ astrophys .
, * 318 * , 78 ( 1987 ) .
a. p. lobanov and j. a. zensus , _ science _ , * 294 * , 128 ( 2001 ) .
e. hardee , r. c. walker and j. l. gmez , _ astrophys .
j. _ , * 620 * , 646 ( 2005 ) .
m. perucho , a. p. lobanov , j. m. mart and p. e. hardee , _ astron .
_ , * 456 * , 493 ( 2006 ) .
a. p. lobanov , t. p. krichbaum , a. witzel and j. a. zensus , _ publ .
_ , * 58 * , 253 ( 2006 ) .
m. perucho and a. p. lobanov , _ astron .
_ , * 469 * , l23 ( 2007 ) .
m. perucho , a. p. lobanov and y. y. kovalev , physical information derived from the internal structure in jets , in _ approaching micro - arcsecond resolution with vsop-2 _ , astrophysics and technologies asp conference series , vol .
402 , eds . y. hagiwara , e. fomalont , m. tsuboi and y. murata ( san francisco , 2009 ) , p. 349
m. perucho , y. y. kovalev , p. e. hardee , a. p. lobanov and i. agudo , in preparation .
aloy , j .- l .
gmez , j.m .
ibez , j. m. mart and e. mller , _ astrophys .
j. _ , * 528 * , 85 ( 2000 ) .
g. bodo , g. , s. massaglia , s. , a. ferrari and e. trussoni , _ astron .
_ , * 283 * , 655 ( 1994 ) .
m. hanasz , phd .
thesis , ( 1995 ) .
m. hanasz , kelvin - helmholtz instability of relativistic jets - the transition from linear to nonlinear regime , in _ relativistic jets in agns _ , proceedings of the international conference , eds .
m. ostrowski , m. sikora , g. madejski , m. begelman ( krakow , 1997 ) , p. 85 .
m. perucho , m. hanasz , j. m. mart and h. sol , _ astron .
_ , * 427 * , 415 ( 2004 ) .
m. perucho , j. m. mart and m. hanasz , _ astron .
_ , * 427 * , 431 ( 2004 ) . m. perucho , m. hanasz , j. m. mart and j. a. miralles , _ phys . rev .
e _ , 75e6312 ( 2007 ) .
e. hardee and p. a. hughes , _ astrophys . j. _ , * 583 * , 116 ( 2003 ) .
v. urpin , _ astron .
_ , * 385 * , 14 ( 2002 ) .
p. e. hardee , _ astrophys .
j. _ , * 664 * , 26 ( 2007 ) .
m. perucho , j. m. mart , j. m. cela , m. hanasz , r. de la cruz and f. rubio , _ astron .
_ , * 519a * , 41 ( 2010 ) . j. xu , p. e. hardee and j. stone , _ astrophys .
j. _ , * 543 * , 161 ( 2000 ) .
r. a. daly and a. p. marscher , _ astrophys .
j. _ , * 334 * , 539 ( 1988 ) .
s. a. e. g. falle , _ mon . not .
r. astron .
_ , * 250 * , 581 ( 1991 ) .
i. agudo , j. l. gmez , j. m. mart et al . ,
_ astrophys .
j. _ , * 549 * , l183 ( 2001 ) .
m. perucho and j. m. mart , _ mon . not .
r. astron .
_ , * 382 * , 526 ( 2007 ) .
a. t. araudo , v. bosch - ramon , g. e. romero , _ astron .
_ , * 522 * , 97 ( 2010 ) . m. v. barkov , v. bosch - ramon , f. aharonian , _ astrophys .
j. _ , * 724 * , 1517 ( 2010 ) .
s. s. komissarov , _ mon . not .
r. astron .
_ , * 269 * , 394 ( 1994 ) .
r. a. laing and a. h. bridle , _ mon . not .
r. astron .
_ , * 336 * , 1161 ( 2002 ) .
a. hubbard and e. g. blackman , _ mon . not .
r. astron .
_ , * 371 * , 1717 ( 2006 ) .
gopal - krishna , and p. j. wiita , _ astron .
_ , * 363 * , 507 ( 2000 ) .
z. meliani , r. keppens and b. giacomazzo , _ astron .
astrophys . , _
* 491 * , 321 ( 2008 ) p. rossi , a. mignone , g. bodo , s. massaglia and a. ferrari , _ astron .
astrophys . ,
_ * 488 * , 795 ( 2008 ) .
m. kadler , e. ros , m. perucho , et al .
, _ astrophys .
j. _ , * 680 * , 867 ( 2008 ) .
m. perucho , i. agudo , j. l. gmez , m. kadler , e. ros and y. y. kovalev , _ astron .
_ , * 489 * , l29 ( 2008 ) .
p. mimica , m. a. aloy , i. agudo , j. m. mart , j. l. gmez and j. a. miralles , _ astrophys . j. _
, * 696 * , 1142 ( 2009 ) .
r. a. laing , a. h. bridle , p. parma , l. feretti , g. giovannini , m. murgia , m. and perley , r. a. , _ mon . not .
r. astron .
_ , * 386 * , 657 ( 2008 ) . v. bosch - ramon , m. perucho and p. bordas , _ astron
. astrophys .
_ , * 528a * , 89 ( 2011 ) . m. perucho and v. bosch - ramon , _ astron .
_ , * 482 * , 917 ( 2008 ) .
m. perucho , v. bosch - ramon and d. khangulyan , _ astron .
_ , * 512 * , l4 ( 2010 )
. m. perucho and v. bosch - ramon , _ astron .
_ , in preparation .
g. giovannini , w. d. cotton , l. feretti , l. lara and t. venturi , _ astrophys .
j. _ * 552 * , 508 ( 2001 ) . c. a. hummel , t. w. b. muxlow , t. p. krichbaum , t.p .
, et al . , _ astron .
_ , * 266 * , 93 ( 1992 ) . y. wang , c. knigge , j. h. croston and g. pavlovski , the entrainment - limited evolution of frii sources : maximum sizes and a possible connection to fris , to appear in _ mon . not
r. astron .
_ ( arxiv:1108.0558 ) .
m. perucho , v. quilis and j. m. mart , intracluster medium reheating by relativistic jets , to appear in _ astrophys . j. _ ( arxiv:1107.5484 ) . | relativistic jets carry energy and particles from compact to very large scales compared with their initial radius .
this is possible due to their remarkable collimation despite their intrinsic unstable nature . in this contribution ,
i review the state - of - the - art of our knowledge on instabilities growing in those jets and several stabilising mechanisms that may give an answer to the question of the stability of jets .
in particular , during the last years we have learned that the limit imposed by the speed of light sets a maximum amplitude to the instabilities , contrary to the case of classical jets . on top of this stabilising mechanism ,
the fast growth of unstable modes with small wavelengths prevents the total disruption and entrainment of jets .
i also review several non - linear processes that can have an effect on the collimation of extragalactic and microquasar jets . within those
, i remark possible causes for the decollimation and deceleration of fri jets , as opposed to the collimated frii s .
finally , i give a summary of the main reasons why jets can propagate through such long distances . |
study of the probability distribution of wealth of the people in a society goes back to 1897 when pareto observed empirically that the distribution is characterized by a power law tail .
probability that an individual member has wealth more than @xmath0 is given by @xmath1 with @xmath2 @xcite .
this type of distribution is known as the pareto distribution @xcite .
this observation reflects the inherent inequality in the economic structure of the society .
a large number of individuals are economically poor . in comparison
the number of wealthier people is less and their number gradually decreases as their wealth increases .
the cut - off of the distribution corresponds to with few very rich individuals .
consequently a sizable fraction of the society s net wealth is infact possessed by a class of top rich people consisting only a few percent of society s entire population . since measuring wealth is difficult , in recent years distributions of income and tax return amounts have been studied in different countries .
for example tax return amount distribution in us and japan shows a log - normal distribution in the middle range followed by a power law for high income people @xcite , uk data of income shows an exponential decay which is followed by a power law in the high income range @xcite and income data in brazil for 2004 shows an almost gaussian law for the low and middle income groups where as high salary groups are described approximately by power law @xcite .
a good amount of research effort has been devoted in recent times to study the wealth distribution using the ideas of statistical physics . in this description
an individual member is called an agent .
the microstate specified by the precise description of wealth of every agent changes after each transaction . their wealth change due to interaction among themselves .
this interaction is the mutual bipartite trade among different pairs of agents .
thus the wealth distribution evolves due to such repeated interactions and finally assumes a time independent form . in a simple model
the total wealth of the entire society has been considered as strictly conserved . in a transaction
the net amount of wealth of the pair of agents is randomly re - shuffled between them and therefore the pairwise interactions also maintain conservation of wealth @xcite .
starting from an arbitrary initial distribution of wealth with a fixed average value the system attains a steady state .
the steady state wealth distribution has been found to have an exponentially decaying tail .
such distribution was also observed in @xcite . in a subsequent modification of
the model saving propensity factor @xmath3 has been introduced where each trader invests only @xmath4 fraction of his wealth to the bipartite trade @xcite . with a fixed value of @xmath3
the steady state wealth distribution becomes very similar to the gamma distribution @xcite .
this model was further extended by assigning quenched saving propensities @xmath5 specific to each individual agent @xcite .
here one gets a power law tail with @xmath6 only when wealth distributions are averaged over different @xmath7 sets @xcite .
all these results exhibit economic inequality in the society .
a review of all these models has been published in @xcite .
apart from these a different idea of extremal dynamics for the evolution of wealth distribution has been studied by pianegonda et .
al . on an one dimensional ( @xmath8 ) lattice @xcite .
it is assumed that always the poorest agent of the society initiates a trade since he feels the strongest urge to raise its economic status . the trade is implemented by locating the poorest agent and refreshing his wealth randomly . to ensure that the model remains strictly conserved , the amount of wealth gained by the minimal site has been taken out equally from its two neighbors .
consequently this model allows an agent to possess a negative wealth . in the stationary state of large systems @xmath9 jumps from zero to a maximal value at a critical value of @xmath10 and
then it decays as @xmath0 increases as the boltzmann - gibbs exponential function .
the pianegonda model is very similar to the non - conservative self - organized extremal model for the ecological evolution of interacting species introduced by bak and sneppen ( bs ) @xcite .
the phenomenon of self - organized criticality ( soc ) is the spontaneous emergence of fluctuations of all length and time scales in a slowly driven system .
this concept was first introduced to describe the formation of a sandpile of a fixed shape @xcite .
later the idea of soc has been applied to a large number of different physical systems @xcite . in a stochastic version of the sandpile model grains
are distributed to randomly selected neighboring sites @xcite . in the soc models fluctuations
are described in terms of avalanches of activities and their size distributions assume power law decaying functions for large system sizes . in bs model an entire species is represented by a single fitness variable . a set of species is represented by the nodes of a graph . using the spirit of darwinian principle in each mutation the fitness of the node with globally minimal value
is searched and is refreshed by a new random value .
effect of this propagates to few neighboring nodes which are also refreshed .
the system eventually reaches a stationary state when the fitness distribution assumes a time independent step like form .
an absorbing state phase transition in presence of a conserved continuous local field has been studied recently @xcite . in this model
a pair of sites is said to be active if at least one of them has energy more than a pre - assigned threshold .
an active pair re - shuffles its net energy between them keeping the energy strictly conserved . beyond a critical value of the threshold
the number of active pairs fluctuates with time in the stationary state and the time averaged density of active pairs has been considered as the order parameter for the problem . as the threshold value is tuned a continuous phase transition is observed from an absorbing phase to an active phase @xcite .
it was claimed that the critical exponents of this model are different from the directed percolation model . in a related model defined for the wealth distribution at least one of the agents in a bipartite trading
is selected from a subset of agents @xcite .
this subset is formed by agents who have wealth less than certain upper cutoff , the other agent being selected randomly from the neighbors of the first agent . in a transaction
the net wealth of the pair of agents is randomly re - shuffled between them .
the order parameter has been defined as the fraction of agents having wealth below a certain threshold value and it is claimed that the system undergoes a continuous phase transition at a critical value of the threshold wealth .
a number of critical exponents have been measured to characterize the transition and some of them are found to be close to corresponding exponents in the manna model of self - organized critically @xcite .
in this paper we present a detailed numerical analysis of the modified pianegonda et .
al . model whose trading rule has been replaced by the stochastic re - shuffling of the net wealth of a pair of agents .
we consider this model as one of the few examples of non - dissipative soc systems where the entire wealth of the society is strictly conserved , for example the fixed energy sandpile @xcite .
estimation of a number of critical exponents of the modified pianegonda et .
model suggests that this model does not belong to the universality class of either bs model or manna model .
we believe that it belongs to a new universality class perhaps because of strict conservation of wealth is maintained in its dynamical rules . the model studied by us in this paper and that in @xcite are essentially same but studied from two different approaches .
an approach which is very typical of a soc system has been followed by us where the critical poverty line spontaneously evolves without any fine tuning .
in contrast in @xcite the position of the poverty line is tuned by hand to arrive at the critical state .
( black ) , @xmath11 ( red ) and @xmath12 ( blue ) .
( a ) the relaxation time @xmath13 has been plotted with the corresponding position of the poverty line @xmath14 .
( b ) the same data has been plotted with deviation @xmath15 on a @xmath16 scale .
the slopes are -2.30 , -2.64 and -2.76 respectively , which on extrapolation give the value of the dynamical exponent @xmath17 .
, width=264 ] in section 2 we describe the minimal wealth model where the trader with minimal wealth initiates the trade . in subsection 2.1
we discuss the relaxation dynamics of the system on its way to the stationary state . in subsection 2.2
the correlation that evolves in the system in the stationary state is studied .
wealth distribution in the stationary state is described in subsection 2.3 .
the statistics of the avalanche life - time distributions has been studied in section 2.4 .
this section ends with the study of persistence time distribution of individual agent s wealth in subsection 2.5 . in section 3
we have described the maximal wealth model , which is opposite to the minimal wealth model , where the trader with maximal wealth initiates the trade .
finally we conclude in section 4
. per agent has been plotted with time @xmath18 for @xmath19 ( black ) , @xmath20 ( red ) and @xmath12 ( blue ) and after a long time they approach the stationary state value @xmath21 .
( b ) the scaled deviation @xmath22n^{0.55}$ ] has been plotted with the scaled variable @xmath23 .
the relaxation exponent @xmath24 .
, width=264 ]
in this paper we have considered a model with a conservative extremal dynamics for studying the evolution of wealth distribution in a society . in a bipartite transaction
one agent is necessarily selected as the one with the globally minimal value of wealth @xmath25 .
the second agent is chosen randomly with uniform probability from neighbors of the first agent .
this neighbor list has been defined in different ways for different graphs .
we have studied this model on four different graphs , namely , ( i ) @xmath8 regular lattice with periodic boundary condition ( ii ) two dimensional @xmath26 square lattice with periodic boundary condition ( iii ) the barabsi - albert ( ba ) scale - free graph @xcite and on an ( iv ) @xmath27-clique graph . on every graph
the nodes represent agents and the nearest neighbors of each node connected by direct links constitute the neighbor list of every agent .
we report elaborately the results of our model on an @xmath8 lattice and mention the key results of the same model studied on other graphs in tables .
the dynamics starts with @xmath27 agents , each having an amount of wealth @xmath28 drawn from a uniform distribution with the average @xmath29 = 1 irrespective of the system size @xmath27 .
the discrete time @xmath18 is the number of bipartite transactions . at an arbitrary time
@xmath18 first the site @xmath30 is searched out which has the minimal wealth @xmath25 .
the other agent @xmath31 is selected randomly with uniform probability from the neighbors of @xmath32 .
both agents invest their entire amount of wealth .
therefore the total invested amount by both the traders is : @xmath33 .
this amount is then randomly divided into two parts and received by them also randomly : @xmath34 where @xmath35 is a freshly generated random fraction and @xmath36 .
these transactions are repeated ad infinitum .
after some relaxation time the system reaches a stationary state when the wealth distribution assumes a time independent form . with globally minimal wealth in successive time steps
are marked for a system of size @xmath27 = 512 in @xmath8 .
the gradually increasing duration of correlation has been exhibited by time windows of increasing lengths : ( a ) 500 , ( b ) 3000 and ( c ) 6000 .
, width=321 ] in @xmath8 a linear chain of @xmath27 sites with periodic boundary condition has been considered where the neighbor list of an arbitrary site @xmath37 consists of the two nearest neighboring sites at @xmath38 .
therefore the second agent @xmath31 is selected randomly with equal probability from this list .
if @xmath25 is very small then the probability that it will be replaced by an even smaller wealth after trade is also small .
however this probability gradually increases as @xmath25 increases . as the sites with the minimal values of wealth are systematically replaced , very soon all nodes with small @xmath0 values are replaced by larger values of @xmath0 resulting a vacancy in the small @xmath0 region .
this is explained pictorially in fig .
1 . on an @xmath8 lattice with @xmath27 = 512
we plot the lattice positions @xmath37 along the abscissa and the corresponding wealth @xmath39 along the ordinate . as time increases a vacant region gradually forms for small values of @xmath0 for all sites .
if on the average the wealth of none of the agent is below a certain threshold value @xmath0 , it is called the ` poverty line ' . in fig .
1 the poverty line gradually moves up with time and finally settles at a critical value @xmath40 at the stationary state .
this behavior is the same for all system sizes @xmath27 but with different values of @xmath41 . unlike the model in @xcite here
the critical poverty line is spontaneously determined by the dynamical evolution of the system where no fine tuning is necessary which is the distinctive signature of self - organization and we will see in the following that the model exhibits critical behavior as well .
that sites with minimal wealth at successive time steps are separated by a distance @xmath42 has been plotted with @xmath42 for the system sizes @xmath43 ( black ) , @xmath20 ( red ) and @xmath44 ( blue ) .
the slopes are -2.64 , -2.81 and -2.89.(b ) the finite - size scaling of @xmath45 vs. @xmath46 with @xmath47 = 2.89 and @xmath48 and therefore @xmath49 .
, width=264 ] to find the agent with minimal wealth a brute - force search takes cpu @xmath50 .
a much faster algorithm to search for the globally extremal ( minimal or maximal ) site was proposed by grassberger @xcite which stores the data in a hierarchical structure .
this takes cpu @xmath51 .
we have used this method for @xmath8 , @xmath52 and for @xmath27-clique graphs . for ba graphs we used the brute - force algorithm .
we first estimate the relaxation time required for the system to reach the stationary state . during this relaxation period
the wealth distribution gradually changes starting from its initial uniform distribution to its time independent form in the stationary state .
the relaxation time has been estimated as a function of deviation of the poverty line from its critical value in two ways . .
( a ) the wealth distribution @xmath53 in the stationary state for @xmath54 ( black ) , @xmath20 ( red ) and @xmath44 ( blue ) .
( b ) a finite - size scaling using @xmath55 vs. @xmath56n^{0.73}$ ] for system sizes @xmath57 ( black ) , @xmath12 ( red ) and @xmath44 ( blue ) .
, width=264 ] for a given system size @xmath27 we have simulated our model up to @xmath58 time steps and calculated the wealth distribution @xmath59 at 100 time instants @xmath13 at the interval of @xmath60 steps .
these distributions are calculated for a sample size of @xmath61 independent runs .
the value of the poverty line @xmath14 at time @xmath13 is determined by the value of @xmath0 where @xmath59 is the maximum .
this estimation is done for all values of @xmath13 . in fig .
2(a ) we plot on a @xmath62 scale @xmath13 vs. @xmath14 for @xmath63 , @xmath11 and @xmath12 .
we see that in each case the relaxation time @xmath13 diverges as @xmath14 approaches its stationary state value @xmath41 .
these data have been replotted in fig .
2(b ) using a @xmath16 scale as @xmath13 vs. @xmath15 where we have used @xmath41 = 0.8242 , 0.8375 and 0.8383 to obtain the best straight line plots .
the slopes of these straight lines are 2.30 , 2.64 and 2.76 respectively which are then extrapolated with @xmath64 to obtain the exponent @xmath65 as @xmath66^{-z}$ ] with @xmath17 . in a second method we calculated @xmath67 with time @xmath18 starting from its initial value
when the distribution is uniform . after a long time
this quantity saturates to its stationary value @xmath21 . in fig .
3(a ) we show the plots of @xmath67 vs. @xmath18 on a @xmath16 scale . in fig .
3(b ) @xmath22n^{0.55}$ ] has been plotted with the scaled value of time @xmath23 using @xmath21 = 0.057 , 0.0607 and 0.061 for @xmath68 and @xmath12 respectively .
a nice data collapse has been obtained with the following scaling form @xmath69n^{0.55 } \sim { \cal f}_{z}(tn^{-2.5})\ ] ] where the scaling function @xmath70 varies as @xmath71 for small @xmath72 . a direct measurement of the slope of the scaled plot gives an estimate of @xmath73 which corresponds to @xmath24 .
we conclude an average value of @xmath74 . in the stationary state for seven different system sizes @xmath54 to @xmath44 in @xmath8 .
the sequence starts with the distribution for @xmath54 at the extreme left and is shifted to the right by 0.2 when system size is multiplied by a factor of 2 .
, width=264 ] starting from an uncorrelated wealth distribution the system becomes more and more correlated as time passes .
this is reflected in the fact that the probability of occurrence of the minimal sites close to each other in successive time steps gradually increases . at the early uncorrelated stage
the position of the minimal site at the next time step is likely to be anywhere in the lattice .
however as time increases , the poverty line moves up , consequently @xmath25 increases and the probability that the minimal site at the next time step occurring at the same site or at the neighboring updated site also increases .
this is shown in fig .
4 using a @xmath8 lattice of @xmath75 . for a single run
it is observed that the locations of @xmath25 are quite random ( fig .
however as time evolves these positions gradually become more correlated ( figs .
4(b ) and 4(c ) ) .
in general one can consider the successive jumps of @xmath32 positions constituting a lvy flight random walk @xcite .
we see below that indeed their flight lengths follow a power law distribution .
the correlation in the stationary state is quantitatively measured by the probability distribution @xmath76 of the distance of separation @xmath77 between successive minimal sites using periodic boundary condition .
this distribution measured in the stationary state has been plotted in fig .
5(a ) for different system sizes @xmath78 and @xmath44 . the value of @xmath76 for @xmath77 = 0 and 1 are approximately 0.4575(1 ) and 0.4820(1 ) and then it decreases as a power law @xmath79 with increasing @xmath77 .
a direct measurement of slope gives @xmath80 .
5(b ) exhibits the finite - size scaling of the same data when the @xmath77 and @xmath81 axes are scaled as : @xmath82 where @xmath83 is a universal scaling function with the scaling exponents @xmath84 and @xmath85 . from this scaling analysis
we get @xmath86 . .
the @xmath87 are the upper plot ( in red ) and @xmath88 are the lower plot ( in blue ) . when they are extrapolated with @xmath89 with @xmath90 and @xmath91 ,
@xmath92 is obtained as 0.8174 and 0.8176 respectively .
we conclude @xmath92 = 0.8175(2 ) .
, width=264 ] there exists a spatial correlation too in this model .
a two point correlation function has been measured in the stationary state .
the average correlation between two sites situated at a distance of separation @xmath72 has been defined as : @xmath93 where @xmath94 is always set equal to unity .
we assume a power law decay for the correlation , i.e. , @xmath95 for @xmath96 .
for @xmath8 a plot of ( not shown ) @xmath97 vs. @xmath72 on a @xmath16 scale indicates a power law for large @xmath72 values .
however considerable variation of slopes exists for system sizes @xmath98 and @xmath12 .
the slopes are : -1.17 , -1.29 .
-1.34 respectively which extrapolates to a value of @xmath99 in the limit of @xmath100 .
our estimate for @xmath101 in @xmath52 is 2.2(2 ) .
system with @xmath27 = 256 and @xmath102 in the stationary state .
it shows that depending on the value of @xmath103 an avalanche can be broken into a hierarchy of avalanches . for this run
the system has been relaxed for the initial @xmath104 time steps .
, width=302 ] next we estimated the probability density distribution @xmath53 of wealth in the stationary state of the system of size @xmath27 .
the distribution grows very rapidly near @xmath41 for all values of @xmath27 and then decays very fast . in fig .
6(a ) we show the plot of @xmath53 vs. @xmath0 for @xmath105 and @xmath44 .
all of them have similar variations but with increasing system size the curves gradually become sharper .
therefore we tried a finite - size scaling analysis in fig .
6(b ) for the growing region and for @xmath106 and @xmath44 .
a nice data collapse is observed when axes are scaled and @xmath55 has been plotted with @xmath56n^{0.73}$ ] . the functional form of the decay of the probability distribution @xmath53 has been studied right after the maximal jump at @xmath41 . this part fits very well with the gaussian form : @xmath107\ ] ] in fig .
7 we showed @xmath53 vs. @xmath0 on a @xmath62 scale for seven different system sizes : @xmath108 . in each case
the fitting curve is indistinguishable from the data .
we observe a systematic variation of @xmath109 , @xmath110 and @xmath111 with system size @xmath27 .
for example @xmath112 , @xmath113 and @xmath114 .
the precise value of @xmath92 is calculated by extrapolating @xmath41 values which are calculated using the following two methods .
we have seen in fig .
6(a ) that the probability distribution of @xmath53 becomes increasingly steeper with increasing @xmath27 . for a certain size @xmath27 we defined @xmath87 as the value of @xmath0 for which @xmath53 is half of its maximum value . in a second method
the @xmath41 has been calculated in the following way .
a pair of successive points on the @xmath53 vs. @xmath0 curve which has the largest slope is found out . a straight line joining these two points
is then extrapolated to meet the @xmath0 axis at @xmath88 .
the pair of values of @xmath87 and @xmath88 for eight @xmath27 values @xmath115 to @xmath116 are then extrapolated with @xmath89 .
a least square fit of straight line has been done for trial values of @xmath117 starting from 0.20 to 1.20 at an interval of 0.001 and the errors have been calculated .
the errors are minimal for @xmath90 and @xmath91 . using these two values of the exponent
@xmath117 we extrapolate @xmath41 values with @xmath89 in fig . 8 to meet the @xmath41 axis at 0.8174 and 0.8176 respectively .
we conclude a value for @xmath92 for @xmath8 model as 0.8175(2 ) . minimal wealth model .
( a ) the avalanche size distribution for @xmath118 ( black ) , @xmath44 ( red ) and @xmath116 ( blue ) .
( b ) a finite - size scaling of the data in ( a ) with scaling exponents @xmath119 and @xmath120 giving the avalanche size exponent @xmath121 . , width=264 ] in the stationary state successive values of minimal wealth @xmath25 fluctuates with time .
if a certain reference wealth is fixed by hand at @xmath122 then the successive @xmath25 appear below and above @xmath103 line .
one defines a @xmath103-avalanche as the sequence of successive bipartite trades whose @xmath25 values are smaller than @xmath103 .
the size @xmath123 of the avalanche measures the duration of the avalanche i.e. , at times @xmath18 and @xmath124 the @xmath125 , whereas at every time step from @xmath126 to @xmath127 the @xmath128 .
when @xmath103 is set equal to @xmath41 it is called a critical avalanche .
this is explained in fig .
9 where part of the @xmath25 time series for @xmath129 and @xmath130 = 0.8010 is displayed discarding the initial @xmath131 time steps . for @xmath132
an avalanche of size 53 breaks into two avalanches of sizes 25 and and 27 when @xmath103 is reduced to 0.66 . on further reduction of @xmath103 to 0.56
these two avalanche break into even smaller avalanches of sizes 18 , 6 and 10 , 16 respectively .
thus any avalanche can be splitted into a hierarchy of smaller avalanches if @xmath103 value is lowered @xcite . on the other hand if @xmath103 is raised the average avalanche size increases and becomes infinite at certain value of @xmath103 .
right at the poverty line @xmath41 has been plotted with system size @xmath27 on a @xmath133 scale .
the slope of this curve gives the exponent @xmath134 . , width=264 ] at the critical point the distribution of the avalanche life - times has a power law tail in the limit of @xmath100 : @xmath135 . in the stationary state we used @xmath136 and
measured life - times of a large number of avalanches for different system sizes and plot the probability distributions @xmath137 vs. @xmath123 using a @xmath16 scale in fig .
each curve has a straight portion in the intermediate regime of the avalanche sizes and this regime becomes gradually larger on increasing @xmath27 .
the direct measurement of slopes in the scaling region gives @xmath138 = 1.086 , 1.091 and 1.096 for @xmath139 and @xmath116 respectively .
a finite - size scaling is very much suitable with the following scaling form : @xmath140 where the scaling function @xmath141 in the limit of @xmath142 and @xmath143 approaches zero very fast for @xmath144 .
the exponents @xmath145 and @xmath146 fully characterize the scaling of @xmath147 in this case .
an immediate way to check the validity of this equation is to attempt a data collapse by plotting @xmath148 vs. @xmath149 with trial values of the scaling exponents .
for @xmath8 the values for obtaining the best data collapse are found to be @xmath119 and @xmath120 ( fig .
the life - time distribution exponent for @xmath8 is therefore @xmath150 .
next we calculated the average value of avalanche life - times @xmath151 right at the critical poverty line . in fig .
11 we plot this quantity with system size @xmath27 on a @xmath16 scale .
the plot fits excellent to a straight line and its slope gives the value of the exponent @xmath152 in : @xmath153 . assuming the distribution @xmath147 of avalanche sizes holds good up to a cut - off @xmath154 one gets a scaling relation @xmath155 and our estimates of @xmath156 , @xmath120 and @xmath157 satisfy this relation very closely .
has been plotted with deviation @xmath158 from the poverty line @xmath41 .
the system sizes used are @xmath159 ( black ) , @xmath20 ( blue ) , @xmath12 ( green ) and @xmath116 ( red ) .
the value of @xmath160 obtained by extrapolation of slopes is 2.67 .
( b ) finite size scaling analysis of the data in ( a ) is shown .
the scaling exponents @xmath161 and @xmath162 gives @xmath163 .
( c ) plot of the data in ( a ) but with @xmath164 .
the slopes for four different system sizes in ( a ) on extrapolation gives @xmath165 .
we conclude @xmath166 .
, width=264 ] the size of the @xmath103-avalanches are smaller when @xmath167 and we have studied how the average avalanche size grows as the deviation @xmath168 decreases .
similar to the bs model we assume @xmath169^{-\gamma}$ ] for @xmath100 .
we measured the average size @xmath170 of the @xmath103-avalanches for different system sizes @xmath27 and plotted them with @xmath158 in fig .
12(a ) with @xmath171 . for all plots on @xmath16 scale
the curves are horizontal as deviation @xmath172 is very small .
however as the deviation increases the curves become straight with negative slopes -1.98 , -2.15 , -2.28 and -2.38 for for @xmath173 and @xmath116 respectively . these values when extrapolated with @xmath174 give @xmath175 for @xmath100 . again
a finite - size scaling has been possible as shown in fig .
12(b ) : @xmath176n^{\zeta_{\gamma}})\ ] ] where @xmath177 is another scaling function . from this data collapse
the scaling exponents @xmath161 and @xmath162 with @xmath178 is obtained . of persistence times @xmath179 at the stationary state .
( a ) the finite - size scaling of the distribution in @xmath8 for @xmath63 ( black ) , @xmath12 ( red ) and @xmath116 ( blue ) . from the scaling exponents
@xmath180 and @xmath181 the persistence exponent @xmath182 is obtained .
( b ) the finite - size scaling of the distribution in @xmath52 for @xmath183 ( black ) , @xmath184 ( red ) and @xmath20 ( blue ) . from the scaling exponents
@xmath185 and @xmath186 the persistence exponent @xmath187 is obtained .
, width=264 ] for every system size @xmath27 there is a value of @xmath188 so that when @xmath103 is raised to this value the avalanche size becomes infinite .
this implies that if we plot the data in fig .
12(a ) with respect to @xmath189 then we should be able to see the divergence of average avalanche size instead of saturation of the avalanche sizes .
we plot this in fig .
12(c ) using @xmath190 = 0.8167 , 0.8169 , 0.8170 , 0.8172 for @xmath159 , @xmath20 , @xmath12 and @xmath116 respectively again on a @xmath16 scale .
each curve is a straight line but with different slopes : -2.31 , -2.43 , -2.51 and -2.56 respectively .
when these slopes are extrapolated with @xmath191 the extrapolated value for @xmath100 is -2.66 .
our conclusion for the value of the exponent @xmath166 . of different agents in the stationary state are plotted with their positions @xmath37 along a @xmath8 lattice for the ( a ) maximal wealth model , the red line is at @xmath192 .
( b ) mixture of the minimal wealth and maximal wealth models with probability @xmath193 , the red lines correspond to peaks at @xmath194 and at 1.3574 .
, width=264 ] the time interval between two successive updates of wealth of an agent is known as the persistence time @xmath179 .
different agents have to wait for different amounts of times in general .
more specifically agents having small amount of wealth have to wait for very little times .
on the other hand potentially rich agents have to wait long enough times . in the stationary state
we measure the persistence times for all sites of the lattice and use this data to plot their probability distribution .
more precisely we set a clock to each site .
whenever there is a change of wealth at this site the time is noted and the clock time is reset to zero . at the stationary state we collect a large number of persistence time data and use these data to calculate the persistence time distribution .
we assume a power law variation of the persistence time distribution as @xmath195 in the limit of @xmath100 . for finite size systems
the distributions @xmath196 vs. @xmath179 are plotted on a @xmath16 scale ( not shown ) and the direct measurement of slopes give the @xmath197 values for finite size systems .
these values are extrapolated as @xmath198 to obtain @xmath199 = 1.539 for @xmath100 in @xmath8 . in a similar analysis for a @xmath52 square lattice of size @xmath200 using an extrapolation with respect to @xmath201 we get @xmath199 = 1.25 for @xmath202 .
persistence exponents are also obtained by the finite - size scaling analysis . in fig .
13(a ) we show the scaling plot of @xmath203 vs. @xmath204 with @xmath205 and @xmath206 .
this gives @xmath207 in @xmath8 .
similar scaling analysis in terms of the system size @xmath200 in @xmath52 square lattice has been performed with @xmath208 and @xmath209 which gives @xmath199 = 1.274 ( fig .
13(b ) ) . averaging @xmath199 values obtained from direct measurement and scaling analysis we conclude @xmath210 in @xmath8 and @xmath211 in @xmath52 .
[ cols="^ , > , > , > , > , > , > , > , > , > , > " , ]
next we studied the maximal wealth model where one agent is necessarily the agent with maximal wealth . the other agent
being selected randomly with uniform probability from the neighbors of the first agent .
random re - shuffling of wealth takes place in the same way as in the minimal wealth model . vs. @xmath0 for the maximal wealth model for @xmath27 = @xmath212 ( black ) , @xmath20 ( red ) , @xmath12 ( green ) and for @xmath116 ( blue ) .
( b ) finite - size scaling of the avalanche size distribution @xmath137 at the critical threshold @xmath41 for @xmath27 = @xmath212 ( black ) , @xmath20 ( red ) and @xmath12 ( blue ) .
the scaling exponents are @xmath213 and @xmath214 which gives the exponent @xmath215 .
, width=264 ] and @xmath216 respectively for @xmath217 and for @xmath218 = 0.02 ( black ) , 0.08 ( red ) , 0.2 ( green ) , 0.4 ( blue ) , 0.6 ( magenta ) and 0.9 ( maroon ) .
( b ) variation of the ratio @xmath219 of heights of the right peak and the left peak with probability @xmath218 .
, width=264 ] in fig.14(a ) we plot again for the maximal wealth model the values of wealth @xmath39 of different agents at a certain instant of time in the stationary state with their positions @xmath37 along an @xmath8 lattice of size @xmath75 .
in contrast to the similar plot of the minimal wealth model in fig .
1 here an upper cut - off for the wealth has been visible at @xmath220 . in this case
the stationary state wealth distribution @xmath53 takes an opposite shape ( fig .
15(a ) ) . a critical wealth @xmath41 exists here as well .
@xmath53 takes a gaussian form elevated by a constant term @xmath221 for @xmath222 , whereas for @xmath223 it sharply decreases to zero .
the parameters of the gaussian function ( eqn .
( 4 ) ) are different for different @xmath27 and they vary very systematically with @xmath27 as : @xmath224 , @xmath225 and @xmath226 and the constant @xmath227 . the critical value of wealth @xmath92 in the asymptotic limit of system sizes has been estimated by the same method as used for the minimal wealth model . the @xmath87 and @xmath88 values have been calculated for @xmath159 , @xmath20 , @xmath12 and @xmath116 , extrapolated with @xmath228 and @xmath229 and the asymptotic values are 1.3610 and 1.3608 respectively .
we conclude @xmath230 for @xmath8 .
a similar analysis gives @xmath231 for @xmath52 square lattice , 1.8895(2 ) for the ba graph and 1.9998(2 ) for the @xmath27-clique .
it may appear that the minimal wealth and maximal wealth models should be symmetric about the average wealth per agent which we have set at @xmath232 .
we have seen above that this indeed not the case since @xmath10 values are 0.8175 and 1.3609 for the minimal wealth and maximal wealth models respectively .
the symmetry between these two models are broken by the presence of a rigid wall at @xmath233 which means that negative value of wealth of an agent is not allowed .
the avalanche size distributions have been studied as well .
a finite - size scaling of these distributions has been done and are plotted in fig .
15(b ) using @xmath16 scale as before for @xmath68 and @xmath12 .
the scaling exponents are @xmath213 and @xmath214 respectively giving the value of the avalanche size exponent @xmath215 .
finally we have studied a mixture of the minimal wealth and maximal wealth models . at every bipartite trade the first agent with minimal wealth
is selected with probability @xmath218 or with maximal wealth with probability @xmath216 .
the second agent is selected with uniform probability from the neighbors of the first agent .
a snapshot of the individual wealth for @xmath193 at the stationary state for different agents has been shown in fig .
14(b ) . here
the wealth values are restriced within a ` wealth - band ' with sharp cut - offs at a high and a low end .
consequently the shape of the wealth distribution @xmath234 has double peaks for all @xmath27 and we plot the distribution in fig .
16(a ) for @xmath217 .
portion of the distribution between the peaks fits excellent to elevated gaussian distributions with different parameter values for different values of @xmath218 in the range between 1/2 and 1 . on the two sides of this region the distribution decays to zero very fast .
the figure shows the plot of @xmath234 vs. @xmath0 for @xmath218 = 0.02 , 0.08 , 0.2 , 0.4 , 0.6 and 0.9 . in fig .
16(b ) we plot the ratio @xmath219 of the heights of right peak and the left peak with the probability @xmath218 .
social inequality in terms of economic strengths is ubiquitous for the people of all countries .
perhaps this inequality acts as the major driving force behind the advancement of society .
consequently the mechanism that establishes this inequality in a society is an important issue and attracted the attention of researchers over the last century .
here we have studied a modified version of the conservative self - organized extremal model introduced by pianegonda et .
al . which is motivated by the wealth distribution in a society . in this model the entire wealth of the society is strictly conserved .
it evolves by a trade dynamics that takes the society from equality ( or any other initial wealth distribution ) to a stationary state where strong social inequality is present .
the dynamics is an infinite sequence of stochastic bipartite tradings where one of the agents has the globally minimal value of wealth , the other one being selected randomly from the neighbors of the first agent .
our numerical study reveals that this model is one of the simplest models of self - organized criticality where the stationary state is non - ergodic .
this model is very similar to the self - organized critical bak - sneppen model for the ecological evolution of interacting species . using numerical simulation
we have estimated a number of critical exponents for this model on an @xmath8 regular lattice , @xmath52 square lattice , the barabsi - albert scale - free graph and on the @xmath27-clique graph .
we present evidences which suggest that this model does not belong to the universality class of either the bak - sneppen model or the manna model of self - organized criticality .
this model belongs to a new universality class perhaps because of strict conservation of wealth is maintained in its dynamical rules .
0.3 cm g. mukherjee thankfully acknowledges the associateship in s. n. bose national centre for basic sciences , kolkata under the extended visitor and linkage programme .
comments from b. k. chakrabarti , a. chatterjee , d. dhar , j. r. iglesias , j. kertesz , p. k. mohanty and p. pradhan are thankfully acknowledged .
90 v. pareto , _
cours deconomie politique _ ( f. rouge , lausanne , 1897 ) . `
en.wikipedia.org/wiki/pareto_distribution ` .
w. souma , fractals , * 9 * , 463 ( 2001 ) .
a. a. drgulescu , v. m. yakovenko , eur .
j. b * 299 * , 213 ( 2001 ) .
j. r. iglesias , science and culture , * 76 * , 437 ( 2010 ) .
a. a. drgulescu , v. m. yakovenko , eur .
j. b * 17 * , 723 ( 2000 ) . v. m. yakovenko and j. b. rosser , rev .
phys . * 81 * , 1703 ( 2009 ) .
j. angle , social forces * 65 * , 293 ( 1986 ) .
a. chakraborti , b. k. chakrabarti , eur .
b , * 17 * 167 ( 2000 ) .
k. bhattacharya , g. mukherjee and s. s. manna , in _ econophysics of wealth distributions _ , ( springer verlag , milan , 2005 ) .
m. patriarca , a. chakraborti , k. kaski , phys .
e * 70 * , 016104 ( 2004 ) .
a. chatterjee , b. k. chakrabarti , s. s. manna , physica a * 335 * 155 ( 2004 ) , a. chatterjee , b. k. chakrabarti , s. s. manna , physica scripta t * 106 * 36 ( 2003 ) .
p. k. mohanty , phys .
e , * 74 * , 011117 ( 2006 ) .
u. basu and p. k. mohanty , eur .
j. b * 65 * , 585 ( 2008 ) .
a. chakraborty and s. s. manna , phys .
e * 81 * , 016111 ( 2010 ) .
a. chatterjee and b. k. chakrabarti , euro .
journal b * 60 * , 135 ( 2007 ) .
s. pianegonda , j. r. iglesias , g. abramson , j. l. vega , physica a , * 322 * , 667 ( 2003 )
. j. r. iglesias , s. concalves , s. pianegonda , j. l. vega and g. abramson , physica a , * 327 * , 12 ( 2003 ) .
p. bak and k. sneppen , phys .
lett . * 71 * , 4083 ( 1993 ) .
p. bak , c. tang and k. wiesenfeld , phys .
lett . * 59 * , 381 ( 1987 ) .
p. bak , _ how nature works : the science of self - organized criticality _ , ( copernicus , new york , 1996 ) .
s.s . manna , j. phys .
a * 24 * , l363 ( 1991 ) .
m. basu , u. gayen and p. k. mohanty , arxiv:1102.1631 .
a. ghosh , u. basu , a. chakraborti and b. k. chakrabarti , phys .
rev . e. * 83 * , 061130 ( 2011 ) .
s. lbeck , int .
mod . phys .
b * 18 * , 3977 ( 2004 ) . a .-
barabsi and r. albert , science , * 286 * , 509 ( 1999 ) .
p. grassberger , phys .
a * 200 * , 277 ( 1995 ) .
s. s. manna , phys .
e. , * 80 * , 021132 ( 2009 ) .
b. b. mandelbrot , _ the fractal geometry of nature _ , ( 1982 ) w.h .
freeman , isbn 0 - 7167 - 186 - 9 .
m. paczuski , s. maslov and p. bak , phys .
e * 53 * , 414 ( 1996 ) .
s. lbeck and p. c. heger , phys .
rev . lett . * 90 * , 230601 ( 2003 ) .
h. n. huynh , g. pruessner and l. y. chew , j. stat .
mech . 09024 ( 2011 ) . | we present a detailed numerical analysis of the modified version of a conservative self - organized extremal model introduced by pianegonda et .
al . for the distribution of wealth of the people in a society . here
the trading process has been modified by the stochastic bipartite trading rule . more specifically in a trade one of the agents is necessarily the one with the globally minimal value of wealth , the other one being selected randomly from the neighbors of the first agent .
the pair of agents then randomly re - shuffle their entire amount of wealth without saving .
this model has most of the characteristics similar to the self - organized critical bak - sneppen model of evolutionary dynamics .
numerical estimates of a number of critical exponents indicate this model is likely to belong to a new universality class different from the well known models in the literature .
in addition the persistence time , which is the time interval between two successive updates of wealth of an agent has been observed to have a non - trivial power law distribution
. an opposite version of the model has also been studied where the agent with maximal wealth is selected instead of the one with minimal wealth , which however , exhibits similar behavior as the minimal wealth model . |
gamma - ray bursts ( grbs ) emit an amount of isotropic - equivalent energy @xmath4 ergs in @xmath5-rays and x - rays in a few seconds and subsequently emit afterglows at x - ray ( costa et al . 1997 ) , optical ( van paradijs et al .
1997 ) and radio bands ( frail et al . 1997 ) , which generally last days to months ( van paradijs , kouveliotou & wijers 2000 ) .
the energetics of grbs , which is comparable to that of supernovae , and their rapid variability strongly suggest compact objects involving black holes , neutron stars and strange stars as the energy source for grbs .
two popular models satisfying this energetics are explosive events of very massive stars , also named hypernovae ( paczyski 1998 ) or collapsars ( woosley 1993 ; macfadyen & woosley 1999 ) , and mergers of neutron - star binaries ( eichler et al . 1989 : narayan , paczyski & piran 1992 ) .
other possible models include phase transitions of neutron stars to strange stars ( cheng & dai 1996 ; dai & lu 1998a ; bombaci & datta 2000 ; wang et al .
2000 ) , births of magnetars ( usov 1992 ; kluniak & ruderman 1998 ; spruit 1999 ; wheeler et al . 2000 ) , and implosions of supra - massive neutron stars ( vietri & stella 1998 , 1999 ) .
there has been considerable evidence linking the progenitors of grbs with massive stars .
first , the sources of grbs with known redshifts lie within the optical radii and central regions of the host galaxies rather than far outside the disks of the galaxies ( bloom , kulkarni & djorgovski 2001 ) , which seem to rule out mergers of neutron - star binaries as the grb central engine .
second , the brightness distribution of grbs is in agreement with the models in which the grb rate tracks the star formation rate over the past 15 billion years of cosmic history ( totani 1997 ; wijers et al .
1998 ; kommers et al .
third , the supernova sn1998bw with an unusual brightness was discovered in the error box of grb 980425 ( galama et al .
1998 ) and a supernova - like component was detected in the afterglows from grb 980326 ( bloom et al .
1999 ) and grb 970228 ( reichart 1999 ; galama et al .
these detections provide the most direct evidence for the relation between grbs and a specific type of supernova . finally , the recent discovery of a transient absorption edge in the x - ray spectrum of grb 990705 ( amati et al .
2000 ) and the observations of x - ray lines from grb 991216 ( piro et al .
2000 ) and grb 000214 ( antonelli et al .
2000 ) provide new evidence that grbs are related to the core collapse of massive stars .
therefore , it has been widely believed that long duration grbs arise from the explosions of massive stars
. however , it seems theoretically difficult to understand the association of grbs with supernovae .
this is because the rapid variability of grbs and their nonthermal spectra ( woods & loeb 1995 ) requires that the lorentz factor of a grb fireball be @xmath6 .
this conclusion was also recently drawn by lithwick & sari ( 2001 ) , who derived the lower limits on @xmath7 due to annihilation of photons and scattering of photons by pair - created electrons and positrons .
thus , the mass of the baryons contaminating the fireball must be less than @xmath8 .
on the other hand , too many baryons in the stellar envelope exist in the vicinity of the collapsing core in a massive star so that an ultrarelativistic fireball or jet forming during the collapse of the core is easy to become non - relativistic and then a bubble due to sideways expansion of the jet .
a choked fireball ( or jet ) has been found numerically by macfadyen & woosley ( 1999 ) here we propose a scenario for the formation of grbs when a newborn neutron star accretes sufficient mass to undergo a phase transition to a strange star .
cheng & dai ( 1996 ) , dai & lu ( 1998a ) , bombaci & datta ( 2000 ) , and wang et al .
( 2000 ) have proposed the conversion from neutron stars to strange stars as a cosmological origin of grbs , but have not investigated the association of grbs with supernovae , the burst rate , and other interesting implications , e.g. , x - ray grbs .
the purpose of this paper is to discuss these questions .
we assume the neutron star to move outward at an initial kick velocity of @xmath0 in the supernova ejecta . in section 2 , we analyze and calculate hypercritically accreted mass . in section 3 , we discuss implications for grbs . in order to produce a grb ,
the phase transition is required to occur near the supernova front .
we estimate the burst rate to be @xmath2 per year per galaxy .
our model explains the grb - supernova connection . in section 4
, we discuss other implications of the model .
it is known that the core collapse of massive stars with @xmath9 produces type ii supernovae accompanying neutron stars whose initial mass is likely near the chrandrasehkar limit @xmath10 .
the neutron stars are believed to have proper velocities .
gunn & ostriker ( 1970 ) first recognized that galactic pulsars have much larger random velocities than their progenitor massive stars .
modern observations and analysis on the proper motion of pulsars even give @xmath11 as an average 3-dimension velocity of neutron stars at birth ( e.g. , lyne & lorimer 1994 ; lorimer et al .
1997 ; hansen & phinney 1997 ; cordes & chernoff 1998 ) , with possibly a significant population having velocities greater than @xmath12 .
direct evidence for pulsar velocities @xmath13 is provided by observations of the bow shock produced by psr b@xmath14 in the interstellar medium ( cordes , romani & lundgren 1993 ) .
the studies of the associations of neutron stars with supernova remnants have , in many cases , indicate large velocities ( e.g. , frail et al .
in particular , the associations of soft gamma - ray repeaters ( sgrs ) with supernova remnants imply that sgr @xmath15 and sgr @xmath16 have velocities of @xmath17 and @xmath18 respectively where @xmath19 is the supernova remnant age , although the associations seem problematic .
since many isolated pulsars have such large proper velocities , it appears necessary to invoke natal kicks " imparted to newborn neutron stars due to asymmetrical processes during supernovae .
several mechanisms have been suggested for natal kicks : local hydrodynamical instabilities , neutrino - magnetic field driven asymmetry , local high - order gravity mode instabilities , and electromagnetic radiation of an off - centered rotating magnetic dipole ( for a recent review see lai 2000 ) .
owing to these mechanisms , we assume a newborn neutron star to have an initial kick velocity of @xmath20 .
the supernova explosion scenarios involve an outgoing shock wave that initially travels with speed of @xmath21 .
however , when the initial outgoing shock wave enters hydrogen envelope , a deceleration of matter is formed ( woosely 1988 ) .
this deceleration sharpens into a reverse shock . as a result
, the final velocity of the supernova ejecta may be slowed down to @xmath22 . for simplicity , we assume that , since this time , the expanding supernova ejecta is spherical , its mass @xmath23 is constant with time
, its density @xmath24 is uniform in space and decreases with time , and the supernova front radius @xmath25 increases at the fixed velocity of @xmath26 .
therefore , the density throughout is given by @xmath27 where @xmath28 , @xmath29 cm is the initial radius of the ejecta , and @xmath30 is the time after the explosion in units of @xmath31 s. since the ejecta is radiation dominated , its temperature @xmath32 scales as @xmath33 . from this scaling law
, we have @xmath34k ( brown & weingartner 1994 ) , and the sound speed of the ejecta @xmath35 the initial outgoing supernova shock may produce a hole with radius of @xmath36 km ( bethe 1993 ) .
so , no matter is accreted by the neutron star in a time of @xmath37 s where @xmath38 .
subsequently , the neutron star will enters the supernova ejecta .
we want to calculate the accretion rate as follows .
if the effect of the neutron star magnetic field is neglected ( in fact , an initial strong magnetic field can rapidly decay due to hypercritical accretion , cf .
geppert , page & zannias 1999 ) , the accretion rate of the neutron star at radius @xmath39 is given by the bondi - hoyle accretion formula , @xmath40 ^ 2+c_s^2\}^{3/2}},\ ] ] where @xmath41 is the schwarzschild radius of the neutron star with @xmath42 , and @xmath43 is the velocity of the supernova matter ( not front ) at radius @xmath44 ( chevalier 1989 ) .
here we have defined @xmath45 ^ 2+c_s^2\}^{1/2}$ ] .
we first estimate the accreted mass . to do this
, @xmath46 is assumed to be approximately constant .
thus , at early times @xmath47 , the accretion rate is approximated by @xmath48 at late times @xmath49 , the accretion rate becomes @xmath50 this accretion rate is at least ten orders of magnitude larger than the eddington accretion rate for a solar - mass star .
the gravitational energy released during such a hypercritical accretion is carried away by neutrinos .
it is neutrino emission that allows accretion of the star at a much higher rate than the eddington rate . from simple analytical arguments , chevalier ( 1989 ) and
brown & weingartner ( 1994 ) estimated a lower limit to steady neutron star accretion with neutrino losses assuming spherical symmetry : @xmath51 .
our estimated accretion rate exceeds @xmath52 for @xmath53 s. we estimate the accreted mass @xmath54 where equation ( 7 ) is obtained by substituting equations ( 4 ) and ( 5 ) into ( 6 ) .
next , we numerically calculate equation ( 6 ) .
figure 1 presents the accreted mass as a function of time .
it can be seen from this figure that the neutron star will be able to accrete considerable matter ( with mass of @xmath55 ) before its conversion to a strange star .
since conversion of neutron stars to strange stars was suggested as a possible origin of cosmological grbs by cheng & dai ( 1996 ) , resulting rotating strange stars with strong magnetic fields have been proposed to explain the observed features of some grb afterglows by some authors ( dai & lu 1998a , 1998b , 2000 , 2001 ; zhang & mszros 2001 ) .
bodmer ( 1971 ) and witten ( 1984 ) conjectured that strange quark matter may be the true ground state of hadrons .
detailed calculations based on the zero - temperature thermodynamics of strange matter show that strange matter is indeed more stable than @xmath56fe for a wide range of the parameters of the mit bag model ( farhi & jaffe 1984 ) .
if this hypothesis is true , strange stars as a kind of compact object may exist in the universe .
how are strange stars produced ?
one natural way is direct collapse of the core of a massive star to a strange star .
if this way is possible , a binary including a strange star and a more massive compact object imaginably exists . during the coalescence of such a binary ,
the strange star is disrupted by its companion . as a result
, the entire galaxy can be comtaminated and all neutron stars " become strange stars ( madsen 1988 ; caldwell & friedman 1991 ; kluniak 1994 ) , which is referred to as the madsen - caldwell - friedman ( mcf ) effect .
this effect conflicts with the post - glitch behavior of pulsars , which is well described by the neutron - superfluid vortex creep theory , and current strange star models can not explain the observed pulsar glitches ( alpar 1987 ; alpar , pines & cheng 1990 ) .
another way to produce strange stars is that a neutron star in a low - mass x - ray binary accretes sufficient mass to undergo a phase transition to a strange star ( cheng & dai 1996 ) .
this way can avoid the mcf effect . in this paper
, we propose a third way to produce strange stars , i.e. , a newborn unbound " neutron star with a kick velocity of @xmath57 will catch up with the outgoing supernova ejecta and will accrete considerable matter so that the neutron star can convert to a strange star ( see section 2 ) .
if such a rapidly moving neutron star arises from the supernova explosion of a massive star in a binary , the neutron star must be able to escape from the binary system because of its too large kick velocity , and thus the resulting strange star will have no companion . if a newborn neutron star does not have an enough large proper velocity to catch up with the supernova ejecta , the star may accrete only a mass of @xmath58 ( chevalier 1989 ) so that it can not undergo a phase transition to become a strange star .
therefore , the third way can also avoid the mcf effect .
cheng & dai ( 1998 , 2001 ) have argued that such a strange star , if it has a strong magnetic field and a superconducting core , may produce soft gamma - ray repeaters in an age of @xmath59 yrs .
here we discuss its implications for grbs .
the conversion of a neutron star to a strange star requires the formation of a strange matter seed , which is produced through the deconfinement of neutron matter at a density of @xmath60 ( where @xmath61 is the saturation nuclear matter density ) ( baym 1991 ) , much larger than the central density of a @xmath62 neutron star with a moderately stiff to stiff equation of state ( as implied by some astrophysical processes , cf .
dai & lu 1998a ) . to reach the deconfinement density ,
cheng & dai ( 1996 ) suggested , a @xmath62 neutron star with a moderately stiff to stiff equation of state should accrete matter with mass of @xmath63 before its phase transition to a strange star .
once the accreted mass is @xmath64 , a strange matter seed may appear in the core of the neutron star , and subsequently the strange matter will begin to swallow its surrounding neutron matter in a hydrodynamically unstable mode ( detonation ) .
thus , the neutron star will convert to a strange star in a timescale of the order of 0.1 ms .
the phase transition includes two processes : ( 1 ) the neutron matter converts to two - flavor quark matter , and ( 2 ) the two - flavor quark matter converts to strange ( three - flavor ) quark matter .
the latter process can release the energy per baryon of a few tens of mev in a timescale of @xmath65 s ( dai et al .
owing to this process , the resulting strange star will be as hot as a few @xmath66 k. how is the energy release due to the phase transition deposited ?
one mechanism for the energy deposition is the neutrino - antineutrino annihilation process @xmath67 and the neutrino absorption processes @xmath68 and @xmath69 ( in the crust ) .
another mechanism is the creation of electron / positron pairs in an extremely strong strong electric field at the quark surface ( usov 1998 , 2001 ) .
the total energy deposition for these two mechanisms , @xmath70 , is a few @xmath1 ergs , which will inevitably lead to a fireball composed of @xmath5 and electron / positron pairs polluted by a small number of baryons .
the contaminating baryon mass is up to the crustal mass of the strange star , @xmath71 .
therefore , the resulting fireball should be accelerated to an ultra - relativistic phase with lorentz factor of @xmath72 .
the newborn strange star is surrounded by the supernova ejecta matter .
if the sn ejecta mass in the solid angle of @xmath73 between the star and the sn front , @xmath74 , greatly exceeds @xmath75 , the resulting fireball can not still be accelerated to ultra - relativistic .
an ultra - relativistic fireball also requires @xmath76 .
we assume that @xmath77 is the minimum distance of the initial site of the strange star to the supernova front .
assuming @xmath78 , we estimate the ratio of @xmath74 to the total sn ejecta mass @xmath79 : @xmath80 since @xmath81 , we have @xmath82 which further requires @xmath83 because the number of the neutron stars with @xmath84 is @xmath85 per galaxy in the hubble time ( lorimer et al . 1997 ) , the number of the strange stars that can produce grbs are estimated to be @xmath86 per galaxy in the hubble time , where we have assumed that the number distribution of the neutron stars is uniform in the velocity space for @xmath87 .
the burst rate in our model is approximated by @xmath88 this rate is enough to explain the observed grb rate , @xmath89
/yr / galaxy .
the latter rate has been estimated due to the evidence that the grb rate is proportional to the star formation rate ( totani 1997 ; wijers et al .
1998 ; kommers et al .
2000 ) . since the accreted mass @xmath90
, accretion of the pre - conversion neutron star should not influence its velocity significantly , we have an approximate relation : @xmath91 , and the ejecta radius @xmath92 .
thus , the expansion timescale of the sn ejecta is @xmath93 this implies that the sn explosion might occur a few days before a grb .
our model has several implications .
first , it can clearly explain the association of grbs with a special kind of supernova .
the phase transition of the neutron star to a strange star is almost isotropic ; about one half of the energy release will result in an ultra - relativistic fireball with baryon contamination of @xmath94 , and another half will be absorbed by the sn ejecta because the photon - electron scattering depth in the sn ejecta @xmath95r\sim 10 ^ 5 \gg 1 $ ] .
so , the phase transition discussed here may also give rise to a hypernova with a high amount of explosive energy ( @xmath96 ergs ) and a bright optical luminosity .
a hypernova - like component has been observed in several cases , e.g. , sn 1998bw ( iwamoto et al .
1998 ) , grb 980326 ( bloom et al . 1999 ) and grb 970228 ( reichart 1999 ; galama et al
second , the observed breaks in the light curves of some optical afterglows are argued to be due to sideways expansion of collimated fireballs ( rhoads 1999 ; sari , piran & halpern 1999 ) . however ,
this argument is still now a matter of considerable theoretical debate ( moderski , sikora & bulik 2000 ; wei & lu 2000 ) .
dai & lu ( 1999 , 2000 ) have proposed the evolution of a fireball to the non - relativistic regime in a dense medium as an alternative explanation for these observed breaks .
our present model meets the second explanation .
third , if the phase transition occurs at @xmath97 , then the resulting fireball has the loading baryon mass of @xmath98 , and its lorentz factor should be @xmath99 .
although this fireball can not produce a classical grb due to its low lorentz factor , it may result in a weak grb like grb 980425 or an x - ray grb ( or called an x - ray flash ) like grb 991106 observed by bepposax ( search http://www.ias.rm.cnr.it/ias-home/sax/xraygrb.html ) .
we predict that the rate of such a kind of grb is @xmath100 per year per galaxy .
finally , the phase transition occurs more possibly at @xmath101 .
all the energy release will be absorbed by the sn ejecta , and the resulting fireball is non - relativistic but much more energetic than a normal supernova .
this fireball may only behave as a hypernova but not emit a grb or even an x - ray grb . in summary
, we have presented a new model for grbs with small baryon contamination .
a key point of our model is that a newborn neutron star with an initial kick velocity of @xmath0 will catch up with the outgoing supernova ejecta , and accrete matter at a hypercritical rate . once the stellar mass increases to some critical mass
, the neutron star will undergo a phase transition to a strange star , resulting in an energy release of a few @xmath1 ergs .
the phase transition may produce a grb when it occurs just near the sn front .
the burst rate is @xmath2 per year per galaxy .
in addition , if the phase transition occurs in the interior of the sn ejecta , it may result in an x - ray grb observed by bepposax or only a burstless hypernova explosion . in this paper
, we have discussed one result for accreting high - velocity neutron stars in supernovae , i.e. , the conversion to strange stars as an origin of grbs . another result for accreting neutron stars
is the collapse to black holes when the stellar mass reaches the maximum mass .
if the pre - collapse neutron stars are millisecond pulsars , the resulting black holes must be rapidly rotating . as suggested by vietri & stella ( 1998 )
, such black holes may produce grbs by extracting their rotational energy or tapping the binding energy of the disk - black hole system .
99 alpar , m. a. 1987 , phys .
lett . , 58 , 2152 alpar , m. a. , pines , d. , & cheng , k. s. 1990 , nature , 348 , 707 amati , l. et al .
2000 , science , 290 , 953 antonelli , l. a. et al .
2000 , apj , 545 , l39 baym , g. 1991 , in neutron stars : theory and observations ( kluwer , dordrecht ) bethe , h. a. 1993 , apj , 419 , 197 bloom , j. s. et al .
1999 , nature , 401 , 453 bloom , j. s. , kulkarni , s. r. , & djorgovski , s. g. 2001 , aj , submitted ( astro - ph/0010176 ) bodmer , a. r. 1971 , phys
d , 4 , 1601 bombaci , i. , & datta , b. 2000 , apj , 530 , l69 brown , g. e. , & weingartner , j. c. 1994 , apj , 436 , 843 caldwell , r. r. , & friedman , j. r. , 1991 , phys .
b , 264 , 143 cheng , k. s. , & dai , z. g. 1996 , phys . rev .
lett . , 77 , 1210 cheng , k. s. , & dai , z. g. 1998 , phys .
, 80 , 18 cheng , k. s. , & dai , z. g. 2001 , astropart . phys . , in press chevalier , r. a. 1989 , apj , 346 , 847 cordes , j. m. , & chernoff , d. f. 1998 , apj , 505 , 315 cordes , j. m. , romani , r. w. , & lundgren , s. c. 1993 , nature , 362 , 133 costa , e. et al .
1997 , nature , 387 , 783 dai , z. g. , & lu , t. 1998a , phys .
lett . , 81 , 4301 dai , z. g. , & lu , t. 1998b , a&a , 333 , l87 dai , z. g. , & lu , t. 1999 , apj , 519 , l155 dai , z. g. , & lu , t. 2000 , apj , 537 , 803 dai , z. g. , & lu , t. 2001 , a&a , 367 , 501 dai , z. g. , peng , q. h. , & lu , t. 1995 , apj , 440 , 815 eichler , d. , livio , m. , piran , t. , & schramm , d. n. 1989 , nature , 340 , 126 farhi , e. , & jaffe , r. l. 1984 , phys .
d , 30 , 2379 frail , d. a. et al .
1997 , nature , 389 , 261 frail , d. a. , goss , w. m. , & whiteoak , j. 1994 , apj , 437 , 781 galama , t. j. et al .
1998 , nature , 398 , 394 galama , t. j. et al .
2000b , apj , 536 , 185 geppert , u. , page , d. , & zannias , t. 1999 , a&a , 345 , 847 gunn , j. e. , & ostriker , j. p. 1970 , apj , 160 , 979 hansen , b. m. s. , & phinney , e. s. 1997 , mnras , 291 , 569 iwamoto , k. et al .
1998 , nature , 395 , 672 kluniak , w. 1994 , a&a , 286 , l17 kluniak , w. , & ruderman , m. a. 1998 , apj , 505 , l113 kommers , j. m. et al .
2000 , apj , 533 , 696 lai , d. 2000 , astro - ph/0012049 lithwick , y. , & sari , r. 2001 , apj , in press lorimer , d. r. , bailes , m. , & harrison , p. a. 1997 , mnras , 289 , 592 lyne , a. g. , & lorimer , d. r. 1994 , nature , 393 , 35 macfadyen , a. , & woosley , s. e. 1999 , apj , 524 , 262 madsen , j. 1988 , phys .
lett . , 61 , 2909 moderski , r. , sikora , m. , & bulik , t. 2000 , apj , 529 , 151 narayan , r. , paczyski , b. , & piran , t. 1992 , apj , 395 , l83 paczyski , b. 1998 , apj , 494 , l45 phinney , e.s . , & kulkarni , s. r. 1994 , ara&a , 32 , 591 piro , l. et al .
2000 , science , 290 , 955 reichart , d. e. 1999 , apj , 521 , l111 rhoads , j. 1999 , apj , 525 , 737 sari , r. , piran , t. , & halpern , j. p. 1999
, apj , 519 , l17 spruit , h. c. 1999 , a&a , 341 , l1 totani , t. 1997 , apj , 486 , l71 usov , v. v. 1992 , nature , 357 , 472 usov , v. v. 1998 , phys .
, 80 , 230 usov , v. v. 2001 , apj , 550 , l179 van paradijs , j. et al .
1997 , nature , 386 , 686 van paradijs , j. , kouveliotou , c. , & wijers , r. a. m. j. 2000 , ara&a , 38 , 379 vietri , m. , & stella , l. 1998 , apj , 507 , l45 vietri , m. , & stella , l. 1999 , apj , 527 , l43 wang , x. y. , dai , z. g. , lu , t. , wei , d. m. , & huang , y. f. 2000 , a&a , 357 , 543 wei , d. m. , & lu , t. 2000 , apj , 541 , 203 wijers , r. a. m. j. , bloom , j. s. , blagla , j. s. , & natarajan , p. 1998 , mnras , 294 , l13 witten , e. 1984 , phys .
d , 30 , 272 wheeler , j. c. , yi , i. , hflith , p. , & wang , l. 2000 , apj , 537 , 810 woods , e. , & loeb , a. 1995 , apj , 453 , 583 woosley , s. e. 1988 , apj , 330 , 218 woosley , s. e. 1993 , apj , 405 , 273 zhang , b. , & mszros , p. 2001 , apj , in press ( astro - ph/0011133 ) | we present a new model for gamma - ray bursts ( grbs ) that are not only associated with supernovae but also have small baryon contamination . in this model , we assume a newborn neutron star to move outward at a kick velocity of @xmath0 in the supernova ejecta .
we find that such a neutron star still hypercritically accretes its surrounding supernova matter . once the stellar mass increases to some critical mass
, the neutron star will undergo a phase transition to become a strange star , leading to an energy release of a few @xmath1 ergs .
the phase transition , if possibly occuring just near the supernova front , will first result in an ultra - relativistic fireball and then a grb .
this provides a plausible explanation for the grb - supernova association .
we estimate the burst rate to be @xmath2 per year per galaxy .
our model also predicts other possiblities .
for example , if the resulting fireballs have a lorentz factor of the order of a few , they will produce x - ray grbs observed by bepposax .
we find the rate of such bursts to be @xmath3 per year per galaxy . |
the ground state ( `` vacuum '' ) of non - abelian gauge theories like qcd is known to be very rich .
it includes topologically non - trivial fluctuations of the gauge fields , carrying an integer topological charge @xmath1 .
the simplest building blocks of topological structure in the vacuum are @xcite _ instantons _ with @xmath2 and _ anti - instantons _ with @xmath3 .
instantons represent gluon field configurations that are localized ( `` instantaneous '' ) in euclidean time and space .
while they are believed to play an important role in various long - distance aspects of qcd , there are also important short - distance implications . in qcd with @xmath4 ( light ) flavours , instantons induce hard processes violating _
chirality _ in accord @xcite with the selection rule @xmath5 _ chirality _ @xmath6 , due to the general chiral anomaly .
while in ordinary perturbative qcd ( @xmath7 ) these processes are forbidden , their experimental discovery would clearly be of basic significance .
the deep - inelastic scattering regime is strongly favoured in this respect , since hard instanton - induced processes are both calculable @xcite within instanton - pertubation theory and have good prospects for experimental detection at hera @xcite .
qcdins @xcite is a monte carlo package for simulating qcd - instanton induced scattering processes in deep - inelastic scattering ( hera ) .
it is designed as an `` add - on '' hard process generator interfaced by default to the monte carlo generator herwig @xcite .
it incorporates the theoretically predicted production rate for instanton - induced events as well as the essential characteristics that have been derived theoretically for the partonic final state of instanton - induced processes : notably , the flavour democratic @xcite and isotropic @xcite production of the final state partons , energy weight factors different for gluons and quarks @xcite , and a high average multiplicity @xmath8 of produced partons with a ( approximate ) poisson distribution of the gluon multiplicity @xcite .
earlier versions of qcdins have been used already to establish first experimental bounds on the rate of instanton - induced events at hera @xcite and to develop instanton search strategies @xcite . in the present report a comprehensive description is given of the theoretical framework built into the program ( section [ theorie ] ) as well as of the various program components ( section [ qcdins ] ) and of their usage ( section [ usage ] ) .
let us briefly summarize in this section the underlying physics picture , some relevant formulae and the main stages involved in qcdins to generate the complete instanton - induced partonic final state .
the remaining formulae may be found under the corresponding descriptions of qcdins routines in section [ qcdins ] .
ll & l + dis variables : + @xmath9 + @xmath10 + @xmath11 + @xmath12 + @xmath13 + @xmath14 + + variables of instanton - subprocess : + @xmath15 + @xmath16 + @xmath17 + in deep - inelastic @xmath18 scattering , instanton - induced events are predominantly associated @xcite with a process structure as sketched in fig . [ kin - var ] : a photon splitting into a @xmath19-pair fuses with a gluon from the proton in the background of an instanton ( @xmath20 ) or an anti - instanton ( @xmath21 )
. for each ( light ) flavour , @xmath22 , a violation of _ chirality _ is induced , @xmath23=\pm\ , 2 \
\mbox{\ for an } \left\{\begin{array}{l}i\\{{\overline{i}}}\end{array } \right . \,,\ ] ] in agreement with the general chiral anomaly relation @xcite . correspondingly , the partonic final state exhibits `` flavour democracy '' , i.e. @xmath24 ( @xmath25 ) pairs of _ all _ @xmath4 light flavours occur precisely once in case of an instanton ( anti - instanton ) , @xmath26 + n_g\,g \,.\ ] ] as illustrated in fig . [ kin - var ] , one of those partons acts as a current - quark ( jet ) @xmath27 , whereas the other @xmath28 ( anti-)quarks and some number @xmath29 of gluons are directly emitted from the instanton ( anti - instanton ) `` blob '' .
instanton - induced processes initiated by a quark from the proton are suppressed by a factor of @xmath30 with respect to the gluon initiated process @xcite .
this fact , together with the high gluon density in the relevant kinematical domain at hera , justifies to neglect quark initiated processes . in instanton - perturbation theory , the dominant instanton - induced contribution to the inclusive @xmath31 cross section at cross sections .
] , subject to appropriate kinematical restrictions and ( theoretical ) fiducial cuts , has a convolution - like structure @xcite , @xmath32 \label{evwgt } & & \times \int\limits_{{\rm max}\left(\frac{q^{\prime 2}}{sx^\prime y_{\rm bj\ , max}},\frac{x_{\rm bj\,min}}{x^\prime } \right)}^{z_{\rm max } } \frac{dz}{z}\,f_g ( z ) \int\limits_{x_{\rm bj\,min}}^{x^\prime z - \frac{m_{k}^2}{s } \frac{1}{y_{\rm bj\,max}-\frac{q^{\prime 2}}{sx^\prime z } } } \frac{dx_{\rm bj}}{x_{\rm bj}}\ , \\[2ex ] \nonumber & & \times \int\limits_{{\rm max}\left ( \frac{q^{\prime 2}}{sx^\prime z}+\frac{m_{k}^2}{s } \frac{1}{x^\prime z -x_{\rm bj}},y_{\rm bj\,min } \right)}^{y_{\rm bj\,max } } \frac{dy_{\rm bj}}{y_{\rm bj}}\ , \theta ( s x_{\rm bj } y_{\rm bj } - q^2_{\rm min})\ , \\[2ex ] \nonumber & & \times \frac{1+(1-y_{\rm bj})^2}{y_{\rm bj}}\ p_{q^\prime}^ { { ( i)}}\ .\end{aligned}\ ] ] it involves integrations over the gluon density @xmath33 , the virtual photon flux and the flux of virtual ( anti-)quarks @xmath34 in the instanton - background @xcite , @xmath35 all relevant kinematical variables in eq .
( [ evwgt ] ) are defined in fig .
[ kin - var ] , and @xmath36 denotes the electric charge squared of the virtual ( anti-)quark @xmath34 in units of the electric charge squared @xmath37 .
[ ht ] in eq .
( [ evwgt ] ) , @xmath38 denotes the instanton - induced total cross section of the @xmath39-subprocess ( c.f .
[ kin - var ] ) and contains the essential instanton dynamics .
its analytical form @xcite used in qcdins may be found in eq .
( [ qgcross - vxi ] ) . as illustrated in fig .
[ isorho ] , @xmath38 is very steeply growing for decreasing values of @xmath40 and @xmath41 , respectively .
eventually , our theoretical predictions based on instanton - perturbation theory @xcite will cease to hold .
therefore , the following cuts inferred from a high - quality lattice simulation of qcd @xcite are implemented by default in qcdins 2.0 ( table [ floating ] ) , @xmath42 a further cut on the photon virtuality , @xmath43 is applied in order to warrant sufficient suppression of non - planar contributions @xcite , which are hard to calculate and may spoil the validity of eq .
( [ evwgt ] ) . the cross section @xmath38 exhibits a rather weak residual dependence on the renormalization scale @xmath44 . as an `` optimal '' choice , the value @xmath45 , corresponding to the minimum @xcite , @xmath46 , is taken by default ( table [ floating ] ) .
however , @xmath38 depends strongly on the qcd scale @xmath47 .
since , strictly speaking , the underlying theoretical framework refers to massless quarks , the ( default ) number of flavours is set to @xmath48 ( table [ floating ] ) .
the respective value @xmath49 is obtained by a standard 3-loop perturbative flavour reduction ( eq . ( 9.7 ) of ref .
@xcite ) from the 1998 world - average of the running qcd coupling at the z - boson mass @xcite , @xmath50 the central values of these parameters are taken as default in qcdins 2.0 ( table [ floating ] ) .
this upgrade of @xmath47 , together with the modified cuts ( [ fiducial ] ) and ( [ fiducialq ] ) , represents an improved understanding of the input parameters and a considerable reduction of uncertainties , as compared to the original publication @xcite and earlier versions of qcdins . note that it also implies a significant change in the predicted rate .
next , let us summarize the various stages of event generation by means of qcdins . in a first stage , the various bjorken variables @xmath51 of the instanton - induced process ( c.f .
[ kin - var ] ) are generated , with a distribution according to the normalized differential cross section from eq .
( [ evwgt ] ) , @xmath52 in the second stage of the event generation , the 4-momenta @xmath53 of the incoming gluon @xmath54 , the virtual photon @xmath55 , the virtual quark @xmath34 and the current quark @xmath27 , respectively , are filled .
sudakov decompositions of these momenta are used to incorporate various constraints , e.g. on the momenta squared .
the 4-momentum @xmath56 of the outgoing lepton is calculated subsequently . in the third stage ,
the partonic final state of the instanton - induced @xmath57-subprocess is generated in its centre - of - mass system ( cms ) as follows .
the number @xmath29 of produced gluons is generated according to a poisson distribution with mean @xmath58 ( for the cuts ( [ fiducial ] ) ) , as calculated theoretically ( eq . ( [ ng ] ) ) in instanton - perturbation theory @xcite .
next , @xmath59 @xmath60$]-``strings '' of partons are set up , each beginning with a quark , followed by a random number @xmath61 of gluons and ending with an anti - quark of randomly chosen flavour
. there are @xmath62 gluons in total and , due to the required flavour democracy ( [ flavour - democr ] ) , @xmath19-pairs of all @xmath4 flavours occur precisely once .
a quark and a gluon among these @xmath63 partons are ( randomly ) marked as incoming .
the momenta @xmath64 of the @xmath65 outgoing partons are then generated in the cms of the instanton subprocess , according to the energy - weighted phase - space @xmath66 these different energy weights for quarks and gluons @xcite , along with the angular isotropy @xcite , are characteristic features of the leading - order partonic final state ( after averaging over colour ) . next , the colour and flavour connections of the partons are set up .
the colour flow is obtained simply by connecting the colour lines of adjacent partons within each of the above - mentioned @xmath4 @xmath60$]-``strings '' in a planar manner ( consistent with the leading order @xmath67 expectation ) .
this choice is inspired by the leading - order partonic final state ( after averaging over colour ) @xcite , but may well deserve further research .
the flavour flow is constructed by connecting the flavour lines of the quark at the beginning of a string with the flavour line of the anti - quark at the end of a string .
the hard subprocess generation ends by boosting the momenta of the final state partons to the laboratory frame .
while the subsequent perturbative evolution of the generated partons is always handled by the herwig @xcite package , the final hadronization step may optionally be performed also by means of jetset @xcite .
this section is devoted to a systematic description of the various routines of the qcdins package that is designed as an `` add - on '' hard process generator , interfaced to the monte carlo generator herwig @xcite .
this reference section is organized as follows : while all subroutines and functions of the qcdins package are described in _ alphabetical order _ in section [ routines ] , a _
logical flow - chart _ is provided in form of tables [ flow - chart ] and [ flow - qihgen ] below .
they should always be used as the main guide through the description of the package . a routine listed in the n - th column and the m - th row of these tables
progressively calls all routines in the ( n+1)-th column starting in the ( m+1)-th row .
all routines called at the level of the main hard process generator * qihgen * and below are documented in table [ flow - qihgen ] .
a specific application requires the writing of a steering program by the user ( c.f . section [ usage ] and appendix a ) .
it must contain the standard herwig - initialization calls as well as the calls to various initialization routines for qcdins .
the latter comprise essentially the four routines listed in the second column of table [ flow - chart ] . by calling the last one of these ( * qcloop * ) , the user starts the proper simulation which comprises the chain of internally called qcdins routines as documented in tables [ flow - chart ] and [ flow - qihgen ] .
a specific and quite extensive example is provided with the qcdins distribution ( qtesthz.f ) and may be found in the directory qcdtest. it also illustrates the use of the event analysis routine * hwanal * called by herwig after each processed event
. meters ( tables [ floating ] , [ flags ] ) & [ -1.5ex]*qiinit * & & & & + flavour reduction : & & & & & + @xmath68 & & [ -1.5ex]*lamnf * & & & + initialize particle " & & & & & + inst in event record & & [ -1.5ex]*qiinih * & & & + print input parame- & & & & & + ters and warnings & [ -1.5ex]*qistat * & & & & + initialize jetset & & & & & + common block data & [ -1.5ex]*gjeini * & & & & + loop over desired & & & & & + number of events & [ -1.5ex]*qcloop * & & & & + generate one event & & * qcdgen * & & & + assign hard process & & & & & + variables ( herwig ) & & & [ -1.5ex]_hwepro _ & & + no action ( modified & & & & & + herwig routine ) & & & & [ -1.5ex]*hwegam * & + call main instanton & & & & & + process generator & & & & [ -1.5ex]*hvhbvi * & + main ( hard ) instanton & & & & & + process generator & & & & & ' '' '' [ -1.5ex ] + generate parton & & & & & + cascades ( herwig ) & & & [ -1.5ex]_hwbgen _ & & + combine jets with & & & & & + correct kinematics & & & & [ -1.5ex]*hwbjco * & + convert herwig to & & & & & + jetset block data & & & [ -1.5ex]*herlund * & & + jetset event record & & & & & + to hepevt common & & & [ -1.5ex]*luhepc * & & + radiation from lepton & [ -1.5ex]*exfrac * & & + check kinematical boundaries & * qicalc & & + generate identity code of current & & & + quark @xmath27 and virtual quark @xmath69 & [ -1.5ex]*qihpar * & & + generate @xmath70 and associated weight & * qihins & & + generate @xmath71 as @xmath72 & & * qirdis & + calculate total cross section of & & & + instanton - induced subprocess @xmath73 & & [ -1.5ex]*q2sig * & + calculate @xmath74-valley action @xmath75 & & & * action + calculate fermionic overlap @xmath76 & & & * omega + calculate lambert w - function & & & * lambertw + calculate saddle - point value of & & & + conformally invariant @xmath74-distance @xmath77 & & & [ -1.5ex]*xi * + * * * * * * ' '' '' calculate inverse running coupling @xmath78 & & & ' '' '' * xqs * + generate number @xmath29 of emitted gluons & * qigmul & & + calculate gluon multiplicity @xmath79 & & * gmult & + * * ' '' '' calculate inverse running coupling @xmath78 & & & ' '' '' * xqs * + calculate @xmath74-valley action @xmath75 & & & * action + calculate lambert w - function & & & * lambertw + calculate flux of virtual quark @xmath69 & * qisplt & & + calculate remaining weight & * qipvwt & & + calculate momentum of incoming gluon & * qikpar & & + generate momentum of virtual photon & * qikgam & & + generate momenta of virtual quark @xmath69 & & & + and current quark @xmath27 & [ -1.5ex]*qikgsp * & & + generate partonic final state & * qistid & & + generate the partons of the instanton & & & + subprocess in form of @xmath80$]-strings & & [ -1.5ex]*qiglst * & + find the incoming partons & & & + in the @xmath80$]-strings & & [ -1.5ex]*qigpar * & + assign masses of outgoing partons & & * qiplst & + generate 4-momenta of outgoing partons & & * qipsgn & + calculate relative energy weight & & & + of outgoing partons & & & [ -1.5ex]*qipswt * + store 4-momenta of outgoing partons & & & + into phep common block of herwig & & [ -1.5ex]*qipsto * & + colour / flavour connections for each string & & * qiccon & + * * * * * * * * * * subroutine * action*(xi4,s , ds , dds ) abcdefghikl calculation of the @xmath74-action as well as its 1st and 2nd derivatives , as function of the conformally invariant @xmath74-distance
. + abcdef conformally invariant @xmath74-distance @xmath77 .
@xmath74-action @xmath75 , eq .
( [ action ] ) .
@xmath81 , @xmath82 the action is calculated according to the exact valley form @xcite , @xmath83 \ln\left [ \frac{1}{2\xi } \bigl ( f(\xi ) + 4\bigr)\right ] \ , , \\[2.4ex ] f(\xi ) & = & \xi^2+\sqrt{\xi^2 - 4}\xi-4 \ , .
\end{aligned}\ ] ] subroutine * exfrac*(a ) abcdefghikl optional account of initial state radiation from the lepton .
+ ab abcdef dummy rescaling factor of the incoming lepton momentum ; double precision output variable .
the actual routine has to be provided by the user .
subroutine * gjeini * abcdefghikl initialization of the jetset @xcite parameter common blocks ludat1 , ludat2 , ludat3 , ludat4 and ludatr .
+ * gjeini * has to be called by the user before any other jetset routine .
* gjeini * is from ref .
@xcite . function * gmult*(xpr , xi_min , xi_max , qlam , kappa , nf , loopfl ) abcdefghikl calculation of the average gluon multiplicity @xmath84 depending on @xmath41 , @xmath85 , @xmath86 , @xmath4 and loop - order . here ,
@xmath44 and @xmath4 denote the renormalization scale and the number of light flavours , respectively .
+ abcdefghi @xmath41 @xmath87 ; lower boundary of @xmath77 used for interpolation .
@xmath88 ; upper boundary of @xmath77 used for interpolation .
@xmath85 @xmath86 @xmath4 ; number of light flavours .
abcd 1-loop renormalization group ( rg ) invariance @xcite along with 1-loop form of @xmath89 .
2-loop rg invariance @xcite along with 2-loop form of @xmath90 .
( default ) 2-loop rg invariance along with 3-loop form of @xmath90 . from an analysis based on the generalized ( mueller @xcite ) optical theorem for the @xmath91 forward scattering amplitude and the @xmath74-valley method , one infers @xcite the differential one - gluon inclusive @xmath92 cross section , normalized by the total cross section @xmath38 .
the mean gluon multiplicity @xcite is then found by phase space integration , @xmath93 the function * gmult * calculates and returns the average gluon multiplicity ( [ ng ] ) .
the stars @xmath94 in eq .
( [ ng ] ) denote the saddle point values of the @xmath74 collective coordinates @xmath95 , @xmath96 and @xmath77 .
their computation proceeds as in the descriptions of the functions * q2sig * and * xi*. the required values of @xmath97 are calculated and returned by the function * xqs*. the @xmath74-action @xmath75 and its @xmath77-derivatives are provided by the subroutine * action*. subroutine * herlund * abcdefghikl conversion of the herwig @xcite event record in the hepevt common block to the respective jetset @xcite common block . + *
herlund * is a modified @xcite copy from the jetset subroutine luhepc .
subroutine * hvhbvi * abcdefghikl call of the main ( hard ) instanton process generator * qihgen*. + replaces a dummy stub in herwig @xcite that was originally used as event generation interface for the monte carlo generator herbvi @xcite for baryon number violating interactions .
used in the qcdins package to select qcd - instanton induced processes via mod(iproc/100,100 ) @xmath98 75 .
the process code iproc (= 17600 ) has to be set in the user s steering program ( c.f .
appendix a ) .
furthermore , the qcdins program header is printed .
subroutine * hwbjco * abcdefghikl modification of herwig @xcite routine to account for instanton - induced scattering .
+ the modifications are @xcite : the logical flag ( dispro ) for keeping the lepton momenta fixed in herwig 5.9 is modified to include also instanton - induced dis , ipro = 76 .
furthermore , a bug concerning energy - momentum conservation in the original routine from herwig 5.9 has been fixed .
subroutine * hwegam*(ihep , zmi , zma , wwa ) abcdefghikl modification to avoid standard generation of the ( virtual ) photon at this stage
. + c.f . ref .
@xcite this is a modified routine from herwig 5.9 ( c.f . ref .
usually , * hwegam * generates an incoming photon from the incoming @xmath99 . within qcdins , however , the photon is generated at a later stage in the subroutine * qikgam*. thus , the herwig routine * hwegam * has been modified to immediately return for instanton - induced processes ( ipro=76 ) .
function * lambertw*(x ) abcdefghikl calculation of the principal branch of the lambert w - function @xmath100 for @xmath101 .
+ x : @xmath102 ; argument of the lambert w - function .
@xmath100 is the ( real ) solution of @xmath103 , analytic at @xmath104 .
the following simple , but accurate approximation is used and returned by * lambertw * : @xmath105 \ln(x-4)-(1-\frac{1}{\ln(x)})\cdot\ln(\ln(x ) ) ; & { \rm\ for\ } x > 500\ , .
\end{array } \right.\ ] ] function * lamnf*(nf , lambda5 ) abcdefghikl calculation of @xmath106 from @xmath107 to order @xmath108
. + abcdefghijklm number of ( light ) flavours , @xmath4 .
input value @xmath107 .
the flavour reduction of @xmath107 to the desired number of light flavours is performed by using eq .
( 9.7 ) of ref .
@xcite .
subroutine * luhepc*(mconv ) abcdefghikl conversion of the jetset @xcite event record contents back to the hepevt common block .
+ mconv = 1 the present routine is a modified @xcite version of the jetset routine * luhepc*. function * omega*(xi4 ) abcdefghikl calculation of the fermionic overlap , as function of the conformally invariant @xmath74-distance .
+ xi4 : conformally invariant @xmath74-distance @xmath77 .
the following simple , but accurate approximation for the fermionic overlap @xcite @xmath109 is used and returned by * omega * : @xmath110 function * q2sig*(xprime , qlam , kappa , loopfl , nf ) abcdefghikl calculation of the total cross section @xmath111\ ] ] for the instanton - induced subprocess , depending on @xmath41 , @xmath85 , @xmath86 , loop - order and @xmath4 . here ,
@xmath44 and @xmath4 denote the renormalization scale and the number of light flavours , respectively .
+ abcdefghi @xmath41 @xmath85 @xmath86 abcd 1-loop renormalization group ( rg ) invariance @xcite along with 1-loop form of @xmath89 .
2-loop rg invariance @xcite along with 2-loop form of @xmath90 .
( default ) 2-loop rg invariance along with 3-loop form of @xmath90 .
@xmath4 ; number of light flavours .
the function * q2sig * calculates and returns the cross section ( [ cs ] ) as derived in ref .
@xcite , @xmath112 & & \times \frac{\omega ( \xi_\ast ) ^{2n_f-1}(\xi_\ast -2)^3 v^{\ast\,5 } } { \sqrt{\frac{1}{2}({\tilde{s}}-v_\ast -2 { d(\tilde{s})})^2 + { \tilde{s}}({\tilde{s}}-v_\ast ) { d\left ( \ln \left ( \frac{{d(\tilde{s})}}{\sqrt{\xi_\ast -2 } } \right)\right ) } } } \\[1.6ex ] & & \times \left ( \frac{4\pi } { \alpha_{\overline{\rm ms}}\left(\mu _ r \right)}\right)^{19/2 } \exp \left [ -\frac{4\pi } { \alpha_{\overline{\rm ms}}\left(\mu _ r \right ) } s^{({i\overline{i}})}\left(\xi_\ast \right ) - 2 \left(1-\ln\left(\frac{v_\ast\mu_r}{q^\prime}\right)\right ) \,{\tilde{s}}\right ] \ , . \nonumber\end{aligned}\ ] ] it is expressed entirely in terms of the saddle point values for the @xmath74 collective coordinates , @xmath113 ( conformally invariant distance ) and @xmath114 ( scaled size ) . for given @xmath115 , @xmath85 and ( scaled ) renormalization scale @xmath86 , these are in turn unique solutions of the saddle point equations @xcite @xmath116 \label{eq1 } v_\ast & = & 2\ , d({\tilde{s}}(\xi_\ast))\ , w\left ( \frac{q^\prime}{\mu_r } \frac{\exp\left\{\frac{1}{2 } \left[\frac{4\pi } { \alpha_{\overline{\rm ms}}\left(\mu _ r \right)}\frac{1}{\delta_1\beta_0 } + \frac{{\tilde{s}}(\xi_\ast)}{d({\tilde{s}}(\xi_\ast ) ) } \right ] \right\ } } { 2\,d({\tilde{s}}(\xi_\ast ) ) } \right),\end{aligned}\ ] ] with @xmath117 and @xmath118 the @xmath74-action @xmath75 as well its @xmath77-derivatives , entering the cross section ( [ qgcross - vxi ] ) and eqs .
( [ eq2 ] ) , ( [ eq1 ] ) through @xmath119 are calculated in the subroutine * action*. the fermionic overlap @xmath120 is calculated and returned by the function * omega*. in eq .
( [ eq1 ] ) , @xmath121 denotes the principal branch of the lambert @xmath121-function , i.e. the ( real ) solution of @xmath103 , analytic at @xmath104 .
the latter is calculated and returned by the function * lambertw*. the first step in the solution of the saddle - point equations ( [ eq2 ] ) , ( [ eq1 ] ) consists in eliminating @xmath122 in eq .
( [ eq2 ] ) by inserting eq .
( [ eq1 ] ) .
next , for given @xmath115 , @xmath123 and @xmath86 , the resulting implicit equation is solved numerically for @xmath113 .
this is done by the function * xi * which provides @xmath124 on return .
the latter is then inserted into eq .
( [ eq1 ] ) providing @xmath125 .
the values of @xmath97 are calculated and returned by the function * xqs*. subroutine * qcdgen * abcdefghikl interface for the generation of one instanton - induced event , including calls to event initialization , hadronization and event termination routines . + * an instanton - induced event is initialized by a call to the herwig @xcite subroutine * hwuine*. * the partonic instanton subprocess is generated by a call to the herwig subroutine * hwepro*. * if the hadronization flag qicont(21 ) is set .true .
( default ) , the event is fully hadronized
. else , the event is finalized ( * hwufne * ) immediately after the call of * hwepro*. depending on the control flag for the hadronization model , qicont(18)=.true./.false .
, the hadronization is either performed by appropriate herwig @xcite or ( modified @xcite ) jetset @xcite routines , respectively ( see table [ hadroutines ] ) . * furthermore , in this routine energy - momentum conservation in the generated event is checked . .[hadroutines ] depending on the flag qicont(18 ) = .true./.false .
, herwig @xcite and jetset @xcite hadronization may be selected .
the respective calls to herwig and ( modified ) jetset routines used in * qcdgen * are displayed . [ cols="^,<,<",options="header " , ] subroutine * qiusps(wt ) * abcdefghikl dummy routine for optional modification of relative phase space weight as calculated in * qipswt*. + wt : modified relative phase space weight .
called from * qipswt * if qicont(2 ) = .false .. calculation of wt has to be provided by the user .
function * xi * ( xpr , xi_min , xi_max , xq , xmu , nf , loopfl ) abcdefghikl the saddle point value of the conformally invariant @xmath74-separation @xmath77 is calculated as function of @xmath115 , @xmath126 , @xmath127 and @xmath4 .
+ + abcdefghi @xmath115 @xmath87 ; lower boundary of @xmath77 used for interpolation .
@xmath88 ; upper boundary of @xmath77 used for interpolation .
@xmath126 @xmath127 @xmath4 ; number of light flavours .
abcd 1-loop renormalization group ( rg ) invariance @xcite along with 1-loop form of @xmath89 .
2-loop rg invariance @xcite along with 2-loop form of @xmath90 .
( default ) 2-loop rg invariance along with 3-loop form of @xmath90 .
first , the saddle point equation ( [ eq2 ] ) is solved analytically for @xmath41 , and the @xmath128-values corresponding to a discrete set of @xmath129 values are calculated as @xmath130 with @xmath131 and @xmath132 from eqs .
( [ eq1])([dadada ] ) inserted : @xmath133 \nonumber & & \hspace{-5ex } w\left ( \frac { \exp\left\ { \frac{1}{2}\frac{{\tilde{s}}}{{d(\tilde{s } ) } } + \frac{1}{2 } \frac{\delta_1 - 1}{\delta_1\beta_0 } \left [ \delta_1 \ln \left ( \frac{\delta_1 x(\mu_r ) } { x(q^\prime ) + ( \delta_1 - 1)x(\mu_r ) } \right ) -1 \right ] x(\mu_r ) + \frac{1}{2}\frac{x(q^\prime ) } { \beta_0 } \right\ } } { 2\,{d(\tilde{s } ) } } \right ) .\end{aligned}\ ] ] in eq .
( [ vastxq ] ) , the explicit dependence of @xmath132 on @xmath134 ( c.f .
( [ eq1 ] ) has been eliminated in favour of an @xmath135 , @xmath136 dependence by means of the standard scale transformation of @xmath90 . here
, we have introduced the shorthand @xmath137 the desired continuous inversion @xmath138 is then achieved by means of numerical interpolation based on the above exact supporting points ( @xmath139 ) .
function * xqs*(qlam , loopfl , nf ) abcdefghikl calculation of @xmath140 as function of @xmath141 , where @xmath142 is a generic mass scale .
+ abcdefghik @xmath141 abcd 1-loop form of @xmath89 .
2-loop form of @xmath90 .
( default ) 3-loop form of @xmath90 .
@xmath4 ; number of light flavours .
the running coupling @xmath90 is calculated according to the explicit formula eq .
( 9.5a ) in ref .
@xcite , which is accurate to 3-loop .
the quantity @xmath97 is returned .
the integer variable loopfl = @xmath143 controls the loop - order at which eq .
( 9.5a ) in ref .
@xcite is truncated .
qcdins 2.0 should be loaded together with herwig @xcite version 5.9 and jetset @xcite version 7.4 , that are part of the cernlib distribution .
the program is a slave system , which the user must call from his own steering program .
a very simple example is provided in appendix a. by default , qcdins is compiled in form of a library , libqcdins.a , that may be linked together with libherwig59.a , libjetset74.a and the library libpdflib.a of parton distribution functions to the steering program .
an extensive demonstration program , including an interface to the hztool @xcite package , and a detailed installation instruction are included in the distribution .
information about the distribution , its update history , an interactive manual , the source code , pictures of typical events etc . can be accessed via the qcdins www site , http://www.desy.de/~t00fri/qcdins/qcdins.html .
first of all , we are grateful to m. gibbs who left physics in may 1995 . without his early contributions
the qcdins package would presumably not exist .
a number of people helped improving the code .
in particular , let us mention t. carli , g. grindhammer , h. jung and m. seymour . specifically , we thank h. jung for his efforts to incorporate an interface to the jetset package .
moreover , we are grateful to m. seymour for an evaluation of the qcdins package during the desy workshop 1998/99 on monte carlo generators for hera physics . finally , we thank g. ingelman for a careful reading of the manuscript .
the following code may serve as the simplest illustration of a steering program for the qcdins / herwig package .
note that the two subroutines * hwanal * and * hwaend * must exist .
for an example of an interface with the jetset hadronization , we refer to the extensive steering program qtesthz included in the qcdins distribution . ....
program qcdins # include " herwig.inc " c force inclusion of block data external hwudat c initialize process number iproc = 17600 c maximal number of events in this run maxev = 1000 c beam particles pbeam1 = 27.5d0 pbeam2 = 820.0d0 part1 = ' e+ ' part2 = ' p ' c initialize common blocks call hwigin c user can reset parameters at this point , otherwise values c set in hwigin will be used .
c no vertex information in event printout prvtx=.false .
c reset number of shots for initial max weight search ibsh = 5000 lrsud=0 lwsud=77 c seeds nrn(1)=106645412 nrn(2)=135135176 c use laboratory frame usecmf = .false .
c compute parameter - dependent constants call hwuinc c number of herwig events to print out maxpr = 1 c call hwusta to make any particle stable call hwusta('pi0 ' ) call hwusta('k_s0 ' ) c initialize default qcdins input parameters call qiinit c print input parameters call qistat c loop over events call qcloop c user 's terminal calculations call hwaend stop end subroutine hwanal return end subroutine hwaend return end ....
below , we display the essential output from a test run of the very simple steering program from appendix a. this test run simulates 1000 complete instanton - induced events in deep - inelastic @xmath144 scattering ( hera ) , with @xmath145 gev and @xmath146 gev , in the laboratory frame .
all the qcdins parameters correspond to the default values as set in the initialization routine * qiinit * ( c.f .
tables [ floating ] and [ flags ] ) . ....
= = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = qcd instanton monte carlo information version 2.0 = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = parameter value = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = default gluon mass ( gev ) 0.7500e+00 minimum allowed value of x_bj 0.1000e-02 maximum allowed value of z ( proton mom.frac . ) 0.1000e+01 maximum allowed value of y_bj 0.1000e+01 minimum allowed value of y_bj 0.1000e+00 maximum allowed value of x prime 0.9000e+00 minimum allowed value of x prime 0.3500e+00 maximum allowed value of q prime * * 2 ( gev**2 ) 0.1652e+04 minimum allowed value of q prime *
* 2 ( gev**2 ) 0.1134e+03 lower cut for me calculation on x prime 0.3500e+00 lowest allowed weight efficiency cut -.1000e-09 minimum instanton invariant mass 0.0000e+00 lambda - ms - bar(nf ) [ gev ] from pdg 1998 0.3459e+00 minimum total k.e . of outgoing partons ( gev ) 0.0000e+00 renormalization point kappa = mu_r / qprime : 0.1500e+00 minimum allowed value of q * * 2 ( gev**2 ) 0.1134e+03 factorization scale ( gev ) 0.1065e+02 = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = i d code assumed for instanton 206 maximum number of mambo iterations per ps wt . 100 maximum number of ps wt . rejections per event 300 number nf of ( light ) flavours 3 maximum average gluon multiplicity in distbn
. 10 maximum number of iterations for me generation 20 maximum number of qigmul iterations 40 number of loops in rg - invariance / alpha_s 3 = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = control flag option setting = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = reweight phase space configurations true use default phase space reweighting true use matrix element weight true disregard instanton minimum mass requirement true generate q prime before x prime true enforce mass of current quark in kinematics true enforce limit on maximum number of gluons true ensure mass less than subprocess energy true kill events with insufficient instanton mass true use z generation as dz / z true use x prime * * -n generation for efficiency true use q prime * * -n generation for efficiency true use herwig rather than jetset hadronization true use random azimuth angle for scattered electron true use full hadronization true = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = process code iproc = 17600 random no .
seed 1 = 1246579 seed 2 = 8447766 number of shots = 5000 * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * qcdins version 2.0 * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * .... .... new maximum weight = 2.4520378923431534e-05 new maximum weight = 6.0858601240747906e-02 new maximum weight = 0.1289870599803515 new maximum weight = 0.3662942637479250 new maximum weight = 0.5705336816482935 new maximum weight = 0.8061467165013312 number of events = 0 number of weights = 5000 mean value of wgt = 2.8721e-02 rms spread in wgt = 6.4377e-02 actual max weight = 7.3286e-01 assumed max weight = 8.0615e-01 ihep i d idpdg ist mo1 mo2 da1 da2 p - x p - y p - z energy mass 1 e+ -11 101 0 0 4 0 0.00 0.00 27.50 27.50 0.00 2 p 2212 102 0 0 0 0 0.00 0.00 -820.00 820.00 0.94 3 cmf 0 103 1 2 0 0 0.00 0.00 -792.50 847.50 300.33 ihep i d idpdg ist mo1 mo2 da1 da2 p - x p - y p - z energy mass 5 e+ -11 121 7 9 21 9 0.00 0.00 27.50 27.50 0.00 6 gluon 21 122 7 20 22 19 0.00 0.00 -50.37 50.38 0.75 7 hard 0 120 5 6 9 20 0.53 0.99
-22.98 77.98 74.51 8 inst 0 3 7 0 0 0 0.23 9.33 -43.54 47.01 15.07 9 e+ -11 123 7 5 26 5 8.20 -7.47 3.48 11.63 0.00 10 ubar -2 124 7 14 27 17 -8.43 -1.86 17.19 19.24 0.31 11 dqrk 1 124 7 12 31 13 -0.07 0.52 -7.65 7.67 0.32 12 gluon 21 124 7 13 33 11 0.16 -0.28 -0.64 1.04 0.75 13 dbar -1 124 7 11 35 12 -0.67 -0.12 -2.65 2.75 0.32 14 uqrk 2 124 7 15 37 10 0.01 4.97 -11.73 12.75 0.32 15 gluon 21 124 7 16 39 14 0.53 0.62 -2.17 2.44 0.75 16 gluon 21 124 7 17 41 15 0.43 -0.43 -8.44 8.49 0.75 17 gluon 21 124 7 10 43 16 -0.72 0.09 -2.40 2.62 0.75 18 sqrk 3 124 7 19 45 20 -0.11 2.50 -4.93 5.55 0.50 19 gluon 21 124 7 6 47 18 0.81 0.93 -1.46 2.06 0.75 20 sbar -3 124 7 18 51 6 -0.14 0.52 -1.47 1.64 0.50 21 z0/gama * 23 3 5 7 0 0 -8.20 7.47 24.02 15.87 -21.17 .... .... check of energy - momentum conservation in the event : = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = sum_i(p_i)[gev ] = ( 4.87e-13 -5.49e-13 1.59e-12
-4.20e-12 ) with 60 stable particles in final state contributing .
number of events = 1000 number of weights = 27258 mean value of wgt = 2.8837e-02 rms spread in wgt = 6.3371e-02 actual max weight = 7.0199e-01 assumed max weight = 8.0615e-01 9 a. belavin , a. polyakov , a. schwarz and yu .
tyupkin , .
t hooft , ; ; ( erratum ) ; .
s. moch , a. ringwald and f. schrempp , .
a. ringwald and f. schrempp , .
a. ringwald and f. schrempp , hep - ph/9411217 , in : quarks 94 , eds .
d. grigoriev , v. matveev , v. rubakov , d. son and a. tavkhelidze ( world scientific , singapore , 1995 ) p. 170 . m. gibbs , a. ringwald and f. schrempp , hep - ph/9506392 , in : workshop on deep inelastic scattering and qcd ( dis 95 ) , eds .
f . laporte and y. sirois ( ecole polytechnique , paris , 1995 ) p. 341 .
t. carli , j. gerigk , a. ringwald and f. schrempp , hep - ph/9906441 , in : monte carlo generators for hera physics , eds . a.t .
doyle , g. grindhammer , g. ingelman and h. jung , desy - proc-1999 - 02 ( desy , hamburg , 1999 ) p. 329 .
g. marchesini et al .
, . s. moch , a. ringwald and f. schrempp , hep - ph/9706400 , in : deep inelastic scattering and qcd ( dis 97 ) , eds . j. repond and d. krakauer ( american institute of physics , woodbury , new york , 1997 ) p. 1007 .
a. ringwald and f. schrempp , hep - ph/9706399 , in : deep inelastic scattering and qcd ( dis 97 ) , eds .
j. repond and d. krakauer ( american institute of physics , woodbury , new york , 1997 ) p. 781 .
s. aid et al . , h1 collaboration , . s. aid et al . , h1 collaboration ,
. t. carli and m. kuhlen , .
s. moch , a. ringwald and f. schrempp : in preparation a. ringwald and f. schrempp , .
c. caso et al . , particle data group , .
h. jung , private communication .
t. sjstrand , .
v.v . khoze and a. ringwald , .
j. verbaarschot , .
g. grindhammer , private communication .
a. mueller , .
m. gibbs and b. webber , .
t. carli , private communication .
j. owens , .
r. kleiss and w. stirling , .
m. seymour , private communication .
j. bromley et al . , in : future physics at hera , eds .
g. ingelman , a. de roeck and r. klanner ( desy , hamburg , 1996 ) vol . 1 , | we describe a monte carlo event generator for the simulation of qcd - instanton induced processes in deep - inelastic scattering ( hera ) .
the qcdins package is designed as an `` add - on '' hard process generator interfaced to the general hadronic event simulation package herwig .
it incorporates the theoretically predicted production rate for instanton - induced events as well as the essential characteristics that have been derived theoretically for the partonic final state of instanton - induced processes : notably , the flavour democratic and isotropic production of the partonic final state , energy weight factors different for gluons and quarks , and a high average multiplicity @xmath0 of produced partons with a poisson distribution of the gluon multiplicity .
while the subsequent perturbative evolution of the generated partons is always handled by the herwig package , the final hadronization step may optionally be performed also by means of the general hadronic event simulation package jetset . and
qcd ; instanton ; deep - inelastic scattering ; monte carlo simulation 11.15.kc ; 12.38.lg ; 13.60.hb * program summary * + _ title of program : _ qcdins 2.0 + _ catalogue identifier : _ + _ program obtainable from : _
http://www.desy.de/~t00fri/qcdins/qcdins.html + _
computer for which the program is designed and others on which it has been tested : _ any computer with a fortran 77 compiler + _ operating systems under which the program has been tested : _
linux 2.0.x ; hp - ux 10.2 + _ programming language used : _ fortran 77 + _ memory required to execute with typical data : _ size of executable program is approximately 2.6 mb .
the size of the qcdins library itself is about 200 kb ; the required routines from the herwig and jetset libraries constitute the dominant portion of the needed memory .
+ _ no . of processors
used : _ 1 + _ has the code been vectorised or parallelized ? : _ no + _ no . of bytes in distributed program ,
including test data , etc . :
_ 1071106 + _ distribution format : _ ascii + _ cpc program library subprograms used : _ herwig [ 1 ] version 5.9 ; jetset 7.4 [ 2 ] + _ keywords : _ qcd ; instanton ; deep - inelastic scattering ; monte carlo simulation + _ nature of physical problem _ + instantons are a basic aspect of quantum chromodynamics .
being non - perturbative fluctuations of the gauge fields , they induce hard processes absent in conventional perturbation theory .
deep - inelastic lepton - nucleon scattering at hera offers a unique possible discovery window for such processes induced by qcd - instantons through their characteristic final - state signature and a sizable rate , calculable within instanton - perturbation theory .
an experimental discovery of such a novel , non - perturbative manifestation of non - abelian gauge theories would be of fundamental significance .
however , instanton - induced events are expected to make up only a small fraction of all deep - inelastic events .
therefore , a detailed knowledge of the resulting hadronic final state , along with a multi - observable analysis of experimental data by means of monte carlo techniques , is necessary .
_ method of solution _ + the qcdins package is designed as an `` add - on '' hard process generator interfaced to the general hadronic event simulation package herwig .
it incorporates the theoretically predicted production rate for instanton - induced events as well as the essential characteristics that have been derived theoretically for the partonic final state of instanton - induced processes : notably , the flavour democratic and isotropic production of the partonic final state , energy weight factors different for gluons and quarks , and a high average multiplicity @xmath0 of produced partons with a poisson distribution of the gluon multiplicity .
while the subsequent perturbative evolution of the generated partons is always handled by the herwig package , the final hadronization step may optionally be performed also by means of the general hadronic event simulation package jetset .
_ restrictions on the complexity of the problem _
+ the default values of the implemented kinematical cuts represent the state of the art limits for the reliability of the generated instanton - induced event rate and event topology . _ typical running time _ + 10 - 100 events per second for a pc with pentium cpu , depending on its clock frequency . on a hp 9000/735 ( 99 mhz ) workstation , 6 events per second
are generated .
_ unusual features of the program _
+ none _ references _ + [ 1 ] g. marchesini et al .
, comput .
phys .
commun .
67 ( 1992 ) 465 .
+ [ 2 ] t. sjstrand , comput .
phys .
commun .
82 ( 1994 ) 74 .
* long write - up * + |
lon foucault built his first pendulum to demonstrate the turning of the earth in the basement of a building , using a roughly two meter long fiber @xcite .
he also soon recognized the problem arising from the intrinsic precession of a spherical pendulum caused by unwanted ellipsoidal motion .
imperfections in the suspension or initial conditions of the pendulum generally cause this to quickly grow to the point that the precession due to the earth s turning is overwhelmed .
the pendulum can come to precess in either sense ( clockwise or counterclockwise ) at almost any rate , or indeed even cease all precession .
these practical problems are mitigated in pendula of great length , and so most are constructed to have lengths of tens of meters , starting with the celebrated 67 m long device built by foucault in paris in 1851 .
the precession of an ideal spherical pendulum with no ellipsoidal motion is caused by the non - inertial nature of the reference frame tied to the surface of the earth .
the so - called coriolis force advances the plane of the pendulum s motion by an amount @xmath0 where @xmath1 is the sidereal rate of rotation of the earth , @xmath2 is the latitude of the pendulum measured from the equator , and @xmath3 is the foucault precession rate ; many textbooks treat this problem . near @xmath4
northern latitude , this amounts to almost @xmath5 of clockwise advance per hour , for an 18 hour half - rotation of the pendulum , after which the motion repeats .
this amount of precession is easily masked , however , by the intrinsic precession of a less than perfect pendulum that develops some non - planar ellipsoidal motion .
the construction of a room - height or table - top version of a foucault pendulum thus presents a technical challenge , first to minimize the amount of ellipsoidal motion that accrues as the pendulum swings , and secondly to compensate in some way for the irreducible amount of this motion that remains . in this paper
we will first discuss the dynamics of the spherical pendulum that lead to the problem ( section ii ) .
then we introduce a method that stops the ellipsoidal motion of the pendulum from causing precession , and show that this immunity is , to first order , independent of the minor axis of the ellipse .
the method hinges on the observation that pushing the pendulum bob away from the origin after it passes , rather than either pulling it in or alternately pulling and pushing it , acts in a way to counter the unwanted intrinsic precession ( section iii ) .
we then present the design of a pendulum and a driving mechanism to exploit this method ( section iv ) , and demonstrate the validity of this approach by discussing the supporting experimental results ( section v ) .
finally , we contrast our results with earlier published work on foucault pendula and summarize how our method and design are new and unique ( section vi ) .
the dynamics of the idealized spherical pendulum are determined by the centrally - directed force of gravity and initial conditions , and lead to approximately elliptical motion with a semi - major axis _ a _ and a semi - minor axis _
b _ , as shown in fig .
[ fig:1 ] .
it is somewhat counter - intuitive but true that the centrally - directed restoring force of gravity results in a constant intrinsic precession rate @xmath6 about the _ z _ axis .
this precession arises from the symmetry - breaking condition of having a finite minor axis , _ b_. the rate is linearly proportional to _ b _ , and its sense is in the same direction as the ellipsoidal motion .
an example of the exact path of a pendulum with a large ratio of _ b / a _ is given in fig .
[ fig:2 ] , which illustrates a numerical solution of the equations of motion of a spherical pendulum .
( color online ) planar view of the approximate path of a spherical pendulum with semi - major axis _ a _ and semi - minor axis _
b _ that is moving in a counterclockwise ellipsoid .
the suspension is centered on the _ z _ axis above the origin .
the pendulum is precessing at rate @xmath6 , and in one full cycle the apex advances by a distance @xmath7 , as suggested by the light dotted and rotated ellipse .
the impulsive driving force is applied at _ x _ = _ d _ , and it is resolved into components parallel and perpendicular to the major axis .
the minor axis can be larger or smaller , resulting in a _
b_-dependent magnitude of the transverse force @xmath8 for a fixed longitudinal force @xmath9 . ]
( color online ) numerical simulation ( using mathematica ) of a spherical pendulum with a large ratio of semi - minor to semi - major ellipse axes .
the arrows indicate the counterclockwise intrinsic precession , in the same sense as the motion of the pendulum . ]
any ellipsoidal motion that develops in a pendulum will result in an intrinsic precession rate @xmath6 that is wholly unrelated to the foucault precession @xmath10 of eq .
this has been worked out in detail by olsson @xcite and pippard @xcite , among others , and it appears in some textbooks , for example in ref @xcite .
the main result is @xmath11 where @xmath12 is the pendulum s angular frequency , _
l _ is the length , _ t _ is the period , and _
g _ is the acceleration due to gravity .
the formula is the lowest - order term in a complicated motion , but it is easily sufficient for our purpose .
the area of an ellipse , _ a _ , is given by @xmath13 , so another way to write the intrinsic precession rate is @xmath14 from the first expression , note that the ratio of @xmath6 ( the very slow intrinsic precession ) to @xmath15 ( the rapid pendular rate ) is proportional to the ratio of the area of the ellipse to the area of a sphere of radius _
l _ , that is , @xmath16 to minimize this ratio , foucault pendula are generally made with _
l _ very large compared to the axes _ a _ and _ b _ , since it is comparatively easy to keep _ b _ small while making _
l _ large .
besides causing precession , ellipsoidal motion changes the central oscillation frequency of the pendulum from @xmath17 to @xmath18 the fractional change in frequency due to a finite semi - minor axis _
b _ turns out to be of order @xmath19 , and therefore negligible for the present discussion .
every foucault pendulum , no matter how carefully constructed to avoid asymmetries in its suspension , and no matter how carefully `` launched '' to make ellipsoidal area _ a _ as small as reasonably achievable will , over time , acquire an intrinsic precession @xmath6 that can easily grow to overwhelm the coriolis - force induced foucault precession @xmath3 . near @xmath4 latitude , for a pendulum of length _
= 3.0 meters and semi - major axis _ a _ = 16 .
cm , the foucault rate is equaled with a semi - minor axis of _ b _ = 3.9 mm .
the unwanted intrinsic precession may add or subtract from the foucault precession ; when it is subtractive , then the pendulum stops all precession when the corresponding value of _ b _ is reached . in our experience
, it is easy to launch such a pendulum with semi - minor axis well under one millimeter , but under free oscillation , on the time scale of only two to three minutes , _ b _ grows enough such that @xmath6 dominates @xmath10 . when the pendulum at an extremum , at @xmath20 in fig .
[ fig:1 ] , its motion is entirely transverse , with momentum @xmath21 as large as it gets .
preferential damping of this component of the motion will reduce the unwanted ellipsoidal excursions . in previous work
, this has been tried using a so - called charron s ring around the suspension wire near the top @xcite , or letting a part of the pendulum bob scrape an annular disk @xcite at _
r = a _ , or using eddy current damping between a permanent magnet in the bob and a non - ferrous metal annular disk @xcite located near _
x = a_. we adopt the touch - free eddy - current damping method in the design we present later in this paper . even in principle
, none of these methods will stop ellipsoidal motion completely , so an additional method is needed to cope with any remaining intrinsic precession .
every pendulum suffers dissipative losses of energy , mainly due to air friction .
long , very massive , museum - type pendula can simply be relaunched once every day or so , but a small pendulum has a free exponential decay time of order one half hour , so a `` driving '' mechanism is needed to restore the lost energy .
various mechanisms have been reported for this task @xcite . like some others , we will use magnetic induction to sense the passage near the origin of a permanent magnet embedded in the pendulum bob .
a carefully - timed electromagnetic impulse then imparts lost momentum to the bob .
we will introduce a quantitative method for using a magnetic push not only to compensate for dissipative losses , but also to compensate for ellipsoidal precession .
the perturbations that lead to ellipsoidal motion are , in our experience , in approximate decreasing order of severity , ( 1 ) internal stresses or other imperfections in the fiber supporting the pendulum bob , ( 2 ) less than perfectly symmetric suspension of the fiber at its upper end , ( 3 ) nearby iron objects that result in asymmetric force on the drive magnet in the bob , ( 4 ) a driving coil that is not sufficiently level and centered under the pendulum .
all but the first of these were straightforward to reduce to insignificance , but the first was persistent .
this led to the necessity of finding a method to evade the problem of ellipsoidal motion rather than to remove it .
there is no simple force law that leads to the intrinsic precession @xmath6 , but rather it is an inescapable feature of the spherical pendulum .
nevertheless , we can apply a separate perturbative force to counteract , i.e. nullify , the intrinsic precession .
we discuss this in the next section .
as shown in fig . [ fig:1 ] , the driving force that pushes the pendulum away from the origin can be thought of as consisting of components parallel and perpendicular to the major axis .
the parallel component @xmath9 is the larger one , and it is adjusted to overcome the dissipative forces such as air resistance .
the perpendicular component of the pushing force @xmath22 counteracts the intrinsic precession , and we will now show that it is possible to select distance _
d _ at which the impulsive drive force is applied to stop the precession @xmath6 . the crucial result will be that this distance is independent of semi - minor axis _
b_. starting from @xmath23 , in one half cycle of duration @xmath24 , intrinsic precession advances the ellipse in angle by @xmath25 .
the pendulum arrives at @xmath26 and @xmath27 where @xmath7 is the transverse displacement at the apex of the ellipse .
however , we arrange to apply an impulsive momentum change at @xmath28 that is calibrated to move the pendulum bob by a distance @xmath29 as it traverses the remaining longitudinal distance @xmath30 .
if this is done , the bob will arrive at location @xmath31 , @xmath32 , as desired .
the impulse and momentum change are related by @xmath33 where @xmath34 is the duration of the applied force and _ m _ is the mass of the bob .
it turns out that @xmath34 is a few milliseconds , compared to hundreds of milliseconds for the duration of the swing , so it is appropriate to use the impulse formulation of newton s second law .
we treat this perturbation in the _ y _ direction as though the bob were free of other forces , which is reasonable since @xmath7 is of order 10 microns compared to _ a _ of order 10 centimeters .
the horizontal component of the motion is simply @xmath35 where @xmath17 is the angular frequency characterizing the pendular oscillations . to good accuracy ,
the time @xmath36 between passage closest to the origin and the instant the pendulum reaches _ x = d _ is thus @xmath37 in one quarter of the full period _
the pendulum reaches its apex .
therefore , one way to write a relation between the distance @xmath7 that the pendulum advances and the transverse velocity @xmath38 that must be imparted by the driver is @xmath39 from the geometry of the situation shown in fig .
[ fig:1 ] we see that that the driving force components are related by @xmath40 and , using the formula for an ellipse @xmath41 we have @xmath42 combining eqs , , , , and we arrive at @xmath43 one sees in eq the crucial cancellation of the factor _
b _ that occurs between the intrinsic precession rate and in the perpendicular component of the force .
this leads to the result that follows being independent of the transverse size of the ellipse .
this , in turn , means the result is insensitive to exactly what non - central forces may act on the pendulum to cause non - vanishing ellipsoidal motion .
an approximation is being made that @xmath44 does not depend on _ b _ ; this will be justified later .
it is now convenient to introduce some dimensionless scaling parameters .
let @xmath45 which is the ratio of the momentum of the pendulum as it crosses the origin to the momentum kick it receives on each half oscillation .
the denominator is the momentum the driver must supply to compensate for the dissipative losses in order to maintain the full amplitude of the swing .
the factor of @xmath46 stems from letting _ q _ represent the conventional `` quality factor '' of an oscillator , specifically that @xmath47 we expect this ratio to be quite large , on the order of 1000
. then let @xmath48 be the scaled amplitude of the pendulum , that is , the amplitude divided by the length ; this parameter is of order 0.1 for a practical pendulum .
next , let @xmath49 be the scaled distance from the origin at which the driving force is applied , written as a fraction of the amplitude ; this value can range from near zero to unity . with these dimensionless variables
, eq can be written @xmath50 equation 19 is the main result of this model .
it relates the scaled distance @xmath51 at which the impulsive driving force must be applied on each half oscillation , to the physical parameters of the pendulum , specifically the scaled amplitude @xmath52 and the quality of the oscillator _
q_. when eq is satisfied , the intrinsic precession is stopped or nullified , and the result is independent of the transverse size of the ellipse .
figure [ fig:3]a is an illustration of this result for the cases of three different lengths of pendulum , each with a maximum excursion of 0.15 meters , as a function of the oscillation parameter , _
q_. it is seen that a low - loss pendulum with a larger value of _ q _ will have to be pushed when it is closer to the origin , while a pendulum with more dissipative losses , and therefore lower _ q _ , will have be driven further away from the origin . a high - quality
, low - loss pendulum needs little energy input .
therefore , the transverse kick , which scales in magnitude with the longitudinal kick at a given @xmath51 , must be applied early in the ellipsoidal swing , when the force has a larger transverse component . part ( b ) of fig .
[ fig:3 ] illustrates the relationship between the location of the impulse , @xmath51 , and the time , @xmath36 , at which it is applied , as per eq .
( color online ) ( a ) for three pendulum lengths ( _ l _ ) with amplitude of 0.15 meter , the relationship of the scaled driving distance ( @xmath53 ) versus the quality factor _ q _ for the oscillation .
curves are for an _ l _ = 3.0 meter ( solid blue ) , 2.0 meter ( dashed red ) and 1.0 meter pendulum ( dotted green ) .
( b ) for the same pendulum lengths as in ( a ) , the distance versus time relationship for the driving pulse . ] taken together , the two parts of the fig .
[ fig:3 ] enable design of a foucault pendulum not plagued by intrinsic precession .
for example , a 3.0 meter long pendulum with a 0.15 meter amplitude ( @xmath54 ) and a quality factor of @xmath55 must receive its impulsive drive at @xmath56 , which corresponds to time @xmath57 msec and a drive location of _ d _ = 7.4 cm from the origin . on the other hand ,
a 1.0 meter long pendulum with the same @xmath52 and _ q _ must receive its drive pulse at @xmath58 , at a time of 28 msec and 1.3 cm from the origin . as introduced in eq ,
_ q _ is proportional to the ratio of the momentum of free oscillation at the origin to the damping / driving momentum that accrues with each half period of oscillation .
in fact , it is quite easy to measure this parameter for an actual pendulum by measuring the period , _ t _ , and the exponential decay time , @xmath59 .
the motion of the pendulum in the presence of velocity - dependent losses such as air friction is given , to a good approximation at low speeds , by the damped harmonic oscillator formula @xmath60 the momentum loss between times _ t _ = 0 and _ t = t_/2 is @xmath61 this leads to @xmath62 from which _ q _ can be determined experimentally . for the actual three meter pendulum discussed below
, we found the decay time to be about 30 minutes and hence _ q _ to be roughly 1600 .
eq can be used with eq in order to express eq in dimensioned variables as @xmath63 one important approximation that was made needs to be examined . in going from eq to eq , the cancellation of the semi - minor ellipse parameter _
b _ was crucial to showing that the result is independent of inevitable changes in the size of the pendulum s ellipsoidal motion .
in fact , @xmath9 does depend very slightly on _ b _ in the design we will discuss in the following sections .
the pushing agent is a magnetic coil that produces a pulsed dipole field .
it acts against a permanent magnetic dipole inside the pendulum bob .
though we operate in the near field of the coil , we presume that the magnetic field has a distance dependence of the dipole form @xmath64 , where @xmath65 and @xmath66 are suitable scales . the dipole - dipole repulsion that drives the pendulum goes as the gradient of this field ,
so the interaction has the form @xmath67 where the components of @xmath68 are what we earlier called @xmath69 and @xmath9 .
it is easy to expand this expression to show that @xmath70 as seen in the second term in the curly brackets , there is a quadratic dependence on the semi - minor axis _
b _ that we ignored earlier .
this term is very small : for typical values of @xmath71 , and @xmath72 , the second term is 0.015 , which is much less than 1.0 .
hence this approximation was justified .
to verify the mathematical model discussed above , a pendulum of about three meter length was built .
the mechanics and drive electronics will be described here . an image of the lower end of the setup is shown in fig .
[ fig:4 ] .
( color online ) image of the copper pendulum bob suspended above the sensing and driving coils ( red wire on a white bobbin ) .
the aluminum damping ring on brass legs is seen .
the white band carries angular position markings . the driver box and an oscilloscope are visible in the background . ] the mass of the pendulum was in the form of a lathe - turned copper cylinder , with a diameter of 2.75 ( 6.99 cm ) and height 3.00 ( 7.62 cm ) .
a 3/8 ( .953 cm ) axial hole was drilled and tapped to accept a small threaded set screw with a 0.040 ( 0.10 cm ) hole in its center .
the small hole was for the fiber suspending the pendulum , and the threading was used to adjust the distance between the center of mass and the suspension point . in practice ,
the fiber was held flush with the top surface of the bob .
the bottom of the cylindrical bob was chamfered by 0.10 ( 0.25 cm ) to allow closer approach to the damping ring . to hold the permanent magnet ,
a recess was milled into the bottom of the mass , of diameter 3/4 ( 1.90 cm ) and depth 1/4 ( 0.635 cm ) , which were the dimensions of the magnet .
the total mass of the pendulum bob , apart from the magnet , was 2.47 kg .
this is a small mass compared to what has generally been reported for foucault pendula .
the fiber suspending the mass was an ordinary and inexpensive polymer material of the type used for yard trimmers .
the diameter was 0.040
( 0.10 cm ) .
the cross section of the fiber was circular and the diameter was uniform to better than 1% .
the polymeric material was thought to be advantageous due to its lack of crystalline internal structure .
it was found that the polymer material was very flexible , yet strong , and did not suffer the bending fatigue and eventual failure of some metal fibers we tried . in practice
, we found that the growth of elliptical motion of the pendulum was dominated by an unobservable non - uniformity of this fiber , but it was less severe than with various metal fibers .
tests wherein the fiber was rotated without changing any other aspect of the pendulum showed this to be the case .
thermal expansion and contraction of the fiber was large enough to necessitate occasional shimming of the pendulum s length at the level of about one millimeter .
the upper end of the fiber passed through a close - fitting drilled hole in an aluminum plate .
the hole had a sharp rim , and no special steps were taken to soften the bend of the fiber as it exited the hole . on the upper side of the plate the fiber was clamped in a way that thin shims could be inserted to fine tune the length .
the plate was leveled and clamped to rigid brackets on the ceiling of the laboratory .
the upper support of the fiber used in this setup , though carefully arranged , was certainly less exacting than what has been found to be necessary for other foucault pendula reported in the literature .
we view this as an advantage of our design .
the permanent magnet placed inside the pendulum bob was a neodymium - iron - boron disk magnet , placed with its dipole axis vertical .
the axial field strength was 2.1 kgauss on contact , but it varied by @xmath7310% from one edge of the disk to the other .
similarly , the radial field was 1.0 kg at the edge of the disk , but varied also by @xmath7310% from one side to the other .
surprisingly perhaps , these variations did not affect the performance of the pendulum .
one purpose of the magnet was to provide eddy - current damping when the bob passed over an aluminum ring near the extrema of its motion .
this `` damping ring '' had an inner diameter of 11 7/8 ( 30 cm ) , and was 1/4 ( 0.64 cm ) thick .
three legs made of threaded brass rod were used to level the ring to a precision of less than half a millimeter of variation around the perimeter .
the amplitude of the pendulum was adjusted such that the magnet passed over the inner edge of the ring with a clearance of three to four millimeters .
this maximized damping of the unwanted transverse precessional speed .
since the eddy current damping force is velocity dependent , it can never entirely stop this motion , but our experience was that it limited the ellipsoidal motion to a semi - minor radius of less than half a centimeter . as a side note ,
lon foucault first identified the eddy - current phenomenon in 1851 , thus we use two phenomena in this investigation that are attributed to him .
two concentric coils under the pendulum sensed and controlled its motion .
both coils were wound from 22 awg copper wire on polyethylene bobbins , and both had 240 turns .
the inner coil was designated as the `` drive '' coil used for pulsing the pendulum on each half cycle .
it had a mean radius of 0.80 ( 2.0 cm ) and a total length of 100 ( 30 m ) of wire .
the outer coil was the `` sense '' coil for detecting the approach of the pendulum .
it had a mean radius of 1.5 ( 3.8 cm ) and a total length of 187 ( 57 m ) of wire .
the supply leads to both coils were twisted pair copper wire near the coils , and rg-174 coaxial cable at distances where the steel ground braid of the latter no longer perturbed the motion . the concentric sense and drive coils were supported by a small aluminum platform that sat on the main table , supported by three brass screws with knurled heads .
these were used to level the coils : imperfect leveling caused the induced signal from the pendulum , described below , to be asymmetric .
the clearance between the bottom of the pendulum and the top of the coils was 6@xmath731 mm .
the sense coil acted to detect the induced emf of the swinging pendulum via faraday induction .
the typical waveform is shown in fig .
[ fig:5 ] , showing that the approaching coil induced a negative voltage that peaked at about @xmath74 mv , as viewed across @xmath75 input impedance on an oscilloscope .
when the pendulum was centered over the coil , the voltage crossed upwards through zero , followed by a positive excursion of inverted shape and the same magnitude @xmath76 as the negative part .
the peak magnitudes of the waveform , v@xmath77 , depended on the speed of the pendulum ( and hence the amplitude of the swing ) , as well as the distance between the pendulum magnet and the coil .
monitoring v@xmath77 over time was a sensitive way to detect small changes in the pendulum s amplitude and/or length . induced voltage at the input stage of the driver circuit , due to the magnet in the pendulum approaching , passing over , and then receding from the sense coil .
upper trace shows a complete cycle of two passages .
lower zoomed trace shows how on the trailing end of the sense pulse the drive pulse causes a large bipolar induced pulse in the sense coil .
screen capture was from a tds 3032b oscilloscope . ] the pendulum driver circuit developed for this study is shown in fig .
[ fig:6 ] , and operated as follows .
the initial signal from the coil was filtered with a 1 @xmath78f capacitor that reduced rf noise to less than one millivolt .
an op - amp circuit of type lm324n level - shifted and amplified the signal by a factor of 200 .
this signal was the input to a sn7414n schmitt trigger that switched at 0.9 and 1.7 volts . in effect , this was a discriminator with a ttl output `` trigger '' pulse when the induced voltage went more negative than @xmath79 mv ; the trigger stayed high until the trailing end of the sense - coil signal shown in fig .
[ fig:5 ] fell .
a nand gate output was used to trigger two 555 timers called delay and inhibit ; the second input of the nand was an inhibit signal that ensured that the delay timer was not restarted by the off - scale pulse induced by activation of the drive coil .
the delay timer set the interval between the approach of the pendulum and the drive pulse in the second coil ; it was adjustable in the range from near zero to several hundred milliseconds .
the inhibit timer duration was set to be longer than the maximum possible delay plus drive times .
the output of the inhibit timer was fed via an inverter back to the nand , so the drive pulse could not re - fire the delay timer .
the output of the delay timer was passed via an inverter to the input of the third 555 timer called drive .
this timer was adjustable over a range of several tens of microseconds , and was used to set the duration of the current through the drive coil .
the output of the drive timer was used to control two high - current mosfet relays that switched current to and from the drive coil .
the circuit was designed to maintain a substantial current draw from the commercial 12v switching power supply at all times , hence the dual opposite - acting relays . a zener diode across the drive coil prevented back - emf from damaging the relay and power supply during switching .
the very brief drive current was estimated to be 7.5 amps .
the circuit was constructed using the wire - wrap method and housed with its power supply in a small box .
external connections were made using lemo - style coaxial connectors , as seen in fig .
[ fig:4 ] .
circuit diagram of the pendulum driver .
the op - amp ( lm324 ) and solid - state relays ( crydom cmx60d10 ) are supplied by + 12v , while all other components used @xmath80v . unlabeled circuit elements
use standard ttl chips .
the `` out '' connections are for monitoring the circuit on an oscilloscope . ]
the pendulum described in the preceding section was used to determine the accuracy to the model that lead to eq .
the prediction that the foucault precession rate should be unaffected by the transverse size of the elliptical motion is to be confirmed .
the measured pendulum parameters were @xmath81sec , and @xmath82 cm .
the free decay period was determined by making an exponential fit to amplitude - vs .- time data , leading to @xmath83min . from these values
we computed @xmath84 m , @xmath85 , and @xmath86 . solving eq numerically leads from these values to @xmath87 .
this value in turn can be converted using eq to an expected drive time @xmath88ms . on the other hand ,
the actual value of @xmath89 was adjusted such that a steady foucault precession , as shown below , was observed .
the direct measurement of this time using an oscilloscope gave @xmath90ms .
thus , the model expectation as given in eq is found to be in excellent agreement with experiment with about 1% precision .
figure [ fig:7 ] shows the result of a set of measurements of the precession rate spanning several days .
the abscissa angles were measured from an origin arbitrarily picked to point north , with positive azimuthal angles to the counterclockwise side , as seen from above .
red circular data points are for counterclockwise and blue squares are for clockwise ellipsoidal excursions in the pendulum s motion .
green diamond points are for readings with no measurable sense , i.e. for @xmath91 . at the latitude of @xmath92 ( pittsburgh , pennsylvania ) , the expected foucault precession rate in one sidereal day
is @xmath93 and this is shown as the blue bar .
the vertical error bars on the points represent the estimated random measurement errors , and these were dominated by a @xmath94 degree precision of the angle measurements .
the horizontal error bars represent the span of angles over which the rate was measured , with a typical span being ten degrees . the scatter of the data points clearly clusters around the expected rate , with no trend in angle or sense of ellipsoidal motion .
the simple average of the measured points gives @xmath95 , where the given uncertainty is the error on the mean , not the standard deviation , and this is the red - dotted band shown in the figure .
the weighted mean of the data gives @xmath96 , which is in very good agreement with the simple average , and both are in excellent agreement with the expectation of eq .
we prefer the former uncertainty because the wide scatter of the data points suggests that perhaps there are some small systematic effects that are not captured using the weighted mean s uncertainty .
hence the former uncertainty is a better estimate of the true uncertainty of the result .
the main candidate for a poorly - controlled systematic effect is the temperature - related changes in the length of the pendulum .
even at the sub - millimeter level , these affected the strength of the magnetic kick the pendulum received , and hence affected the value of @xmath52 , which in turn could affect control of the precession rate .
we fine - tuned the length of the pendulum with shims , as needed , to minimize this effect , but the compensation was not perfect , and this probably led to some loss of reproducibility . nevertheless , we have found the expected foucault precession rate with an angle - integrated precision of 1.5% .
( color online ) precession rate as a function of azimuthal angle of the pendulum .
red circles are for counterclockwise , while blue squares are for clockwise ellipsoidal perturbation .
green diamond points were for measurements with no discernible elliptical motion .
the expected rate is shown as the thick blue line
. the measured average precession rate ( solid ) and its uncertainty band ( dotted ) are shown as red lines . ]
as discussed in connection with eq , if @xmath36 is correctly chosen , the rate of precession in this pendulum should not depend on the size of the semi - minor axis , _ b_. we could not control _ b
_ since it depended on the small asymmetries of the fiber itself and any other perturbations of the pendulum s motion .
of course it was damped by the effect of eddy current braking against the damping ring , but there was always some irreducible amount of ellipsoidal motion to contend with .
we assign clockwise motion to have negative _ b _ values and counterclockwise motion to have positive _ b _ values .
figure [ fig:8 ] shows our measured precession rates , the same data set as in the previous figure , as a function of _ b_. the size of _ b _ was measured by watching the pendulum pass over a ruler placed under the pendulum , and the precision of doing this was no better than @xmath97 mm .
we plot the points at the mean value of the ellipse size during the measurement interval .
it is clear from the figure that there is no correlation between the two variables , thus proving the claim that they are in fact independent when the pendulum is properly arranged .
the foucault precession was clockwise in this northern - hemisphere setup , _
i.e. _ toward decreasing angles .
note that even when the pendulum was moving in a counterclockwise ellipse ( positive values of _ b _ ) that would normally precess in a counterclockwise direction , the action of the driver was such that the clockwise foucault precession rate was unaffected .
this again shows how our method nullifies the unwanted intrinsic precession .
* * ( color online ) precession rate as a function of the size of the semi - minor axis _
b _ of elliptical motion .
positive values of _ b _ are for counterclockwise ellipses , negative values for clockwise ellipses .
the horizontal lines are the same as in the previous figure . ] as further evidence that our method succeeds in controlling the motion of the pendulum , we intentionally reduced the value of the driving time by @xmath98 , so that the transverse kick given by @xmath69was larger than optimal .
this means that any ellipsoidal precession was overcompensated by the drive system , resulting in forced precession in the opposite sense to what free oscillation would produce .
figure [ fig:9 ] shows the precession rate as a function of angle for @xmath99 msec .
there is , in this new situation , substantial systematic deviation from the expected rate as a function of azimuthal angle .
the angle - integrated average rate is now @xmath100 , which is still in agreement with the expected foucault rate , but now with a much wider error band , as shown .
one also sees the propensity for clockwise ellipsoidal motion ( blue squares ) to decrease the magnitude of the precession rate , while counterclockwise motion ( red circles ) increases the magnitude of the precession rate .
this is as expected in our mathematical model .
there are again some cases of irreproducibility of the data points at a given angle .
we believe this is the result of occasional shimming of the pendulum s length during a multi - day run ; it leads to uncontrolled small variations in the performance of the system .
( color online ) precession rate as a function of azimuthal angle of the pendulum when the driving time @xmath36 is reduced by @xmath98 from the ideal value . as in fig .
[ fig:7 ] , red circles are for counterclockwise , while blue squares are for clockwise ellipsoidal motion .
green diamond points were for measurements with no discernible elliptical motion .
the expected rate is shown as the thick blue line . the measured average precession rate ( solid ) and its uncertainty band ( dotted )
are shown as red lines .
the heavy black dashed line is a guide to the eye , as discussed in the text . ]
the heavy black dashed line in fig .
[ fig:9 ] is a guide to the eye to illustrate what happens when the ellipsoidal precession is not perfectly canceled .
there is a preferred axis near @xmath101 , where the measured rate crosses the foucault rate , because here the non - central forces present in the system happen to vanish .
for more negative azimuthal angles ( which then wrap to the positive side of the diagram ) the precession rate is slower .
this is because the unwanted forces act to cause clockwise ellipsoidal motion , but the drive system is overcompensating and adding a component of counterclockwise precession . at about @xmath102 , which is @xmath103 away from the preferred axis
, there is a tipping point where the non - central forces act to push the pendulum in the other direction , counterclockwise .
but here the overcompensating drive system makes the pendulum precess more quickly clockwise .
we do not claim that the departure from the foucault rate is strictly linear , as shown , only that the trends are consistent with our understanding of the physical process . finally , in fig . [ fig:10 ]
we show the effect of reducing the drive time @xmath36 by @xmath104 on the relationship between precession rate @xmath6 and the semi - minor axis _
b_. one sees clearly a correlation between them , whereas in fig .
[ fig:8 ] there was no correlation . for positive values of _ b _ ( counterclockwise ellipses ) ,
the precession rate has increased magnitude since the overcompensating drive system `` adds '' to the foucault rate . for negative @xmath105 values ( clockwise ellipses ) the precession rate
is decreased in magnitude since the overcompensating drive system `` subtracts '' from the foucault rate .
thus , we have shown that when deviating from the prediction of eq for the correct drive distance , and therefore the correct drive time , the pendulum shows marked departure from the constant foucault precession rate that is expected .
( color online ) precession rate as a function of the size of the semi - minor axis _
b _ when the driving time @xmath36 is reduced by @xmath98 from the predicted value .
positive values of _ b _ are for counterclockwise ellipses , negative values for clockwise ellipses .
this figure is to be compared with fig .
[ fig:8 ] . ]
in the previous work of crane @xcite , he recognized the importance of nullifying the intrinsic precession that remains after damping as well as possible .
however , he used a `` push - pull '' drive system to mitigate problems of alignment of the coils with the pendulum .
this led him to introduce a carefully - placed fixed permanent magnet at the origin to provide the desired stopping of intrinsic precession .
no quantitative understanding of how to predict the placement of this magnet was offered .
it seems to us that his very delicate _
ad hoc _ adjustment of this auxiliary magnet is difficult , and , as we have shown , not necessary .
our method is simpler and more direct , in that it does not require this additional magnet .
alignment of the driving and sense coils was not found to be a problem , and the results were not sensitive to the alignment at the level of about a millimeter .
this was because the method we have introduced is in a sense self - correcting : if the drive coil causes some small amount of ellipsoidal motion , the very action of the method prevents this motion from causing unwanted precession . the work of mastner _ et al . _
@xcite made note of the benefits of a `` push only '' driving force , but they did not offer the quantitative explanation as to how and why it worked .
they built a traditional , very long , very massive pendulum , taking great care to minimize asymmetries . in conclusion , we have demonstrated the workings of a `` short '' foucault pendulum that was designed on a quantitative basis to avoid the unwanted precession due to ellipsoidal motion .
the design formula we have derived , eq , was shown to agree with measured values in our setup .
we have shown that a driving mechanism that pushes , not pulls , the pendulum is the key to canceling the intrinsic precession for all values of the semi - minor axis of the ellipse .
our driver system used faraday induction and magnetic repulsion to control the pendulum , using a circuit based on simple op - amp , logic , and timer chips .
eddy current damping was used to reduce the ellipse size , but active compensation did the rest .
the design is immune to small non - central forces that are difficult to control in a short pendulum .
we plan to further test this method on even shorter pendula , since there is no lower limit at which the model given here should apply .
we thank mr .
gary wilkin for his expert help in the machine shop .
we thank mr .
michael vahey for construction of the pendulum driver , and we thank mr .
chen ling for help with exploratory initial trials in construction of a foucault pendulum .
see for example c.f .
moppert and w. j. bonwick , `` the new foucault pendulum at monash university '' , q. jl .
* 21 * , 108 - 118 ( 1980 ) , and references therein . also haym kruglak _ et al _ , `` a short foucault pendulum for a hallway exhibit '' , am .
* 46 * , 438 - 440 ( 1978 ) . h. richard crane , `` foucault pendulum `` wall clock '' '' , am .
* 63 * , 33 - 39 ( 1995 ) ; `` short foucault pendulum : a way to eliminate the precession due to ellipticity '' , am . j. phys .
* 49 * , 1004 - 1006 ( 1981 ) . | a quantitative method is presented for stopping the intrinsic precession of a spherical pendulum due to ellipsoidal motion . removing this unwanted
precession renders the foucault precession due to the turning of the earth readily observable .
the method is insensitive to the size and direction of the perturbative forces leading to ellipsoidal motion .
we demonstrate that a short ( three meter ) pendulum can be pushed in a controlled way to make the foucault precession dominant .
the method makes room - height or table - top foucault pendula more accurate and practical to build . |
the abundance evolution of galaxy structures is a major prediction of cosmological models ( e.g. oukbir @xmath11 blanchard 1992 , 1997 , romer et al .
2001 ) . the more distant the structure , the strongest the constraint . until now
, however , we are limited to the study of the most massive distant structures .
this is because these structures are the easiest to detect at high redshift using for example x - ray selected samples ( e.g. borgani et al . 1999 , burke et al .
1997 , ebeling et al .
2000 , gioia et al .
1990 , romer et al .
2000 , vikhlinin et al .
these distant massive structures are the progenitors of the most massive z@xmath120 clusters .
the hierarchical models predict , however , that nearby intermediate mass clusters have formed from low mass high redshift systems as groups of galaxies
. the new generation of deep galaxy surveys ( vlt and keck surveys ) will be able to detect such high redshift low mass structures .
this will add an element to the large scale structure formation models .
we present in this paper a new efficient method to detect and study such structures in very large galaxy redshift samples ( section 2 ) .
we applied this method to the ssrs2 ( da costa et al .
1998 ) galaxy redshift catalog in order to quantify our detection efficiency at low redshift .
we have created a catalog of groups from this survey and we have studied the properties of these structures ( section 3 ) , comparing our results to literature studies ( carlberg et al .
2001 , girardi et al .
2000 , merchan et al .
we conclude in section 4 .
the structure detection we used is an hybrid method derived from the method used for the enacs sample of galaxy clusters ( e.g. katgert et al .
1996 ) and from classical friend - of - friend detection methods ( e.g. huchra @xmath11 geller 1982 ) .
the friend - of - friend algorithm is applied to a preprocessed catalog of elementary galaxy associations and not to individual galaxies . in this way
we reduce the computation time by a large factor ( see section 3.3 ) .
the procedure consists of four steps : * we apply a running window ( with an overlap of the half of the window size ) to the original catalog of galaxies to divide the total area covered by the survey in several elementary lines of sight .
the angular size @xmath13 of the running window is chosen as a function of the kind of structures we are trying to detect ( @xmath14 has to be close to the typical size of the structures ) and as a function of the redshift of these structures .
* we search for compact galaxy redshift associations in each of these elementary beams by using a technique similar to katgert et al .
this technique is basically detecting gaps ( fixed redshift separations between successive galaxies ) greater than a given value ( noted `` gap analysis '' in the following ) .
this value is given by the typical maximal velocity dispersion of the structures we try to detect .
for groups , we searched galaxies separated by more than 600 km / s in each elementary beam . if two galaxies are separated by more than this value , these galaxies belong to different structures .
an elementary structure has more than 2 galaxies .
* we apply a friend - of - friend algorithm to merge the elementary structures into larger real structures .
two elementary structures were assumed to be connected if they were closer than @xmath14 .
this value is an approximation of the maximal diameter of a group in real space .
this allows to generate a preliminary catalog of interesting regions with potential structures inside . *
finally , we re - apply the gap method for each of these regions in order to detect real structures in the catalog .
the ssrs2 catalog is a sample of 5369 galaxies , covering a region of 1.70 sr of the southern sky ( @xmath15 and @xmath16 for the southern galactic cap and @xmath17 and @xmath18 for the northern galactic cap ) .
this catalog is 99% complete down to m@xmath19 .
the m@xmath20 magnitude is close to a b magnitude ( see da costa et al .
the precision of the individual galaxy redshift measurements is @xmath1240 km / s .
detailed information on this catalog can be found in da costa et al .
we limited our analysis to the 2000@xmath21cz@xmath2120000 km / s range . to detect groups
, we used @xmath22 .
this value is an approximation of the maximal diameter of a group and gives the most similar results compared to the ssrs2 group analysis by merchan et al .
we kept only the structures with more than 4 galaxies and less than 40 galaxies ( to be consistent with merchan et al .
structures with less than or with 3 galaxies are possibly close superposition effects and not real dynamical structures ( ramella et al .
structures with more than 40 galaxies are likely to be rich clusters or diffuse elongated filaments of galaxies . the detected structures are described in table 1 and 2 .
the mean number of galaxies per structure is @xmath23 and 90% of the structures have less than 11 galaxies .
the largest group has 32 galaxies .
the richest groups are the closer ones , in agreement with the detection of a uniform richness class of groups .
this is because , due to apparent limiting magnitude , the closest groups have more galaxies brighter than the ssrs2 catalog magnitude limit .
the groups are more or less circular : they have the same angular extension in @xmath24 ( right ascension ) and @xmath25 ( declination ) .
we have @xmath26arcmin ( computed with 80 groups ) .
this means that we are not detecting systematically elongated structures such as poor filaments .
the mean extension of the detected groups ( computed as the dispersion of the coordinates of the galaxies of a given group ) is @xmath27 @xmath28 .
we have no variation of the group size as a function of the redshift , but the richest groups are also the larger ones . the groups with more than 10 galaxies have a size in the range [ 0.45;0.8 ] @xmath28 .
the groups with less than 10 galaxies have a size in the range [ 0.05;0.65 ] @xmath28 . finally , the mean virial radius r@xmath29 of the groups is @xmath30 ( in the range [ 0.06;0.87 ] @xmath28 , see table 1 and 2 ) .
we computed velocity dispersions using a robust estimator ( beers et al . 1990 ) .
this gives a more realistic estimate of the velocity dispersion for groups with more than 10 galaxies , but still a low number , compared to classical estimators . for groups with less than 10 galaxies
, there is no significant difference between this robust estimator and the classical estimator of the velocity dispersion ( lax 1985 ) .
the mean velocity dispersion of the groups is @xmath31km / s .
none of them has a velocity dispersion greater than 500 km / s ( see fig 1 ) .
these values are consistent with several other computations for groups ( e.g. carlberg et al .
2001 with a median velocity dispersion of 200 km / s ) .
this means that we detected no rich clusters .
the groups with the highest velocity dispersions are also the richest ( see fig 2 ) but we have no significant relation between the group mean redshift and velocity dispersion .
this is , one more time , in favor of a uniform detection of groups as a function of redshift . in order to check how many spurious groups we are detecting ( groups with a large velocity dispersion , but a low number of galaxies : chance alignments ) , we plotted the velocity dispersion as a function of the number of galaxies inside the groups ( fig 2 ) .
only 10% of the groups have less than 5 galaxies and a velocity dispersion greater than 300 km / s .
this can be assumed to be the contamination rate of our catalog . combining the virial radius and the velocity dispersion , we computed the mass using the standard virial mass estimator ( m @xmath32 virial radius @xmath33 ( velocity dispersion)@xmath34 ) ( e.g. ramella et al . 1997 ) . about 85% of the sample has a mass in the range [ @xmath35;@xmath36 solar masses and
97.5% has a mass in the range [ @xmath37;@xmath36 solar masses .
none of the groups has a mass greater than @xmath38 solar masses .
the mean mass is ( @xmath39)@xmath40 solar masses .
all the masses computed by merchan et al .
( 2000 ) are in the range [ @xmath35;@xmath36 solar masses .
this means that we have slightly lower masses in the sample ( for 15% of the sample ) .
this is probably due to the fact that we are using robust estimators to compute velocity dispersions .
such estimators are less biased toward high values for small samples of galaxies .
the velocity dispersions and masses are given in table 1 and 2 . in order to estimate the mass uncertainty , we used galaxy groups with more than 10 galaxies . for these groups ,
a reliable error can be computed for velocity dispersion using 100 bootstrap resamplings .
this is obviously a lower limit of the velocity dispersion uncertainty as this uncertainty is probably larger for smaller groups .
the 1-@xmath8 uncertainty is : @xmath41 .
the mass uncertainty can be written as : @xmath42 assuming that the velocity dispersion uncertainty is the dominant factor , we have the 1-@xmath8 error on the mass : @xmath43 .
this large error raises the question of the validity of virial mass estimates for groups of galaxies as these structures are probably not all in equilibrium ( e.g. dos santos @xmath11 mamon 1999 ) .
we compared our catalog with the one of merchan et al .
( 2000 ) that used the same galaxy sample .
fig 3 shows the group redshift distribution of the two catalogs .
we are detecting similar numbers of groups up to cz@xmath128000 km / s .
the mean difference per bin is 5% ( with redshift bins of 1250 km / s width ) .
merchan et al .
are detecting more groups at higher redshift .
however , their catalog begins to be incomplete above 5000 km / s .
this means that we have the same detection completeness level , in agreement with press @xmath11 schechter ( 1974 ) predictions ( merchan et al .
2000 ) . finally , we estimated the 2-point correlation function for the sample . using a power law approximation of the form @xmath44
, we have @xmath45 and @xmath46 .
this is consistent at the 1@xmath8 level with the values of merchan et al . (
2000 ) : @xmath47 and @xmath48 , girardi et al .
( 2000 ) : @xmath49 and @xmath50 or carlberg et al .
( 2001 ) : @xmath51 and @xmath52 ( fixed ) .
we detect a positive signal up to @xmath1290 @xmath28 ( similar to the values of merchan et al .
these estimates are all in good agreement and give confidence in the detection method .
the only significant difference is a lower group detection rate at the faint end of the galaxy catalog .
however , this is not affecting the completeness level of our group sample as we detect the same number of groups up to cz@xmath128000 km / s ( assuming that the merchan et al .
completeness limit is correct ) .
the redshift completeness limit close to 5000 km / s is directly related to the magnitude limit of the catalog : m@xmath53 .
it gives an absolute magnitude of -18 + 5logh .
this is the magnitude limit for the galaxies lying inside groups before the completeness limit .
we can extrapolate this redshift completeness limit assuming the typical depth of the new generation of redshift surveys using 10 meter class telescopes ( keck and vlt ) .
for example , the deep part of the virmos survey should provide a redshift survey 80% complete down to b@xmath1226 , and the shallow part of this survey a 50% complete redshift catalog down to b@xmath1224 . assuming a k - correction proportional to the redshift
, it should provide a catalog of groups complete up to [email protected] for the deep catalog and [email protected] for the shallow catalog .
this is enough to sample nearly all classes of galaxy structures up to [email protected] .
this will allow to put strong constraints on cosmological models ( e.g. romer et al .
2001 ) as well as sampling the internal structure evolution of groups ( and more massive clusters ) up to these redshifts .
the time required to treat these future samples with our method is very short .
we used the simulations of steve hatton ( private communication ) in order to estimate these times .
these simulations cover 1 deg@xmath34 and reproduce an @xmath55=0.3 and @xmath56=0.7 universe , including several structures as clusters or groups .
we used a single 6 running window for simplicity and we selected randomly different sub - samples of galaxies in this catalog .
the computation times are given in fig 4 .
we see that a catalog of 150000 galaxies , comparable to the spectroscopic catalogs which will be produced by the vlt and keck surveys , is completely analyzed in about 2 minutes ( using an es40 compaq , processor eb67 at 600 mhz ) .
we computed the luminosity function of the group galaxies of the detected sample .
the galaxy catalog is complete down to m@xmath57 with groups up to 8000 km / s . limiting ourselves to groups with
velocity lower than 6000 km / s ( this velocity limit is close to the completeness limit of the group sample ) , we have a galaxy catalog complete down to m@xmath20=-18.4 + 5logh . using the estimate of zabludoff @xmath11 mulchaey ( 2000 )
, this corresponds to about m@xmath58 + 1.5 .
we fitted a schechter function on the magnitude distribution down to m@xmath20=-18.4 + 5logh and for groups with cz@xmath596000 km / s using a maximum likelihood technique ( e.g. lobo et al
. 1997 ) .
the absolute magnitudes were k - corrected ( but this is a minor correction due to the low redshift of the sample : see table 1 and 2 ) using the k - correction -3.5@xmath60z ( rauzy et al .
we also corrected for galactic extinction using the work by schlegel et al .
the correction we applied for the magnitudes was -4.325@xmath33e(b - v ) .
the mean correction was 16% of the luminosity ( but up to 60% for the worse cases : see table 1 and 2 ) .
the best fit parameters of a schechter function are m@xmath58=([email protected])+5logh and @[email protected] ( estimates with 1-@xmath8 error ) .
this is consistent at the 1-@xmath8 level with the estimates of zabludoff @xmath11 mulchaey ( 2000 ) : m@xmath61=([email protected])+5logh ( assuming m@[email protected] ) and @[email protected] for groups in 2800@xmath21cz@xmath217700 km / s .
these values are also similar at the 1-@xmath8 level with the estimates of rauzy et al .
( 1998 ) for rich clusters of galaxies . in order to check the robustness of our estimates
, we split the sample in two parts : the northern and the southern galactic cap .
the southern cap gives m@xmath58=([email protected])+5logh and @[email protected] and the northern cap gives m@xmath58=([email protected])+5logh and @[email protected] . despite the larger uncertainty , these values are still consistent , with , however , a mild tendency to have more faint galaxies in the southern galactic cap .
we computed the total luminosity of each group summing up all the individual magnitudes ( see table 1 and 2 ) .
we used m@xmath64=5.53 ( approximation of m@xmath65 ) .
we corrected these values for incompleteness due to galaxy catalog magnitude limit by using the ratio between the luminosity function integrated from m@xmath66=-23 + 5logh . to -10 + 5logh and integrated between the faintest magnitude of each group and m@xmath66=-10 + 5logh .
this is because we assumed that the faintest galaxies were not fainter than -10 + 5logh .
this limit has , however , a moderate influence on this correction factor .
we used the luminosity function computed in the previous section with m@xmath58=([email protected])+5logh and @[email protected] .
the mean correction was 135% of the raw group luminosity .
the total luminosities and the individual completeness corrections are given in table 1 and 2 .
combining estimates of the mass and of the luminosity , we computed the mass to light ratio ( see table 1 and 2 ) .
the mean value is 550 hm@xmath7/l@xmath67 and the median is 250 hm@xmath7/l@xmath7 . we compared these estimates with those of carlberg et al .
( 2001 ) . in the r band , they computed median values in the range [ 150;250 ] hm@xmath7/l@xmath7 . assuming m@xmath68-m@xmath69=-1.17 ( allen 1973 ) and m@xmath70-m@xmath71=-1.5 for galaxies in nearby structures ( e.g. katgert et al .
1998 ) , it corresponds to median values in the range [ 204;340 ] hm@xmath7/l@xmath67 in the ssrs2 magnitude passband ( multiplication factor of 10@xmath72 ) .
this is in good agreement with our estimates .
we found a relation between the mass to light ratio and the velocity dispersion @xmath73 for our sample ( see fig . 5 and table 3 ) .
these relations are qualitatively similar to the relation we found between rich cluster velocity dispersion and cluster mass to light ratio ( adami et al .
1998 ) : @xmath74 .
the slope is , however , significantly steeper for groups than for clusters .
the slope of these relations is also similar to that of carlberg et al .
( 2001 ) which is close to 3.5 .
the more massive the group ( or the cluster ) , the higher the mass to light ratio , and therefore , the larger the amount of dark matter inside the group ( or cluster ) .
another explanation would be a very efficient stripping of the gas of the galaxies in rich groups ( as opposed to very low mass groups ) .
this would be due to a denser intra group medium or a more efficient tidal stripping due to the higher number of galaxies ( e.g. dos santos @xmath11 mamon 1999 ) in rich groups .
it would lower the star formation rate in galaxies , inducing a higher mass to light ratio for the richest groups ( see also carlberg et al .
the dispersion of the relation is also smaller for richer groups .
this is probably due to a better computation of the velocity dispersion due to the larger number of available galaxies .
schaeffer et al . ( 1993 ) or adami et al .
( 1998 ) have shown , using optical data , that clusters of galaxies populate a well defined area in the [ total luminosity : l , characteristic radius of a beta profile : r@xmath75 , velocity dispersion : @xmath8 ] space .
adami et al . ( 1998 ) found the following relation : l@xmath77 r@xmath78@xmath79@xmath80 with a dispersion of the relation of 5@xmath81 using the virial radius instead of the characteristic radius of a beta profile would give : l@xmath77 r@xmath82@xmath79@xmath83 this relation called the fundamental plane is also detected for elliptical galaxies , but with different coefficients ( see adami et al .
1998 for a comparison ) .
this is interpreted as the different relaxation state of elliptical galaxies and clusters of galaxies .
we searched for the same kind of relation in the groups we detected . using all the groups
, we found no relation between l , r@xmath84 and @xmath8 . using only groups with more than 8 galaxies
, we found : l@xmath77 r@xmath85@xmath79@xmath86 with a dispersion of 58@xmath81 the coefficients are consistent with those of clusters of galaxies , but the uncertainties are very large , and the intrinsic dispersion of the relation is 10 times larger than for clusters .
we conclude that a mean relation between l , r@xmath29 and @xmath8 probably exists for groups , but this relation is much less well defined than for rich clusters .
moreover , poor groups ( less than 7 galaxies ) probably have too large a velocity dispersion uncertainty to allow any detection of this relation .
crrccrccccrrr n & @xmath87 & @xmath88 & z & obs . lum . &
e(b - v ) & ext . & comp . & tot .
mass & m / l & @xmath89 & rvir + 5 & 3.2488 & -24.3504 & 0.0255 & 5.33e+10 & 0.02220 & 1.09 & 3.20 & 1.86e+11 & 1.90e+12 & 10.2 & 82 & 0.13 + 4 & 3.2868 & -7.4360 & 0.0181 & 4.54e+10 & 0.03389 & 1.14 & 1.94 & 1.01e+11 & 4.91e+13 & 486.9 & 286 & 0.27 + 4 & 5.5908 & -4.3517 & 0.0138 & 1.13e+10 & 0.03145 & 1.13 & 1.60 & 2.04e+11 & 2.13e+14 & 1043.7 & 401 & 0.6 + 5 & 6.8999 & -9.0373 & 0.0192 & 2.99e+10 & 0.03553 & 1.15 & 2.19 & 7.56e+10 & 1.44e+14 & 1905.7 & 359 & 0.51 + 4 & 7.0446 & -37.3208 & 0.0245 & 3.91e+10 & 0.02375 & 1.10 & 3.13 & 1.34e+11 & 4.04e+12 & 30.1 & 86 & 0.25 + 4 & 7.1985 & -22.9825 & 0.0268 & 7.17e+10 & 0.01305 & 1.05 & 3.06 & 2.31e+11 & 4.44e+12 & 19.2 & 113 & 0.16 + 5 & 7.3508 & -10.9007 & 0.0122 & 5.79e+10 & 0.03334 & 1.14 & 2.15 & 1.42e+11 & 2.45e+13 & 172.4 & 189 & 0.31 + 5 & 8.0602 & -3.3412 & 0.0196 & 3.70e+10 & 0.03581 & 1.15 & 2.23 & 9.51e+10 & 1.98e+13 & 208.3 & 215 & 0.2 + 4 & 8.4047 & -28.5031 & 0.0234 & 3.47e+10 & 0.02017 & 1.08 & 2.62 & 9.86e+10 & 2.45e+13 & 248.4 & 149 & 0.5 + 4 & 8.8372 & -23.7145 & 0.0128 & 1.90e+10 & 0.01679 & 1.07 & 1.80 & 3.65e+10 & 3.89e+12 & 106.5 & 99 & 0.18 + 7 & 10.6675 & -9.3056 & 0.0197 & 4.22e+10 & 0.03604 & 1.15 & 1.89 & 9.20e+10 & 6.88e+13 & 747.4 & 281 & 0.4 + 4 & 12.7327 & -13.4234 & 0.0376 & 9.85e+10 & 0.02390 & 1.10 & 7.94 & 8.60e+11 & 2.06e+13 & 24.0 & 286 & 0.12 + 4 & 14.2872 & -9.1834 & 0.0150 & 2.88e+10 & 0.08095 & 1.38 & 2.07 & 8.22e+10 & 1.35e+13 & 164.2 & 113 & 0.48 + 9 & 18.0011 & -32.3378 & 0.0192 & 1.25e+11 & 0.02365 & 1.10 & 2.18 & 3.00e+11 & 5.37e+13 & 179.0 & 346 & 0.2 + 4 & 18.0096 & -6.6967 & 0.0204 & 2.76e+10 & 0.11385 & 1.57 & 2.37 & 1.03e+11 & 1.10e+13 & 107.0 & 285 & 0.06 + 14 & 20.7478 & -35.3655 & 0.0193 & 1.23e+11 & 0.01880 & 1.08 & 2.17 & 2.88e+11 & 5.47e+13 & 190.1 & 212 & 0.56 + 5 & 20.7672 & -38.5622 & 0.0205 & 4.90e+10 & 0.01960 & 1.08 & 2.18 & 1.15e+11 & 7.60e+12 & 65.8 & 159 & 0.14 + 4 & 25.2701 & -34.4858 & 0.0128 & 1.14e+10 & 0.01997 & 1.08 & 1.57 & 1.93e+10 & 4.52e+12 & 233.7 & 190 & 0.06 + 5 & 25.3051 & -4.2930 & 0.0180 & 3.59e+10 & 0.02266 & 1.09 & 2.05 & 8.05e+10 & 1.92e+13 & 238.4 & 157 & 0.36 + 5 & 26.2856 & -13.8575 & 0.0181 & 2.74e+10 & 0.01640 & 1.07 & 2.08 & 6.09e+10 & 1.94e+13 & 318.4 & 175 & 0.29 + 4 & 28.2577 & -4.2047 & 0.0164 & 2.91e+10 & 0.02386 & 1.10 & 1.85 & 5.92e+10 & 1.25e+14 & 2110.9 & 331 & 0.52 + 4 & 28.7924 & -9.3172 & 0.0173 & 1.91e+10 & 0.02394 & 1.10 & 1.95 & 4.10e+10 & 1.08e+14 & 2635.4 & 265 & 0.7 + 4 & 29.5184 & -25.5656 & 0.0144 & 3.66e+10 & 0.01053 & 1.04 & 1.98 & 7.56e+10 & 2.34e+13 & 309.5 & 256 & 0.16 + 5 & 30.2063 & -32.1871 & 0.0183 & 4.08e+10 & 0.01977 & 1.08 & 2.33 & 1.03e+11 & 1.17e+14 & 1138.0 & 288 & 0.64 + 4 & 30.6148 & -29.3072 & 0.0169 & 3.96e+10 & 0.01656 & 1.07 & 2.15 & 9.10e+10 & 1.34e+14 & 1471.9 & 289 & 0.73 + 4 & 31.4930 & -6.8843 & 0.0169 & 1.60e+10 & 0.01961 & 1.08 & 1.83 & 3.16e+10 & 3.71e+13 & 1172.2 & 271 & 0.23 + 4 & 31.5281 & -23.1068 & 0.0178 & 5.10e+10 & 0.01674 & 1.07 & 2.12 & 1.16e+11 & 9.35e+13 & 808.6 & 316 & 0.43 + 4 & 34.8054 & -38.3532 & 0.0167 & 2.95e+10 & 0.02016 & 1.08 & 1.83 & 5.85e+10 & 4.24e+12 & 72.4 & 55 & 0.64 + 7 & 36.4329 & -11.4888 & 0.0153 & 3.37e+10 & 0.02209 & 1.09 & 1.75 & 6.44e+10 & 6.41e+12 & 99.6 & 181 & 0.09 + 9 & 39.0671 & -13.1909 & 0.0150 & 5.33e+10 & 0.03001 & 1.13 & 1.68 & 1.01e+11 & 4.99e+13 & 494.9 & 243 & 0.39 + 9 & 40.8864 & -25.7116 & 0.0229 & 1.15e+11 & 0.01903 & 1.08 & 3.60 & 4.45e+11 & 8.34e+13 & 187.4 & 219 & 0.79 + 8 & 41.4914 & -31.8348 & 0.0221 & 1.13e+11 & 0.02305 & 1.10 & 2.55 & 3.17e+11 & 1.66e+14 & 523.7 & 393 & 0.49 + 5 & 42.5306 & -8.9778 & 0.0178 & 2.59e+10 & 0.02907 & 1.12 & 2.04 & 5.93e+10 & 9.15e+13 & 1542.9 & 274 & 0.56 + 14 & 46.6818 & -10.4633 & 0.0155 & 8.66e+10 & 0.07630 & 1.36 & 1.67 & 1.96e+11 & 1.17e+14 & 597.3 & 422 & 0.3 + 5 & 47.4664 & -25.3934 & 0.0212 & 2.99e+10 & 0.01680 & 1.07 & 2.33 & 7.45e+10 & 3.07e+12 & 41.2 & 113 & 0.11 + 4 & 48.1498 & -7.5486 & 0.0177 & 3.10e+10 & 0.06842 & 1.31 & 2.15 & 8.76e+10 & 7.86e+13 & 896.8 & 203 & 0.87 + 10 & 49.0948 & -4.7276 & 0.0079 & 2.40e+10 & 0.06228 & 1.28 & 1.26 & 3.87e+10 & 8.08e+13 & 2087.3 & 387 & 0.25 + 4 & 52.7168 & -4.8493 & 0.0198 & 3.00e+10 & 0.04634 & 1.20 & 2.24 & 8.10e+10 & 8.13e+13 & 1004.3 & 235 & 0.07 + 4 & 53.8719 & -18.4927 & 0.0142 & 2.26e+10 & 0.06259 & 1.28 & 2.06 & 5.97e+10 & 1.94e+12 & 32.5 & 67 & 0.2 + 11 & 55.2961 & -4.5489 & 0.0138 & 1.01e+11 & 0.06435 & 1.29 & 1.58 & 2.07e+11 & 3.45e+13 & 166.7 & 216 & 0.34 + 5 & 149.4540 & -2.7081 & 0.0210 & 5.70e+10 & 0.04201 & 1.18 & 2.13 & 1.43e+11 & 2.49e+13 & 173.6 & 308 & 0.12 + 7 & 150.3229 & -6.2040 & 0.0164 & 4.93e+10 & 0.03796 & 1.16 & 1.78 & 1.02e+11 & 3.36e+13 & 329.0 & 249 & 0.25 + 5 & 151.6062 & -2.4202 & 0.0202 & 3.65e+10 & 0.04465 & 1.19 & 2.37 & 1.03e+11 & 2.66e+12 & 25.8 & 58 & 0.36 + 5 & 155.7194 & -2.3807 & 0.0186 & 5.69e+10 & 0.04830 & 1.21 & 1.93 & 1.33e+11 & 4.88e+13 & 366.3 & 218 & 0.47 + 6 & 160.6563 & -10.2180 & 0.0077 & 1.29e+10 & 0.00910 & 1.04 & 1.38 & 1.77e+10 & 6.25e+13 & 3522.1 & 245 & 0.48 + 4 & 162.5583 & -12.2600 & 0.0154 & 1.79e+10 & 0.03559 & 1.15 & 1.86 & 3.84e+10 & 2.77e+13 & 721.7 & 248 & 0.21 + 4 & 164.2253 & -9.5503 & 0.0270 & 6.25e+10 & 0.03783 & 1.16 & 3.15 & 2.29e+11 & 2.44e+13 & 106.6 & 191 & 0.31 + 17 & 165.7276 & -14.5042 & 0.0144 & 1.09e+11 & 0.00944 & 1.04 & 1.52 & 1.66e+11 & 2.97e+14 & 1787.6 & 475 & 0.6 + 5 & 168.0586 & -13.6485 & 0.0171 & 3.23e+10 & 0.07506 & 1.35 & 1.93 & 8.42e+10 & 1.25e+14 & 1484.7 & 351 & 0.46 + 4 & 170.6570 & -13.4858 & 0.0173 & 5.77e+10 & 0.03418 & 1.15 & 1.85 & 1.22e+11 & 7.72e+13 & 630.9 & 244 & 0.59 + [ ] crrccrccccrrr n & @xmath87 & @xmath88 & z & raw lum . & e(b - v ) & ext . & comp . & tot . lum . &
mass & m / l & @xmath89 & rvir + 4 & 171.4233 & -11.2592 & 0.0174 & 2.83e+10 & 0.04699 & 1.21 & 1.99 & 6.79e+10 & 3.97e+13 & 585.1 & 165 & 0.66 + 4 & 175.3333 & -11.9775 & 0.0174 & 4.43e+10 & 0.03384 & 1.14 & 1.97 & 9.98e+10 & 1.35e+14 & 1352.6 & 395 & 0.39 + 6 & 179.4168 & -20.3306 & 0.0222 & 4.46e+10 & 0.04524 & 1.20 & 2.31 & 1.23e+11 & 4.77e+13 & 386.6 & 272 & 0.29 + 4 & 184.8120 & -12.4777 & 0.0144 & 2.48e+10 & 0.03033 & 1.13 & 1.67 & 4.14e+10 & 1.98e+13 & 477.9 & 127 & 0.56 + 4 & 190.5120 & -20.5779 & 0.0224 & 5.86e+10 & 0.05480 & 1.24 & 2.45 & 1.79e+11 & 8.54e+13 & 478.2 & 313 & 0.4 + 4 & 191.9568 & -4.7924 & 0.0148 & 4.57e+10 & 0.01745 & 1.07 & 1.93 & 8.82e+10 & 1.28e+14 & 1450.5 & 326 & 0.54 + 5 & 192.5408 & -26.9663 & 0.0115 & 2.64e+10 & 0.07663 & 1.36 & 2.16 & 5.69e+10 & 2.71e+13 & 476.1 & 166 & 0.45 + 26 & 192.5900 & -13.3802 & 0.0148 & 2.04e+11 & 0.05606 & 1.25 & 1.47 & 3.01e+11 & 3.63e+14 & 1207.8 & 466 & 0.76 + 4 & 193.6383 & -20.2686 & 0.0229 & 4.89e+10 & 0.06569 & 1.30 & 2.66 & 1.69e+11 & 5.99e+12 & 35.5 & 223 & 0.06 + 4 & 195.6448 & -4.6870 & 0.0103 & 1.72e+10 & 0.01346 & 1.06 & 1.55 & 2.66e+10 & 2.04e+12 & 76.7 & 79 & 0.15 + 9 & 199.7299 & -16.4609 & 0.0229 & 9.04e+10 & 0.08047 & 1.38 & 2.41 & 3.00e+11 & 9.28e+13 & 309.1 & 291 & 0.5 + 5 & 200.9558 & -11.9613 & 0.0223 & 6.17e+10 & 0.03190 & 1.14 & 2.31 & 1.62e+11 & 3.86e+13 & 238.4 & 238 & 0.31 + 17 & 201.6564 & -20.3991 & 0.0183 & 1.10e+11 & 0.08540 & 1.41 & 1.88 & 2.06e+11 & 5.93e+13 & 287.5 & 242 & 0.46 + 4 & 201.9665 & -1.8782 & 0.0136 & 2.51e+10 & 0.03450 & 1.15 & 1.69 & 4.25e+10 & 4.04e+13 & 951.6 & 262 & 0.27 + 4 & 201.9803 & -21.4628 & 0.0243 & 2.69e+10 & 0.09882 & 1.48 & 2.52 & 1.00e+11 & 2.59e+12 & 25.8 & 93 & 0.14 + 11 & 202.8497 & -24.7016 & 0.0159 & 7.93e+10 & 0.06650 & 1.30 & 1.74 & 1.38e+11 & 1.39e+14 & 1007.2 & 278 & 0.82 + 4 & 229.1948 & -13.2268 & 0.0073 & 1.25e+10 & 0.03372 & 1.14 & 1.43 & 1.78e+10 & 1.56e+12 & 87.6 & 104 & 0.07 + 4 & 318.7207 & -23.0257 & 0.0269 & 6.63e+10 & 0.05581 & 1.25 & 3.3 & 2.73e+11 & 1.06e+13 & 38.8 & 228 & 0.09 + 32 & 329.3656 & -33.4588 & 0.0087 & 1.01e+11 & 0.03357 & 1.14 & 1.26 & 1.46e+11 & 2.54e+13 & 174.4 & 166 & 0.42 + 6 & 331.6321 & -27.8423 & 0.0238 & 1.49e+11 & 0.02030 & 1.08 & 2.64 & 4.28e+11 & 8.80e+12 & 20.6 & 209 & 0.09 + 5 & 332.2561 & -27.0691 & 0.0086 & 6.75e+09 & 0.02354 & 1.10 & 1.32 & 9.78e+09 & 5.11e+11 & 52.3 & 30 & 0.26 + 4 & 332.4736 & -22.8598 & 0.0178 & 4.14e+10 & 0.02640 & 1.11 & 2.5 & 1.15e+11 & 2.63e+12 & 22.9 & 52 & 0.44 + 4 & 332.6806 & -30.0202 & 0.0139 & 4.61e+10 & 0.01484 & 1.06 & 2.39 & 1.17e+11 & 6.94e+13 & 593.3 & 262 & 0.46 + 4 & 332.9872 & -27.7711 & 0.0178 & 5.87e+10 & 0.01950 & 1.08 & 2.74 & 1.74e+11 & 1.01e+13 & 58.1 & 122 & 0.31 + 7 & 333.7648 & -21.2985 & 0.0087 & 1.61e+10 & 0.02895 & 1.12 & 1.33 & 2.40e+10 & 5.09e+11 & 21.2 & 59 & 0.07 + 4 & 335.6633 & -31.4416 & 0.0285 & 6.64e+10 & 0.01282 & 1.05 & 3.9 & 2.72e+11 & 6.78e+13 & 248.9 & 295 & 0.36 + 16 & 341.0783 & -22.2565 & 0.0107 & 9.25e+10 & 0.02412 & 1.10 & 1.38 & 1.41e+11 & 2.88e+13 & 204.9 & 130 & 0.78 + 4 & 343.0948 & -34.2753 & 0.0291 & 6.90e+10 & 0.01265 & 1.05 & 3.86 & 2.80e+11 & 1.57e+12 & 5.6 & 93 & 0.08 + 14 & 356.4653 & -28.3663 & 0.0282 & 2.60e+11 & 0.01613 & 1.07 & 3.79 & 1.05e+12 & 3.56e+13 & 33.9 & 315 & 0.16 + 4 & 357.4948 & -29.5501 & 0.0291 & 5.24e+10 & 0.01603 & 1.07 & 3.97 & 2.22e+11 & 1.62e+13 & 73.1 & 143 & 0.36 + [ ] rrrr selection & slope & constant & dispersion + all ssrs2 groups & [email protected] & 294@xmath90570 & 8@xmath81 + ssrs2 groups with & [email protected] & 462@xmath90431 & 6@xmath81 + more than 7 galaxies & & & + ssrs2 groups with & [email protected] & 579@xmath90470 & 4@xmath81 + more than 8 galaxies & & & + cluster sample & [email protected] & 12@xmath90236 & - + [ ]
we demonstrated that we are able to detect similar populations of groups with our method compared to classical friend - of - friend algorithms .
the groups detected with our method have the same physical properties than those detected by merchan et al .
( 2000 ) : nearly same mass distribution ( most of our groups are in the range [ @xmath35;@xmath36 solar masses ) , same abundances up to cz@xmath128000 km / s ( and , therefore , same completeness limit : @xmath125000 km / s and abundances in agreement with press @xmath11 schechter models ) , and same 2-point correlation function ( modeled by a power law : @xmath91 with @xmath1 and @xmath2 ) .
this redshift completeness limit close to 5000 km / s is directly related to the magnitude limit of the catalog : m@xmath53 . extrapolating these limits assuming the typical depth of the new generation of redshift surveys using 10 meter class telescopes ( keck and vlt ) leads to a redshift completeness limit of [email protected] for the groups in these samples .
we found for the ssrs2 catalog a similar luminosity function for nearby group galaxies to that of zabludoff et al .
( 2000 ) : m@xmath58=([email protected])+5logh and @[email protected] .
we computed a similar mass to light ratio compared to carlberg et al .
( 2001 ) ( median value of 250 hm@xmath7/l@xmath7 in the b ssrs2 magnitude passband ) and we deduced a similar relation between the mass to light ratio and velocity dispersion ( @xmath92 ) .
this relation is qualitatively similar to that detected for rich clusters of galaxies ( adami et al .
1998 ) , but with a significantly steeper slope .
the more massive the group ( or the cluster ) , the higher the mass to light ratio , and therefore , the larger the amount of dark matter inside group ( and cluster ) .
another explanation is a significant stripping of the gas of the galaxies in rich groups as opposed to poorer groups .
finally , we detected a fundamental plane for these groups ( @xmath93 for groups with more than 8 galaxies ) but much less narrow and barely significant compared to clusters of galaxies .
we conclude that a mean relation between l , r@xmath29 and @xmath8 probably exists for groups , but this relation is much less well defined compared to clusters . | we present an automated method to detect populations of groups in galaxy redshift catalogs .
this method uses both analysis of the redshift distribution along lines of sight in fixed cells to detect elementary structures and a friend - of - friend algorithm to merge these elementary structures into physical structures .
we apply this method to the ssrs2 galaxy redshift catalog .
the groups detected with our method are similar to group catalogs detected with pure friend - of - friend algorithms .
they have similar mass distribution , similar abundance versus redshift , similar 2-point correlation function ( modeled by a power law : @xmath0 with @xmath1 and @xmath2 ) and the same redshift completeness limit , close to 5000 km / s .
if instead of ssrs2 , we use catalogs of the new generation ( deep redshift surveys obtained with 10 meters class telescopes ) , it would lead to a completeness limit of [email protected] .
we model the luminosity function for nearby galaxy groups by a schechter function with parameters m@xmath4=([email protected])+5logh and @[email protected] to compute the mass to light ratio .
the median value of the mass to light ratio is 360 hm@xmath7/l@xmath7 ( in the ssrs2 band , close to a b band magnitude ) and we deduce a relation between mass to light ratio and velocity dispersion @xmath8 ( @xmath9 ) .
the more massive the group , the higher the mass to light ratio , and therefore , the larger the amount of dark matter inside the group .
another explanation is a significant stripping of the gas of the galaxies in massive groups as opposed to low mass groups .
this extends to groups of galaxies the mild tendency already detected for rich clusters of galaxies .
finally , we detect a barely significant fundamental plane for these groups ( @xmath10 for groups with more than 8 galaxies ) but much less narrow than for clusters of galaxies .
galaxies : clusters : general ; cosmology : large - scale structure of universe |
blazars are an extreme subclass of active galactic nuclei ( agns ) , seen at small angle ( @xmath5 ) to the relativistic jet emanating from very close to the black hole @xcite .
they are characterized by strong variability in flux and polarization at almost all the frequencies in the electromagnetic spectrum and their variability is often used to probe the central engine and nature of physical processes in agns .
blazars are known to show variations on different time scales- ranging from years to months to days to hours and minutes .
the variations which occur during the course of night ( few hours ) are known as intra - night variations ( inov ) while microvariability is few tenths of magnitude change in brightness during hours or less .
some authors use these terms interchangeably @xcite .
fast variation , or microvariabilty , was first discovered in 1960s by @xcite , who found the brightness of bl lac object 3c48 change by 0.04 mag in v band in 15 minutes but their result was not taken seriously due to instrumental errors .
however , now microvariabilty has been confirmed as intrinsic nature of the agns , especially for blazars @xcite and has become a subject of intense activity as its physical mechanisms are not understood well .
several models ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) have been proposed in order to explain such fast variations . in order to constrain these models ,
long - term continuous observations with high temporal resolution ( few minutes ) are required .
the blazar s5 0716 + 71 ( pks 0716 + 714 , redshift z= 0.31 ) is one of the brightest bl lac objects which is highly variable from the radio to @xmath6-rays with very high duty cycle @xcite .
this source has been the target of a number of monitoring campaigns ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
for the first time , inov was detected in this source by @xcite . a decay in j band brightness by 0.5 mag during 2003 dec 10 - 12
was reported by @xcite .
variations on various time scales have also been reported by many authors .
@xcite reported 0.02 mag per hour variation during their 52 nights monitoring .
very recently , @xcite reported b and i bands microvaribility study of s5 0716 + 71 based on their 5-nights observations during 2003 march 5 - 9 . in their systematic statistical study , they detected variations at several time scales from days to few tens of minutes .
@xcite claim microvariability down to 6 mins time scale but their sampling time is large ( 4 to 7 mins ) . in the present communication
we report day to day and rapid variations in the optical brightness of bl lac object s5 0716 + 71 obtained in high temporal resolution observations during the 5 nights in march 2010 .
the paper is organized as follows : section 2 describes the observations and data analysis procedures while section 3 presents the results and their discussion .
conclusions from the study are presented in section 4 .
the photometric observations were carried out by using the liquid nitrogen cooled ccd - camera mounted at the f/13 cassegrain focus of the 1.2 m telescope at mt .
abu infrared observatory , gurushikhar , rajasthan , operated by the physical research laboratory ahmedabad , india .
the pixellant ccd camera has 1296x1152 square pixels each with 22 micron size and a total read out time of about 13 seconds . with a scale of 0.29 arcsec per pixel , the total field of view is about 6.5x 5.5 arcmin@xmath7 .
the ccd read - out noise is 4 electrons and the dark current is negligible when cooled . the ccd - photometric system is equipped with johnson - cousin ubvri filter set .
the source was observed in two observing slots ; during 2010 march 08 - 10 and 2010 march 19 - 20 .
all the observation nights were photometric with a seeing better than 1.6 .
several bias frames were taken every night at the beginning and the end of the observations . to construct master flats ,
we have taken large number of evening twilight sky flats in all the bands each night .
observation strategy was to take 4-frames each in b , v , r and i band and then to monitor the source in r - band for several hours .
the field of view was large enough to accommodate several standard stars in the target frame to facilitate calibration .
the exposure times were 30 secs in i , r and v bands and 60 secs in b band .
table 1 presents the details of the observations , giving date , time ( ut ) of starting observation , duration of monitoring ( hours ) and total number of observation points on the source .
the data reduction is performed using standard routines in iraf ( image reduction and analysis facility ) software . on the bias
subtracted , flat fielded images , differential aperture photometry was performed using daophot package available in iraf .
the photometry is carried out using several aperture radii , ranging from 1 to 9 times the fwhm .
the right size of the aperture is chosen keeping in mind the optimum value of s / n ratio and the prescription of @xcite to avoid spurious variations .
based on these criteria we use 4.5 arcsec as aperture radius for the target and other stars used in the differential photometry .
we have used standard stars 6 ( r=13.26 mag ) & 5(r=13.18 amg ) from @xcite , having apparent magnitudes close to that of the source to check the variability of the blazar .
such a choice of the comparison and control stars is necessary to avoid any disparity in the measured dispersions of the target - comparison and comparison - control light curves @xcite due to photon statistics .
one star is used to correct the source magnitude and the other as a control star to check the stability . to check the significance of intra - night variability
, we performed the f - test incorporated in the r statistical package .
the f - statistics is the ratio of the sample variances , or @xmath8 where @xmath9 is the variance in the blazar magnitude and @xmath10 is that in the standard stars during the whole night of observations .
for all the five nights , the f - values are more than 9 with a significance level of 0.999995 or @xmath11 .
we have also calculated the intra - night variability amplitude which is given by @xmath12 where @xmath13 are the maximum and minimum values in the light curves and @xmath14 is given as follows @xmath15 where @xmath16 is the differential magnitude of stars 6 and 5 for the i - th observation point while @xmath17 is their differential magnitudes averaged over the entire dataset , and @xmath18 is the number of observation points obtained that night in a particular band . as the errors are subtracted from the total measured variability , eq .
1 gives fairer estimate of the amplitude of variability in the source .
the results from such analysis of our data for each night are discussed in next section .
the light curves for the source were obtained by adopting above mentioned analysis procedure .
the observed magnitudes of the source in b , v , r , & i bands are calculated with respect to the standard star 5 and nightly averaged values are plotted as a function of time in mjd in fig 1 for 2010 march 8 - 10 and 19 - 20 . in fig 2 , we plot r - band light curves showing intra - night variations during individual nights .
the bottom curve in each panel shows differential light curve of the stars .
the observational uncertainties are the @xmath19 errors of the nightly differential magnitudes of the calibration star 5 and check star 6 as given in eq .
the typical @xmath19 errors for the r - band are less than 0.008 mag .
these r - band magnitudes for s5 0716 + 71 for all the nights of observation are given in table 2 in truncated form .
full data table is available from the authors .
table 3 lists the result of the f - test , giving the date of observation , f - value , standard deviation in the differential magnitudes of stars and the amplitude of variation in the source each night . on march 19 ,
we have five data points at the beginning and 24- data points at the end of the night , covering about 3.5 hrs . the f - value for this night , therefore , is obtained from these limited measurements .
the tabulated values for all the nights indicate that the source is significantly variable during present observing run , showing 100% duty cycle for variation .
similar result has been reported by other workers ( eg * ? ? ?
the f - test results are in very good agreement with the values obtained from the variability test @xcite . here
confidence level of variability is defined by the parameter c=@xmath20 , where @xmath21 is the standard deviation in the differential light curve of the source and comparison star .
the source is considered variable at 99% confidence level if @xmath22 .
our values for the variability parameter for 8 , 9 , 10 and 20 march are , 7.66 , 3.35 , 4.72 and 4.45 , respectively , confirming significant variability in the source on all the nights . to further support the genuine nature of the microvaribility reported here
, we note that the host galaxy is more than 4-magnitude fainter as compared to the source @xcite ruling out any significant effect on the source variability .
the variation rates ( mag / hour ) for various peaks appearing in the light curves of each night of observation are calculated by fitting a straight line in the rising and falling segments using least square fitting algorithm .
@xcite studied intra - night variability of s5 0716 + 71 for 52 nights and claimed a variation rate of 0.02 mag / hour along with a maximum rising rate of 0.16 mag / hour while @xcite reported equally fast variation rates of 0.1 - 0.16 mag / hour . in the following we discuss results obtained in the present study .
figure 1 shows nightly averaged b , v , r & i magnitudes as a function of time ( mjd ) for all 5-nights .
it is evident that s5 0716 + 71 brightens by 0.34 , 0.3 , 0.3 & 0.24 magnitudes in b , v , r and i bands , respectively , during 2010 march 8 - 10 . during march 19 - 20
, source decreases in brightness by 0.23 , 0.22 and 0.21 mags in b , r and i bands , respectively .
evidently , during our observations blazar s5 0716 + 71 was brightest on march 10 ( r@xmath23 mag ) and faintest on march 20 ( r@xmath24 mag ) .
the figure also gives clear indication that the source is mildly bluer when brighter and redder when fainter . during the period of eight days ( march 11 - 18 ) , when we do not have observations , s5 0716 + 71 became fainter by about 1.10 ( r - band ) and 1.49 ( b - band ) magnitudes , clearly showing increase in the amplitude of variation with the frequency . from the light curves in fig 2
, it is evident that the blazar s5 0716 + 71 is showing significant inov on almost all the nights it was monitored . here
we discuss the variability behaviour night by night . on 2010
march 8 , the source brightens up to @xmath25 13.185 mag at mjd = 55263.40 .
it then decays to @xmath25 13.335 mag at mjd = 55263.52 , a variation of 0.15 mag during 2.88 hrs ( decay rate @xmath26 mag / hour ) .
this smoothly falling curve is superposed by a flicker ( between mjd55263.44 and mjd55263.465 ) brightening the source by more than 2@xmath14 within a time scale of about 15min .
the blazar s5 0716 + 71 appears to be very active during the 2010 march 9 night ( cf fig 2 ) .
the source is in brightening phase with several rapid fluctuations modulating a smoothly varying intra - night light curve . during the first microvariability event of the night
, source fades by 0.032 mag ( @xmath27 ) with a time scale of @xmath0 15 mins .
it then brightens to its highest value ( r=12.976 mag ) at mjd=55264.456 within a time span of about 1.24 hrs ( rise rate 0.063 mag / hr ) .
the source then decays by 0.05 mag within 0.48 hr towards the end of the observations .
the statistical analysis of the light curves shows fast variation rates ( up to @xmath4 0.38 mag / hour ) for several segments on march 10 . during this whole night
, source remains in bright state with rapid fluctuations in the intensity .
we have recorded most rapid fluctuation during this night . for the most significant peak
we have calculated the rise and fall rates @xmath4 0.38 mag / hour & @xmath4 0.08 mag / hour , respectively . towards the end ( from mjd 55265.502 to mjd 55265.539 ) , s5 0716 + 71 brightens by 0.09 mag in r - band within about 53 mins .
the source also shows several microvariability events on a time scale of about 15 mins .
on march 19 , we have few observation points in the beginning and end of the night .
however , the trend shows significant variation , about 0.1 mag in 3.5 hours of duration .
the light curve for this night is not shown here .
on march 20 , source is initially stable but later ( at mjd= 55275.43 ) starts brightening , changing by 0.12 mag in 1.68 hrs ( 0.07 mag / hr ) .
source is generally faint during these two nights and microvariability , if any , is washed out in the relatively large scatter ( 0.008 mag ) .
let us now discuss our results in detail .
the present observations reveal significant variability at intra - night ( 1.24 hr to 2.88 hrs ) , inter - night ( night - to - night to 9 nights ) time scales as well as microvariability ( 15 mins or more ) or fast fluctuations .
similar behaviour for s5 0716 + 71 is reported by @xcite in their 2003 march 5 - 9 observations made in b and i bands .
as mentioned above , to avoid the spurious variations caused by the variation in the seeing and/or contamination by the thermal emission from unresolved host galaxy , we have carefully chosen aperture and comparison / control stars . in order to delineate small scale fluctuations , we have used best temporal resolution ( @xmath0 45 seconds ) reported so far .
many authors have reported inov and microvariability for this source at similar time scales ( e.g. * ? ? ? * ; * ? ? ?
* ; * ? ? ?
* ) ) but with few minutes to tens of minutes temporal resolutions .
@xcite have reported fast variations at 6 mins to 33 mins time scales with unusually large ( @xmath28 magnitude ) amplitude of variations .
they do not mention about the aperture size adopted but have used exposures ranging from 4 mins to 7 mins and have calibrated the source magnitude with the brightest star in the field .
some of these are prescriptions for spurious variations as investigated by @xcite .
the observed optical emission in blazars originates in a part of the accretion disk and the inner ( pc - scale ) regions of the jet . in the light of this
, one can discuss the possible reasons behind the variations over various time scales .
we should also keep in mind that optical variability time scales shorter than a few hours would imply emitting regions to be smaller than the schwarzschild radius for certain objects , depending upon their mass .
there are a host of models to explain extrinsic and intrinsic variability in blazars ; microlensing effect @xcite , light house effect @xcite , accretion disk models @xcite and shock - in - jet model @xcite .
so far as our observations are concerned , we notice mild chromatic behaviour ( bluer when brighter ) in the inter - night light curves .
our intra - night light curves do not show any symmetry or periodicity and the variability time scales are short , ranging from few tens of minutes to few hours .
such fast variations with amplitude of variations reported here are difficult to explain by the accretion disk models .
since the blazar emission is dominated by the jet radiation , we concentrate on relativistic jet models .
a mild chromatic behaviour in the long term variation is explained by @xcite and @xcite for bl lacertae , using the data obtained during several webt ( whole earth blazar telescope ; www.to.astro.it/blazars/webt/ ) campaigns covering the period from 1997 to 2002 , by the variation in the doppler factor due to the change in the viewing angle .
they interpreted flux variability in terms of two components ; long - term ( few days time scale ) variation component as a mildly chromatic event and a fast ( intra - day ) varying component characterizing strong bluer when brighter chromatic behaviour .
if the intrinsic source spectrum is well described by a power law , a doppler factor variation does not imply a colour change .
but a mildly chromatic behaviour could be due to doppler factor variation on a spectrum slightly deviating a power law .
a change in doppler factor changes both , the flux ( @xmath29 ) and the frequency ( @xmath30 ) of emission .
the inter - night variations reported here for s5 0716 + 71 , which show mild chromatic behaviour , could be the result of the change in doppler factor .
rapid variability can be produced when a relativistic shock wave or a blob propagates down the jet with turbulent plasma @xcite .
synchrotron emission gets enhanced when the shock encounters particle or magnetic field over - densities . the amplitude and the time scale of the variation
depend on the turbulence and shock thickness .
fast variations , thus require shocks to be very thin and emission to originate from very close to the central engine . based on our shortest intra - night variation time scale ( @xmath31 ) of 1.24 hrs , the upper limit on the size of the emission region , with doppler boosting and cosmological corrections , @xmath32 is @xmath33 cm where @xmath34 is the speed of light and @xmath35 the doppler factor ( taken as 10 here ) .
considering this size as a bound on the schwarzschild radius of the central engine , one can estimate its mass using @xmath36 , which comes out to be @xmath37 .
however , estimation of black hole mass using such fast optical variations must be taken with caution@xcite . the variations shorter than inov ( microvariations )
amounting to few tens of minutes as reported here and by many other authors can not possibly be explained by the shock - in - jet model . these are perhaps due either to small fluctuations intrinsic to the jet or imprinted by small fraction of black hole horizon @xcite .
such fast variations may not represent linear dimension of an emission region .
here we have reported intra - night variations and microvariability in blazar s5 0716 + 71 as observed in the high temporal resolution observations carried out during 2010 march 8 - 10 & 19 - 20 .
we note that source was variable with 100% duty cycle at various time scales .
inter - night behaviour of the source appears to be bluer when brighter from the limited observations .
it is evident that s5 0716 + 71 was highly active during 2010 march 9 , 10 and 20 with rapid flickers superposed on the slowly varying intra - night light curves . on march 19
it shows substantial decay but we do not have full coverage of the night to comment on the nightly variation behaviour .
the source shows various time scales for the variation , ranging from close to two hours to 15 minutes .
the inter - night variations showing mild bluer when brighter nature could be due to variations in the doppler factor . while intra - night variations with few hours time scales are probably due to interaction of fast moving shocks in the jet with local small scale inhomogeneities , it is very difficult to associate faster variations with spatial extent of the emitting region .
perhaps these variations originate in a small region of the blob . taking the shortest intra - night variability time scale of 1.24 hour , the linear size of the emitting region
is estimated to be @xmath38 cm and with corresponding schwarzschild radius , @xmath39 is mass of the black hole . , k. s. , joshi , u. c. , & ganesh , s. ( 2005 ) , _ bulletin of the astronomical society of india _ , * 33 * , 181 , m. c. , fabian , a. c. , & rees , m. j. ( 2008 ) , _ _ , * 384 * , l19 , m. , & krockenberger , m. ( 1992 ) , _ _ , * 255 * , 59 , m.t . , walters , r. , & hopper , l. ( 2011 ) , _ _ , * 141 * , 49 , s. a. , romero , g. e. , & combi , j. a. ( 2000 ) , _ _ , * 119 * , 1534 , s. k. , & wiita , p. j. ( 1993 ) , _ _ , * 411 * , 602 , k. , & refsdal , s. ( 1979 ) , _ _ , * 282 * , 561 , & subramanian , k. ( 1991 ) , _ _ , * 349 * , 766 , & wiita , p. j. ( 1992 ) , _ _ , * 259 * , 109 , a. c. , srivastava , a. k. , & wiita , p. j. ( 2009 ) , _ _ , * 690 * , 216 , j. , & wagner , s. j. ( 1996 ) , _ _ , * 305 * , 42 , s. b. , warnock , i. a. , & mitchell , k. j. ( 1988 ) , _ _ , * 95 * , 247 , d. et al.(2011 ) , _ _ , * 192 * , 12 , m. , & miller , h. r. ( 1997 ) , _ _ , * 114 * , 565 , a. v. , & wiita , p. j. ( 1993 ) , _ _ , * 406 * , 420 , a. p. ( 1996 ) , in asp conf . ser . 110 , blazar continuum variability , ed . h. r. miller , j. r. webb , & j. c. noble ( san francisco , ca : asp ) , 248 , a. p. , & gear , w. k. ( 1985 ) , _ _ , * 298 * , 114 , a. p. , gear , w. k. , & travis , j. p. ( 1992 ) , in _ variability of blazars_,ed
. e. vataoja , & m. valtonen ( cambridge univ .
press , cambridge),85 , t. a. , & sandage , a. r. ( 1963 ) , _ _ , * 138 * , 30 , h. r. , carini , m. t. , & goodrich , b. d. ( 1989 ) , _ _ , * 337 * , 627 , f. , maselli , a. , massaro , e. , nesci , r. , sclavi , s. , & maesano , m. ( 2006 ) , _ _ , * 451 * , 435 , r. , massaro , e. , & montagni , f. ( 2002 ) , _ _ , * 19 * , 143 , k. , pursimo , t. , sillanp , a. , takalo , l. o. , & lindfors , e. ( 2008 ) , _ _ , * 487 * , l29 , i.e. , villata , m. , & raiteri , m. ( 2007 ) , _ _ , * 470 * , 857 , b. , tao , j. , & fan , j. ( 2002 ) , _ _ , * 123 * , 678 , s. j. , quirrenbach , a. , witzel , a. , krichbaum , t.p . ,
hummel , c.a .
, & zensus , j. a. ( 1991 ) , _ _ , * 241 * , 15 , a. , et al . ( 1991 ) , _ _ , * 372 * , l71 , c. m. , et al . ( 2003 ) , _ _ , * 402 * , 151 , c. s. , et al . ( 2009 ) , _ _ , * 399 * , 1357 , c. m. , & padovani , p. ( 1995 ) , _ _ , * 107 * , 803 , m. , raiteri , c. m. , lanteri , l. , sobrito , g. , & cavallone , m. ( 1998 ) , _ _ , * 130 * , 305 , m. , et al . , ( 2004 ) , _ _ , * 421 * , 103 , m. , et al . , ( 2008 ) , _ _ , * 481 * , l79 , s. j. , & witzel , a. ( 1995 ) , _ _ , * 33 * , 163 , s. j. , et al .
( 1996 ) , _ _ , * 111 * , 2187 , x. , zheng , y.g .
zhang , h.j . ,
liu , y. , wen , y.b .
, & wang , w. ( 2008 ) , _ _ , * 136 * , 1846 l c c c date & @xmath40(ut ) & duration ( @xmath41 ) & no of images + 2010 march 08 & 14:53:51 & 3.2 & 248 + 2010 march 09 & 14:39:36 & 3.4 & 265 + 2010 march 10 & 14:44:44 & 4.0 & 294 + 2010 march 19 & 14:24:16 & 0.06 & 5 + & 17:24:27 & 0.30 & 24 + 2010 march 20 & 14:11:53 & 3.3 & 201 + r c c c date & mjd ( 55200 + ) & mag & @xmath14 ( mag ) + 2010 march 08 & 63.39160 & 13.211 & 0.007 + & 63.3921 & 13.215 & 0.007 + & 63.3926 & 13.211 & 0.007 + & 63.3931 & 13.208 & 0.007 + & 63.3936 & 13.204 & 0.007 + 2010
march 09 & 64.381700 & 13.054 & 0.006 + & 64.3822 & 13.041 & 0.006 + & 64.3827 & 13.048 & 0.006 + & 64.3832 & 13.049 & 0.006 + & 64.3837 & 13.045 & 0.006 + 2010 march 10 & 65.38520 & 12.906 & 0.005 + & 65.3857 & 12.911 & 0.005 + & 65.3863 & 12.914 & 0.005 + & 65.3868 & 12.921 & 0.005 + & 65.3873 & 12.920 & 0.005 + l c c c date & @xmath42 & @xmath14(mag ) & @xmath43 + 2010 march 08 & 38.82 & 0.006 & 16.8 + 2010 march 09 & 11.27 & 0.006 & 9.1 + 2010 march 10 & 20.51 & 0.005 & 9.9 + 2010 march 19 & 98.54 & 0.004 & 12.8 + 2010 march 20 & 19.73 & 0.008 & 14.4 + | we report rapid optical variability for the blazar s5 0716 + 71 during 2010 march 08 - 10 & 19 - 20 in the ccd observations made from mt .
abu infrared observatory .
the light curves are constructed for the duration longer than 3-hours each night , with very high temporal resolution(@xmath0 45 seconds in r - band ) . during 2010 march 08 source smoothly decayed by about 0.15 mag in 2.88 hours , apart from a fast flicker lasting about 30 mins .
s5 0716 + 71 brightened up during march @xmath1 and @xmath2 showing high activity while it was relatively faint ( @xmath3 14 mag in r ) albeit variable during march 19 - 20 . during march 9
& 10 , rapid flickers in the intensity modulate the long term intra - night ( @xmath4 3 hours ) variation .
the present observations suggest that the blazar s5 0716 + 71 showed night - to - night and intra - night variability at various time scales with 100% duty cycle for variation along with microvariability at significant levels . on night - to - night basis ,
the source exhibits mild bluer when brighter nature .
the interaction of shocks with local inhomogeneities in the jet appears to cause intra - night variations while microvariations could be due to small scale perturbations intrinsic to the jet . |
under nuclear medicine , positron emission tomography ( pet ) is an important imaging technique in which a radio - pharmaceutical having positron emitter is injected into the object of study .
this radioactive element can be used to collect the morphological information from the object . inside the object , these positrons annihilate with the available electrons to produce two nearly back - to - back photons each having energy of 511 kev .
the simultaneous detection of these photons with efficient detectors can lead to the identification of annihilation point . as these positron emitters are injected via some physiological substances , mapping of the density of the positron sources
can give the information about the activity inside the object of interest .
although pet gives various important information which the other imaging techniques are not able to provide , limitations due to a short field of view ( fov)(@xmath016 cm in standard pet scanners ) , long dead time of the electronics among others affect the sensitivity and the spatial resolution of the currently available pet scanners .
moreover , to obtain an image , in a reasonable examination time a high radiation dose is required to deliver within the object to collect sufficient amount of data . in a time - of - flight ( tof ) pet scanner ,
the arrival times of each coincident pair of photons are more precisely known , and the ( extremely small ) difference between them is used to localize the annihilation event along the line between the two detected photons .
this additional tof information helps to improve the image quality by reducing the length over which the coincident events are back - projected . instead of being back - projected over the entire distance through the body ,
the events are back - projected over a smaller distance , determined by the timing resolution of the scanner .
for example , for a coincidence timing resolution of 600 ps the positional uncertainty is 9 cm full width at half maximum along the line pair .
the average diameter of a typical patient ( torso ) is 25 - 30 cm , or approximately three times the tof positional uncertainty .
the ratio of patient diameter ( d ) to the positional uncertainty ( @xmath1x ) has been reported to be representative of the noise reduction , or the sensitivity gain , with tof .
this benefit in tof increases as the pet system timing resolution improves .
the detectors that are generally used worldwide for pet imaging are the scintillator - based detectors , e.g. , bismuth germanate ( bgo ) or lutetium orthosilicate ( lso ) , as they have excellent light output and very good energy resolution .
but because of its high cost and the above - said limitations , several attempts are being made to find an alternative .
multi - gap resistive plate chambers ( mrpcs ) are considered to be a good alternative to the scintillator - based pet imaging systems because of their excellent time resolution ( @xmath0 50 ps ) and position ( @xmath0 1 mm ) resolution in addition to the relatively lower cost of fabrication . in mrpc - based system
, the location of the origin of photons can be obtained in 3-dimensions .
the detector cell provides the position in transverse plane ( x - y plane ) and the time of arrival on the detector gives the distance in z - direction .
if the time resolution of the mrpc is known then an estimation of the spread for the position of origin of photon source can be obtained by the formula @xcite : @xmath2 \approx \sigma_{t}[ps]/2\ ] ] where @xmath3 is the fwhm in the position accuracy and @xmath4 is the rms time accuracy per photon .
reconstruction of events however improves the resolution therby producing the desired image granularity ( few mm ) .
several attempts have been made in building mrpc - based prototype pet imaging systems using varying number of gaps . however , like many other gas - based detector systems , the efficiency of conversion of 511 kev photons is very low .
mrpc , due to its layered structure , has a natural configuration where proper conversion materials can be placed to obtain better efficiency . however , very thick converter will likely to absorb the produced electrons before they can reach the gas volume , whereas for very thin one it is very difficult to have a good conversion efficiency .
efforts have been made to optimize the converter - mrpc combination for obtaining best possible efficiency of single 511 kev photon detection .
a maximum value up to @xmath0 20 - 25% has been reported , as mentioned in ref @xcite . in this work ,
we have performed a detailed simulation to optimize the detection efficiency of a pair of 511 kev photons likely to be generated from positron annihilation .
the work includes the response of the passage of the photons inside the material by geant4 and a monte - carlo ( mc ) based simulation package for obtaining the mrpc signals giving their timing properties due to produced electrons .
the simulation is done in two stages .
first , geant4 has been used to simulate the conversion of the incident photons .
the electrons obtained by photon conversion are considered to be the particles ionizing the gas .
the processes for ionization to signal generation have been implemented in another mc code .
we organize the article as follows . in sec .
2 , we discuss the simulation procedure , in sec .
3 we give the results obtained from the simulation and in sec .
4 we present the results of the performance of a four gap bakelite - base mrpc , operated in streamer mode .
geant4 ( version 4.9.4 ) has been used to simulate the response of photons through the detector thereby generating electrons which reach the gas volume .
geant4 @xcite is the efficient toolkit where different aspects of simulation process , such as the geometry of the system , the materials involved , the generation of secondary particles , the tracking of particles through materials , the physics processes governing particle interactions , the response of sensitive detector components , etc .
can be implemented . for gamma ray conversion ,
the data file g4emlow version 6.19 , containing cross - sections for low energy electromagnetic processes , was used . the photon - pairs are considered to be back to back lying on the same plane . for attaining good conversion efficiency , we have used a set of lead - mrpc combinations as shown schematically in fig .
[ fig:1 ] .
two such combinations are placed facing each other considering the source of photon - pair to be in between as shown in fig .
[ fig:1 ] . in this work , as an input of geant4 , we have used different materials including lead as the converter , glass as detector material and the mixture c@xmath5f@xmath6h@xmath5/i - c@xmath6h@xmath7/sf@xmath8(85/5/10 ) as the gas volume in the gas gap .
we have studied the variation of the conversion efficiency by varying the converter thickness .
schematic representation of the detector configuration , title="fig : " ] + two photons with momenta of same magnitude ( 511 kev ) but opposite directions are used as the input for every event .
any particle ( photon or electrons ) backscattered from one detector to the other are not considered for further study .
photons get converted mainly in the lead volume mostly by the compton and the photoelectric processes @xcite .
a number of low energy electrons get absorbed inside the lead volume . only the electrons reaching any one of the gas volume are considered for signal generation . increasing the number of pb - mrpc combinations
improves the conversion probability , which subsequently decreases due to the absorption of electrons .
the * conversion efficiency * of each detector is defined as the ratio of the number of the photons that generate electrons after conversion in any of the gas gaps to the number of the incident photons .
the conversion electrons resulting from geant4 are used as the input to the monte - carlo based programme , developed to study the response in the gas volume for signal generation .
when the electron passes through the gas volume it ionizes the gas , which leads to the avalanche formation . during primary
ionization the average number of clusters formed into the gas mixture is considered to be around nine/0.2 mm @xcite , which is the higher for slow electrons from converted photons than the value commonly used for minimum ionizing particles for the gas of given composition .
the avalanche development from the primary cluster within the gas volume is governed by the townsend coefficient ( @xmath9 ) and the attachment coefficient ( @xmath10 ) , which in this work are simulated by heed @xcite .
the probability to obtain _
n _ avalanche electrons at position _
x _ within the gas volume , considering the electrode to be at x=0 , is given by the relation @xcite @xmath11 where @xmath12 and @xmath13 .
to reduce the computational time in avalanche formation by a large number of secondary electrons , in our case for @xmath14 , we have used the central limit theorem ( clt ) , as mentioned in ref @xcite , to obtain the average number of electrons and their spread after avalanche at a particular position . the space charge effect is also implemented by the application of a cut off on the number of avalanche electrons , after which multiplication stops . in the present case ,
the threshold applied is 1.6@xmath1510@xmath16 .
finally the avalanche electrons induce current signals onto the mrpc pick - up strips .
the particle is said to be detected if the induced charge crosses a threshold value , which is 20 fc in our case .
the * detection efficiency * of the mrpc , in this work , is defined as the ratio of the number of particles crossing the threshold charge value to the number of converted particles entering the gas gap .
the concept of multi - gap has been implemented within the mc code by introducing a number of sub - gaps , keeping the total width of the gas volume same . during this study , we have simulated the detection efficiency and the time response of mrpc for varying number of gaps .
no effort has been made to implement the response of the readout electronics .
we have defined the * avalanche growth time * as the time taken by the avalanche , starting from time t=0 , to produce sufficient number of electrons within the gas volume so that it can induce a charge greater than 20 fc to the mrpc electrodes .
the standard deviation of the distribution of the avalanche growth time taken over a large number of particles is called the * time resolution * ( tr ) of the detector for single photon . for a pair of detectors , facing each other , as shown in fig .
[ fig:1 ] . , the difference of the time ( t@xmath17 ) taken by one photon to produce a detectable avalanche in one detector and the time ( t@xmath5 ) taken by the other photon to produce the same in another detector is defined as the * pair time difference * [ @xmath18 t = t@xmath19 t@xmath5 ] .
the standard deviation of pair time difference taken over a large number of photon pairs is defined as * pair time resolution * ( ptr ) for pair of photons .
for the case , when both the detectors are equally spaced from the point of origin of the photon pair , ptr is governed mainly by the folding of tr of two detectors .
the conversion efficiency is already defined in the previous section . fig .
. shows the conversion efficiency over the number of gas gap , where we have varied the converter thickness .
conversion efficiency as a function of number of gaps , title="fig : " ] + by increasing the number of gaps , the effective thickness increases .
it is seen clearly that with increasing number of gaps , the conversion efficiency increases .
however for larger thickness , e.g. 0.1 mm , etc . , the efficiency does not go higher than 18% thereby suggesting the converted electrons to stop inside the lead volume .
use of thinner converter in large numbers however increases the efficiency even further .
the plot indicates that an efficiency > 25% can be achieved with a configuration having 120 gaps , the converter thickness being 0.002 cm . from the plot
it can be seen that the converter having thickness 0.0008 cm is also giving an impressive conversion efficiency beyond 140 layers .
a saturation of the conversion efficiency for more than 140 gaps is observed for almost all the converter thickness studied .
it is of utmost importance to optimize the converter thickness for having a good conversion efficiency . during all through our study we have considered the optimized converter thickness as 0.004 cm .
. indicates the photon conversion efficiency with the configuration prescribed earlier for different materials each having the same thickness of 0.004 cm .
the plot clearly indicates that lead is the most suitable converter for 511 kev photons for 0.004 cm thickness of the converter layer .
conversion efficiency as a function of number of gaps , title="fig : " ] + we therefore conclude that a suitable tuning of converter material and thickness can help to achieve reasonable conversion efficiency in a multi - gap configuration . in fig .
[ fig:4 ] .
we have plotted the distribution of the signal collection time for single photon incident on the detector having 20 gas gaps as obtained from mc simulation for mrpc response .
the total number of entries in this plot indicate the total number of the particles after conversion .
the time resolution ( @xmath21 ) is found to be @xmath0 17 ps , which is in good agreement with earlier results @xcite .
time resolution of the detector , title="fig : " ] + the pair time resolution was calculated considering the start time to be zero .
. shows the distribution of the pair time difference collected from both the detectors .
the standard deviation of the plot , 24 ps in this case , has a good agreement with the fact that both the detector have the same intrinsic time resolution .
pair time resolution for both the detectors , title="fig : " ] + we have observed the effect of the number of gaps in time resolution by varying the number of the gas gaps , keeping the total width of the gas gap to be the same . it is seen that the time resolution decreases with the number of gas gaps . in this work , we have also estimated the position resolution of the detector in the prescribed configuration for an extended annihilation source . for this purpose ,
we have introduced a gaussian spread of the source position in one direction , e.g. , in z - direction keeping mean at zero and sigma of 0.5 mm .
the photon pairs are back to back on same plane .
we have then obtained the corresponding time difference measured by the pair of detectors . in fig .
[ fig:6 ] .
we have plotted the ratio of the distance of the source from the origin to the corresponding measured time difference .
the spread gives an estimation of the resolution of the position measurement .
ratio of the diatance of the point of generation for the origin to the photon detection time by the system , title="fig : " ] + in this case the time of detection of a photon by the detectors is the sum of the time taken by the photon to reach the detector from the point of its origin and the signal generation time for each detector .
it should be mentioned that the contribution from the second plays the leading role in the spread of the detection time .
the rpc technology mentioned in @xcite was used to build a four - gap bakelite - based mrpc .
bakelite plates of 30 cm @xmath15 30 cm in dimension and thickness of 1.2 mm had been used as electrodes and a uniform spacing of 0.6 mm each had been maintained by a high resistive material g-10 ( resistivity @xmath0 @xmath22-cm ) .
a mixture of ar / i - butane / freon(r134a ) in the ratio of 55/7.5/37.5 was used as the working gas .
a very thin coating of silicone has been applied on the inner surfaces of the bakelite plates to improve the surface roughness .
the detector has been operated in streamer mode as it gives larger signal size and does not need any preamplifier .
the detector has been tested with cosmic muons using the signals from three scintillators for trigger .
even though for these application , glass - based detectors operating in avalanche mode are usually used , the performance of 4-gap bakelite - based detector will provide the lower limit of resolution for a least expensive option of background filtering pet imaging .
[ fig:7 ] . shows the efficiency as a function of the high voltage .
an efficiency plateau > 95% had been obtained for an applied high voltage greater than 13.5 kv .
the distribution of the time difference between the master trigger and the signal from one mrpc strip is shown in fig .
[ fig:8 ] . from the time difference spectrum ,
the full width at half maximum ( fwhm ) and the corresponding standard deviation ( @xmath23 ) , where i and j refer to scintillators and the mrpc , were obtained by fitting a gaussian function .
@xmath23 s were similarly obtained for 3 different pairs of the scintillators , taking start signal from one and stop signal from the other .
the intrinsic time resolutions of the mrpc and the scintillators were obtained from the individual standard deviations @xmath24 , @xmath25 , which were extracted by solving the equation : @xmath26 .
efficiency as a function of applied voltage , title="fig : " ] + distribution of the time difference between the mrpc and the master trigger , title="fig : " ] + time resolution and time delay as a function of applied voltage , title="fig : " ] + the variation of fwhm and the time delay for the mrpc is shown in fig .
[ fig:9 ] . the extracted time resolution ( @xmath21 ) of the mrpc at 14 kv operating voltage for a typical run
was found to be @xmath0 870 @xmath27 0.03 ns .
mrpc - based pet sysytem is potentially an attractive alternative to the scintillator - based sysytem . in this work
we have performed detailed simulation for the detection of photon pairs for mrpc - based system . a layered structure of mrpc with thickness 0.004 cm lead converter can give conversion efficiency greater than 25% , which is higher compare to normally available values for gas detector systems .
for a 20 gap system , time resolution of photon pair can reach upto 17 ps .
we have also performed simulation for the position measurement for an extended source of 500 @xmath28 m width and found that the spread of the photon source can be measured with the addition of 3.17% extra fluctuation .
this suggests the appreciability of thse systems for measurements of photon - pair position to micron level . as an effort to obtain a least expensive pet imaging system capable of filtering the background ,
we have built a 4-gap bakelite - based mrpc giving time resolution of @xmath0 900 ps . for applications requiring relatively lower resolution , such a system can be useful .
this also provides a platform to build a full glass - based high resolution mrpc system for pet imaging .
the authors gratefully acknowledge the technical contributions from mr .
g. das and mr .
j. kumar and also acknowledge dr .
s. biswas , dr .
t. ghosh and dr .
sanjay pal for their help in detector test as well as in the simulation .
a. blanco et al . , nucl .
instr . and meth . a 508 ( 2003 ) 88 .
c. lippmann et al . , nucl .
instr . and meth . a 602 ( 2009)735 .
m. abbrescia et al . , nucl .
instr . and meth . a 593 ( 2008 ) 263 .
a. blanco et al . , nucl .
instr . and meth . a 478 ( 2002 )
a. blanco et al . , nucl .
instr . and meth . a 580 ( 2007 )
s. giani , et al . ,
gimage 4 : an object - oriented toolkit for simulation in hep , cern / lhcc 98 - 44 , 1998 , gimage 4 web page : http://cern.ch/geant4 .
see physics reference manual at the gimage 4 web page [ 1 ] under documentation .
w. blum , l. ronaldi , particle detection with drift chambers , springer , berlin , 1993 isbn3 - 540 - 56425-x .
w.riegler , et al.,nucl .
instr . and meth . a 500 ( 2003 ) 144 .
s. an , et al . , nucl .
instr . and meth . a 594 ( 2008 ) 39 .
s. biswas , et al .
instr . and meth . a 604 ( 2009 ) 310 .
s. biswas , et al .
. instr . and meth . a 602 ( 2009 ) | the low cost and high resolution gas - based multi - gap resistive plate chamber ( mrpc ) opens a new possibility to find an efficient alternative detector for time of flight ( tof ) based positron emission tomography , where the sensitivity of the system depends largely on the time resolution of the detector .
suitable converters can be used to increase the efficiency of detection of photons from annihilation . in this work ,
we perform a detailed geant4 simulation to optimize the converter thickness thereby improving the efficiency of photon conversion .
also we have developed a monte carlo based simulation of mrpc response thereby obtaining the intrinsic time resolution of the detector , making it possible to simulate the final response of mrpc - based systems for pet imaging .
the result of the cosmic ray test of a four - gap bakelite - based mrpc operating in streamer mode is discussed . |
chemical abundance of long - lived stars is a powerful tool to probe nucleosynthesis in supernovae ( sne ) .
in fact , detailed chemical compositions for very metal - poor stars in the galaxy revealed by pioneering observations ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) for more than a decade have brought a remarkable progress in the understanding of nucleosynthesis in core - collapse sne , i.e. , type ii sne ( sne ii ) which first contribute to the chemical enrichment in the universe . on the other hand , owing to a long time delay until the explosion , the nucleosynthesis products from type ia sne ( sne ia ) are unavoidably superimposed on interstellar matter ( ism ) already enriched by numerous sne ii , which prevents us from directly assessing sn ia nucleosynthesis through an analysis of the observed data of stellar abundances
. however , some circumstances will give birth to a pocket of sn ia enriched gas , from which stars inheriting their abundance pattern almost entirely from sne ia could be formed @xcite .
the most compelling relic of such an exclusive sn ia enrichment must be a very low [ mg / fe ] ( or [ o / fe ] ) since nucleosynthesis in sne ia predicts [ mg / fe]@xmath1 @xcite while individual sne ii normally give [ mg / fe]@xmath20.1 - 0.5 @xcite . @xcite discuss the origin of the globular cluster ( gc ) ngc 1718 in the large magellanic cloud ( lmc ) exhibiting [ mg / fe]=@xmath3 @xcite , and found that its ratios of fe - group elements ( cr , mn , ni ) to fe are identical to those predicted from the sn ia nucleosynthesis . besides ngc 1718
, two stars are found in the carina dwarf spheroidal ( dsph ) galaxy to have very low [ mg / fe ] ratios less than @xmath4 @xcite .
in addition to them , @xcite obtained detailed elemental abundance of the star ( _ car_-612 ) with [ mg / fe][email protected] in the carina , which is measured as [ mg / fe]=@xmath6 by @xcite .
surprisingly , _
low [ fe - group / fe ] ratios such as [ cr / fe ] , [ mn / fe ] , [ ni / fe ] = @xmath7 , @xmath8 , @xmath9 largely different from sn ia - like ratios as in ngc 1718 showing [ cr / fe ] , [ mn / fe ] , [ ni / fe ] = @xmath10 , @xmath11 , @xmath12 .
in fact , these low [ fe - group / fe ] ratios are seen in stars for [ fe / h ] @xmath13 in the lmc @xcite , which are a remarkable difference in the observed feature between the galaxy and the lmc as shown in figure 1 .
for instance , the [ cr / fe ] ratios among galactic stars broadly reside within a narrow range of @xmath14 to + 0.1 while the lmc stars have [ cr / fe ] extending below @xmath15 .
we also see such low ratios in the fornax dsph for cr and ni @xcite and for mn @xcite .
these observations lead to two hypotheses that the elemental feature of _
car_-612 represents nucleosynthesis of a subclass of sne ia , and that these sne ia preferentially drive the chemical enrichment in dwarf galaxies .
this issue is emphasized by recent results regarding the delay time distribution ( dtd ) of sne ia yielded by the studies on the sn ia rate in distant and nearby galaxies .
these studies dramatically shorten the sn ia s delay time , compared with its conventional timescale of @xmath161 gyr @xcite .
@xcite find that about 50 % of sne ia explode soon after their stellar birth , and further works reveal that the dtd is proportional to @xmath17 with its peak at around 0.1 gyr extending to @xmath16 10 gyr @xcite .
thus , the current view on sne ia is organized such that a majority of sne ia explode promptly after the bursting explosions of sne ii ( prompt sne ia ) , and the rest gradually emerge with a long interval of gyrs ( slow sne ia ) .
to aim at unraveling two distinct chemical pathways of fe - group elements seen in the galaxy and the lmc , we make two predictions : ( i ) nucleosynthesis of slow sne ia is characteristic of low [ fe - group / fe ] ratios as observed in _
car_-612 , and ( ii ) the ejecta of prompt sne ia , which synthesize fe - group elements more efficiently as already predicted in the existing sn models , can easily escape from the gravitational potential , if it is shallow , owing to an inactive cooling in the low - density ism after the bursting sn ii explosions . regarding prediction ( i )
, it is crucial to validate whether some theoretical models for sne ia give low abundances for fe - group elements , which will be discussed in 4 .
on the other hand , prediction ( ii ) implies that the total number of sne ia contributing to the chemical enrichment in dwarf galaxies should be smaller than that in the galaxy .
this is fully compatible with the implication from the observed high [ ba / fe ] ratios in the fornax dsph and the lmc , which demands about one third of the sn ia rate of the galaxy @xcite . in this letter
, we incorporate different nucleosynthesis yields from prompt and slow sne ia into the chemical evolution model , and present a unified scheme that explains the chemical feature among fe - group elements for both the lmc and the galaxy .
it is further shown to give a nice explanation for the observed low - ca feature associated with the relatively high mg abundance in the lmc .
we start with a brief review on the recent advancement in the understanding of theoretical models for sne ia ( 2 ) , followed by modeling the chemical evolutions for two galaxies ( 3 ) . in 4 , we perform one - dimensional numerical simulations for spherically symmetric white dwarfs ( wds ) accreting matter composed of c+o to investigate detailed nucleosynthesis proceeding in merging wds .
from a theoretical point of view , sne ia might be composed of two types .
one is the explosion of a wd with the mass of the chandrasekhar limit accreting matter from the normal companion star ( single degenerate scenario : * ? ? ?
* ; * ? ? ?
* ) and the other is the explosion resulting from the merger of two wds ( double degenerate scenario : * ? ? ?
* ; * ? ? ?
. the former scenario might be consistent with the similarity in light curves and spectra of sne ia in contrast to the variety of sne ii @xcite , though there is no direct observational evidence to support this model ( e.g. , * ? ? ?
the latter scenario naturally explains the observed diversity in sne ia by changing the combination of the masses of two wds @xcite .
a binary system of wds with the secondary wd less massive than 0.6 is associated with a time scale longer than 1 gyr .
if the merger ignites c in the primary wd , this may explain dim sne ia in elliptical galaxies @xcite because the mass of the primary is likely to be comparable to that of the secondary .
on the other hand , a pair of massive wds both exceeding 1 could explode as a sn ia within 0.1 gyr after their stellar birth .
furthermore , @xcite reported that they did not detect any stars to a visual magnitude limit of 26.9 in the central region of the supernova remnant snr 0509 - 67.5 in the lmc , which is classified as ia based on the spectrum of the light echo @xcite .
this observation seems to rule out single degenerate models and indicates that this particular supernova is the result of a violent merging of two wds , supporting the double degenerate scenario .
a recent three - dimensional simulation done by @xcite showed that a violent merging of two wds ( 1.1 and 0.9 ) results in a supernova whose light curve and spectrum are in agreement with those of observed sne ia .
therefore numerical simulations have shown that the double degenerate model actually works .
the question then arising is whether merging wds synthesize heavy elements with the right abundance ratios observed in the star _
car_-612 as well as in the lmc .
a novel ingredient in our chemical evolution model is the use of different nucleosynthesis yields for prompt and slow sn ia . for the prompt sne ia , we adopt fe - group - rich yields considered as a standard property of sn ia nucleosynthesis .
we use the delayed detonation model in a single - degenerate scenario from @xcite , which predicts the [ fe - group / fe ] ratios identical to those of ngc 1718 as shown in figure 1 .
the yields of this model are tabulated in table 1 and incorporated into the chemical evolution model . for the slow sne ia
, we assume a factor of @xmath165 lower yields of fe - group elements .
we assign prompt sne ia to the first 1 gyr in the adopted dtd @xmath18 with the range of @xmath19 gyr so that they make up @xmath1670% of sne ia .
the dtd is normalized so that 8% of the primary stars in binaries with the initial masses in the range of @xmath20explode as sne ia @xcite .
evolutions against [ fe / h ] for the lmc ( green lines , dashed ones for the cases with actual nucleosynthesis yields ) and the galaxy ( blue lines for the thick disk and red lines for the thin disk ) , compared with the observations . for the observed data of the lmc
, the gcs and field stars are denoted by open circles ( johnson et al .
2006 ; mucciarelli et al .
2008 , 2010 ; colucci et al . 2012 ) and filled circles ( * ? ? ?
* but the fornax dsph data by north et al .
2012 for mn / fe ) , respectively .
for the galaxy data , disk stars and halo stars are denoted by crosses @xcite and pluses @xcite , respectively .
the data of ngc 1718 and _
car_-612 are also shown by blue filled squares @xcite .
theoretical nucleosynthesis ratios in normal sne ia @xcite and the newly proposed ratios are indicated by dashed lines and dash - dotted lines , respectively . , width=264 ] we construct models of chemical evolution for the galaxy and the lmc .
the galaxy is modeled separately for the thick and thin disks , and each modeling is the same as in @xcite .
the lmc model is modified from the galaxy s in respect of the star formation rate ( sfr ) and the initial mass function ( imf ) ( see * ? ? ?
the essence of these models is briefly described below . _ the galaxy _ : first , the thick disk is formed rapidly , enriched by metal , and then the thin disk is gradually formed from the thick disk s debris gas mixed with the accreted metal - poor gas .
we set the initial value of [ fe / h ] to -1.5 which is implied from the metallicity distribution of thick disk stars .
all sne ia contribute to chemical enrichment and a salpeter imf ( @xmath21=@xmath22 ) is adopted . _
the lmc _ : star formation proceeds slowly , parameterized by a low sfr coefficient , @xmath23=0.15 gyr@xmath24 ( the mass fraction of the gas converted into stars per gyr , and note that @xmath23=0.4 , 2 gyr@xmath24 for the thin and thick disk , respectively ) .
it is assumed that only slow sne ia contribute to the chemical enrichment whereas the ejecta of prompt sne ia are all dispersed into the halo without being incorporated in the ism . in addition , we reduce the contribution of heavy elements from sne ii by adopting a steep imf ( @xmath21=@xmath25 ) , which is predicted in particular from the ba / fe evolution @xcite .
this reduction of sn ii enrichment mimics a partial removal triggered by galactic winds associated with sne ii .
figure 1 shows the evolution of [ cr / fe ] , [ mn / fe ] , and [ ni / fe ] calculated for the thick disk ( blue line ) , the thin disk ( red line ) , and the lmc ( green line ) . in the thick disk ,
the [ fe - group / fe ] ratios gradually increase owing to high fe - group yields from prompt sne ia ( dashed lines ) for the first @xmath161 gyr and then take a slight downturn due to the contribution from slow sne ia ( dashed - dotted line ) .
subsequently , the thin disk stars start to be formed from the thick disk s remaining gas with [ fe / h]@xmath16 + 0.2 .
first , [ fe / h ] decreases owing to dilution by metal - poor infalling gas .
this reverse evolution comes to an end when the chemical enrichment by star formation exceeds the effect of gas dilution and the usual evolution appears . on the other hand , in the lmc case , the [ fe - group / fe ] ratios decrease from a very low metallicity ( [ fe / h]@xmath26 ) due to the absence of chemical enrichment by prompt sne ia . in reality , however , prompt sne ia are likely to contribute to the chemical enrichment at different degrees in different places in the lmc , which results in the observed large scatter broadly ranging between the lines calculated for the lmc and the galaxy disks presented in figure 1 .
-[fe / h ] ( upper ) and [ ca / fe]-[fe / h ] ( lower ) correlations for both the lmc and the galaxy .
the symbols are the same as in figure 1 .
the green dotted line is the prediction from the model in which the fraction of our proposed sne ia is set at twice the value adopted in standard model .
, width=264 ] for the lmc , we see a puzzling feature not only in the [ cr , mn , ni / fe ] trends but also in the differential trend between [ mg / fe ] and [ ca / fe ] . though the common property for both is their low ratios in comparison with those of the galaxy , the deficiency levels of mg and ca are quite different .
[ mg / fe ] moderately decreases with most values staying above the solar ratio , while [ ca / fe ] is severely low , reaching as low as [ ca / fe]@xmath15 .
such observed features can be consistently understood in the framework of our proposed sn ia model .
since the mg yield of sn ia is completely negligible as compared with that of sn ii , slow sne ia do not function to lower the mg abundance .
in contrast , the contribution of ca from sne ia to chemical enrichment is large ( @xmath1630% in the present - day local ism ; * ? ? ? * ) .
therefore , if slow sne ia yield a low ca abundance like fe - group elements , [ ca / fe ] will become lower than [ mg / fe ] in the lmc . here
we assume that the ca yield of slow sne ia is 1/5 that of prompt sne ia , keeping the same yield for mg .
the results for both the galaxy and the lmc are shown in figure 2 .
the predicted [ ca / fe ] in the lmc stars is not sufficiently lowered as a reflection of the low contribution of ca from sne ia relative to sne ii as compared to the case of cr , mn , and ni .
then , for instance , if we change the frequency ratio of prompt / slow from 0.7/0.3 to 0.4/0.6 , the corresponding [ ca / fe ] evolution denoted by dashed line is in better agreement with the observation .
cccccc & 0.41 & @xmath27 & @xmath28 & @xmath29 & + & & @xmath30 & @xmath31 & @xmath32 & + & 0.71 & @xmath33 & @xmath34 & @xmath35 & + here we investigate if a wd merger can satisfy the nucleosynthesis conditions in slow sne ia required from the above chemical evolution model . as an initial model , an isothermal spherical wd
is constructed by numerically integrating the equation of hydrostatic equilibrium from the center for a given central density .
the used equation of state includes electrons with arbitrary degeneracy , ideal ions , and thermal radiation .
the temperature is assumed @xmath36 k to approximate the core temperature during the he - shell burning .
the wd is embedded in a stationary accretion flow with a constant mass accretion rate .
the velocity of the flow at the surface of wd is given by the free fall velocity .
the flow is truncated where the total mass becomes the specified value .
we calculate the subsequent evolution of the flow with a one - dimensional lagrangian ppm code including the energy generation due to nuclear reactions and energy loss due to neutrino emissions @xcite .
the code is a revised version used in @xcite .
the nuclear energy generation rate is evaluated by solving a nuclear reaction network including 13 @xmath37-elements with the code developed by @xcite .
the initial composition of matter is assumed to be half c and half o in this simplified nucleosynthesis calculation .
the entire flow is assumed to be spherically symmetric .
the mass coordinate is discretized with a uniform interval @xmath38 .
the time interval @xmath39 is controlled by the courant condition and the nuclear reaction timescale .
using the thermal history of each grid point from the hydrodynamical calculation , we use a detailed nuclear reaction network involving 1266 nuclides up to @xmath40kr to obtain abundances of various elements @xcite .
the initial abundances are assumed as follows .
the mass fractions of elements heavier than o are 10% of the solar .
the remaining matter is half c and half o by mass .
the reaction network equations are implicitly integrated .
the thermonuclear reaction rates in both nuclear reaction networks are taken from @xcite and @xcite . . , width=283 ] the intense accretion generates two shock waves propagating outward in the accreted matter and inward in the wd .
the shocked matter incinerates c and o up to si and s near the contact surface .
the subsequent rarefaction waves prohibits further nuclear reactions .
when the shock generated in the star eventually arrives at the center without igniting c or o along the way , it ignites these elements at the center and generates a detonation shock wave .
this shock propagates outward and incinerates a significant fraction of c and o in the star up to fe - group elements . here
, we show the results of the explosion of a 0.8 wd accreting 0.6 of a mass accretion rate of 0.07 s@xmath24 ( fig . 3 ) .
this yields 0.41 fe as shown in table 1 , which is a factor 1.5 smaller than the fe masses from prompt sne ia .
a 0.6 wd is the final product of a @xmath162 star @xcite , and its lifetime is about 1 gyr .
therefore , taking into account the time it takes for the orbital decay of merging wds due to the gravitational radiation , this explosion results in a dim sn ia with a delay time @xmath41 1 gyr
. the resulting abundance ratios of mn / fe and ni / fe in the ejecta meet the requirement from the chemical evolution model . on the other hand ,
the model shows overabundance of cr as compared to the predicted value , but cr / fe is reduced by @xmath1630 % from that of the prompt sn ia and becomes closer to the ratio predicted from the chemical evolution model .
the results of chemical evolution models adopting these actual nucleosynthesis yields are shown for the lmc case by dashed green lines in figure 1 .
the amount of ca from this model is 0.032 , which results in a ca / fe ratio comparable to that from prompt sne ia .
all these features are inherent in the spherical detonation of a wd with the mass significantly lower than the chandrasekhar limit .
in fact , ca and cr are synthesized in the outer layers of the primary wd . since the actual accretion flow forms a disk , the assumption of spherical accretion overestimates the compression in the primary star .
the resulting shallower densities might inhibit synthesis of elements heavier than s , which is realized in the accreted flow of the present spherical model at @xmath42 in terms of the enclosed mass coordinate @xmath43 ( see fig .
quantitative arguments need to wait for detailed nucleosynthesis calculations based on three - dimensional simulations like @xcite .
we have assessed the diversity of sne ia yielding different nucleosynthesis products from two theoretical approaches ; ( i ) modeling the chemical evolution of two galaxies exhibiting distinct chemical features , i.e. , the galaxy and the lmc , and ( ii ) performing nucleosynthesis calculations in the scheme of the double degenerate scenario .
accordingly , we propose a subclass of sne ia which releases low abundances of fe - group elements with a long timescale and which leaves little relics in the chemical abundances of the galaxy . the presence of dim sne ia has been already predicted as a result of the merger of two wds @xcite . in dwarf galaxies , however , it is difficult to trap the ejecta from prompt sne ia that synthesize a large amount of fe - group elements and explode in a burst .
hence , the proposed slow sne ia characterize the chemical features of a late evolution which leads to low ratios of [ cr , mn , ni / fe ] and to a large scatter in these ratios .
our scenario should be validated by further studies both observationally and theoretically , in particular by determination of detailed elemental abundances for more stars exhibiting very low [ mg / fe ] together with three - dimensional simulations for sn ia nucleosynthesis .
angulo , c. , arnould , m. , rayet , m. , et al . 1999 , nuclear physics a , 656 , 3 bekki , k. , & tsujimoto , t. 2012 , apj , in press bensby , t. , feltzing , s. , lundstrm , i. , & ilyin , i. 2005 , a&a , 433 , 185 cayrel , r. , depagne , e. , spite , m. , et al .
2004 , a&a , 416 , 1117 colella , p. , & woodward , p. r. 1984 , journal of computational physics , 54 , 174 colucci , j. e. , bernstein , r. a. , cameron , s. a. , & mcwilliam , a. 2012 , apj , 746 , 29 dobbie , p. d. , napiwotzki , r. , burleigh , m. r. , et al .
2006 , , 369 , 383 filippenko , a. v. 1989 , , 101 , 588 gilfanov , m. , & bogdn , . 2011 , american institute of physics conference series , 1379 , 17 gratton , r. g. , carretta , e. , claudi , r. , lucatello , s. , & barbieri , m. 2003 , a&a , 404 , 187 hflich , p. , mller , e. , & khokhlov , a. 1993 , , 268 , 570 iben , i. , jr . , & tutukov , a. v. 1984 , , 54 , 335 itoh , n. , hayashi , h. , nishikawa , a. , & kohyama , y. 1996 , , 102 , 411 kasen , d. 2010 , , 708 , 1025 iwamoto , k. , brachwitz , f. , nomoto , k. , kishimoto , n. , umeda , h. , hix , w. r. , & thielemann , f .- k . 1999 , apjs , 125 , 439 johnson , j. a. , ivans , i. i. , & stetson , p. b. 2006 , apj , 640 , 801 koch , a. , grebel , e. k. , gilmore , g. f. , et al .
2008 , aj , 135 , 1580 lemasle , b. , hill , v. , tolstoy , e. , et al .
2012 , a&a , 538 , a100 letarte , b. , hill , v. , tolstoy , p. , et al .
2010 , a&a , 523 , 17 mannucci , f. , della valle , m. , & panagia , n. 2006 , mnras , 370 , 773 maoz , d. , sharon , k. , & gal - yam , a. 2010 , apj , 722 , 1879 mcwilliam , a. , preston , g. w. , sneden , c. , & searle , l. 1995 , aj , 109 , 2757 mucciarelli , a. , origlia , l. , & ferraro , f. r. 2010 , apj , 717 , 277 mucciarelli , a. , carretta , e. , origlia , l. , & ferraro , f. r. 2008 , aj , 136 , 375 mller , e. 1986 , , 162 , 103 north , p. , cescutti , g. , jablonka , p. , et al .
2012 , a&a , 541 , a45 pagel , b. e. j. , & tautvaiien , g. 1995 , mnras , 276 , 505 pakmor , r. , kromer , m. , taubenberger , s. , et al .
2012 , , 747 , l10 pakmor , r. , hachinger , s. , rpke , f. k. , & hillebrandt , w. 2011 , a&a , 528 , a117 pakmor , r. , kromer , m. , rpke , f. k. , sim , s. a. , ruiter , a. j. , & hillebrandt , w. 2010 , nature , 463 , 61 pompia , l. , hill , v. , spite , m. , et al .
2008 , a&a , 480 , 379 rauscher , t. , & thielemann , f .- k .
2009 , + http://download.nucastro.org/astro/reaclib/ reddy , b. e. , tomkin , j. , lambert , d. l. , & allende prieto , c. 2003 , mnras , 340 , 304 rest , a. , matheson , t. , blondin , s. , et al .
2008 , , 680 , 1137 salaris , m. , serenelli , a. , weiss , a. , & bertolami , m. m. 2009 , apj , 692 , 1013 schaefer , b. e. , & pagnotta , a. 2012 , , 481 , 164 shigeyama , t. , nomoto , k. , yamaoka , h. , & thielemann , f .- k .
1992 , , 386 , l13 shigeyama , t. , nakamura , k. , tsujimoto , t. , & moriya , t. 2010 , american institute of physics conference series , 1279 , 415 totani , t. , morokuma , t. , oda , t. , doi , m. , & yasuda , n. 2008 , pasj , 60 , 1327 tsujimoto , t. 2011 , apj , 736 , 113 tsujimoto , t. , & bekki , k. 2012a , apj , 751 , l35 tsujimoto , t. , & bekki , k. 2012b , apj , 747 , 125 tsujimoto , t. , nomoto , k. , yoshii , y. , hashimoto , m. , yanagida , s. , & thielemann , f .- k .
1995 , mnras , 277 , 945 venn , k. a. , shetrone , m. d. , irwin , m. j. , et al .
2012 , apj , 751 , 102 webbink , r. f. 1984 , , 277 , 355 whelan , j. , & iben , i. , jr .
1973 , , 186 , 1007 yong , d. , norris , j. e. , bessell , m. s. , et al .
2012 , apj , in press yoshii , y. , tsujimoto , t. , & nomoto , k. 1996 , apj , 462 , 266 | a time delay of type ia supernova ( sn ia ) explosions hinders the imprint of their nucleosynthesis on stellar abundances .
however , some occasional cases give birth to stars that avoid enrichment of their chemical compositions by massive stars and thereby exhibit a sn ia - like elemental feature including a very low [ mg / fe ] ( @xmath0 ) .
we highlight the elemental feature of fe - group elements for two low - mg / fe objects detected in nearby galaxies , and propose the presence of a class of sne ia that yield the low abundance ratios of [ cr , mn , ni / fe ] .
our novel models of chemical evolution reveal that our proposed class of sne ia ( slow sne ia ) is associated with ones exploding on a long timescale after their stellar birth , and gives a significant impact on the chemical enrichment in the large magellanic cloud ( lmc ) . in the galaxy , on the other hand , this effect is unseen due to the overwhelming enrichment by the major class of sne ia that explode promptly ( prompt sne ia ) and eject a large amount of fe - group elements .
this nicely explains the different [ cr , mn , ni / fe ] features between the two galaxies as well as the puzzling feature seen in the lmc stars exhibiting very low ca but normal mg abundances .
furthermore , the corresponding channel of slow sn ia is exemplified by performing detailed nucleosynthesis calculations in the scheme of sne ia resulting from a 0.8 + 0.6 white dwarf merger . |
it is well understood that the highly super - eddington luminosities associated with grb are liable to put most of their energy into a baryonic wind if the energy release is in a position to drag matter outward .
one solution for this ( meszaros and rees 1994 ) is that at distances @xmath5 cm from the central burster , internal shocks in a baryonic outflow release some fraction of the bulk expansion energy by accelerating particles , which then radiate @xmath6-rays .
this probably predicts that typically 10 to 50 percent of the energy can be recovered and put back into radiation .
it also predicts that the @xmath6-rays are always accompanied by baryonic outflow along the same direction to within @xmath7 .
alternatively , it may be supposed that the energy release is originally devoid of baryons ( e.g. if the energy emerged along event horizon -threading field lines [ eichler and levinson , 1993 ] ) and the the baryon content of the grb fireball is whatever it swept up subsequently , either from the sides ( eichler and levinson 1999 , levinson and eichler 2003 ) , or from the ambient material into which the fireball expands ( e.g. meszaros and rees 1992 , lyutikov and blandford 2004 ) . if the fireball were able to sweep up ambient material without having any baryons originally , then afterglow would be `` guaranteed '' provided that the ambient medium has sufficient density .
however , if the asymptotic lorentz factor @xmath8 of the fireball were too large , then it would not pick up ambient matter ; early baryon loading or an initial baryon content , which keeps @xmath8 from getting too large , is probably necessary for the grb to have afterglow .
moreover , the poynting flux may be considerably less than the @xmath6-ray flux from the central object , and baryon loading from the side downstream of such a fireball s point of origin could reclaim some of the @xmath6-ray energy for generating afterglow if its optical depth exceeded unity .
the case can thus be made that baryon loading from the sides of such a fireball , as it exits a baryon rich surrounding , can enhance the blast efficiency until it is of order unity .
( in this paper , blast efficiency , @xmath9 refers to the isotropic equivalent kinetic energy @xmath10 divided by the isotropic equivalent @xmath6-ray energy output , @xmath11 .
the @xmath6-ray efficiency , @xmath0 , refers to the inverse of the blast efficiency , and either quantity can be greater than unity . ) there is no guarantee that early baryon loading from the sides penetrates the entire fireball ; there exists the logical possibility that one could have a @xmath6-ray bright grb with little or very weak afterglow .
previous estimates ( levinson and eichler 2003 , eichler and levinson 2004 [ el04 ] ) suggested a picture in which the penetration is only 1/3 to 1/2 of the angular distance to the center from the outside , depending on the duration .
longer bursts allow greater penetration because the penetrating baryons are freely streaming neutrons , and the larger the radius they are allowed to penetrate , the lower the lorentz factor of the fluid at the surface from which free streaming begins .
observers close to the axis of symmetry might therefore see afterglow only if the spread in the afterglow beam , which is smeared by an additional @xmath12 ( where @xmath13 is the lorentz factor of the blast at observer time t ) beyond the angle into which the baryons have penetrated , covers their line of sight .
typical numbers might be a beam opening angle of @xmath14 radians , a penetration angle of 0.03 to 0.04 radians , corresponding to @xmath15 ( where @xmath15 is the lorentz factor of the penetrating baryons ( levinson and eichler 2003 ) , and @xmath12 for typical afterglow observation times of @xmath1610 hours is about 1/30 to 1/10 .
although this is about enough to cover the entire range of viewing angles over which the grb would be seen , it is just barely so .
given the scatter in parameters so natural to astrophysical systems , we might expect to see , every now and then , a baryon - underloaded grb with little or very weak afterglow .
delayed - afterglow ( granot et al 2002 ) or briefly - delayed afterglow ( eichler 2005 ) is another possibility . in this letter
we discuss whether the observations of several dozen afterglows are consistent with the hypothesis that some afterglows have far less afterglow efficiency than the majority .
we find that they are ; there are four or five obvious outliers relative to an otherwise expected distribution of afterglow efficiencies clustering `` near '' unity ( but see below ) .
it can not of course be proved that the reason for the poor afterglow efficiency is baryon underloading .
it may be due to a lower ambient density ( e.g. fan et al and references therein ) that has the effect of spreading the afterglow over a longer timescale thus lowering the afterglow luminosity .
however , this could be resolved with sufficiently thorough observations and a sufficiently large database . in order to minimize the likelihood of indirect correlations , we first recall that afterglow efficiency is correlated with location of spectral peak @xmath3 ( lloyd - ronning and zhang 2004 [ lrz ] ) .
softer grb seem to have lower blast efficiencies ; the efficiencies scale roughly as @xmath1 with @xmath17 ( see below ) .
this , and its possible physical interpretations , are discussed in section 2 .
we then plot the afterglow efficiency corrected for this correlation against burst duration and show a ) that the data appear better organized after the correction , b ) that the majority of grb have inferred blast efficiencies of roughly 1/7 , which could possibly be identified with baryon saturation , given the uncertainties and c ) that there is no conclusive correlation with burst duration , with the present sample .
we also note that several grb are outliers to this correlation and all of them have anomalously _ high _ values for @xmath0 .
it is emphasized that the results are not meant to be convincing beyond reasonable doubt .
they are meant to show trends that we suggest should be checked with the much richer data set that swift should provide .
the significance of the trends , if real , would be some or all of the following implications : a ) most grb have blast energies that are at least somewhat lower than the @xmath6-ray energies .
previous estimates may have been influenced by the preferential underrepresentation of the @xmath6-ray energies , relative to afterglow energies , by off - beam observers .
b ) while most grb in the data set cluster around a value of @xmath0 of order a few , several have extremely large values of @xmath0 .
these could plausibly be interpreted as baryon - underloaded grb . according to freedman and waxman ( 2001 ) ,
the blast energy estimate is independent of ambient density and they can not be interpreted as grb that took place in an underdense environment if the observed x - ray frequency is above the cooling frequency , though it can be posited that the ambient density and/or magnetic field energy was anomalously low and that the cooling frequency was anomalously high .
c ) there is some indication that some of the anomalous grb with very high @xmath0 tend to be short , and could thus be attributed to a qualitatively different type of phenomenon and/or environment .
the search for afterglow from short grb that can be undertaken with swift will thus be important .
however , three of the five lasted longer than 25 seconds and have no apparent distinguishing characteristics other than a weak afterglow .
d ) various explanations for the amati et al correlation can be tested with a good enough data set .
the values of @xmath3 and @xmath18 correlate according to the relation @xmath19 ( amati et al 2002 , atteia et al 2004 ) .
two possible accounts of the @xmath20 correlation are a ) the dirty fireball model ( e.g. dermer 1999 , qin et al 1998 ) , in which baryon overloading delays transparency until photons have softened to x - ray energies and b ) off - beam viewing , in which the observed @xmath3 is lessened by kinematic effects , viz .
the reduced blue shift at the observer s viewing angle relative to that seen by an observer in the beam ( el04 ) . in the viewing angle model for the amati
et al relation proposed by ( el04 ) , the apparent total isotropic equivalent fluence is viewer angle dependent .
it is lowered by a viewing angle offset from the closest part of the beam by angle @xmath21 , approximately as @xmath22 , where @xmath23 .
this is opposed to the @xmath24 dependence that would apply to a thin pencil beam because the solid angle that makes a significant contribution to what is detected by observers just outside the beam is roughly proportional to the factor @xmath25 .
earlier discussions of off - beam viewing ( yamazaki et al 2002 , 2004 ) predict a different @xmath26 relation and attribute the amati et al relation to an unspecified intrinsic correlation .
a second distinction to be noted in the context of this paper is that here we are considering the possibility of baryon poor lines of sight which nevertheless emit @xmath6-rays .
the apparent afterglow fluence is also be reduced by off - beam viewing , but generally not as much .
freedman and waxman ( 2001 ) noted that x - ray afterglow fluence at @xmath2710 hours could be used as a calorimeter for the blast energy .
the lorentz factor after 10 hours , the typical time for bepposax measurements of afterglow , @xmath28 , is expected to be about a factor of 10 less ( if the expansion is into a uniform medium ) than the lorentz factor at 100 seconds , @xmath29 .
( the subscript p is for `` prompt '' , which refers to @xmath30s . )
hence the reduction in prompt fluence _ relative _ to the fluence of x - ray afterglow is given by @xmath31 ^ 2/\left[d(\theta,\gamma_x)/d(0,\gamma_x)\right]^2 \label{effa}\ ] ] by hypothesis that @xmath3 is established by viewing angle effects , @xmath32 after using ( 2 ) to eliminate the viewing angle in favor in @xmath3 , equation ( 1 ) becomes : @xmath33 ^ 2\ ] ] over the range of viewing angles @xmath34 , @xmath35 ^ 2 % \label{epsa}\ ] ] thus the viewing angle explanation for the amati et al relation predicts that the apparent ratio of gamma ray energy to blast energy @xmath36 should decrease as @xmath37 decreases , as described by equation ( 4 )
. weak correlation in the intrinsic @xmath3 with the opening angle ( see fig . 6 below ) and the fact that the beam probably does not have a sharp edge could cause the correlation to deviate somewhat from equation ( 4 ) .
so might other indirect correlations .
also , pole to equator energy transfer , a true physical effect ( kumar & granot 2003 ) , may play some role . in any event
, we expect the qualitative correlation to survive these considerations .
we have plotted @xmath6-ray efficiency @xmath0 against @xmath37 , defining efficiency as @xmath38 where @xmath10 is the isotropic equivalent kinetic energy of the grb ejecta .
we have used the @xmath10 values as presented in lrz based on x - ray afterglow luminosity in berger et al ( 2003 ) we add more data to the plot in figure 7 of lrz , using the correlation between x - ray afterglow at ten hours and blast energy to estimate @xmath10 for grb 980326 , grb 980329 and grb 000214 .
the efficiency grb 990506 has been taken from freedman & waxman ( 2001 ) .
the data for grb 980329 have a large uncertainty as the redshift is uncertain ( @xmath39 ) .
we have used the estimates by ghirlanda et al for @xmath40 and @xmath41 . using equation ( 1 ) in berger et al ( 2003 ) , and the relation between @xmath42 and @xmath43 of lrz , . ]
this would give grb 980329 an isotropic blast kinetic energy value of @xmath44 . the radiative efficiency we calculate differs from that as calculated by freedman & waxman ( 2001 ) who used an assumed value for the redshift for their calculations .
for all grbs , we have used values for @xmath45 and @xmath41 from ghirlanda et al ( 2004 ) , bloom et al ( 2003 ) , and the _ hete _ webpage . we also add grb 020124 whose blast energy we estimate below .
finally , the discovery of the host galaxy of grb 040924 has allowed the redshift to be measured at @xmath46 ( starling et al : gcn grb observation report ) as well as @xmath47 .
we have taken @xmath3 from the _ hete _ webpage .
below we estimate the kinetic energy of the grb 040924 .
+ + _ grb 040924 : _ the afterglow of grb 040924 is reproduced in fan et al ( 2004 ) , and can be extrapolated for @xmath48 , using @xmath49 ( sari et al . , 1999 ) . in this case @xmath50 , and @xmath51 ( fan et al 2004 ) . the @xmath10 and its uncertainty
have been calculated using the equations of adiabatic afterglow evolution ( sari et al . , 1999 ) as arranged below , with @xmath52 : @xmath53 @xmath54 where @xmath55 is in hz and @xmath56 is in @xmath57jy .
this allows upper and lower bounds to @xmath10 to be calculated , with reasonable ranges assumed for the unknown parameters : @xmath58 $ ] , @xmath59 $ ] and @xmath60 ( cm^{-3})$ ] .
+ + these equations are plotted below in fig.1 to show the bounds imposed on @xmath10 as a function of @xmath61 .
we include in this plot , the lines corresponding to equation ( 6 ) above for the where the circumburst density @xmath62 .
the most likely value for @xmath43 has been taken as the center of the polygon bounded by the equations and the line @xmath63 .
we add the extreme cases as well as the limits on @xmath64 that result from assuming @xmath65 , for comparison in our results .
we find the best value of @xmath43 to be @xmath66 with extremes at @xmath67 and a range for the case @xmath68 of @xmath69 .
+ + _ _ grb 020124:__by the same method employed for grb 040924 , we have used the afterglow lightcurve produced in berger et al ( 2002 ) to calculate the kinetic energy of grb 020124 over the same ranges for the unknown parameters , using @xmath70 .
we found that in the limits of uncertainty @xmath71 , and for the case @xmath65 , @xmath72 , and the most likely blast energy to be : @xmath73 .
in fig.2 , the radiative efficiency of each grb is plotted as a function of the square of the spectral peak .
there are 17 grbs closely correlated along the dashed - dotted line of best fit , with the remaining 5 outliers being xrf 020903 ( eight standard deviations @xmath74 above the line ) , our estimate for grb 040924 @xmath75 , grb 011211 @xmath76 , grb 990705 @xmath77 , and grb 980329 @xmath78 , where the standard deviation @xmath79 is computed for log@xmath80 relative to the corresponding value given by the linear fit ; @xmath81 for the group of 17 well correlated grbs .
the existence of these bursts with exceptional radiative efficiency , roughly an order of magnitude above the majority for a given @xmath3 , provides evidence of a distinct subclass of grbs .
the slope of the line of best fit in this plane reveals a correlation whereby @xmath82 .
+ + _ radiative efficiency , spectral peak and @xmath83 : _ we include a graph ( fig.3 ) of @xmath6-ray efficiency corrected for the correlation noted above , as a function of @xmath83 , taking @xmath83 from ghirlanda et al ( 2004 ) or the _ hete _ webpage . in this plane , there appears to be a general population of well correlated grbs and several scattered outliers .
the majority of bursts appear to settle on a roughly constant value @xmath84 , with a scatter of less than one order of magnitude .
contrast this with fig 4 .
, where the naive efficiency ( i.e. uncorrected for @xmath85 ) has either a larger scatter or some dependence on @xmath83 . the outliers xrf 020903 and grb 040924 are rather short bursts , though still at least @xmath86s . however , grb 990705 , grb 980329 and grb 011211 all lasted at least 25s
. + _ viewing angle calculations and radiative efficiency : _ if we can assume that viewing angle on the jet is the _ only _ factor which reduces an otherwise standard radiative efficiency and standard spectral peak of grbs , then we can compare equation ( 3 ) with the plot we have made in fig .
2 , normalising both the efficiency and @xmath3 to be unity for observers where @xmath87 .
the question of whether short bursts have afterglows is a longstanding one .
if they result from neutron star coalescence ( goodman 1986 , paczynski 1986 , eichler et al 1989 , paczynski 1990 ) , then they might take place in regions of low ambient density ( fan et al 2004 and references therein ) , which would weaken and prolong their afterglow .
here we have called attention to several weak afterglows whose grbs were not so short , such as grbs 011211 , 990705 , and 980329 .
we know of no _ a priori _ particular reason for them to have had weak afterglows and suggest that they may have been the occasional bursts that we view along baryon - poor lines of sight . with mounting evidence that grbs may be divided into subclasses which shed light on jet structure or the cause of grbs ( see bloom et al 2003 ) ,
our hypothesis would add yet another distinction between populations of grbs .
relying on previous work , we find that for the grbs with available data , 5 of 22 appear unassociated with what is otherwise a closely clustered population in blast efficiencies .
all of them have a weak afterglow , ( high radiative efficiency ) , whereas none were particularly deviant in the opposite direction .
we have speculated on possible trends in this small sample , but stress our anticipation of future data .
the majority of burst efficiencies follow @xmath88 .
ghirlanda et al ( 2004 ) find a correlation between collimation corrected @xmath89 and @xmath3 whereby @xmath90 .
( the slightly shallower than @xmath37 dependence found by ghirlanda et al . could be attributed at least in part to a removal in the intrinsic scatter of @xmath18 , if wider beams tend to have lower @xmath3 . from fig .
6 , it appears as though they do but we do not yet know the reason .
perhaps a wider beam allows more opportunity for off - beam viewing . ) in any case , there may be additional causes for scatter in @xmath3 such as a dirty fireball effect , which would lower both @xmath3 and @xmath89 but not @xmath10 .
however , it would do so at the cost of imposing an extremely small radiative efficiency . curiously , we find that @xmath6-ray efficiency has nearly the same dependence on @xmath3 as does @xmath89 in the correlation noted by ghirlanda et al .
this could be attributed to the simple fact that apparent @xmath6-ray luminosity depends more on viewing angle than does apparent afterglow luminosity .
it is anticipated that swift data will reveal whether there are more grbs distinct from the majority by their high radiative efficiency .
the sample we have used is not free of all bias ; on the other hand , the sample of available redshifts may itself suffer a possible selection bias against weak afterglows , so the eventual fraction of weak afterglow grb may be considerably different from the 5 of 22 portrayed here .
the significance of this result , if valid , is that the blast energy as a fraction of the total is only about 1/7 , and that instances where it is greater can be largely attributed to viewing angle dependent reduction of the apparent radiative efficiency .
theoretical estimates for the dissipation efficiency of the internal shocks vary ( kumar 1999 , guetta , spada , waxman , 2001 , beloborodov , 2000 , kobayashi , sari , 2001 ) , but in principle this efficiency can be large . given the uncertainties and possible systematic errors , both in afterglow observations and in the theory , the value of 7 for the ratio of isotropic equivalent @xmath6-ray energy to isotropic equivalent kinetic energy could be interpreted as a not implausible value for a baryon - saturated outflow .
however , it may be uncomfortably large for the scenario in which internal shocks in a baryonic outflow convert kinetic energy to @xmath6-ray energy .
+ the authors acknowledge a center of excellence grant from the israel science foundation and the grant from from the israel - u.s . binational science foundation and support from the arnow chair of theoretical astrophysics .
99 amati , l. , et al 2002 a & a , 390 , 81 atteia , j - l .
, et al 2004 aipc , 727 , 37a beloborodov , a.m. 2000 , apj , 539 , l25 berger , e. , kulkarni , s.r . , and frail , d.a . , 2003 , apj .
590 , 379 berger , e. , et al , 2002 , apj .
581 , 981 bloom , j.s . , frail , d.a . , and kulkarni , s.r .
, 2003 , apj .
594 , 674 dermer , c. 1999 _ astron .
suppl ser _ , 138 , 519 eichler , d. , et al 1989 , nature , 340 , 126 eichler , d. , & levinson , a. 1999 apj , 521 , l117 eichler , d. & levinson , a. , 2003 apj , 594 , l119 eichler , d. & levinson , a. , 2004 apj , 614 , l13 fan , y.z . , zhang , b. , kobayashi , s. , and meszaros , p. , 2004 american astronomical society meeting 205 , 160.07 freedman , d.l . ,
& waxman , e. , 2001 apj , 547 , 922 ghirlanda , g. , ghisellini , g. , and lazzati , d. , 2002 , apj , 616 , 331 goodman , j. , 1986 , apj , 308 , l47 granot , j. , panaitescu , a. , kumar , p. and woosley , s. ( 2002 ) , apj , 570 , l61 guetta , d. , spada , m. , & waxman , e. 2001 , apj , 557 , 399 katz , j. , & canel , l. , 1996 , apj , 471 , 915 kobayashi , s. , & sari , r .
2001 , apj , 551 , 934 kumar , p. , 1999 , apj , 523 , l113 kumar , p. , & granot , j. , 2003 , apj , 591 , 1075 lloyd - ronning , n. , and zhang . ,
b , 2004 , apj , 613 , 477 lyutikov , m. , & blandford , r. , 2004 astro - ph/0312347 meszaros , p. , & rees , m. , 1992 , apj , 397 , 570 meszaros & rees 1994 _ mon not astron soc _ , 269 , l41 narayan , r. , paczynski , d. , & piran , t. , 1992 , apj , 395 , l83 paczynski , b. , 1986 , apj ( letters ) , 308 , l43 paczynski , b. , 1990 , apj , 348 , 485 qin , b. et al 1998 apj 494 , l57 rees & meszaros 1992 _ mon .
not r. astr soc _ , 258 , 41p sari , r. , piran , t. , and narayan , r. , 1998 , apj , 497 , l17 yamazaki , r. , ioka , k. , & nakamura , t. 2004 , apj , 606 , l33 yamazaki , r. , ioka , k. , & nakamura , t. 2002 , apj , 571 , l31 zand , j. , et al , 2000 , apj , 545 , 266 http://space.mit.edu/hete/bursts/grb040924/ frontera , f. et al , gcn grb observation report 1215 , http://gcn.gsfc.nasa.gov/other/011211.gcn3 starling , r. et al , gcn grb observation report 2800 , http://gcn.gsfc.nasa.gov/gcn/other/040924.gcn3 | using data mostly assembled by previous authors , we consider the linear correlation between the apparent radiative efficiency @xmath0 ( defined as the ratio of isotropic equivalent radiative output to inferred isotropic equivalent kinetic energy of the blast ) and @xmath1 where @xmath2 , for 17 of 22 grbs ( lloyd - ronning and zhang , 2004 ) .
we note in a quantitative manner that this is consistent with the hypothesis that @xmath0 and @xmath3 are influenced by viewing angle .
we suggest a more general theoretically derived expression for this correlation that could be tested with a richer data set . if the reduction in both @xmath0 and @xmath3 is due to viewing angle effects , then the actual radiative efficiency is @xmath4 .
we also find preliminary evidence ( with a small sample ) for a separate class of weak grb afterglows . |
when electron collides elastically with open - shell atom or molecule , they can exchange their spins .
thus , spin polarization of the electron beam is in general reduced after scattering with unpolarized open - shell targets .
we can obtain more precise information of the scattering process by studying this depolarization @xcite , which is difficult to observe in usual experiment with spin - averaging procedure .
collisions of spin - polarized electrons with atoms have been studied for long years ( see hegemann et al.@xcite and references therein ) .
in contrast , number of experiment on electron molecule system is limited .
ratliff et al.@xcite measured rate constants for electron exchange in elastic electron collisions with o@xmath0 @xmath3 and no @xmath4 molecules in thermal energies .
their spin - exchange rate constants are substantially smaller than those in electron hydrogen - atom or alkali - metal - atom collisions .
hegemann et al.@xcite studied exchange process in elastic electron collisions with o@xmath0 @xmath3 and no @xmath4 molecules and na @xmath5 atoms .
they measured ratio of spin - polarizations in electron beams before and after collisions , i.e. , polarization fraction , which is directly related to the spin - exchange differential cross sections .
as in the work of ratliff et al.@xcite , hegemann et al.@xcite confirmed that the exchange cross sections of electron molecule collisions are much smaller than those of electron collisions with atoms .
although absolute value of spin - exchange cross section is small , the degree of spin - exchange in electron o@xmath0 collisions becomes larger at 100 degrees with collision energies between 8 and 15 ev compared to the other angles and energies , which they attributed to the existence of the o@xmath6 @xmath2 resonance .
theoretical study of spin - exchange in electron o@xmath0 collisions was performed by da paixao et al.@xcite .
they used the schwinger multichannel method with the three lowest electronic states of o@xmath0 in their model , and confirmed that spin - exchange cross section is small in electron o@xmath0 @xmath3 elastic scatterings .
although exchange cross sections are small for electron collisions with randomly oriented o@xmath0 , they observed large depolarization at some scattering angles when electrons were scattered from spatially oriented o@xmath0 molecules .
the profile of depolazation as a function of scattering angle depends strongly on molecular orientation .
based on these resuls , da paixao et al.@xcite explained for the first time that the experimental exchange cross section in electron - molecule collisions is small because averaging over molecular orientation washes out depolarization effects .
fullerton et al.@xcite , nordbeck et al.@xcite and wste et al.@xcite used the r - matrix method to calculate polarization fractions in electron o@xmath0 collisions .
the calculations of fullerton et al.@xcite and nordbeck et al.@xcite employed the fixed bond approximation with t - matrix elements obtained by the nine - state r - matrix calculation of noble and burke @xcite , whereas wste et al.@xcite used vibrational averaging of t - matrices to include the effect of nuclear motion .
the fixed bond r - matrix calculations of fullerton et al.@xcite and nordbeck et al.@xcite confirmed the results of da paixao et al.@xcite .
agreement with experimental results at energies from 10 to 15 ev is marginal , as in the calculation of da paixao et al.@xcite .
the vibrational averaging procedure of wste et al.@xcite improved agreement with the experimental results in this energy region .
machado et al.@xcite applied the schwinger variational iterative method combined with the distorted - wave approximation and obtained similar elastic e - o@xmath0 polarization fractions to those of fullerton et al.@xcite .
other than electron o@xmath0 collisions , theoretical work of spin exchange in electron molecule collisions is scarce .
da paixao et al.@xcite calculated polarization fractions in electron no @xmath4 elastic collisions as they did in electron o@xmath0 collisions @xcite .
calculated exchange effect was small in e - no elastic collisions , in agreement with the experimental results of hegemann et al.@xcite .
sartori et al.@xcite studied spin - exchange in the superelastic electron collisions with h@xmath0 c @xmath7 state using the schwinger multichannel method . large depolarization was observed in their results , intermediate between depolarizations in e - na and e - o@xmath0 collisions .
recently , fujimoto et al.@xcite performed the iterative schwinger variational calculation of spin - exchange effect in elastic electron c@xmath0o @xmath8 collisions .
they found modest depolarization near resonances , however , spin - exchange effect was very small in other energy region .
recently , we have studied electron o@xmath0 scatterings by the r - matrix method with improved molecular orbitals and increased number of target states@xcite compared to the previous theoretical studies .
our results are in good agreement with the previous experimental results .
since the previous theoretical polarization fractions of elastic e - o@xmath0 collisions at energies between 10 and 15ev agree not so well with the experimental results , it would be interesting to examine how spin - exchange cross section will change in this energy region by our improved calculational parameters . at the same time
, it is important to understand general behaviour of spin - exchange cross sections , polarization fractions in other words , in elastic electron molecule collisions . until now ,
spin - exchange effect in low - energy electron molecule elastic scattering has been studied only for no and c@xmath0o molecules other than o@xmath0 .
thus it is desirable to study spin - exchange in other electron - molecule scattering systems as well . in this work ,
we study spin - exchange in electron o@xmath0 collisions with the same calculational parameters as we used in our previous works@xcite .
in addition , we calculate spin - exchange cross section in elastic electron collisions with b@xmath0 , s@xmath0 and si@xmath0 molecules .
these b@xmath0 , s@xmath0 and si@xmath0 are stable homo - nuclear molecules with @xmath9 symmetry in their ground states , as in o@xmath0 molecule . in this paper , details of the calculations are presented in section 2 , and we discuss the results in section 3 comparing our results with previous theoretical and available experiments . then the summary is given in section 4 .
in this work , we consider elastic scattering of spin - polarized electrons from randomly oriented unpolarized molecules . when the incident electrons have spin polarization @xmath10 and the scattered electrons have polarization @xmath11 , with the polarization direction perpendicular to the scattering plane , the polarization fraction , the ratio of @xmath10 and @xmath11 , is a measure of spin - exchange and is related to the spin - flip differential cross section ( dcs ) @xmath12 as @xcite , @xmath13 here @xmath14 is the dcs obtained by unpolarized electrons .
the dcss of @xmath14 and @xmath12 are evaluated by the spin - specific scattering amplitude @xcite , @xmath15 where @xmath16 and @xmath17 specify the states of the target molecule as well as the scattering electron in the initial and final channels , respectively . @xmath18 and @xmath19 are the initial and final wavenumber of the electron , @xmath20 is the rotation matrix with the euler angles @xmath21 representing orientation of the target molecule in the laboratory frame .
the electron is scattered to the direction @xmath22 in the laboratory frame in this expression .
the t - matrix elements @xmath23 are prepared for all possible spin @xmath24 of the electron - molecule system as well as all irreducible representation @xmath25 of the symmetry of the system .
we used @xmath26 in the r - matrix calculations .
since the target molecules have triplet spin symmetry in this work , we only include @xmath27 and @xmath28 in our calculations .
the matrix element @xmath29 relates the spherical harmonics @xmath30 to the real spherical harmonics @xmath31 .
the explicit expression of @xmath29 can be found in our previous paper @xcite . in this paper
, we consider elastic scattering of electron from molecule with triplet spin symmetry .
then , @xmath14 and @xmath12 are expressed by the spin specific amplitude @xmath32 as @xcite , @xmath33 and @xmath34 here summations over channel indices are omitted for notational simplicity .
note that the expression of the spin - flip dcs contains interference of amplitudes with different spin multiplicities .
since the target molecules are randomly oriented , these dcss are averaged over all possible molecular orientations in space . the t - matrix elements @xmath23 were obtained by a modified version of the polyatomic programs in the uk molecular r - matrix codes @xcite .
general procedure of calculation is almost the same as in our previous works@xcite .
since the r - matrix method itself has been described extensively in the literature @xcite and references therein , we do not repeat general explanation of the method here . in this work ,
elastic electron collisions with o@xmath0 , b@xmath0 , s@xmath0 and si@xmath0 molecules were studied . for electron o@xmath0 scattering
, we used the same parameter set as we used in the previous works@xcite .
specifically , we employed the equilibrium bond length of 2.300 a@xmath35 for o@xmath0 , the r matrix radius of 10 a@xmath35 .
the angular quantum number of the scattering electron was included up to @xmath36=5 .
the atomic basis set for bound molecular orbitals , number of the target states included in the model as well as choice of the configurations in the inner region calculation were the same .
for the electron b@xmath0 , s@xmath0 and si@xmath0 scatterings , we included 14 , 13 and 15 target electronic states in the r - matrix calculation , respectively .
symmetries and spin - multiplicities of these states are given in table [ tab1 ] .
these target states were represented by valence configuration interaction wave functions constructed by the state averaged complete active space scf ( sa - casscf ) orbitals .
fixed - bond approximation was employed with internuclear distances of 3.036 , 3.700 and 4.400 a@xmath35 for b@xmath0 , s@xmath0 and si@xmath0 , respectively .
although we study only elastic scattering in this work , we included these excited target states to improve quality of the r - matrix calculations .
also , by including these excited states , we can suppress artificial structure coming from pseudo - resonance . in this study ,
the sa - casscf orbitals were obtained by calculations with molpro suites of programs @xcite .
the target orbitals of b@xmath0 , s@xmath0 and si@xmath0 were constructed from the cc - pvtz basis set@xcite .
the radius of the r - matrix sphere was chosen to be 13 a@xmath35 , which is larger than the r - matrix sphere used in the electron o@xmath0 calculation .
we need this extended r - matrix sphere to avoid overlap of b@xmath0 , s@xmath0 and si@xmath0 molecular orbitals with the r - matrix boundary . in order to represent the scattering electron , we included diffuse gaussian functions up to @xmath36 = 4 , with 13 functions for @xmath36 = 0 , 11 functions for @xmath36 = 1 , 10 functions for @xmath36 = 2 , 8 functions for @xmath36 = 3 , 6 functions for @xmath36 = 4 .
exponents of these diffuse gaussians were taken from faure et al . @xcite .
the construction of the configuration state functions ( csfs ) for the electron - molecule system is the same as in our previous e - o@xmath0 papers @xcite .
two different kind of ( @xmath37)-electron configurations are included , where @xmath38 is a number of electrons in the target molecule .
the first type of the ( @xmath37)-electron csfs is constructed from @xmath38 target molecular orbitals ( mos ) plus one continuum orbital . the second type of csfs is constructed from the @xmath37 target mos .
these target mos are just the sa - casscf orbitals , whereas the continuum orbitals are obtained by orthogonalization of the diffuse gaussian functions to the target mos @xcite .
since only the continuum orbitals have overlap with the r - matrix sphere , the first type of csfs mainly contributes the cross sections .
however , the second type of csfs is also important , as it is crucial to describe resonance . for reference
, we show orbital set used in the e - b@xmath0 calculation in table [ tab2 ] .
the orbital sets for e - s@xmath0 and e - si@xmath0 scatterings are very similar .
more detailed explanation can be found in our previous paper@xcite .
in this section , we show excited state energies of b@xmath0 , s@xmath0 and si@xmath0 molecules . since o@xmath0 energies have been shown in our previous paper@xcite , we do not discuss them here . in table
[ tab3 ] , calculated excitation energies of b@xmath0 molecule are compared with full configuration interaction ( fci ) vertical excitation energies of hald et al.@xcite . although they employed different basis set , aug - cc - pvdz , and shorter internuclear distance of 3.005a@xmath35 , our casscf values agree reasonably well with their results . in table
[ tab4 ] , our casscf energies of s@xmath0 molecule are compared with mrd ci vertical excitation energies of hess et al.@xcite and mrci adiabatic excitation energies of kiljunen et al.@xcite . in this case ,
our results agree well with the previous calculations for the lowest two excitations . for excitation energies to the three higher states ,
deviations become larger because kiljunen et al.@xcite studied adiabatic excitation energies whereas we calculated vertical excitation energies . in table
[ tab5 ] , calculated energies of si@xmath0 molecule are compared with mrd ci vertical excitation energies of peyerimhoff and buenker @xcite . since they employed shorter internuclear distance of 4.3 a@xmath35 compared to 4.4 a@xmath35 of our calculation , precise comparison is difficult .
however , our casscf results agree reasonably well with their results . in figure [ fig1 ] ( a ) , integral cross sections ( icss ) for elastic electron collision with o@xmath0 molecules are shown .
the sharp peak around 0.2 ev comes from the o@xmath6 @xmath39 resonance .
also , @xmath2 resonance causes a small rise of cross section around 13ev .
the details of these e - o@xmath0 icss were discussed in the previous paper@xcite , however , they are shown here for comparison with the icss of the electron b@xmath0 , s@xmath0 and si@xmath0 collisions . in figure [ fig1 ] ( b ) , elastic icss for electron b@xmath0 collisions are shown . in this case , very large cross section is observed near zero energy , about @xmath40 , compared to @xmath41 in the e - o@xmath0 collisions .
the partial cross sections of @xmath42 and @xmath43 symmetries equally contribute to this enhancement .
there is a broad peak around 3 ev , which comes from @xmath44 symmetry partial cross section .
an analysis of configuration state functions ( csfs ) suggests that this peak is related to the configuration @xmath45 , which is the ground state b@xmath0 with a scattering electron attached to the @xmath46 orbital . by removing this @xmath47 configuration from the r - matrix calculation , this peak vanishes from the icss . in figure [ fig2 ] ( a ) , icss for elastic electron scattering with s@xmath0 molecules
are shown .
the magnitude of the ics increases from @xmath48 at zero energy to @xmath49 at 10 ev , then it decreases to @xmath50 at 15 ev .
although the magnitudes are different , the profiles of the @xmath42 and @xmath43 symmetry partial cross sections are very similar to those partial cross sections in the e - o@xmath0 elastic collision .
a broad peak is observed around 4.5 ev , which comes from the @xmath51 symmetry partial cross section .
the csf analysis suggests that this peak is related to the configuration @xmath52 , and likely belongs to the s@xmath6 @xmath51 resonance .
the @xmath53 symmetry partial cross section has also a small rise around 6 ev ( not shown in the figure ) , its contribution to the total ics is small .
two anomalous structures are observed in the ics , a kink at 2.7 ev and a cusp at 5 ev .
the former kink belongs to the @xmath54 partial cross section , whereas the cusp at 5 ev comes from the @xmath51 symmetry .
we analyzed the csfs and found that the kink at 2.7 ev is likely related to a resonance with configuration @xmath55 , which is obtained from an attachment of the scattering electron to the excited @xmath56 , @xmath57 and @xmath58 states of s@xmath0 with configuration @xmath59 .
the position of the cusp coincides with the s@xmath0 b@xmath60 state , thus it is associated with opening of this excitation channel . in figure [ fig2 ] ( b ) ,
icss for elastic electron scattering with si@xmath0 molecules are shown .
the magnitude of the ics is about @xmath61 between 0 and 15 ev .
there are two sharp peaks below 1 ev .
the peak at 0.55 ev is from the @xmath39 symmetry partial cross sections and the other peak at 0.12 ev is from the @xmath44 symmetry .
we checked the csfs of the @xmath39 and @xmath44 symmetry calculations and found that the configuration @xmath62 has dominant contribution to these resonances . in figure [ fig3 ] , calculated polarization fractions ( pfs ) for elastic electron o@xmath0 collisions
are shown for scattering energies of 5 , 10 , 12 and 15 ev with the previous theoretical results of fullerton et al.@xcite , machado et al.@xcite , da paixo et al.@xcite and wste et al.@xcite .
these theoretical results are also compared with the experimental values of hegemann et al.@xcite in the figure . as we can see from eq.[eq1 ] , deviation of pf from unity is a measure of spin - exchange .
our e - o@xmath0 pfs are close to 1 at all scattering energies , indicating the degree of spin - exchange is relatively small . for 5 ev
, our results are very similar to the previous r - matrix results of fullerton et al.@xcite .
the results of machado et al.@xcite are also similar , but smaller at low angles below 30 degrees .
our pfs at 10ev are slightly smaller than the results of fullerton et al.@xcite and machado et al.@xcite .
the pfs of da paixao et al.@xcite at 10 ev are smaller than our results in all angles , especially 120 - 180 degrees .
our calculation can not reproduce the drop of experimental pfs at 10 ev at 100 degrees .
the pfs of wste et al.@xcite have a minimum at this position and their value at 100 degree is the closest to the experimental result , although there is still some deviation in magnitude . at collision energy of 12 ev ,
the results of our calculation , wste et al.@xcite and fullerton et al.@xcite have similar angular behaviour , though our results are smaller than the others at all angles .
all of these three theoretical results agree reasonably well with the experimental results . for 15ev ,
our pfs are larger than the results of the other theoretical calculations in all scattering angles and are closer to the experimental results .
the pfs of fullerton et al.@xcite and machado et al.@xcite are very similar in shape and magnitude , whereas the results of da paixao et al.@xcite are slightly smaller at higher angles above 110 degrees .
the deviation of our pfs from the previous theoretical results is the largest around 90 - 110 degrees , where there is a dip in the profile .
although the magnitude of the pfs are different , the shape of the our pf profiles itself is similar to the previous calculations . in figure
[ fig4 ] , the pfs for elastic electron o@xmath0 collisions are shown as a function of energy at a scattering angle of 100 degrees .
our result has a minimum at 13 ev , however , it is located at 15 ev and 12 ev in the result of fullerton et al.@xcite and wste et al.@xcite , respectively .
the magnitude of the pf at the minimum is larger in wste et al.@xcite than in our calculation and fullerton et al.@xcite . for comparison of the e - o@xmath0 pfs with the pfs of electron b@xmath0 , s@xmath0 and si@xmath0 collisions in the following figures ,
the pfs of elastic electron o@xmath0 collisions are again shown in the figure [ fig5 ] ( a ) for collision energies of 3 , 5 , 7 , 10 and 15 ev .
the depolarization , i.e. , deviation of pf from 1 , is only prominent at 10 and 15 ev where the @xmath2 resonance exists as shown in fig.[fig1 ] ( a ) . in order to check the relation of the @xmath2 resonance and the pfs at 15 ev , we have carried out the r - matrix calculation with modified configurations , removing the @xmath63 configuration from the original calculation . by this procedure we can suppress the effect of the resonance .
the results in the fig .
[ fig5 ] ( a ) indicates that the pfs become very close to 1 by removing the configuration of the @xmath2 resonance . in figure [ fig5 ] ( b ) , calculated pfs for elastic electron b@xmath0 collisions are shown for scattering energies of 3 , 5 , 7 , 10 and 15 ev . for most of the scattering energies and angles , the depolarization in electron b@xmath0
collision is larger than that in electron o@xmath0 collisions .
between 80 and 180 degrees , the magnitude of the pfs are about 0.8 - 0.9 at all energies . in contrast , the e - o@xmath0 pfs are larger than 0.9 .
the e - b@xmath0 pfs show large depolarization effect at scattering energies of 3 and 5 ev , which are close to the @xmath44 resonance .
in order to understand the origin of the large depolarizations at 3 and 5 ev , we excluded the effect of the b@xmath6 @xmath44 resonance around 3.5 ev and re - calculated the e - b@xmath0 pfs .
specifically , we removed @xmath45 configuration from the r - matrix calculation and erased the @xmath44 resonance .
as shown in the fig .
[ fig5 ] ( b ) , the effect of the resonance on the pfs is evident . with the resonance effect ,
the lowest value of the pf at 3 ev is about 0.7 at 90 degrees , but it becomes about 0.95 without the resonance contribution . also , the depolarizations at 5 and 7 ev become less pronounced when we remove the effect of the resonance .
the calculated pfs for elastic electron s@xmath0 collisions are shown in figure [ fig6 ] ( a ) .
in general , the degree of depolarization is smaller than the e - b@xmath0 case , but is larger than the e - o@xmath0 case .
the profiles of the pfs at 7 , 10 and 15 ev look similar to each other .
however , the pfs at 3 and 5 ev behave differently .
the degree of depolarization is larger at forward angles for 3 ev case , however ,
it is larger at backward angles at 5 ev . to understand the effect of resonance on electron s@xmath0 pfs
, we removed the @xmath52 configuration and erased the @xmath51 resonance at 4.5ev , then re - calculated the pfs .
the results are shown in the same figure .
as in the case of electron b@xmath0 pfs , the degree of depolarization becomes smaller when we removed the resonance effect .
the calculated pfs for elastic electron si@xmath0 collisions are shown in figure [ fig6 ] ( b ) . in this case , relatively large depolarization is observed for 3 ev at 100 degrees . the depolarization becomes smaller as the collision energy increases , however , some degree of depolarization remains near 80 and 180 degrees .
the effect of resonance on the electron si@xmath0 pfs were examined by removing the @xmath62 configuration , which is responsible for the @xmath39 and @xmath44 resonances at 0.12 and 0.55 ev , respectively . by removing this configuration ,
these two sharp peaks in the icss disappear .
the pfs without the resonances are shown in the fig .
[ fig6 ] ( b ) . for 3 ev case , the depolarization becomes smaller at all angles .
however , the decrease of depolarization is not so large compared to the cases of e - b@xmath0 and e - s@xmath0 collisions . for 5 ev case , the depolarization at 85 degrees becomes larger , though it becomes smaller at backward angles .
thus , in this case , the association of the resonances with the depolarization is not so straightforward as in the e - o@xmath0 , b@xmath0 and s@xmath0 cases .
as we show in the figures [ fig5 ] and [ fig6 ] , existence of resonance and behaviour of pf is closely related each other .
when resonance exists at some energy , relatively large depolarization is observed compared to the other energies . also , when the resonance is artificially removed by deleting specific configuration in the r - matrix calculation , depolarization becomes smaller in general . in case of electron si@xmath0 collisions , this trend is partly broken at 5 ev around 80 degrees , however , depolarization generally becomes smaller at the other region after removing the resonance effect .
the association of resonance and pf has been discussed in the previous theoretical and experimental papers @xcite , and we have confirmed this association more clearly by explicitly studying the effect of resonance on the pf of four different electron - molecule systems . even if collision occurs away from the resonance energy , some degree of depolarization is observed in all cases of e - o@xmath0 , b@xmath0 , s@xmath0 and si@xmath0 collisions . in e - b@xmath0 case , depolarization is relatively large even outside of the resonance energy region .
in contrast , the pfs in e - o@xmath0 collisions are very close to 1 when collision energy is distant from the resonance energy .
the degree of depolarization in e - s@xmath0 and si@xmath0 collisions is intermediate between e - b@xmath0 and e - o@xmath0 depolarizations .
it is unclear why different degree of depolarization is seen in these four electron - molecule collisions when collision energy is distant from resonance .
the extent of molecular orbitals may be related to this difference , as discussed by sartori et al.@xcite on the electron h@xmath0 superelastic collisions . for the electron o@xmath0 elastic scattering ,
the previous theoretical and experimental pfs are available for comparison with our results .
our low energy pfs at 5 and 10 ev are similar to the other theoretical results .
however , the pfs at 12ev are smaller than the previous results at all angles , and our pfs at 15ev are much closer to unity than the other theoretical results as shown in fig.[fig3 ] .
the reason of these deviations can be attributed to the shift of o@xmath6 @xmath2 resonance position . in our calculation , the position of the @xmath2 resonance peak is located around 13.0 ev , whereas it is around 14.1 ev in the previous r - matrix calculation . as discussed in wste et al.@xcite ,
the position of the @xmath2 resonance is sensitive to the internuclear distance . in this work ,
the internuclear length is fixed to be 2.3 a0 and it is the same as in the calculation of fullerton et al.@xcite .
so the choice of basis set , molecular orbitals and number of target states is important to the difference in the position of the resonance . probably the position of resonance is stabilized by inclusion of more target states in the present r - matrix calculations compared to the previous calculations of fullerton et al.@xcite , machado et al.@xcite and da paixo et al.@xcite . in figure [ fig1 ] ( b ) , sharp increase of cross section is observed in electron b@xmath0 elastic scattering near zero energy .
this increase of cross section is similar to the case of electron polar - molecule collision , although b@xmath0 molecule has no dipole moment .
the cross section of electron co@xmath0 elastic collision also has similar sharp increase near zero energy , and several experimental and theoretical works have been performed to understand this behaviour .
morrison analyzed this problem and suggested that this behaviour is related to the existence of a virtual state @xcite .
morgan has shown that the correlation and polarization effect is important for this sharp peak @xcite .
similar mechanism may exist for the electron b@xmath0 elastic collisions .
for the electron o@xmath0 elastic collisions , we put larger number of target electronic states and better quality molecular orbitals in the r - matrix calculations , and obtained improved results around @xmath2 resonance region compared to the previous fixed - bond calculations . however , the results of wste et al.@xcite agree better with the experimental pfs at 10 ev .
they achieved this good agreement by vibrational averaging of the t - matrices to include the effect of the nuclear motion . by extending the present r - matrix calculation to include the vibrational effect using vibrational averaging procedure or the non - adiabatic r - matrix method , we may obtain better agreement with the experimental pfs .
also , inclusion of nuclear motion effect may improve the quality of the calculations on electron b@xmath0 , s@xmath0 and si@xmath0 elastic collisions .
we have calculated the polarization fractions ( pfs ) on low - energy elastic collisions of spin - polarized electrons with open - shell molecules , o@xmath0 , b@xmath0 , s@xmath0 and si@xmath0 , all of them having @xmath9 symmetry in their ground states . as in our previous works , we employed the fixed - bond r - matrix method based on state - averaged complete active space scf orbitals .
our pfs for electron o@xmath0 collisions agree better with the previous experimental result , especially around the @xmath2 resonance , compared to the previous theoretical calculations .
larger spin - exchange effect is observed in the electron b@xmath0 and si@xmath0 collisions than in the e - o@xmath0 collisions . in e - s@xmath0 collisions ,
degree of depolarization is similar to the e - o@xmath0 collisions . in all four electron - molecule collisions ,
the pfs deviate larger from 1 near resonances .
this association of resonance and pf was explicitly confirmed by the r - matrix calculations removing configurations responsible for the resonance .
panel ( a ) : the elastic integrated cross sections ( icss ) of electron scattering by o@xmath0 molecules .
panel ( b ) : the elastic icss of electron scattering by b@xmath0 molecules .
thick full lines represent total cross sections .
the partial cross sections are represented by thin lines .
symmetries with minor contributions are not shown in the figure . ]
panel ( a ) : the elastic icss of electron scattering by s@xmath0 molecules .
panel ( b ) : the elastic icss of electron scattering by si@xmath0 molecules .
other details are the same as in the figure [ fig1 ] . ]
polarization fractions of electron o@xmath0 elastic scattering .
panel ( a ) : 5ev , ( b ) 10ev , ( c ) 12ev , ( d ) 15ev .
our results are shown as thick full lines .
experimental results of hegemann et al.@xcite are shown as open circles with error bars and theoretical pfs of fullerton et al.@xcite , machado et al.@xcite , da paixao et al.@xcite , and wste et al.@xcite are shown as thin lines . ]
panel ( a ) : polarization fractions ( pfs ) of electron o@xmath0 elastic scattering .
panel ( b ) : pfs of electron b@xmath0 elastic scattering .
calculated pfs are shown as thick lines .
thin lines with symbols represent pfs without the effect of resonances , obtained by the r - matrix calculation with modified configurations ( see text for details ) . ] | the spin - exchange effect in spin - polarized electron collisions with unpolarized open - shell molecules , o@xmath0 , b@xmath0 , s@xmath0 and si@xmath0 , has been studied by the r - matrix method with the fixed - bond approximation .
all of these molecules have @xmath1 symmetry in their ground states .
usual integrated cross sections with unpolarized electrons has also been studied .
we used the complete active space self consistent field orbitals and put more than 10 target electronic states in the r - matrix models . in electron o@xmath0 elastic collisions ,
calculated polarization fractions agree well with the experimental results , especially around the @xmath2 resonance . in e - b@xmath0 , s@xmath0 and si@xmath0 elastic collisions ,
larger spin - exchange effect is observed compared to the e - o@xmath0 elastic collisions . in all four cases ,
spin - exchange effect becomes prominent near resonances .
this association of resonance and magnitude of the spin - exchange effect was studied by explicitly removing the resonance configurations from the r - matrix calculations . in general , spin - exchange effect is larger in e - b@xmath0 collisions than in e - s@xmath0 and si@xmath0 collisions , and is smallest in e - o@xmath0 collisions . |
medical imaging including x - rays , magnetic resonance imaging ( mri ) , computer tomography ( ct ) , ultrasound etc .
are susceptible to noise @xcite .
reasons vary from use of different image acquisition techniques to attempts at decreasing patients exposure to radiation . as the amount of radiation is decreased , noise increases @xcite .
denoising is often required for proper image analysis , both by humans and machines .
image denoising , being a classical problem in computer vision has been studied in detail .
various methods exist , ranging from models based on partial differential equations ( pdes ) @xcite , domain transformations such as wavelets @xcite , dct @xcite , bls - gsm @xcite etc . , non local techniques including nl - means @xcite , combination of non local means and domain transformations such as bm3d @xcite and a family of models exploiting sparse coding techniques @xcite .
all methods share a common goal , expressed as @xmath0 where @xmath1 is the noisy image produced as a sum of original image @xmath2 and some noise @xmath3 .
most methods try to approximate @xmath2 using @xmath1 as close as possible . in most cases ,
@xmath3 is assumed to be generated from a well defined process . with recent developments in deep learning @xcite ,
results from models based on deep architectures have been promising .
autoencoders have been used for image denoising @xcite .
they easily outperform conventional denoising methods and are less restrictive for specification of noise generative processes . denoising autoencoders constructed using convolutional layers have better image denoising performance for their ability to exploit strong spatial correlations . in this paper we present empirical evidence that stacked denoising autoencoders built using convolutional layers work well for small sample sizes , typical of medical image databases . which is in contrary to the belief that for optimal performance , very large training datasets are needed for models based on deep architectures
we also show that these methods can recover signal even when noise levels are very high , at the point where most other denoising methods would fail .
rest of this paper is organized as following , next section discusses related work in image denoising using deep architectures .
section iii introduces autoencoders and their variants .
section iv explains our experimental set - up and details our empirical evaluation and section v presents our conclusions and directions for future work .
although bm3d @xcite is considered state - of - the - art in image denoising and is a very well engineered method , burger et al .
@xcite showed that a plain multi layer perceptron ( mlp ) can achieve similar denoising performance .
denoising autoencoders are a recent addition to image denoising literature .
used as a building block for deep networks , they were introduced by vincent et al .
@xcite as an extension to classic autoencoders .
it was shown that denoising autoencoders can be stacked @xcite to form a deep network by feeding the output of one denoising autoencoder to the one below it .
jain et al .
@xcite proposed image denoising using convolutional neural networks . it was observed that using a small sample of training images , performance at par or better than state - of - the - art based on wavelets and markov random fields can be achieved .
xie et al .
@xcite used stacked sparse autoencoders for image denoising and inpainting , it performed at par with k - svd .
agostenelli et al .
@xcite experimented with adaptive multi column deep neural networks for image denoising , built using combination of stacked sparse autoencoders .
this system was shown to be robust for different noise types .
an autoencoder is a type of neural network that tries to learn an approximation to identity function using backpropagation , i.e. given a set of unlabeled training inputs @xmath4 , it uses @xmath5 an autoencoder first takes an input @xmath6^d$ ] and maps(encode ) it to a hidden representation @xmath7^{d'}$ ] using deterministic mapping , such as @xmath8 where @xmath9 can be any non linear function .
latent representation @xmath3 is then mapped back(decode ) into a reconstruction @xmath1 , which is of same shape as @xmath2 using similar mapping .
@xmath10 in , prime symbol is not a matrix transpose .
model parameters ( @xmath11 ) are optimized to minimize reconstruction error , which can be assessed using different loss functions such as squared error or cross - entropy .
basic architecture of an autoencoder is shown in fig .
[ autoencoder_fig ] @xcite here layer @xmath12 is input layer which is encoded in layer @xmath13 using latent representation and input is reconstructed at @xmath14 . using number of hidden units lower than inputs forces
autoencoder to learn a compressed approximation .
mostly an autoencoder learns low dimensional representation very similar to principal component analysis ( pca ) .
having hidden units larger than number of inputs can still discover useful insights by imposing certain sparsity constraints .
denoising autoencoder is a stochastic extension to classic autoencoder @xcite , that is we force the model to learn reconstruction of input given its noisy version .
a stochastic corruption process randomly sets some of the inputs to zero , forcing denoising autoencoder to predict missing(corrupted ) values for randomly selected subsets of missing patterns .
basic architecture of a denoising autoencoder is shown in fig .
[ denautoencoder_fig ] denoising autoencoders can be stacked to create a deep network ( stacked denoising autoencoder ) @xcite shown in fig .
[ sautoencoder_fig ] @xcite .
output from the layer below is fed to the current layer and training is done layer wise .
convolutional autoencoders @xcite are based on standard autoencoder architecture with convolutional _ encoding _ and _ decoding _ layers .
compared to classic autoencoders , convolutional autoencoders are better suited for image processing as they utilize full capability of convolutional neural networks to exploit image structure . in convolutional autoencoders ,
weights are shared among all input locations which helps preserve local spatiality .
representation of @xmath15th feature map is given as @xmath16 where bias is broadcasted to whole map , @xmath17 denotes convolution ( 2d ) and @xmath9 is an activation . single bias per latent map is used and reconstruction is obtained as @xmath18 where @xmath19 is bias per input channel , @xmath20 is group of latent feature maps , @xmath21 is flip operation over both weight dimensions .
backpropogation is used for computation of gradient of the error function with respect to the parameters .
we used two datasets , mini - mias database of mammograms(mmm ) @xcite and a dental radiography database(dx ) @xcite .
mmm has 322 images of 1024 @xmath22 1024 resolution and dx has 400 cephalometric x - ray images collected from 400 patients with a resolution of 1935 @xmath22 2400 .
random images from both datasets are shown in fig .
[ random_real ] .
all images were processed prior to modelling .
pre - processing consisted of resizing all images to 64 @xmath22 64 for computational resource reasons .
different parameters detailed in table [ datasets ] were used for corruption .
.dataset perturbations [ cols="<,<",options="header " , ] + @xmath23 represents 50% corrupted images with @xmath24 , @xmath25 are images corrupted with @xmath26 , @xmath27 are corrupted with @xmath28 and @xmath29 are corrupted with a poisson noise using @xmath30 also , as the noise level is increased the network has trouble converging .
[ troublecon ] shows the loss curves for gaussian noise with @xmath31 . even using 100 epochs ,
model has not converged .
we have shown that denoising autoencoder constructed using convolutional layers can be used for efficient denoising of medical images . in contrary to the belief , we have shown that good denoising performance can be achieved using small training datasets , training samples as few as 300 are enough for good performance .
our future work would focus on finding an optimal architecture for small sample denoising .
we would like to investigate similar architectures on high resolution images and the use of other image denoising methods such as singular value decomposition ( svd ) and median filters for image pre - processing before using cnn dae , in hope of boosting denoising performance .
it would also be of interest , if given only a few images can we combine them with other readily available images from datasets such as imagenet @xcite for better denoising performance by increasing training sample size .
agostinelli , forest , michael r. anderson , and honglak lee .
`` adaptive multi - column deep neural networks with application to robust image denoising . '' _ advances in neural information processing systems_. 2013 . burger , harold c. , christian j. schuler , and stefan harmeling .
`` image denoising : can plain neural networks compete with bm3d ? . '' _ computer vision and pattern recognition ( cvpr ) _ ,
2012 ieee conference on .
ieee , 2012 .
rudin , leonid i. , and stanley osher .
`` total variation based image restoration with free local constraints . '' image processing , 1994 . proceedings .
_ icip-94_. , ieee international conference .
1 . ieee , 1994 .
vincent , pascal , et al .
`` stacked denoising autoencoders : learning useful representations in a deep network with a local denoising criterion . ''
_ journal of machine learning research _ 11.dec ( 2010 ) : 3371 - 3408 .
yaroslavsky , leonid p. , karen o. egiazarian , and jaakko t. astola .
`` transform domain image restoration methods : review , comparison , and interpretation . ''
_ photonics west 2001-electronic imaging_. international society for optics and photonics , 2001 . | image denoising is an important pre - processing step in medical image analysis .
different algorithms have been proposed in past three decades with varying denoising performances . more recently ,
having outperformed all conventional methods , deep learning based models have shown a great promise .
these methods are however limited for requirement of large training sample size and high computational costs . in this paper
we show that using small sample size , denoising autoencoders constructed using convolutional layers can be used for efficient denoising of medical images .
heterogeneous images can be combined to boost sample size for increased denoising performance .
simplest of networks can reconstruct images with corruption levels so high that noise and signal are not differentiable to human eye .
image denoising , denoising autoencoder , convolutional autoencoder |
a scattering zipper describes consecutive scattering events with a fixed number @xmath0 of incoming and out - going channels each .
it is specified by a sequence @xmath1 of unitary scattering matrices @xmath2 in the unitary group @xmath3 as well as two unitaries @xmath4 modeling the boundary scattering . the size @xmath5 of the system is supposed to be either even or infinite .
then the scattering zipper operator acting on @xmath6 is defined as @xmath7 where the two unitaries @xmath8 and @xmath9 are given by @xmath10 the main hypothesis on each of the scattering matrices @xmath2 is that its upper right entry of size @xmath11 is invertible . in the notation of section [ sec - prel ] below ,
@xmath2 is in a subset @xmath12 of the unitary group .
this condition assures that the scattering is effective so that @xmath13 does not decouple into a direct sum of two or more parts .
the terminology _ scattering zipper _ is best understood by looking at figure 1 illustrating the model .
it shows the first scattering events of a semi - infinite model @xmath14 for which we also drop the indices on @xmath15 , @xmath16 and @xmath17 .
it is also possible to consider periodic scattering zippers , see section [ sec - periodic ] below for finite operators and section [ sec - periodic - infinite ] for infinite ones .
furthermore , by placing either of the boundary conditions @xmath18 and @xmath19 into @xmath8 , one can consider the case of odd @xmath5 , but we refrain from doing so . and @xmath20 satisfy @xmath21 , which is equivalent to @xmath22 and @xmath23 .
the picture corresponds to proposition [ prop - solutions ] below .
the boundary conditions @xmath24 and @xmath25 as well as @xmath26 can be introduced for convenience , but one can also just keep @xmath18 , _ c.f . _
section [ sec - transfer ] .
the main message of this paper is that scattering zippers are the unitary analogs of jacobi matrices with matrix entries . here
are the structural results supporting this claim : * the matrix @xmath13 is not tridiagonal , but five - diagonal .
nevertheless , solutions of the associated eigenvalue equation can be calculated by transfer matrices having the same symmetries as in the jacobi case , but no further restrictions ( see section [ sec - transfer ] ) . *
the matrix element of the resolvent of @xmath13 corresponding to the left boundary site @xmath27 has a simple expression in terms of the entries of the transfer matrix , namely it is given by a mbius transformation of the other boundary condition ( theorem [ theo - green ] ) .
* these matrix elements of the resolvent lie on a weyl surface which is a matrix ball ( theorems [ theo - weyl1 ] and [ theo - weylsurf ] ) .
* all semi - infinite scattering zipper operators @xmath15 with fixed boundary condition @xmath18 are in the limit point case .
fixing an appropriate gauge for each @xmath2 , the semi - infinite scattering zipper operators are in bijection with their spectral measures which are all matrix - valued probability measures on the unit circle ( theorem [ theo - bijection ] ) . *
the eigenvalues of the finite scattering zipper operators can be calculated using matrix prfer phases by sturm - liouville type oscillation theory ( theorems [ theo - oscillation ] and [ theo - osci2 ] ) .
this is also an efficient tool to calculate the spectrum of infinite periodic scattering zippers ( theorem [ theo - oscillationper ] ) . for jacobi matrices with matrix entries
, it is well - known how to use transfer matrices , and how to calculate the resolvents ( _ e.g. _ @xcite , but this is classical ) .
weyl surfaces are also known for jacobi matrices ( see @xcite which also contains earlier references ) .
the reason why there is a simple bijection in theorem [ theo - bijection ] is that all semi - infinite scattering zipper operators are in the weyl limit point case , namely the weyl surfaces shrink to one point in the large @xmath5 limit .
this is in strong contrast with jacobi matrices where there are limit discs which lead to all the issues related to the moment problem ( see _ e.g. _ the beautiful paper by simon @xcite ) , nevertheless there are close connections between probability measures on the real line and jacobi matrices .
finally , oscillation theory of jacobi matrices with matrix entries was developed in @xcite , but again the scalar case is well - known .
now let us present our personal and without doubt very restricted view on connections of this work to the literature .
first of all , the scattering zipper itself is a generalization of three well - known models , the blatter - browne model @xcite and the chalker - coddington network model @xcite of the solid state physics community as well as the cmv matrices @xcite of the mathematical literature .
the blatter - browne model is scalar , namely @xmath28 . in the chalker - coddington model , the aim is to model a higher dimensional lattice of scattering events . in terms of the scattering zipper
this means that there is supplementary structure in each of the scattering matrices @xmath2 and that they are infinite - dimensional and very sparsely filled .
motivated by applications to the quantum hall transitions , the main focus in the chalker - coddington model has been on random scattering events .
first rigorous works on the analysis of the blatter - brown model and the chalker - coddington model on a strip have appeared @xcite . on the other hand , the cmv matrices in its matricial version
@xcite only consider scattering blocks of the type @xmath29 where the @xmath30 verify @xmath31 which is equivalent to @xmath32 ( here and below we always use the operator norm ) .
these @xmath30 are called the verblunsky coefficients @xcite .
there is a huge literature on cmv matrices ( see @xcite for a long list of references ) .
they form a subclass of the scattering zipper models considered here . of course
, the transfer matrix techniques also apply and have very efficiently been used in most works on the subject . however , the weyl discs at finite @xmath5 as presented below seem to have been studied only in the scalar case @xcite . nevertheless , it was possible to prove in @xcite that every sequence of verblunsky coefficients corresponds to a unique matrix - valued measure on the unit circle ( verblunsky s theorem ) .
theorem [ theo - bijection ] extends this theorem in that it exhibits a bijection between all probability measures on the unit circle and the semi - infinite scattering zippers with fixed boundary condition @xmath18 .
finally , it seems that oscillation theory for cmv matrices was only developed in the scalar case @xmath28 ( * ? ? ?
* theorem 8.3.3 . ) .
* acknowlegements : * h. s .- b . wants to thank mihai stoiciu for introducing him to the world of cmv matrices and for a number of discussions about oscillation theory at a very early stage of this work .
he also let us know about the reference @xcite .
we also acknowlege financial support of the dfg .
as explained in the introduction , the following subset of the even - dimensional unitary group will play a role : @xmath33 in , equivalent to the condition that @xmath34 is invertible is either the invertibility of @xmath35 or the condition @xmath36 or the condition @xmath37 .
furthermore , one has the representation @xmath38 where @xmath39 * proof .
* the equations @xmath40 give @xmath41 from three of these identities the first claims can be deduced immediately . for ,
let us first of all note that @xmath42 implies that @xmath43 has a unique polar decomposition @xmath44 with some unitary @xmath18 .
similarly , @xmath45 shows that @xmath46 for some unitary @xmath19 .
but then @xmath47 . @xmath48 recall that the lorentz group @xmath49 of signature @xmath50 is defined to be the set of @xmath51 matrices conserving the form @xmath52 the following well - known result on the passage from scattering matrices to transfer matrices is illustrated in figure 2 .
[ prop - stransferequiv ] the formula @xmath53 defines a bijection from @xmath12 onto @xmath54 . for any vectors @xmath55 , one has the equivalence @xmath56 moreover , if @xmath57 form an inhomogeneity , then @xmath58 * proof . *
first of all , note that @xmath59 is well - defined because @xmath34 is invertible .
next one readily checks that @xmath60 by using the relations and .
moreover , the inverse of @xmath59 is given by @xmath61 next let us check , which generalizes .
the upper equation on the l.h.s .
is @xmath62 which can be rewritten as @xmath63 this is already the lower equation of the r.h.s .. solving it for @xmath64 and replacing in the lower equation of the l.h.s .
gives the upper equation on the r.h.s .. @xmath48 .
figure 2(i ) is the usual graphical representation of a scattering event with incoming amplitudes @xmath65 and @xmath64 and outgoing amplitudes @xmath66 and @xmath67 .
figures 2(ii ) and 2(iii ) corresponding to the left and right hand side of equation show different representations of the same event as they are used as building blocks in figure 1 and 3 respectively .
_ ] let us conclude this section with a few comments .
first of all , one readily checks @xmath68 comparing with , one way to characterize cmv matrices is therefore to say that its individual scattering events give rise to self - adjoint transfer matrices . of course
, this does not imply that products of such transfer matrices have the same property .
this means that cmv matrices are not specified by some symmetry given by a combination of time - reversal , particle - hole or sublattice symmetry , and it is not clear to us whether there is a deeper reason to consider scattering events of the type @xcite . on the other hand ,
it is possible to implement all the above symmetries also in scattering zippers , just as for jacobi matrices with matrix entries ( see @xcite where only even and odd time - reversal symmetry is dealt with ) . finally , let us briefly discuss degenerate scattering events which do nt mix all incoming and outgoing amplitudes .
there are many possibilities to do this , but only two are relevant for the boundary conditions in the next section . in one
@xmath66 only depends on @xmath65 ( left and right in figure 1(i ) decoupled ) , in the other only on @xmath64 ( top and bottom in figure 1(i ) decoupled ) .
let us focus on the latter .
then there are two unitaries @xmath18 and @xmath19 such that @xmath69 and @xmath70 . in this case ,
@xmath71 and @xmath72 are given by @xmath73
in this section , the formal solutions @xmath74 for the eigenvalue equation @xmath21 at @xmath75 will be constructed . here
all the @xmath76 are @xmath11 matrices and the index @xmath77 runs to infinity and it is formal in the sense that @xmath78 is typically not square integrable .
the construction of @xmath78 is done such that the left boundary condition ( at site @xmath27 ) is satisfied . for finite @xmath5
, the solution @xmath78 in general does not satisfy the right boundary condition , namely @xmath79 .
[ prop - solutions ] let @xmath80 .
then the following assertions are equivalent : \(i ) @xmath81 and @xmath82 for @xmath83 .
\(ii ) @xmath82 and @xmath84 for @xmath83 .
\(iii ) for any @xmath85 , @xmath86 where the transfer matrices and initial condition are @xmath87 into a transfer matrix multiplication as done in proposition [ prop - solutions ] . _ ] * remark * the application @xmath59 is used here even if @xmath88 is not on the unit circle so that @xmath89 is not unitary .
of course , @xmath90 is in the group @xmath49 only if @xmath88 is on the unit circle . also let us point out that @xmath91 is actually independent of @xmath88 .
moreover , one could also use the initial condition @xmath92 and add in ( iii ) one more equation , namely @xmath93 this results from the discussion at the end of section [ sec - prel ] . @xmath94 * proof * of proposition [ prop - solutions ] : the equivalence of ( i ) and ( ii ) results immediately from @xmath95 .
the equivalence of ( ii ) and ( iii ) can be checked by successive application of the identity in proposition [ prop - stransferequiv ] .
more precisely , the two equations of ( ii ) mean that for any @xmath96 @xmath97 applying the transformation @xmath59 from proposition [ prop - stransferequiv ] to both of these equations than shows the equivalence with ( iii ) .
@xmath48 as usual , the transfer matrices can be nicely iterated .
let us set for @xmath98 @xmath99 and in order to nicely write the solutions , let us also introduce the notations @xmath100 then the solutions are simply given by @xmath101
as a preparation for the weyl theory , this section calculates the entries of the resolvents of @xmath102 corresponding to the site @xmath27 in terms of the entries of the transfer matrix from @xmath103 to @xmath5 : @xmath104 the first object of study is the green matrix @xmath105 where @xmath106 is the partial isometry onto the @xmath77th site and @xmath88 is in the open unit disc @xmath107 . for self - adjoint matrices ,
the green matrix has a herglotz property , which it is laking in the present situation . actually , for the unitary operator @xmath102 it is more natural to consider @xmath108 one readily checks that this analytic function @xmath109 has a positive imaginary part @xmath110 .
it also satisfies @xmath111 and is thus a so - called caratheodory function ( up to the factor @xmath112 ) .
if one also takes the cayley transform of the domain @xmath107 to the upper half - plane @xmath113 , one has again a herglotz function .
a little more information on this is resembled in appendix b. [ theo - green ] the matrix @xmath114 is invertible and @xmath115 lies in the siegel disc @xmath116 .
moreover , the resolvent matrix @xmath117 is given by @xmath118 and the green matrix by @xmath119 let us note that the formulas for @xmath120 and @xmath117 can also be written using the ( inverse ) matrix mbius transformation ( see appendix a ) : @xmath121 where the cayley transformation is the @xmath51 unitary matrix defined by @xmath122 based on and the results of appendix a , one obtains further identities , _ e.g. _
@xmath123 for the proof of theorem [ theo - green ] , one needs a number of lemmata which will also be useful later on for other purposes .
[ lem - lemmap ] let @xmath124 for some @xmath125 and @xmath126 , and set @xmath127 then @xmath128 and @xmath129 moreover , for any @xmath130 , the mbius transformation @xmath131 is well - defined and lies in @xmath132 .
furthermore , only for @xmath133 , @xmath134 * proof . * with the notations of the lemma , the matrix @xmath135 satisfies the defining equation @xmath136 of the group @xmath54 so that @xmath137 thus a direct computation using the identities contained in leads to . note that alternatively @xmath138 can be expressed in terms of @xmath139 and @xmath140 only .
it remains to show the positivity of @xmath138 .
let @xmath141 be with , say , @xmath142 .
then @xmath143 where in the last inequality the bound @xmath142 was used . the case @xmath144 is dealt in the same manner . as to the last claims ,
let us begin by noting @xmath145 thus we need to check the invertibility of @xmath146 for @xmath130 and @xmath133 .
for that purpose let us use again so that , in particular , @xmath147 . thus @xmath148 and @xmath140 is invertible .
also @xmath149 from which follows @xmath150 .
therefore @xmath151 is indeed invertible .
thus @xmath152 is well - defined and @xmath153 but using @xmath154 \ , ( { { \cal t}}^z)^{-1 } \begin{pmatrix } z \\ { { \bf 1}}\end{pmatrix } \\ & < & \begin{pmatrix } z \\ { { \bf 1}}\end{pmatrix}^*\,{{\cal l}}\ , \begin{pmatrix } z \\ { { \bf 1}}\end{pmatrix } \;=\;z^*z-{{\bf 1}}\;\leq\;0\;,\end{aligned}\ ] ] which together with shows the last claim .
@xmath48 [ coro - positive ] with the positive matrix @xmath155 defined in terms of entries of @xmath156 as in lemma [ lem - lemmap ] , @xmath157 * proof .
* this follows by iterating lemma [ lem - lemmap ] and using @xmath158 .
@xmath48 [ lem - invertible ] let @xmath159 be @xmath160 matrices and set @xmath161 . if @xmath162 , then @xmath163 is invertible . on the other hand , if @xmath164 , then @xmath165 is invertible .
* @xmath162 implies that @xmath166 . if there is a vector @xmath167 such that @xmath168 , then @xmath169 which is impossible and therefore @xmath163 is invertible .
if @xmath170 , the same argument implies invertibility of @xmath165 .
@xmath48 [ lem - invertible2 ] let @xmath171 and @xmath172 be two @xmath173 matrices of maximal rank satisfying @xmath174 and either @xmath164 or @xmath162
. then @xmath175 is invertible
. * proof .
* the claimed invertibility does not depend on normalization so that we may assume that @xmath176 and @xmath177 .
now the fact that @xmath171 is @xmath178-lagrangian implies @xmath179 applying @xmath180 and @xmath172 to the left and right of this equation shows @xmath181 thus let us argue that @xmath182 because this then shows that @xmath183 and thus also @xmath175 is invertible . now @xmath184 is given by the squares of the scalar products of the vectors in the planes spanned by @xmath172 and @xmath171 and its eigenvalues are thus the squares of the cosines of the principal angles between these planes .
there is an eigenvalue @xmath27 if and only if one angle vanishes and therefore if and only if @xmath172 and @xmath171 have a direction in common .
this would mean that there are non - vanishing vectors @xmath185 such that @xmath186 , which is incompatible with @xmath174 and @xmath164 ( or @xmath162 ) .
@xmath48 [ lem - solution ] suppose @xmath187 .
given @xmath188 with @xmath189 , the solution @xmath66 of the equation @xmath190 is , for even @xmath77 , given by @xmath191 where @xmath192 and @xmath193 are the entries of @xmath194 and @xmath195 has to be chosen in the unique manner such that @xmath196 given by satisfies @xmath197 * proof . * as above
, the equation @xmath190 is solved using an auxiliary vector @xmath65 satisfying @xmath198 and @xmath199 .
this is equivalent that , for @xmath200 and respectively @xmath201 , @xmath202 together with the condition @xmath203 and the constraint that @xmath204 stemming from the other boundary condition .
each of the two equations in ( the first one divided by @xmath88 ) is transformed using proposition [ prop - stransferequiv ] : @xmath205 iterating and replacing @xmath203 gives , for even @xmath77 , @xmath206 while for odd @xmath77 the entries on the l.h.s .
are simply exchanged ( _ cf . _ the definition of @xmath207 ) .
now one , moreover , has to satisfy the constraint @xmath208 . for that purpose ,
one takes the last equation for @xmath209 and extracts @xmath196 which is then set equal to @xmath210 .
this leads to equation which can indeed be solved for @xmath195 because the matrix @xmath211 is invertible by lemma [ lem - invertible ] .
indeed , corollary [ coro - positive ] shows that @xmath212 satisfies @xmath164 which is the hypothesis in lemma [ lem - invertible ] . @xmath48 * proof * of theorem [ theo - green ] . by the last claim of lemma [ lem - lemmap ] ,
@xmath213 is well - defined and lies in the siegel disc @xmath214 . iterating this shows that @xmath215 exists and lies in @xmath214 .
this shows the first claim .
let us note that the invertibility @xmath216 also follows from the identity @xmath217 because @xmath218 satisfies @xmath164 as shows corollary [ coro - positive ] , so that lemma [ lem - invertible2 ] applies .
the green matrix @xmath219 is the component @xmath195 of the solution of @xmath220 with inhomogeneity @xmath221 .
hence by lemma [ lem - solution ] @xmath222 and @xmath223 now as @xmath224 , one checks that using the identities in @xmath225 @xmath226 therefore the above two equations become @xmath227 now these two equations have to be solved for @xmath219 . for that purpose , the invertibility of @xmath228 is needed .
it follows from lemma [ lem - invertible2 ] because @xmath229 is @xmath178-lagrangian and @xmath230 satisfies @xmath164 as already argued in the proof of lemma [ lem - solution ]
. therefore @xmath231 now the formula for @xmath219 follows .
finally let us calculate @xmath117 .
we start from @xmath232 let us denote @xmath233 , thus we can calculate @xmath234 using the same procedure than with @xmath235 .
notably , @xmath234 is the equal to the component @xmath195 of the matrix - valued solution of @xmath190 with inhomogeneity @xmath236 .
hence @xmath237 and @xmath238 using again @xmath224 and the identities in @xmath225 , one checks that @xmath239 therefore the above two equations become @xmath240 and @xmath241 solving yields @xmath242^{-1 } \bigl [ ( d^z_n - vb^z_n)(-z^{-1}\alpha_2\gamma_2^{-1 } ) - ( c^z_n - va^z_n)u^*\gamma_2^{-1 } \bigr ] \;.\ ] ] replacing this and @xmath235 in the form into leads to the desired formula for @xmath117 . @xmath48
for @xmath133 , let us set @xmath243 we consider @xmath244 as quadratic form on @xmath245 and are particularly interested in its maximally isotropic subspaces , also called lagrangian subspaces .
it will shortly be shown that the signature of @xmath244 is @xmath50 so that the dimension of these lagrangian subspaces is @xmath246 . by definition
the weyl surface is then the image of ( the inverse cayley transform of ) these subspaces under an adequate chart on the grassmannian called the stereographic projection . recall that the grassmannian @xmath247 of all @xmath246-dimensional subspaces of @xmath245 can be introduced as set of equivalent classes of @xmath248 matrices of maximal rank with respect to the equivalence relation @xmath249 for some @xmath250 .
elements of @xmath247 will be denoted by @xmath251_\sim$ ] .
let us consider the subset @xmath252 of subspaces represented by some @xmath253 with an invertible @xmath11 matrix @xmath163 .
this set is the domain of the stereographic projection @xmath254 defined by @xmath255_\sim ) \;=\ ; a\,b^{-1}\ ; , \qquad \phi \;=\ ; \binom{a}{b } \;.\ ] ] the weyl surface is defined by @xmath256_\sim \in \mathbb{g}_l \;\big|\ ; \phi\ ; \mbox{\rm isotropic for } \mathcal{q}_n^z \right\}\bigr ) \;,\ ] ] and the closed weyl disc by @xmath257_\sim \in \mathbb{g}_l \;\big|\ ; \phi^*\mathcal{q}_n^z\phi\leq 0 \right\}\bigr ) \;.\ ] ] of course , we have to check below that this is well - defined , namely that @xmath258_\sim$ ] is in the domain of @xmath259 for all lagrangian subspaces of @xmath260 .
then the next aim will be to show that the weyl surface is the surface of a matrix ball and this will ultimately allow to prove estimates on the dependence of @xmath219 and @xmath117 on the boundary condition @xmath261 .
[ prop - qformulas ] the quadratic form @xmath262 has the following properties : \(i ) @xmath263 and @xmath264 . furthermore @xmath265 where @xmath155 is positive matrix given in term of entries of @xmath156 defined in lemma [ lem - lemmap ] .
\(ii ) @xmath266 \(iii ) @xmath267 \(iv ) @xmath268 * proof . *
( i ) is just a reformulation of lemma [ lem - lemmap ] and corollary [ coro - positive ] .
( ii ) results from .
( iii ) is a consequence of ( i ) and @xmath269 .
finally , from the definition of @xmath270 , one deduces that @xmath271 which is ( iv ) .
@xmath48 [ prop - isovector ] let @xmath272 be given as in theorem [ theo - green ] .
if @xmath172 is lagrangian for @xmath270 , then there is a unique unitary @xmath19 and an invertible matrix @xmath163 with @xmath273 moreover , @xmath258_\sim$ ] is in the domain of @xmath259 .
* let @xmath161 satisfy @xmath274 .
this implies that @xmath275 by proposition [ prop - qformulas](i ) and thus by lemma [ lem - invertible ] that @xmath163 is invertible .
as @xmath276 is isotropic if and only if @xmath277 is isotropic , we may assume that @xmath278 .
now @xmath274 is equivalent to @xmath279 using polar decomposition , it thus follows that there exists a unique unitary @xmath19 such that @xmath280 as @xmath281 is invertible by theorem [ theo - green ] , it follows indeed that @xmath282 .
furthermore , @xmath283 by theorem [ theo - green ] , @xmath120 is in the siegel disc so that @xmath284 is invertible and therefore @xmath258_\sim$ ] is in the domain of @xmath259 .
@xmath48 [ theo - weyl1 ] one has @xmath285 * proof . *
the first equality follows immediately from and in theorem [ theo - green ] .
the second one then follows by combining and .
@xmath48 the next aim is to analyze the geometry of the weyl surface .
this can be done in complete analogy with @xcite if one works with the cayley transform of the quadratic form : @xmath286 where @xmath287 [ def - radial ] the radial and central operators are defined by @xmath288^{-1}\ ; , \qquad s_{n}^z \ ; = \ ; -\ , r_{n}^z\;\binom{{{\bf 1}}}{0}^ * { \widetilde{\mathcal q}}^z_n \binom{0}{{{\bf 1}}}\;.\ ] ] [ prop - snrn ] let @xmath133 .
then @xmath289 is well - defined , positive and decreasing in @xmath5 .
also @xmath290 is well - defined , but negative .
moreover : @xmath291 * proof . *
all claims on @xmath292 follow by taking the matrix element of because @xmath293 and the map @xmath294 is operator monotone . for the second claim ,
one has to adapt lemma [ lem - lemmap ] and corollary [ coro - positive ] to the case @xmath295 , but then the proof is identical .
next let us note that the cayley transform of proposition [ prop - qformulas](ii ) reads @xmath296 using @xmath297 , this yields @xmath298 the upper left and upper right entries of this equation give @xmath299 these two equations lead to the remaining two claims .
@xmath48 [ theo - weylsurf ] the weyl discs are strictly nested matrix discs , namely one has @xmath300 and @xmath301 , as well as @xmath302 * proof . *
the first claim is an immediate corollary of proposition [ prop - qformulas](iii ) .
let @xmath303 be a @xmath304lagrangian plane .
this is equivalent to @xmath305 where in the second equality definition [ def - radial ] and proposition [ prop - snrn ] were used .
rewriting gives @xmath306 by proposition [ prop - isovector ] and theorem [ theo - green ] , @xmath307 so that @xmath308 is invertible .
hence @xmath309 therefore there exists a unique unitary @xmath310 such that @xmath311 as the l.h.s .
is precisely @xmath312_\sim)$ ] , this concludes the proof . @xmath48 the following is now an immediate consequence .
[ coro - gweyl ] for any boundary condition @xmath19 and @xmath133 ,
there is a unitary @xmath313 such that @xmath314 and @xmath315 by now the beautiful theory of weyl discs is complete .
let us finally come to its main analytical application , namely the control of the dependence of the green matrix @xmath316 and the resolvent @xmath317 ( both matrix elements of the resolvent at the boundary site @xmath27 ) on the boundary condition @xmath19 on the ( other ) boundary site @xmath5 .
let us start by noting the following .
[ coro - gfbound ] for any @xmath318 ,
@xmath319 this becomes useful in combination with the following result .
[ prop - rbound ] one has @xmath320 * proof . *
using and the bound from lemma [ lem - lemmap ] , @xmath321 now @xmath322 and for @xmath323 @xmath324 ( this is rough , but sufficient for our purposes . ) therefore @xmath325 from which the upper bound on @xmath326 follows . using a version of lemma [ lem - lemmap ] for @xmath327 , the bound on @xmath328
is shown similarly .
let now @xmath15 be a semi - infinite scattering zipper operator associated to a sequence @xmath329 of scattering matrices @xmath330 and a boundary condition @xmath331 .
then it follows from proposition [ prop - rbound ] combined with corollary [ coro - gfbound ] that the limit @xmath332 exists and is independent of the choice of @xmath19 . in the terminology of weyl theory ,
a semi - infinite scattering zipper operator is always in the limit point case .
moreover , the convergence is uniform in @xmath88 on compact subsets of @xmath107 .
therefore , @xmath333 is analytic . as @xmath334 and @xmath335 it follows that also @xmath336 and @xmath337 .
therefore the riesz - herglotz representation theorem recalled in appendix b can be applied to show that there is a unique matrix - valued probability measure @xmath338 on @xmath339 such that @xmath340 this measure is called the spectral measure of @xmath15 .
it dominates all other spectral measure ( because the range of @xmath341 is a cyclic subspace for @xmath15 ) . resuming ,
there is a map @xmath342 that associates to every semi - infinite scattering zipper operator a matrix - valued probability measure @xmath343 on @xmath339 .
there is still some gauge freedom though allowing several operators @xmath15 to have the same spectral measure , as already pointed out in @xcite .
this is dealt with in the following result .
[ theo - bijection ] let @xmath344 denote the set of all semi - infinite scattering zipper operators with left boundary condition @xmath345 and scattering matrices @xmath329 given by @xmath346 where @xmath32 and @xmath347 .
then the map @xmath342 establishes a bijection between @xmath344 and the matrix - valued probability measures on @xmath339 . for the proof , it remains to construct an inverse to @xmath342 , namely to construct a scattering zipper with boundary condition @xmath18 from a given matrix - valued probability measure @xmath338 on @xmath339 , such that @xmath338 is again its spectral measure .
adapting the approach in @xcite , this will be done by a gram - schmidt procedure associated an adequate scalar product @xmath348 ( with values in the @xmath11 matrices ) and an adequate basis for the functions on @xmath339 ( this will be given by laurent polynomials in an adequate order as suggested by @xcite ) .
let us define @xmath349 where @xmath350 and @xmath351 are matrix - valued functions . at first sight
, there seems to be something wrong with this definition .
indeed , one can define another scalar product by exchanging @xmath350 and @xmath351 on the r.h.s .. when the gram - schmidt procedure below is done w.r.t .
this other scalar product , one obtains different orthonormal polynomials , and only their so - called szeg transformation @xcite will lead to the polynomials @xmath78 and @xmath352 used in section [ sec - transfer ] . somewhat anticipating , already the notations @xmath78 and @xmath352
will be chosen already now .
using allows to avoid using the szeg transformation and also simplifies the calculations below .
the product is left matrix - linear in the second argument and anti - linear in the first , namely it satisfies for matrix valued functions @xmath350 ,
@xmath351 and @xmath353 on @xmath339 as well as a matrix @xmath354 that @xmath355 now two sequences @xmath356 and @xmath20 of orthonormal families of matrix - valued laurent polynomials in @xmath88 will be constructed by the gram - schmidt algorithm w.r.t . .
for @xmath78 , one orthonormalizes the sequence @xmath357 .
then @xmath80 and the @xmath77th element of the resulting orthonomal sequence is @xmath76 .
simarly , @xmath352 is obtained by orthonormalizing @xmath358 .
hence @xmath359 .
the orthonormality relations read : @xmath360 let us note that the @xmath78 satisfy @xmath361 where the leading coefficient @xmath362 are invertible matrices , @xmath363 is a matrix polynomial in the span of @xmath364 and @xmath365 a polynomial in the span of @xmath366 . similarly , @xmath367 with invertible @xmath368 and polynomials @xmath369 and @xmath370 in the span of @xmath371 and @xmath372 respectively .
now let us also define @xmath373 [ lem - zsegorecurs ] the orthonormal polynomials @xmath78 and @xmath352 have the following properties .
\(i ) there exist @xmath160 matrices @xmath374 , such that @xmath375 and @xmath376 \(ii ) one has @xmath377 in the following , we will simply denote @xmath378 by @xmath379 and @xmath380 by @xmath381 .
\(iii ) there are unique unitaries @xmath382 and @xmath383 with unit determinant such that @xmath384 \(iv ) one has @xmath385 . * proof .
* all four relations in and are dealt in the same manner , so let us focus on the first one . by and , @xmath386 is in the span of @xmath387 .
moreover , this polynomial is orthonormal to @xmath388 as @xmath389 and @xmath390 are by construction .
this implies that @xmath386 is a left multiple of @xmath391 .
for item ( ii ) , let us first calculate as follows : @xmath392 on the other hand , @xmath393 thus @xmath394 .
the other equality is checked in a similar manner .
next let us check the first identity of for odd index : @xmath395 thus , there exists a unique unitary of unit determinant denoted @xmath396 such that @xmath397 the other cases are dealt with in the same way . as each @xmath398 and @xmath399
is invertible , the identities also imply ( iv ) .
@xmath48 * proof * of theorem [ theo - bijection ] . from the density of laurent polynomials
, it follows that @xmath400 and @xmath401 are orthonormal basis of the square integrable matrix - valued functions on @xmath339 w.r.t .
thus any @xmath160 matrix valued function @xmath350 can be expanded as follows : @xmath403 now let us define the matrix entries of semi - infinite matrix - valued operators by @xmath404 then it follows from that @xmath405 it remains to show that @xmath16 and @xmath17 defined above have the same structure as their homonyms from section [ sec - transfer ] .
for @xmath406 it follows from that @xmath407 where the identity @xmath408 was used . recalling the notation
, it follows that @xmath409 together with @xmath410 , one obtains @xmath411 similarly , implies @xmath412 thus , for @xmath406 , @xmath413 and one obtains @xmath414 in conclusion , @xmath15 is a semi - infinite scattering zipper operator . @xmath48
the transfer matrices allow to calculate formal solutions of @xmath415 by equation .
as already pointed out in section [ sec - transfer ] , the @xmath66-component of @xmath416 does not yet lead to a matrix - valued solution of @xmath415 because there is a supplementary constraint on the two components of @xmath417 . indeed , because @xmath5 is even , the last equation of @xmath418 is @xmath419 . for each vector
@xmath420 satisfying @xmath421 one finds an eigenvector of @xmath13 .
let us reformulate this in terms of the dimension of an intersection of two @xmath246-dimensional subspaces of @xmath245 : @xmath422 where @xmath423 now the intersection of the planes @xmath424 and @xmath425 can be conveniently calculated using the fact that both planes are @xmath178-lagrangian for @xmath426 , namely they both satisfy @xmath427 .
in fact , the initial condition @xmath428 as well as @xmath429 are @xmath178-lagrangian as shows a direct calculation . as the transfer matrices @xmath430 for @xmath426 all conserve the form @xmath178
, also @xmath431 is @xmath178-lagrangian for all @xmath77 . before analyzing the intersection of two @xmath178-lagrangian planes ,
let us study the set @xmath432 of lagrangian planes , namely those @xmath251_\sim\in{{\mathbb g}}_l$ ] satisfying @xmath427 .
[ prop - lagrangian ] @xmath433 and @xmath434 is a bijection .
every @xmath251_\sim\in{{\mathbb l}}_l$ ] has a representation of the form @xmath435 with @xmath345 .
* let @xmath161 be @xmath178-lagrangian .
then @xmath436 because otherwise @xmath172 would not be of rank @xmath246 .
moreover , @xmath437 so that both @xmath165 and @xmath163 are invertible . also @xmath438 and @xmath439 .
from this all claims follow .
@xmath48 now follows a general result about the intersection of two @xmath178-lagrangian planes .
[ prop - intersection ] let @xmath440 and @xmath441 be @xmath178-lagrangian frames and set @xmath442_\sim)^*\pi([\psi]_\sim)$ ] . then @xmath443 * proof .
* let us begin with the inequality @xmath444 of the first equality .
suppose there are two @xmath445 matrices @xmath446 of rank @xmath447 such that @xmath448 . then @xmath449 so that the kernel of @xmath450 is at least of dimension @xmath447 . inversely , given a @xmath445 matrix @xmath451 of rank @xmath447 such that @xmath452 , one deduces that @xmath453 .
as the column vectors of @xmath172 and @xmath454 are orthogonal and span @xmath455 , it follows that the column vectors of @xmath456 lie in the span of @xmath172 , that is , there exists an @xmath445 matrix @xmath167 of rank @xmath447 such that @xmath457 .
this shows the other inequality and hence proves the first equality of the lemma . for
the second , let us first note that the dimension of the kernel of @xmath450 does not depend on the choice of the representatives .
choosing the representatives @xmath458_\sim)}{{{\bf 1}}}$ ] and @xmath459_\sim)}{{{\bf 1}}}$ ] then shows the second equality .
@xmath48 for @xmath426 ,
let us define @xmath460 to be the unitary associated by proposition [ prop - intersection ] to @xmath461 and @xmath462 .
taking into account the explicit form of @xmath429 as well as , one finds @xmath463 these unitaries are the analogs of matrix - valued prfer phases and can conveniently be calculated by iterated mbius transformations with the transfer matrices .
now the main facts of oscillation theory can be stated .
[ theo - oscillation ] let @xmath464 be even . for @xmath426 ,
one has @xmath465 furthermore , setting @xmath466 , @xmath467 so that all eigenvalues of @xmath468 rotate in the positive sense as a function of @xmath469 .
* the equality follows immediately from , the definition of @xmath468 and proposition [ prop - intersection ] .
further , let us note that @xmath470 .
thus one calculates @xmath471(\phi_n^z)^{-1}v \\ & = & v^*\bigl((\phi_n^z)^{-1}\bigr)^ * ( \phi^z_n)^*(\imath{{\cal l}})(\partial_\theta \phi^z_n)\,(\phi_n^z)^{-1}v\;.\end{aligned}\ ] ] it is therefore sufficient to check the positivity of @xmath472 . by and @xmath473 ,
one now finds @xmath474 for odd @xmath77 , the summands vanish because @xmath430 is independent of @xmath88 and thus also @xmath475 .
we shall now show that for every even @xmath77 there is a strictly positive definite contribution .
for that purpose , let us drop the index @xmath77 on @xmath430 and @xmath2 .
now @xmath476 is given by @xmath477 where @xmath478 , @xmath479 , @xmath139 and @xmath140 are the coefficients of @xmath480 .
then one verifies @xmath481 it remains to check that this matrix is positive . on first sight
, it is not even hermitian , but actually the defining relations of @xmath49 are @xmath482 the last one shows that the above matrix is indeed hermitian .
now let @xmath141 be with , say , @xmath142 .
then @xmath483 where in the second step we used the cauchy - schwarz inequality .
this completes the proof .
@xmath48 * remark * as @xmath13 has exactly @xmath484 eigenvalues , the rotation number of @xmath485 is equal to @xmath484 .
this can also be shown independently by calculating the maslov index as in @xcite .
it is possible to associate to a sequence @xmath486 of scattering matrices @xmath487 with @xmath5 still even , a periodic scattering zipper operator @xmath488 by setting @xmath489 where @xmath490 the aim of this section is to calculate the spectrum of @xmath491 .
this parallels the calculations in @xcite and is useful for the calculation of the spectrum of infinite periodic scattering zippers , as explained in section [ sec - periodic - infinite ] .
the solutions @xmath492 of the eigenvalue equation @xmath493 can again be constructed with the transfer matrices .
indeed , every eigenvector of @xmath494 to the eigenvalue @xmath27 allows to construct a periodic eigenvector so that @xmath495 hence one needs to find those @xmath426 for which @xmath27 is eigenvalue of @xmath496 . by considering the fact that the graph of @xmath496 is also lagrangian w.r.t .
an adequate quadratic form , this can be done by means of intersection theory in a similar manner as in the previous section .
several notations need to be introduced .
let us associate to @xmath497 the @xmath498 matrix @xmath499 where the @xmath500 denotes the checker board sum given by @xmath501 this allows to define the quadratic form @xmath502 .
for @xmath426 the matrices @xmath503 conserve the quadratic form @xmath504
. therefore also @xmath504-lagrangian planes are mapped onto @xmath504-lagrangian planes . by proposition [ prop - lagrangian ] , the stereographic projection maps the set @xmath505 of @xmath504-lagrangian planes is diffeomophically to @xmath506 .
let us denote the stereographic projection by @xmath507 .
[ prop - ttou0 ] to a given @xmath508 , let us associate the unitary @xmath509_\sim)^*\ ; \widehat{\pi}([\widehat{\psi}_0]_\sim ) \;\in\ ; \mbox{\rm u}(2l ) \;,\ ] ] where the @xmath504-lagrangian plane is given by the @xmath510 matrix @xmath511 then @xmath512 * proof .
* first of all , it can readily be checked that @xmath513 .
now let us suppose that this frame and the lagrangian frame @xmath514 have a non - trivial intersection .
this means that there exist vectors @xmath515 such that such @xmath516 .
the first and third line of this vector equality imply @xmath517 and @xmath518 , the other two that @xmath519 .
this shows @xmath520 but now proposition [ prop - intersection ] can be applied to calculate the r.h.s . and completes the proof .
@xmath48 it is now natural to introduce the following @xmath504-lagrangian frames : @xmath521 with @xmath522 as in .
associated are then the unitaries @xmath523_\sim\bigr)^*\ ; \widehat{\pi}\bigl ( [ \widehat{\psi}_0]_\sim\bigr ) \;.\ ] ] [ theo - osci2 ] the multiplicity of @xmath524 as eigenvalues of @xmath491 is equal to the multiplicity of @xmath27 as eigenvalue of @xmath525 . moreover , @xmath526 so that the eigenvalues of @xmath527 rotate around the unit circle in the positive sense and with non - vanishing speed as function of @xmath475 . * proof . *
the first claim follows directly from proposition [ prop - ttou0 ] . for
the proof of the second one , let us denote the upper and lower entries of @xmath528 by @xmath529 and @xmath530 .
these are @xmath51 matrices such that @xmath531_\sim\bigr)=\psi_-^z(\psi_+^z)^{-1}=((\psi_-^z)^{-1})^*(\psi_+^z)^*$ ]
. then @xmath532_\sim\bigr)^*\ ; ( ( \psi_+^z)^{-1})^ * \bigl[\ , ( \psi_+^z)^*\partial_\theta\psi_+^z \,-\ , ( \psi_-^z)^*\partial_\theta\psi_-^z \,\bigr ] ( \psi_+^z)^{-1 } \;\widehat{\pi}\bigl ( [ \widehat{\psi}_0]_\sim\bigr ) \;.\ ] ] thus it is sufficient to check positive definiteness of @xmath533 \;=\ ; \imath \ ; ( \widehat{\psi}^z_n)^*\,\widehat{{{\cal l}}}\,\partial_\theta \widehat{\psi}^z_n \;.\ ] ] from the product rule follows that @xmath534 this implies that @xmath535 as @xmath536 and the matrices @xmath537 do not mix first and third columns and lines with the rest , it follows from evaluation in the state @xmath522 that @xmath538 but positivity of @xmath539 was already checked in the proof of theorem [ theo - oscillation ] .
this completes the proof .
in this section , we consider a two - sided infinite scattering zipper @xmath540 defined on @xmath541 which are @xmath5-periodic where again @xmath5 is even .
it is specified by a sequence of scattering matrices @xmath330
satisfying @xmath542 for all @xmath543 . one can partially diagonalize such periodic operators by the bloch - floquet transform defined next .
the bloch - floquet transform @xmath544 is defined by @xmath545 where @xmath546 $ ] and @xmath547 .
its inverse @xmath548 is given by @xmath549 [ prop - fourier ] one has the following properties : \(i ) @xmath550 , namely @xmath551 is unitary .
\(ii ) let @xmath552 be the shift on @xmath553 defined by @xmath554 and let @xmath555 be the cyclic shift on @xmath556
. then @xmath557 * proof .
* this follows from direct computations .
@xmath48 for any @xmath558 and
@xmath559 let us set @xmath560 now @xmath561 and @xmath562 are defined as section [ sec - periodic ] using @xmath563 instead of @xmath564 and then @xmath565 is a finite periodic scattering zipper , the spectrum of which can be calculated by the technique of section [ sec - periodic ] . by the following result this allows to calculate the spectrum of @xmath540 .
[ theo - oscillationper ] the operators @xmath566 , @xmath567 and @xmath568 are fibered after bloch - floquet transformation , precisely @xmath569 therefore , @xmath570 * proof . * by definition , for @xmath571 : @xmath572 in case @xmath573 , one obtains from the structure of @xmath574 : @xmath575 by the same calculation for @xmath576 , @xmath577 together this shows @xmath578 the second equality is proved using proposition [ prop - fourier](ii ) and previous computation : @xmath579 as the product of fibered operators is fibered , this also implies the formula for @xmath580 . @xmath48
this appendix resembles some basic properties of the mbius transformation as they are used in the main text
. a lot of references to the literature can be found in @xcite .
the mbius transformation ( also called canonical transformation or fractional transformation ) is defined by @xmath581 whenever the appearing inverse exists .
if @xmath497 is @xmath582-unitary and @xmath583 , then @xmath584 exists and is in @xmath585 ( see appendix b for the definition of the upper half - plane @xmath585 ) . for @xmath497 as in and as long as the appearing inverse exists , the inverse mbius transformation is defined by @xmath586 the mbius transformation is a left action , namely @xmath587 as long as all objects are well - defined .
the inverse mbius transformation is a right action in the sense of the following proposition , the algebraic proof of which is left to the reader .
[ prop - moebinv ] under the condition that all the mbius und inverse mbius transformations as well as matrix inverses below exist , one has the following properties .
\(i ) @xmath588 \(ii ) @xmath589 \(iii ) @xmath590
let @xmath585 denote the upper half plane of matrices @xmath591 such that @xmath592 .
it is well - known ( _ e.g. _ section 4.5 . of @xcite ) that the cayley transform maps @xmath585 via mbius transformation to the siegel disc @xmath214 , namely @xmath593 .
an analytic function @xmath594 is called a herglotz function .
then @xmath595 is an analytic function on the unit disc @xmath596 having positive imaginary part . if , moreover , @xmath111 , then such a function is called a caratheodory function .
the scalar version of the following classical theorem can be found in text books such as @xcite .
the matrix version is an immediate corollary of it .
( riesz - herglotz representation theorem ) let @xmath597 be analytic satisfying @xmath598 and @xmath111 .
then there exists a unique matrix - valued probability measure @xmath338 on @xmath339 such that @xmath599 s. clark , f. gesztesy , m. zinchenko , _ weyl - titchmarsh theory and borg - marchenko - type uniqueness results for cmv operators with matrix - valued verblunsky coefficients _ , oper .
matrices * 1 * , 535 - 592 ( 2007 ) . | a scattering zipper is a system obtained by concatenation of scattering events with equal even number of incoming and out going channels .
the associated scattering zipper operator is the unitary equivalent of jacobi matrices with matrix entries and generalizes blatter - browne and chalker - coddington models and cmv matrices .
weyl discs are analyzed and used to prove a bijection between the set of semi - infinite scattering zipper operators and matrix valued probability measures on the unit circle .
sturm - liouville oscillation theory is developed as a tool to calculate the spectra of finite and periodic scattering zipper operators . |
cosmological and astrophysical probes suggest that @xmath023% of the universe is composed of non - baryonic dark matter ( dm ) , commonly assumed to be in the form of weakly interacting massive particles ( wimps ) arising in extensions of the standard model of particle physics ( for reviews see , e.g. @xcite ) . amongst
the most widely discussed dm candidates are the lightest neutralino in supersymmetric extensions of the standard model @xcite and the first excitation of the kaluza - klein bosons ( lkp ) in universal extra dimension theories.@xcite the annihilation of wimp pairs can produce in the final state a continuum of gamma - rays whose flux extends up to the dm particle mass , from the hadronization and decay of the cascading annihilation products . in supersymmetric models ,
the gamma - ray spectrum from neutralino annihilation is not uniquely determined and the branching ratios ( brs ) of the open annihilation channels are not determined since the dm particle field content is not known _
a priori_. in contrast , in kaluza - klein scenarios where the lightest kaluza - klein particle ( lkp ) is the first kk mode of the hypercharge gauge boson , the brs of the annihilation channels can be computed given that the field content of the dm particle is known .
the gamma - ray flux from annihilations of dm particles of mass @xmath1 accumulating in a spherical dm halo can be expressed in the form : @xmath2 as a product of a particle physics component with an astrophysics component .
the particle physics part contains @xmath3 , the velocity - weighted annihilation cross section , and @xmath4 , the differential gamma - ray spectrum summed over the whole final states with their corresponding branching ratios .
the astrophysical part corresponds to the line - of - sight - integrated squared density of the dm distribution j , averaged over the instrument solid angle @xmath5 usually matching the angular resolution of the instrument : @xmath6)ds \hspace{2 cm } \bar{j}(\delta\omega)\,=\,\frac{1}{\delta\omega}\,\int_{\delta\omega } \,psf\times j\ , d\omega\ ] ] where @xmath7 stands for the point spread function of the instrument .
the annihilation rate being proportional to the square of the dm density integrated along the line of sight , regions with enhanced dm density are primary targets for indirect searches . among them are the galactic halo , external galaxies , galaxy clusters , substructures in galactic haloes , and the galactic center .
we report here on recent results on dark matter searches with current iacts such as h.e.s.s . and magic , towards the galactic center , dwarf galaxies from the local group , galactic globular clusters and dm mini - spikes around intermediate mass black holes ( imbhs ) .
h.e.s.s . observations towards the galactic center have revealed a bright pointlike gamma - ray source , hess j1745 - 290 @xcite , coincident in position with the supermassive black hole sgr a * , with a size lower than 15 pc .
diffuse emission along the galactic plane has also been detected @xcite and correlates well with the mass density of molecular clouds from the central molecular zone , as traced by cs emission @xcite . according to recent detailed studies @xcite
, the source position is located at an angular distance of [email protected]@xmath98.5@xmath10 from sgr a*. the pointing accuracy allows to discard the association of the very high energy ( vhe ) emission with the center of the radio emission of the supernova remnant sgr a east but the association with the pulsar wind nebula g359.95 - 0.04 can not be ruled out . from 2004 data
set , the energy spectrum of the source is well fitted in the energy range 160 gev - 30 tev to a power - law spectrum dn / de @xmath11 e@xmath12 with a spectral index @xmath13 = [email protected]@[email protected]@xmath16 .
no deviation from a power - law is observed leading to an upper limit on the energy cut - off of 9 tev ( 95% c.l . ) .
the vhe emission from hess j1745 - 290 does not show any significant periodicity or variability from 10 minutes to 1 year @xcite .
besides plausible astrophysical origins ( see e.g. @xcite and references therein ) , an alternative explanation is the annihilation of dm in the central cusp of our galaxy .
the spectrum of hess j1745 - 290 shows no indication for gamma - ray lines .
the observed gamma - ray flux may also result from secondaries of dm annihilation .
the left hand side of fig .
[ aba : gc]shows the h.e.s.s . spectrum extending up to masses of about 10 tev , which requires large neutralino masses ( @xmath1710 tev ) .
they are unnatural in phenomenological mssm scenarios .
the kaluza - klein models provide harder spectra which still significantly deviate from the measured one .
non minimal version of the mssm may yield flatter spectrum with mixed 70% @xmath18 and 30% @xmath19 final states .
even this scenario does not fit to the measured spectrum .
the hypothesis that the spectrum measured by h.e.s.s .
originates only from dm particle annihilations is highly disfavored .
the magic observations were carried out towards the galactic centre since 2004 and revealed a strong emission .
the observed excess in the direction of the gc has a significance of 7.3 standard deviations @xcite and is compatible with a pointlike source .
large zenith observation angle ( @xmath2060@xmath21 ) implies an energy threshold of @xmath0400 gev . the source position and the flux level are consistent with the measurement of hess @xcite within errors . the right hand side of fig .
[ aba : gc ] shows the reconstructed vhe gamma - ray energy spectrum of the gc after the unfolding with the instrumental energy resolution
. the differential flux can be well described by a power law of index @xmath22 .
the systematic error is estimated to be 35% in the flux level determination .
the flux level is steady within errors in the time - scales explored within these observations , as well as in the two year time - span between the magic and hess observations .
an interpretation in term of dm require a minimum value of mass higher than 10 tev .
most probably , if dm signal exists is overcome by other astrophysical emitters .
dwarf spheroidal galaxies in the local group are considered as privileged targets for dm searches since they are among the most extreme dm - dominated environments .
measurements of roughly constant radial velocity dispersion of stars imply large mass - to - luminosity ratios .
nearby dwarfs are ideal astrophysical probes of the nature of dm as they usually consist of a stellar population with no hot or warm gas , no cosmic ray population and little dust .
indeed , these systems are expected to have a low intrinsic gamma - ray emission .
this is in contrast with the galactic center where disentangling the dominant astrophysical signal from possible more exotic one is very challenging .
prior to the year 2000 , the number of known satellites was eleven . with the sloan digital sky survey ( sdss ) , a population of ultra low - luminosity satellites has been unveiled , which roughly doubled the number of known satellites . among them
are coma berenices , ursa major ii and willman 1 .
iacts have started observation campaigns on dwarf galaxies for a few years .
table [ aba : tbl1 ] presents a possible list of the preferred targets for dm searches .
@xmath23see ref . .
@xmath24see ref . .
@xmath25see ref . .
+ [ aba : tbl1 ] the star velocity dispersions in draco reveal that this object is dominated by dm on all spatial scales and provide robust bounds on its dm profile , which thus decreases uncertainties on the astrophysical contribution to the gamma - ray flux .
the magic collaboration searched for a steady gamma - ray emission from the direction of draco @xcite .
the analysis energy threshold after cuts is 140 gev .
no significant excess is found . for a power law with spectral index of 1.5 , typical for a dm annihilation spectrum , and assuming a pointlike source , the 2@xmath26 upper limit is @xmath27 .
the measured flux upper limit is several orders of magnitude larger than predicted for the smooth dm distribution in msugra models .
the limit on the flux enhancement caused by high clumpy substructures or a black hole is around @xmath28 .
the whipple collaboration has also observed draco @xcite .
[ aba : globularcluster ] shows the 95% c.l .
exclusion curve on @xmath29 .
the upper limit is at the level of 10@xmath30 @xmath31s@xmath32 for 1 tev neutralino annihilating with brs of 90% in @xmath18 and 10% in @xmath19 .
the sagittarius ( sgr ) dwarf galaxy , one of the nearest galaxy satellites of the local group , has been observed by h.e.s.s . since 2006 . the annihilation signal from sgr is expected to come from a region of 1.5 pc , which is much smaller than the h.e.s.s .
point spread function .
thus , a pointlike signal has been searched for .
no significant gamma - ray excess is detected at the nominal target position .
a 95% c.l .
upper limit on the gamma - ray flux is derived : @xmath33 , assuming a power - law spectrum of spectral index 2.2 @xcite .
a modelling of the sgr dm halo has been carried out .
two models of the mass distribution for the dm halo have been studied : a cusped nfw profile and a cored isothermal profile , to emcompass a large class of profiles .
the cored profile has a small core radius due to a cusp in the luminous profile .
the value of the line - of - sight - integrated squared density is then found to be larger for the cored profile than for the nfw profile ( see ref . for more details ) .
the left hand side of fig .
[ aba : dwarf ] presents the constraints on the velocity - weighted annihilation cross section @xmath29 for a cusped nfw and cored profiles in the solid angle integration region @xmath34 sr , for neutralino dm .
predictions for susy models are displayed .
for a cusped nfw profile , h.e.s.s . does not set severe constraints on @xmath29 . for a cored profile , due to a higher central density ,
stronger constraints are derived and some pmssm models can be excluded in the upper part of the scanned region . although the nature of canis major is stilll debated , hess observed this object and results on canis major @xcite are presented in ref . .
the recently discovered willman 1 dwarf galaxy has been observed by magic @xcite .
willman 1 has a a total mass of @xmath05@xmath3510@xmath36 m@xmath37 , which is about an order of magnitude smaller than those of the least massive satellite galaxies previously known .
this object has one of the highest mass - to - luminosity ratio .
no significant gamma - ray excess beyond 100 gev above the background was observed in 15.5 hours of observation of the sky region around willman 1 .
this is shown in the right hand side of fig .
[ aba : dwarf ] , where the @xmath38-plot@xmath39 is reported -parameter is the angular distance between the shower image main axis and the line connecting the image barycenter and the camera center . ] .
the signal is searched with a cut slightly larger than for a pointlike source to take into account a possible source extension which makes a fiducial region to be @xmath40 @xmath4112@xmath21 .
flux upper limits of the order of 10@xmath42 @xmath43s@xmath32 for benchmark msugra models are obtained @xcite .
boost factors in flux of the order of 10@xmath44 are required in the most optimistic scenarios , i.e. in the funnel region or in region where the internal bremsstrahlung may play an important role . however
, uncertainties on the dm distribution or the role of substructures may significantly reduce this boost .
globular clusters are not believed to be dm - dominated objects .
their mass - to - luminosity ratios are well described by purely king profile which suggets no significant amount of dm .
however , the formation of globular clusters fits in the hierarchical strucure formation scenario in which globular clusters may have formed in dm overdensities . during their evolution
, they may have hold some dm in their central region .
m15 is a nearby galactic globular cluster with a core radius of 0.2 pc and a central density of @xmath010@xmath45m@xmath37pc@xmath46 .
even if the mass - to - light ratio of m15 does not suggest a significant component of dm , the compact and dense core of m15 does not prevent from looking for a hypothetical dm signal .
the presence of stars and dm in globular clusters may lead to a dm enhancement in their inner part .
this is usually treated with the adiabatic contraction model @xcite .
the whipple collaboration obsered m15 in 2003 .
the modelling of dm profile assuming adiabatic contraction of dm from an initial nfw distribution leads to an enhancement in the astrophysical factor of @xmath010@xmath47 to 10@xmath44 .
[ aba : globularcluster ] shows the exclusion curve on @xmath48 ( dotted line ) versus the neutralino mass assuming a nfw profile after adiabatic contraction @xcite .
the limit reaches @xmath010@xmath49 @xmath31s@xmath32 for tev neutralinos .
astrophysical systems such as the virgo galaxy cluster @xcite have been considered as target for dm annihilation searches .
the elliptical galaxy m87 at the center of the virgo cluster has been observed by h.e.s.s .
it is an active galaxy located at 16.3 mpc which hosts a black hole of 3.2@xmath3510@xmath50m@xmath37 . from 2003 to 2006 ,
h.e.s.s . detected a 13@xmath26 signal in 89 hour observation time at a location compatible with the nominal position of the nucleus of m87 @xcite . with the angular resolution of h.e.s.s .
, the vhe source is pointlike , with an upper limit on the size of 3 at 99% confidence level .
given the distance of m87 , this corresponds to a size of 13.7 kpc .
[ aba : m87 ] shows the energy spectrum for 2004 and 2005 data .
both data sets are well fitted to a power - law spectrum with spectral indices
@xmath13 = [email protected] ( 2004 ) and @xmath13 = [email protected] ( 2005 ) .
the 2005 average flux is found to be higher by a factor @xmath05 to the average flux of 2004 .
the integrated flux above 730 gev shows a yearly variability at the level of 3.2@xmath26 @xcite .
variability at the day time scale has been observed in the 2005 data at the level of 4@xmath26 .
this fast variability puts stringent constraints on the size of the vhe emission region .
the position centered on m87 nucleus excludes the center of the virgo cluster and outer radio regions of m87 as the gamma - ray emission region .
the observed variability at the scale of @xmath02 days requires a compact emission region below 50r@xmath51 , where r@xmath5210@xmath53 cm is the schwarzschild radius of the m87 supermassive black hole @xcite .
the short and long term temporal variability observed with h.e.s.s .
excludes the bulk of the tev gamma - ray signal to be of dm origin .
mini - spikes around intermediate mass black holes have been recently proposed as promising targets for indirect dark matter detection @xcite .
the growth of massive black holes inevitably affects the surrounding dm distribution .
the profile of the final dm overdensity , called mini - spike , depends on the initial distribution of dm , but also on astrophysical processes such as gravitational scattering of stars and mergers . ignoring astrophysical effects , and assuming adiabatic growth of the black hole , if one starts from a nfw profile , a spike with a power - law index 7/3 is obtained , as relevant for the astrophysical formation scenario studied here characterized by black hole masses of @xmath010@xmath36 m@xmath37 @xcite .
mini - spikes might be detected as bright pointlike sources by current iacts .
data collected between 2004 and 2007 during the galactic plane survey have allowed to accurately map the galactic plane between @xmath143@xmath21 in galactic latitude and from -30@xmath21 to 60@xmath21 in galactic longitude with respect to the galactic center position .
the study of the h.e.s.s .
sensitivity in a large field of view to dark matter annihilations has been performed @xcite .
[ aba : sensitivity_hess ] shows the experimentally observed sensitivity map in the galactic plane from galactic longitudes l=-30@xmath21 to l=+60@xmath21 and galactic latitudes b=-3@xmath21 to b=+3@xmath21 , for a dm particle of 500 gev mass annihilating into the 100% br @xmath18 channel .
the h.e.s.s .
sensitivity depends strongly on the exposure time and acceptance maps which are related to the choice of the pointing positions .
the flux sensitivity varies along the latitude and longitude due to inhomogeneous coverage of the galactic plane . in the band between -2@xmath21 and 2@xmath21 in galactic latitude ,
a dm annihilation flux sensitivity at the level of 10@xmath42 @xmath43s@xmath32 is achieved for a 500 gev dm particle annihilating in the @xmath18 channel .
deeper observations of the galactic center and at galactic longitude of @xmath020@xmath21 allow the flux sensitivity to be of @xmath05@xmath3510@xmath54 @xmath43s@xmath32 . for b@xmath202@xmath21 , the sensitivity is deteriorated due to a weaker effective exposure . for b=0@xmath21 and l=-0.5@xmath21 ,
the flux sensitivity is @xmath010@xmath54 @xmath43s@xmath32 in @xmath18 .
reached the required sensitivity to be able to test dm annihilations from mini - spikes in the context of one relatively favorable scenario for imbh formation and adiabatic growth of the dm halo around the black hole ( e.g. scenario b of ref . ) .
observations ( 2004 - 2006 ) of the galactic plane allowed to discover more than 20 very high energy gamma - ray sources @xcite .
some of them have been identified owing to their counterparts at other wavelengths , but almost half of the sources have no obvious counterpart and are still unidentified @xcite .
an accurate reconstruction of their energy spectra shows that all the spectra are consistent with a pure power - law , spanning up to two orders of magnitude in energy above the energy threshold .
none of them exhibits an energy cutoff , characteristic of dm annihilation spectra , in the energy range from @xmath0100 gev up to 10 tev .
furthermore , the detailed study of their morphology @xcite shows that all the sources have an intrinsic spatial extension greater than @xmath05 arcminutes , while mini - spikes are expected to be pointlike sources for h.e.s.s .
no imbh candidate has been detected so far by h.e.s.s . within the survey range .
based on the absence of plausible imbh candidates in the h.e.s.s .
data , constraints are derived on the scenario b of ref . for neutralino or lkp annihilations , shown as upper limits on @xmath29 @xcite .
[ aba : imbh_hess ] shows the exclusion limit at the 90% c.l . on @xmath29 as a function of the neutralino mass .
the neutralino is assumed to annihilate into @xmath18 and @xmath19 with 100% br , respectively .
predictions for susy models are also displayed .
the limits on @xmath29 are at the level of 10@xmath55 cm@xmath44s@xmath32 for the @xmath18 channel for neutralino masses in the tev energy range .
limits are obtained one mini - spike scenario and constrain on the entire gamma - ray production scenario .
dark matter searches will continue and searches with the phases 2 of h.e.s.s . and
magic will start soon .
the phase 2 of h.e.s.s . will consist of a new large 28 m diameter telescope located at the center of the existing array , and will allow to lower the analysis energy threshold down to less than 50 gev .
the installation of the second telescope on the magic site will allow to perform more sensitive searches by the use of the stereoscopic mode .
the upcoming generation of iacts , the cherenkov telescope array , is in the design phase .
this array composed of several tens of telescopes will permit to significantly improve the performances on both targeted dm searches and wide - field - of - view dm searches .
99 g. bertone , d. hooper and j. silk , _ phys .
rept . _ * 405 * , 279 ( 2005 ) .
l. bergstrom , _ rept .
phys . _ * 63 * , 793 ( 2000 ) .
g. jungman , m. kamionkowski and k. griest , _ phys .
* 267 * , 195 ( 1996 ) t. appelquist , h .- c .
cheng and b. a. dobrescu , phys .
d * 64 * , 035002 ( 2001 ) .
f. aharonian _ et al .
_ [ h.e.s.s .
collaboration ] , _ phys .
lett . _ * 97 * , 221102 ( 2006 ) [ erratum - ibid . *
97 * , 249901 ( 2006 ) ] .
f. aharonian _ et al . _
[ h.e.s.s .
collaboration ] , _ nature _ * 439 * , 695 ( 2006 ) .
m. tsuboi , h. tsohihiro and n. ukita , _ astrophys . j. suppl .
_ 120 ( 2006 ) 675 .
icrc 2007 contributions , arxiv:0710.4057 [ astro - ph ] .
f. aharonian and a. neronov , _ astrophys .
j. _ * 619 * , 306 ( 2005 ) .
j. albert _ et al .
_ [ magic collaboration ] , _ astrophys . j. _ * 638 * , l101 ( 2006 ) . f. aharonian _ et al . _
[ hess collaboration ] , _ astron .
astrophys . _ * 425 * , l13 ( 2004 ) .
m. mateo , _ ann .
astrophys . _ * 36 * , 435 ( 1998 ) .
j. d. simon and m. geha , _ astrophys .
j. _ * 670 * , 313 ( 2007 ) .
a. helmi and s. d. m. white , _ mon . not .
soc . _ * 323 * , 529 ( 2001 ) .
j. albert _ et al .
_ [ magic collaboration ] , _ astrophys . j. _ * 679 * , 428 ( 2008 ) .
m. wood _ et al .
_ , arxiv:0801.1708 [ astro - ph ] . f. aharonian _ et al .
_ [ hess collaboration ] , _ astropart . phys . _ * 29 * , 55 ( 2008 ) .
f. aharonian _ et al . _
[ hess collaboration ] , _ astrophys .
j. _ * 691 * , 175 ( 2009 ) .
m. vivier for the h.e.s.s .
collaboration , _ these proceedings_. e. aliu _ et al .
_ [ magic collaboration ] , arxiv:0810.3561 [ astro - ph ] . g. r. blumenthal , s. m. faber , r. flores and j. r. primack , _ astrophys . j. _ * 301 * , 27 ( 1986 ) .
e. a. baltz , c. briot , p. salati , r. taillet and j. silk , _ phys .
_ * d 61 * , 023514 ( 2000 ) .
d. berge _ et al .
_ , _ science _ * 314 * , 1424 ( 2006 ) .
g. bertone , a. r. zentner and j. silk , _ phys . rev . _ * d 72 * , 103517 ( 2005 ) . f. aharonian _ et al . _
[ hess collaboration ] , _ phys .
* d 78 * , 072008 ( 2008 ) .
f. aharonian _ et al .
_ [ hess collaboration ] , _ astrophys . j. _ * 636 * , 777 ( 2006 ) .
f. aharonian _ et al .
_ [ hess collaboration ] , _ astron . astrophys . _ * 477 * , 353 ( 2008 ) . | the annihilations of wimps produce high energy gamma - rays in the final state .
these high energy gamma - rays may be detected by imaging atmospheric cherenkov telescopes ( iacts ) . amongst
the plausible targets are the galactic center , the centre of galaxy clusters , dwarf sphreroidal galaxies and substructures in galactic haloes .
i will review on the recent results from observations of ongoing iacts . |
the swift progress in molecular electronics achieved during the past decade has mostly been centered around a detailed understanding of charge transport through single - molecule junctions,@xcite where quantum effects generally turn out to be important .
when two _ superconducting _ ( instead of normal ) electrodes with the same chemical potential but a phase difference @xmath1 are attached to the molecule , the josephson effect@xcite implies that an equilibrium current @xmath2 flows through the molecular junction . over the past decade ,
experiments have observed gate - tunable josephson currents through nanoscale junctions , @xcite including out - of - equilibrium cases , and many different interesting phenomena have been uncovered .
in particular , the current - phase relation has been measured by employing a superconducting quantum interference device.@xcite for weakly coupled electrodes , the current - phase relation is@xcite @xmath3 with the _ critical current _ @xmath4 .
the above questions have also been addressed by many theoretical works .
it has been shown that the repulsive electron - electron ( e - e ) interaction @xmath5 , acting on electrons occupying the relevant molecular level , can have a major influence on the josephson current.@xcite for intermediate - to - strong coupling to the electrodes , an interesting interplay between the kondo effect and superconductivity takes place.@xcite in the present paper , we address the opposite limit where , for sufficiently large @xmath6 , a so - called @xmath0-phase can be realized , with @xmath7 in eq .
( [ cpr ] ) . in the @xmath0-regime
, @xmath8 corresponds to the ground state of the system ( or to a minimum of the free energy for finite temperature ) , in contrast to the usual @xmath9-state with @xmath10 , where @xmath11 in the ground state.@xcite the sign change in @xmath4 arises due to the blocking of a direct cooper pair exchange when @xmath6 is large .
double occupancy on the molecular level is then forbidden , and the remaining allowed processes generate the sign change in @xmath4.@xcite the most natural way to explain the @xmath0-junction behavior is by perturbation theory in the tunnel couplings connecting the molecule to the electrodes .
experimental observations of the @xmath0-phase were recently reported for inas nanowire dots@xcite and for nanotubes,@xcite but a @xmath0-junction is also encountered in superconductor - ferromagnet - superconductor structures.@xcite accordingly , theoretical works have also analyzed spin effects in molecular magnets coupled to superconductors.@xcite the impressive experimental control over supercurrents through molecular junctions reviewed above implies that modifications of the supercurrent due to vibrational modes of the molecule play a significant and observable role.@xcite we have recently discussed how a _ two - level system _ ( tls ) coupled to the dot s charge is affected by the josephson current carried by andreev states.@xcite for instance , two conformational configurations of a molecule may realize such a tls degree of freedom .
experimental results for molecular break junctions with normal leads were interpreted using such models , @xcite but the tls can also be created artifically using a coulomb - blockaded double dot.@xcite a detailed motivation for our model , where the pauli matrix @xmath12 in tls space couples to the dot s charge , and its experimental relevance has been given in refs . .
while our previous work@xcite studied the josephson - current - induced switching of the tls , we here address a completely different parameter regime characterized by weak coupling to the electrodes , and focus on the josephson current itself .
we calculate the critical current @xmath4 in eq .
( [ cpr ] ) using perturbation theory in these couplings , allowing for arbitrary e - e interaction strength and tls parameters .
a similar calculation has been reported recently,@xcite but for a harmonic oscillator ( phonon mode ) instead of the tls .
our predictions can be tested experimentally in molecular break junctions using a superconducting version of existing@xcite setups .
the remainder of this paper has the following structure . in sec .
[ sec2 ] , we discuss the model and present the general perturbative result for the critical current . for tunnel matrix element @xmath13 between the two tls states ,
the result allows for an elementary interpretation , which we provide in sec .
the case @xmath14 is then discussed in sec
. [ sec4 ] , followed by some conclusions in sec .
technical details related to sec . [ sec3 ] can be found in an appendix .
we mostly use units where @xmath15 .
we study a spin - degenerate molecular dot level with single - particle energy @xmath16 and on - site coulomb repulsion @xmath5 , coupled to the tls and to two standard @xmath17-wave bcs superconducting banks ( leads ) .
the tls is characterized by the ( bare ) energy difference @xmath18 of the two states , and by the tunnel matrix element @xmath19 .
the model hamiltonian studied in this paper is motivated by refs . where it was employed to successfully describe break junction experiments ( with normal - state leads ) .
it reads @xmath20 where the coupled dot - plus - tls part is @xmath21 with the occupation number @xmath22 for dot fermion @xmath23 with spin @xmath24 .
note that the tls couples with strength @xmath25 to the dot s charge .
indeed , assuming some reaction coordinate @xmath26 describing molecular conformations , the dipole coupling to the dot is @xmath27 , just as for electron - phonon couplings.@xcite if the potential energy @xmath28 is bistable , the low - energy dynamics of @xmath26 can be restricted to the lowest quantum state in each well and leads to eq . .
the tls parameters and the dipole coupling energy @xmath25 can be defined in complete analogy to refs . , and typical values for @xmath25 in the mev range are expected , comparable to typical charging energies @xmath6 .
moreover , the electron operators @xmath29 , corresponding to spin-@xmath17 and momentum-@xmath30 states in lead @xmath31 , are governed by a standard bcs hamiltonian with complex order parameter @xmath32 ( with @xmath33 ) , respectively , @xmath34 where @xmath35 is the ( normal - state ) dispersion relation . finally , the tunneling hamiltonian is @xmath36 where @xmath37 describes tunneling of an electron with spin @xmath17 from the dot to lead @xmath38 with tunnel amplitude @xmath39 , and the reverse process is generated by @xmath40 .
the josephson current @xmath2 at temperature @xmath41 follows from the equilibrium ( imaginary - time ) average , @xmath42 where @xmath31 and @xmath24 can be chosen arbitrarily by virtue of current conservation and spin-@xmath43 invariance , and @xmath44 is the time - ordering operator .
equation ( [ jos1 ] ) is then evaluated by lowest - order perturbation theory in @xmath45 .
the leading contribution is of fourth order in the tunnel matrix elements and can be evaluated in a similar manner as in ref . .
we assume the usual wide - band approximation for the leads with @xmath30-independent tunnel matrix elements , and consider temperatures well below both bcs gaps , @xmath46 .
putting @xmath47 and @xmath48 , after some algebra , the josephson current takes the form ( [ cpr ] ) with the critical current @xmath49 we define the hybridizations @xmath50 , with ( normal - state ) density of states @xmath51 in the leads . the function @xmath52 in eq .
( [ ic ] ) can be decomposed according to @xmath53 with contributions @xmath54 for fixed dot occupation number @xmath55 . for given @xmath56 , the two eigenenergies ( labeled by @xmath57 ) of the dot - plus - tls hamiltonian @xmath58 in eq .
( [ h0 ] ) are @xmath59 with the scale @xmath60 the occupation probability for the state @xmath61 is @xmath62 where @xmath63 ensures normalization , @xmath64 . with the propagator @xmath65
we then find the contributions @xmath66 in eq .
( [ cn ] ) , @xmath67,\end{aligned}\ ] ] @xmath68,\end{aligned}\ ] ] @xmath69.\end{aligned}\ ] ] here , we have used the matrix elements @xmath70 with the @xmath71 matrices ( in tls space ) @xmath72 for @xmath73 , it can be shown that @xmath74 and @xmath75 are always positive , while @xmath76 yields a negative contribution to the critical current .
when @xmath77 outweighs the two other terms , we arrive at the @xmath0-phase with @xmath78 .
below , we consider identical superconductors , @xmath79 , and assume @xmath80 .
it is useful to define the reference current scale @xmath81 within lowest - order perturbation theory , the hybridizations @xmath82 and @xmath83 only enter via eq .
( [ i0 ] ) and can thus be different . equation ( [ ic ] ) provides a general but rather complicated expression for the critical current , even when considering the symmetric case @xmath84 . in the next section
, we will therefore first analyze the limiting case @xmath85 .
when there is no tunneling between the two tls states , @xmath13 , the hilbert space of the system can be decomposed into two orthogonal subspaces @xmath86 , with the fixed conformational state @xmath57 in each subspace .
equation ( [ h0eig ] ) then simplifies to @xmath87 one thus arrives at two decoupled copies of the usual interacting dot problem ( without tls ) , but with a shifted dot level @xmath88 and the `` zero - point '' energy shift @xmath89 . as a result , the critical current @xmath4 in eq .
( [ ic ] ) can be written as a weighted sum of the partial critical currents @xmath90 through an interacting dot level ( without tls ) at energy @xmath91 , @xmath92 where @xmath93 with eqs .
( [ pns ] ) and ( [ ensigma ] ) denotes the probability for realizing the conformational state @xmath94 .
the current @xmath95 has already been calculated in ref .
( in the absence of phonons ) , and has been reproduced here . in order to keep the paper
self - contained , we explicitly specify it in the appendix . in order to establish the relevant energy scales determining the phase diagram
, we now take the @xmath73 limit .
then the probabilities ( [ pns ] ) simplify to @xmath96 , where @xmath97 is the ground - state energy of @xmath98 for @xmath85 . depending on the system parameters
, the ground state then realizes the dot occupation number @xmath99 and the tls state @xmath100 .
the different regions @xmath101 in the @xmath102 plane are shown in the phase diagram in fig .
the corresponding critical current in each of these regions is then simply given by @xmath103 .
ground - state phase diagram in the @xmath102 plane for @xmath13 .
different regions @xmath104 are labelled according to the ground - state dot occupation number @xmath105 and the conformational state @xmath106 .
dark areas correspond to @xmath0-junction behavior .
the charge - degeneracy line @xmath107 is indicated as dashed line .
main panel : @xmath108 .
inset : @xmath109 , where no @xmath0-junction behavior is possible for @xmath110 .
, scaledwidth=40.0% ] by analyzing the dependence of the ground - state energy on the system parameters , one can always ( even for @xmath111 ) write the function @xmath112 in eq .
( [ ic ] ) as @xmath113 where @xmath114 is the heaviside function and the energies @xmath115 are the boundaries enclosing the @xmath0-phase region with @xmath116 , i.e. , @xmath117 ( @xmath118 ) denotes the boundary between the @xmath119 and @xmath116 ( the @xmath116 and @xmath120 ) regions , see fig . [ fig1 ] .
explicit results for @xmath121 follow from eq .
( [ ensigma ] ) for @xmath85 . for @xmath122 ( @xmath123 ) and arbitrary @xmath124 ,
the ground state is realized when @xmath125 @xmath126 , leading to @xmath127 ( @xmath128 ) . in both cases ,
the other boundary energy follows as @xmath129 . in the intermediate cases , with @xmath130 we find for @xmath131 , @xmath132 while for @xmath133
, we obtain @xmath134 these results for @xmath121 are summarized in fig .
remarkably , in the @xmath102 plane , the phase diagram is inversion - symmetric with respect to the point @xmath135 .
furthermore , we observe that for many choices of @xmath136 , one can switch the tls between the @xmath137 states by varying @xmath138 , see fig .
[ fig1 ] .
( color online ) ground - state critical current @xmath139 as a function of @xmath138 for @xmath85 .
@xmath139 is given in units of @xmath140 , see eq .
( [ i0 ] ) . in all figures ,
the energy scale is set by @xmath141 . dashed ( red ) , dotted ( blue ) and solid ( black ) curves represent the partial critical currents @xmath142 , @xmath143 , and the realized critical current @xmath139 , respectively .
main panel : @xmath144 , such that @xmath145 .
this corresponds to the @xmath0-phase region with @xmath146 in the main panel of fig .
inset : @xmath147 , where @xmath148 and no @xmath0-junction behavior is possible .
this corresponds to @xmath149 , see inset of fig .
, scaledwidth=45.0% ] we now notice that eq .
( [ eqc ] ) implies the same decomposition for the critical current ( [ ic ] ) . we can
therefore immediately conclude that the @xmath0-junction regime ( where @xmath150 can exist only when @xmath151 .
this condition is always met away from the window @xmath152 . however , inside that window , eqs .
( [ eq1 ] ) and ( [ eq2 ] ) imply that for sufficiently strong dot - tls coupling , @xmath109 , the @xmath0-phase may disappear completely .
indeed , for @xmath149 , no @xmath0-phase is possible for any value of @xmath138 once @xmath25 exceeds @xmath6 .
the resulting ground - state critical current is shown as a function of the dot level @xmath138 for two typical parameter sets in fig .
the inset shows a case where the @xmath0-phase has been removed by a strong coupling of the interacting dot to the tls .
the above discussion shows that the @xmath0-junction regime is very sensitive to the presence of a strongly coupled tls .
next we address the case of finite tls tunneling , @xmath153 . due to the @xmath154 term in @xmath98 , the critical current can not be written anymore as a weighted sum , see eq .
( [ ic00 ] ) , and no abrupt switching of the tls happens when changing the system parameters .
nevertheless , we now show that the size and even the existence of the @xmath0-phase region still sensitively depend on the tls coupling strength ( and on the other system parameters ) .
in particular , the @xmath0-phase can again be completely suppressed for strong @xmath25 .
( color online ) phase diagram and boundary energies @xmath155 enclosing the @xmath0-phase for finite @xmath19 .
main figure : @xmath156 and @xmath157 , where a @xmath0-phase is present ; @xmath158 and @xmath159 , for solid ( blue ) , dashed ( black ) and dash - dotted ( red ) curves , respectively .
inset : @xmath160 and @xmath157 , where the @xmath0-phase vanishes ; @xmath161 and @xmath162 , for solid ( blue ) , dashed ( black ) and dash - dotted ( red ) curves , respectively.,scaledwidth=50.0% ] for finite @xmath19 , the ground - state critical current is obtained from eq .
( [ eqc ] ) , where the @xmath54 are given by eqs .
( [ c0_gen])([c2_gen ] ) and the @xmath0-phase border energies @xmath155 are replaced by @xmath163 the @xmath164 are defined in eq .
( [ phin ] ) . compared to the @xmath85 case in fig .
[ fig1 ] , the phase diagram boundaries now have a smooth ( smeared ) shape due to the tls tunneling .
nevertheless , the critical current changes sign abruptly when the system parameters are tuned across such a boundary .
the energies ( [ eqb ] ) are shown in fig .
[ fig3 ] for various values of @xmath165 in the @xmath102 plane . in between the @xmath117 and @xmath118 curves ,
the @xmath0-phase is realized . from the inset of fig .
[ fig3 ] , we indeed confirm that the @xmath0-phase can again be absent within a suitable parameter window . just as for @xmath13 ,
the @xmath0-phase vanishes for @xmath166 , and the transition between left and right @xmath9-phase occurs at @xmath167 . for @xmath168
, we effectively recover the phase diagram for @xmath85 , since the tls predominantly occupies a fixed conformational state .
( color online ) listline plots of the @xmath111 ground - state critical current @xmath139 ( in units of @xmath140 ) in the @xmath102 plane , with @xmath169 and @xmath157 .
the boundaries @xmath121 enclosing the @xmath0-phase , see also fig .
[ fig3 ] , are indicated as solid ( blue ) curves .
top panel : small tunnel amplitude , @xmath170 .
bottom panel : large tunnel amplitude , @xmath171.,title="fig:",scaledwidth=45.0% ] ( color online ) listline plots of the @xmath111 ground - state critical current @xmath139 ( in units of @xmath140 ) in the @xmath102 plane , with @xmath169 and @xmath157 .
the boundaries @xmath121 enclosing the @xmath0-phase , see also fig .
[ fig3 ] , are indicated as solid ( blue ) curves .
top panel : small tunnel amplitude , @xmath170 .
bottom panel : large tunnel amplitude , @xmath171.,title="fig:",scaledwidth=45.0% ] the corresponding critical current @xmath139 is shown in fig . [ fig4 ] for both a small and a very large tls tunnel matrix element @xmath165 . in the limit of large @xmath172 , see lower panel in fig .
[ fig4 ] , the dot and the tls are effectively decoupled , since @xmath173 and @xmath174 . while this limit is unrealistic for molecular junctions , it may be realized in a side - coupled double - dot system.@xcite finally we note that , unlike for @xmath85 , the perturbative result for the critical current _ diverges _ at the point where the @xmath0-phase vanishes , i.e. , for @xmath175 .
this divergence is an artefact of perturbation theory and is caused by the appearance of the factor @xmath176 in eqs .
( [ c0_gen ] ) and ( [ c2_gen ] ) .
in this paper , we have presented a perturbative calculation of the critical josephson current , @xmath139 , through an interacting single - level molecular junction side - coupled to a two - level system ( tls ) .
such a tls is a simple model for a bistable conformational degree of freedom , and has previously been introduced in the literature.@xcite our perturbative calculation assumes very weak coupling to attached superconducting reservoirs .
the ground - state critical current can then be computed exactly for otherwise arbitrary parameters .
our main finding is that the @xmath0-phase with @xmath78 is quite sensitive to the presence of the tls . in particular , for strong coupling @xmath25 of the molecular level to the tls as compared to the coulomb energy @xmath6 on the level , the @xmath0-phase can disappear altogether .
we thank t. novotn ' y for discussions .
this work was supported by the sfb tr 12 of the dfg and by the eu networks instans and hyswitch .
in this appendix , we provide the partial critical current @xmath177 which appears in the calculation for @xmath85 , see sec .
[ sec3 ] . in the absence of tls tunneling , the matrix elements ( [ tdef ] ) simplify to @xmath178 where @xmath179 .
we now rename @xmath180 to denote the conformational state ( eigenstate of @xmath181 ) .
the partial current @xmath90 corresponding to fixed conformational state @xmath57 is then given by @xmath182 where @xmath183 with @xmath184 such that @xmath185 .
moreover , the @xmath186 are given by @xmath187,\ ] ] @xmath188,\end{aligned}\ ] ] @xmath189.\end{aligned}\ ] ] a.yu .
kasumov , k. tsukagoshi , m. kawamura , t. kobayashi , y. aoyagi , k. senba , t. kodama , h. nishikawa , i. ikemoto , k. kikuchi , v.t .
volkov , yu.a .
kasumov , r. deblock , s. gueron , and h. bouchiat , phys .
b * 72 * , 033414 ( 2005 ) . | we compute the critical josephson current through a single - molecule junction . as a model for a molecule with a bistable conformational degree of freedom ,
we study an interacting single - level quantum dot coupled to a two - level system and weakly connected to two superconducting electrodes .
we perform a lowest - order perturbative calculation of the critical current and show that it can significantly change due to the two - level system . in particular , the @xmath0-junction behavior , generally present for strong interactions , can be completely suppressed . |
the investigation of magnetism in the molecule - based solid prussian blue , fe@xmath5[fe(cn)@xmath2@xmath6h@xmath4o , and related analogs has a rich history @xcite , dating back to 1928 @xcite .
measurements down to liquid helium temperatures identified the transition to long - range ferromagnetic order @xcite , but an understanding of the magnetic interactions remained elusive until the 1970s , when x - ray @xcite and neutron @xcite diffraction data identified the crystal structure and the spin delocalization from high - spin fe(iii ) to low - spin fe(ii ) .
interest in these materials was renewed in the 1990s with the synthesis of several mixed - metal prussian blue analogs with higher magnetic ordering temperatures , along with the 1996 discovery of long - lived photoinduced magnetism in k@xmath7co@xmath8[fe(cn)@xmath2@xmath9h@xmath4o @xcite .
a flurry of experimental and theoretical research has elucidated the fundamental nature of the light - induced effects in three - dimensional bulk materials @xcite , and recent efforts have been made to integrate the photoinduced magnetism into films @xcite and nanoparticles , which have technical and biophysical applications .
there have been several efforts to synthesize nanoparticles of prussian blue analogs @xcite , but only a few examples of photoinduced magnetism in these particles have been reported , including work that isolated k@xmath0co@xmath1[fe(cn)@xmath2@xmath3h@xmath4o particles with typical diameters of @xmath10 nm within a silica xerogel @xcite and other research producing 11 nm @xmath11 70 nm nanorods of mo(cn)@xmath12cu@xmath4 protected by polyvinylpyrrolidone ( pvp ) @xcite . in each case , although photoinduced magnetism was observed , the particles did not exhibit long - range order .
the purpose of this paper is to report our ability to synthesize rb@xmath0co@xmath1[fe(cn)@xmath2@xmath3h@xmath4o nanoparticles protected by pvp that exhibit photoinduced magnetism for all sizes and that possess long - range ordering , with coercive fields ranging between 0.25 @xmath13 1.5 kg , in the larger particles .
the magnetic properties evolve with size , where at 2 k , the particles with diameters less than approximately 10 nm provide a curie - like magnetic signal .
c c c r r c c r r batch & starting & resulting & pvp : co & diameter & @xmath14 & @xmath15 & @xmath16 & @xmath17 + & pvp ( g ) & chemical formula & ratio & ( nm ) & ( k ) & ( k ) & ( g ) & ( g ) + a & 1.0 & rb@xmath18co@xmath5[fe(cn)@xmath2@xmath19 @xmath20 @xmath21h@xmath4o & 360 & @xmath22 & @xmath23 & @xmath23 & @xmath24 & @xmath24 + b & 0.5 & rb@xmath25co@xmath5[fe(cn)@xmath2@xmath19 @xmath20 @xmath21h@xmath4o & 200 & @xmath26 & 10 & @xmath27 & @xmath28 & @xmath29 + c & 0.2 & rb@xmath30co@xmath5[fe(cn)@xmath2@xmath19 @xmath20 @xmath21h@xmath4o & 60 & @xmath31 & 13 & 17 & 250 & 330 + d & 0.1 & rb@xmath32co@xmath5[fe(cn)@xmath2@xmath33 @xmath20 @xmath34h@xmath4o & 20 & @xmath35 & 19 & 22 & 1000 & 1500 +
our rb@xmath0co@xmath1[fe(cn)@xmath2@xmath3h@xmath4o nanoparticles were synthesized by modifying the procedure for prussian blue nanoparticles described by uemura and coworkers @xcite . a 2 ml solution containing both 28 mg k@xmath36fe(cn)@xmath37 ( 0.085 mm ) and 6.8 mg rbno@xmath36 ( 0.046 mm )
was added dropwise to an 8 ml solution containing 30 mg co(no@xmath36)@xmath38h@xmath4o ( 0.103 mm ) and pvp while stirring rapidly . by varying the amount of pvp ( table 1 ) ,
the protocol produced samples with different particle sizes and size distributions .
after 30 minutes of stirring , the solution was allowed to sit for one week . for transmission electron microscopy ( tem ) studies ,
a 50 @xmath39l aliquot of the suspension was diluted 2000 times , and 8 @xmath39l of the diluted suspension was placed on a holey carbon grid .
a representative image is shown in fig . 1 , and selected area electron diffraction
was compared to powder x - ray diffraction patterns to confirm the structure @xcite . in addition
, infrared spectra displayed a c@xmath13n stretch of 2124 @xmath40 , consistent with a co(iii)@xmath13fe(ii ) low - spin pair . using image j imaging software @xcite ,
the tem images were analyzed to obtain the particle size distributions shown in fig .
these data were fit to a log - normal function that yielded the characteristic diameters shown in fig . 2 and table 1 .
to isolate the particles , three volumes of acetone were added to the synthesis solution , which was centrifuged , and then further washed with acetone and dried under vacuum .
chemical analysis was obtained from a combination of chn combustion analysis and inductively coupled plasma mass spectrometry ( icp - ms ) , and the resulting formulae are listed in table 1 , along with the ratio of the pvp repeat unit per cobalt .
the extent of h@xmath4o coordination to the co was estimated by considering the fe vacancies implicit with the measured co : fe ratio .
co@xmath5[fe(cn)@xmath2@xmath41(h@xmath4o)@xmath42 nanoparticles from batch d , see table 1 .
[ fig1],width=288 ] , height=432 ]
the temperature dependences of the dc magnetic susceptibilities , @xmath43 , of the four batches of particles are shown in fig .
3 . a standard commercial squid magnetometer was employed , and the samples were mounted to commercial transparent tape and could be irradiated with light from a room - temperature , halogen source by using a homemade probe equipped with a bundle of optical fibers @xcite .
small diamagnetic background contributions from the holder and tape ( representing @xmath44% of the total signal ) were independently measured and have been subtracted from the data .
the magnetic signals are expressed per mole of sample using the chemical formula listed in table 1 .
the dark state zfc data were obtained after cooling in zero applied field from 300 k , while the dark state fc data were taken after cooling in 100 g from 300 k. the light state was established after field cooling the samples from 300 k to 5 k in 100 g and subsequently irradiating with light for 5 hours , which saturated the photoinduced response . in order to avoid spin glass - like relaxation
, the light state fc data were obtained after cycling the sample to 30 k in 100 g. ) , field - cooled ( fc ) dark ( @xmath45 ) , and fc light ( @xmath46 ) states of each batch produced .
the insets display the differences between the fc light and dark states , as described in the text .
finite values for this difference can only arise from photoinduced magnetism .
[ fig3],width=288 ]
it is noteworthy to recall that the photoinduced magnetism of the prussian blue analogs depends upon a charge transfer induced spin crossover event ( ctist effect ) , requiring the presence of vacancies to allow crystalline flexibility @xcite .
more specifically for rb@xmath0co@xmath1[fe(cn)@xmath2@xmath3h@xmath4o , the spins of an co@xmath13fe pair can exist in either of two arrangements .
the low - spin state consists of co(iii ) ( @xmath47 ) and fe(ii ) ( @xmath47 ) , while the high - spin state possesses co(ii ) ( @xmath48 ) and fe(iii ) ( @xmath49 ) .
the presence of the ctist effect depends on the local chemical environment defined by the values of @xmath50 , @xmath51 , @xmath52 and @xmath53 in the chemical formula . for limiting ratios , co and fe spins
can be locked into either their high - spin or low - spin states for all accessible temperatures . alternatively , with the proper tuning of the chemical ratios
, the co and fe ions can exist in high - spin states at room temperature and then experience a crossover to their low - spin states at approximately 150 k. the spin crossover phenomenon generates the pairs involved in the photoinduced magnetism .
however , since the ctist effect may not be 100% efficient , some regions remain locked in their high - spin states , and we refer to them as _ primordial _ spins @xcite . for example , these primordial spins are responsible for the magnetic signals observed for the dark state of the largest particles . when irradiated , the low - spin regions , near the primordial high - spin clusters , are photoinduced to the high - spin magnetic state , resulting in a growth of the magnetic domain .
this scenario is supported by our description of the anisotropic photoinduced magnetism observed in thin films @xcite and by the local probe investigations of others @xcite .
all samples show a photoinduced increase in their magnetic signals , and the strength of the change is correlated with the size of the particles .
the differences between the fc susceptibilities of the light and dark states , @xmath54 are plotted in the insets of fig .
3 , and finite values can only arise from the photoinduced magnetism .
it is important to stress that photoinduced magnetism is observed even in small particles possessing only curie - like behavior . ) and fc ( @xmath45 ) states show a signal associated with long - range order for the two batches displaying a mix of magnetic behavior after subtracting the curie - like contribution .
the susceptibility of batch a only exhibits a curie - like contribution , while the magnetic signal for batch d is dominated by spins that are ordered , see text for details . for each sample ,
a curie - like contribution was obtained by fitting the dark state data above 30 k , and this magnetic signal was subtracted from the fc and zfc dark state data shown in fig . 3 . [ fig4],width=288 ] two main features are seen when considering the evolution of material properties due to the increasing average size of the separate batches , namely the onset of long - range magnetic order and an increasing net magnetization .
this scaling of magnetization is linked to an increased diamagnetic - surface to magnetically - active - volume ratio at smaller particle sizes . at low temperatures ,
the high - spin co and fe ions interact antiferromagnetically , giving rise to a ferrimagnetic transition at @xmath55 , which is about 24 k for bulk powder specimens with a similar chemical formula @xcite . for the magnetic data shown in fig . 3 ,
the onset of this transition can be estimated , and these macroscopic temperatures are listed in table 1 .
we contend that particles larger than a critical size will allow domains large enough to approach bulk - like magnetic properties .
conversely , smaller particles may put limits on allowed domain size , suppressing the ordering temperature .
microscopically , if the size of the magnetic domains is less than or of the order of the magnetic coherence length , then a spectrum of @xmath55 values can be expected , and evidence of this trend is observed in batches c and d , see the insets of fig .
3 . in other words
, the size distributions may include particles both above and below the critical size that leads to bulk - like properties . as a result ,
the magnetic response of each batch may include contributions from curie - like and bulk - like fractions .
the percentage of each can be estimated by subtracting a curie - like signal from each @xmath56 plot , and the results for batches b and c are shown in fig .
in contrast , batch a follows curie - like behavior , whereas the active sites in batch d are almost entirely ferrimagnetically ordered .
this analysis of the data suggests the curie - like contribution for each of the four batches of nanoparticles ( smallest to largest ) is 100% , 90% , 50% , and 10% . consequently ,
at least down to 2 k , nanoparticles with sizes below @xmath57 nm do not exhibit long - range order .
these interpretations are consistent with our @xmath58 versus @xmath59 measurements performed at 2 k , where coercive fields , @xmath60 , and remnant magnetization values are observed for the largest sets of particles but not for the smallest set of particles , fig . 5 and table 1 .
furthermore , the differences between the fc and zfc data for the dark states in batches b , c , and d are consistent with spin glass or cluster glass behavior @xcite , and our ac susceptibility data are presented in the appendix .
k magnetization versus magnetic field sweeps for the two largest sizes of nanoparticles , batches c and d , are shown for the light ( @xmath61 and @xmath46 ) and dark ( @xmath62 and @xmath45 ) states .
the coercive fields , @xmath60 for the light and dark states for each batch are listed in table 1 , and the lines are guides for the eyes .
[ fig5],width=480 ]
in conclusion , we have synthesized four different sizes of rb@xmath0co@xmath1[fe(cn)@xmath2@xmath63h@xmath4o nanoparticles protected by pvp .
each batch of particles exhibits photoinduced magnetism , but the strength of this effect , as well as other global properties , e.g. @xmath55 and @xmath60 , are correlated with the intrinsic particle size distributions of each batch .
the combination of photoinduced magnetism and nanosized prussian blue analog particles with finite coercive fields is unique and establishes a length scale limit of @xmath57 nm for these properties .
the next advances in the field will undoubtedly employ prussian blue analogs with higher transition temperatures @xcite that may have a variety of technical applications , including , for example , the exploitation of optically controlled magnetocaloric properties @xcite .
) and imaginary ( @xmath64 ) ac susceptibilities of the zfc dark states , i.e. primordial states , are shown for all four batches of nanoparticles .
all batches were measured with no applied static field and an alternating field of 4 g , except for batch d , which was measured in 1 g. the frequency dependence was studied at 1 hz ( ) , 10 hz ( ) , 100 hz ( @xmath65 ) , and 1 khz ( ) for all batches , except for batch d , which has an additional measurement at 333 hz ( @xmath66 ) .
arrows are guides for the eyes and pass through the peaks.,width=480 ] the temperature dependences of the real ( @xmath67 ) and imaginary ( @xmath64 ) ac susceptibilities of the zfc dark states , i.e. primordial states , of all four batches are shown in fig .
the phenomenological parameter , @xmath68 , given by @xcite @xmath69 where @xmath70 is the freezing temperature given by the cusp in @xmath71 and @xmath72 is the angular frequency , is @xmath73 for batches c and d , and this observation is consistent with spin glass or cluster glass behavior @xcite .
10 dunbar k r and heintz r a 1997 in _ progress in inorganic chemistry _
karlin k d ed ( new york ny : wiley and sons ) vol 45 p 283 verdaguer m and girolami g 2004 in _ magnetism : molecules to materials v _ miller j s and drillon m eds ( weinheim : wiley - vch ) p 284 davidson d and welo l a 1928 1191 holden a n , matthias b t , anderson p w and lewis h w 1956 1463 bozorth r m , williams h j and walsh d e 1956 572 ito a , suenaga m and no k 1968 _ j. chem .
phys . _ * 48 * 3597 buser h j , schwarzenbach d , petter w and ludi a 1977 _ inorg . chem .
_ * 16 * 2704 day p , herren f , ludi a , gudel h u , hulliger f and givord d 1980 _ hel .
. acta _ * 63 * 148 sato o , iyoda t , fujishima a and hashimoto k 1996 _ science _ * 272 * 704 ohkoshi s and hashimoto k 2001 _ j. photochem .
c : photochem .
* 2 * 71 varret f , nogues m and goujon a 2002 in _ magnetism : molecules to materials i _ miller j s and drillon m eds ( weinheim : wiley - vch ) p 257 yamamoto t , umemura y , sato o and einaga y 2004 _ chem .
* 33 * 500 park j - h , cimr e , meisel m w , huh y d , frye f , lane s and talham d r 2004 _ appl .
lett . _ * 85 * 3797 yamamoto t , umemura y , sato o and einaga y 2004 _ chem . mater . _
* 16 * 1195 park j - h , frye f , lane s , cimr e , huh y d , talham d r and meisel m w 2004 _ polyhedron _ * 24
* 2355 yamamoto t , umemura y , sato o and einaga y 2005 _ j. am . chem . soc . _ * 127 * 16065 park j - h 2006 _ ph.d .
( gainesville , university of florida ) frye f a , pajerowski d m , lane s m , anderson n e , park j - h , meisel m w and talham d r 2007 _ polyhedron _ * 26 * 2281 vaucher s , li m and mann s 2000 _ angew .
engl . _ * 39 * 1793 ng c w , ding j , chow p y , gan l m and quek c h 2000 _ j. appl . phys .
_ * 87 * 6049 vaucher s , fielden j , li m , dujardin e and mann s 2001 _ nano lett . _
* 2 * 225 moore j g , lochner e j , ramsey c , dalal n s and stiegman a e 2003 _ angew .
chem . _ * 115 * 2847 uemura t and kitagawa s 2003 _ j. am .
soc . _ * 125 * 7814 uemura t , ohba m and kitagawa s 2004 _ inorg .
_ * 43 * 7339 yamada m , arai m , kurihara m , sakamoto m and miyake m 2004 _ j. am .
soc . _ * 126 * 9482 catala l , mathonire c , gloter a , stephan o , gacoin t , boilot j - p and mallah t 2005 _ chem .
_ 746 johansson a , widenkvist e , lu j , boman m and jansson u 2005 _ nano lett .
_ * 5 * 1603 xian yu , zhou y , xian ya , zhou l , wang h and jin l 2005 _ anal .
* 546 * 139 taguchi m , yamada k , suzuki k , sato o and einaga y 2005 _ chem .
* 17 * 4554 taguchi m , yagi i , nakagawa m , iyoda t and einaga y 2006 _ j. am .
soc . _ * 128 * 10978 catala l , gloter a , stephan o , rogez g and mallah t 2006 _ chem .
_ 1018 frye f a , pajerowski d m , anderson n e , long j , park j - h , meisel m w and talham d r 2007 _ polyhedron _ * 26 * 2273 rasband w s imagej http://rsb.info.nih.gov/ij/ ( accessed october 1 , 2006 ) u.s .
national institutes of health , bethesda , maryland , usa yoshizawa k , mohri f , nuspl g and yamabe t 1998 _ phys .
_ b * 102 * 5432 kawamoto t , asai y and abe s 2001 _ phys .
* 86 * 348 bleuzen a , lomenech c , escax v , villain f , varret f , cartier dit moulin c and verdaguer m 2000 _ j. am .
soc . _ * 122 * 6648 ( independently , we have generated bulk samples with light state @xmath74 k. ) cartier dit moulin c , villain f , bleuzen a , arrio m - a , sainctavit p , lomenech c , escax v , baudelet f , dartyge e , gallet j - j and verdaguer m 2000 _ j. am .
_ * 122 * 6653 escax v , bleuzen a , cartier dit moulin c , villiam f , goujon a , varret f and verdaguer m 2001 _ j. am .
_ * 123 * 12536 champion g , escax v , cartier dit moulin c , bleuzen a , villian f , baudelet f , dartyge e and verdaguer m 2001 _ j. am .
soc . _ * 123 * 12544 shimamoto n , ohkoshi s , sato o and hashimoto k 2002 _ inorg .
chem . _ * 41 * 678 escax v , champion g , arrio m - a , zacchigna m , cartier dit moulin c and bleuzen a 2005 _ angew .
_ * 44 * 4798 hanawa m , moritomo y , kuriki a , tateishi j , kato k , takata m and sakata m 2003 _ j. phys .
jpn . _ * 72 * 987 salman z , parolin t j , chow k h , keeler t a , miller r i , wang d and macfarlane w a 2006 b * 73 * 174427 mydosh j a 1993 _ spin glasses : an experimental introduction _
( london : taylor and francis ) pejakovi d a , manson j l , miller j s and epstein a j 2000 1994 ruiz e , rodrguez - fortea a , alvarez s and verdaguer m 2005 _ chem . eur .
j. _ * 11 * 2135 manuel e , evangelisti m , affronte m , okubo m , train c and verdaguer m 2006 b * 73 * 172406 | nanoparticles of rubidium cobalt hexacyanoferrate ( rb@xmath0co@xmath1[fe(cn)@xmath2@xmath3h@xmath4o ) were synthesized using different concentrations of polyvinylpyrrolidone ( pvp ) to produce four different batches of particles with characteristic diameters ranging from 3 to 13 nm . upon illumination with white light at 5 k , the magnetization of these particles increases .
the long - range ferrimagnetic ordering temperatures and the coercive fields evolve with nanoparticle size . at 2 k , particles with diameters less than approximately 10
nm provide a curie - like magnetic signal . |
let @xmath0 be a compact oriented connected surface of genus @xmath1 with one boundary component .
_ homology cylinders _ over @xmath0 , each of which consists of a homology cobordism @xmath2 between two copies of @xmath0 and markings of both sides of the boundary of @xmath2 , appeared in the context of the theory of finite type invariants for 3-manifolds ( see goussarov @xcite , habiro @xcite , garoufalidis - levine @xcite and levine @xcite ) , and play an important role in a systematic study of the set of 3-manifolds . in our previous paper @xcite , we observed their relationship to knot theory by introducing _
homologically fibered knots_. the set @xmath3 of isomorphism classes of homology cylinders over @xmath0 becomes a monoid by the natural stacking operation .
it is known that the monoid @xmath3 contains the mapping class group @xmath4 of @xmath0 as the group of units ( see example [ ex : mgtocg ] ) . moreover ,
many techniques and invariants to study @xmath4 such as johnson homomorphisms and the magnus representation can be extended to @xmath3 ( see @xcite , @xcite ) . by using them
, we can observe that @xmath3 and @xmath4 hold many properties in common .
taking account of the similarity between @xmath3 and @xmath4 , we now pay our attention to abelian quotients of them .
it is known that @xmath4 is a perfect group for @xmath5 , namely it has no non - trivial abelian quotients . in this paper , however , we will show that the opposite holds for @xmath3 .
the outline of this paper is as follows . after introducing homology cylinders in section [ section : cylinder ] ,
we see that @xmath3 has a big abelian quotient arising from the reducibility of a homology cylinder as a 3-manifold ( theorem [ prop : notfg1 ] ) .
however this fact seems not to be suitable for our purpose of comparing @xmath4 and @xmath3 because all homology cylinders coming from @xmath4 have irreducible underlying 3-manifolds .
therefore we shall introduce the submonoid @xmath6 of @xmath3 consisting of _ irreducible homology cylinders_. the main result is theorem [ thm : notfg2 ] that @xmath6 has a big abelian quotient originating in a quite different context from that of @xmath3 mentioned above . in fact , we prove it in section [ sec : proof ] as an application of sutured floer homology theory .
sutured floer homology was defined first in @xcite by juhsz , then an alternative definition was given in @xcite by ni .
it is a variant of heegaard floer homology theory defined by ozsvth and szab ( here we only refer to @xcite , which contains the results we use later , for details ) . in the last section ,
we discuss our results from the viewpoint of homology cobordisms of homology cylinders .
the authors would like to thank dr .
motoo tange for giving them a lecture about sutured floer homology theory .
they also would like to thank the referee for his / her careful reading of the manuscript and valuable comments .
the final publication is available at www.springerlink.com .
we first recall the definition of homology cylinders , following garoufalidis - levine @xcite and levine @xcite .
a _ homology cylinder _ @xmath7 _ over _ @xmath8 consists of a compact oriented 3-manifold @xmath2 with two embeddings @xmath9 such that : 1 .
@xmath10 is orientation - preserving and @xmath11 is orientation - reversing ; 2 .
@xmath12 and @xmath13 ; 3 .
@xmath14 ; and 4 .
@xmath15 are isomorphisms .
two homology cylinders @xmath16 and @xmath17 over @xmath0 are said to be _ isomorphic _ if there exists an orientation - preserving diffeomorphism @xmath18 satisfying @xmath19 and @xmath20 .
we denote by @xmath21 the set of all isomorphism classes of homology cylinders over @xmath0 .
we define a product operation on @xmath3 by @xmath22 for @xmath16 , @xmath23 , so that @xmath21 becomes a monoid with the identity element @xmath24 , \mathrm{id } \times 1 , \mathrm{id } \times 0)$ ] .
[ ex : mgtocg ] for each diffeomorphism @xmath25 of @xmath8 which fixes @xmath26 pointwise , we can construct a homology cylinder by setting @xmath27 , \mathrm{id } \times 1 , \varphi \times 0),\ ] ] where collars of @xmath28 and @xmath29 are stretched half - way along @xmath30 $ ] .
it is easily checked that the isomorphism class of @xmath24 , \mathrm{id } \times 1 , \varphi \times 0)$ ] depends only on the ( boundary fixing ) isotopy class of @xmath25 and that this construction gives a monoid homomorphism from the mapping class group @xmath31 of @xmath8 to @xmath21 .
in fact , it is an injective homomorphism ( see garoufalidis - levine ( * ? ? ?
* section 2.4 ) , levine ( * ? ? ?
* section 2.1 ) and also ( * ? ? ?
* proposition 2.3 ) ) .
it is known that the image of this homomorphism coincides with the group of units of @xmath21 ( see habiro - massuyeau ( * ? ? ?
* proposition 2.4 ) for example ) .
as seen in example [ ex : mgtocg ] , we may regard @xmath3 as an enlargement of @xmath4 .
we recall here that @xmath4 is a perfect group for @xmath5 ( see korkmaz s survey @xcite and papers listed there for details ) . in comparing the structures of @xmath3 and @xmath4
, it seems interesting to discuss abelian quotients of @xmath3 . note that we need to be careful when mentioning abelian quotients of @xmath3 since it is not a group but a monoid .
we avoid this by considering its universal group @xmath32 and its abelian quotients .
recall that for every monoid @xmath33 , there exist ( uniquely up to isomorphism ) a group @xmath34 together with a monoid homomorphism @xmath35 satisfying the following : for every monoid homomorphism @xmath36 to a group @xmath37 , there exists a unique group homomorphism @xmath38 such that @xmath39 .
one of the possible constructions of @xmath34 is to regard a monoid presentation of @xmath33 as a group presentation . in our discussion below , the case where @xmath40 is exceptional .
we can check that @xmath41 is isomorphic to the monoid @xmath42 of all ( integral ) homology 3-spheres whose product is given by connected sums .
indeed , an isomorphism @xmath43 is given by assigning to each homology 3-sphere @xmath44 the homology cylinder @xmath45)\sharp x , { \mathop{\mathrm{id}}\nolimits}\times 1 , { \mathop{\mathrm{id}}\nolimits}\times 0)$ ] with the inverse homomorphism given by taking _ closures _ ( see example [ ex : knot ] ) .
consequently , @xmath41 is an abelian monoid which is not finitely generated .
we begin our main argument by the following observations .
[ lem : incompressible ] for each @xmath46 with @xmath47 , the surfaces @xmath48 and @xmath49 are incompressible in @xmath2 .
it suffices to show that @xmath50 is injective .
since @xmath51 induces an isomorphism on homology , it follows from stallings theorem ( * ? ? ?
* theorem 3.4 ) that @xmath52 induces isomorphisms on all stages of nilpotent quotients . combining it with the fact that @xmath53 is free , in particular residually nilpotent
, we see that @xmath54 is injective .
[ prop : notfg1 ] the monoid @xmath21 is not finitely generated , for every @xmath55 .
in fact , the abelianization of @xmath32 has infinite rank .
the case where @xmath40 is as mentioned above .
we now assume @xmath47 . for each homology cylinder @xmath46
, the underlying 3-manifold @xmath2 has a prime decomposition of the form @xmath56 where @xmath57 is the unique prime factor containing @xmath58 and @xmath59 are homology 3-spheres . using this decomposition
, we can define the _ forgetting _ map @xmath60 by @xmath61 the uniqueness of the prime decomposition of a 3-manifold shows that @xmath62 is well - defined .
we now claim that @xmath62 is a surjective monoid homomorphism .
let @xmath63 @xmath64 .
we decompose @xmath2 into @xmath65 , where @xmath57 is the prime factor containing @xmath58 and @xmath66 is a homology 3-sphere .
similarly , we have @xmath67 . by lemma [ lem : incompressible ] ,
the underlying 3-manifold of the product @xmath68 has a decomposition @xmath69 such that @xmath70 is the prime factor containing the boundary .
this shows that @xmath62 is a monoid homomorphism .
the surjectivity of @xmath62 follows from the existence of a section @xmath71 defined by @xmath72 ) \sharp x , { \mathop{\mathrm{id}}\nolimits}\times 1 , { \mathop{\mathrm{id}}\nolimits}\times 0)$ ] .
the result follows from the fact that @xmath42 satisfies the conditions mentioned in the statement .
proposition [ prop : notfg1 ] says that the monoid @xmath21 has a different property about its abelian quotients from the mapping class group @xmath31 , which arises from the reducibility of the underlying 3-manifold of a homology cylinder .
however the underlying 3-manifolds of homology cylinders obtained from @xmath31 are all product @xmath73 $ ] and , in particular , irreducible .
therefore it seems reasonable to consider the following subset of @xmath21 .
[ def : irred ] a homology cylinder @xmath16 is said to be _
irreducible _ if the underlying 3-manifold @xmath2 is irreducible .
we denote by @xmath74 the subset of @xmath21 consisting of all irreducible homology cylinders .
note that @xmath74 is a submonoid of @xmath21 , for @xmath75 .
in particular , @xmath76 is the trivial monoid .
note also that we have an injective monoid homomorphism @xmath77 .
the following is the main result of this paper , whose proof will be given in the next section .
[ thm : notfg2 ] the monoid @xmath74 is not finitely generated , for every @xmath47 .
in fact , the abelianization of @xmath78 has infinite rank .
our proof of theorem [ thm : notfg2 ] will be obtained as an application of sutured floer homology theory due to juhsz @xcite and ni @xcite . for each homology cylinder @xmath46 , we have a natural decomposition @xmath79 of @xmath58 .
such a decomposition defines a _ sutured manifold _
@xmath80 with the suture @xmath81 .
sutured manifolds were originally defined by gabai @xcite , to which we refer for details . [ ex : knot ] for a knot @xmath82 in a closed oriented connected 3-manifold @xmath83 with a seifert surface @xmath33 , let @xmath2 be the manifold obtained from the knot exterior @xmath84 by cutting open along @xmath33 and @xmath85 the annulus @xmath86 .
then , @xmath80 is called the _ complementary sutured manifold for _ @xmath33 . the core curve of @xmath85 is denoted by @xmath87 , and called the _ suture_. the suture @xmath87 and @xmath82 are parallel . on the other hand , for each @xmath46 ( or , more generally , each marked cobordism of @xmath0 ) , we have a closed oriented connected @xmath88-manifold @xmath89 called the _ closure _ and a knot @xmath90 with a seifert surface @xmath91 in @xmath92 .
sutured floer homology is an invariant of _ balanced _ sutured manifolds ( see juhsz ( * ? ?
* definition 2.11 ) ) , where all of the sutured manifolds mentioned above satisfy this condition .
it assigns a finitely generated abelian group @xmath93 to each balanced sutured manifold @xmath80 .
we rely on papers of juhsz @xcite and ni @xcite for the definition and fundamental properties of sutured floer homology , and concentrate on using this theory .
juhsz ( * ? ? ?
* theorem 1.5 ) showed that @xmath94 holds for any sutured manifold @xmath80 mentioned in example [ ex : knot ] , where the right hand side is the genus @xmath95 part of the knot floer homology of the knot @xmath87 in @xmath92 with the seifert surface @xmath33 of genus @xmath95 . for each homology cylinder @xmath46
, we put @xmath96 with @xmath97 .
[ prop : nontrivial ] @xmath98 contains @xmath99 for any @xmath46 .
we first assume that @xmath2 is irreducible . by juhsz (
* theorem 1.4 ) , all we have to do is to check that @xmath16 gives a _
taut _ sutured manifold .
that is , * @xmath2 is irreducible ; and * @xmath48 and @xmath49 are incompressible and thurston norm minimizing in their homology classes in @xmath100 with @xmath97 .
the condition ( i ) is automatic and the first half of ( ii ) follows from lemma [ lem : incompressible ] . for the latter half of ( ii )
, it suffices to show that @xmath101 is thurston norm minimizing .
suppose that we have a proper embedding of a surface @xmath102 representing @xmath103 \in h_2 ( m,\gamma)$ ] with smaller norm than that of @xmath101 .
we may assume that @xmath102 does not have any closed component , for such a component comes from @xmath104 and hence removing it does not change the class @xmath105 \in h_2 ( m,\gamma)$ ] nor increase the norm .
next , consider the intersection @xmath106 of @xmath102 and the annulus @xmath85 .
generically , it consists of oriented circles , each of which is an essential simple loop in @xmath85 or bounds a disk in @xmath85 . for an innermost circle bounding a disk
, we attach this disk to @xmath102 and move it away from @xmath58 by an isotopy .
repeating this , we can eliminate all circles bounding disks in @xmath106 , so that the intersection consists of parallel copies of the essential simple loop in @xmath85 .
if @xmath106 is disconnected , we can find an annulus in @xmath85 bounding ( with coherent orientations ) two adjacent components of @xmath106 .
then we attach this annulus to @xmath102 and move it away from @xmath58 .
( when a closed component is produced , remove it . ) repeating this , @xmath106 becomes connected .
note that the above procedure does not change @xmath105 \in h_2 ( m,\gamma)$ ] nor increase the norm since disks and annuli in @xmath85 are trivial in @xmath107 .
consequently , we may assume that @xmath102 is a connected surface with one boundary component .
now suppose that we have a proper embedding @xmath108 with @xmath109 and satisfying @xmath110=[i_+({\ensuremath{\sigma_{g,1 } } } ) ] \in h_2(m,\gamma)$ ] .
we take a basis @xmath111 of @xmath112 such that @xmath113 \in [ \pi_1(m),\pi_1(m)]\ ] ] under suitable orientations . here
we set the basepoint on @xmath87 . by stallings (
* lemma 3.1 ) , we see that @xmath114}{[\pi_1 ( { \ensuremath{\sigma_{g,1}}}),[\pi_1 ( { \ensuremath{\sigma_{g,1}}}),\pi_1({\ensuremath{\sigma_{g,1 } } } ) ] ] } \longrightarrow \frac{[\pi_1 ( m),\pi_1(m)]}{[\pi_1 ( m),[\pi_1 ( m),\pi_1(m)]]}\ ] ] is an isomorphism , and we pull back ( [ eq : boundary ] ) to @xmath115}{[\pi_1 ( { \ensuremath{\sigma_{g,1}}}),[\pi_1 ( { \ensuremath{\sigma_{g,1}}}),\pi_1({\ensuremath{\sigma_{g,1}}})]]}$ ] , which is known to be isomorphic to @xmath116 .
then we obtain an equality @xmath117 ) \wedge i_+^{-1}([\delta_{2i } ] ) \in \wedge^2 ( h_1({\ensuremath{\sigma_{g,1}}})),\ ] ] where @xmath118 \in h_1(m)$ ] denotes the homology class of @xmath119 . on the other hand , we have @xmath120 , the symplectic form , for any symplectic basis @xmath121 of @xmath122 .
define an endomorphism of @xmath122 by @xmath123 ) , \ \
y_i \mapsto i_+^{-1}([\delta_{2i } ] ) \quad \mbox{for \ $ 1 \le
i \le h$},\\ & x_j , y_j \mapsto 0 \quad \mbox{for \ $ h+1 \le j \le g$}. \end{aligned}\ ] ] by definition , this endomorphism is not injective , but preserves the symplectic form .
however , such an endomorphism does not exist since the symplectic form embodies the intersection form on @xmath124 , which is nondegenerate , a contradiction .
therefore @xmath101 is thurston norm minimizing and we finish the proof when @xmath2 is irreducible .
when @xmath2 is not irreducible , we take a prime decomposition @xmath125 as in section [ section : cylinder ] , where @xmath59 are all homology 3-spheres .
then we obtain the conclusion by an argument similar to the proof of ( * ? ? ?
* corollary 8.3 ) using the connected sum formula ( * ? ? ?
* proposition 9.15 ) . by formulas of juhsz (
* proposition 8.6 ) and ni ( * ? ? ? * theorem 4.1 , 4.5 ) together with the fact ( [ eq : sfh - hfk ] ) , we have @xmath126 for @xmath16 , @xmath23 .
hence by taking the rank of @xmath127 , we obtain a monoid homomorphism @xmath128 defined by @xmath129 where @xmath130 is the monoid of positive integers whose product is given by multiplication .
we call @xmath131 the _ rank homomorphism_. note that the restriction of @xmath131 to @xmath4 is trivial since every element of @xmath4 has its inverse . by the uniqueness of the prime decomposition of an integer
, we can decompose @xmath131 into prime factors @xmath132 where @xmath133 is a copy of @xmath134 , the monoid of non - negative integers whose product is given by sums , corresponding to the power of the prime number @xmath135 .
we now restrict the above homomorphisms to @xmath6 .
[ prop : independent ] for @xmath47 , the set @xmath136 contains infinitely many non - trivial homomorphisms that are linearly independent as elements of @xmath137 . to prove this proposition
, we have to consider the images of the homomorphisms @xmath138 .
more specifically , we need many homology cylinders whose ranks of @xmath127 are known .
we now use _ homologically fibered knots _ defined in the previous paper @xcite to construct such homology cylinders .
recall that a knot @xmath82 in @xmath139 is said to be _ homologically fibered _ if @xmath82 satisfies the following two conditions : * the degree of the normalized alexander polynomial @xmath140 of @xmath82 is @xmath141 ; and * @xmath142 , where @xmath143 is the genus of @xmath82 and @xmath140 is normalized so that its lowest degree is @xmath144 . in (
* theorem 3.4 ) , we showed that @xmath82 is homologically fibered if and only if @xmath82 has a seifert surface @xmath33 whose complementary sutured manifold is a homology product .
here , a homology product means a homology cylinder without markings .
note that crowell and trotter observed in @xcite this essentially .
( see also @xcite . )
we also showed that if @xmath82 is homologically fibered , then any minimal genus seifert surface gives a homology product . for a homologically fibered knot @xmath82 of genus @xmath1
, we obtain an irreducible homology cylinder @xmath145 by fixing a pair of markings of the boundary of the complementary sutured manifold @xmath146 for a minimal genus seifert surface . by ( [ eq : sfh - hfk ] )
, @xmath147 holds for such a homology cylinder .
we first give a proof of the case where @xmath148 .
we now consider pretzel knots @xmath149 with @xmath150 .
as depicted in figure [ pretzel1 ] , each of such knots has a genus @xmath151 seifert surface .
it is well known that the normalized alexander polynomial of @xmath149 is given by @xmath152 using this formula , we see that the sequence @xmath153 consists of homologically fibered knots of genus @xmath151 since @xmath154 .
moreover , a computation due to ozsvth - szab ( * ? ? ?
* latter formula of theorem 1.3 ) gives @xmath155 from which we see that @xmath156 .
we now analyze the sequence @xmath157 of positive integers . the first supplement to quadratic reciprocity
says that there exists a positive integer @xmath158 such that @xmath159 for any odd prime @xmath135 satisfying @xmath160 . in this case , set @xmath161 then , if @xmath158 is odd , we have @xmath162 and if @xmath158 is even , we also have @xmath163 hence we can conclude that @xmath138 is non - trivial if @xmath160 . by dirichlet
s theorem on arithmetic progressions , there exist infinitely many such prime numbers
. for a homology cylinder @xmath164 , let @xmath165 take a sequence of homology cylinders @xmath166 such that @xmath167 .
then we can see that @xmath168 are linearly independent by evaluating them on the @xmath169 s .
this concludes the proof of the case where @xmath148 .
let @xmath170 be a homologically fibered knot of genus @xmath171 obtained from @xmath172 by taking connected sums with @xmath173-tuples of trefoils . by ( * ?
* theorem 1.1 ) and ( * ? ? ?
* corollary 8.8 ) together with the fact that the trefoil is a fibered knot of genus @xmath151 , so that @xmath174 , we have @xmath175 therefore , the cases where @xmath176 follow from the same argument as above .
suppose @xmath6 was finitely generated .
then except for finitely many primes , the homomorphisms @xmath138 are trivial on any finite set of generators , and hence on whole @xmath6 .
this contradicts proposition [ prop : independent ] and we have proved the first half of our claim .
the latter half follows from the construction that uses infinitely many homomorphisms whose targets are abelian .
in @xcite , garoufalidis - levine introduced _ homology cobordisms _ of homology cylinders , which give an equivalence relation of homology cylinders .
we finish this paper by two observations concerning our results ( proposition [ prop : notfg1 ] and theorem [ thm : notfg2 ] ) from the viewpoint of this equivalence relation .
we denote by @xmath182 the quotient set of @xmath21 with respect to the equivalence relation of homology cobordism .
the monoid structure of @xmath21 induces a group structure of @xmath182 .
it is known that @xmath4 can be embedded in @xmath183 ( see ( * ? ? ?
* section 2.4 ) , ( * ? ? ? * section 2.1 ) ) .
one important problem is to determine whether @xmath183 is perfect or not .
has @xmath184 as an abelian quotient , for every @xmath47 .
] in fact , no non - trivial abelian quotients of @xmath182 are known at present .
we now observe that it is difficult to give an answer to this problem by using the homomorphisms used in this paper .
first we consider the forgetting homomorphism @xmath185 discussed in section [ section : cylinder ] .
it follows from myers result ( * ? ? ?
* theorem 3.2 ) that every homology cylinder in @xmath3 with @xmath47 is homology cobordant to an irreducible one , whose image by @xmath62 is trivial by definition .
hence if @xmath189 factors through @xmath183 for a monoid homomorphism @xmath190 , then @xmath191 must be trivial .
next we consider the rank homomorphisms @xmath192 discussed in section [ sec : proof ] .
it induces a group homomorphism @xmath193 on universal groups .
note that the quotient group of @xmath194 by homology cobordism relation is also @xmath183 as mentioned in the proof of theorem [ thm : ab_quot1 ] .
s. garoufalidis , j. levine , _ tree - level invariants of three - manifolds , massey products and the johnson homomorphism _ , graphs and patterns in mathematics and theorical physics , proc .
pure math .
73 ( 2005 ) , 173205 . | a homology cylinder over a surface consists of a homology cobordism between two copies of the surface and markings of its boundary .
the set of isomorphism classes of homology cylinders over a fixed surface has a natural monoid structure and it is known that this monoid can be seen as an enlargement of the mapping class group of the surface .
we now focus on abelian quotients of this monoid .
we show that both the monoid of all homology cylinders and that of irreducible homology cylinders are not finitely generated and moreover they have big abelian quotients .
these properties contrast with the fact that the mapping class group is perfect in general . the proof is given by applying sutured floer homology theory to homologically fibered knots studied in a previous paper . |
in the recent decade and a half there has been a lot of activity on the topic of non - gibbsian measures .
most of the original studies were based on the question of whether renormalised hamiltonians exist as properly defined objects , @xcite , with an emphasis on discrete - spin models .
another issue , which also arose in physics but somewhat later @xcite , was the following question : apply a ( stochastic ) dynamics which converges to a system at a temperature @xmath1 to an initial state at temperature @xmath2 for a finite time .
is the resulting measure in this transient non - equilibrium regime a gibbs measure ? could it be described in terms of an effective temperature ( hopefully between initial and final one ) ?
again the first results @xcite were for discrete spins . afterwards more general dynamics and
also unbounded spins were investigated in @xcite .
although the work of @xcite was about continuous spins , there remains something of a problem , in that for unbounded spins the notion of what one should call gibbsianness for a `` reasonable '' interaction is less clear than in the compact case .
thus it turned out to be of interest to see how a model with compact but continuous spins behaves .
another extension of the original investigations was the investigation into the question of what the proper mean - field version of the gibbs - non - gibbs question might be .
for this , see in particular @xcite .
this question has a particular charm for systems with a general local spin space . as there have recently been a number of reviews on other aspects of the non - gibbsian problem @xcite , we here want to emphasize what has been found for @xmath3-vector spins .
the results as yet are less complete than what is known for ising or potts spins , but it has also become clear that , although many things are similar , such systems have traits of their own which are somewhat different and require new ideas .
we have mostly worked on transformations such as stochastic evolution , which does not rescale space , such as renormalisation group transformations do .
note that in a statistical interpretation , such maps for discrete spins model imperfect observations , that is observations in which with some probability one makes a mistake , an interpretation which already was mentioned in @xcite . for continuous spins ,
the probability of staying exactly at the initial value is zero , but for short times the map is close to the identity in the sense that the distribution of an evolved spin is concentrated on a set close to the initial value .
we obtained conservation of gibbsianness under stochastic evolutions when either the time is short , or when both initial and final temperature are high .
we also found that loss of gibbsianness occurs if the initial temperature is low , and the dynamics is an infinite - temperature one .
if the initial system is in an external field , after a long time the measure can become gibbsian again .
in fact , here we extend the regime where such results can be proven .
another question we could address is the discretisation question . if one approximates a continuous model by a discrete one , is the approximation still a gibbs measure , now for discrete spins ?
morally , this question is somewhat related to renormalisation - type questions , as in both cases some coarse - graining takes place , in which the transformed system only contains part of the initial information .
it turns out that the transformed measure is gibbsian , once the discretisation is fine enough .
all of these questions , the high - temperature and short - time gibbsianness for stochastic evolutions , as well as the loss and recovery properties , can also be addressed in the mean - field setting , and we find that the results are similar as in the lattice case .
again , for transformations which in some sense are close enough to the identity , the transformed model is gibbsian .
finally , one may ask which of our results depend on the fact that our local state space is a sphere and not just a compact space ? the regularity results ( preservation of gibbsianness ) do not , as they are based on absence of phase transitions .
in fact , such extensions have been proved , for which we refer to the original papers .
when it comes to a failure of gibbsianness , an internal phase transition has to be exhibited .
the mechanism of this is usually very model - dependent and this is where the intricacy but also the charm of the @xmath3-vector models lies .
in this section we review some recent developments in the study of gibbsianness and non - gibbsianness for @xmath0vector models subjected to various transformations .
the review is mainly based on the recent papers @xcite .
before we plunge into details let us fix some definitions , notation and give some background from the theory of lattice spin systems . for general information on gibbs measures for lattice spins systems
we refer the reader to @xcite . in this review
we will focus attention on models living on a @xmath4dimensional lattice @xmath5 ( @xmath6 ) .
we will take @xmath7 , the @xmath0dimensional sphere , as the single - site spin space equipped with a borel probability measure @xmath8 ( the a priori measure ) .
the measures we will study shall be given by hamiltonians .
the hamiltonians in a finite volume @xmath9 , with boundary condition @xmath10 outside @xmath11 , will be given by @xmath12 where the interaction @xmath13 is a translation - invariant family of functions @xmath14 , with @xmath15 depending only on the spins in the finite volume @xmath16 .
it satisfies the following absolute summability condition @xmath17 the * gibbs measures * for the interaction @xmath13 are the measures @xmath18 on @xmath19 whose finite - volume conditional distributions are given by @xmath20 where @xmath21 is the product measure of @xmath8 over the sites in @xmath11 .
another , equivalent , way of defining a gibbs measure was identified by kozlov @xcite , via two properties of the family of conditional distributions @xmath22 .
these properties are * uniform nonnullness * and * quasilocality*. the latter property holds for a measure @xmath18 if for all @xmath23 , @xmath24 and configurations @xmath25 there exists a @xmath26 and @xmath27 such that for all pairs of configurations @xmath28
@xmath29 a collection @xmath30 of everywhere defined conditional distributions @xmath31 satisfying all the above conditions is referred to as a * gibbsian specification*. now what can be said about the gibbs properties of transformed gibbsian @xmath0vector models ? in @xcite the gibbs properties of various transformations acting on @xmath0vector models were investigated and we will review the results below .
we discuss conservation of gibbsianness for initial gibbsian @xmath0vector lattice models subjected to local transformations close to the identity .
the discussion will mainly follow @xcite .
though these two papers use different techniques , the results proved therein have some common ground and we will compare the advantages and disadvantages of both methods . we will mainly address conservation of gibbsianness for transformed initial gibbs measures in this subsection .
we start with a gibbs measure @xmath18 of an @xmath0vector model and apply local transformations to it .
examples of such local transformations are infinite - temperature diffusive dynamics ( sitewise independent brownian motions on spheres ) , fuzzification or discretisation of the local spin space , etc .
the natural question that comes to mind is whether such a transformed measure @xmath32 is a gibbs measure . for transformations close to identity
the above question can be answered in the affirmative .
this we make precise in the sequel by first stating a theorem which is the intersection of the results found in @xcite .
[ mainthm ] suppose @xmath18 is the gibbs measure for a translation invariant interaction @xmath13 on @xmath33 .
further , assume that @xmath13 is twice continuously differentiable and of finite range .
let @xmath34 be the transformed ( time - evolved ) measure obtained by applying infinite - temperature diffusive dynamics to @xmath18 .
then for short times the time - evolved measure @xmath34 is a gibbs measure .
theorem [ mainthm ] can be proved either by using cluster expansion techniques as in @xcite or by dobrushin uniqueness techniques @xcite .
the results proved in these papers generalise the above theorem in different directions . in the following
we will review some of the main issues discussed in them .
let us start with the approach of @xcite .
the advantage of using cluster expansion techniques is that we can prove short - time gibbsianness for more general dynamics beyond the independent brownian motion on the circles . in particular
, one can handle a whole class of systems which are modeled via the solution @xmath35 of the following system of _ interacting _ stochastic differential equations : @xmath36 where @xmath37 denotes a family of independent brownian motions moving on a circle , @xmath38 and @xmath39 is the `` dynamical '' inverse temperature .
we assumed that the `` dynamical '' hamiltonian @xmath40 is built from an absolute summable `` dynamical '' interaction which is again of finite range and at least twice continuously differentiable .
let @xmath41 denote the semigroup of the dynamics defined in .
then one can prove that for all values of @xmath42 the time - evolved measure @xmath43 is gibbsian for short times .
note that the statement of theorem [ mainthm ] corresponds to the case where @xmath44 .
we note that the cluster expansion technology was heavily influenced by @xcite .
extensions to different graphs are also immediate . + the proof in @xcite makes also use of the fact that @xmath45 where
@xmath46 and @xmath47 are considered to be the same points .
consequently we can work on the real line and do not have to worry about more general compact manifolds @xmath48 .
although it is in principle possible to write a cluster expansion for @xmath48 and we believe that short - time gibbsianness for general interacting dynamics holds also in higher spin dimensions , this has not been done so far .
next let us review the results in @xcite .
the dobrushin uniqueness technique employed in that work applies to more general interactions on general @xmath3-spheres and also to more general graphs aside from @xmath5 .
moreover , we expect that it provides better bounds for the gibbsian regime than the cluster expansion approach does . on the other hand ,
no results for interacting dynamics have been obtained via this approach , although we believe in principle this should be possible .
to be precise , one considers initial gibbs measures for interactions @xmath13 with finite triple norm , i.e. @xmath49 note that this summability condition implies the one in and here we do nt require that @xmath13 is translation invariant .
initial gibbs measures of such interactions were subjected to local ( one - site ) transformations given by @xmath50 , with @xmath51 . here
@xmath25 represents the spin variable for the transformed system taking values in a compact separable metrizable space @xmath52 , which now needs not be the same as @xmath7 . in the language of renormalization group transformations
, one could think of the transformed system as the renormalized system obtained via the single - site renormalization map @xmath53 .
the map @xmath53 can also be thought of as the transition kernel for an infinite - temperature dynamics , where the variable in the second slot will be the configuration of the system at some time after starting the dynamics from the configuration in the first slot of @xmath53 .
sometimes we will refer to the time direction as the `` vertical '' direction .
starting with an initial gibbs measure @xmath18 for an interaction @xmath13 with finite triple norm , in @xcite it was studied to what extent the transformed measure @xmath54 will be gibbsian . in the above we have set @xmath55 .
the study in @xcite uses dobrushin uniqueness techniques .
the paper also provides continuity estimates for the single - site conditional distributions of the transformed system whenever it is gibbsian . to introduce these estimates
the authors made use of a so - called `` goodness matrix '' , which describes the spatial decay of the conditional distributions of the transformed measure .
+ in the sequel we will write @xmath56 for @xmath57 and @xmath58 for @xmath59 .
in particular the following definition from @xcite will be used .
[ otto ] assume that @xmath60 is a metric on @xmath7 and @xmath61 is a non - negative matrix with @xmath62 .
a gibbsian specification @xmath30 is said to be of goodness @xmath63 if the single - site parts @xmath64 satisfy the continuity estimates @xmath65 here @xmath66 is the variational distance between the measures @xmath67 and @xmath68 .
the matrix @xmath69controls the influnce on the specification due to variations in the conditioning when we measure them in the metric @xmath60 .
the faster @xmath69 decays , the better , or `` more gibbsian '' , the system of conditional probabilities is .
we note , without going into details , that a fast decay of @xmath69 also implies th existence of a fast decaying interaction potential , but in our view an estimate of the form ( 7 ) is more fundamental than a corresponding estimate on the potential .
we are restricting our attention to single - site @xmath70 s since all @xmath71 for finite @xmath11 can be expressed by an explicit formula in terms of the @xmath64 s with @xmath72 . for the solution of this `` reconstruction - problem '' see @xcite .
the dobrushin interdependence matrix @xmath73 of a gibbsian specification @xmath30 @xcite , is the matrix with smallest matrix - elements for which the specification @xmath30 is of the goodness @xmath74 . here
@xmath60 is the discrete metric on @xmath7 given by @xmath75 .
one says that @xmath30 satisfies the * dobrushin uniqueness condition * whenever the dobrushin constant @xmath76 , and such a gibbsian specification @xmath30 admits a unique gibbs measure @xcite .
let us now introduce some notation for our discussion on conservation of gibbsianness for transforms of gibbs measures .
set for each @xmath77 , @xmath78 , the a priori measures on the initial spin space which are obtained by conditioning on transformed spin configurations .
we call @xmath79 the posterior ( pseudo-)metric , associated to @xmath80 on the transformed spin space @xmath52 .
@xmath81 satisfies non - negativity and the triangular inequality , but we may have @xmath82 for @xmath83 ( which happens e.g. if @xmath84 and @xmath85 are independent under @xmath80 ) .
for any given @xmath13 with finite triple norm write @xmath86 for @xmath87 where @xmath88 is as in .
consider the matrix @xmath89 here we have denoted by @xmath90 , for @xmath91 a real - valued observable on @xmath92 , the oscillation of @xmath91 given by @xmath93 the above quantity @xmath94 , can be small if either the initial interaction @xmath13 is weak or the measures @xmath95 are close to delta measures . for example
this is the case for short - time evolution of the initial gibbs measure associated with @xmath13 , as we will point out later .
@xmath96 is an upper bound on the dobrushin matrices for the joint systems consisting of the initial and the transformed spins vertically coupled via the map @xmath53 , and having fixed transformed configurations .
a specification for this sytem is generated by @xmath97 by replacing in equation ( 3 ) the a priori measure @xmath98 by the @xmath99 s .
the main tool used in @xcite to show gibbsianness of the transformed measure was the lack of phase transitions in the conditional joint system discussed above .
this lack of phase transitions will follow if the dobrushin constant of the matrix , @xmath100 , is strictly less than @xmath101 .
more precisely the following theorem was proved ( @xcite : theorem 2.5 ) .
[ mainthm - lat ] suppose that @xmath18 is a gibbs measure associated with a lattice interaction @xmath13 with finite triple norm .
suppose further that @xmath102
. then the transformed measure @xmath32 is gibbsian and the transformed gibbsian specification @xmath103 has goodness @xmath104 , where @xmath105 with @xmath106 .
thus the transformed measure @xmath32 will be gibbsian if either the initial interaction @xmath13 is weak or the a priori measures @xmath95 are close to delta measures .
furthermore , in the above theorem the goodness of the transformed specification is expressed in terms of the posterior metric @xmath81 .
can one have local transformations where this metric could be expressed in terms of more familiar metrics on @xmath7 ? in what follows we present two examples where the above question will have a positive answer . to do this
we pay a price of putting further restrictions on the class of allowed interactions for the initial system .
let us equip @xmath7 with a metric @xmath60 .
denote by @xmath107 the smallest constants such that the @xmath108-variation of the hamiltonian @xmath109 satisfies @xmath110 we say that @xmath97 satisfies a lipschitz property with constants @xmath111 , if all these constants are finite . for this class of interactions
it is not hard to see from that @xmath112 this follows from replacing the @xmath113 in by @xmath114 .
let us now see some concrete examples . to a gibbs measure @xmath18 for a lattice interaction
@xmath13 we apply sitewise independent diffusive dynamics given by @xmath115 in the above @xmath116 is the equidistribution on @xmath7 and @xmath117 is the heat kernel on the sphere , i.e. @xmath118 where @xmath119 is the laplace - beltrami operator on the sphere and @xmath120 is any test function .
@xmath117 is also called the _ gauss - weierstrass kernel_. for more background on the heat - kernel on riemannian manifolds , see the introduction of @xcite .
let @xmath121 be the matrix with entries @xmath122 with the above notation we have the following generalization of theorem 2.7 of @xcite .
[ genthm ] suppose @xmath60 is the euclidean metric on @xmath7 and @xmath13 is an interaction for which there are finite constants @xmath123 such that @xmath124 assume further that @xmath125 .
then the transformed measure @xmath34 obtained from a gibbs measure @xmath18 for @xmath13 and @xmath126 is gibbsian and the specification for @xmath34 has goodness @xmath127 with @xmath128 here @xmath129 is defined in the same way as in but has @xmath96 replaced by @xmath130 .
the inequality implies . but using gives a better bound on the dobrushin interdependence matrix . in we keep all the interactions a given site @xmath56 has with the rest of its environment , but in only the interaction between @xmath56 and a reference site @xmath108 is kept .
note also that the entries of @xmath130 will be small if either the initial interaction is weak or @xmath131 is small enough .
this is a generalization of the corresponding ising and planar rotor results found in @xcite to more general interactions on any @xmath0dimensional sphere , subjected to infinite - temperature diffusive dynamics .
+ + the above theorem was proved in @xcite for some special pair interaction .
the proof there followed from three steps , namely : 1 ) an application of theorem [ mainthm - lat ] to obtain continuity estimates in terms of the posterior metric @xmath81 , 2 ) a comparison result between @xmath81 and @xmath60 ( the euclidean metric ) , see e.g proposition 2.8 of @xcite and 3 ) a telescoping argument over sites in the conditioning .
the first two steps hold for any general interaction on the sphere .
+ in the third step one uses the continuity property to proceed . in particular
one replaces the constant @xmath132 in inequality ( 100 ) of @xcite by @xmath133 .
@xmath134 in the next subsection we consider another class of transformations which was studied in @xcite .
consider a gibbs measure @xmath18 for an interaction @xmath13 satisfying .
furthermore , partition @xmath7 into countably many pairwise disjoint subsets with non - zero @xmath8 measure , indexed by elements in a countable set @xmath52 .
thus we have disjoint subsets @xmath135 , such that @xmath136 and @xmath137 for all @xmath138 . for each such decomposition of @xmath7
we define @xmath139 where @xmath140 is the diameter of the set @xmath135 .
we refer to @xmath141 as the fineness of the decomposition . in this set - up the conditional a priori measure
is given by @xmath142 .
the above decompositions of @xmath7 define natural maps from the space of probability measures on @xmath92 to probability measures on @xmath143 .
the question now is : `` which of these maps will lead to a gibbsian image measure @xmath32 upon their application to the gibbs measure @xmath18 ? '' this question is partially answered in theorem 2.9 of @xcite , which we state below .
[ fineapp ] suppose @xmath13 is as above and @xmath144 then for any gibbs measure @xmath18 of @xmath13 the transformed measure @xmath32 , associated with the decomposition with fineness @xmath141 , is a gibbs measure for a gibbsian specification @xmath103 of goodness @xmath145 . here @xmath146 is the discrete metric on @xmath52 and @xmath69
is given in with @xmath100 given by @xmath147 observe from the above theorem that the quantity in can be small if either the initial interaction is weak or the fineness @xmath141 of the decomposition is small enough .
thus for any strength of the initial interaction , the transformed measures will remain gibbsian if our decomposition is fine enough .
we note that if we make a decomposition of the circle into equal intervals , the resulting models resembles a clock model . on could in fact also apply the theorem starting with discrte spins , such as a large-@xmath148 clock model , but the advantage of considering continuous spin spaces is that the theorem can always be applied ( in other words , there is _ always _ a fine enough decomposition ) .
note also that such a discretisation map has strong similarities with fuzzification maps such as have been studied for potts models , see e.g. @xcite , in which one also decomposes the single - site spin space , into a smaller number of fuzzy spin values .
this section deals with what is known about conservation , loss and recovery of the gibbs property in time - evolved gibbsian measures of vector models on the lattice @xmath149 .
the conservation part will focus on large - time results , as the short - time results have been described in the previous section .
we will concentrate here on the work done in @xcite .
moreover we will present some new arguments which extend the results in @xcite . in the previous section we defined gibbs measures , see equation , and furthermore we gave an equivalent description which we stated in equation .
let us now focus on the latter expression .
in words it says if a measure @xmath18 is gibbsian , every configuration @xmath25 is * good * , in the sense that for every @xmath25 the measure is + continuous w.r.t . a change in the conditioning .
we referred to this property as the * quasilocality * property .
but what does it mean for a measure @xmath18 to be _
loss of gibbsianness means essentially the failure of this quasilocality property .
it is enough to find at least one point of essential discontinuity @xmath150 w.r.t .
the conditioning , namely a point satisfying @xmath151 for @xmath152 , uniformly in @xmath153 , to prove that a measure is non - gibbsian .
physically the failure of quasilocality means the following : + the spin at the origin @xmath154 is influenced by far away configurations @xmath155 and @xmath156 even when the spins in between are frozen in the configuration @xmath157 . for a measure which is gibbsian ,
the spin @xmath154 is shielded off from spins far away when intermediate spins are fixed .
so there are no hidden fluctuations transmitting information from infinity to the origin .
+ typically , in the analysis one considers a double - layer system with the initial spin space in the first layer and the transformed system ( or image - spin space ) in the second layer .
the question of the gibbsianness of the measure on the second layer then can be shown to reduce to the question : is it possible to end up in this particular configuration coming from one initial gibbs measure only ?
it turns out that if the original spin system * conditioned * on a particular image spin configuration @xmath150 exhibits a phase transition , this implies for @xmath158 that this measure is not gibbsian .
the configuration @xmath150 is often called a * bad configuration*. we want to stress the difference between a phase transition of the initial system and a phase transition of the conditioned double - layer system . even if the initial system exhibits a phase transition and the time - evolved measure at time @xmath131 is a gibbs measure , it means that conditioned at every possible image spin configuration @xmath25 at time @xmath131 , there is no phase transition for the conditioned system . in other words , for every possible @xmath25 there is only one possible initial measure leading to this image spin configuration .
hence , a phase transition of the initial system does not necessarily imply non - gibbsianness , nor does non - gibbsianness imply a phase transition for the initial measure .
+ in the case of time - evolved xy - spins on the lattice @xmath5 , in @xcite and @xcite some results about conservation , loss and recovery of gibbsianness were proved . in @xcite results
are proven for conservation and loss of gibbsianness during time - evolution .
in particular , loss of gibbsianness could be proven for zero initial external fields .
the paper @xcite deals with loss and recovery of gibbsianness in a situation where there is a positive initial external field . as we already discussed in the previous section
, the gibbsian property is conserved for short times for all initial gibbs measures evolving under diffusive dynamics consisting of brownian motions on the circle , either with or without gradient hamiltonian drifts , at all temperatures ( for all values of @xmath1 and all values of the initial temperature @xmath2 ) .
moreover , conservation for all times holds if the system starts with a high - or infinite - temperature gibbs measure and evolves under high- or infinite - temperature dynamics ( @xmath159 ) .
+ let us make the statement on the loss of gibbsianness result from @xcite more precise .
systems in @xmath160 are considered which start in a low - temperature initial measure with nearest neighbour interaction and zero external field , @xmath161 and evolve under independent brownian motion dynamics on the circle .
then there is a time interval , depending on the initial ( inverse ) temperature , such that the time - evolved measure is not gibbsian .
the idea of the proof is the following .
consider the double - layer measure and condition it on the alternating north - south configuration .
the ground states of the conditioned system then are two symmetric configurations of spins pointing either to the east with a small correction @xmath162 or to the west with a small correction .
let us give a schematic picture .
west - pointing ground state east - pointing ground state the potential of the conditioned model is of c - type ( nearest neighbour , invariant under reflection in vertical and horizontal direction and invariant under translation ) , so that the corresponding measure is reflection positive . using a percolation argument for low - energy clusters ( clusters of vertices connected by low - energy bonds ) from @xcite , in two dimensions
one proves that at low temperatures there exist two distinct extremal gibbs measures for the thus - conditioned system .
this implies that the conditioned double - layer system undergoes a phase transition via discrete symmetry breaking and therefore the time - evolved measure is not gibbsian .
this phase transition is called of spin - flop " type .
let us remark on a special feature of this result .
in the ising spin case see @xcite , where one also finds at zero external field loss of gibbsianness , the initial system itself is already not unique .
the xy spin model , however , does not have a first order phase transition in two dimensions .
so even though starting from a unique gibbs measure , there is a time interval where gibbsianness is lost . +
we also mention that the result can be extended to @xmath163 and arbitrary large times . in that case
the initial measure is not unique , and there is long - range order for any strength of the alternating magnetic field , including zero .
unfortunately , the techniques which are used rely on the reflection positivity property of the measure , and therefore can not be applied to a system which evolves with high - temperature dynamics , since then the conditioned measure is not reflection positive any more . also for higher - component spins the proof breaks down .
+ for discrete spins , the authors in @xcite prove that loss of gibbsianness appears also for high - temperature dynamics , for rotor spins we believe the same is true but this has not yet been proven . by some pirogov - sinai type arguments
one might hope to extend the above result to high - temperature dynamics .
but this seems a technically hard question .
+ in the presence of an initial external field @xmath164 loss and also recovery results were obtained in @xcite .
similar to the situation for discrete spins in @xcite , one finds that also in the presence of a small initial external field there can be a time interval @xmath165 where gibbsianness is lost in @xmath166 dimensions .
moreover there exists a time @xmath167 such that for all @xmath168 , the time - evolved measure is again a gibbs measure .
this re - entrance result was obtained for strong initial external fields in @xmath169 lattice dimensions .
+ intuitively , for an intermediate time interval , the strength of the initial field is compensated by the induced field coming from the time - evolution .
this compensation makes the system behave like in a modified zero field situation .
the system looks like a zero field system plus some rest terms which have a discrete symmetry instead of a continuous one . for low enough initial temperatures
there is a time interval where this symmetry is broken for the conditioned double - layer system , so therefore gibbsianness is lost . after some time the influence of the time - induced fields decreases and the system follows the initial field again which brings it back to the gibbsian regime .
thus the presence of the initial external field is responsible for the recovery of the gibbsian property .
+ the proof in @xcite is similar to the one in @xcite .
one considers a double - layer system and conditions on spins pointing all southwards .
then the two ground states of the conditioned system are again symmetric , pointing either to the east or to the west .
we present a schematic picture of the ground states .
east - pointing ground state west - pointing ground state since the interaction of the conditioned system is also of c - type , i.e. among other properties invariant under reflections , one can use again the percolation of low - energy clusters argument of georgii , see @xcite .
+ the proof for the * loss * of gibbsianness works only for @xmath166 and for small initial fields .
+ in the following we propose an argument for a loss of gibbsianness result which works for a general initial field @xmath164 already in a two - dimensional lattice .
moreover we will prove a recovery result valid for all strengths of the initial field at low enough temperatures .
let @xmath164 be given .
for @xmath170 big enough , there exists a time interval + @xmath171 such that for @xmath172 , the time - evolved measure @xmath173 is not gibbsian .
* proof : * + let us consider the double - layer joint measure at time 0 and time @xmath131 simultaneously .
the dynamical hamiltonian @xmath174 ( the inverse temperature is in this case absorbed into the definition of the hamiltonian ) is formally given by @xmath175 where @xmath176 , @xmath177 is the transition kernel on the circle and the initial hamiltonian @xmath178 is formally given by @xmath179 while @xmath180 equals @xmath181 we condition the system at time @xmath131 to point down alternatingly with a small correction @xmath182 ( which we will specify later ) to the east , resp . with a small correction to the west ,
i.e. @xmath183 let us look at the corresponding dynamical hamiltonian .
it can be written as a sum over all nearest neighbor pairs of the pair interaction - potential @xmath184 where @xmath185 are parameters ( at @xmath186 fixed ) , and @xmath187 denote the values of neighboring spins .
the single - site terms coming from the dynamics play the role of dynamical alternating external fields .
let us rewrite the terms as @xmath188 and similarly for the second one .
we choose @xmath131 of order @xmath189 , or more precisely in such a way that @xmath190 where @xmath191 is an error term going to 0 for @xmath131 small .
call @xmath192 and similarly for @xmath193 .
observe that the pair interaction potential is equal to @xmath194 we end up with a ferromagnetic system with alternating dynamical external fields of strength @xmath195 coming from the terms @xmath196 which effectively point perpendicular to the original fields .
let us assume @xmath197 .
then the strength of the fields is in fact of order @xmath198 and and the direction alternates , pointing almost to the east or almost to the west while the strength is relatively weak compared to the nearest - neighbour interaction
. we will be able to show that the spin flop mechanism causes a phase transition to occur .
+ in order to understand the phases of such a hamiltonian we will look at first at its ground states .
we want that @xmath199 is minimal at @xmath200 and @xmath201 , so the ground states point almost to the north , namely in north - east and north - west direction , or almost to the south . in general one of them
is a local minimum and one is global .
one determines @xmath202 and @xmath203 in such a way that the configurations @xmath200 and @xmath201 are the only minima . to make them both global
, we specify @xmath204 such that the following equation is true @xmath205 in contrast to the zero - field situation , the spin flip between @xmath206 and @xmath207 is not a symmetry of the hamiltonian any more .
indeed , for the particular choice of the time @xmath131 and @xmath208 two ground states occur which are not related by any symmetry . as we described above , the conditioning more or less cancels in the direction of the original field , and one ends up with a model having alternating single - site terms ( external fields ) , which are pointing almost perpendicular to the original fields .
the coexistence of the ground states can , for temperature t low enough , be extended to coexistence of two gibbs measures , now not related by any spin - flip symmetry , for a slightly different choice of @xmath204 .
as the choice of the `` bad '' conditioning configuration which contains the @xmath209 can be continuously varied , we can deduce the existence of a time - interval of non - gibbsianness .
+ we remark that , unlike the two previous cases where one could use the reflection positivity property of the measure ( as well as the spin - flip symmetry ) in this case unfortunately we can not .
the interaction is not invariant under lattice reflections , so the measure is not reflection positive .
we have to use other techniques .
we will use a general contour argument from @xcite .
let us recall the statement .
let @xmath210 \subset \r$ ] and let @xmath211 be an ( n-1)-parameter family of potentials defined for @xmath212 varying in a neighbourhood of 0 in @xmath213 .
assuming the family @xmath214 satisfies the following conditions 1 .
the function @xmath214 is smooth in all its variables 2 .
for @xmath215 the function @xmath216 has n absolute minima at points situated on the diagonal of the square @xmath217 , i.e. @xmath218 3 . at the minima
@xmath219 the second differential of the function @xmath216 is strictly positive and moreover @xmath220 where @xmath25 is a sufficiently small constant , 4 . at points
@xmath219 , the differentials of @xmath214 at @xmath215 are nonzero , then there exists a point @xmath221 such that for the system described by the potential @xmath222 there exist at least n distinct limit gibbs distributions .
we want to apply the above theorem . for the assumptions to be satisfied we have to transform and shift our potential @xmath223 , to apply the statement about translation invariant potentials to a statement which also applies to periodic ones .
we will define our new potential @xmath224 on @xmath225 instead of @xmath217 as required in the assumptions .
this does not affect the proof in any essential way .
our spin space @xmath226 is isomorphic to @xmath227 $ ] by the isomorphy @xmath228 , where 0 and 1 are considered to be the same points .
we abbreviate @xmath229 .
let @xmath230 be a smooth function around a small neighbourhood of 0 in @xmath231 and let @xmath232 .
we define the new potential @xmath233 as being a sufficiently smooth function of all its variables .
furthermore let the differentials of @xmath234 at @xmath215 be nonzero . for @xmath215
the potential is given by @xmath235 note that it is isomorphic to @xmath236 .
then obviously @xmath237 inherits the two minima from @xmath236 which we call @xmath238 and @xmath239 .
the second assumption is satisfied by the definition of @xmath237 .
let us further examine the determinant of the hessian matrix to check the third condition .
we call @xmath240 the hessian matrix for the function @xmath241 .
then one observes that for the determinant of the hessian @xmath242 which is strictly positive at the minimal points @xmath243 and @xmath244 for the parameters @xmath131 and @xmath208 chosen above and @xmath170 big enough . then using the theorem
we deduce that for a sufficiently large @xmath170 there exists a @xmath245 such that for the system described by @xmath246 there exist at least 2 distinct gibbs measures . since @xmath246 and @xmath247 are physically equivalent this follows also for our potential @xmath236 .
@xmath248 let us now present a recovery result which will be valid for all strengths of the initial field at sufficiently low temperatures .
let @xmath164 be given , then for @xmath131 large enough and @xmath170 large enough , for example of order @xmath249 , there is a unique time - evolved measure .
* proof : * + we want to prove that for large enough times the time - evolved measure is unique . to prove this we want to use theorem 7 from @xcite which is a pirogov - sinai type argument for continuous models with one ground state .
let us cite their theorem 7 .
let @xmath250 \subset \mathbb{r}$ ] and let us consider the lattice @xmath149 .
suppose that the function @xmath251 is smooth in a neighbourhood of @xmath252 and on @xmath217 achieves an absolute minimum at @xmath252 .
let us also suppose that @xmath253 .
moreover let the expansion of @xmath251 in a neighbourhood of @xmath252 have the form @xmath254 where @xmath25 is a small ( positive or negative constant ) .
then there exists a temperature @xmath255 such that for @xmath256 there exists a unique limit gibbs distribution which depends analytically on @xmath170 .
@xmath257 is the state space of the spins and @xmath224 is the potential on the product space @xmath217 .
so all we have to do is again rewrite our potential and prove the assumptions given in the theorem . our original potential without approximation
is given by @xmath258 it is defined including the inverse temperature @xmath170 , which does not pose a problem . for @xmath131
large enough the unique minimum of @xmath223 is equal to @xmath252 .
let us rescale the potential @xmath259 by @xmath260 and consider the isomorphy @xmath261/2\pi \simeq [ -1,1]$ ] where @xmath262 and @xmath101 are considered to be the same points .
moreover we subtract a constant from the potential to assure that @xmath263 , i.e. @xmath264 we call @xmath265 and write the above potential as @xmath266 - \frac{\beta h}{4}\biggl [ \cos(\sigma^{\prime}_{i } ) -1 + \cos(\zeta^{\prime}_{i+1 } ) -1 \biggr ] \cr & + f^t(\sigma^{\prime}_{i},\zeta^{\prime}_i ) + f^t(\zeta^{\prime}_{i+1},\eta^{\prime}_{i+1 } ) .
\end{split}\ ] ] around the absolute minimum @xmath252 , we have the following expansion of @xmath267 , using the abbreviation @xmath268 : @xmath269 we clearly see that the expansion gives us for @xmath131 large enough , at least bigger than @xmath270 , the desired quadratic form required for theorem 7
. @xmath248
mean - field models are spin systems whose distribution is permutation - invariant . in @xcite
the gibbs properties of various mean - field models ( with finitely many spin values ) were investigated when subjected to various transformations . in the recent study in @xcite ,
extensions to more general mean - field models with compact polish spaces as their single - site spin space are discussed .
we describe these results in this section , restricting to the case where the spins take values on a sphere .
let us start by recalling the general notion of mean - field models and what it takes for them to be gibbsian .
we now present the definition of general mean - field models and mean - field gibbsianness for such models for @xmath0vector spins @xcite .
we write @xmath271 for the volume at size @xmath272 . [ mean - field - model ] for each @xmath273 , let @xmath274 be a probability measure on the space @xmath275 . 1 .
we refer to the sequence of the probability measures @xmath276 as a * mean - field model * if the @xmath274 s are permutation invariant .
a mean - field model @xmath276 is said to be gibbsian if the following holds : 1 .
the limit of conditional probabilities @xmath277 exists for any sequence @xmath278 of conditionings for which the empirical distribution converges weakly as @xmath272 tends to infinity , @xmath279 .
the function @xmath280 is weakly continuous . in the above , gibbsianity of mean - field models
is defined in terms of the asymptotic behavior of a sequence of measures instead of a single limiting measure .
this is in contrast to the lattice case where we just investigated the single infinite - volume measure .
the results one would get by only looking at the infinite - volume limit measures for mean - field systems would provide a lot less , and in some sense misleading , information .
indeed , such limit measures are either product measures , and thus trivially gibbsian , or nontrivial mixtures of product measures and thus highly non - gibbsian ( see for this fact @xcite ) .
the notion of gibbsianness given in definition [ mean - field - model ] is equivalent to the one considered in @xcite for the corresponding curie - weiss model ( for which of course one has a simpler single - site spin space and measure ) .
this is the case since the distribution of a binary random variable is uniquely determined by its mean .
hence for the curie - weiss model conditioning on the empirical averages gives the same information as conditioning on empirical measures .
for the rest of this section the term `` gibbsian '' should be taken in the sense provided by definition [ mean - field - model ] . in the above
we have defined general mean - field models .
we are now going to prescribe a class of mean - field models given via some potential functionals defined on the space of measures on the single - site spin space introduced in @xcite . in the following
we have denoted by @xmath281 and @xmath282 the spaces of finite signed measures and finite measures on @xmath7 respectively .
[ mfinteraction ] we shall refer to a map @xmath283 as a proper mean - field interaction if : 1 .
it is weakly continuous , 2 .
it satisfies the uniform directional differentiability condition , meaning that , for each @xmath284 the derivative @xmath285 at @xmath286 in direction @xmath18 exists and we have @xmath287 with @xmath288 uniformly in @xmath289 for which @xmath290 , for @xmath291 $ ] and 3 .
@xmath285 is a continuous function of @xmath292 standard examples are the quadratic interactions for the @xmath293state mean - field potts and the curie - weiss model ( which are defined on the product spaces of finite single - site spin spaces , instead of @xmath3-spheres , with symmetric a priori measure ) and which are respectively given by @xmath294 where @xmath295 is the mean of the measure @xmath286 . for each mean - field interaction @xmath13 and each @xmath273
we define the * finite - volume hamiltonian * @xmath296 ( a real - valued function on the product space @xmath275 ) as @xmath297 where @xmath298 is the empirical measure of the configuration @xmath299 . observe from the permutation invariance of the empirical measures that @xmath296 is permutation invariant . with this notation the sequence of probability measures
@xmath300 associated with the finite - volume hamiltonians @xmath296 at inverse temperature @xmath170 given by @xmath301 is a mean - field model ( on @xmath7 associated with @xmath13 and the a priori measure @xmath8 ) . in the above we
have used @xmath302 to denote the tensor product of measures . in the following ,
unless otherwise stated , the inverse temperature @xmath170 will be absorbed into the interaction @xmath13 and we write @xmath274 instead of @xmath303 .
we will , with abuse of notation , write @xmath274 for the sequence @xmath276 .
it is show in proposition 2.4 of @xcite that the mean - field models obtained in this way are gibbsian .
having disposed of the discussion on gibbsianness for general @xmath0vector mean - field models , we now turn our attention to discussing gibbs properties of transforms of gibbsian @xmath0vector mean - field models .
we now review the notion of gibbsianness for transformed gibbsian @xmath0vector mean - field models as found in @xcite .
we take @xmath52 as the single - site spin space for the transformed system , which we also assume to be a compact complete separable metrizable space .
further , we let @xmath304 be some appropriately chosen a priori probability measure on @xmath52 .
now we take @xmath305 as the joint a priori probability measure on @xmath306 given by @xmath307 where @xmath308 just as we had before for the corresponding transformed lattice spin models . now
the question of interest is the following .
starting with an initial sequence of mean - field gibbs measures @xmath274 , associated to a fixed general mean - field interaction @xmath13 , will the transformed sequence of measures @xmath309 with @xmath310 density @xmath311 be gibbsian ?
in other words , will the transformed single - site kernel a ) exist , and b ) be a continuous function of the empirical measures of the conditionings ?
it is shown in theorem 3.10 of @xcite that the transformed mean - field model @xmath309 will remain gibbsian if a certain constrained potential function has unique global minimizer , uniformly over the domain of the constraint variable .
the ideology behind this theorem is the same as in the lattice : absence of hidden phase transition in the initial system , constrained to be mapped to a fixed configuration in the transformed system implies gibbsianity for the transformed system . in the mean - field situation estimates can be made explicitly . to see something concrete , the authors in @xcite focused attention on mean - field interactions @xmath13 of the form @xmath312,\ldots,\nu[g_l]),\ ] ] where @xmath313 are some fixed bounded non - constant real - valued measurable functions defined on @xmath7 , @xmath314 and @xmath315 is some twice continuously differentiable function . in the above we
have denoted by @xmath316 $ ] the expectation of @xmath313 with respect to @xmath286 .
we further set @xmath317 and @xmath318 .
additionally , we assume that @xmath319 is a lipschitz function from @xmath7 to @xmath320 , with lipschitz - norm @xmath321 where @xmath60 is the metric on @xmath7 .
we also denote by @xmath322 the sum of the oscillations of the components of @xmath323 , i.e. @xmath324 for any @xmath323 satisfying the above conditions we set @xmath325:\nu\in\pp(\s^n)\right\}}.\ ] ] note that @xmath326 is compact subset of @xmath320 by the boundedness of @xmath323 . in the sequel
we will write @xmath327 for the supremum of the matrix max norm of the hessian @xmath328 .
i.e. @xmath329 furthermore , we also set @xmath330 with the above interaction , it is proved in @xcite that the transformed system associated to any @xmath331 will remain gibbsian if a certain quantity is small . before we make the above result more precise , some more notation is in order .
we set @xmath332 with these notation we have the following theorem . [ generalint ] consider the transformed system @xmath309 associated to the initial mean - field model @xmath274 ( given by the interaction @xmath13 satisfying the above conditions ) and joint a priori measure @xmath80 described above .
suppose that @xmath333 then 1 .
the transformed system is gibbsian and 2 .
the single - site kernel @xmath334 of the transformed system satisfies the continuity estimate @xmath335 where @xmath336 and @xmath337 is the variational distance between @xmath338 and @xmath339 .
the above theorem is found in @xcite as theorem 4.3 .
the smallness of the quantity @xmath340 may come from two sources ; namely + 1 ) the smallness of @xmath341 , arising from the weakness of the interaction @xmath13 among the components of the initial system + and + 2 ) the smallness of @xmath342 , coming from the good concentration property of the conditional measures @xmath343 , uniformly in @xmath344 .
+ thus we could start with a very strong interaction , but if the measures @xmath345 are close to delta measures then the transformed system will be gibbsian .
an advantage of this result is that it provides explicit continuity estimates for @xmath334 whenever the transformed system is gibbsian , which were lacking in all the results before .
however , it has the drawback that the estimates it provides for the regions in parameter space where the transformed system is gibbsian might not be sharp , as techniques employed in @xcite and @xcite do provide .
we now review two examples discussed in @xcite , which are reminiscent of some of the results found in @xcite . here
we present the result found in @xcite but for general mean - field interactions @xmath13 given in terms of @xmath346 and @xmath323 .
we study the gibbs properties of the transformed ( time - evolved ) system @xmath347 obtained upon application of infinite - temperature diffusive dynamics to the initial gibbsian mean - field model @xmath274 , associated with @xmath13 . in this set - up @xmath348 .
the joint single - site a priori measure @xmath80 is then given as in of subsection [ stimegibbs ] .
the following theorem is the result about the short - time conservation of gibbsianness for the time - evolved system @xmath347 .
[ thm2 ] suppose we have @xmath349 , then the time - evolved system @xmath347 will be gibbsian and @xmath350 , the single - site kernel for @xmath347 , has the continuity estimate @xmath351 observe from theorem [ thm2 ] that the time - evolved measure will be gibbsian whenever either the initial interaction is weak or @xmath131 is small enough .
the above result was only stated in @xcite for the corresponding curie - weiss model .
we present this case below .
for the curie - weiss rotator model the interaction for the initial system is given by @xmath352,\cdots,\nu[\s^{n+1}_i]\bigr)=-\dfrac{\beta \sum_{j=1}^{n+1 } \nu[\s^j_i]^2}{2},\ ] ] where @xmath353 is the @xmath108th coordinate of the point @xmath354 and @xmath355 .
as a corollary to theorem [ thm2 ] we have the following short - time gibbsianness result for the curie - weiss rotator model under diffusive time evolution . [ cor3 ]
suppose we have @xmath356 , then the time - evolved system @xmath347 will be gibbsian and @xmath350 , the single - site kernel for @xmath347 has the continuity estimate @xmath357 corollary [ cor3 ] is found in @xcite as lemma 5.1 .
this corollary is reminiscent of the result in theorem 2.2 of @xcite , where the curie - weiss model under independent spin - flip dynamics was studied .
it is shown therein that if @xmath170 is small enough ( weak initial interaction ) , then the time - evolved system will be gibbsian forever , but if @xmath170 is large , then the time - evolved system will only be gibbsian for short times .
consider general @xmath346 and @xmath323 as above , and decompose @xmath7 into countably many pairwise disjoint subsets ( countries ) as in subsection [ fine ] above .
then with this notation it follows from theorem [ generalint ] that [ prop1 ] if the quantity @xmath358 , then the transformed system is gibbsian and the single - site kernel @xmath334 satisfies the continuity estimate @xmath359 the above proposition can be found in lemma 5.2 of @xcite .
thus the transformed system @xmath309 will be gibbsian if either the initial interaction @xmath13 is weak or the local coarse - grainings ( i.e. the @xmath135 ) are fine enough .
in other words : if we have initial gibbsian mean - field system with spins living on the sphere and we partition the sphere into countries , representing each country by a distinct point in @xmath52 , then the resultant transformed system will be gibbsian if the countries are small enough .
let us mention in this context the result of theorem 1.2 of @xcite for the corresponding fuzzy potts mean - field model .
in that paper it was shown that the transformed system will be gibbsian at all temperatures whenever the sets of points contracted into single points by the fuzzy map have cardinality at most @xmath360 . in this section
specialize to the quadratic mean - field rotator model on the circle , where we focus now on the interesting case @xmath361 .
in fact , although we do not treat the simpler case @xmath362 here , one can in a very similar way prove loss of gibbsianness , again just as in the lattice situation .
+ we start with the measure @xmath363 where @xmath364 is a vector - sum in @xmath365 and @xmath366 is the equidistribution .
we take a time - evolution with the transition kernels @xmath367 describing brownian motion on the circle , as above .
we are interested in the gibbsian character of the time - evolved measures @xmath368 in the sense of continuity of limiting conditional kernels , as described above .
the virtue of mean - field models is that we can describe the limiting kernels explicitly . by this
we mean a description in terms of a minimization problem of an explicit expression .
this has been done in the general setup of site - wise independent transformations in @xcite .
for the present case we get the following concrete results .
the limiting kernels @xmath369 of the time - evolved mean - field models @xmath370 are given by the formula @xmath371 for all choices of the ( non - negative ) parameters @xmath372 and the conditioning @xmath373 ( in the probability measures on the circle ) , for which the minimizer ( in the closed unit disk ) @xmath374 is unique with @xmath375 we do not give a proof of @xmath376 here ( which can be deduced from the general result in @xcite ) , but we sketch a fast heuristics which explains what happens : note first that @xmath377 denotes the rate function ( up to an additive constant ) of the initial model , constrained to have an empirical distribution @xmath373 in the transformed model .
conditioning the empirical distribution of the transformed spins outside the site @xmath101 to @xmath373 produces a quenched system involving the initial spins which acquires the magnetisation @xmath378 .
the propagation of the corresponding distribution of @xmath379 to @xmath380 with the kernel @xmath381 gives the desired conditional probability distribution @xmath382 .
compare the rate function to the well - known rate - function of the initial model given by @xmath383 the map @xmath384 has a unique minimizer @xmath385 , if @xmath361 is arbitrary , pointing in the direction of @xmath164 . the input to understand the large time - behavior is the fact that the kernel @xmath386 converges to the equidistribution @xmath366 , uniformly in @xmath85 .
from this we see that , at fixed @xmath387 , the functions @xmath388 converge to @xmath384 , uniformly in @xmath373 .
the same holds for higher derivatives w.r.t .
these statements imply that , for @xmath131 sufficiently large , for all choices of @xmath373 there is only one minimizer @xmath378 .
looking at the linear appearance of the measure @xmath373 in , we see that @xmath378 changes continuously under a variation of @xmath373 . by the form of this
implies gibbsianness . to prove non - gibbsianness at the parameter - triple @xmath390 we use the formula for the limiting kernels for those quadruples @xmath391 where they are well - defined and , for fixed @xmath390 we show that there exists a @xmath392 at which the limiting kernels are not continuous .
to do so , it suffices to exhibit a one - parameter trajectory @xmath393 which is continuous in the weak topology s.t . 1 .
@xmath394 has unique minimizers for @xmath208 in a neighborhood of @xmath395 , + excluding @xmath395 2 .
@xmath396 so far , the reasoning is general . now , to create a phase transition in the constrained model , a suitable choice of @xmath373 which is able to balance the influence of the external magnetic field @xmath164 has to be found .
we choose conditionings of the type @xmath397 where @xmath398 denotes the vector on the circle corresponding to the angle @xmath399 .
this conditioning mimicks the choice of conditionings on @xmath400 obtained by putting @xmath401 on the even sublattice and @xmath402 on the odd sublattice .
let @xmath403 large enough , and @xmath404 be given .
then there exists a time - interval for which there exists an @xmath405 such that @xmath406 has two different global minimizers of the form @xmath407 , pointing in the direction or in the opposite direction of @xmath164 .
we provide an explanation of this phenomenon .
let us look at the rate - function for magnetisation - values pointing in the direction of @xmath164 , in the conditioning @xmath408 which reads @xmath409 with the diffusion kernel on the sphere written in angular coordinates @xmath399 as : @xmath410 for fixed parameters @xmath411 , we use the new magnetization variable @xmath412 to rewrite @xmath413 where @xmath414 the second term on the l.h.s .
is an unimportant constant .
this choice of parameters is handy because we have separated the influence of the parameters , and moreover , two of them are appearing only linearly in our 4-parameter family .
let us explain how a balance between @xmath208 and @xmath415 can be used to create a situation of a pair of different equal depth - minimizers , without going into the details of the analysis of the regions in parameter - space where this can be done . for this heuristic argument ,
let us fix the @xmath208 first .
we note that @xmath416 is convex , so @xmath417 has a chance to have more than one local minimum , for good choices of @xmath418 .
having found such a situation , a tuning of the @xmath415 will result in a tilting of the rate - function which can create a situation where this pair has an equal depth in the full function @xmath419 .
the mechanism described provides us with a curve in the space of @xmath208 and @xmath415 where the two minima have equal depth .
now , fixing a value of @xmath415 and varying the @xmath420 across this curve , yields a jump of the global minimizer which implies non - gibbsianness . in the pictures , showing the plot of @xmath421
we see this mechanism at work .
, right : @xmath422,title="fig:",height=170 ] , right : @xmath422,title="fig:",height=170 ] @xmath423 it is clear from the above diagram that for @xmath424 there is a choice of @xmath425 such that @xmath426 has two global minimizers .
numerically we find @xmath427 .
hence at such values for @xmath428 and @xmath131 , the transformed system will be non - gibbsian .
we thank roberto fernndez and frank redig for many discussions .
the possibility thereof was due to support from the ndns mathematics research cluster .
v.e . thanks sacha friedli and bernardo de lima for inviting him to a wonderful school in ouro preto and maria eulalia vares for the invitation to write a review on the topics presented there .
van enter , r. fernndez and a.d .
sokal : regularity properties and pathologies of position - space renormalization - group transformations : scope and limitations of gibbsian theory , j. stat . phys .
72 , 879 - 1167 ( 1993 ) | we review some recent developments in the study of gibbs and non - gibbs properties of transformed @xmath0vector lattice and mean - field models under various transformations .
also , some new results for the loss and recovery of the gibbs property of planar rotor models during stochastic time evolution are presented .
* ams 2000 subject classification : * 82b20 , 82b26 , 60k35 . *
keywords : * gibbs measures , non - gibbsian measures , @xmath0vector lattice models , @xmath0vector mean - field models , transformed model , dobrushin uniqueness , cluster expansion , spin - flop transitions . |
the elucidation of dynamo action is a problem of central importance in nonlinear dynamics because it has implications for a variety of physical systems
. dynamo instabilities , which amplify weak magnetic fields in a turbulent conducting fluid , are believed to be the principal mechanism for the generation of magnetic fields in celestial bodies and in the interstellar medium @xcite , and in liquid - metal systems @xcite studied in laboratories . in these situations
the kinematic viscosity @xmath3 and the magnetic diffusivity @xmath4 can differ by several orders of magnitude , so the magnetic prandtl number @xmath5 can either be very small or very large ; e.g. , @xmath6 at the base of the sun s convection zone , @xmath7 in the liquid - sodium system , and @xmath8 in the interstellar medium .
this prandtl number is related to the reynolds number @xmath9 and the magnetic reynolds number @xmath10 that characterize the conducting fluid ; here @xmath11 and @xmath12 are typical length and velocity scales in the flow ; clearly @xmath13 .
two dissipative scales play an important role here ; they are the kolmogorov scale @xmath14 [ @xmath15 at the level of kolmogorov 1941 ( k41 ) phenomenology @xcite ] and the magnetic - resistive scale @xmath16 [ @xmath17 in k41 ] . for large prandtl numbers ,
i.e. , @xmath18 , @xmath19 so the magnetic field grows predominantly in the dissipation range of the fluid till it is strong enough to affect the dynamics of the fluid through the lorentz force .
this behavior is a characteristic of a small - scale turbulent dynamo , in which dynamo action is driven by a smooth , dissipative - scale velocity field . in the initial stage of growth ,
called the kinematic stage of the dynamo , the magnetic field is not large enough to act back on the velocity field .
dynamo action can be obtained for values of @xmath1 that are large enough to overcome joule dissipation ; and the dynamo - threshold value @xmath20 decreases as @xmath0 increases @xcite .
@xmath21 in liquid - metal flows @xcite so they lie in the small - prandtl - number region , @xmath22 , for which the growth of the magnetic energy occurs initially in the _ inertial scales _ of fluid turbulence , because @xmath23 ; here the velocity field is not smooth and the local strain rate is not uniform : at the k41 level the turnover velocity of an eddy of size @xmath24 is @xmath25 , so the rate of shearing @xmath26 .
direct numerical simulations ( dns ) are playing an increasingly important role in developing an understanding of such dynamo action .
most dns studies of mhd turbulence @xcite have been restricted , because of computational constraints , either to low resolutions or to the case @xmath27 ; small - scale dynamos , with @xmath28 have also been studied via dns @xcite .
however , given the large range spanned by @xmath0 in the physical settings mentioned above , some recent dns studies of the mhd equations have started to explore the @xmath0 dependence of dynamo action ; the range of @xmath29 covered by such pure dns studies @xcite is quite modest ( @xmath30 ) . to explore the dynamo boundary in the @xmath2 plane over a large range of @xmath0 ,
one recent study @xcite has used a combination of numerical methods , some of which require small - scale models like large - eddy simulations ( les ) or lagrangian - averaged mhd ( lamhd ) , and others , like a pseudospectral dns , in which the only approximations are the finite number of collocation points and the finite step used in time marching ; yet another dns study @xcite has introduced hyperviscosity of order @xmath31 to study the low @xmath0 regime ; by using this combination of methods these studies has been able to cover the range @xmath32 and to obtain the boundary between dynamo and no - dynamo regions but with fairly large error bars .
we have carried out extensive , high - resolution , numerical studies that have been designed to explore in detail the boundary between the dynamo and no - dynamo regimes in the @xmath2 plane in a shell model for three - dimensional mhd @xcite .
this shell model allows us to explore a much larger range of @xmath0 than is possible if we use the mhd equations .
although our study uses a simple shell model , it has the virtue that it can explore the boundary between dynamo and no - dynamo regions in great detail without resorting to the modelling of small spatial scales .
shell - model studies of dynamo action have also been attempted in refs .
@xcite but these have concentrated on aspects of the dynamo problem that are different from those we consider here .
our study suggests that it is natural to think of the boundary between dynamo and no - dynamo regimes in the @xmath2 plane as a first - order phase boundary that is the locus of first - order , nonequilibrium phase transitions from one nonequilibrium statistical steady state ( ness ) to another .
the first ness is a turbulent , but statistically steady , conducting fluid in which the magnetic energy is negligibly small compared to the kinetic energy ; the second ness is also a statistically steady turbulent state but one in which the magnetic energy is comparable to the kinetic energy . indeed , the ratio of the magnetic and fluid energies @xmath33 turns out to be a convenient order parameter for this nonequilibrium phase transition since it vanishes in the no - dynamo phase and assumes a finite , nonzero value in the dynamo state .
the other , intriguing result of our study is that the boundary between these phases is very intricate and might well have a fractal character ; this provides an appealing explanation for the large error bars in earlier attempts to determine this boundary @xcite .
the analogy with first - order transitions that we have outlined above is not superficial . as in any first - order transition
we find that our order parameter shows hysteretic behavior @xcite as we scan through the dynamo boundary by changing the forcing term at a nonzero rate .
we also find some evidence of nucleation - type phenomena : the closer we are to the dynamo boundary , the longer it takes for a significant magnetic field to nucleate and thus lead to dynamo action .
we compare our results with earlier studies such as ref .
@xcite , which have suggested that dynamo action occurs because of a subcritical bifurcation .
the remaining part of this paper is organised as follows : in sec .
[ models ] we describe the shell model for mhd @xcite and the numerical method we employ . sec .
[ results ] is devoted to our results and sec .
[ conclusion ] contains a concluding discussion .
to study dynamo action it is natural to use the equations of magnetohydrodynamics(mhd ) . in three dimensions
the mhd equations are @xmath34 where @xmath3 and @xmath4 are the kinematic viscosity and the magnetic diffusivity , respectively , the effective pressure @xmath35 , and @xmath36 is the pressure . for low - mach - number flows ,
to which we restrict ourselves , we use the incompressibility condition @xmath37 ; and @xmath38 .
as we have mentioned above , a dns of the mhd equations poses a significant computational challenge , even on the most powerful computers available today , if we want to cover a large part of the @xmath2 plane and to locate the dynamo boundary accurately .
therefore one study @xcite has used a combination of les , lamhd , and dns to obtain this boundary .
we employ a complementary strategy : we use a simple shell model for mhd @xcite that allows us to carry out very extensive numerical simulations to probe the nature of the dynamo boundary without using les or lamhd .
shell models comprise a set of ordinary differential equations with nonlinear coupling terms that mimic the advection terms in and respect the shell - model analogs of the conservation laws of the parent hydrodynamic equations in the inviscid , unforced limit @xcite . for the case of mhd
each shell @xmath39 is characterized by a complex velocity @xmath40 and a complex magnetic field @xmath41 in a logarithmically discretized fourier space with wave vectors @xmath42 ; furthermore , there is a direct coupling only between velocities and magnetic fields in nearest and next - nearest neighbor shells .
the mhd shell model equations @xcite are @xmath43^\ast+f_n^u , \\
\frac{db_n}{dt}= & -&\eta k_n^2b_n + i[d_n(u_{n+1}b_{n+2}-b_{n+1}u_{n+2})\nonumber \\ & + & e_n(u_{n-1}b_{n+1}-b_{n-1}u_{n+1})\nonumber \label{shell2}\\ & + & f_n(u_{n-2}b_{n-1}-b_{n-2}u_{n-1})]^\ast+f_n^b , \end{aligned}\ ] ] where @xmath44 denotes complex conjugation , @xmath45 , with @xmath46 the total number of shells , the wave numbers @xmath47 , with @xmath48 , and @xmath49 and @xmath50 the forcing terms in the equations for @xmath40 and @xmath41 , respectively . in our studies of dynamo action , we set @xmath51 .
the parameters @xmath52 , are obtained by demanding that these equations conserve all the shell - model analogs of the invariants of 3dmhd , in the inviscid , unforced case , and reduce to the well - known gledzer - ohkitani - yamada ( goy ) shell model @xcite for fluid turbulence if @xmath53 , @xmath54 . in particular , to ensure the conservation of shell - model analogs of the total energy @xmath55 , cross helicity @xmath56 , and magnetic helicity @xmath57 , in the unforced and inviscid case , and to obtain the goy - model limit for the fluid , we choose @xmath58 the only adjustable parameters are the forcing terms and @xmath3 and @xmath4 .
the ratio @xmath59 yields the magnetic prandtl number @xmath0 .
grashof numbers yield nondimensionalized forces @xcite but , for easy comparison with earlier studies @xcite , we use the fluid and magnetic integral - scale fluid and magnetic reynolds numbers whose shell - model analogs are , respectively , @xmath60 , where @xmath61 and @xmath62 , @xmath63 and @xmath64 .
we use the following boundary conditions : @xmath65 we set @xmath66 and use a fifth - order , adams - bashforth scheme for solving the shell - model equations , i.e. , for an equation of the type @xmath67 we use @xmath68,\end{aligned}\ ] ] where @xmath69 is the time step .
we have found that this numerical scheme works well for the integration of eqs.([shell1 ] ) and ( [ shell2 ] ) so long as @xmath70 and @xmath71 . in all our calculations we use @xmath72 .
characteristic time scales include the time scale for diffusion @xmath73 and the large - eddy - turnover time @xmath74 , where @xmath75 is the box - size length scale and @xmath76 is the root - mean - square velocity .
the initial conditions we use are as follows : we first obtain a statistically steady state for the goy - shell - model equations , which are obtained from eq.([shell1 ] ) by setting all @xmath77 ; the forcing terms are chosen to be @xmath78 , with @xmath79 in all our runs , except ones in which we study hysteretic behavior , and @xmath51 .
we choose the goy - model shell velocities at time @xmath80 to be @xmath81 , with @xmath82 a random phase distributed uniformly on the interval @xmath83 . to make sure we have a statistically steady state we evolve the shell velocities @xmath40 till @xmath84 .
this yields the shell - model energy spectrum @xmath85 that has the k41 form @xmath86 if we ignore intermittency corrections .
we now introduce a small seed magnetic which is such that @xmath87 and then follow the temporal evolution of @xmath40 and @xmath41 that is given by eqs.([shell1 ] ) and ( [ shell2 ] ) .
versus time @xmath88 , with @xmath89 the magnetic - diffusion time , in the dynamo region ( blue , dashed curve ) , near the dynamo boundary ( green , full line ) , and in the no - dynamo regime ( red , full line in the inset).,title="fig : " ] + 50 30 50 30 [ fig : energy ] + 50 30 plane : red circles indicate dynamo action ; green stars are used if no dynamo occurs .
the boundary between the two regions shows an intricate , interleaved pattern of fine , dynamo and no - dynamo regimes ( see inset for a detailed view ) .
we have drawn two black , dashed lines ; the region above the upper one of these lines is predominantly in the dynamo regime ; the area below the lower one of these lines is principally in the no - dynamo regime .
, title="fig : " ] 50 30 50 30 50 30 given the numerical scheme that we have described in the previous section , we obtain the time series for @xmath40 and @xmath41 from the mhd - shell - model equations .
an analysis of these time series shows two types of nonequilibrium statistical steady states ( ness ) .
we refer to the first as the no - dynamo state and to the second as the dynamo state .
these states have been found in several earlier studies such as refs .
our main goal is to explore in detail the phase boundary between these two states .
this can be done most easily by the introduction of a dynamo order parameter a natural candidate for which is the ratio @xmath33 , where the fluid and magnetic energies are , respectively , @xmath90 in fig .
[ fig : tseries ] : the blue , dashed curve shows the evolution of @xmath33 in the dynamo regime ; note that here the dynamo order parameter rises rapidly , fluctuates significantly for @xmath91 , and finally reaches a statistical steady state with equipartition , i.e. , @xmath92 .
the red , full curve in the inset of fig .
[ fig : tseries ] shows how @xmath33 vanishes rapidly in the no - dynamo state .
the behavior of the dynamo order parameter is more complicated than these two simple possibilities in the vicinity of the phase boundary between dynamo and no - dynamo states as shown by the green , full curve in fig .
[ fig : tseries ] ; @xmath33 rises much more slowly from zero than in the dynamo regime and then it fluctuates significantly for a long time ; the difficulty of pinpointing the dynamo boundary is a consequence of these fluctuations .
the time series for the dynamo order parameter are obtained from those for @xmath93 and @xmath94 ; representative plots for these are shown , via red and blue - dashed curves , in figs .
[ fig : energy](a ) , ( b ) , ( c ) , ( d ) , ( e ) , and ( f ) for a very large range of magnetic prandtl numbers , namely , @xmath95 and @xmath96 , respectively .
the values of @xmath3 are ( a ) @xmath97 , ( b ) @xmath97 , ( c ) @xmath96 , ( d ) @xmath98 , ( e ) @xmath99 , and ( f ) @xmath98 ; and the corresponding values of the diffusion time scale @xmath100 and @xmath101 , respectively .
clearly dynamo action occurs in figs .
[ fig : energy](a)-(d ) but not figs .
[ fig : energy](e ) and ( f ) . by obtaining many such plots we can identify the dynamo boundary in the @xmath102 plane as we discuss later . in the dynamo regime
the shell - model kinetic and magnetic energy spectra defined , respectively , by @xmath103 and @xmath104(a ) , ( b ) , and ( c ) for @xmath105 and @xmath106 , respectively , we show the evolution of @xmath107 with time : the curves with red stars , green diamonds , blue hexagons , cyan circles , and magenta triangles , are obtained , respectively , for @xmath108 and @xmath109 ; the analogs of these plots for @xmath110 are given in figs .
[ fig : spectra](d ) , ( e ) , and ( f ) .
note that the initial growth of @xmath107 occurs principally at large values of @xmath42 if @xmath0 is large , i.e. , we have a small - scale dynamo ; this growth of @xmath107 moves to low values of @xmath42 as @xmath0 decreases ; earlier studies @xcite have observed similar trends but not over the large range of @xmath0 we cover .
as @xmath107 grows , the velocity spectra are also affected but much less than their magnetic counterparts as can be seen by comparing figs .
[ fig : spectra](d ) , ( e ) , and ( f ) with figs .
[ fig : spectra](a ) , ( b ) , and ( c ) , respectively . in all these plots
the curves with black squares indicate @xmath110 from the initial steady state for the goy shell model ; and the black lines with no symbols show the k41 @xmath111 spectrum for comparison . from this line
we see that the ness that is obtained , once dynamo action has occurred , is such that both velocity and magnetic - field energy spectra display a substantial inertial range with k41 scaling ; these inertial ranges are not large enough , at least near the dynamo boundary in our runs , for a reliable estimation of multiscaling corrections to the @xmath112 exponent .
if @xmath113 then the scaling ranges in velocity and magnetic - field spectra are comparable ; as @xmath0 decreases ( increases ) , the scaling range for the magnetic spectrum decreases ( increases ) relative to its counterpart in the velocity spectrum ; these trends are clearly visible in the representative plots in fig .
[ fig : spectra ] .
we return now to the identification of the dynamo boundary . a close scrutiny of the plots in fig .
[ fig : energy ] shows that the initial growth of @xmath94 is not monotonic .
it is important , therefore , to set a threshold value of the magnetic energy @xmath114 : for a given pair of values for @xmath0 and @xmath1 , if @xmath115 for @xmath116 , where @xmath117 is the time at which the threshold value is crossed , we conclude that dynamo action occurs ; if not , then there is no dynamo formation . by examining the growth of @xmath118 we
can , therefore , map out the dynamo boundary in the @xmath119 plane .
the crossing time @xmath117 depends on @xmath0 and @xmath120 .
note that , if @xmath121 , the length of time for which we integrate eqs.([shell1 ] ) and ( [ shell2 ] ) , we would conclude , _ incorrectly _ , that no dynamo action occurs for this of values of @xmath0 and @xmath1 .
in other words the dynamo boundary depends on @xmath122 ; we have checked this explicitly in several cases .
an important questions arises now : is there a well - defined dynamo boundary in the @xmath2 plane as @xmath123 ?
earlier studies @xcite have began to answer this question .
they find that , if @xmath124 , then a well - defined dynamo boundary is obtained .
however , since they work with the mhd equations the error bars on this boundary are large and the range of values of @xmath0 and @xmath1 rather limited .
the simplicity of our model allows us to carry out a systematic study of the dynamo boundary .
we find that , at least in our shell model for mhd , we can obtain an asymptotic dynamo boundary ( see fig . [
fig : stability ] ) if we choose @xmath125 , i.e. , we conclude that dynamo action has occurred if @xmath118 exceeds @xmath126 ; furthermore , if @xmath118 falls below @xmath127 we say that dynamo action will never be achieved .
we continue the temporal evolution of eqs.([shell1 ] ) and ( [ shell2 ] ) till one of these criteria is satisfied . for all values of @xmath0 and @xmath120 that we have used
we find that this @xmath122 , the run time required to decide whether or not dynamo action occurs , is several orders of magnitude lower than @xmath89 .
we have also checked for several representative pairs of values for @xmath0 and @xmath1 that runs of length @xmath124 do not change our conclusions about such dynamo action . the dynamo boundary that we obtain is shown in the stability diagram of fig .
[ fig : stability ] .
red circles indicate parameter values at which we obtain dynamo action whereas green stars are used for values at which no dynamo occurs .
the most important result that follows from this stability diagram is that the boundary between dynamo and no - dynamo regimes is very complicated .
it seems to be of fractal - type , with an intricate pattern of fine , dynamo regions interleaved with no - dynamo regimes .
this is especially apparent in the inset of fig .
[ fig : stability ] , which shows a detailed view of the stability diagram in the vicinity of the dynamo boundary .
earlier studies seem to have missed this fractal - type of boundary because they have not been able to examine the transition in as much detail as we have for our shell model .
however , fractal - type boundaries between different dynamical regimes have been suggested in other extended dynamical systems ; recent examples include the transition to turbulence in pipe flow @xcite and different forms of spiral - wave dynamics in mathematical models for cardiac tissue @xcite . in fig .
[ fig : stability ] we have drawn two black , dashed lines ; the region above the upper one of these lines is predominantly in the dynamo regime ; the area below the lower one of these lines is predominantly in the no - dynamo regime .
these two lines give an approximate indication of the error bars we might expect in the determination of the dynamo boundary in a study that can not scan through points in the @xmath128 plane as finely as we have . from fig .
[ fig : tseries ] we see that the order parameter @xmath33 jumps from a very small value in the no dynamo region to a value @xmath129 in the dynamo state .
it is natural , therefore , to think of the dynamo boundary as a nonequilibrium , first - order boundary . in an equilibrium ,
first - order transition the order parameter shows hysteretic behavior if we scan through a first - order boundary by , say , changing , at a finite rate , the field that is conjugate to the order parameter @xcite .
it is natural to ask if we see such hysteretic behavior at the dynamo boundary .
indeed , we do , as we show in fig . [
fig : hysterisis ] where we cross the dynamo boundary by changing the amplitude @xmath130 of the forcing term in eq.([shell1 ] ) .
figure [ fig : hysterisis ] shows representative plots of the dynamo order parameter @xmath33 versus @xmath130 ; these illustrate the hysteretic behavior that occurs when @xmath130 is cycled at a finite , nonzero rate across the dynamo boundary ; here @xmath131 and @xmath132 .
as @xmath130 increases , @xmath33 follows the blue , full line : it increases and then saturates ; fluctuations are superimposed on these mean trends .
if we now decrease @xmath130 , then @xmath33 follows the red dotted line , and not the blue one , i.e. , we have a hysteresis loop .
the faster the rate at which we change @xmath130 the wider is the hysteresis loop as is known from studies of hysteresis in spin systems @xcite . here
we increase @xmath130 in steps of @xmath133 from an initial value of @xmath133 ; we keep @xmath130 constant for a time period @xmath134 in fig . [
fig : hysterisis](a ) and @xmath135 in fig .
[ fig : hysterisis](b ) ; the red , dotted - line segments of the hysteresis loops are obtained by decreasing @xmath130 at the same rates as for the blue , full - line segments ; the loop in the former case is narrower than in the latter . given the analogy with first - order transitions that we have outlined above , it is natural to ask if nucleation - type phenomena @xcite are also associated with dynamo formation .
it would be interesting to check this in a dns of the mhd equations . at the level of our shell model , the best we can do is to try to see
if , for a given @xmath0 , when we obtain a dynamo , the time required for dynamo action @xmath117 diverges as we approach the dynamo boundary .
our data are consistent with an increase of @xmath117 as we approach this boundary from the dynamo side as shown by the representative plots in fig .
[ fig : time ] .
however , it is hard to fit a precise form to the behavior of @xmath117 near the dynamo boundary partly because of the complicated nature of this boundary which makes it difficult to estimate the position @xmath20 reliably ( the plot in fig . [
fig : time ] is motivated by the form of eq .
( 27 ) in ref .
we have presented a detailed study of dynamo action in a shell model of turbulence @xcite .
our study has been designed to explore the nature of the boundary between dynamo and no - dynamo regimes in the @xmath119 plane over a much wider range of @xmath0 than has been attempted in earlier numerical studies .
the dynamo boundary emerges as a first - order nonequilibrium phase boundary between one turbulent , nonequilibrium statistical steady state ( ness ) and another @xcite .
this point of view is implicit in earlier work , e.g. , in studies of the kazantsev dynamo @xcite or in studies that view dynamo generation as a subcritical bifurcation @xcite .
one of these studies @xcite has remarked that when dynamo action ... is obtained in a fully turbulent system , where fluctuations are of the same order of magnitude as the mean flow ... the traditional concept of amplitude equation may be ill - defined and one may have to generalize the notion of subcritical transition `` for turbulent flows ... '' .
we believe that the natural generalization is the nonequilibrium , first - order transition we suggest above .
we have explored the explicit consequences of such a view in far greater detail than has been attempted hitherto . in particular , the ratio @xmath33 is a convenient order parameter for this nonequilibrium phase transition ; it shows hysteresis across the dynamo boundary like order parameters at any first - order transition ; and nucleation - type phenomena also seem to be associated with dynamo formation .
last , and perhaps most interesting , we find that the dynamo boundary seems to have a fractal character ; this provides a natural explanation for the large error bars in earlier attempts to determine this boundary @xcite .
furthermore , this fractal - type boundary might well be the root cause of magnetic - field reversals discussed , e.g. , in refs .
@xcite .
it is important to check , of course , that our shell - model results carry over to the mhd equations .
this requires large - scale dns that might well be beyond present - day computing capabilities if we want to explore issues like the possible fractal nature of the dynamo boundary .
however , analogs of the hysteretic behavior we mention above have been obtained in dns studies of the mhd equations @xcite ; hysteresis has also been seen in a numerical simulation that includes turbulent convection @xcite . in some of these studies
hysteretic behavior is obtained by changing the viscosity of the magnetic prandtl number .
we have obtained hysteresis by changing the forcing ; this change of forcing might be easier to effect in experiments than a change of the viscosity or magnetic diffusivity . to the best of our knowledge ,
earlier studies have not noted the increase in the dynamo - formation time @xmath117 as the dynamo boundary is approached from the dynamo side . we have suggested that this is akin to the increase in the time required to form a critical nucleus as we approach a first - order boundary @xcite .
it would be interesting to see if such an increase of @xmath117 can be obtained in dns studies of dynamo formation with the mhd equations .
it is worth noting here that some dns studies @xcite have suggested that simulation times comparable to the diffusion time scale @xmath89 are required to confirm dynamo formation ; by contrast our shell - model study yields dynamo action on a much shorter time @xmath117 , which increases as we approach the dynamo boundary . perhaps the large simulation times required for dynamo action in full mhd simulation might have arisen because these simulations have been carried out in the vicinity of the dynamo boundary . to settle completely whether the dynamo boundary is of fractal - type , very long simulations might be required to make sure that the apparent fractal nature is not an artifact of long - lived metastable states
to make sure that our calculations do not suffer from such an artifact , we have carried out very long runs for representative points in the region of the dynamo boundary in fig .
[ fig : stability ] ; we have found that these long runs do not change our results .
furthermore , it is useful to check whether , instead of one dynamo boundary , there is a sequence of transitions , with more and more complicated temporal behaviors for the order parameter , as has been seen in the turbulence - induced melting of a nonequilibrium vortex crystal @xcite .
we have not found any conclusive evidence for this but , in the vicinity of the dynamo boundary , the order parameter can oscillate for fairly long times ( see , e.g. , the green full curve in fig . [ fig : tseries ] ) . to decide conclusively whether these oscillations characterize a new nonequilibrium oscillating state , different from the simple dynamo and no - dynamo nesss we have mentioned , requires extensive numerical studies that lie beyond the scope of this paper . in equilibrium
statistical mechanics different ensembles are equivalent ; in particular , we may determine a first - order phase boundary by using either the canonical or the grand - canonical ensemble .
however , such an equivalence of ensembles does not apply to transitions between different nonequilibrium statistical steady states ( nesss ) ; examples may be found in driven diffusive systems @xcite or in the turbulence - induced melting of a nonequilibrium vortex crystal @xcite .
given that the dynamo boundary separates two turbulent nesss , we might expect that this boundary might depend on precisely how the system is forced .
evidence for this exists already : for example , the dynamo boundary depends on whether a stochastic external force is used @xcite or whether a taylor - green force is used @xcite ; furthermore , this boundary is different if the fluid is helical @xcite , as in most astrophysical dynamos .
we thank r. karan , s.s . ray , and s. ramaswamy for discussions , serc(iisc ) for computational resources and dst , ugc and csir india for support .
one of us is a member of the international collaboration for turbulence research ( ictr ) .
et al . _ , new j. phys . * 4 * , 84 ( 2002 ) ; phys .
lett . * 92 * , 054502 ( 2004 ) ; new j. phys .
* 9 * , 300 ( 2007 ) ; a.a .
schekochihin , s.c .
cowley , and s.f .
taylor , astrophys .
j. * 612 * , 276 ( 2004 ) .
f. ptrlis and s. fauve , europhys . lett .
* 22 * , 273 ( 2001 ) ; * 76 * , 602 ( 2006 ) ; s. fauve and f. ptrlis , in _
peyresq lectures on nonlinear phenomena _ , edited by j .- a .
sepulchre ( world scientific , singapore , 2003 ) , vol .
2 , pp . 164 ; s. fauve , f. ptrlis , c.r .
physique * 8 * , 87 ( 2007 ) .
minnini , d.c .
montgomery , and a. pouquet , phys .
fluids * 17 * , 035112 ( 2005 ) ; p.d .
_ , astrophys .
j. * 626 * , 853 ( 2005 ) ; p. mininni , a. alexakis , and a. pouquet , phys .
e * 72 * , 046302 ( 2005 ) ; p.d .
mininni and a. pouquet , phys .
lett . * 99 * , 254502 ( 2007 ) .
j. leorat , p. lallemand , j.l .
guermond , and f. plunian , in _ dynamo and dynamics : a mathematical challenge _ , edited by p. chossat _
( kluwer academic , dordrecht , 2001 ) , p. 25 - 33 ;
r. stepanov and f. plunian , j. turbulence , * 7 * , n 39 ( 2006 ) .
shajahan , s. sinha , and r. pandit , phys .
e * 75 * , 011929 ( 2007 ) ; t.k .
shajahan , a.r .
nayak , and r. pandit , plos one * 4*(3 ) , e4738 ( 2009 ) .
oxtoby , j. phys . condens .
matter * 4 * , 7627 ( 1992 ) .
we use the expression first - order transition as in equilibrium statistical mechanics since the order parameter ( the first derivative of a free energy ) jumps at the transition . since there is no free energy for the nesss we consider , one can
, just as well , refer to this as a discontinuos transition . | we carry out systematic and high - resolution studies of dynamo action in a shell model for magnetohydrodynamic ( mhd ) turbulence over wide ranges of the magnetic prandtl number @xmath0 and the magnetic reynolds number @xmath1 .
our study suggests that it is natural to think of dynamo onset as a nonequilibrium , first - order phase transition between two different turbulent , but statistically steady , states .
the ratio of the magnetic and kinetic energies is a convenient order parameter for this transition . by using this order parameter
, we obtain the stability diagram ( or nonequilibrium phase diagram ) for dynamo formation in our mhd shell model in the @xmath2 plane .
the dynamo boundary , which separates dynamo and no - dynamo regions , appears to have a fractal character .
we obtain hysteretic behavior of the order parameter across this boundary and suggestions of nucleation - type phenomena .
# 1 # 2 # 3 ( # 1,#2)(0,0 ) ( 0,0)(#1,#2)#3 |
majorana fermion , particle which is own antiparticle , was originally predicted by ettore majorana in high energy physics@xcite .
recently , however , physics of majorana fermion has been a hot issue in condensed matter physics since the emergence of majorana fermion was pointed out at surfaces of topologically nontrivial superconductors@xcite . detecting a majorana fermion and controlling of majorana bound states(mbss ) have been a desired subject to realize the fault - tolerant topological quantum computation@xcite .
there are several suggested systems hosting mbss such as @xmath4 wave superconductors@xcite , topological insulator / superconductor heterostructures@xcite , semiconductor / superconductor junctions with strong spin - orbit interaction@xcite , helical superconductors@xcite , and superconducting topological insulators@xcite .
the most practical system among them is a semiconductor nanowire fabricated on top of a superconductor because of its controllability for the emergence of mbs by changing the chemical potential in the nanowire and by applying the zeeman field onto it@xcite .
the coexistence of the rashba spin - orbit coupling and the zeeman potential enables a topologically nontrivial superconducting state in the nanowire in the presence of proximity induced pair potential there .
even so , it is still very difficult to demonstrate convincing evidences of majorana fermions in experiments@xcite because we need to tune the number of propagating channels in nanowires @xmath2 .
the @xmath2 should be unity @xcite when the zeeman field is parallel to the nanowire .
alternatively , @xmath2 should be odd integer numbers when the zeeman field is applied perpendicular direction to the nanowire @xcite . in the latter case
, the zeeman field may destroy the pair potential .
in such situation , we seek an alternative way of realizing majorana fermion by tuning the spin - orbit interactions .
the dresselhaus spin - orbit interactions are caused by breaking the lattice inversion symmetry@xcite . in insb or gaas , for example ,
the dresselhaus [ 110 ] spin - orbit interactions can be large on their film growing along the [ 110 ] direction .
a theoretical study has shown that such artificial superconductor hosts the dispersionless surface andreev bound states which is nothing other than the mbss@xcite .
to have topologically nontrivial superconducting state , the zeeman field should be applied in plane , which is an advantage of this method .
the [ 110 ] dresselhaus nanowire superconductor is unitary equivalent to the two - dimensional polar state in @xmath5he @xcite .
it has been well known that the polar state has surface andreev bound states @xcite as a result of the sign change of the pair potential on the fermi surface @xcite . today
such surface states are recognized as the topologically protected edge states reflecting the topologically nontrivial character of the superconducting phase @xcite .
although two of authors have reported the anomalous proximity effect of superconductors in the polar state @xcite , such superconducting state has never been experimentally confirmed in any compounds .
this paper suggests a way of artificially realizing the 2d polar superconductor by combining existing materials . the anomalous proximity effect in the dresselhaus nanowire superconducting junctions is strongly related to the physics of the odd - frequency cooper pairs @xcite . in this paper
, we theoretically study the transport properties of nanowires with strong dresselhaus [ 110 ] spin - orbit interaction by using the lattice green function method on the two - dimensional tight - binding lattice .
we first calculate the local density of states ( ldos ) at the edge of the semi - infinite nanowire .
the dresselhaus [ 110 ] nanowire with in - plane magnetic field shows the large zero - energy peak independent of @xmath2 .
the zero - bias differential conductance in normal - metal / superconductor ( ns ) junctions on the nanowire shows the quantization at @xmath6 irrespective of the degree of disorder in the normal segment .
we also show the fractional current - phase ( @xmath7 ) relationship in superconductor / normal - metal / superconductor ( sns ) junctions on the nanowire .
the resonant transmission through the mbs in the normal segment is responsible for such unusual low energy transport in nanowires@xcite .
in addition to numerical simulation , we solve the bogoliubov - de - gennes ( bdg ) equation analytically and discuss the stability of mbss in the the dresslhaus[110 ] nanowire with in - plane magnetic fields .
we find that chiral symmetry of the bdg hamiltonian protects the mbss at the zero - energy@xcite .
our results indicate a way of detecting the mbss in experiments . this paper is organized as follows . in sec .
ii , we compare the local density of states at the edge of the dresselhaus noanowire superconductors with those of the rashba nanowire superconductors .
the numerical result for the transport properties are also presented . in sec .
iii , we discuss the stability of the mbss based on the analytical solution of the bdg equation . in sec .
iv , effects of disorder on the mbs in the normal metal are discussed .
the conclusion is given in sec .
-wave superconductor.,scaledwidth=45.0% ] let us consider a nanowire with the strong spin - orbit coupling fabricated on a metallic superconductor as shown in fig .
[ fig : model1 ] .
the nanowire is in the superconducting state due to the proximity - induced @xmath0-wave pair potential .
the thickness of nanowire is sufficiently small so that only the lowest subband in the @xmath8 direction for each spin degree of freedom is occupied .
we describe the present nanowire by using the tight - binding model in two - dimension .
a lattice site is pointed by a vector @xmath9 , where @xmath10 and @xmath11 are the unit vectors in the @xmath12 and the @xmath13 directions , respectively .
we consider the nanowire as the semi - infinite system in the @xmath12 direction ( i. e. , @xmath14 ) . in the @xmath13 direction ,
the number of the lattice site is @xmath15 and the hard - wall boundary condition is applied .
the nanowire is described by the bogoliubov - de - gennes(bdg ) hamiltonian , @xmath16 where @xmath17(@xmath18 ) is the creation(annihilation ) operator of an electron at the site @xmath19 with spin @xmath20 or @xmath21 , @xmath22 denotes the hopping integral , @xmath23 is the chemical potential , @xmath24 represents the strength of the dresselhaus [ 110 ] spin - orbit interaction , @xmath25 is the proximity - induced @xmath0-wave pair potential at the zero temperature .
the pauli s matrices in spin space are represented by @xmath26 for @xmath27 and the unit matrix in spin space is @xmath28 . by tuning the magnetic field @xmath29 in the @xmath12 direction as shown in fig .
[ fig : model1 ] , it is possible to introduce the external zeeman potential @xmath30 .
we measure the energy and the length in the units of @xmath22 and lattice constant , respectively . throughout this paper , we fix several parameters as @xmath31 , @xmath32 and @xmath33 . at first , we focus on the local density of states ( ldos ) at the edge of the nanowire .
the ldos averaged over @xmath15 lattice sites in the @xmath13 direction is defined by @xmath34 , \label{eq : ldossm}\end{aligned}\ ] ] where @xmath35 is the normal green s function at the site @xmath19 with the energy @xmath36 measured from the fermi energy , and @xmath37 represents the trace in spin space . to calculate the ldos ,
we add the small imaginary part @xmath38 to the energy in the green s function .
we calculate the green s function by using the lattice green s function method@xcite .
cccc ) of superconductors is plotted as a function of the energy .
( a ) dresselhaus [ 110 ] nanowire for @xmath39 ( @xmath40 ) .
( b ) dresselhaus [ 110 ] nanowire for @xmath41 ( @xmath42 ) .
( c ) rashba nanowire for @xmath39 ( @xmath40 ) .
( d ) rashba nanowire for @xmath41 ( @xmath42 ) .
the samll imaginary part of energy in the green s function @xmath43 is chosen as @xmath44.,scaledwidth=100.0% ] ) of superconductors is plotted as a function of the energy .
( a ) dresselhaus [ 110 ] nanowire for @xmath39 ( @xmath40 ) .
( b ) dresselhaus [ 110 ] nanowire for @xmath41 ( @xmath42 ) .
( c ) rashba nanowire for @xmath39 ( @xmath40 ) .
( d ) rashba nanowire for @xmath41 ( @xmath42 ) .
the samll imaginary part of energy in the green s function @xmath43 is chosen as @xmath44.,scaledwidth=100.0% ] + ) of superconductors is plotted as a function of the energy .
( a ) dresselhaus [ 110 ] nanowire for @xmath39 ( @xmath40 ) .
( b ) dresselhaus [ 110 ] nanowire for @xmath41 ( @xmath42 ) .
( c ) rashba nanowire for @xmath39 ( @xmath40 ) .
( d ) rashba nanowire for @xmath41 ( @xmath42 ) .
the samll imaginary part of energy in the green s function @xmath43 is chosen as @xmath44.,scaledwidth=100.0% ] ) of superconductors is plotted as a function of the energy .
( a ) dresselhaus [ 110 ] nanowire for @xmath39 ( @xmath40 ) .
( b ) dresselhaus [ 110 ] nanowire for @xmath41 ( @xmath42 ) .
( c ) rashba nanowire for @xmath39 ( @xmath40 ) .
( d ) rashba nanowire for @xmath41 ( @xmath42 ) .
the samll imaginary part of energy in the green s function @xmath43 is chosen as @xmath44.,scaledwidth=100.0% ] in fig . [
fig : ld - en](a ) and ( b ) , we plot the ldos at @xmath45 as a function of the energy .
the width of system @xmath15 is chosen as @xmath46 .
the results are normalized by the density of states at the fermi energy in the clean normal nanowire @xmath47 .
the amplitude of zeeman potential @xmath30 is @xmath48 and @xmath49 in fig .
[ fig : ld - en](a ) and ( b ) , respectively . as a result , the number of propagating channels @xmath50 , is @xmath51 and @xmath52 in fig .
[ fig : ld - en](a ) and ( b ) , respectively .
the ldos at the edge of the dresselhaus nanowires shows the single zero - energy peak irrespective of @xmath50 . this is a robust feature appearing as far as the zeeman potential @xmath30 is larger than a critical value @xmath53 .
the critical value of the zeeman potential is discussed in sec.iii c. for comparison , we also plots the results for the nanowire with rashba spin - oribit coupling in fig . [
fig : ld - en](c ) and ( d ) . to describe the rashba nanowires
, we replace the @xmath54 by @xmath55,\end{aligned}\ ] ] and replace @xmath56 by @xmath57 representing the magnetic field in the @xmath8 direction .
we chose @xmath30 as @xmath48 resulting @xmath58 in fig .
[ fig : ld - en](c ) and @xmath59 resulting @xmath60 in ( d ) , where @xmath61 and @xmath62 . as already discussed in ref .
, the ldos of the rashba nanowire shows the zero - energy peak only when @xmath50 is odd integer numbers . in addition , the number of peaks in the subgap energy window is equal to @xmath50 . therefore , in the rashba nanowires , we need the delicate tuning of the wire width and the zeeman field to have the mbs indicated by the zero - energy peak .
the zeeman potential @xmath30 is chosen as @xmath48 the results for the dresselhaus nanowire and for the rashba nanowire are plotted in ( a ) and ( b ) , respectively . the small imaginary part for the green s function @xmath43 is chosen as @xmath44,scaledwidth=100.0% ] .
the zeeman potential @xmath30 is chosen as @xmath48 the results for the dresselhaus nanowire and for the rashba nanowire are plotted in ( a ) and ( b ) , respectively . the small imaginary part for the green s function @xmath43 is chosen as @xmath44,scaledwidth=100.0% ] in fig . [ fig : ld - v](a ) , we plot the ldos with @xmath63 at the edge of superconductors as a function of the wire width @xmath15 for @xmath64 .
the ldos for the dresselhaus nanowire is almost constant independent of @xmath15 .
namely , the number of the zero - energy states at the edge increases proportionally to @xmath15 . in fig .
[ fig : ld - v ] ( b ) , we also plot the results for the rashba nanowire .
the ldos for the rashba nanowire takes the zero and the nonzero values alternatively as a function of @xmath15 .
the envelop function of its amplitude gradually decreases with increasing @xmath15 , which reflects a fact that the number of the zero - energy states is at most unity in the rashba case .
secondly we study the conductance in the ns junctions of nanowire superconductors as shown in fig .
[ fig : junct](a ) .
a nanowire is fabricated on an insulator / metallic superconductor junction .
the segment on the insulator and that on the superconductor are in the normal and the superconducting state , respectively .
the present junction consists three segments : an ideal lead wire ( @xmath65 ) , a normal disordered segment ( @xmath66 ) and a superconducting segment ( @xmath67 ) .
the superconducting segment is described by the bdg hamiltonian in eq .
( [ eq : bdg ] ) .
the normal - metal segment(@xmath68 ) is described by eq .
( [ eq : bdg ] ) by setting the pair potential @xmath25 to zero .
in addition , we introduce the potential disorder in @xmath69 by @xmath70 the amplitude of impurity potentials is given randomly in the range of @xmath71 . we calculate the differential conductance @xmath72 of the ns junctions based on the blonder - tinkham - klapwijk formula@xcite @xmath73_{ev = e},\end{aligned}\ ] ] where @xmath74 and @xmath75 denote the normal and andreev reflection coefficients at the energy @xmath36 , respectively .
the index @xmath76 and @xmath77 label the outgoing channel and the incoming one , respectively .
these reflection coefficients are calculated by using the lattice green s function method@xcite . in fig .
[ fig : gns - en](a ) , we present the differential conductance of the dresselhaus nanowires as a function of the bias voltage for several choices of the length of the disordered segments @xmath78 , where we choose the parameters as @xmath62 , @xmath79 , and @xmath64 . in the present parameter choice , @xmath50 becomes @xmath51 .
the results are the normalized to @xmath80 .
the differential conductance decreases with increasing @xmath78 for the finite bias voltage .
however , the zero - bias conductance is quantized at @xmath81 irrespective of @xmath78 .
the results suggest that the perfect transmission channels exist in the disordered normal segment @xcite and their number is equal to @xmath2 . for comparison
, we also plot the results for the rashba nanowire in fig .
[ fig : gns - en](b ) .
the results show that the zero - bias conductance is not quantized and decrease with increasing @xmath78 .
even if a mbs appears at the zero - energy , its contribution to the zero - bias conductance is relatively small .
therefore it is difficult to demonstrate the presence of mf by the conductance measurement in experiments .
cc , where @xmath79 , @xmath64 , and @xmath62 .
the results of the dresselhaus nanowire and those of the rashba nanowire are plotted in ( a ) and ( b ) , respectively .
the number of propagating channels @xmath50 is @xmath51 in both ( a ) and ( b ) . the number of samples used for the random ensemble average is @xmath82,scaledwidth=100.0% ] , where @xmath79 , @xmath64 , and @xmath62 .
the results of the dresselhaus nanowire and those of the rashba nanowire are plotted in ( a ) and ( b ) , respectively .
the number of propagating channels @xmath50 is @xmath51 in both ( a ) and ( b ) .
the number of samples used for the random ensemble average is @xmath82,scaledwidth=100.0% ] in fig . [ fig : gns - w ] , we plot the zero - bias conductance as a function of the width of nanowire @xmath15 for @xmath64 .
the length of disordered segment @xmath78 is chosen as @xmath83 and @xmath84 . in the dresselhaus nanowires , the perfect quantization of the zero - bias conductance
can be seen irrespective of the width of the nanowire @xmath15 .
this result reflects the presence of a mbs for each propagating channel .
the presence of mbss can be checked by the quantized value of the zero - bias conductance . in the case of the rashba nanowire ,
on the other hand , the zero - bias conductance slowly decreases with increasing @xmath15 as shown in fig .
[ fig : gns - w ] .
the results are away from the @xmath6 for all @xmath15 . for @xmath85 and @xmath84 .
the results are normalized to @xmath81 . the number of samples used for the random ensemble average is @xmath82,scaledwidth=35.0% ] finally , we study the josephson effect in the junction shown in fig . [
fig : junct](b ) .
the junction consists three segment : a disordered normal segment ( @xmath66 ) and two dresselhaus superconducting nanowires ( @xmath65 and @xmath67 ) .
the pair potential for the left superconductor is described by @xmath86 where @xmath87 corresponds to the phase difference of the pair potential between two superconductors .
the josephson current is calculated by using the green s function method@xcite . in fig .
[ fig : jcs - p ] , we plot the josephson current at @xmath88 as a function of the phase difference for several choices of @xmath89 such as @xmath90 , @xmath91 and @xmath92 , where @xmath62 , @xmath85 , and @xmath59 . for comparison
, we also plot the result for the conventional @xmath0-wave superconductor junction ( i.e. , @xmath93 and @xmath94 ) with a dashed line .
the current - phase relationship for the conventional junction slightly deviate the sinusoidal function .
the results suggest the small contribution of the higher harmonics such as @xmath95 and @xmath96 to the josephson current . this is the well know behavior of the josephson current in diffusive sns junction of the metallic superconductor@xcite . on the other hand ,
the josephson current in the dresselhaus nanowires indicates the large contribution of the higher harmonics to the josephson current at a low temperature . as a consequence ,
the results are close to the fractional current - phase relationship of @xmath97 irrespective of @xmath89 .
such fractional relationship also indicates the perfect transmission through the disordered normal segment @xcite .
is plotted as a function of the phase difference , where @xmath85 , @xmath62 , and @xmath59 .
we plot the result for several choices of @xmath89 such as @xmath90 , @xmath91 and @xmath92 .
we also plot the results for the conventional @xmath0-wave superconductor ( i.e. , @xmath93 and @xmath94 ) at @xmath79 with a dashed line .
the number of samples used for the random ensemble average is @xmath98 for all the plots.,scaledwidth=35.0% ]
in this section , we analyze the properties of the majorana bound states appearing at the edge of the dresslhaus nanowire as shown in fig . [ fig : model1 ] . here
we consider the nanowire in the continuous space for simplicity .
the bdg hamiltonian of the dresselhaus nanowire is represented by @xmath99 , \label{eq : original}\end{aligned}\ ] ] @xmath100 where @xmath101 denotes the effective mass of an electron . in what follows
, we assume large enough zeeman potential so that @xmath102 is satisfied with @xmath103 . by applying the unitary transformations in appendix [ sec : trbdg ]
, we obtain the deformed bdg hamiltonian @xmath104 , \label{eq : bdganap } \\
\check{h}_{p_{x } } = & \left [ \begin{array}{cc } \hat{h}_{p_{x},\uparrow } & 0 \\ 0 & \hat{h}_{p_{x},\downarrow } \\
\end{array } \right],\label{h : pxpm } \\
\hat{h}_{p_{x } , \sigma } = & \left
[ \begin{array}{cc } \xi + s_s v_{ex } & -s_s \frac{\lambda_d \delta_{0}}{v_{ex } } \partial_{x } \\
s_s \frac{\lambda_d \delta_{0}}{v_{ex } } \partial_{x } & - \xi -s_s v_{ex } \\
\end{array } \right ] , \label{eq : hpxsigma } \\
\check{v}_{\delta } = & \left [ \begin{array}{cc } 0 & i \delta_{0 } \hat{\sigma}_{2 } \\ - i \delta_{0 } \hat{\sigma}_{2 } & 0\\ \end{array } \right ] , \label{eq : spinmix } \\
s_s=&\left\ { \begin{array}{cl } 1 & \text{for}\ ; \sigma=\uparrow \\ -1 & \text{for}\ ; \sigma=\downarrow . \end{array}\right.\end{aligned}\ ] ] since @xmath102 , we ignore the higher order terms indicated by @xmath105 $ ] in the argument below .
the diagonal components @xmath106 are equivalent to the hamiltonian of the spin - triplet @xmath107-wave superconductor .
thus the bdg hamiltonian represents the spin - full @xmath107-wave superconductor @xmath108 with the spin - mixing term @xmath109 .
we first solve the bdg equation for the zero - energy edge states by neglecting the spin - mixing term @xmath109 .
we will discuss the effects of @xmath109 later on .
the bdg equation at the zero - energy reads @xmath110 where @xmath111^{\mathrm{t } } \nonumber\end{aligned}\ ] ] is the eigen wave function of the zero - energy states labeled by @xmath112 . under the hard - wall boundary condition in the @xmath13 direction ,
the wave function is represented as @xmath113 where @xmath114 indicates the transmission channel in the nanowires and @xmath115 is a vector with the four components . in the @xmath12 direction , we assume that the length of the nanowire is @xmath116 ( i.e. , @xmath117 ) and we apply the hard - wall boundary conditions at the edge of the nanowire , @xmath118 we show how to solve the bdg equation in appendix [ sec : trbdg ] . here
we summarize the results and discuss important properties of the solution .
the bdg equation can be solved for each transmission channel indicated by @xmath114 in which @xmath119 represents the effective chemical potential . when @xmath120 , there is no solution of eq .
( [ eq : hbdgzero ] ) with eq ( [ bcx2 ] ) . for @xmath121
, we can find the two solutions for each transmission channel at each edge : one is in the spin - up sector and the other is in the spin - down one in eq .
( [ h : pxpm ] ) . at the edge around @xmath122 , for example ,
two zero - energy states in the two spin sectors are degenerate .
however such doubly - degenerate zero - energy states are unstable in the presence of the @xmath109 because the spin - mixing terms hybridize the two zero - energy states and lift the degeneracy . finally ,
for @xmath123 , we obtain the only one zero - energy edge state for each transmission channel at each edge .
namely the zero - energy state in spin - up sector disappears and only the zero - energy state in the spin - down sector remains at each edge .
since the up - spin state is absent , @xmath109 does not affect such zero - energy edge states .
the wave function at the left edge @xmath124 and that at the right edge @xmath125 can be represented as @xmath126 { \rm sin } \left [ \sqrt{k_{n , \downarrow}^{2 } - \xi_{d}^{-2 } } ( x+l ) \right ] e^{-x/\xi_{d } } , \label{eq : phil2 }
\\ \varphi_{n}^{r } ( x ) = & c_{n}^{r } \left [ \begin{array}{cccc } 0\\ 0 \\ 1 \\ 1 \\ \end{array } \right ] { \rm sin } \left [ \sqrt{k_{n , \downarrow}^{2 } - \xi_{d}^{-2 } } ( x - l)\right ] e^{x/\xi_{d}},\label{eq : phir2}\end{aligned}\ ] ] where @xmath127 with @xmath128 and @xmath129 being the normalization coefficients .
when @xmath130 , the two zero - energy states localizing at @xmath131 are decoupled from each other .
the number of the zero - energy states at each edge is equal to the number of channels which satisfies @xmath123 .
the bdg hamiltonian in eq .
( [ eq : bdganap ] ) preserves chiral symmetry , @xmath132 .
\label{eq : gammsmsec}\end{aligned}\ ] ] in the presence of chiral symmetry , it is possible to introduce the eigen states of @xmath133 @xmath134 where @xmath135 is also the eigen states for @xmath136 @xmath137 the eigenvalue @xmath138 is either @xmath139 or @xmath140 .
see also appendix [ sec : stzs ] for details . by using these eigen states @xmath135 ,
the states belonging to the zero - energy @xmath141 can be represented by @xcite @xmath142 where @xmath143 satisfies @xmath144 the index @xmath112 labels the eigen states belonging to the zero energy . from the results above ,
we conclude that @xmath145 is the eigen state of @xmath136 belonging to @xmath138 , which is an important fact leading to the stability of the zero - energy states .
the situation in the nonzero - energy states is different that in the zero - energy states .
as shown in appendix [ sec : stzs ] , the nonzero - energy states are always described by the linear combination of two states : one belongs to @xmath146 ( i.e. , @xmath147 ) and the other belongs to @xmath148 ( i.e. , @xmath149 ) . generally speaking
, perturbations may lift zero - energy states to nonzero - energy ones .
such modification happens only when the perturbations couple the two zero - energy states belonging to opposite @xmath138 as schematically illustrated in fig .
[ fig : sta00](a ) and ( b ) .
this argument is valid as far as the perturbations preserve chiral symmetry .
therefore the zero - energy states belong to the same eigenvalue of @xmath138 are stable and remain at the zero - energy when perturbations preserve chiral symmetry in eq .
( [ eq : gammsmsec ] ) . in the dresselhaus nanowires for @xmath123 , it is easy to confirm that @xmath150 in eq .
( [ eq : phil2 ] ) belongs to @xmath148 , whereas @xmath151 in eq .
( [ eq : phir2 ] ) belongs to @xmath146 .
since they are spatially separated , the zero - energy states at the two edges are robust under perturbations preserving chiral symmetry as illustrated in fig .
[ fig : sta00](c ) . finally , we note that chiral symmetry in the original basis represented by @xmath152 , \label{chiral_h0}\end{gathered}\ ] ] where @xmath153 is the original hamiltonian in eq .
( [ eq : original ] ) .
this fact implies that the zero - energy states are robust under perturbations preserving chiral symmetry even if we take the higher order terms of @xmath154 into account .
and the one remaining state belongs to @xmath146 .
( b ) two states with opposite @xmath138 are coupled with each other and form the nonzero - energy states .
the remaining two states stay at the zero - energy because they do not have the coupling partners .
( c ) the schematic picture of the zero - energy edge states of the dresselhaus nanowire . the number of the zero - energy states at either edge is equal to @xmath2 .
all of the edge states at the left ( right ) belong to @xmath148 ( @xmath146 ) .
, scaledwidth=48.0% ] the anomalous properties in the low energy transport in sec .
[ sec : ldos ] appear when the zeeman field is larger than a critical value ( i.e. , @xmath155 ) . here
we discuss why the low energy spectra of the edge states drastically changes at @xmath156 .
we consider the periodic boundary condition in the @xmath12 direction to obtain the momentum representation of @xmath157 in eq ( [ eq : bdganap ] ) , the state vectors can be written as @xmath158 where @xmath114 labels the channels in the @xmath13 direction and @xmath159 denotes the wave number in the @xmath12 direction .
cc and @xmath160 are plotted as a function of the zeeman potential @xmath30 at @xmath62 .
@xmath161 becomes zero at @xmath162 .
( b ) we plot the local density of states at the edge of the dresselhaus nanowire as a function of energy .
the zeeman potential @xmath30 is chosen as @xmath163 ( i.e. , @xmath164 ) .
the small imaginary part in the green s function @xmath43 is chosen as @xmath165.,scaledwidth=100.0% ] and @xmath160 are plotted as a function of the zeeman potential @xmath30 at @xmath62 .
@xmath161 becomes zero at @xmath162 .
( b ) we plot the local density of states at the edge of the dresselhaus nanowire as a function of energy .
the zeeman potential @xmath30 is chosen as @xmath163 ( i.e. , @xmath164 ) .
the small imaginary part in the green s function @xmath43 is chosen as @xmath165.,scaledwidth=100.0% ] the hamiltonian in eq .
( [ eq : hpxsigma ] ) in the momentum space is represented by @xmath166 , % \\
\hat{h}_{p_{x } , \sigma}(k , n ) = & \left [ \begin{array}{cc } \xi_{n}(k ) + s_s v_{ex } & -s_s i \frac{\lambda_d \delta_{0}}{v_{ex } } k \\
s_s i \frac{\lambda_d \delta_{0}}{v_{ex } } k & - \xi_{n}(k ) -s_s v_{ex } \\
\end{array } \right ] , \label{eq : px7 } \\
\xi_{n}(k ) = & \frac{\hbar^{2 } k^{2}}{2 m } - \mu_{n } , % \\ % s_s=&\left\ { \begin{array}{cl } % 1 & \text{for}\ ; \sigma=\uparrow \\ % -1 & \text{for}\ ; \sigma=\downarrow .
% \end{array}\right.\end{aligned}\ ] ] with @xmath167 being represented in eq .
( [ eq : defmun3 ] ) . as discussed in sec .
[ sec : stblityz ] , the zero - energy edge states are stable when the condition @xmath168 is satisfied .
this condition corresponds to the situation in which the dispersion of spin - down sector @xmath169 remains at the fermi level and the dispersion of spin - up sector leaves away from the fermi level ( i.e. , @xmath170 ) . the number of the zero - energy states becomes equal to the number of the propagating channel @xmath50 in the spin - down sector . in fig .
[ fig : nc - vv](a ) , we plot the number of the propagating channels as a function of the zeeman potential , where @xmath171 represents the number of the propagating channels in spin - up ( spin - down ) sector . in the tight binding model ,
the effective chemical potential @xmath172 should be replaced by @xmath173 .
\end{aligned}\ ] ] in the spin - up sector , @xmath161 becomes zero at @xmath174 at the present parameter choice . for @xmath175 , the dispersions in both the spin - up and the spin - down sectors remain at the fermi level .
in fig . [
fig : nc - vv](b ) , we show the ldos at the edge of the dresselhaus nanowire for @xmath176 .
the resulting channel numbers are @xmath177 and @xmath178 .
the edge states also arise from the two spin sectors .
but they are interact with each other due to @xmath109 . as a result ,
the energy of such interacting states leave away from the zero energy .
the results of ldos show that two peaks appear at @xmath179 in addition to the large zero - energy peak .
thus the number of the zero - energy edge states is @xmath180 which is less than the number of propagating channels @xmath181 .
for @xmath155 , the hamiltonian of the dresselhous nanowire becomes unitary equivalent to that of the two - dimensional spinless @xmath182-wave superconductor .
it is possible to realize such polar states by tuning the spin - orbit coupling in an alternative way as shown in appendix [ sec : eabsoi ] .
the field operator of an electron for the bdg hamiltonian @xmath157 is described as @xmath183,\\ \xi = \gamma { \cal k},\end{gathered}\ ] ] where @xmath184 ( @xmath185 ) is the creation ( annihilation ) operator of the bogoliubov quasiparticle belonging to @xmath186 and @xmath187 is the charge conjugation operator with @xmath188 representing the complex conjugation . here
we focus on the electron operator at the zero - energy states for @xmath155 .
as discussed in sec .
[ sec : aatmbs ] b , the wave function of the zero - energy states are well characterized by the channel index @xmath114 and described by @xmath189 where @xmath190 ( @xmath191 ) correspond to the left(right ) edge states .
these wave function satisfy @xmath192 as illustrated in fig .
[ fig : sta00](c ) .
therefore , the electron operator of the zero - energy state is written as @xmath193 , \\ \gamma_{n}^{r } ( \boldsymbol{r } ) = & \left [ \varphi_{n}^{r}(\boldsymbol{r } ) \gamma_{n + } + \left ( \varphi_{n}^{r}(\boldsymbol{r } ) \right)^ { * } \gamma_{n + } ^{\dagger } \right],\end{aligned}\ ] ] where the field operator @xmath194 and @xmath195 correspond to the edge state on the left hand side and that on the right hand side , respectively .
the operator @xmath194 is pure imaginary while @xmath195 is real in the present gauge choice .
it is easy to show that they satisfy the majorana relation @xmath196 we conclude that both @xmath197 and @xmath198 fields describe the majorana fermions .
the above relations hold for each propagating channel @xmath114 .
therefore the number of majorana fermions at each edge is @xmath199 which is equal to @xmath2 for @xmath155 .
the numerical results in sec .
ii show that the majorana bound states penetrate into the diffusive normal segment and form the resonant transmission channels there .
the perfect transmission through such majorana bound states at the zero energy are responsible for the anomalous transport properties . here
we discuss why the majorana bound states remain at the zero energy even in the presence of impurity potentials in the normal segment .
we first analyze the wave functions in the normal segment using in eq .
( [ eq : px7 ] ) . in the absence of impurity potential ,
the wave function in the normal segment at @xmath63 is decribed by @xmath200 e^ { ik_{n,\uparrow}x } + \left [ \begin{array}{c } 0 \\ 0 \\ 1 + r^{ee}_{\downarrow,\downarrow } \\ r^{he}_{\downarrow,\downarrow } \end{array } \right ] e^ { ik_{n,\downarrow}x } , \\
k_{n,\sigma}= & \sqrt{2m(\mu_n - s_sv_{ex})}/\hbar \end{aligned}\ ] ] where @xmath201 is the normal reflection coefficients and @xmath202 is the andreev reflection coefficients . here
we consider the wave function at the channel @xmath114 .
the @xmath182 superconductors causes the perfect andreev reflection at @xmath63 , which results in @xmath203 . at the same time , we obtain @xmath204 and @xmath205 . we find that the wave function in each spin sector is the eigen state of @xmath136 in eq .
( [ eq : gammsmsec ] ) . therefore all states in spin - up ( spin - down ) sector belong to @xmath146 ( @xmath148 ) in the normal segment .
this conclusion is also true even when we introduce impurity potential into the normal segment because the impurity potentials preserve chiral symmetry in eq .
( [ eq : gammsmsec ] ) . to describe the disordered ns junctions in fig .
[ fig : junct](a ) , we introduce the random potential @xmath206 in the normal segment , ( i.e. , @xmath207 ) .
the hamiltonian @xmath208 in eq .
( [ eq : original ] ) is replaced by @xmath209 the hamiltonian of the ns junction reads , @xmath210^\ast \\ \end{array } \right ] .
\label{eq : ns}\end{aligned}\ ] ] it is easy to show that @xmath211 satisfies the relations , @xmath212 .
\label{eq : chiral4}\end{aligned}\ ] ] therefore , all of the zero - energy states at the ns interface keep staying at the zero - energy even in the presence of disorder because all them belong to @xmath213 as shown in fig . [
fig : sta00](c ) .
although the channel index @xmath114 is no longer a good quantum number under the potential disorder , the number of the zero - energy states is still equal to the number of the propagating channels in the spin - down sector @xmath214 .
this argument can be applied to the zero - energy states penetrating into the normal segment .
the resonant transmission channels at the zero energy in the normal segment is protected in the presence of chiral symmetry .
this explains the perfect quantization of the zero - bias conductance at @xmath6 . to confirm the argument above
, we calculate ldos in the disordered normal segment . in fig .
[ fig : ns - ld2](a ) , we plot the ldos at the center of the disordered segment ( i.e. , @xmath215 ) as a function of the energy , where @xmath216 , @xmath62 , @xmath79 .
the zeeman potential @xmath30 is chosen as @xmath217 and @xmath48 . in the case of @xmath218 ,
the states with @xmath219 remain at the fermi level in the normal metal .
the random impurity potentials mix the penetrated majorana bound states with @xmath213 and the normal states with @xmath219 . as a result , the ldos in the disordered normal segment for @xmath220 is almost flat around the zero - energy .
when @xmath155 , all of the normal states with @xmath219 pinch off from the fermi level .
therefore , penetrated majorana bound states stably remain at the fermi level . as a result ,
the ldos in the disordered normal segment for @xmath221 shows the large zero - energy peak reflecting the existence of the penetrated majorana bound states .
it is possible to demonstrate how chiral symmetry protects the zero - energy states in the normal segment by analyzing the details of ldos at the zero - energy in fig .
[ fig : ns - ld2](a ) . to do this
, we calculate the green s function around the zero energy .
the hamiltonian satisfies , @xmath222 , \label{eq : charge4}\end{aligned}\ ] ] where @xmath223 represents the charge conjugation . when @xmath224^{\mathrm{t}}$ ] is the wave function belonging to @xmath225 , @xmath226 is the wave function belonging to @xmath227 . using these wave functions , the green s function
is represented by @xmath228.\end{aligned}\ ] ] when we consider @xmath229 , the wave functions at the zero energy mainly contribute to the green s function , @xmath230}{e+i\delta},\end{aligned}\ ] ] where @xmath112 indicates the states at the zero energy .
we can immediately confirm that @xmath231= [ h_{ns}^{2},\xi_0 ] = [ \gamma_0 , \xi_0 ] = 0.\end{aligned}\ ] ] therefore , as shown in appendix a , we find that @xmath232 is the eigen vector of @xmath233 and @xmath234 at the same time .
> from this fact , it is possible to represent the wave function by the linear combination of following vectors @xmath235 , \end{aligned}\ ] ] where @xmath236 is the eigen function of @xmath234 belonging to @xmath237 and is also the eigen function of @xmath233 belonging @xmath238 with @xmath239 .
we assume @xmath240 . by using these vectors ,
the normal green s function at the electron space becomes @xmath241 .
\nonumber\end{aligned}\ ] ] the normal green s function @xmath242 ( @xmath243 ) are the green function derived from the wave function with @xmath146 ( @xmath148 ) . in the same way , the anomalous green function is represented as @xmath244 . \nonumber\end{aligned}\ ] ] by using the relations , @xmath245 @xmath246 is described by @xmath247 and @xmath248 obtained in the numerical simulation , @xmath249 .
\label{eq : rofag}\end{aligned}\ ] ] we can separate the numerical results of ldos at the zero - energy into two contributions @xmath250.\end{gathered}\ ] ] in fig .
[ fig : ns - ld2](b ) and ( c ) , we plot the ldos corresponding to the states with @xmath146 ( i.e. , @xmath251 ) and with @xmath148(i.e . ,
@xmath252 ) as a function of the zeeman potential @xmath30 , respectively . here
the ldos are calculated at the center ( i.e. , @xmath215 ) of the disordered segment with @xmath216 . when @xmath253 , the penetrated majorana bound states with @xmath148 and the normal states with @xmath146 are coupled by the random disordered potentials . as a result ,
the amplitude of the ldos at the zero energy remains at a small value . as shown in fig .
[ fig : ns - ld2](b ) , @xmath251 become zero for @xmath254 , which means the all zero - energy states belong to @xmath146 disappear .
as shown in fig [ fig : ns - ld2](c ) , @xmath252 suddenly becomes large for @xmath254 .
all of the penetrated majorana bound states belong to @xmath148 . as discussed in sec .
iiic , the number of the zero - energy states at the edge of superconductor is equal to the number of the propagating channels @xmath50 for @xmath254 . in the normal segment ,
the number of the zero - energy states also becomes @xmath2 even in the presence of disorder because the impurity potentials preserve chiral symmetry .
such majorana bound states at the zero energy form the resonant transmission channels in the normal segment .
this explains the perfect transmission in the presence of disorder . )
is plotted as a function of the energy .
we chose @xmath62 , @xmath216 and @xmath79 .
we plot the results for @xmath255 and @xmath48 .
the ldos corresponding to @xmath256 and the ldos corresponding to @xmath148 is plotted as a function of the zeeman potential in ( b ) and ( c ) , respectively .
the energy for the green s function is fixed to zero .
we chose @xmath62 , @xmath216 and @xmath79 in both ( b ) and ( c ) .
the number of samples used for random ensemble average is @xmath257 for all the results .
the small imaginary part for the green s function @xmath43 is chosen as @xmath258 for all the results.,scaledwidth=100.0% ] ) is plotted as a function of the energy .
we chose @xmath62 , @xmath216 and @xmath79 .
we plot the results for @xmath255 and @xmath48 .
the ldos corresponding to @xmath256 and the ldos corresponding to @xmath148 is plotted as a function of the zeeman potential in ( b ) and ( c ) , respectively .
the energy for the green s function is fixed to zero .
we chose @xmath62 , @xmath216 and @xmath79 in both ( b ) and ( c ) .
the number of samples used for random ensemble average is @xmath257 for all the results .
the small imaginary part for the green s function @xmath43 is chosen as @xmath258 for all the results.,scaledwidth=100.0% ] in the normal segment in fig .
4(a ) , the zero energy states consist of two contributions : one is the penetrated mbss from the superconductor and the other is the usual metallic states at the fermi level .
the ldos in fig .
9(a ) indicates that the former is much dominant than the latter for @xmath155 .
according to the argument in sec .
iii b , such mbss in the normal segment belong to @xmath148 .
therefore the anomalous green s function can be represented by @xmath259 . by applying the analytic continuation
, we obtain @xmath260 .
\nonumber\end{aligned}\ ] ] the anomalous function satisfies @xmath261^{\mathrm{t}}.\end{aligned}\ ] ] thus the pairing correlation is spin - triplet even - parity . as a result ,
the pairing correlation is the odd function of @xmath262 .
therefore majorana fermions always accompany the odd - frequency cooper pairs @xcite .
the odd - frequency cooper pairs support the quantization of the zero - bias conductance in sec .
ii b @xcite .
the josephson current shown in sec .
ii c is carried by the odd - frequency cooper pairs @xcite .
we have studied transport properties in junctions consisting of a superconducting nanowire with dresselhaus[110 ] spin - orbit coupling .
the local density of states at the edge of the isolated nanowire shows the large zero - energy peak when the zeeman potential is larger than a critical value .
this single peak structure reflects the existence of the majorana bound states .
we show that the number of such majorana bound states is equal to the number of the propagating channels @xmath2 . when we attach a normal nanowire to the superconducting one
, the majorana bound states penetrate into the normal segment and form @xmath2 resonant transmission channels there .
all of the majorana bound states remains at the zero energy because of chiral symmetry of the junction . as a result ,
the majorana bound states in the normal segment are responsible for the perfect transmission .
we numerically show that the zero - bias differential conductance of the normal - metal / superconductor junction are quantized at @xmath263 irrespective of the disorder .
the josephson current in disordered superconductor / normal - metal / superconductor junctions shows the fractional current - phase relationship @xmath264 at a low temperature .
the superconducting nanowires with dresselhaus[110 ] spin - orbit coupling are two - dimensional analog of the spin - triplet superconductor in the polar state .
our results indicate a way of detecting majorana fermions in experiments .
the authors are grateful to j. d. sau for useful discussion .
this work was partially supported by the `` topological quantum phenomena '' ( nos . 22103002 , 22103005 )
grant - in aid for scientific research on innovative areas and kakenhi ( no . 26287069 ) from the ministry of education , culture , sports , science and technology ( mext ) of japan and by the ministry of education and science of the russian federation ( grant no .
14y.26.31.0007 ) .
here , we briefly summarize the argument in ref . which shows the important properties of the zero - energy states under the chiral symmetry .
we consider the bdg hamiltonian @xmath265 which preserves the chiral symmetry @xmath266 where eq .
( [ eq : c - cs ] ) is equivalent to @xmath267 = 0 .
\label{eq : c - cs2}\end{aligned}\ ] ] the bdg equation is given by @xmath268 when we consider the eigen equation @xmath269 eq .
( [ eq : c - cs2 ] ) suggest that the states @xmath270 is also the eigen states of @xmath136 at the same time .
since @xmath271 , we find that the eigen value of @xmath136 is @xmath139 or @xmath140 .
namely the eigen equation @xmath272 holds for @xmath273 . by multiplying @xmath265 to eq .
( [ eq : c - gc ] ) from the left side and by using eq .
( [ eq : c - cs ] ) , we obtain the equation @xmath274 we find that @xmath275 is the eigen state of @xmath136 belonging to @xmath276 .
thus we can connect @xmath277 and @xmath278 as @xmath279 where @xmath280 is a constant .
as shown in ref . , the one - to - one correspondence exists between @xmath281 and @xmath282 .
at first , we consider zero energy states @xmath283 which satisfies @xmath284 in eq .
( [ eq : chidefa ] ) .
the integration of @xmath19 after multiplying @xmath285 from the left results in @xmath286 this means that the norm of @xmath287 is zero .
therefore we conclude that @xmath288 as a result , we find the relation @xmath289 when a zero energy state is described by @xmath290 , the relations in eqs .
( [ eq : c - chr ] ) and ( [ eq : hchizero ] ) suggest that @xmath291
. therefore the zero - energy states are always the eigen states of @xmath136 . for @xmath292 ,
it is possible to represent @xmath293 by @xmath294 @xcite . by calculating the norm of @xmath275
, we obtain @xmath295 multiplying @xmath265 to eq .
( [ eq : c - chr ] ) from the left alternatively gives a relation @xmath296 therefore , we find the relation @xmath297 although we can not fix the phase factor @xmath298 , it is possible to express the states @xmath281 for @xmath292 as @xmath299 the nonzero - energy states are constructed by a pair of eigen states for @xmath136 : one belongs to @xmath146 and the other belongs @xmath148 .
therefore , the states with @xmath292 are not the eigen states of @xmath136 . on the contrary to the nonzero - energy states ,
the the zero - energy states are the eigen states of @xmath136 .
the bdg hamiltonian of the dresselhaus nanowire is represented by @xmath153 in eq .
( [ eq : original ] ) . by using the unitary matrix @xmath300,\ ; \hat{r}= \frac{1}{\sqrt{2 } } \left [ \begin{array}{cc } e^{-i \pi/4 } & -e^{-i \pi/4 } \\ e^{i \pi/4 } & e^{i \pi/4 } \\
\end{array } \right ] , \label{eq : unir}\end{aligned}\ ] ] the bdg hamiltonian @xmath153 is first transformed to @xmath301 , \label{eq : bdg1}\\ \hat{h}^{\prime } & = \xi \hat{\sigma}_{0 } + v_{ex } \hat{\sigma}_{3 } + i \lambda_d \partial_{x } \hat{\sigma}_{2}.\end{aligned}\ ] ] the hamiltonian in this basis is represented only by real numbers .
next we apply a transformation which is similar to the foldy - wouthysen transformation @xcite to the bdg hamiltonian in eq .
( [ eq : bdg1 ] ) . using a unitary matrix @xmath302 , \label{eq : uniu}\\ \hat{u } = & { \rm exp } [ i \hat{s } ] , \ ; \hat{s } = \frac{\lambda_d}{2 \hbar v_{ex } } p_{x } \hat{\sigma}_{1},\end{aligned}\ ] ] with @xmath303 , we transform @xmath304 into @xmath305.\end{aligned}\ ] ] the diagonal term of eq .
( [ eq : bdg1 ] ) can be expanded as @xmath306 + \frac{i^{2}}{2 ! } [ \hat{s } , [ \hat{s } , \hat{h}^{\prime } ] ] + \cdots,\end{aligned}\ ] ] with using the baker - housdorff formula .
we assume large enough zeeman potential so that @xmath307 is satisfied where @xmath308 denotes fermi wave number .
> from this assumption , we obtain @xmath309,\end{aligned}\ ] ] within the first order of @xmath310 .
the off - diagonal term corresponding to the pair potential is transformed to @xmath311 + \cdots \nonumber\\ & = i \delta_{0 } \hat{\sigma}_{2 } - i \frac{\lambda_d \delta_{0}}{\hbar v_{ex } } p_{x } \hat{\sigma}_{3 } + o\left(\frac{(\lambda_d k_{f})^{2}}{v_{ex}^{2}}\right),\end{aligned}\ ] ] where we assume the uniform pair potential ( i.e. , @xmath312=0 $ ] ) .
as the result , the bdg hamiltonian can be written as @xmath313 \nonumber\\ & + o\left(\frac{(\lambda_d k_{f})^{2}}{v_{ex}^{2}}\right).\end{aligned}\ ] ] by interchanging the second column and the third one , and by interchanging the second row and the third one , the hamiltonian can be deformed as @xmath314 , \\
\hat{h}_{p_{x } , \sigma } = & \left [ \begin{array}{cc } \xi + s_s v_{ex } & - s_s i \frac{\lambda_d \delta_{0}}{\hbar v_{ex } } p_{x } \\
s_s i \frac{\lambda_d \delta_{0}}{\hbar v_{ex } } p_{x } & - \xi -s_s v_{ex } \\
\end{array } \right ] , \\
\check{v}_{\delta } = & \left [ \begin{array}{cc } 0 & i \delta_{0 } \hat{\sigma}_{2 } \\ - i \delta_{0 } \hat{\sigma}_{2 } & 0\\ \end{array } \right],\\ s_s=&\left\ { \begin{array}{cl } 1 & \text{for}\ ; \sigma=\uparrow \\ -1 & \text{for}\ ; \sigma=\downarrow . \end{array}\right.,\end{aligned}\ ] ] where the diagonal components @xmath106 are equivalent to the hamiltonian of the spin - triplet @xmath107-wave superconductor .
therefore , eq . ( [ eq : dfbdg ] ) corresponds to the bdg hamiltonian of the spin - full @xmath107-wave superconductor @xmath108 with the spin - mixing term @xmath109 .
in addition , we find that the bdg hamiltonian @xmath157 preserves chiral symmetry @xmath132.\end{aligned}\ ] ] now we seek the wave function of the zero - energy edge states in @xmath157 .
we first neglect the spin - mixing term @xmath109 . after solving the bdg equation at the zero - energy with @xmath315 @xmath316 we will discuss effects of @xmath109 . as shown in appendix [
sec : stzs ] , the zero - energy states under chiral symmetry is also the eigen states of @xmath136 @xmath317 where @xmath273 .
> from eq .
( [ eq : add - ch ] ) , the zero - energy states can be written as @xmath318^{\mathrm{t}}. \label{eq : spi0}\end{aligned}\ ] ] by substituting eq .
( [ eq : spi0 ] ) into the bdg equation , we obtain @xmath319 = 0,\end{aligned}\ ] ] which is equivalent to @xmath320 u_{\nu_{0 } \lambda , \sigma}(\boldsymbol{r } ) = 0 .
\label{eq : eqspin}\end{aligned}\ ] ] when we apply the hard - wall boundary condition in the @xmath13 direction , the wave function can be expanded as @xmath321 where @xmath322 denotes the width of the nanowire . by substituting eq .
( [ eq : spin ] ) into eq .
( [ eq : eqspin ] ) , we obtain the equations for each transmission channel @xmath114 , @xmath323 u_{n , \lambda , \sigma}(x ) = 0 , \label{eq : upspin}\end{aligned}\ ] ] where @xmath324 in what follows , we consider the nanowire with the length @xmath116 in the @xmath12 direction ( i.e. , @xmath117 ) and apply the hard - wall boundary condition at the edge of the nanowire , @xmath325 the length of the nanowire is long enough so that @xmath130 is satisfied . for @xmath326 , there is no solution which satisfies the boundary conditions in eq .
( [ eq : bc2 ] ) . for @xmath123 , we find the two solutions in spin - down sector as @xmath327 { \rm sin } \left [ \sqrt{k_{n , \downarrow}^{2 } - \xi_{d}^{-2 } } ( x+l ) \right ] e^{-x/\xi_{d}},\label{eq : leftedg }
\\ \varphi_{n,\downarrow}^{r } ( x ) = & \frac{c_{r}}{\sqrt{2 } } \left [ \begin{array}{cccc } 0\\ 0 \\ 1 \\ 1 \\ \end{array } \right ] { \rm sin } \left [ \sqrt{k_{n , \downarrow}^{2 } - \xi_{d}^{-2 } } ( x - l)\right ] e^{x/\xi_{d } } , \label{eq : rightedg}\end{aligned}\ ] ] where @xmath328 and @xmath329 are the normalization coefficients .
it is easy to show that @xmath124 localizing at the left edge belongs to @xmath148 and @xmath125 localizing at the right edge belongs to @xmath219 . in the spin - up sector , on the other hand , there is no solution . for @xmath121 ,
there are solutions at either edges in both spin - up and spin - down sectors . for the left edge
, we find @xmath330 { \rm sin } \left [ \sqrt{k_{n , \uparrow}^{2 } - \xi_{d}^{-2 } } ( x+l ) \right ] e^{-x/\xi_{d } } , \label{eq : leftedg2}\end{aligned}\ ] ] in addition to @xmath331 in eq .
( [ eq : leftedg ] ) .
the solution @xmath332 belongs to @xmath146 .
the two zero - energy states at the left edge ( i.e. , @xmath332 and @xmath331 ) belong to the opposite spin sector and to the opposite @xmath138 to each other . at the right edge , we find @xmath333 { \rm sin } \left [ \sqrt{k_{n , \uparrow}^{2 } - \xi_{d}^{-2 } } ( x - l ) \right ] e^{x/\xi_{d}}\end{aligned}\ ] ] in addition to @xmath334 in eq .
( [ eq : rightedg ] ) .
the two zero - energy states at the right edge ( i.e. , @xmath335 and @xmath334 ) belong to the opposite spin sector and to the opposite @xmath138 to each other . next , we consider the effects of the spin - mixing term @xmath109 in eq .
( [ eq : dfbdg ] ) . for @xmath121 ,
the two zero - energy states exist for each propagating channel at each edge . at the left edge , for example , the spin - up state with @xmath146 and the spin - down state with @xmath148 coexist . the spin - mixing term @xmath109 couples the two states . as a result , the energy of the coupled states shifts away from the zero - energy because the linear combination of a state with @xmath146 and a state with @xmath148 can form the nonzero - energy state as discussed in eq
( [ eq : epm2 ] ) . in this case , all the edge states can leave away from the zero - energy . for @xmath123 , on the other hand , the only one zero - energy state with spin - down exists for each propagating channel at each edge because the zero - energy state with spin - up is absent .
the edge states remain at the zero - energy even in the presence of @xmath109 because coupling partners are absent . in other words
, @xmath109 does not affect the zero - energy edge states .
and @xmath336 are orthogonal to each other in the spin - space .
the direction of @xmath337 is denoted by @xmath338 .
( b ) at @xmath339 , the spin - orbit interaction proportional to @xmath340 vanishes .
when we chose the zeeman potential as @xmath341 , @xmath342 becomes perpendicular to @xmath343 .
( c ) schematic picture of the effective @xmath107-wave superconductor with the coexistence of the rahba and dresselhaus [ 100 ] spin - orbit coupling.,scaledwidth=48.0% ] we propose an alternative nanowire whose hamiltonian is unitary equivalent to eq .
( [ eq : original ] ) . let us consider the two dimensional electron system where the rashba spin - orbit coupling and the dresselhaus[100 ] one coexist .
the hamiltonian is represented by @xmath344 where @xmath101 , @xmath23 , @xmath345 and @xmath346 denotes the effective mass of an electron , the chemical potential , the strength of the rashba spin - orbit coupling and the strength of the dresselhaus[100 ] spin - orbit coupling , respectively .
the unit vectors in spin space is denoted by @xmath347 with @xmath348 .
the part of the hamiltonian @xmath349 denotes the zeeman potential induced by the in - plane external magnetic field .
when we define @xmath350 the commutation relations among @xmath351 and @xmath352 become @xmath353 = i \hbar \delta_{\nu,\nu } , \\ & \left [ x_{\nu},x_{\nu } \right ] = \left [ p_{\nu},p_{\nu } \right ] = 0.\end{aligned}\ ] ] in this basis , the hamiltonian is represented as @xmath354 where @xmath355 as shown in fig .
[ fig : appc](a ) , the direction @xmath356 and @xmath357 are orthogonal to each other in spin space .
when the strength of the rashba spin - orbit coupling @xmath345 and the strength of the dresslhaus spin - orbit coupling @xmath346 are equal to each other ( i.e. , @xmath358 ) , @xmath359 becomes zero .
such electron systems have been studied in spintronics because they show unusual spin property so called spin - helix@xcite .
in addition , we tune the zeeman potential as @xmath360 . as shown in fig .
[ fig : appc](b ) , the hamiltonian is constructed by the two orthogonal components in the spin space , @xmath361 by multiplying a unitary matrix , @xmath362,\end{aligned}\ ] ] the hamiltonian is transformed into @xmath363 the hamiltonian @xmath364 is equivalent to the hamiltonian with the dresselhaus [ 110 ] spin - orbit coupling and with the in - plane magnetic field .
next , we introduce the proximity - induced @xmath0-wave pair potential . the bdg hamiltonian for the original basis in eq .
( [ eq : abvhm ] ) is given by @xmath365.\end{aligned}\ ] ] when @xmath366 and @xmath360 , the bdg hamiltonian is transformed into @xmath367,\\ \check{d } & = \left [ \begin{array}{cc } \hat{d } & 0 \\ 0 & \hat{d}^ { * } \\ \end{array } \right].\end{aligned}\ ] ] the bdg hamiltonian @xmath368 is unitary equivalent to the bdg hamiltonian in eq .
( [ eq : original ] ) .
therefore , by applying the unitary transformations introduced in appendix [ sec : trbdg ] , we can obtain the bdg hamiltonian of the effective @xmath107-wave super conductor in eq .
( [ eq : bdganap ] ) . as shown in fig .
[ fig : appc](c ) , we have to prepare the nanowire along the @xmath369 direction and apply the magnetic field in the @xmath370 direction .
more generally , by applying the appropriate unitary transformation , we also obtain the hamiltonian @xmath368 with @xmath371,\end{aligned}\ ] ] where @xmath372 is the arbitrary angle of the magnetic field .
alternatively , it is possible to reach bdg hamiltonian in eq .
( [ eq : original ] ) when we prepare the nanowire along the @xmath370 direction with @xmath373 and @xmath374.\end{aligned}\ ] ] 32 e. majorana , nuovo cimento , * 14 * , 171 ( 1937 ) .
f.wiczek , nature phys . * 5 * , 614 ( 2009 ) d. a. ivanov , phys .
lett . * 86 * , 268 ( 2001 ) .
j. d. sau , d. j. clarke , and s. tewari , phys .
rev b * 84 * , 094505 ( 2011 ) .
n. read and d. green , phys .
b * 61 * , 10267 ( 2000 ) .
a. y. kitaev , phys .
usp . * 44 * , 131 ( 2001 ) .
l. fu and c. l. kane , phys .
rev . lett . * 100 * , 096407 ( 2008 ) .
m. sato , y. takahashi , and s. fujimoto , phys .
rev . lett . * 103 * , 020401 ( 2009 ) .
j. d. sau , r. m. lutchyn , s. tewari , and s. dassarma , phys .
104 * , 040502 ( 2010 ) .
j. alicea , phys .
b * 81 * , 125318 ( 2010 ) .
r. m. lutchyn , j. d. sau , and s. dassarma , phys . rev . lett .
* 105 * , 077001 ( 2010 ) .
y. oreg , g. refael , and f. von oppen , phys .
* 105 * , 177002 ( 2010 ) . t. d. stanescu , r. m. lutchyn , and s. dassarma , phys .
b * 84 * , 144522 ( 2011 ) . m. sato and s. fujimoto , phys . rev .
b * 79 * , 094504 ( 2009 ) .
s. sasaki , m. kriener , k. segawa , k. yada , y. tanaka , m. sato , and y. ando , phys .
lett . * 107 * , 217001 ( 2011 ) .
a. c. potter and p. a. lee , phys .
* 105 * , 227003 ( 2010 ) ; a. c. potter and p. a. lee , phys . rev .
b * 83 * , 094525 ( 2011 ) .
r. m. lutchyn , t. d. stanescu , and s. dassarma , phys .
lett . * 106 * , 127001 ( 2011 ) .
m. gibertini , f. taddei , m. polini , and r. fazio , phys .
b * 85 * , 144525 ( 2012 ) .
g. dresselhaus , phys .
rev . * 100 * , 580 - 586 ( 1955 ) .
j. you , c. h. oh , and v. vedral , phys .
b * 87 * , 054501 ( 2013 ) .
a. j. leggett , rev .
* 47 * , 331 ( 1975 ) .
l. j. buchholtz and g. zwicknagl , phys .
b * 23 * , 5788 ( 1981 ) .
j. hara and k. nagai , prog .
* 74 * , 1237 ( 1986 ) . c. r. hu , phys .
lett . * 72 * , 1526 ( 1994 ) .
y. tanaka and s. kashiwaya , phys .
. lett . * 74 * , 3451 ( 1995 ) .
m. sato , y. tanaka , k. yada , and t. yokoyama , phys .
b * 83 * , 224511(2011 ) .
y. tanaka , y. asano , a. a. golubov , and s. kashiwaya , phys .
b * 72 * , 140503(r ) ( 2005 ) .
y. tanaka and s. kashiwaya , phys .
b * 70 * , 012507 ( 2004 ) .
y. asano , y. tanaka , and s. kashiwaya , phys .
rev.lett . * 96 * , 097007 ( 2006 ) .
y. asano , y. tanaka , a. a. golubov , and s. kashiwaya , phys .
lett . * 99 * , 067005 ( 2007 ) .
y. asano , a. a. golubov , ya .
v. fominov , and y.tanaka , phys .
lett . * 107 * , 087001 ( 2011 ) . y. asano and y. tanaka , phys .
rev b * 87 * , 104513 ( 2013 ) .
y. tanaka , s. kashiwaya , and t. yokoyama , phys .
b * 71 * , 094513 ( 2005 ) .
p. a. lee and d. s. fisher , phys .
lett . * 47 * , 882 - 885 ( 1981 ) .
t. ando , phys .
b * 44 * , 8017 - 8027 ( 1991 ) .
g. e. blonder , m. tinkham , and t. m. klapwijk , phys .rev .
b * 25 * , 4515 ( 1982 ) . y. asano , phys .
b * 63 * , 052512 ( 2001 ) . k. k. likharev , rev .
phys . * 51 * , 101 ( 1979 ) .
l. l. foldy and s. s. wouthuysen , phys .
rev . * 78 * , 29 - 36 ( 1950 ) .
b. a. bernevig , j. orenstein and s. c. zhang , phys .
lett . * 97 * , 236601 ( 2006 ) . | we theoretically study transport properties of nanowires with the dresselhaus [ 110 ] spin - orbit coupling under the in - plane zeeman potential and the proximity - induced @xmath0-wave pair potential . in the topologically nontrivial phase
, the nanowire hosts the majorana fermions at its edges and the number of the majorana bound states is equal to the number propagating channels @xmath1 . when we attach a normal metal to the superconductor , such majorana bound states penetrate into the dirty normal segment and form the @xmath2 resonant transmission channels there .
we show that chiral symmetry of the electronic states protects the majorana bound states at the zero energy even in the presence of impurities . as a result
, we find that the zero - bias conductance of normal - nanowire / superconducting - nanowire junctions is quantized at @xmath3 independent of the random potentials . |
near @xmath23 , the electron wavefunction may be taken as proportional to @xmath141 , with small @xmath142 , independent of layer index .
only @xmath143 is relevant to the spectrum in this small range of momentum , and small terms of order @xmath0 can be treated perturbatively .
the form of @xmath143 is simplified by the transformation @xmath144 with @xmath145}$ ] , and the transformed @xmath143 now has only one nonzero matrix element . with the unitary transformation , ^+_1,2=-^-_1,2=w , where @xmath147 is positive and we assume @xmath148 , @xmath149 . defining ket vectors in terms of eigenstates of @xmath150 , @xmath151 and the constant @xmath152 ,
the normalized eigenvectors are given by < ^+_1|h_1|^+_2>=-k_2 .
these matrix elements define an approximate projection of @xmath158 into the positive eigenvalue subspace of the @xmath26 hamiltonian in eq .
[ hsp ] . a correction to this
can be included in the diagonal elements in eq .
( [ diagonal ] ) using second - order perturbation theory , with the negative energy states , eqs .
[ eigenstate_m1 ] and [ eigenstate_m2 ] , being the intermediate states . using for the purpose of computing the @xmath1-matrix , we may choose any convenient gauge .
the wavefunctions well away from the saddle point have distinct in - coming and out - going characters on either side of it , so that a gauge transformation does not affect the @xmath1-matrix itself . to compute the @xmath1-matrix
we adopt landau gauge , so that introducing the vector potential can be implemented via the substitution @xmath161 , while @xmath67 remains unchanged .
the energy reference is set by @xmath162 .
the corresponding equations for the two - band model become -+(e-v_a)(v_b - e)=0+o(_b^-4).[webber ] for the band containing the saddle point , @xmath166 .
the factor @xmath167 on the left - hand side of the above equation in this situation has the form of an inverted parabola in the neighborhood of @xmath96 , which can be approximated as = _ b^2 . from this expression one
may compute the @xmath1-matrix using standard methods as described in the text .
incoming and outgoing states on either side of @xmath168 correspond to such states for the original hamiltonian , and this same @xmath1-matrix connects the amplitudes for those states .
the expressions for @xmath169 and @xmath170 in the main text are obtained by setting @xmath171 for the purpose of simplifying the expressions ; the actual numerical computation still uses the expressions listed here . ) _ , [ prop1 ] in which the parameter @xmath172 encodes the boundary condition between the edges of the doubled bz , and the basic @xmath1-matrix @xmath106 is given in the main text .
the 2x2 unitary matrices , ) , specify the boundary conditions . the subsequent phase accumulation between saddle points @xmath23 and @xmath173 ( arc 1 and 6 in first bz ) and that between @xmath107 and @xmath34 ( arc @xmath174 and @xmath175 in the second bz ) is represented by ) _ , [ prop2 ] in which @xmath176 .
the quantities @xmath177 and @xmath178 appearing in the exponent combine to give the areas between the numbered arcs and the @xmath67 axis .
we shall represent these areas in terms of the five elementary areas @xmath179 defined in fig .
[ bzbox ] .
before we proceed to show how the representation of area is done , one should notice that arcs 1 and 6 ( see fig .
1 in the main text ) in first bz should sweep out identical areas since the orbits are symmetric about the @xmath67 axis and move in opposite directions .
the same is true for arcs @xmath174 and @xmath175 in the second bz .
one may show that @xmath180 is the shading area associated with arc 1 in the top - left of fig .
[ bzbox ] .
it can be seen that @xmath181 is the negative ( brown ) area contributed from the left side of circle , and @xmath182 is the positive ( green ) shaded area .
similarly , one may show that the area associated with arc @xmath175 is @xmath183 note that the overall minus sign is due to the choice of circulation of those closed trajectories specified by the arrow in fig.[bzbox ] . continuing the same procedure ,
one can write down the relations at the saddle point @xmath34 and @xmath173 , ) . the reordering matrix @xmath184 implements the property that the @xmath34 and @xmath173 sp s scatter trajectories between different bz s .
@xmath4 is another phase angle parameter encoding the eigenvalues under the mt operators .
the next phase accumulation is given by ) _ , [ prop4 ] with the area @xmath185 specifying the phase along arcs 3 and 5 ( top - right in fig .
[ bzbox ] ) , and @xmath186 specifying the phase along arcs @xmath187 and @xmath188 ( bottom - right in fig . [ bzbox ] ) .
finally , the scattering at @xmath35 and @xmath189 may be written as ) _ .
[ prop6 ] the corresponding areas for arcs 2 and 4 are the same as those for arcs 1 and 6 , and arcs @xmath190 and @xmath191 are the same as for arcs @xmath174 and @xmath175 .
putting together eqs .
[ prop1 ] , [ prop2 ] , [ prop3 ] , [ prop4 ] , [ prop5 ] and [ prop6 ] leads to eq .
8 in the main text
. a_ab+a_bc+a_caa=2a+2b+c . for energies below that of the saddle point ,
one finds ( excluding corrections of order @xmath192 ) @xmath193 , which corresponds to the circular area associated with the trajectories of energy below that of the saddle point . moreover , for energy sufficiently below the saddle point , @xmath103 , the basic s - matrix reduces to @xmath195 , which leads to all @xmath196 s equal identity matrix as well . because the sum of @xmath125 s vanishes and the sum of @xmath124 s equals the dirac circle area , it is easy to show that eq . 8 in the text reduces to ( i)^3 e^iae^-i e^i e^-i=(-i)e^iae^-ix , where we have written @xmath200
we have also used the facts that @xmath201 in the brackets anticommutes with @xmath202 , that its square is the unit matrix .
eq . 8 in the text
then reduces to = 0 .
note that we have set @xmath203 in all expressions in this supplement except the last one .
the inclusion of boundary conditions specified by @xmath4 and @xmath172 can be shown to yield identical spectra away from the saddle point . however , for energies close to the saddle point , the magnetic states are indeed altered by these parameters .
this is illustrated in fig .
[ ll1 ] below , which show how the states in a range of energies behave for various values of @xmath204 .
.representation of the area between the portions of trajectories ( see fig . 1 in main text for numbering ) and @xmath67 axis using the five elementary areas shown in fig.[bzbox ] of the supplement .
the pair of numbered arcs appearing in the same row correspond to the same area due to the orbit symmetry with respect to @xmath67 axis and the fact that they move in opposite directions .
arcs 1 , 3 , @xmath175 , and @xmath188 are four representative orbits for demonstrating the areas in terms of the elementary ones @xmath179 in fig .
[ bzbox ] . [ cols="^,^,^,^,^,^,>",options="header " , ] s and @xmath125 s in eqs .
[ prop2 ] , [ prop4 ] and [ prop6 ] by the shaded areas associated with the representative arcs .
circles in top row are the trajectories of top layer in first bz , while those in bottom row are for the trajectories of bottom layer in second bz ( see fig . 1 in main text ) .
because the orbits are symmetric about the @xmath67 axis , arc 1 in top - left is the representative of arcs @xmath205 all of which under the specified circulation correspond to the same area listed in table [ area_table ] .
similarly , arc 3 on top - right , arc @xmath175 on bottom left , and arc @xmath188 on bottom right are the representative ones .
the five elementary areas @xmath179 specified by either shape or color are listed in the right column . referring to the circle in bottom - right , the circle of area @xmath206 with @xmath207
is divided into three distinct parts , the gold box in the middle of area @xmath208 , the side of area @xmath209 and the top / bottom portion of area @xmath210 .
the green box of area @xmath211 in top - right and the purple box of area @xmath212 in bottom - right are different because the dirac points ( the center of circle ) in top and bottom rows have different distances , @xmath213 and @xmath214 , respectively , from the @xmath67 axis.,scaledwidth=65.0% ] and @xmath172 .
the spectrum away from saddle point at @xmath215 does not change with these parameters .
, title="fig:",scaledwidth=30.0% ] and @xmath172 .
the spectrum away from saddle point at @xmath215 does not change with these parameters .
, title="fig:",scaledwidth=30.0% ] and @xmath172 .
the spectrum away from saddle point at @xmath215 does not change with these parameters .
, title="fig:",scaledwidth=30.0% ] and @xmath172 .
the spectrum away from saddle point at @xmath215 does not change with these parameters .
, title="fig:",scaledwidth=30.0% ] and @xmath172 .
the spectrum away from saddle point at @xmath215 does not change with these parameters .
, title="fig:",scaledwidth=30.0% ] and @xmath172 .
the spectrum away from saddle point at @xmath215 does not change with these parameters .
, title="fig:",scaledwidth=30.0% ] | we consider magnetic breakdown in twisted bilayer graphene where electrons may hop between semiclassical @xmath0-space trajectories in different layers .
these trajectories within a doubled brillouin zone constitute a network in which an @xmath1-matrix at each saddle point is used to model tunneling between different layers .
matching of the semiclassical wavefunctions throughout the network determines the energy spectrum .
semiclassical orbits with energies well below that of the saddle points are landau levels of the dirac points in each layer .
these continuously evolve into _ both _ electron - like and hole - like levels above the saddle point energy .
possible experimental signatures are discussed .
_ introduction _ the fundamental description of electron dynamics in a crystal and a uniform magnetic field involves orbital motion in a plane perpendicular to the field , along contours of constant energy @xcite as a function of crystal momentum * k*. this behavior can be significantly modified when tunneling from one trajectory to another becomes important , a phenomenon known as magnetic breakdown ( mb ) @xcite .
mb is important when the closest approach between @xmath0-space trajectories is on the order of the inverse of magnetic length , @xmath2}$ ] nm , where @xmath3 is the magnetic field in tesla .
mb sometimes leads to the formation of open orbits , with dramatic transport signatures @xcite .
mb effects in bulk metals can be challenging to observe because saddle points in a band structure , where mb initially sets in as the electron energy changes @xcite , are often quite far from the fermi energy .
recently , excellent candidates to observe mb phenomena have become available in the form of twisted graphene bilayers @xcite and graphene deposited on boron nitride substrates @xcite .
these two - dimensional systems can support large unit cells in real space ( `` moir patterns '' ) , and correspondingly small brillouin zones , for which critical points in the energy dispersion can be at relatively low energy @xcite .
such large unit cells have allowed the recent observation of the self - similar hofstadter spectrum @xcite , and may in principle nucleate unusual many - body states @xcite for fermi energies near that of a saddle point .
these interesting behaviors are among the reasons that twisted bilayers have attracted so much attention @xcite . a twisted bilayer graphene system is characterized by a rotation angle @xmath4 of the layers principal axes relative to an aa - stacked bilayer . in momentum space
, this relative rotation separates the dirac points associated with each layer by a distance @xmath5 . at low magnetic field and for energies just above those of the dirac points , momentum space trajectories are circular and surround one or the other dirac point .
allowed areas enclosed by these trajectories are quantized in units of @xmath6 , are electron - like ( increase in energy with field ) , and yield a spectrum essentially the same as for uncoupled layers . at higher energy
these trajectories approach one another , and interlayer tunneling becomes qualitatively important .
to understand how the spectrum evolves it is crucial to recognize that the coupling results in _ three _ distinguishable , degenerate saddle points . in the presence of the field , as the energy is raised above that of the saddle points the semiclassical orbits break apart and reconnect .
the new orbits are topologically distinct from the lower energy ones in that they enclose _ neither _ of the dirac points .
instead they surround a local maximum and navely should be hole - like , _ i.e. _ decrease in energy with field @xcite .
this suggests a very large accumulation of levels at the saddle point energy at high field .
below we demonstrate by a careful treatment of the magnetic translational symmetries that such singular behavior is avoided .
the spectrum necessarily contains _ both _ hole - like and electron - like orbits above the saddle point ( see fig .
2 ) , with the latter sweeping the levels to high energy at large field .
we expect this mechanism to be generic to band structures in which there is a sharp transition energy between hole - like and electron - like semiclassical orbits . the model we adopt for the system , first introduced in ref . , contains one dirac hamiltonian for each layer and three interlayer hopping terms , two of which contain scattering momenta @xmath7 and @xmath8 .
these become reciprocal lattice vectors for the system ( see fig . [ 2bz ] ) , and define the brillouin zone ( bz ) for the zero field energy spectrum . in the presence of a magnetic field
, energy eigenstates can simultaneously be eigenstates of two magnetic translation operators , but , as we show below , the resulting states can be represented as @xmath0-space trajectories only if one includes a minimum of two bz s in the representation .
this turns out to be a crucial element in understanding the energy spectrum above the saddle point : one finds two `` star - like '' semiclassical orbits , illustrated in fig .
[ 2bz ] .
a very unusual property of these orbits is that they involve periodic oscillations of the electrons between layers , and their quantization conditions leads to the interpenetrating electron- and hole - like levels .
we discuss possible experimental consequences of these properties below .
_ hamiltonian and saddle point dispersions _ our starting point is the zero - field hamiltonian @xcite h= ( cccc h_t & w_i=0,1,2v_i + w_i=0,1,2v_i^ & h_b ) , [ basicmodel ] in which @xmath9 $ ] are the dirac hamiltonians for uncoupled top / bottom layers , with dirac points located at @xmath10 , @xmath11 are components of the momentum operator , and the pauli matrices @xmath12 act on the sublattice index .
the coupling terms , @xmath13 , @xmath14 , and @xmath15 , are the largest interlayer hopping terms expected in a continuum model @xcite .
these introduce discrete translational symmetry characterized by reciprocal lattice vectors @xmath16 .
the hopping matrices are then specified by @xmath17 , @xmath18 , and @xmath19 . here
@xmath20 is the two - dimensional unit matrix , @xmath21 , and @xmath22 is its complex conjugate . to understand the behavior of this system , we treat the interlayer hopping as a weak periodic perturbation .
this has important qualitative effects for nearly degenerate states in the top and bottom layers that are coupled by the perturbation .
for example ( see fig .
[ 2bz ] ) , a degeneracy between the top and bottom dirac bands in the neighborhood of @xmath23 is split by the interlayer term @xmath24 . setting @xmath25 to unity , we find that the two states at @xmath26 with energy @xmath27 in the absence of @xmath24 splits into states of energies @xmath28 . near @xmath23 ,
one may treat terms involving ( small ) momenta @xmath29 perturbatively , to obtain a two - band effective hamiltonian @xcite h_sp=(k_/2+k_1 ^ 2/k_)_2 + , [ 2band ] where @xmath30 is small .
the eigenstates of @xmath31 include a parabolic band at higher energy , and a lower band with a saddle point ( sp ) for which the dispersion is @xmath32 , leading to a van hove singularity at @xmath33 .
this is similar to numerical results found in ref . .
there are two other saddle points in the first bz , near @xmath34 and @xmath35 in fig .
[ 2bz ] .
dispersions for these can be obtained in a way very similar to that of @xmath23 by employing an appropriate unitary transformation , shifting the zero of momentum for one of the two layers by @xmath7 or @xmath8 .
up to 120@xmath36 rotations , the resulting spectra are essentially identical to that of @xmath23 . , @xmath34 , @xmath35 .
symbols of cross represent the dirac points .
, scaledwidth=40.0% ] _ magnetic translation ( mt ) operators _ to incorporate a uniform perpendicular magnetic field we introduce a vector potential @xmath37 . to study the small @xmath38 limit it is convenient to work with momentum - space wavefunctions , so that the momentum operators @xmath39 entering eq .
[ basicmodel ] are replaced by @xmath40 . in the momentum representation , the interlayer tunneling terms are @xmath41 where @xmath42 , with momentum translation operators @xmath43 to exploit the translational symmetries of the problem we define mt operators t_1(g_x)= , + t_2(g_y)=,[mmt ] which commute with @xmath44 .
the combinations @xmath45 moreover commute with the _ full _ hamiltonian , as well as with one another , if 4 ^ 2_bg_xg_y=2n , [ fluxcondition ] for any integer @xmath46 .
we focus on magnetic fields satisfying this equality .
note such fields have the form @xmath47 , so that our analysis applies to a dense set of small magnetic fields .
eigenfunctions of the hamiltonian can also be expressed as eigenfunctions of mt operators that commute with @xmath48 , and it is convenient to choose the particular combination @xmath49 and @xmath50 for this purpose . to see how this plays out , we consider spinor wavefunctions written in the form @xmath51 . in the absence of interlayer coupling
, @xmath52 is a good quantum number and eigenfunctions of the hamiltonian involve harmonic oscillator states whose centers lie near @xmath52 .
thus @xmath52 can be viewed as a momentum - space guiding center coordinate .
more generally , the requirement that wavefunctions be eigenvectors of @xmath53 dictates that @xmath54 be the same for all the @xmath55 s entering a wavefunction .
the integral over @xmath52 then becomes a discrete sum . to see
the effect of interlayer coupling one needs to notice that the action of momentum shift operator @xmath56 appearing in the interlayer coupling on @xmath57 becomes @xmath58 this is consistent with the allowed discrete values of @xmath52 for a given wavefunction provided eq .
[ fluxcondition ] is obeyed .
thus @xmath59 can be written as a sum over wavefunctions @xmath60 with @xmath61 and @xmath62 .
the set of @xmath52 s one must retain is further reduced by use of a second mt symmetry condition , @xmath63 this becomes the condition @xmath64 .
ultimately one needs to only compute _ two _ functions , e.g. , @xmath65 and @xmath66 .
some comments are in order .
first , the reduction of the wavefunction to two functions of @xmath67 was possible because of our gauge choice @xcite .
secondly , since @xmath68 involves a single continuous variable , @xmath67 , it can be approximated conveniently in a semiclassical approach . because we need to retain two values of @xmath52
, these wavefunctions must be represented in two bz s @xcite .
finally , while the two - bz semiclassical description is strictly valid only for fields satisfying eq.[fluxcondition ] , we will treat @xmath38 as a continuous variable .
this captures the broad shape of the spectrum , but misses small gaps in what turn out to be narrow bands in the low field limit @xcite .
_ semiclassical wavefunctions _
assuming @xmath69 is larger than any other length scale in the problem ( weak fields ) , we may use a gradient expansion for the wavefunctions @xcite , @xmath70 $ ] .
we again start with uncoupled layers . defining @xmath71 , with @xmath72 , and @xmath73 , @xmath74 for the top ( bottom ) layer ,
the lowest non - trivial contribution has the form _~ e^i_b^2^k_1dk_x q_y^(k_x ) .
[ semiclassical ] the ( spinor ) coefficient of the wavefunction is determined at higher order in @xmath6 @xcite , and is not included in our analysis .
the set of momenta @xmath75 represent contours of constant energy above and below a dirac point .
when @xmath76 approaches 0 , these two curves approach one another , and the semiclassical approximation breaks down . to account for this one employs matching conditions @xcite at each turning point .
these work simultaneously at certain discrete energies , yielding a spectrum with spacing matching the exact result for landau levels of a single dirac point hamiltonian .
this result is essentially correct even in the presence of interlayer tunneling when one considers levels close in energy to that of the dirac points . for energies near those of the saddle points
, one must develop further connection formulae among the different semiclassical trajectories @xcite .
this is most easily implemented for @xmath77 , which connects trajectories near @xmath23 in fig .
[ 2bz ] .
the cases of @xmath78 are somewhat more complicated @xcite .
@xmath79 connects the wavefunction for @xmath52 in the top layer with with the bottom layer for @xmath80 through the saddle point @xmath35 via the operator @xmath81 .
the problem becomes closely analogous to that of the @xmath23 saddle point if one applies a unitary transformation , shifting the bottom component of the wavefunctions by @xmath82 .
this is represented conveniently by placing one quarter of the bz for @xmath80 continuously onto the upper right side of the @xmath52 bz .
similar constructions for @xmath83 , @xmath84 , and @xmath85 bring in another quarter of the @xmath80 bz on the lower right , and half of the @xmath86 bz on the left , yielding a doubled bz in the form of a rectangle .
this is illustrated in fig .
[ 2bz ] along with relevant semiclassical orbits , which are labeled with unprimed ( primed ) numbers for the @xmath52 ( @xmath87 ) bz . wavefunctions for the full system involve amplitudes multiplying functions of the form in eq .
[ semiclassical ] , with the caveat that @xmath88 represents the location of the dirac point around which an orbit is centered .
we assign an amplitude for each trajectory that enters or exits a saddle point , which are related to one another in several ways .
( _ i _ ) each trajectory has an amplitude @xmath89 to exit from some saddle point and an amplitude @xmath90 to enter another .
these are related by @xmath91 , where @xmath92 is the area between the trajectory ( which begins and ends at the points of closest approach to the saddle points ) and the @xmath93 axis in fig .
[ 2bz ] .
this area is taken to be positive ( negative ) if the trajectory is above @xmath67-axis and moves to the right ( left ) .
factors of @xmath94 must be inserted if there is a left or right turning point in the trajectory @xcite . _
( ii ) _ at each saddle point shown in fig .
[ 2bz ] , there are two incoming trajectories and two outgoing ones .
these are related by an @xmath1-matrix , which we discuss in more detail below . _
( iii ) _ trajectories exiting the doubled bz on the left or right are related to ones entering on the opposite side due to the periodicity imposed by the mt operators .
the effect of this can be incorporated in the matrices relating different amplitudes with some added ( energy independent ) phase factors @xcite . in practice , their presence only impacts the spectrum for energies rather close to that of the saddle points .
the @xmath1-matrix associated with the saddle points can be obtained through the two - band model , eq .
( [ 2band ] ) . introducing the magnetic field by adding a vector potential to the momentum @xmath95 , one finds the eigenvalue equation can be reduced to a single component problem in the neighborhood of the saddle point at @xmath96 . with a gauge transformation to landau gauge ,
one obtains an eigenvalue equation involving a massive particle in an inverted parabolic potential , @xmath97\psi=0 $ ] with @xmath98 and @xmath99^{\frac{1}{4}}$ ] @xcite . here
we define @xmath100 .
the resultant @xmath1-matrix is then obtained by standard methods @xcite , yielding s_0= ( cccc 1 & ie^-/2 + ie^-/2 & 1 ) , [ smatrix ] with @xmath101 .
eq .
( [ smatrix ] ) suggests that @xmath102 for @xmath103 and @xmath104 for @xmath105 .
these two limits define intra- and inter - layer dominant scattering regimes , respectively .
as written , @xmath106 applies directly to @xmath23 ; for @xmath107 , @xmath108 and @xmath109 , certain matrix elements are multiplied by phase factors related to the eigenvalues of @xmath110 and @xmath111 @xcite .
these only have noticeable affect quite close to the saddle point energy . the description above yields 24 independent amplitudes and 24 equations relating them .
it is convenient to group these amplitudes into the six four - component columns , @xmath112 , @xmath113 , and @xmath114 .
the labels for portions of the trajectories are displayed in fig .
[ 2bz ] . requiring single - valued wavefunctions then leads to the condition @xcite det=0.[final_expression ] in this representation , the @xmath115 s are 4 @xmath116 4 matrices which treat scattering through the @xmath117 and @xmath118 sp s together ( @xmath119 ) .
the unitary matrix @xmath120 $ ] with @xmath121 encodes areas swept out by electron orbits between the @xmath122th and @xmath123th sp .
precise definitions of the @xmath115 s , @xmath124 s , and @xmath125 s are given in ref . . .
here the reference of energy corresponds to zero of @xmath126 .
energy units are @xmath127 , magnetic field units are @xmath128 . , scaledwidth=50.0% ] numerical solutions to the problem described above are illustrated in fig .
[ ll ] . and are consistent with direct numerical diagonalization @xcite of the hamiltonian in eq .
[ basicmodel ] @xcite .
relatively simple behavior is apparent well below and above the saddle point energy , which may be understood analytically . below the saddle point ,
interlayer tunneling is negligible , leading to @xmath129 .
moreover , @xmath130 , which is the area enclosed by a trajectory in an uncoupled layer , and @xmath131 .
this leads to the standard dirac - landau level spacing . above the saddle point
, one finds @xcite @xmath132 and @xmath133 .
these anticommute with @xmath134 , and one can show that eq .
[ final_expression ] is satisfied if @xmath135 for either one of the two signs in the exponent . in this expression ,
@xmath136 , and @xmath137 corresponds to the area of half of a single bz .
the quantities @xmath138 are areas related to the star - like orbits , which increase ( decrease ) in magnitude with energy , leading to coexisting electron- and hole - like levels .
_ discussion _ the spectrum predicted above resolves some apparent inconsistencies among recent results .
studies which include only one saddle point @xcite yield purely electron - like spectra .
by contrast , one expects hole - like orbits surrounding local maxima to come down towards the saddle point , as shown in tight - binding studies of the twisted bilayer @xcite .
these pictures are in a sense both correct .
when several sp s are degenerate in energy , the necessity to include multiple bz s allows electron- and hole - like orbits to coexist .
importantly , this structure explains how levels rising from below the sp and levels falling from above with increasing @xmath38 evolve : the levels anti - cross , and all ultimately move to high energy when the field is sufficiently large , as is evident in fig .
[ ll ] .
this behavior should appear in many systems where degenerate , distinguishable sp s allow a transition between topologically distinct semiclassical orbits , including graphene on boron nitride substrates @xcite , and in single layer graphene at high energy @xcite .
this behavior is also apparent in the surface states of crystalline topological insulators in a magnetic field @xcite . the peculiar landau level structure and the associated semiclassical orbits in our model should have a number of experimental ramifications . for sufficiently clean samples , the level structure itself could be detected directly in tunneling @xcite .
cyclotron resonance @xcite brings another interesting perspective : since star - like orbits tunnel periodically between layers , electromagnetic waves with electric field perpendicular to the layers should couple to them and allow absorption , whereas in truly two - dimensional systems this would not be possible .
( preliminary calculations @xcite demonstrate that this is indeed the case . )
thermodynamically , converging hole - like and electron - like orbits at the saddle point energy should lead to cusp - like behavior in magnetic susceptibility @xcite .
finally , breaking the symmetry among the saddle points , for example by strain or a periodic potential @xcite , can in principle induce open orbits , which might be observed in transport as a metal - insulator transition .
_ acknowledgements _ the authors are grateful to luis brey and pablo san jose for useful discussions , and to s .- x .
zhang for pointing out ref . .
this work was supported by the nsf through grant no .
dmr-1005035 , and the us - israel binational science foundation , through grant no .
2008256 .
plain n. w. ashcroft and n. d. mermin , solid state physics ( holt , rinehart and winston , new york , 1976 ) .
e.m . lifshitz and l.p .
pitaevskii , _ statistical physics _ part ii , vol
. 9 of _ landau and lifshitz course of theoretical physics _
( butterworth - heinemann , oxford , u.k . , 1980 ) .
m. h. cohen and l. m. falicov , phys .
rev .
lett . * 7 * , 231 ( 1961 ) .
a. b. pippard , proc .
r. soc .
lond .
a * 270 * , 1 ( 1962 ) .
w. g. chambers , phys . rev . * 149 * , 493 ( 1966 ) .
j. hass , f. varchon , j. e. milln - otoya , m. sprinkle , n. sharma , w. a. de heer , c. berger , p. n. first , l. magaud , and e. h. conrad , phys .
rev .
lett . * 100 * , 125504 ( 2008 ) g. li , a. luican , j. m. b. lopes dos santos , a. h. castro neto , a. reina , j. kong , and e. y. andrei , nature phys .
* 6 * , 109 ( 2010 ) .
g. giovannetti , p. a. khomyakov , g. brocks , p. j. kelly , and j. van den brink , phys .
rev .
b * 76 * , 073103 ( 2007 ) . c. r. dean , a. f. young , i. meric , c. lee , l. wang , s. sorgenfrei , k. watanabe , t. taniguchi , p. kim , k. l. shepard , and j. hone , nat . nano . * 5 * , 722 ( 2010 ) .
j. xue , j. sanchez - yamagishi , d. bulmash , p. jacquod , a. deshpande , k. watanabe , t. taniguchi , p. jarillo - herrero and b. j. leroy , nature mat .
* 10 * , 282 ( 2011 ) .
r. a. deutschmann , w. wegscheider , m. rother , m. bichler , and g. abstreiter , c. albrecht , and j. h. smet , phys .
rev .
lett . * 86 * , 1857 ( 2001 ) . j. m. b. lopes dos santos , n. m. r. peres , and a. h. castro neto , phys .
rev .
lett . * 99 * , 256802 ( 2007 ) .
d. hofstadter , phys .
rev .
b * 14 * , 2239 ( 1976 ) .
l. a. ponomarenko , r. v. gorbachev , g. l. yu , d. c. elias , r. jalil , a. a. patel , a. mishchenko , a. s. mayorov , c. r. woods , j. r. wallbank , m. mucha - kruczynski , b. a. piot , m. potemski , i. v. grigorieva , k. s. novoselov , f. guinea , v. i. falko and a. k. geim _ nature _ * 497 * , 594 ( 2013 ) . c. r. dean ,
l. wang , p. maher , c. forsythe , f. ghahari , y. gao , j. katoch , m. ishigami , p. moon , m. koshino , t. taniguchi , k. watanabe , k. l. shepard , j. jone , and p. kim , _ nature _ * 497 * , 598 ( 2013 ) .
r. nandkishore , l. levitov , and a. chubukov , nat . phys .
* 8 * , 158 ( 2012 ) .
k. f. mak , j. shan , and t. f. heinz , phys .
rev .
lett . * 106 * , 046401 ( 2011 ) .
j. gonzlez , arxive:1307.6745 for a recent review , see e. j. mele , j. phys .
d : appl .
phys .
* 45 * , 154004 ( 2012 ) .
s. shallcross , s. sharma , and o.a .
pankratov , phys .
rev .
lett .
* 101 * , 056803(2008 ) .
p. san - jose , j. gonzlez and f. guinea , phys .
rev .
lett . * 108 * , 216802 ( 2012 ) .
r. de gail , m. o. goerbig , f. guinea , g. montambaux , and a. h. castro neto , phys .
rev .
b * 84 * , 045436 ( 2011 ) .
m .- y .
choi , y .- h .
hyun , and y. kim , phys .
rev .
b * 84 * , 195437 ( 2011 ) . r. bistritzer and a. h. macdonald , phys . rev .
b * 84 * , 035440 ( 2011 ) . p. moon and m. koshino , phys .
rev .
b * 85 * , 195458 ( 2012 ) .
d. lee , c. riedl , t. beringer , a. h. castro neto , k. von klitzing , u. starke , and j. h. smet , phys .
rev .
lett . * 107 * , 216602 ( 2011 ) i. brihuega , p. mallet , h. gonzlez - herrero , g. trambly de laissardire , m. m. ugeda , l. magaud , j. m. gmez - rodrgues , f. yndurin , and j .- y .
veuillen , phys .
rev .
lett . * 109 * , 196802 ( 2012 ) .
t. ohta , j. t. robinson , p. j. feibelman , a. bostwick , e. rotenberg , and t. e. beechem , phys .
rev .
lett . * 109 * , 186807 ( 2012 ) .
w. yan , m. liu , r .- f .
dou , l. meng , l. feng , z .- d .
chu , y. zhang , z. liu , j .- c .
nie , and l. he , phys .
rev .
lett .
* 109 * , 126801 ( 2012 ) .
k. kim , s. coh , l. z. tan , w. regan , j. m. yuk , e. chatterjee , m. f. crommie , m. l. cohen , s. g. louie , and a. zettel , phys . rev .
lett . * 108 * , 246103 ( 2012 ) .
k. sato , r. saito , c. cong , t. yu , and m. s. dresselhaus , phys .
rev .
b * 86 * , 125414 ( 2012 ) . r. bistritzer and a. h. macdonald , pnas * 108 * , 12233 ( 2011 ) .
j. m. b. lopes dos santos , n. m. r. peres , and a. h. castro neto , phys . rev .
b * 86 * , 155449 ( 2012 ) .
details may be found in the supplementary material for this article .
for example , had we chosen @xmath139 , one would need to retain three values of @xmath52 in the calculation .
the choice of circular gauge minimizes the number of @xmath52 s needed .
a similar unit cell doubling occurs in the context of superconducting vortex lattices .
see a. melikyan and z. tesanovic , phys .
rev .
b * 76 * , 094509 ( 2007 )
. the extension will be similar to the construction in , for example , p. carmier and d. ullmo , phys .
rev .
b * 77 * , 245413 ( 2008 ) . c. m. bender and s. a. orszag , _ advanced mathematical methods for scientists and engineers _ , ( springer , new york , 1999 ) . h. a. fertig and chi - ken lu , unpublished .
j. n. l. connor , chem .
. lett . * 4 * , 419 ( 1969 ) .
h. a. fertig and b. i. halperin , phys .
rev .
b * 36 * , 7969 ( 1987 ) . m. ya .
azbel , sov .
phys .
jetp * 12 * , 891 ( 1961 ) .
h. a. fertig , phys .
rev .
b * 38 * , 996 ( 1988 ) . y. okada et al . , arxive:1305.2823 . see fig .
4 .
r. s. deacon , k .- c .
chuang , r. j. nicholas , k. s. novoselov , and a. k. geim , phys .
rev .
b * 76 * , 081406 ( 2007 )
. e. a. henriksen , z. jiang , l .- c .
tung , m. e. schwartz , m. takita , y .- j .
wang , p. kim , and h. l. stormer , phys .
rev .
lett . * 100 * , 087403 ( 2008 ) . p. moon and m. koshino , arxive:1308.0713 .
g. vignale , phys .
rev .
lett . * 67 * , 358 ( 1991 ) .
l. brey and h. a. fertig , phys .
rev .
lett . * 103 * , 046809 ( 2009 ) . |
quantum theory is generally formulated relative to a given causal structure .
this is the case both in quantum field theory , where the spacetime metric is given from the beginning , and in the operational framework of quantum information , where protocols and computations consist in sequences of operations performed at different times .
relative to the given causal structure , quantum theory has to satisfy the _ causality principle _ @xcite , stating that the probability of a measurement outcome at a given time be independent of the choice of operations performed at later times .
the causality principle forbids any form of signalling from the future to the past : a sender in the future can not deterministically transfer the state of his system to a receiver in the past .
however , it is easy to imagine situations where the state of a system is transferred from the future to the past _ with some probability _ , without leading to signalling
. this is what happens , for example , if we eliminate the communication of classical data in the original quantum teleportation protocol @xcite : the receiver in the past ( say , alice ) can prepare a maximally entangled state and the sender in the future ( say , bob ) can perform a bell measurement on the state to be teleported together with half of the entangled state , as in fig . [
fig : originaltele ] .
( color online ) probablilistic teleportation from the future to the past .
today alice prepares a maximally entangled two - qubit state @xmath3 of the ancilla and the output system , while tomorrow bob performs a measurement on the ancilla and his input qubit .
the measurement has four possible outcomes , for one of which ( here denoted by @xmath4 ) successful teleportation will occur . with probability @xmath5 alice obtains bob s input state today .
however , she will not know whether teleportation has been successful until bob performs his measurement and communicates the outcome . ] in this way , upon observing the right measurement outcome , bob will know that the state has been teleported successfully to the past , and alice will not need to perform any correction operation on the output .
of course , to know that teleportation has succeeded , bob needs to wait until he sees the outcome of the bell measurement .
however , the important fact here is that , even before the measurement is performed , one half of the entangled state is already ready to be used by alice in a quantum circuit : alice can perform any desired computation on her system , and in the end , if teleportation is successful , bob will know that alice s computation has been performed on the input state he provided in the future . conditionally to the occurrence of the right outcome , everything behaves as though the input state has travelled from the future to the past through an ideal quantum channel . note
that this fact is not in contradiction with causality : the probability of teleportation is small enough that no signal can be sent from bob to alice .
we will refer to the above use of probabilistic teleportation as a _
probabilistic simulation of the identity channel from the future to the past_. the idea of probabilistic teleportation as a simulated time - travel was originally proposed by bennett and schumacher in an unpublished work @xcite , and , more recently , by svetlichny @xcite . in different terms
, this idea also appeared implicitly in the context of the graphical language of categorical quantum mechanics @xcite , where the reinterpretation of probabilistic teleportation as information flow from the future to the past follows from a stretching of wires in the basic teleportation diagram .
interestingly , similar ideas also appeared in the work of horowitz and maldacena @xcite in the context of black hole evaporation , as an attempt to reconcile the unitarity of the black hole @xmath6-matrix with hawking s semiclassical arguments ( see also the discussion by gottesmann and preskill @xcite ) .
recently , the role of probabilistic teleportation in simulating closed timelike curves has been further explored by lloyd _ et al _ in refs .
@xcite . in particular , ref .
@xcite reports an experiment that uses probabilistic teleportation to simulate a quantum computation within a closed timelike curve .
de silva , galvao and kashefi @xcite showed that some patterns in measurement - based quantum computation can be interpreted as deterministic simulations of closed timelike curves .
all the works mentioned so far focused on the use of teleportation for the probabilistic simulation of an ideal quantum channel from the future to the past . however , there are many interesting scenarios where one needs to consider the simulation of more general quantum channels . for example
, suppose that we have @xmath0 identical copies of a the same state and that we want to teleport just one copy to the past .
does the probability of success increase with the number if input copies ? and , if it does , what is the the asymptotic value of the success probability in the limit @xmath7 ? to answer these questions we have to address the probabilistic simulation of channels that trace all systems but one .
in this paper we will address the general problem of the probabilistic simulation of a given channel from the future to the past , showing how the causality principle determines the maximum probability of success . to find the maximum probability we will optimize over all generalized teleportation schemes where a bipartite state is prepared and the input of the channel is jointly measured along with half of the bipartite state , so that , for a particular outcome , the desired channel is simulated , as in fig .
[ genchan ] .
( color online ) probabilistic simulation of a given quantum channel from the future to the past via a generalized teleportation scheme . the output system and an ancilla
are first prepared in a bipartite state @xmath3 .
when the input becomes available , the input state and the ancilla are measured jointly . for a particular measurement outcome @xmath4
the simulation will be successful : the effective transformation from input to output will be proportional to the desired quantum channel , with proportionality constant @xmath8 equal to the probability of successful simulation . ]
the spirit of this work is similar to the spirit of the early works on the optimal cloning of non - orthogonal quantum states @xcite : in that case , one knew from the no - cloning theorem @xcite that it is impossible to produce perfect copies of the input state and the goal was to find the optimal physical process that approximates the impossible cloning transformation . in our case
, we know from causality that it is impossible to have a channel that deterministically transfers information from the future to the past and our goal is to find the optimal process that achieves the desired channel with maximum probability of success .
in the same way in which the study of optimal cloning shed light on the process of copying information in the quantum world , we expect that the study of optimal simulations of channels to the past will shed light the interplay between causal structure and quantum information flow .
the main message of this paper is that the maximum probability of simulating a quantum channel from the future to the past is a decreasing function of the amount of information that the channel can transmit .
this general feature will be illustrated in several examples .
first , we will consider ideal classical channels ( those that perfectly transmit the states of an orthonormal basis ) and , second , ideal quantum channels ( those that perfectly transmit all states of a given quantum system ) .
in the first case we find the maximum probability @xmath9 where @xmath10 is the dimension of the hilbert space , while in the second case we find @xmath11 which is exactly the probability of the outcome that does not require correction operations in the original teleportation protocol @xcite .
we will then focus on the probabilistic simulation of measure - and - prepare channels , which are a class of channels that transmit only classical information . in this case
, we will find that the maximum probability of success is at least equal to the probability @xmath12 for a classical channel with the same output hilbert space .
the exact value of the probability of success will be computed in two relevant cases of measure - and - prepare channels : the channel associated to the optimal estimation of a pure state , and the universal not channel @xcite .
finally , we will analyze in detail the case of trace channels , sending @xmath0 copies of a given pure state to @xmath1 copies of the same state , and the case of universal cloning channels @xcite , sending @xmath0 copies of a pure state to @xmath13 optimal approximate copies . in both cases
we find that the maximum probability of success is given by @xmath14 where @xmath10 is the dimension of the one - particle hilbert space and the superscript in @xmath15 reminds that we are restricting ourselves to the symmetric subspaces , in which the @xmath0 input copies and the @xmath16 output copies of the given pure state live . in the particular case of @xmath17 eq .
( [ gennmbound ] ) yields the value @xmath18 which increases with the number of input copies , starting from @xmath19 and reaching the classical value @xmath20 in the asymptotic limit @xmath7 .
this result has to be contrasted with the classical scenario , where having more input copies of a pure state can not lead to any improvement : since classical pure states can be perfectly cloned , there is no difference in having more input copies .
the convergence of the probability of success to the classical value @xmath12 will be explained as a consequence of the convergence of the trace channel to a measure - and - prepare channel @xcite . concerning the probability @xmath21 in eq .
( [ gennmbound ] ) , it is also worth noting that it is symmetric in @xmath0 and @xmath16 : in other words , the probability of successful simulation for a trace channel from @xmath0 to @xmath22 copies is equal to the probability of successful simulation for a universal cloning channel from @xmath16 to @xmath23 copies .
we will explain the symmetry of the probabilities as a consequence of a _ time - reversal duality _ between trace and universal cloning .
as we anticipated , the message emerging from our results is that the maximum probability for a given channel is a decreasing function of the amount of information that the channel can transmit : the smallest value @xmath24 corresponds to the identity channel , while measure - and - prepare channels have probability @xmath25 , and the erasure channels @xmath26 ~ \rho_0 $ ] have probability @xmath27 .
we will also make the connection more quantitative , providing a set of _ statistical information bounds _ , stating that the probability of success in the simulation of a channel is upper bounded by the inverse of the amount of information that the channel can transmit .
the amount of information will be quantified here as the maximum payoff that two parties can achieve in a communication game .
the structure of the paper is the following : in section [ sec : general ] we present the general method and , in particular , the _ causality bound _ , a necessary and sufficient condition for the existence of a probabilistic simulation with probability @xmath8 .
then , we will analyze the probabilistic simulation of several channels : sections [ sec : ideal ] , [ sec : measprep ] , and [ sec : opttele ] will focus on ideal classical and quantum channels , on measure - and - prepare channels , and on partial trace channels from @xmath0 to @xmath28 copies , respectively .
section [ sec : opttele ] will also discuss the case of universal cloning channels from @xmath0 to @xmath29 copies , exploiting the existence of a time - reversal duality between partial trace channels and cloning channels .
finally , in section [ sec : summary ] we will derive the statistical information bound and use it to explain the asymptotic behaviour of the maximum probability for partial trace and cloning channels .
the conclusions of the paper are drawn in section [ sec : conclusion ] .
to find the maximum probability to simulate a given channel we will use formalism of the choi operators @xcite and the framework of quantum combs @xcite . in this section we present the basic notions used in the paper and we derive the _ causality bound _
, a necessary and sufficient condition for the simulation of a given channel with probability @xmath8 . any quantum circuit can be represented as a collection of preparations , measurements , transformations , and quantum wires along which states travel undisturbed . in this paper
we only consider generalized teleportation circuits , which consist of one joint preparation of output and ancilla and one joint measurement on the input and ancilla , as in the circuit of fig .
[ gentele ] .
( color online ) generalized teleportation circuit consisting of a joint state preparation and of a joint measurement . ] here the preparation is represented by a joint quantum state @xmath30 of the ancilla and the output system , and the quantum measurement is represented by a _ positive operator valued measure ( povm ) _
@xmath31 , namely a collection of positive operators @xmath32 acting on the hilbert space @xmath33 with the normalization property @xmath34 @xmath35 and @xmath36 representing the identity operators on the input hilbert space @xmath37 and on the ancilla hilbert space @xmath38 , respectively .
each possible outcome @xmath39 corresponds to a different linear map @xmath40 , mapping density matrices on the input hilbert space @xmath41 to density matrices on the output hilbert space @xmath42 , and given by @xmath43 , \end{aligned}\ ] ] @xmath44 denoting the partial trace over the hilbert spaces @xmath37 and @xmath45 . by construction
, each linear map @xmath46 is completely positive and trace non - increasing , namely it is a _ quantum operation _
note that , however , in this case the output of the quantum operation precedes the input in time , differing to the standard case originally considered by kraus @xcite .
consider now the sum over all possible maps @xmath47 corresponding to different measurement outcomes . summing over all possible measurement outcomes is equivalent to ignoring the measurement outcome , thus obtaining the completely positive trace - preserving map ( also known as _ quantum channel _ ) @xmath48 given by @xmath49 \\ & = \sum_{i \in x } \tr_{in , a } [ \rho \otimes |\psi\>\<\psi| ] \\ & = \rho_0 ~ \tr_{in}[\rho ] \qquad \rho_0 : = \tr_a [ |\psi\>\<\psi|],\end{aligned}\ ] ] the third equality coming from the normalization of eq .
( [ eq : alloraaa ] ) .
note that the trace - preserving map @xmath48 is an _ erasure channel _ : it always produces the same output state @xmath50 independently of the input .
the fact that @xmath48 is an erasure channel is a manifestation of the causality principle stating that no signal can be sent from the future back to the past . in the following
we will use the notation @xmath51 to mean that @xmath48 is the map defined by @xmath52 $ ] , for every quantum state @xmath53 on @xmath37 .
since the linear maps @xmath54 are all completely positive , we must have @xmath55 where the inequality @xmath56 for two linear maps @xmath57 and @xmath58 means that the linear map @xmath59 is completely positive .
we will refer to eq .
( [ causbound ] ) as the _ causality bound _ for a quantum operation @xmath46 in a generalized teleportation scheme
. we can now focus on the a probabilistic simulation of a given quantum channel @xmath60 from the future to the past . in this case
we want that for one particular measurement outcome @xmath61 the transformation @xmath62 is proportional to @xmath60 , namely @xmath63 . in this case , the proportionality constant @xmath8 represents the probability that the simulation succeeds . the maximum value of the success probability @xmath8 must be compatible with the causality bound @xmath64 we will refer to the above bound as the _ causality bound for the simulation of channel @xmath60_. an obvious consequence of the causality bound is the following : a channel @xmath60 can be simulated with probability @xmath65 in a generalized probabilistic teleportation protocol if and only if @xmath60 is an erasure channel , namely @xmath66 .
* suppose that we have @xmath65 in the causality bound of eq .
( [ mapsinequality ] ) and define the quantum operation @xmath67 , so that we have @xmath68 . applying both sides of the equality to a generic state @xmath53 and taking the trace , we obtain @xmath69 + \tr[\map d(\rho ) ] = \tr[\rho]$ ] . since @xmath60 is a quantum channel , by definition @xmath70=\tr[\rho]$ ] .
that is , @xmath71= 0 $ ] for every @xmath53 .
hence , we conclude that @xmath72 , and , therefore @xmath73 .
@xmath74 in the following section we will show that the validity of the causality bound is not only necessary , but also sufficient for the existence of a probabilistic simulation with probability @xmath8 .
a very convenient tool to study quantum operations and channels is the choi correspondence , which associates linear operators on he hilbert space @xmath75 to linear maps sending operators on @xmath41 to operators on @xmath42 .
the choi operator @xmath76 corresponding to a linear map @xmath77 is given by @xmath78 where @xmath79 indicates the unnormalized maximally entangled state @xmath80 , @xmath81 being an orthonormal basis for @xmath37 .
we will often use the
double ket " notation of ref.@xcite for bipartite states on a tensor product hilbert space @xmath82 : this notation is based on the one - to - one correspondence between operators @xmath83 from @xmath84 to @xmath85 and states @xmath86 in @xmath82 , given by @xmath87 where @xmath88 is an orthonormal basis for @xmath89 and @xmath90 is an orthonormal basis for @xmath84 .
the double ket " notation has the useful properties @xmath91~,\end{aligned}\ ] ] for every pair of operators @xmath92 from @xmath84 to @xmath85 , and @xmath93 for every operator @xmath94 on @xmath85 , @xmath95 on @xmath84 , and @xmath83 from @xmath84 to @xmath85 . here
latexmath:[$b^t : = \sum_{i , j } ~ \<\psi_i| b @xmath95 in the given basis @xmath90 .
since the correspondence between linear maps and choi operators is one - to - one every property of a linear map corresponds to a property of the associated choi operator . in particular , we have choi s theorem , stating that a linear map @xmath97 is completely positive if and only if its choi operator @xmath76 is positive semi - definite @xcite .
moreoever , a linear map is trace preserving if and only if its choi operator obeys @xmath98=i_{in}$ ] .
let us now re - cast the causality bound of eq .
( [ mapsinequality ] ) in terms of choi operators .
the choi operator of the erasure channel @xmath99 is given by @xmath100 denoting by @xmath83 the choi operator of the quantum channel @xmath101 , the causality bound of eq .
( [ mapsinequality ] ) becomes @xmath102 this operator form of the causality bound is much simpler to handle than the version with linear maps . in the following
we will use eq .
( [ choiinequality ] ) to find the maximum probability @xmath8 for the simulation of a channel @xmath60 to the past .
we now use the method of quantum combs @xcite to show that any value of @xmath8 compatible with the causality bound of eq .
( [ choiinequality ] ) can be achieved in a suitable generalized teleportation protocol .
the key result that we use is theorem 4 of ref .
@xcite , whose statement , specialized to the case considered in this paper , can be rephrased as [ lem : shortcomb ] let @xmath103 be a collection of positive operators on @xmath104 satisfying the property @xmath105 for some state @xmath50 on @xmath106 .
then , there exists an ancillary system @xmath94 , a pure state @xmath107 , and a povm @xmath108 on @xmath109 such that , for every @xmath110 , @xmath111 is the choi operator of the quantum operation @xmath46 given by @xmath112 , \end{aligned}\ ] ] for every state @xmath53 on @xmath37 .
a simple consequence of lemma [ lem : shortcomb ] is that every value of @xmath8 compatible with the causality bound is attainable in some generalized teleportation protocol : [ theo : existence ] let @xmath113 be the choi operator of the channel @xmath60 . if there exists a state @xmath50 on @xmath106 such that @xmath114 , then there exists a generalized teleportation scheme such that simulates @xmath60 with probability @xmath8 .
* to prove the thesis it is enough to define @xmath115 and @xmath116 and to apply lemma [ lem : shortcomb ] to the collection of operators @xmath117 .
the simulation of channel @xmath60 will succeed when the measurement outcome is @xmath118 .
thanks to theorem [ theo : existence ] we can derive the maximum probability of successful simulation from the causality bound of eq .
( [ choiinequality ] ) : we will know automatically that there exists a generalized teleportation protocol succeeding with that maximum probability .
in this section we will illustrate the general method of the causality bound in two simple examples .
the first example is about the probabilistic simulation of an ideal classical channel , which preserves the elements of an orthonormal basis , but destroys all off - diagonal elements of the density matrix . in this case
, we will show that the probability of success is given by @xmath119 where @xmath10 is the dimension of the hilbert space .
the second example is about the probabilistic simulation of an ideal quantum channel , represented by the identity map . in this case , using the causality bound we will show that the probability of success is given by @xmath120 @xmath10 being again dimension of the hilbert space .
this also proves that the original teleportation protocol @xcite is optimal for the probabilistic simulation of an ideal channel from the future to the past .
[ subsect : idealclassical ] an ideal classical channel can be modelled in quantum theory as a von neumann measurement on an orthonormal basis , represented by the channel @xmath121 , where @xmath53 is the state of a @xmath10-dimensional quantum system . by the definition of choi operator in eq .
( [ correspondence ] ) , the choi operator of @xmath122 is given by @xmath123 inserting the above expression in the causality bound of eq .
( [ choiinequality ] ) , we obtain @xmath124 then , taking on both sides the expectation value on the state @xmath125 , we get @xmath126 , for every @xmath127 , which also implies @xmath128 .
choosing @xmath129 we then achieve the maximum value @xmath130 a generalized teleportation protocol that simulates the ideal classical channel with maximum probability @xmath131 can be easily obtained by preparing the classically correlated state @xmath132 and by measuring the two - outcome povm @xmath133 given by @xmath134 and @xmath135 .
another instructive application of the general method is the proof that the value of the probability of successful teleportation @xmath136 in the original teleportation protocol @xcite is the optimal value determined by causality . in the case of teleportation ,
the quantum channel that we want to simulate is just the identity : @xmath137 . by definition of choi operator
( [ correspondence ] ) , the choi operator of the identity channel is @xmath138 plugging this into the causality bound of eq .
( [ choiinequality ] ) , we then obtain @xmath139 to find the maximum value of the simulation probability @xmath8 we must ensure that the difference between the right hand side and the left hand side is a positive operator . in general , this is not a trivial issue because the operators on the two sides are not necessarily diagonalizable in the same basis .
to resolve this problem , we notice a symmetry of the left hand side . defining the complex conjugate of an operator as @xmath140 and using eq .
( [ doubleketalg ] ) we have @xmath141 and , therefore , @xmath142 . applying the same transformation on both sides of the inequality ( [ originalineq ] ) , we obtain @xmath143 since the inequality must hold for every unitary @xmath144 , it also holds for the integral over the normalized haar measure @xmath145 , @xmath146 having used the schur s lemma for the last equality .
now , both sides of the inequality are diagonalizable in the same basis . since the operator on the left hand side has eigenvalue @xmath147 and the operator on the right hand side has eigenvalue of @xmath148 , the inequality becomes @xmath149 .
hence , the maximum probability of teleportation compatible with the causality bound is @xmath150 thus , we have shown from causality that the original teleportation protocol is the optimal probabilistic simulation of an identity channel from the future to the past .
the probability of success for ideal channels provides a lower bound for the probability of success for arbitrary channels : indeed , if we have protocols that simulate ideal channels from the future to the past , then we can use these protocols to simulate a desired channel @xmath60 in two ways : _
i ) _ we can apply @xmath60 to the input in the future and then send the output to the past through the ideal channel , or _
ii ) _ we can send the input to the past through the ideal channel and apply the desired channel @xmath60 to the input in the past .
this argument leads to the following : let @xmath60 be a quantum channel sending states on the hilbert space @xmath37 to states on the hilbert space @xmath106 .
then , the maximum probability @xmath151 to simulate the action of @xmath60 from the future to the past satisfies the bound @xmath152 for classical channels , of the form @xmath153 and @xmath154 are orthonormal bases for @xmath37 and @xmath106 , respectively , the bound becomes @xmath155
in this section we focus on measure - and - prepare channels , namely channels @xmath156 of the form @xmath157 ~,\end{aligned}\ ] ] where @xmath158 is a povm on the input hilbert space @xmath37 , that is , @xmath159 and @xmath160 is a collection of normalized density matrices on the output hilbert space @xmath106 , that is , @xmath161
= 1 \quad \forall i \in x~.\end{aligned}\ ] ] measure - and - prepare channels are interesting as an example of channels that transmit only classical information : indeed , it is well known that the quantum capacity of a measure - and - prepare channel is always zero ( this follows , e.g. from the general upper bound presented by holevo and werner in ref .
@xcite ) . in this section
we will first show that the maximum probability to simulate a given measure - and - prepare channel from the future to the past satisfies the general lower bound @xmath162 where @xmath163 is the dimension of the output hilbert space .
the bound is achieved for @xmath164 by von neumann measurements on an orthonormal basis , and in that case coincides with the bound for the simulation of an ideal classical channel ( cf .
subsection [ subsect : idealclassical ] ) .
( [ mpbound ] ) is quite intuitive : if a channel can only transmit classical information , then the probability to simulate it is at least equal to the probability to simulate an ideal classical channel with the same output
. then we will compute the exact value of the probability of success for two particular examples of measure - and - prepare channels : the first example will be the channel associated to the optimal estimation of a pure state , namely the linear map @xmath165 sending states on @xmath89 to states on @xmath89 given by @xmath166 \qquad & \rho_u : = u |0\>\<0| u^\dag\\ & p_u : = d ~
u |0\>\<0| u^\dag~,\end{aligned}\ ] ] where @xmath10 is the dimension of @xmath89 and @xmath167 is a fixed state . in this case
the general bound is tight : the probability of success is given by @xmath168 the second example will consist in the _ universal not _
channel @xcite , namely the linear map @xmath169 sending states on @xmath89 to states on @xmath89 given by @xmath170 \qquad & \rho^{\perp}_u : = \frac{u ( i - |0\>\<0| ) u^\dag}{d-1 } ~~ \\ & p_u : = d ~ u |0\>\<0 | u^\dag ~~ \end{aligned}\ ] ] in this case we will show that the probability of success is given by @xmath171 well above the lower bound @xmath172 .
the intuition that the probability of success depends on the amount of information that a channel can transmit suggests that the universal not channel can transmit less classical information than the state estimation channel .
the choi operator of the generic measure - and - prepare channel @xmath173 in eq .
( [ eq : mp ] ) is given by @xmath174 let us consider the causality bound @xmath175 , where @xmath50 is a state on @xmath106 to be optimized .
a ( possibly suboptimal ) choice for the state @xmath50 is @xmath176 , for which the bound becomes @xmath177 the bound will be satisfied as long as we choose @xmath178 , where @xmath179 is the largest eigenvalue of the choi operator @xmath180 .
we now show that @xmath181 can not be larger than @xmath182 . to this purpose
, we consider an arbitrary normalized state @xmath183 and show that @xmath184 . using eq .
( [ doubleketprod ] ) , the normalization of the state @xmath185 reads @xmath186=1 .
\end{aligned}\ ] ] we then obtain @xmath187.\\ & = \sum'_i \tr [ \psi p_i \psi^{\dagger } ] \tr [ \rho_i \sigma_i ] \quad \sigma_i : = \frac { \psi p_i \psi^{\dagger } } { \tr[\psi p_i \psi^{\dagger}]}\\ & \le \sum_{i \in x } \tr [ \psi p_i \psi^{\dagger } ]
\\ & = \tr[\psi^\dag \psi]\\ & = 1.\end{aligned}\ ] ] here we used eq .
( [ doubleketalg ] ) for the first equality and eq .
( [ doubleketprod ] ) for the second .
the sum in the third equation is restricted to the values of @xmath188 such that @xmath189 \not = 0 $ ] .
the inequality is due to the fact that the overlap between the density matrices @xmath190 and @xmath191 is bounded by @xmath182 , the fourth equality is due to the normalization condition @xmath192 , and the last equality is just eq .
( [ psinorm ] ) .
since the bound @xmath193 holds for arbitrary @xmath194 , we conclude that the maximum eigenvalue of @xmath195 satisfies @xmath196 .
this implies that the value @xmath197 is compatible with the causality bound , and therefore , it is achievable with a suitable generalized teleportation protocol .
hence , for the maximum probability @xmath198 we have the lower bound @xmath199 .
the choi operator for the pure state estimation channel is @xmath200 clearly , @xmath201 enjoys the symmetry @xmath202 for every @xmath203 , which implies that the causality bound has the form @xmath204 ( the proof is the same as the proof for the ideal quantum channel in subsection [ sec : original ] ) .
since the maximum eigenvalue of @xmath201 is @xmath205 , we have that the maximum probability is @xmath206 . using eqs .
( [ cest ] ) and ( [ unot ] ) , the universal not channel can be written as @xmath207 hence , its choi operator is given by @xmath208 again , we observe that @xmath209 enjoys the symmetry @xmath210 for every @xmath203 , which implies that the causality bound has the form @xmath211 now , the maximum eigenvalue of @xmath209 is @xmath212 .
hence , the maximum probability is given by @xmath213 .
in this section we consider probabilistic protocols that take @xmath0 input copies of a pure state in the future and produce @xmath1 copies of the same state in the past . in this case , the channel that we want to simulate is the partial trace over @xmath214 systems , symmetrized over all possible choices of the systems that are traced : @xmath215 , \end{aligned}\ ] ] here the partial trace is over all input hilbert spaces except the first @xmath16 , @xmath216 is an element of the group @xmath217 of the permutations of @xmath0 objects and @xmath218 is the unitary operator that permutes the @xmath0 input hilbert spaces according to the permutation @xmath216
. the goal will be to see how the probability of success , denoted by @xmath219 , varies with the number of input and output copies .
in particular , we will show that the probability of success increases with the number of available input copies .
note that in the classical world there can not be any increase of the probability of success with the number of copies of an pure input state : since we can perfectly clone the classical pure states corresponding to an orthonormal basis , we must have @xmath220 the classical value @xmath221 is clearly an upper bound for the probability of success in the quantum case : indeed , if we have a probabilistic teleportation protocol that succeeds with probability @xmath15 in transforming @xmath0 copies of a pure quantum state into @xmath16 copies of the same quantum state , then the protocol will work in particular on the states of an orthonormal basis , thus simulating the classical trace channel from @xmath0 to @xmath16 copies .
hence , we must have @xmath222 this bound is very intuitive : if there existed a quantum protocol that succeeded with probability @xmath223 , then two parties , alice and bob , could use the protocol to reliably win a lottery where the goal is to guess today a winning number @xmath224 between 1 and @xmath10 that will be extracted tomorrow . to win the lottery , alice and bob can agree to use an orthonormal basis @xmath225 such that the basis vector @xmath226 corresponds to the number @xmath224 .
tomorrow , when the winning number is announced , bob inputs into the teleportation machine @xmath0 identical systems , each of which is prepared in the state @xmath226 corresponding to the winning number .
today alice measures one of the @xmath16 copies in the agreed basis and with probability @xmath15 she obtains the winning number , just in time to submit her lottery ticket .
if @xmath227 were larger than the classical value @xmath228 , then alice would win the lottery with probability larger than the probability @xmath229 to win the lottery by a random guess .
in the following we will show that , unlike in the classical case , in the quantum case the probability of successful teleportation increases with the number of input copies .
first , we will consider the simplest case @xmath230 , where the calculation of the maximum probability of success can be done explicitly even without constraining the input copies to be in a pure state . in this case
, we find out that the probability increases from the value @xmath231 in the @xmath232 case to the value @xmath233 .
this result suggests that the quantum improvement is a general feature that does not depend on the assumption of pure input states .
we will then move to the case of general @xmath0 , keeping the number of output copies fixed to @xmath17 . in this case , we will that the success probability approaches the classical value @xmath234 as @xmath235
. an explanation of this asymptotic behaviour will be provided in sec .
[ sec : summary ] .
finally , we will extend the analysis to general @xmath16 , also including the case @xmath236 , where optimal clones of the input state are teleported to the past .
surprisingly , in the case of cloning channels we will show that for fixed @xmath0 the probability of success is an increasing function of the number of output copies @xmath16 .
also this fact will be explained in sec .
[ sec : summary ] . for two copies of input and one copy of output ,
the input hilbert space is @xmath237 and the output hilbert space is @xmath238 . in this case ,
the quantum channel to be simulated is @xmath239 where the labels 1 and 2 label on the right hand side label the two input spaces . using eq .
( [ correspondence ] ) , we find that the choi operator of this channel is given by @xmath240 where we have used 0 to label the output hilbert space . plugging this into the causality bound of eq .
( [ choiinequality ] ) we obtain @xmath241 like in subsection [ sec : original ] , we can notice a symmetry of the choi operator : @xmath242 for every @xmath243 . applying this transformation to both sides of the inequality ( [ oramuoio ] ) and integrating over the haar measure we obtain @xmath244 it now remains to find the eigenvalues of the choi operator @xmath245 . to do this ,
we define the following set of orthonormal vectors in @xmath246 : @xmath247 in terms of these vectors , the choi operator can be rewritten as @xmath248 thus we see that the eigenvalues of this choi operator are @xmath249 and @xmath250 . to ensure that eq .
( [ oramuoio ] ) holds , we must have @xmath251 thus , the maximum teleportation probability from 2 copies of input is @xmath252 which is greater than the maximum probability of teleportation from 1 to 1 copy , @xmath253 . in other words , adding an extra copy of the input yields improvement in the teleportation probability .
it is worth stressing that , since we are simulating the symmetrized trace channel @xmath254 , any bipartite state @xmath255 on @xmath37 with the property @xmath256 + \tr_{2 } [ \sigma ] ) /2 = \rho $ ] will be transformed with probability @xmath257 into @xmath53 .
this holds in particular when @xmath258 , corresponding to two identical copies of the mixed state @xmath53 .
we now calculate the maximum probability of teleportation from @xmath0 input copies to one output copy . in this case
, we will only require the teleportation protocol to work perfectly when we are inputting @xmath0 copies of the same pure state .
since the density matrix of @xmath0 identical copies of a pure state has support in the symmetric subspace of @xmath260 , henceforth denoted as @xmath261 , we will focus on the restriction of the trace channel to the symmetric subspace , namely on the channel @xmath262 given by @xmath263 where @xmath264 is the projector on @xmath261 and @xmath265 is the symmetrized partial trace channel defined in eq .
( [ symmtrace ] ) .
note that the map @xmath262 is trace - preserving only for states with support in @xmath261 .
it is useful to observe that , due to the projection on the symmetric subspace , the sum over all permutations in eq .
( [ symmtrace ] ) can be omitted : indeed , we have @xmath266 \\ & \qquad = \tr_{n } \dots \tr_{2 } \left [ p_+^{(n ) } \rho p_{+}^{(n ) } \right ] , \end{aligned}\ ] ] having used the property @xmath267 for every permutation @xmath268 .
( color online ) probabilistic teleportation from @xmath0 copies to @xmath17 copy of a pure state . here
, @xmath101 represents the symmetrized partial trace over all but one input copy , restricted to the symmetric subspace . ] by the definition of choi operator in eq .
( [ correspondence ] ) we find that the choi operator @xmath269 is given by @xmath270 where the output hilbert space is labelled by @xmath118 , the input hilbert spaces are labelled with numbers from @xmath182 to @xmath0 , and the projector @xmath271 acts on the tensor product of all input hilbert spaces except the first .
we want to find the maximum value @xmath272 compatible with the causality bound of eq .
( [ choiinequality ] ) , which now reads @xmath273 to this purpose , we notice that the choi operator @xmath269 has the symmetry @xmath274 for every @xmath275 . following again the group theoretic argument of subsection [ sec : original ] we can rewrite eq . ( [ eq : causaln1 ] ) as @xmath276 hence , it only remains to find the eigenvalues of @xmath269 .
expanding the projector @xmath277 as @xmath278 where @xmath279 is an orthonormal basis for @xmath280 , we can express @xmath281 as @xmath282 : indeed , we have @xmath283 |\varphi_m\ > \\ & = \frac{d^{(n)}_+}{d^{(n-1)}_+ } ~ \ < \varphi_n| p^{(n-1)}_{+ } |\varphi_m\>\\ & = \frac{d^{(n)}_+}{d^{(n-1)}_+ } ~ \delta_{n , m } , \end{aligned}\ ] ] that is , the vectors @xmath284 are mutually orthogonal and have the same norm
. therefore , @xmath269 only has one non - zero eigenvalue , given by @xmath285 . plugging this value into the causality bound of eq .
we obtain @xmath286 .
the maximum probability of success is then given by @xmath287 note that the probability of successful teleportation increases with the number @xmath0 of input copies , unlike in the classical case .
for @xmath232 we retrieve the value @xmath288 of the standard teleportation protocol , while in the limit of @xmath0 going to infinity we observe that @xmath272 tends towards the classical limit @xmath289 .
such an asymptotic behaviour will be explained in section [ sec : summary ] .
here we calculate the maximum probability of teleportation from @xmath0 input copies to @xmath1 output copies of a generic pure state .
again , since for every integer number @xmath290 the density matrix of @xmath290 identical copies of a pure state has support in the symmetric subspace @xmath291 , we will restrict the input of the partial trace channel @xmath292 to be in @xmath293 and the output to be in @xmath294 .
in other words , we will focus on the probabilistic simulation of the channel @xmath295 given by @xmath296 p^{(m)}_+ ~ , \end{aligned}\ ] ] where @xmath297 is the symmetrized partial trace channel defined in eq .
( [ symmtrace ] ) .
again , thanks to the projection on the symmetric subspace we can write @xmath298 \right ) p^{(m)}_+ ~. \end{aligned}\ ] ] ( color online ) probabilistic teleportation from with @xmath0 input copies in the future to @xmath1 output copies in the past .
here , @xmath101 represents the symmetrized partial trace over all but @xmath16 of the @xmath0 input systems , where the input and output are constrained to be in the symmetric subspace . ] by the definition of choi operator in eq .
( [ correspondence ] ) , we then obtain @xmath299 where the projector @xmath300 acts on the tensor product of the last @xmath301 input hilbert spaces .
our goal is to find the maximum probability of success compatible with the causality bound @xmath302 where @xmath303 is a suitable state on @xmath304 and @xmath305 is the identity on @xmath306 .
now , it is easy to see that the choi operator @xmath307 has the symmetry @xmath308 for every @xmath243 . due to this symmetry
, the state @xmath303 in the causality bound can be chosen without loss of generality to be the invariant state @xmath309 , so that the bound becomes @xmath310 to find the maximum probability it is enough to find the maximum eigenvalue of @xmath311 .
expanding the projector @xmath312 as @xmath313 for some orthonormal basis @xmath314 for @xmath315 , we can rewrite eq .
( [ cnmlast ] ) as @xmath316 having defined @xmath317 and @xmath318 : indeed , we have @xmath319 & = \frac{d^{(n)}_+}{d^{(n - m)}_+ } ~ \ < \varphi_m| p^{(n - m)}_{+ } |\varphi_n\ > \\ & = \frac{d^{(n)}_+}{d^{(n-1)}_+ } ~ \delta_{mn } , \end{aligned}\ ] ] that is , the vectors @xmath320 are mutually orthogonal and have the same norm .
therefore , @xmath321 only has one non - zero eigenvalue , given by @xmath322 . the causality bound of eq .
( [ boundnm ] ) is equivalent to @xmath323 , which implies that the maximum probability is given by @xmath324 note that @xmath325 converges to the value @xmath326 in the limit of @xmath0 going to infinity .
such an asymptotic behaviour will be explained in section [ sec : summary ] .
we now generalize the results of the previous subsection to the case @xmath236 .
of course , in this case we can not expect to obtain @xmath16 perfect output copies of the input state , because that would violate the no - cloning theorem@xcite .
what we can do is to consider the best possible approximation of the ideal cloning process .
such an optimal approximation is given by the optimal _ universal cloning _
channel @xcite , here denoted by @xmath327 for consistency of notation .
the cloning channel @xmath327 is given by @xmath328 note that the linear map @xmath295 is trace - preserving only for density matrices @xmath53 with support within the symmetric subspace @xmath329 .
indeed , @xmath295 should be regarded as a quantum channel sending states on @xmath330 to states on @xmath331 . from eqs .
( [ cnm ] ) and ( [ clonnm ] ) , we can note the existence of an important duality between universal cloning and partial trace : for two arbitrary operators @xmath332 on @xmath333 and @xmath334 on @xmath335 one has @xmath336
= \frac { d_+^{(m ) } } { d_+^{(n ) } } ~ \tr [ \map c^+_{q , m \to n } ( \mu ) ~ \nu ] . \end{aligned}\ ] ] in other words , the map representing the universal cloning from @xmath0 to @xmath16 copies in the heisenberg picture is the map representing the partial trace from @xmath16 to @xmath0 copies in the schrdinger picture , up to a positive scaling factor
. we will refer to this property as the _ time - reversal duality _ between partial trace and universal cloning . in terms of choi operators ,
the duality can be written as @xmath337 where @xmath338 is the operator that exchanges the @xmath0 output hilbert spaces with the @xmath16 input hilbert spaces for the channel @xmath339 .
( [ clontracechoiduality ] ) allows us to easily find the probability of success for the cloning channel @xmath295 .
first of all , eq . ( [ clontracechoiduality ] ) implies that @xmath340 has the symmetry @xmath341 for every @xmath342 .
using this symmetry the causality bound becomes @xmath343 now , eq .
( [ clontracechoiduality ] ) implies that the above equation is equivalent to @xmath344 which is nothing but the causality bound for the trace channel @xmath345 [ cf .
( [ boundnm ] ) ] .
therefore , we have that the causality bound for the cloning channel @xmath346 is satisfied with probability @xmath8 if and only if the causality bound for the trace channel @xmath347 is satisfied with probability @xmath348 .
hence , we have @xmath349 quite surprisingly , the above value increases with the number @xmath16 of output copies , reaching the asymptotic value @xmath350 in the limit @xmath351 .
this asymptotic behaviour will be explained in section [ sec : summary ] .
the results obtained so far indicate a dependence of the maximum probability of success on the amount and type ( quantum or classical ) of information that the channel can transmit : first , we noted that the only channels that can be simulated with unit probability are the erasure channels , that produce a fixed output state independently of the input , that is , they do not transmit any information .
then , we saw that measure - and - prepare channels , that only transmit classical information , can be simulated with probability @xmath352 , @xmath163 being the dimension of the output hilbert space .
finally , channels that transmit both quantum and classical information can be simulated with probability @xmath353 , as in eq .
( [ generallower ] ) in the following we will make a quantitative connection between the amount of information transmitted by a channel and the probability of simulating it with a generalized teleportation protocol : we will show that the probability of success is upper bounded by the inverse of a measure of information defined as the maximum payoff in a communication game .
we will then use this result to explain the asymptotic behaviour of the probability of success for the simulation of a partial trace channel with a large number of input systems and of a universal cloning channel with a large number of output systems . in this subsection
we relate probability of success in the simulation of a channel to the past with the amount of information transmitted by the channel . in this context , a convenient measure of information is the expected payoff obtained by two parties in a communication game .
suppose that the channel @xmath60 is used to transfer states from alice s side to states on bob s side .
alice can encode a set of messages @xmath354 in a set of states @xmath355 on @xmath356 and communicate a message to bob by sending the corresponding state through the channel @xmath60 .
on his side , bob can try to decode alice s message from the output state @xmath357 .
generally , the decoding will be described by a povm @xmath358 on the output hilbert space @xmath106 , with the outcome @xmath359 corresponding to the decoded message .
usually , but not necessarily , the set of decoded messages @xmath354 coincides with the set of encoded messages @xmath360 . to assess the success of the communication
, we can introduce a payoff function @xmath361 , that quantifies the gain ( or the loss ) associated to decoding @xmath362 while the encoded message was @xmath188 . for given @xmath188 ,
the expected payoff will be @xmath363 .
\end{aligned}\ ] ] now , suppose that alice is in the future and bob is in the past , and that they are using a probabilistic simulation of channel @xmath60 .
the causality bound of eq .
( [ mapsinequality ] ) imposes @xmath364 , and , therefore @xmath365 having defined @xmath366 $ ] . introducing a prior probability distribution @xmath367 on alice s messages we can then consider the average expected payoff @xmath368 hence , the bound of eq .
( [ fidboundi ] ) becomes @xmath369 note that the value @xmath370 depends only on the payoff function @xmath371 and on the prior probabilities @xmath367 .
we have now arrived at the bound @xmath372 which holds for every given payoff function @xmath371 , prior probabilites @xmath373 , encoding @xmath374 , and decoding @xmath375 . maximizing the average payoff over all possible choices of encoding @xmath376 and decoding @xmath377
we then get the maximum value @xmath378 which quantifies the maximum payoff that can be achieved when using the channel @xmath60 in a communication game with given payoff function @xmath371 and prior probabilities @xmath379 .
we then conclude that the maximum probability to simulate the channel @xmath60 , denoted by @xmath151 , has to satisfy the bound @xmath380 for every payoff function @xmath371 and for every prior @xmath379 . the above bound states that the probability of success can not be greater than the inverse of the maximum amount of statistical information ( average payoff ) transmitted by the channel . for this reason
, we refer to this bound as the _ statistical information bound_. notice that the derivation of the information bound in eq .
( [ infobound ] ) can be extended to the case where alice and bob are allowed to communicate through an ideal quantum channel on a reference system with hilbert space @xmath381 , in addition to the channel @xmath60 . in this case
, alice can encode her message @xmath382 in a bipartite state @xmath383 on @xmath384 .
later , she will send the input system through the channel @xmath60 and the reference through the ideal channel .
bob s task will then be to decode the message from the output state @xmath385 , through a bipartite povm @xmath386 on @xmath387 .
following the same steps as above , in this case we obtain the _ entanglement - assisted statistical information bound _ @xmath388 having defined @xmath389 \right\}\\ \mathbb e^{(r)}_{ave } ( f ) : = \sum_{i\in x } \pi_i \left\ { \sum_{j\in y } f(i , j ) ~ \tr\left [ \rho_0 \otimes \tr_{in}[\rho_i ] ) p_j \right ] \right\}.\end{aligned}\ ] ] note , however , that both statistical information bounds are just upper bounds : in general there is no guarantee about their achievability .
the aim of the following discussion is to provide an explanation for the asymptotic behaviour of the probability of success in the simulation of the partial trace channel @xmath391 in the limit @xmath7 .
first , we apply the statistical information bound of eq .
( [ infobound ] ) to show that the probability of success satisfies the bound @xmath392 consider a communication scenario where alice encodes a message in state from the set @xmath393 and sends it to bob through the partial trace channel @xmath394 .
suppose that bob s decoding is given by the povm @xmath395 where @xmath396 is the normalized invariant measure on the pure states of @xmath89 . as a payoff function ,
let us assume the dirac delta @xmath397 , so that the expected payoff for input @xmath398 is just the probability ( density ) to make the correct guess @xmath399 p^{(m)}_\psi \right\ } \\ &
= \tr [ \left ( |\varphi\>\<\varphi| \right)^{\otimes m } p^{(m)}_\varphi ] \\ & = d_+^{(m ) } \qquad \forall |\varphi\ > \in { \mathcal{h}}. \end{aligned}\ ] ] since the expected payoff is the same for every state , we have @xmath400 for every prior distribution . on the other hand , choosing the uniform distribution @xmath401 for the input states
, we have @xmath402 \\ & = 1 .
\end{aligned}\ ] ] the statistical information bound of eq .
( [ infobound ] ) then gives @xmath403 . in subsection
[ subsect : tracenm ] we observed that the probability of success @xmath15 asymptotically approaches the value @xmath404 in the limit @xmath7 .
an explanation of this behaviour is the following : from ref .
@xcite , we know that in the limit @xmath7 the partial trace channel @xmath405 converges to the estimation channel @xmath406 defined by @xmath407 . \end{aligned}\ ] ] with @xmath408 .
hence , the probability of success @xmath15 must converge to @xmath409 , the probability of success for the estimation channel @xmath410 . on the other , hand , since @xmath411 is a measure - and - prepare channel , by the general bound of eq .
( [ mpbound ] ) we must have @xmath412 . in conclusion , we obtained @xmath413 that is , @xmath414 .
in other words , the asymptotic behaviour of @xmath15 is dictated by the statistical information bound and by the convergence of the trace channel @xmath415 to a measure - and - prepare channel .
an illustration of the situation for the case of @xmath17 is given in fig .
[ fig : channelsgraph ] .
it is worth stressing , once more , that no improvement with the number of input copies would be possible for a classical partial trace channel : since in that case the probability is already @xmath289 , any improvement would violate the causality bound for ideal classical channels . hence , the increase of the probability of teleportation with the number @xmath0 of input copies is intrinsically a quantum phenomenon . a graph of the maximum probability @xmath8 to simulate the partial trace channel from @xmath0 copies to @xmath17 copy of a pure input state .
the maximum probabilities for erasure channels , measure - and - prepare channels , and ideal classical and quantum channels are also located in the plot . ] here we consider the universal cloning channel @xmath405 .
as we did for the trace channels , we will now provide an explanation for the asymptotic behaviour of the probability of success in the limit @xmath351 .
a direct explanation comes from the time - reversal duality between cloning and partial trace , which implies the relation @xmath417 and thus allows one to reduce the asymptotic behaviour of the probability of success of a cloning channel with many output copies to the asymptotic behaviour of the probability of success for a trace channel with many input copies .
a less direct , but perhaps more insightful explanation can be obtained by following the same steps used in the previous subsection .
let us start by showing that the statistical information bound of eq .
( [ infobound ] ) implies that the probability of success satisfies the bound @xmath418 consider a communication scenario where alice encodes a message in state from the set @xmath393 and sends it to bob through the cloning channel @xmath394 .
suppose that bob s decoding is given by the povm @xmath419 , where @xmath396 is the normalized invariant measure on the pure states of @xmath89 . as a payoff function
we assume again the dirac delta @xmath420 , so that the expected payoff for input @xmath398 is the probability ( density ) of the correct guess : @xmath421 p^{(m)}_\psi \right\ } \\ & = \tr\left\ { \map c_{q , n \to m}^+ \left [ \left ( |\varphi\>\<\varphi| \right)^{\otimes n } \right ] p^{(m)}_\varphi \right\ } \\ & = \frac { d_+^{(n ) } } { d_+^{(m ) } } \tr\left [ \left ( |\varphi\>\<\varphi| \right)^{\otimes n } \map c^+_{q , m \to n } \left ( p^{(m)}_\varphi \right ) \right ] \\ & = \tr [ \left ( |\varphi\>\<\varphi| \right)^{\otimes n } p^{(n)}_\varphi ] \\ & = d_+^{(n ) } \qquad \forall |\varphi\ > \in { \mathcal{h}}. \end{aligned}\ ] ] since the expected payoff is the same for every state , for every prior distribution we have @xmath422 . on the other hand , we already know that @xmath423 , so that the information bound of eq .
( [ infobound ] ) gives @xmath424 . on the other hand , from ref .
@xcite we know that in the limit @xmath425 the cloning channel @xmath405 converges to the estimation channel @xmath411 . hence , the probability of success @xmath15 must converge to @xmath409 . again , we obtain the relation @xmath426 that is , @xmath427 . also in this case , the asymptotic behaviour of the probability @xmath428 is dictated by the statistical information bound and by the convergence of the cloning channel @xmath429 to a measure - and - prepare channel .
this work showed how causality determines the maximum probability of simulating a given quantum channel from the future to the past . since causality states that it is not possible to signal from the future to the past
, the probability of success in the simulation should be small enough to prevent any signal from being sent . as a consequence , channels that are able to transmit more
information will have a smaller probability of being simulated , while channels that transmit less information will have a larger probability .
in particular , erasure channels , that transmit no information at all , can be simulated with unit probability , while measure - and - prepare channels , that transmit only classical information , can be simulated with probability at least equal to @xmath430 the inverse of the output hilbert space dimension .
the hardest channels to simulate are the ideal quantum channels , for which the probability is equal to @xmath431 , the inverse of the square of the hilbert space dimension .
a quantitative connection between the amount of information transmitted by a channel and the probability of successful simulation is given by the statistical information bounds , which state that the probability of success is upper bounded by the inverse of the statistical information transmitted by the channel .
in the paper we computed explicitly the maximum probability for the simulation of several channels , including partial trace channels from @xmath0 input systems to @xmath28 output systems , and universal cloning channels from @xmath0 input systems to @xmath432 output systems .
we pointed out a time - reversal duality between trace and cloning and we exploited it to show that the value of the maximum probability of success is given by @xmath433 . note that for fixed @xmath16 , the probability of success is an increasing function of @xmath0 , the number of input copies .
similarly , for fixed @xmath0 , the probability of success is an increasing function of @xmath16 , the number of output copies .
these are genuinely quantum features , that are impossible in the classical world : indeed , classically the probability of success for all trace channels and for all cloning channels is given by @xmath434 , independently of @xmath0 and @xmath16 . in the case of a single output copy
, @xmath435 the probability of success for a quantum partial trace channel tends to the classical value @xmath434 in the limit of @xmath0 going to infinity .
this asymptotic behaviour can be explained from the fact that a trace channel with asymptotically large number of input copies converges to a measure - and - prepare channel @xcite .
the same explanation can be provided for the asymptotic behaviour of the probability of success for cloning channels .
also in this case the asymptotic behaviour is dictated by the fact that a universal cloning channel with asymptotically large number of output copies converges to a measure - and - prepare channel @xcite .
note that in this paper we focused on determining the maximum probability of success for many channels , but we did not show explicitly the generalized teleportation schemes that achieve the simulation of those channels . in particular , it is interesting to ask which kind of entangled states and which kind of measurements allow one to achieve teleportation from @xmath0 input copies to @xmath16 output copies .
moreover , it is worth asking whether these probabilistic teleportation protocols can be extended to deterministic protocols ( now , from the past to the future ) by reintroducing classical communication and correction operations , as in the case of tele - cloning @xcite .
the fact that the probability of successful teleportation from @xmath0 copies to @xmath17 copy converges to @xmath436 suggests that our probabilistic protocol could be extended to a deterministic protocol that needs only @xmath437 bits of classical communication , instead of the @xmath438 bits of the original teleportation protocol .
however , it is quite possible that in general there exists a trade - off between maximizing the probability of success for a particular outcome and being able to implement a correction for all outcomes .
all these issues will be investigated in a forthcoming work @xcite .
another interesting direction of future research is to consider the simulation of multipartite channels under the constraint that different input and output systems can arrive at different instants of time , and to show how the probability of simulation depends on the possible causal arrangements of the input / output systems .
for example , one could consider the case of a cloning channel from @xmath232 input copy to @xmath439 approximate copies , one copy being sent to the past and the other being sent to the future .
examples of this kind are expected to shed light on the interplay between causal structure and quantum information flow .
90 g. chiribella , g. m. dariano , and p. perinotti , _
probabilistic theories with purification " _ , phys .
a * 81 * , 062348 ( 2010 ) g. chiribella , g. m. dariano , and p. perinotti , _ informational derivation of quantum theory " _ , phys . rev .
a * 84 * , 012311 ( 2011 ) .
bennett , g. brassard , c. crepeau , r. jozsa , a. peres , and w.k .
wooters , `` _ _ teleporting an unknown quantum state via dual classical and einstein - podolski - rosen channels _ _ '' , phys . rev
. lett . * 70 * , 1895 ( 1993 ) .
s. lloyd , l. maccone , r. garcia - patron , v. giovannetti , y. shikano , s. pirandola , l. a. rozema , a. darabi , y. soudagar , l. k. shalm , a. m. steinberg _ `` closed timelike curves via post - selection : theory and experimental demonstration '' _ , phys .
rev .lett . *
106 * , 040403 , ( 2011 ) g. m. dariano and c. macchiavello , _ optimal phase - covariant cloning for qubits and qutrits " _ ,
a * 67 * , 042306 ( 2003 ) .
w. k. wooters and w. h. zurek , _ a single quantum can not be cloned _ , nature * 299 * , 802 ( 1982 ) g. chiribella , _ `` on quantum estimation , quantum cloning , and finite quantum de finetti theorems '' _ , theory of quantum computation , communication , and cryptography , lecture notes in computer science , * 6519/2011 * , 9 ( 2011 ) | we introduce the study of quantum protocols that probabilistically simulate quantum channels from a sender in the future to a receiver in the past .
the maximum probability of simulation is determined by causality and depends on the amount and type ( classical or quantum ) of information that the channel can transmit .
we illustrate this dependence in several examples , including ideal classical and quantum channels , measure - and - prepare channels , partial trace channels , and universal cloning channels . for the simulation of partial trace channels ,
we consider generalized teleportation protocols that take @xmath0 input copies of a pure state in the future and produce @xmath1 output copies of the same state in the past . in this case , we show that the maximum probability of successful teleportation increases with the number of input copies , a feature that was impossible in classical physics . in the limit of asymptotically large @xmath0
, the probability converges to the probability of simulation for an ideal classical channel .
similar results are found for universal cloning channels from @xmath0 copies to @xmath2 approximate copies , exploiting a time - reversal duality between universal cloning and partial trace . |
dark energy , discovered via the observed luminosity distances out to high-@xmath5 type ia supernovae @xcite , is the most mysterious element in physics today ( see * ? ? ?
* for a recent review ) .
as dark energy primarily affects the expansion rate of the universe , one can gain information on the nature of dark energy by measuring the cosmological distances as well as the expansion rates of the universe accurately .
while the cosmic microwave background ( cmb ) and the type ia supernovae can be used for measuring the angular diameter distance out to @xmath6 and the luminosity distances out to @xmath7 , respectively , the power spectrum of matter distribution in the universe can be used to measure the angular diameter distances _ as well as _ the expansion rates of the universe out to a wider range of redshifts . _
two _ length scales are encoded in the matter power spectrum , @xmath8 ( see , e.g. , * ? ? ?
* ) : * the comoving hubble horizon size at the matter - radiation equality , @xmath9 $ ] . *
the comoving sound horizon size at the so - called drag epoch at which baryons were released from photons , @xmath10 , where @xmath11 $ ] is the sound speed of photon - baryon fluid .
the former determines the overall shape of the power spectrum of dark matter including the location of the peak of @xmath8 at @xmath12 , whereas the latter determines the location of the baryonic features called the baryon acoustic oscillations ( baos ) .
these length scales can be predicted from the 5-year data of the wilkinson microwave anisotropy probe ( wmap ) @xcite , the wmap 5-year data yield @xmath13 , @xmath14 , @xmath15 , and @xmath16 . ] : @xmath17 and @xmath18 these lengths can be used as the `` standard rulers , '' which give us the angular diameter distances as well as the expansion rates of the universe @xcite .
we , as observers who measure the angular and redshift distribution of galaxies , can measure _ four _ distance ratios given by @xmath19 where @xmath20 is the proper ( i.e. , not comoving ) angular diameter distance .
we measure @xmath21 and @xmath22 by comparing the predicted lengths with the corresponding observed lengths perpendicular to the line of sight , and @xmath23 and @xmath24 from the lengths parallel to the line of sight . , and that perpendicular to the line of sight , @xmath25 , i.e. , @xmath26 .
the angular observables , @xmath27 and @xmath28 , are measured from @xmath25 , while the line - of - sight observables , @xmath29 and @xmath30 , are measured from @xmath31 . ]
the baos have been detected in the current galaxy redshift survey data from the sloan digital sky survey ( sdss ) and the two - degree field galaxy redshift survey ( 2dfgrs ) @xcite .
however , the current data are not yet sensitive enough to yield @xmath20 and @xmath32 separately @xcite ; thus , one can only determine a combined distance scale ratio from the spherically averaged power spectrum .
since two spatial dimensions are available on the sky and one dimension is available along the line of sight , one can measure @xmath33^{1/3 } = \frac{r_s(z_{drag})}{[(1+z)^2d_a^2(z)c / h(z)]^{1/3}}.\ ] ] @xcite have measured this quantity at @xmath34 from the sdss luminous red galaxies ( lrg ) , and @xcite have extended their analysis to include more data from the sdss lrg , as well as the sdss main galaxy samples and the 2dfgrs galaxies at @xmath35 .
@xcite have combined these measurements with the cmb distance ratios determined from the wmap 5-year data , the `` wmap distance priors , '' to obtain the constraints on dark energy properties .
the analysis performed in @xcite is a proto - type of what one can do in the future .
it is clear that we can gain more information if we can measure @xmath20 and @xmath32 simultaneously at various redshifts .
therefore , in the future we should be able to perform a much more sensitive test of dark energy properties by combining @xmath20 and @xmath32 from the future galaxy survey data , and the cmb distance priors from the future cmb experiments such as planck .
moreover , the baos capture only a part of information encoded in the shape of @xmath8 .
one would miss another baryonic feature , the silk damping scale , by only measuring baos .
a more serious drawback is that one would miss the other prominent standard ruler , @xmath36 , completely , by only measuring baos .
nevertheless , there is one major advantage of using baos : the phases ( not the amplitude ) of baos are less sensitive to the distortion of the shape of @xmath8 due to non - linear matter clustering , non - linear galaxy bias , or non - linear redshift space distortion @xcite . as a result ,
many studies have focused on developing various ways to extract the distance information from baos .
most of the previous work focused only on extracting the baos from the spherically averaged @xmath8 ( which gives @xmath2 ) ( e.g. , * ? ? ? * ) .
@xcite have studied the monopole and quadrupole moments in the galaxy power spectrum and their implications for determinations of the dark energy equation of state parameter , @xmath37 , and concluded that even in the worst case scenario ( i.e. , absence of the baos feature on the observed power spectrum ) , galaxy survey can still provide useful limits on @xmath37 from a combination of the monopole and quadrupole power spectra .
recently , @xcite have explored an extraction of the quadrupole moment of the two - dimensional power spectrum , @xmath38 , which gives a different distance combination , @xmath39 . in this paper
, we shall develop a method for extracting @xmath0 and @xmath1 simultaneously from the two - dimensional baos .
since we do not use spherical averaging or truncate the legendre expansion of baos at arbitrary orders , our method uses more information than most of the previous methods . to our knowledge
, the full two - dimensional extraction of @xmath0 and @xmath1 from baos has been explored only by @xcite .
this paper is organized as follows . in
[ sec : fitex ] we give a brief account of the original one - dimensional `` fit - and - extract '' ( fitex ) method , which was developed by @xcite for extracting baos from a spherically averaged one - dimensional @xmath8 .
we then extend this method to the two - dimensional fitex-2d method by including the full two - dimensional information without spherical averaging . in
[ sec : montecarlo ] we extract @xmath0 and @xmath1 from simulated noisy data using the fitex-2d method , and show that the fitex-2d yields unbiased estimates of @xmath0 and @xmath1 . in
[ sec : millennium ] we repeat the same analysis for a more realistic simulation , using the millennium simulation @xcite . in [ sec : darkenergy ] we propagate errors in @xmath32 and @xmath20 to those in the dark energy equation of state with the parametrization of @xmath40 .
we conclude in [ sec : conclusion ] . throughout this paper
we shall use the cosmological parameters given by @xmath41 , @xmath42 , @xmath43 , @xmath44 , and @xmath45 @xcite , which are the maximum likelihood values inferred from the wmap 5-year data @xcite combined with the current bao data @xcite and type ia supernova data @xcite .
we develop a method for extracting @xmath0 and @xmath1 simultaneously from the two - dimensional baos without spherical averaging .
our method builds upon the existing `` fit - and - extract '' ( fitex ) method developed by @xcite for extracting @xmath2 from a spherically averaged , one - dimensional @xmath8 .
the fitex method extracts baos by fitting and removing the non - oscillatory part of @xmath8 , which leaves only the oscillatory component , i.e. , baos .
@xcite model the non - oscillatory , smooth part by the following functional form : @xmath46 ^ 2 k^{n_s } , \label{eq : psmooth1d}\ ] ] where @xmath47 is the primordial tilt , while @xmath48 , @xmath49 , @xmath50 , @xmath51 , and @xmath52 are free parameters .
@xcite have shown that this function is flexible enough to fit out the smooth part of the spherically averaged @xmath8 measured from the hubble volume simulation @xcite .
they have tested the fitex method particularly for a large scale , @xmath53 , at high redshifts , @xmath54 , that are relevant to the hobby eberly dark energy experiment ( hetdex ; * ? ? ?
we make a simple extension of the one - dimensional fitex method by including angular dependence .
we model the two - dimensional smooth power spectrum by @xmath55 , \label{eq : psmooth2d}\end{aligned}\ ] ] where @xmath56 is the cosine of the angle @xmath57 between @xmath58 and the line of sight , i.e. , @xmath59 and @xmath60
. therefore , @xmath61 and @xmath62 for @xmath63 and @xmath64 , respectively . here
, @xmath65 is the legendre polynomials : @xmath66 the odd multipoles must vanish by symmetry .
one may include @xmath67 if necessary , but we find it sufficient to include the terms only up to @xmath68 . finally , @xmath69 is given by the 6th - order polynomials with only even powers of @xmath70 : @xmath71 where all of @xmath72 s are varied simultaneously for each @xmath73 .
the odd powers must vanish because they are not analytic in @xmath58 @xcite .
we include the terms only up to @xmath74 , as we include the multipoles up to @xmath68 .
if , for instance , @xmath75 is included , then @xmath76 may also be included for consistency . aside from the primordial tilt , @xmath47
, the fitex-2d contains 17 free parameters ( 5 for @xmath77 plus @xmath78 for the angular dependence ) . while it may sound like many , the number of data points available on the _ two - dimensional
_ power spectrum is usually much larger , and thus our fit is well behaved
. it may be instructive to use the conventional model for the redshift space power spectrum to show what these parameters are supposed to capture .
the leading order angular distortion is given by the so - called kaiser effect , which arises from coherent converging velocity flow toward the linear overdensity region @xcite .
the linear kaiser power spectrum is given by @xmath79p^{linear}(k),\end{aligned}\ ] ] where @xmath80 is a @xmath70-independent function that depends on the linear galaxy bias , @xmath81 , and the cosmological parameters ( mainly @xmath82 ) via @xmath83 where @xmath84 is the growth factor of linear density fluctuations .
we therefore find @xmath85 and the other terms are zero .
another example is the so - called finger - of - god ( fog ) effect , which arises from random motion within virialized halos
. when the distribution of the pairwise peculiar velocity within a halo is given by an exponential distribution with the velocity dispersion @xmath86 @xcite , one finds @xcite @xmath87 while the fog yields many terms when expanded into the legendre polynomials , it is still a good approximation to truncate the expansion at @xmath68 if @xmath70 is sufficiently smaller than @xmath88 .
note that the fog effect yields terms in the form of powers of @xmath89 ; thus , it makes sense to use the same number for the maximum power of @xmath70 ( see eq .
( [ eq : gl ] ) ) and the maximum multipole ( see eq .
( [ eq : psmooth2d ] ) ) of the fitex-2d fitting function . in general ,
neither of these two expressions are adequate .
the linear kaiser formula is valid only on very large scales , while the exponential fog formula is valid only on very small scales . at the intermediate scales we find more complicated expressions from , e.g. , the 3rd - order perturbation theory @xcite . to account for these complications
we have included @xmath70-dependent coefficients for the legendre polynomials . in figure [ fig1 ] and [ fig2 ]
we show the performance of @xmath90 . in figure [ fig1 ]
we show a simple analytical model .
] for the non - linear galaxy power spectrum in redshift space given by @xmath91\\ & & \times\frac1{1+f^2k^2\mu^2\tilde{\sigma}_v^2 } , \label{eq : pkred}\end{aligned}\ ] ] where @xmath92 , @xmath93 , and @xmath94 are the density - density , density - velocity , and velocity - velocity power spectra computed from the 3rd - order perturbation theory , and they are given by eq .
( 63 ) , ( 64 ) , and ( 65 ) in @xcite , respectively .
this form is similar to eq .
( 71 ) of @xcite , but we have replaced @xmath95 and @xmath96 in his formula by @xmath97 and @xmath98 , respectively , where @xmath99 is the 1-d peculiar velocity dispersion with an empirical fudge factor of @xmath100 calibrated off our simulations presented in @xcite . here , @xmath86 is given by @xmath101 we chose @xmath102 and @xmath103 .
the contour of power spectrum is anisotropic in fig .
[ fig1 ] due to the redshift space distortion ; however , we recover isotropy after subtracting the best - fitting @xmath90 from the anisotropic data ( see fig . [ fig2 ] ) .
we see that the baos have been extracted successfully , with isotropy of the oscillation phases recovered well .
in [ sec : err_fitex2d ] we show how well we can estimate @xmath0 and @xmath1 from the two - dimensional baos extracted from noisy data using the fitex-2d method . in
[ sec : err_full ] we compare the bao results to the accuracy one would obtain from the full modeling of @xmath38 , including the overall shape . in other words , for the former ( baos ) we only use @xmath28 and @xmath30 for measuring @xmath0 and @xmath1 , while for the latter ( full modeling ) we can use @xmath28 , @xmath30 , @xmath27 , @xmath104 , as well as the silk damping scale for measuring @xmath0 and @xmath1 , provided that non - linear effects ( non - linear matter clustering , non - linear redshift space distortion , and non - linear bias ) are under control .
note that the treatment of non - linear effects in this section is too simple to be realistic . for a more realistic treatment
we shall use the galaxy power spectrum from the millennium simulation @xcite in [ sec : millennium ] . to estimate errors in @xmath0 and @xmath1 from the fitex-2d method
, we use simple monte carlo simulations . for the underlying spectrum we use the same data as shown in fig .
[ fig1 ] , which includes a simplified modeling of non - linear matter clustering and non - linear redshift space distortion as given by eq .
( [ eq : pkred ] ) . as for the galaxy bias ,
we use a linear bias with @xmath103 . once the underlying spectrum is specified , it is straightforward to compute the errors in @xmath105 , @xmath106 , provided that the distribution of @xmath105 is a gaussian .
we use the standard formula that includes sampling variance as well as shot noise ( see , e.g. , * ? ? ?
* ) @xmath107 where @xmath108 is the number density of galaxies , @xmath109 is the survey volume , @xmath110 and @xmath111 are the fundamental wavenumbers , i.e. , the resolution in @xmath25 and @xmath31 .
we take these to be @xmath112 .
we use @xmath106 from eq .
( [ eq : sigp ] ) to calculate the r.m.s . error in @xmath105 , and generate 1000 monte carlo realizations .
we then apply the fitex-2d method to remove the smooth component from each realization to extract baos . for each realization , we measure @xmath0 and @xmath1 simultaneously by fitting the phases of extracted two dimensional baos to those of the reference baos extracted from either ( i ) the linear power spectrum , or ( ii ) the non - linear power spectrum given by eq .
( [ eq : pkred ] ) , with known @xmath113 and @xmath114 .
( later we find that using the linear spectrum as the reference bao yields the biased estimates of @xmath113 and @xmath114 . )
we use a simplex downhill method for @xmath115-minimization in the two - dimensional parameter space .
the number of free parameters for this analysis is two , i.e. , @xmath0 and @xmath1 , and we do not include the amplitude in the fit . we have checked that including the amplitude does not change the results very much , as the amplitude and the phases of baos are nearly uncorrelated ( see appendix [ sec : cc_no_beta ] for more details ) .
this is true in both real and redshift space .
when we apply fitex-2d to the simulated data , we perform a fit out to @xmath116 .
we choose the survey parameters , @xmath109 , @xmath5 , and @xmath108 , such that they roughly match those expected for the hobby - eberly dark energy experiment ( hetdex ) @xcite : @xmath117 , and @xmath118 with the sky coverage of 420 deg@xmath119 , which yields @xmath120 . emitting galaxies between @xmath118 over 420 deg@xmath119 in 3 years of observations on the hobby - eberly telescope . ]
we find that , when the phases extracted by fitex-2d are compared with the reference baos extracted from the linear power spectrum , the best - fitting values of @xmath0 and @xmath1 averaged over 1000 simulations disagree with the underlying , `` true '' values by 0.05% and 0.63% for @xmath0 and @xmath1 , respectively , due to the phase shift of baos caused by non - linearities ( including non - linear redshift space distortion ) .
this result extends the previous study by @xcite , who studied a spherically averaged 1-d power spectrum and found that the bias was less than 1% in @xmath121 . on the other hand , when the phases are compared with the reference baos extracted from the _ non - linear _ power spectrum ( eq .
( [ eq : pkred ] ) ) , the best - fitting values of @xmath0 and @xmath1 agree with the true values to well within the monte carlo sampling error ; thus , we confirm that the fitex-2d method yields unbiased estimates of @xmath0 and @xmath1 . in figure [ fig3 ]
we show the projected error ellipses on @xmath0 and @xmath1 from the baos extracted with the fitex-2d ( larger , dotted contours ; same in all four panels ) .
we find 1.8% and 2.5% errors on @xmath0 and @xmath1 , respectively , with the cross - correlation coefficient of @xmath122 , from the monte carlo simulations .
for the same survey parameters , the bao fisher matrix proposed by @xcite yields 1.5% and 2.5% errors on @xmath0 and @xmath1 , respectively , with @xmath123 .
therefore , we conclude that the fitex-2d method yields the results that nearly saturate the fisher matrix bound , i.e. , it is nearly an optimal method in a sense that it can yield the smallest errorbars one can obtain with the bao - only analysis . to calculate the errors in @xmath0 and @xmath1 expected from the full modeling of the two - dimensional galaxy power spectrum , @xmath124
, we use the fisher matrix given by ( see , e.g. , * ? ? ?
* ; * ? ? ?
* ) @xmath125 where @xmath126 for @xmath127 , 2, ... ,7 , respectively , @xmath128 , where @xmath48 is the overall amplitude of the power spectrum , @xmath98 is the linear redshift distortion parameter , @xmath129 is the calibrated 1-d velocity dispersion ( see eq .
( [ eq : pkred ] ) ) , and @xmath47 and @xmath130 describe the shape of the initial ( primordial ) power spectrum : @xmath131 here , the weight function , @xmath132 , is one half of the so - called `` effective volume , '' @xmath133 v_{survey } \equiv \frac12v_{eff}(k,\mu ) .
\label{eq : weight_fm}\ ] ] the effective volume is equal to the actual survey volume , @xmath109 , in the sampling variance dominated regime , @xmath134 , whereas it is small in the shot - noise dominated regime , @xmath135 . in figure [ fig4 ]
we show @xmath136 for @xmath117 and @xmath103 as a function of @xmath5 .
the factor of @xmath137 accounts for symmetry in @xmath138 .
the derivatives of @xmath139 with respect to @xmath140 are calculated and given in the appendix [ sec : derivatives ] .
cccccccc @xmath141 ( % ) & 0.279 & 0.877 & 0.317 & 0.282 & 0.479 & 0.416 & 1.100 + @xmath142 ( % ) &
0.437 & 0.786 & 1.124 & 0.801 & 0.509 & 0.539 & 1.134 + @xmath143 & 0.382 & @xmath144 & @xmath145 & 0.082 & @xmath146 & @xmath147 & 0.038 + @xmath148 ( % ) & 0.187 & 0.762 & 0.317 & 0.259 & 0.386 & 0.363 & 0.775 + & @xmath98 , @xmath129 & @xmath149 , @xmath129 & @xmath149 , @xmath98 , @xmath129 & @xmath149 , @xmath47 , @xmath130 & @xmath149 , @xmath98 , @xmath47 & @xmath149 , @xmath98 , @xmath129 , + & & & & & @xmath130 & @xmath47 , @xmath130 + @xmath141 ( % ) & 0.327 & 0.891 & 1.101 & 1.089 & 1.233 & 1.250 + @xmath142 ( % ) & 1.457 & 1.101 & 1.468 & 0.984 & 1.362 & 1.530 + @xmath143 & @xmath150 & @xmath151 & 0.005 & @xmath152 & @xmath153 & @xmath154 + @xmath148 ( % ) & 0.322 & 0.869 & 0.879 & 0.974 & 1.000 & 1.014 [ tb : fisher ] unlike for baos , which are insensitive to the parameters that affect the overall shape , for the full modeling we need to make sure that we take into account potential degeneracy between @xmath0 and @xmath1 and any other parameters that affect the overall shape . in this paper
we include @xmath149 , @xmath98 , @xmath155 , @xmath47 , and @xmath130 .
( we shall comment on the effects of non - linear bias in [ sec : caveat ] ) .
we study the effects of marginalization over various parameter combinations by taking the submatrix , @xmath156 , of the full @xmath157 matrix with the index , @xmath158 , of @xmath140 running from 1 to 7 , such that the submatrix includes the matrix components of desired parameters to be marginalized .
in other words , the parameters that are not included in the submatrix are fixed and not marginalized over .
then , we compute the marginalized errors in @xmath159 and @xmath160 as @xmath161 to simplify the analysis , we fix all the other cosmological parameters , such as @xmath162 , @xmath163 , etc .
these cosmological parameters will be determined by the future cmb mission , planck , accurately , and therefore it is a good approximation to simply fix them , and vary only @xmath159 , @xmath160 , @xmath48 , @xmath98 , @xmath129 , @xmath47 and @xmath130 .
the fiducial value for the bias is set to @xmath103 and @xmath164 is computed from the fiducial cosmological model . for the following reason : while @xmath96 can be calculated from the cosmological parameters assuming the validity of general relativity , one might choose to let @xmath96 free and use it for testing the validity of general relativity .
in this paper we chose to assume the validity of general relativity , but one can extend our analysis to let @xmath96 free in a straightforward manner .
] we expect that the analysis of the bispectrum ( fourier transform of three - point function ) will give a precise determination of @xmath81 ( as well as non - linear bias parameters such as @xmath165 ) @xcite , and therefore it is also a good approximation to simply fix it .
however , we also explore a more conservative case where we do not know what @xmath81 is , i.e. , we marginalize over the overall amplitude as well as @xmath98 . in the future work
we also plan to investigate the effect of marginalization over @xmath165 , using a joint analysis of the power spectrum and bispectrum .
therefore , our calculation presented here will provide the lower limit to the errors in @xmath159 and @xmath160 expected from the full modeling of the power spectrum measured in a survey like hetdex .
we use the same survey parameters that we have used in [ sec : err_fitex2d ] , and we integrate eq .
( [ eq : fisher_fm ] ) up to @xmath116 . in figure [ fig3 ]
we show the resulting error ellipses from the full modeling , in the smaller , solid contours , with four choices of marginalization .
( we present the results from more choices of marginalization in table [ tb : fisher ] . ) first , for all choices of marginalization we find that the sizes of the errors in both @xmath0 and @xmath1 are substantially smaller than those from the bao - only analysis with the fitex-2d . for example , determinations of both @xmath0 and @xmath1 are improved by more than a factor of two in the case of the amplitude marginalization .
this is expected , as we are able to use more information encoded in the power spectrum ; namely , the hubble horizon at the matter - radiation equality epoch and the silk damping scale .
second , @xmath0 and @xmath1 are _ anti - correlated _ for the amplitude marginalization , with the cross - correlation coefficient of @xmath166 ( see top - left panel of fig . [ fig3 ] ) , as opposed to a positive correlation seen in the bao - only analysis .
this is due to the marginalization over the overall amplitude : if we fixed the overall normalization , then we would still find a positive correlation between @xmath0 and @xmath1 with @xmath167 .
the origin of the negative correlation is the so - called alcock - paczynski ( ap ) test @xcite : when the redshift space distortion is known perfectly well , the departure of the power spectrum in redshift space from isotropy , i.e. , dependence of @xmath38 on @xmath168 , can be used to determine @xmath39 , resulting in @xmath169 for a power - law power spectrum .
the contributions from departures of @xmath8 from a pure power - law , i.e. , the existence of `` standard rulers , '' such as baos , the hubble horizon at the matter - radiation equality and the silk damping scale , make @xmath170 bigger than @xmath171 .
( see appendix [ sec : cc ] for more details . ) when @xmath149 and @xmath98 are marginalized over simultaneously , the correlation between @xmath0 and @xmath1 nearly disappears : the ap test no longer works when we marginalize over the linear redshift space distortion .
we find @xmath172 ( see top - right panel of fig . [ fig3 ] ) for this case .
when @xmath149 is marginalized over while the other parameters ( @xmath98 , @xmath173 , @xmath47 , and @xmath130 ) are held fixed , we find 0.88% and 0.79% errors on @xmath0 and @xmath1 , respectively , with @xmath166 .
the more parameters we marginalize over , the greater the cross - correlation coefficient between @xmath0 and @xmath1 as well as the errors on @xmath0 and @xmath1 become . note that the increase in the errors does not necessarily imply the decrease in the statistical power in constraining dark energy properties : since the cross - correlation coefficient is also reduced , the error in the combined 1-d distance scale , @xmath174 , is much less affected by the marginalization ( see table [ tb : fisher ] ) .
the error in @xmath175 has been computed as @xcite : @xmath176 finally , the errors in @xmath0 , @xmath1 , and @xmath174 for various choices of marginalization are : ( @xmath177 ) = ( 0.88% , 0.79% , 0.76% ) , ( 1.10% , 1.13% , 0.78% ) , ( 1.10% , 1.47% , 0.88% ) , and ( 1.25% , 1.53% , 1.01% ) for the marginalization over @xmath149 , @xmath149 and @xmath98 , @xmath149 , @xmath98 and @xmath129 , and @xmath149 , @xmath98 , @xmath129 , @xmath47 and @xmath130 , respectively ( see table [ tb : fisher ] for more comprehensive list ) .
this result should be compared with that from the bao - only analysis : ( @xmath177)=(1.76% , 2.47% , 1.08% ) .
it is clear that the full analysis , even with a generous set of marginalization choices , beats the bao - only analysis with a significant gain in the distance determination accuracies .
our analysis presented in [ sec : err_full ] is too simplistic and optimistic , as it ignores any systematic errors due to our lack of understanding of the effects of various non - linearities in the power spectrum . among the three major non - linearities ,
non - linear matter clustering is under control , at least for high redshifts , i.e. , @xmath178 , as one can model non - linear evolution of matter fluctuations almost exactly by the 3rd - order perturbation theory @xcite . while the nominal 3rd - order perturbation theory breaks down at lower redshifts , @xmath179 ,
there have been a number of studies aiming at improving upon our ability to compute @xmath8 at @xmath179 or even lower redshifts @xcite .
therefore , it is quite possible that the non - linear matter clustering will be fully under control in the near future , at least for the scales that are relevant to the bao scales , i.e. , @xmath180 . in a separate paper @xcite ,
we show that non - linear galaxy biasing is also under control in the weakly non - linear regime .
one can use the perturbation theory approach combined with the local bias assumption @xcite to model the galaxy power spectrum with non - linear bias .
the most problematic one is the non - linear redshift space distortion .
our understanding of non - linear redshift space distortion , especially the finger - of - god ( fog ) effect , is limited @xcite . therefore , whether one can achieve the accuracy of @xmath0 and @xmath1 ( @xmath1 in particular ) reported in fig .
[ fig3 ] depends crucially on our ability to correct for the fog effect .
this is work in progress .
note that the marginalization over @xmath173 should capture some of the increase in the errors in distance scales due to our ignorance of fog .
how realistic is our result for the determinations of @xmath0 and @xmath1 from the bao phases using the fitex-2d method ?
since our monte carlo simulations used in [ sec : err_fitex2d ] are too simple , in this section we test the fitex-2d method further by using the millennium simulation @xcite .
we use the millennium galaxy catalogue , generated by the semi - analytical galaxy formation code @xcite .
we have measured the two - dimensional power spectrum of galaxies in redshift space from the millennium simulation , and applied the fitex-2d method to remove the smooth component .
we then find the best - fitting @xmath0 and @xmath1 from the bao phases extracted from the fitex-2d .
again , we use the data up to @xmath116 . in figure [ fig5 ]
we show the result . the best - fitting value that we find from the millennium simulation corresponds to one point at the center of the contours .
we find the errors from the monte carlo simulations that we described in [ sec : err_fitex2d ] with the survey parameters replaced by those of the millennium simulation : @xmath181 , @xmath182 , and @xmath183 .
( there are 17,238,935 galaxies in the millennium simulation at @xmath183 . ) for the theoretical power spectrum that we use for generating monte carlo simulations , we use the best - fitting power spectrum for the galaxy catalogue of the millennium simulation found in @xcite .
since the volume of the millennium simulation is @xmath184 times as small as that would be surveyed by hetdex , the uncertainties in @xmath0 and @xmath1 are larger for the millennium simulation .
( compare fig .
[ fig5 ] with the larger contours of fig .
[ fig3 ] . )
we find 5.1% and 6.8% errors on @xmath0 and @xmath1 , respectively , with the cross - correlation coefficient of @xmath185 , from the monte carlo simulations .
these errors are larger than those from hetdex monte - carlo simulation by a factor of two ( rather than @xmath186 ) as the shot noise on the power spectrum of the millennium simulation is much smaller than that of hetdex simulation .
the best - fitting values of @xmath0 and @xmath1 are well within 68% c.l .
region , which indicates that the fitex-2d is able to yield unbiased estimates of the bao phases from the millennium simulation .
these results indicate that the fitex-2d method that we have developed in this paper can be used for extracting the baos and measuring @xmath0 and @xmath1 safely from the real data
. it would be interesting to apply the fitex-2d method to the two - dimensional power spectrum measured from the sdss lrg samples @xcite , and extract @xmath0 and @xmath1 from them .
in [ sec : montecarlo ] and [ sec : millennium ] , we have estimated errors in @xmath0 and @xmath1 from two different approaches , i.e. , the bao fitting using the fitex-2d method and the full modeling . in this section ,
we propagate errors in @xmath0 and @xmath1 to those in the dark energy equation of state parameters .
we parametrize @xmath187 using the linear model , @xmath40 @xcite .
we propagate the errors in @xmath0 and @xmath1 to those in @xmath3 and @xmath4 by @xmath188 where @xmath189 is the fisher matrix for the dark energy parameters , @xmath190 is the fisher matrix for @xmath0 and @xmath1 , @xmath191 for @xmath127 and 2 , and @xmath192 for @xmath193 and 2 .
partial derivatives of @xmath0 and @xmath1 with respect to @xmath3 and @xmath4 are given by @xmath194f(z^{\prime})g(z^{\prime})^{-3/2}dz^{\prime } \over\int_0^z g(z^{\prime})^{-1/2}dz^{\prime}},\\ { \partial \ln{h } \over \partial w_0}&=&{3\over2}\omega_\lambda \ln(1+z){f(z)\over g(z)},\\ { \partial \ln{h } \over \partial w_a}&=&{3\over2}\omega_\lambda \left[\ln(1+z)-\frac{z}{1+z}\right]{f(z)\over g(z)},\end{aligned}\ ] ] where @xmath162 and @xmath195 are given by @xmath196 figure [ fig6 ] shows the derivatives as a function of @xmath5 between @xmath197 in two different cosmological models , ( @xmath3 , @xmath4)=(@xmath198 , @xmath199 ) and ( @xmath200 , @xmath201 ) .
the former is the @xmath202cdm model , while the latter resembles the maximum likelihood values of @xmath3 and @xmath4 from the wmap+bao+sn+bbn @xcite .
the derivatives are similar for these cosmological models , and therefore we use the @xmath202cdm model as the fiducial model for computing the derivatives .
we add the distance information from cmb as @xmath203 where we assume that the cmb experiment yields 1% determination of the angular diameter distance out to @xmath204 , i.e. , we use @xmath205 we are interested in how the bao - only analysis compares with the full modeling . in fig .
[ fig7 ] we show the projected error contours on @xmath3 and @xmath4 calculated from the bao - only analysis with the fitex-2d and those from the full analysis at four redshift bins : @xmath206 , @xmath207 , @xmath208 , and @xmath209 .
the survey area and the number of galaxies are 420 @xmath210 and @xmath211 for all redshift bins . from the bao - only
analysis we find @xmath212 , @xmath213 , @xmath214 , and @xmath215 , whereas from the full modeling we find @xmath216 , @xmath217 , @xmath218 , and @xmath219 , for @xmath206 , @xmath207 , @xmath208 , and @xmath209 , respectively . we therefore conclude that the full analysis yields much better constraints on @xmath3 and @xmath4 than the bao - only analysis .
in this paper we have developed a method , called the fitex-2d method , to extract the two - dimensional phases of baos from galaxy power spectra in redshift space .
our model builds on and extends the existing one - dimensional algorithm , called fitex , developed by @xcite .
our method removes the smooth , non - oscillating component from the observed galaxy power spectrum in redshift space .
the fitting function consists of the smooth one - dimensional spectrum that depends only on @xmath70 , @xmath77 given by eq .
( [ eq : psmooth1d ] ) , multiplied by the angle - dependent function expanded in the legendre polynomials with even multipoles .
the coefficients of the legendre polynomials contain even powers of @xmath70 .
the resulting function , given by eq .
( [ eq : psmooth2d ] ) , is able to capture the non - oscillating part of the galaxy power spectrum well .
we have tested the fitex-2d method using the analytical model without any noise , the monte carlo realizations with noise expected from the hetdex experiment @xcite , and the galaxy catalogue created from the millennium simulation @xcite . in all cases
the fitex-2d method yields unbiased estimates of the angular diameter distance , @xmath0 , and the expansion rate , @xmath1 .
however , the baos capture only a part of distance information encoded in the galaxy power spectrum . to exploit the distance information , especially the equality scale , @xmath220 , we have explored the constraints on @xmath0 and @xmath1 from the full modeling of the galaxy power spectrum in redshift space .
provided that three key non - linearities ( non - linear matter clustering , non - linear galaxy bias , and non - linear redshift space distortion ) are under control , we find that the full modeling yields the constraints that are better than the bao - only analysis by more than a factor of two both in @xmath0 and @xmath1 , and the dark energy parameters such as @xmath3 and @xmath4 . while the effects of non - linear matter clustering @xcite and non - linear galaxy bias @xcite are being understood in the weakly non - linear regime that is relevant to the future galaxy surveys , the effects of non - linear redshift space distortion are poorly understood .
while the fitex-2d method that we have developed in this paper are useful for obtaining _ robust _ constraints on @xmath0 and @xmath1 , hence the dark energy properties , one must understand non - linear redshift space distortion to fully exploit the full information content of the galaxy power spectrum in redshift space
. we would then be able to reduce the errors in @xmath0 and @xmath1 by more than a factor of two .
we thank the anonymous referee for illuminating comments on the correlation coefficients , which motivated our doing more thorough analysis .
this material is based in part upon work supported by the texas advanced research program under grant no .
003658 - 0005 - 2006 .
ek acknowledges support from an alfred p. sloan research fellowship . the millennium simulation databases used in this paper and the web application providing online access to them were constructed as part of the activities of the german astrophysical virtual observatory .
in this appendix we describe what we have implemented in our fisher matrix code , which is publicly available as a part of `` cosmology routine library ( crl ) , '' developed by one of the authors ( ek ) .
this code includes the non - linear matter power spectrum in both real and redshift space , as well as marginalization over the amplitude , the linear redshift space distortion , the velocity dispersion of fingers - of - god , the primordial tilt and running index . in the future release we plan to include non - linear galaxy bias and primordial non - gaussianity .
a simple , approximate formula of the fisher matrix for galaxy survey is given by ( e.g. , * ? ? ?
* ) @xmath221 where @xmath124 is the galaxy survey power spectrum calculated theoretically as a function of parameters , @xmath140 are the parameters to be extracted from the data , and @xmath132 is a function given by @xmath133 v_{survey } \label{eq : weight}\ ] ] here , @xmath108 and @xmath109 are the number density of galaxies and the volume of survey , respectively . in linear theory , @xmath124 is given by @xmath222 where @xmath81 is the scale independent linear bias factor , @xmath223 is the linear matter power spectrum , and @xmath224 describes the linear redshift space distortion effect ( kaiser effect ) : @xmath225 where @xmath84 is the growth factor of the linear density fluctuations , and @xmath226 is the scale factor . to calculate the logarithmic derivatives of @xmath8 in eq .
( [ eq : fisher ] ) , let us write down the non - linear galaxy power spectrum ( with linear bias ) as ( eq . ( [ eq : pkred ] ) ) : @xmath227 \times\frac1{1+f^2k^2\mu^2\tilde{\sigma}_v^2}.\end{aligned}\ ] ] we compute the derivatives with respect to the following seven parameters : the angular diameter distance , @xmath0 , the hubble expansion rate , @xmath1 , the overall amplitude of the galaxy power spectrum , @xmath48 , the linear redshift space distortion factor , @xmath80 , the velocity dispersion with an empirically calibrated fudge factor , @xmath129 , the tilt of the primordial power spectrum , @xmath47 , and the running index , @xmath130 ( @xmath228}$ ] ) .
we choose the convention such that @xmath229 we compute the derivatives with respect to @xmath159 and @xmath160 in a two step process .
first , we write @xmath231 where @xmath232 finally , we need to know the `` effective spectral index '' , @xmath233 , given by @xmath234 or explicitly @xmath235 where @xmath236 in this subsection we explore the behaviour of the cross - correlation coefficient between @xmath0 and @xmath1 in various cases .
in particular we focus on the effect of the marginalization over the overall amplitude with ( [ sec_cc_beta ] ) and without ( [ sec : cc_no_beta ] ) the additional marginalization over the redshift space distortion .
let us evaluate the fisher matrices in the limit that the redshift space distortion is absent , i.e. , @xmath237 . in this limit , the weighting function in eq .
( [ eq : fisher ] ) and the effective spectral index in eq .
( [ eq : n_eff ] ) become independent of @xmath56 , i.e. , @xmath238 and @xmath239 .
we obtain @xmath240 ^ 2w(k ) \int^1_0d\mu ( 1-\mu^2)^2\\ f_{12}&=&\int^{k_{max}}_{k_{min}}\frac{k^2dk}{2\pi^2}[n_{eff}(k)]^2w(k ) \int^1_0d\mu ( 1-\mu^2)(-\mu^2)\\ f_{13}&=&\int^{k_{max}}_{k_{min}}\frac{k^2dk}{2\pi^2}n_{eff}(k)w(k ) \int^1_0d\mu ( 1-\mu^2)\\ f_{22}&=&\int^{k_{max}}_{k_{min}}\frac{k^2dk}{2\pi^2}[n_{eff}(k)]^2w(k ) \int^1_0d\mu ( -\mu^2)^2\\ f_{23}&=&\int^{k_{max}}_{k_{min}}\frac{k^2dk}{2\pi^2}n_{eff}(k)w(k ) \int^1_0d\mu ( -\mu^2)\\ f_{33}&=&\int^{k_{max}}_{k_{min}}\frac{k^2dk}{2\pi^2}w(k ) \int^1_0d\mu\end{aligned}\ ] ] now , in order to understand the effect of the structure of @xmath241 , let us assume that the galaxy power spectrum is a pure power - law , i.e. , @xmath242 and @xmath243 is the independent of @xmath70 . in this limit , we obtain the marginalized errors of parameters and the correlation coefficients are computed from the inverse of the fisher matrix .
however , one can show that the matrix given in eq .
( [ eq:3by3fisher ] ) is singular . in other words ,
@xmath0 and @xmath1 are completely degenerate with the amplitude for a power - law power spectrum .
this result shows that only the departure of the power spectrum from a pure power - law , i.e. , the existence of characteristic scales , can break the degeneracy between @xmath0 and @xmath1 , and @xmath48 .
these scales are often called the `` standard rulers . '' to understand the structure of the fisher matrix in eq .
( [ eq:3by3fisher ] ) better , let us add small perturbations , @xmath246 , to the diagonal elements , and invert the matrix .
the result is @xmath247 we find that the correlation coefficient between @xmath0 and @xmath1 is @xmath248 as @xmath249 .
therefore , @xmath159 and @xmath160 are totally anti - correlated , which implies that , although we can not determine @xmath159 and @xmath160 simultaneously , we can determine @xmath250 , even for a power - law power spectrum . and
@xmath251 for @xmath252 and @xmath253 respectively .
] this is known as the alcock - paczyski ( ap ) test @xcite .
there is a special case in which the covariance between @xmath48 and @xmath0 or @xmath1 may be ignored .
one may imagine the situation where @xmath254 depends upon @xmath70 such that @xmath48 is uncorrelated with @xmath0 or @xmath1 .
for example , if @xmath254 oscillates about zero , then @xmath255 would be small compared with @xmath256 ^ 2w(k)$ ] or @xmath257 .
therefore , @xmath258 and @xmath259 may be ignored , making @xmath48 de - correlated with @xmath0 and @xmath1 . in this case , the fisher matrix is a 2-by-2 matrix given by @xmath260 the inverse of this matrix is then @xmath261 the correlation coefficient between @xmath0 and @xmath1 is thus given by @xmath262 this result has been derived by @xcite , and justifies the use of baos as a way to measure @xmath0 and @xmath1 with a correlation coefficient of @xmath263 . * when the information is dominated by baos , the correlation coefficient between @xmath0 and @xmath1 is @xmath264 .
the amplitude of the baos contributes little to the errors on @xmath0 and @xmath1 , as the amplitude information is de - correlated with @xmath0 and @xmath1 .
* when the information is dominated by the ap test , @xmath265 .
* in reality , as we have shown in this paper , baos contribute less than the overall shape of the power spectrum . also , the shape of the power spectrum is not exactly a power - law . as a result ,
the correlation coefficient from the full analysis is usually negative ( or small positive ) , but always greater than @xmath171 ( see table [ tb : fisher ] ) .
next , let us consider the case where the redshift space distortion can not be ignored . in this case , we see from eq .
( [ eq : weight ] ) and eq .
( [ eq : n_eff_exp ] ) that the weighting function , @xmath132 , and the effective spectral index , @xmath233 , are no longer independent of @xmath56 .
the analytical treatment is also possible for this case , although the results are too complicated to be useful .
we therefore report on the numerical results . here , we choose the survey parameters given in
[ sec : err_full ] with the non - linear power spectrum of eq .
( [ eq : pkred ] ) .
the results from the numerical calculations of the fisher matrix are given in table [ tb : fisher ] .
we find that the marginalization over the amplitude information , and that over the amplitude and the shape of the primordial power spectrum ( i.e. , @xmath47 and @xmath130 ) give the cross - correlation close to @xmath171 ; thus , one relies on the ap test .
the marginalization over the amplitude and the linear redshift space distortion ( i.e. , @xmath98 ) drive the cross - correlation towards zero , as the ap test no longer works when the linear redshift space distortion is marginalized over
. however , in both cases the errors in the combined 1-d distance scale , @xmath174 , are about the same . in other words , while one changes the orientation of the ellipse , the area is approximately preserved . in summary , when the amplitude information is marginalized over , the information is mostly coming from the dependence of @xmath38 on @xmath168 , which yields a constraint on @xmath39 via the ap test , while when both the amplitude and the linear redshift space distortion are marginalized , the most information is coming from the standard rulers , which can constrain @xmath0 and @xmath1 separately , driving the cross - correlation towards zero . finally , in fig .
[ fig8 ] we show how different choices of marginalization over parameters influence the error contours of @xmath3 and @xmath4 : @xmath267 , @xmath268 , @xmath269 , and @xmath270 , for the cases of no marginalization , marginalization over @xmath149 , marginalization over @xmath149 , @xmath98 and @xmath129 , and marginalization over @xmath149 , @xmath98 , @xmath129 , @xmath47 and @xmath130 respectively . * linear power spectrum with the linear redshift space distortion ( kaiser effect ) , * non - linear power spectrum from the 3rd - order perturbation theory with the linear redshift space distortion , * non - linear power spectrum from the 3rd - order perturbation theory with the non - linear redshift space distortion given by eq .
( 71 ) of @xcite , or * non - linear power spectrum from the 3rd - order perturbation theory with the non - linear redshift space distortion given by eq .
( [ eq : pkred ] ) .
a given galaxy survey can be sliced up into multiple redshift bins . after entering the survey area in units of @xmath271 ,
one is asked to enter the following parameters at each redshift bin : the redshift range ( @xmath272 ) , the number of galaxies in units of millions in the bin , @xmath81 , @xmath273 in units of @xmath274 , and the redshift error in units of km / s .
the linear power spectrum at @xmath275 has been precomputed using the camb code @xcite for the maximum likelihood parameters given in table 1 of @xcite ( wmap+bao+sn ) .
the ingredients of the non - linear power spectra , @xmath276 , @xmath277 , and @xmath278 , have been precomputed from the linear spectrum at @xmath275 .
these spectra are then evolved to a specified redshift by the appropriate growth factor obtained by solving the differential equation given in eq .
( 76 ) of @xcite . | we present a method for extracting the angular diameter distances , @xmath0 , and the expansion rates , @xmath1 , of the universe from the _ two - dimensional _ baryon acoustic oscillations ( bao ) in the galaxy power spectrum .
our method builds upon the existing algorithm called the `` fit - and - extract '' ( fitex ) method , which allows one to extract only @xmath2 from a spherically averaged one - dimensional power spectrum .
we develop the fitex-2d method , an extension of the fitex method , to include the two - dimensional information , which allows us to extract @xmath0 and @xmath1 simultaneously .
we test the fitex-2d method using the millennium simulation as well as simplified monte carlo simulations with a bigger volume .
the baos , however , contain only a limited amount of information .
we show that the full modeling , including the overall shape of the power spectrum , yields much better determinations of @xmath0 and @xmath1 , hence the dark energy equation of state parameters such as @xmath3 and @xmath4 , than the bao - only analysis by more than a factor of two , provided that non - linear effects are under control . |
fully packed loop configurations ( fpls ) are configurations on a square grid to which a certain link pattern @xmath1 is attached ( see section [ sub : fpl ] for definitions ) . as was first shown in @xcite , fpl configurations in a triangle ( tfpls ) occur naturally when studying certain properties of the usual fpls .
the connection is explained in more detail in @xcite , where in several properties of tfpls are also shown .
the study of tfpls was initially motivated by the desire to obtain an expression for the number @xmath2 of fpls with link pattern @xmath1 ; also , as conjectured in @xcite and proved in @xcite , there exist linear relations between fpls whose coefficients are defined in terms of tfpls .
these relations are in fact recurrence formulas allowing to compute all @xmath3 .
the famous razumov - stroganov correspondence @xcite gives rise to a different set of relations characterizing the @xmath2 ; they concern fpls all on the same grid size , while the recurrence formulas express fpls of size @xmath4 in terms of fpls of size @xmath5 . boundaries of tfpls can be encoded by dyck words @xmath6 , to which partitions @xmath7 are attached ; our main result ( theorem [ th : main ] ) is : _ if @xmath8 , then the number of tfpl configurations with boundary @xmath6 is given by the littlewood - richardson coefficient @xmath9 .
_ relevant definitions can be found in section [ sec : prelim ] .
it is quite surprising to come across such coefficients when starting with a purely enumerative problem : objects counted by littlewood
richardson coefficients do not seem to appear `` by chance '' , as they are usually designed to give a combinatorial interpretation of these numbers .
such is the case of knutson
tao puzzles , which are the objects we shall put into one - to - one correspondence with our tfpl configurations .
we will see that although the map itself is quite easy to define , proving that it is indeed a bijection requires some work .
let us give a brief outline : in section [ sec : prelim ] we define fpls both on a grid and in a triangle , and state theorem [ th : main ] which is our main result .
we also explain briefly the connection between fpls and tfpls , which gives rise to formula .
section [ sec : ingredients ] introduces three major ingredients in the proof of the main theorem , which is itself proved in section [ sec : bij ] .
_ note : this work was first presented at the fpsac 2010 conference in san francisco @xcite .
we fix a positive integer @xmath4 , and let @xmath10 be the square grid with @xmath11 vertices .
we impose _ periodic boundary conditions _ on @xmath10 , which means that we select every other external edge on the grid , starting by convention with the topmost on the left side ; we number these @xmath12 external edges counterclockwise .
a _ fully packed loop ( fpl ) _
configuration @xmath13 of size @xmath4 is a subgraph of @xmath10 such that each vertex of @xmath10 is incident to two edges of @xmath13 .
an example of fpl configuration is given on figure [ fig : fplexample ] ( left ) .
, scaledwidth=80.0% ] define a _
link pattern _ @xmath1 of size @xmath4 as a partition of @xmath14 in @xmath4 _ pairwise noncrossing _
pairs @xmath15 , which means that there are no integers @xmath16 such that @xmath17 and @xmath18 are both in @xmath1 .
a fpl configuration @xmath13 on @xmath10 naturally defines non - crossing paths between its external edges , so we can define the link pattern @xmath19 as the set of pairs @xmath15 where @xmath20 label external edges which are the extremities of the same path in @xmath13 .
this is illustrated on the right of figure [ fig : fplexample ] .
if @xmath1 is a link pattern , we denote by @xmath3 the number of @xmath21 configurations @xmath13 of size @xmath4 such that @xmath22 .
a link pattern @xmath1 on @xmath14 can be encoded by a binary word of length @xmath12 , where for each pair @xmath23 in @xmath1 we set @xmath24 and @xmath25 .
such words @xmath1 form the following subset of @xmath26 : we denote by @xmath27 the set of words @xmath28 of length @xmath12 , such that @xmath29 , and each prefix @xmath30 of @xmath28 verifies @xmath31 .
these are known as dyck words , and are counted by the catalan numbers @xmath32 .
note also that if @xmath33 , then it contains a factor @xmath34 , and the removal of such a factor leaves a shorter dyck word in @xmath35
. we will identify link patterns with words in @xmath36 .
there is also a bijection @xmath37 from @xmath36 to the set of ferrers diagrams included in the staircase shape @xmath38 : the letters of @xmath28 encode the southeast boundary of the ferrers shape from bottom to top , cf .
figure [ fig : dn ] . as a link pattern and a ferrers diagram .
[ fig : dn],scaledwidth=80.0% ] the _ degree _ @xmath39 of @xmath28 is the number of indices @xmath40 such that @xmath41 , and is equal to the number of boxes @xmath42
. for instance we have @xmath43 for the example of figure [ fig : dn ] .
the _ conjugate _ @xmath44 of @xmath45 is the word of length @xmath4 defined by @xmath46 ; the diagram @xmath47 is obtained from @xmath48 by a reflection about the main diagonal of @xmath49 .
we define the partial order @xmath50 if @xmath51 in the diagram representation . a _ semistandard young tableau _ of shape @xmath52 is a filling of the boxes of @xmath52 by positive integers , in which entries are nondecreasing across each row from left to right and increasing down each column .
we denote by @xmath53 the number of such tableaux where entries are in @xmath54 .
let @xmath30 be a box in @xmath52 which is in the @xmath55th row from the top and @xmath56th column from the left .
the _ content _ @xmath57 of @xmath30 is defined as @xmath58 ; its _ hook - length _
@xmath59 is the number of boxes in @xmath52 which are below @xmath30 and in the same column , or right of @xmath30 and in the same row ( @xmath30 itself being counted just once ) .
define then @xmath60 where the product is over all cells @xmath30 of the diagram @xmath52 .
we have then the _ hook content formula _ : @xmath61 so that @xmath53 is given by a polynomial in @xmath62 with leading term @xmath63 . for the diagram in figure [ fig : dn ]
we get @xmath64 .
we define the triangle @xmath65 as the set of points @xmath66 which verify @xmath67 and @xmath68 .
we also include the following edges in the definition of @xmath69 : @xmath12 vertical edges @xmath70 below the vertices @xmath71 for @xmath72 ; @xmath73 horizontal edges between @xmath74 and @xmath75 , as well as between @xmath76 and @xmath77 , for @xmath78 .
these edges are in bold on figure [ fig : tfpl ] , left .
a vertex of @xmath69 is _ even _ or _ odd _ depending on the parity of its sum of coordinates : we represent even vertices by filled squares and odd ones by squares with a white interior .
we also call _ inner vertices _ the vertices of @xmath69 which are not of the form @xmath79 ( _ left vertices _ ) or @xmath80 ( _ right vertices _ ) .
we now impose extra conditions given by @xmath81 words in @xmath36 . if @xmath82 , we add a vertical edge below @xmath79 for each @xmath83 such that @xmath84 , and forbid such an edge if @xmath85 .
if @xmath86 , we add a vertical edge below @xmath80 for each @xmath83 such that @xmath87 , and forbid such an edge if @xmath88 .
let @xmath89 be the corresponding triangle . and an example of tfpl .
[ fig : tfpl ] ] [ defi : tfpls ] a _ fpl configuration @xmath90 in a triangle ( tfpl ) _ with boundary conditions @xmath6 in @xmath36 is a graph on @xmath89 , such that all inner vertices are imposed to be of degree @xmath91 , and furthermore ( i ) for each pair @xmath15 in @xmath1 , the edges @xmath92 and @xmath93 are linked by a path in @xmath69 , and ( ii ) the paths starting from a left ( _ resp . _
right ) vertex must end at a right ( _ resp . _ left ) vertex .
an example is shown on figure [ fig : tfpl ] , right .
the set of these tfpls is denoted by @xmath94 , and we let @xmath95 be its cardinality .
we have clearly a _ left - right symmetry _ in tfpls : the reflection of a tfpl configuration about a vertical axis is a tfpl configuration .
if @xmath90 belongs to @xmath94 , then its reflection belongs to @xmath96 . note that because of condition ( ii ) there are three kinds of paths in a tfpl : _ left - right paths _ which join a left vertex to a right vertex , _ bottom paths _ which link two edges @xmath92 and @xmath93 , and _ closed paths_. the main result of this paper is the following enumeration , in which @xmath9 is a littlewood richardson coefficient , cf .
section [ sub : puzzles ] : [ th : main ] suppose @xmath97 verify @xmath98 .
then we have @xmath99 we will prove this in section [ sec : bij ] ; in section [ sec : ingredients ] we give the key tools of this proof .
let us note that theorem [ th : main ] was independently conjectured in ( * ? ? ?
* lemma 2 and following remark ) , as a consequence of the main conjecture of that paper which is a formula for @xmath95 _ for any @xmath97_. given a link pattern @xmath1 on @xmath14 , and an integer @xmath100 , define @xmath101 as the link pattern on @xmath102 given by the nested pairs @xmath103 for @xmath104 , and the pairs @xmath105 for each @xmath106 . as words ,
note that this is simply @xmath107 .
we introduce the notation @xmath108 , so that in particular @xmath109 . the structure of tfpls
arises naturally when considering fpl configurations enumerated by @xmath110 for @xmath111 big enough , see @xcite .
it was shown in @xcite that we have the following formula : for @xmath112 integers , @xmath113 using this expression and certain properties of tfpls , one can show in particular the following result , which was conjectured in @xcite : [ th : ckln ] @xmath110 is a polynomial in @xmath111 with leading term @xmath114 . from equations and
, we see that an expression for the numbers @xmath95 will lead to a formula for @xmath110 .
what theorem [ th : main ] shows is that the numbers @xmath95 are also interesting on their own .
in section [ sub : otfpl ] we introduce certain variations on tfpl configurations called _ oriented _ tfpl configurations ; whereas the former use connectivity conditions , which are _ global _ , the latter only necessitate _ local _ conditions which make them easier to manipulate .
then we introduce knutson tao puzzles in section [ sub : puzzles ] , which are combinatorial objects counted by littlewood
richardson numbers ; the proof of theorem [ th : main ] will consist of a bijection between these puzzles and tfpls . the correctness of this bijection will make use in an essential way of the identity proved in section [ sub : summation ] .
we define @xmath115 as the graph @xmath89 with all edges on the left side oriented up and right , and all edges on the right side oriented down and right ; see example on figure [ fig : otfpls ] , left .
[ defi : otfpls ] an _ oriented tfpl configuration @xmath90 _ with boundary conditions @xmath6 in @xmath36 is a directed graph on @xmath115 such that all inner vertices have exactly one incoming edge and one outgoing edge , and in which the edge @xmath92 is directed upwards ( _ resp . _
downwards ) if @xmath24 ( _ resp . _
@xmath116 ) . and an example of oriented tfpl .
[ fig : otfpls ] ] the set of oriented tfpls is denoted by @xmath117 and its cardinality @xmath118 . like tfpls
, oriented tfpls also possess a _ left - right symmetry _ : by reflecting the underlying non - oriented graph of an oriented tfpl about a vertical axis , and inverting the orientation of up and down edges while keeping left and right orientations , one obtains another oriented tfpl configuration .
this is easily seen to be an involution exchanging @xmath117 and @xmath119 . [
prop : otfpls_to_tfpls ] let us be given an oriented tfpl configuration @xmath120 .
a. its underlying ( non - oriented ) graph is a tfpl configuration @xmath90.[ita ] b. the directed edges induce a global orientation on each path of @xmath90 , where all left - right paths are oriented from left to right .
[ itb ] c. the tfpl @xmath90 belongs to @xmath121 for a certain link pattern @xmath122 , and @xmath123 if and only all bottom paths are oriented from left to right.[itc ] * proof : * given an oriented configuration , it is clear that the underlying graph is such that each inner vertex has degree @xmath91 , and the boundary conditions are the same as those of tfpls .
it is also clear that the directed edges induce a global orientation on paths , and that indeed left - right paths oriented from left to right by the definition of @xmath115 , so is clear . to prove
, it remains to prove condition ( ii ) in definition [ defi : tfpls ] .
we only need to show that paths starting from a left vertex necessarily end at a right vertex , the converse being then true by the left - right symmetry of oriented tfpls .
consider a path @xmath124 starting at a left vertex .
it can not end at another such vertex because of the orientation of @xmath115 .
suppose @xmath124 ends at an edge @xmath92 , so that by definition [ defi : otfpls ] one has @xmath116 .
since @xmath1 is a dyck path , its prefix @xmath125 contains strictly more @xmath126s than @xmath127s , so that there are more up - edges than down - edges among the @xmath128 first bottom edges @xmath129 .
so there is a path starting in @xmath129 that does not end there .
this is a contradiction , since this path can not end at a left vertex ( the orientations do nt match ) , or anywhere else on the boundary since it would then have to cross the path @xmath124 .
therefore all paths starting from a left vertex must end at a right vertex , and as argued above this proves .
so @xmath90 is a tfpl , and , since @xmath115 is simply a certain orientation of @xmath89 , we know that @xmath90 belongs in fact to @xmath121 for a certain @xmath130 .
assume @xmath90 verifies the link pattern @xmath1 , so that for any pair @xmath23 in @xmath1 , @xmath90 contains a path between @xmath92 and @xmath93 .
now , considering @xmath1 as a word one has @xmath24 and @xmath25 and therefore @xmath92 is an up - edge and @xmath93 is a down - edge in @xmath13 , so that the path between them is oriented from left to right .
conversely assume all bottom paths in @xmath13 are oriented from left to right , and consider a pair @xmath131 which exists since @xmath132 .
this means that @xmath13 contains a path from @xmath92 going right , and a path arriving to @xmath133 from the left .
since paths are non - crossing , this must be the same path , and thus @xmath90 respects the pair @xmath134 of the link pattern @xmath1 .
consider now the dyck word @xmath122 , obtained by removing @xmath135 from @xmath1 , and the bottom edges @xmath136 ; we can then use the same argument , and by immediate induction we find that bottom paths in @xmath90 verify the pattern @xmath1 , proving .
@xmath137 we refer to @xcite for background on symmetric functions .
let @xmath138 be commuting variables , and let @xmath139 be the algebra of symmetric functions in @xmath140 . for @xmath52 a ferrers diagram ,
the schur function @xmath141 can be defined by @xmath142 where @xmath143 runs through all semistandard tableaux of shape @xmath52 , and @xmath144 is the number of entries equal to @xmath83 in the tableau @xmath143 .
the schur functions form a linear basis of @xmath139 , and the _ littlewood richardson ( lr ) coefficients _ @xmath145 are the corresponding structure constants defined by @xmath146 the lr coefficient @xmath145 is @xmath126 unless @xmath147 and @xmath148 ; also , they are known to be nonnegative integers by character theory @xcite .
many combinatorial descriptions of them are also known , the most famous being the original littlewood
richardson rule @xcite .
we will here use knutson tao puzzles @xcite : let @xmath4 be an integer , and @xmath6 words in @xmath36 . consider a triangle with edge size @xmath12 on the regular triangular lattice , where unit edges on left , bottom and right side are labeled by @xmath6 respectively when read from left to right .
a _ knutson tao ( kt ) puzzle _ with boundary @xmath6 is a labeling of each internal edge of the triangle with @xmath126,@xmath127 or @xmath91 , such that the labeling induced on each of the @xmath149 unit triangles is composed either of three @xmath126s , or of three @xmath127s , or of @xmath150 in counterclockwise order . ,
scaledwidth=90.0% ] we let @xmath151 be the set of knutson - tao puzzles with boundary @xmath6 . the list of all ten authorized labelings of unit triangles is given on the left of figure [ fig : puzzle ] , while on the right appears a puzzle with boundaries @xmath152 , @xmath153 , @xmath154 .
kt puzzles give a combinatorial interpretation for lr coefficients : [ th : ktlr ] given @xmath6 in @xmath36 , the number of kt puzzles with boundary @xmath6 is given by @xmath9 .
in particular , there are no such puzzles unless @xmath155 and @xmath98 .
the goal of this section is to prove the following formula which holds for any @xmath132 : @xmath156 this will be a consequence of the two formulas and . [ prop : thapperdegree ] one has @xmath157 unless @xmath158 .
furthermore , for every @xmath132 we have @xmath159 we use an argument of ( * ? ? ?
* lemma 3.7 ) , in which the first part of the proposition is proved .
assume @xmath111 is an even integer ; equation with @xmath160 becomes @xmath161 this is a polynomial identity in @xmath111 . by theorem [ th : ckln ]
, @xmath110 is a polynomial with leading term @xmath162 .
therefore the coefficients of degree @xmath163 must vanish on the r.h.s .
, which implies the first part of the proposition .
since a polynomial @xmath164 has leading term @xmath165 , the second part follows by taking the coefficient of degree @xmath166 on both sides of . from equation
, we have that @xmath141 specializes to @xmath53 under the substitutions @xmath167 for @xmath168 , and @xmath169 otherwise .
now if one considers the schur function @xmath170 in the variables @xmath171 , then we have @xmath172 ( see @xcite ) . by specializing at @xmath173 for @xmath104 and @xmath174 for @xmath175
, we thus get the following polynomial identity in @xmath111 : @xmath176 which in top degree becomes : @xmath177 now we use with @xmath178 for @xmath132 . since @xmath179 unless @xmath180 and @xmath148 , we can replace @xmath181 in with @xmath182 and sum over words @xmath183
. we then subtract it from : after removing the common factor @xmath184 , we obtain .
given three words @xmath185 verifying @xmath98 , we will define here a bijection between kt puzzles and tfpl configurations with the same boundary conditions @xmath6 . by theorem
[ th : ktlr ] , this will then give a bijective proof of theorem [ th : main ] .
given three words @xmath185 verifying @xmath98 , let @xmath186 be a kt puzzle with boundary @xmath6 . by definition each labeled unit triangle in @xmath186 must be labeled as one of the ten possibilities in figure [ fig : puzzle ] .
now to each of these triangles we apply the transformations described in figure [ fig : rulesoriented ] , and delete the original puzzle .
some cosmetic modifications are necessary : rescale the resulting configuration vertically by a factor @xmath187 ; remove the @xmath12 odd vertices coming from the left boundary of the puzzle , together with the possible horizontal edges to their right ; finally , double the length of the bottom vertical edges .
notice that edges @xmath188 and @xmath189 of the original unit triangles are now sent to odd and even vertices of @xmath190 respectively ; see figure [ fig : fpltopuzzle ] , right , for the end result
. looking at figure [ fig : rulesoriented ] , one sees that , when two triangles of a puzzle share a common horizontal edge , the oriented vertical half edges drawn inside these triangles are compatible .
indeed one notices that horizontal edges labeled @xmath191 , and @xmath91 are respectively associated to oriented edges pointing up , pointing down , and to no such edge .
therefore the transformations produce a certain directed graph on @xmath190 . the proof is given in the next section .
notice that the definition of @xmath195 does not need the introduction of @xmath196 , as one can perform the transformation rules of figure [ fig : rulesoriented ] while simply ignoring the arrows .
but as we will see , the orientation is essential in the demonstration of the theorem . to prove we check the conditions of definition [ defi : otfpls ]
let us first show that at each inner vertex of @xmath190 , the graph @xmath192 has one incoming edge and one outgoing edge .
for this one simply needs to check , on the local rules of figure [ fig : rulesoriented ] , how the transformation acts on diagonal edges of the triangles .
looking at the three possible labels @xmath191 , and @xmath91 and the two edges @xmath188,@xmath189 of triangles , one checks that in each of these six cases the corresponding inner vertex has indeed one incoming edge and one outgoing edge .
now we need to verify that @xmath6 are the boundary labels of the oriented graph configuration @xmath192 . in this case again , a direct inspection of the rules shows that this is indeed the case : it is immediate for @xmath1 and @xmath197 , and for @xmath28 one must remember that the vertices coming from the left side of @xmath186 must be deleted by the definition of @xmath196 . in conclusion @xmath192 is an oriented tfpl and is proved . to prove , it remains to verify verify that the bottom paths in @xmath193 follow the link pattern @xmath1 : this is indeed the case thanks to proposition [ prop : otfpls_to_tfpls ] and the second part of the following lemma : * proof of lemma [ lem : noloops ] : * suppose there is a path @xmath124 oriented from right to left in @xmath192 , and consider the largest @xmath198-coordinate @xmath199 visited by @xmath124 .
there is necessarily a horizontal edge @xmath200 on the line @xmath201 , and @xmath200 must be oriented to the left by the ( discrete ) jordan curve theorem .
now by observing the rules in figure [ fig : rulesoriented ] , the edge @xmath200 is necessarily connected either to two edges going down or to two edges going up , and both cases contradict the definition of @xmath199 , and therefore all bottom paths are oriented from left to right .
now suppose @xmath192 has a closed loop @xmath56 .
if @xmath56 is oriented counterclockwise , then the previous analysis leads to a contradiction .
if @xmath56 is oriented clockwise , we consider the minimal @xmath198-coordinate visited by @xmath56 ; there must be a left oriented edge at this height , which leads to a contradiction here also .
we now deal with : notice first that @xmath196 is clearly injective .
indeed , in each local transformation of figure [ fig : rulesoriented ] , the ten local configurations of oriented edges are distinct , and thus determine uniquely the labeling of the unit triangle .
to prove the injectivity of @xmath195 comes down to proving that `` erasing the arrows '' in a configuration @xmath192 is an injective operation . by lemma [ lem : noloops ] , this is indeed the case : orienting all bottom paths and left - right paths paths in @xmath193 from left to right gives back @xmath192 .
this achieves the proof of the proposition .
* proof of theorems [ th : main ] and [ th : bij ] : * let @xmath185 be such that @xmath98 . by proposition [ prop :
proof ] , @xmath195 is an injective mapping between @xmath151 and @xmath94 .
therefore , we have by theorem [ th : ktlr ] that @xmath202 now since @xmath203 , equation expresses that a sum of nonnegative terms is equal to zero
. therefore all terms in this sum must be equal zero , which proves that @xmath204 and achieves the proof of theorem [ th : main ] .
it also shows that @xmath195 is a injection between two sets with the same cardinality , therefore @xmath195 is a bijection , which is the content of theorem [ th : bij ] .
we describe first a nice consequence of the bijection @xmath195 .
let the _ canonical orientation _ of a tfpl @xmath90 be the oriented tfpl @xmath205 obtained by orienting all bottom paths and left - right paths of @xmath90 from left to right , and all its closed paths clockwise . if @xmath206 , then clearly @xmath207 .
say that a vertical ( oriented ) edge _ even _ if its lower vertex is even .
* proof : * it is a byproduct of the proof of theorem [ th : bij ] that , when @xmath98 , then @xmath205 has the form @xmath192 or a puzzle @xmath208 . by inspection of the local transformations of figure [ fig : rulesoriented ] applied to @xmath186 ,
no two consecutive left edges can appear , and no even edge either , which proves the direct implication .
consider now an element @xmath209 .
we reverse the passage between triangular grid and square grid given in [ sub : defi_phi ] : we add one horizontal edge to the left of each left - right path , divide by @xmath91 the length of the bottom edges @xmath92 , and rescale the tfpl vertically by a factor @xmath210 , so that one can superimpose / and @xmath189 edges of the triangular lattice to odd and even vertices respectively .
if @xmath90 avoids consecutive left edges and even vertical edges , then the fillings of the unit triangles belong to one of the ten possibilities of figure [ fig : rulesoriented ] , and the boundary words of @xmath90 correspond to the labels from the same figure .
therefore , there exists a puzzle @xmath211 such that @xmath212 .
but such puzzles exist only if @xmath213 by theorem [ th : ktlr ] , which completes the proof .
the preceding proposition thus gives a nice characterization of tfpls .
however the proof of this result is slightly unsatisfying from a combinatorialist s point of view , since one would like a direct proof of this fact by studying the structure of tfpl configurations . this would also give a direct proof of the bijectivity of @xmath195 , that is to say a proof which does not rely on identity .
in current work together with ilse fischer @xcite , we achieve this by studying oriented tfpl configurations . for a given @xmath214 and @xmath6 such that @xmath215 ,
we are able to characterize elements @xmath13 of @xmath117 in terms of the occurrence of @xmath55 specific local sub - configurations in @xmath13 . in the special case of @xmath216
, we obtain the result of proposition [ prop : csq_bij ] _ for all oriented tfpl configurations _ , and not only the canonically oriented ones .
it follows that all elements of @xmath117 are of the form @xmath193 , from which one deduces easily that all maps in the following commutative diagram are bijections ( the unlabeled map is simply the removal of the orientation ) : | in this work we continue our study of fully packed loop ( fpl ) configurations in a triangle .
these are certain subgraphs on a triangular subset of @xmath0 , which first arose in the study of the usual fpl configurations on a square grid .
we show that , in a special case , the enumeration of these fpls in a triangle is given by littlewood
richardson coefficients .
the proof consists of a bijection with knutson tao puzzles . |
after a long arduous effort , photoluminescence from silicon has been achieved from its nanocrystalline form @xcite .
a critical debate , however , continues over the nature of the interface chemistry of silicon nanocrystals ( si - ncs ) embedded in amorphous silica which has direct implications on the optical activity of the interface @xcite .
_ reported that the oxidation of porous silicon quantum dots results in a red shift in the photoluminescence ( pl ) spectra which indicates the importance of oxygen - related interface bond toplogy @xcite .
along this line , puzder and co - workers compared pl calculations of nanoclusters with different passivants and surface configurations and proposed the main reason for the red shift to be double si = o bonds @xcite . countering this , luppi _
et al . _ reported excitonic luminescence features caused by si - o - si bridge bonds at the surface of silicon nanoclusters @xcite .
as a supporting evidence for the latter , gatti _ et al .
_ recently demonstrated that si - o - si is the most stable isomer configuration @xcite .
to reconcile , vasiliev _ et al .
_ claimed that bridge bonds and double bonds have similar effect on pl @xcite .
all of the work cited above represent density functional theory ( dft)-based calculations with small si clusters of less than 100 atoms surrounded by either passivants like hydrogen @xcite or oxygen @xcite . but actual samples are profoundly different : the fabricated systems consist of si - nc with a diameter larger than @xmath0 nm , embedded in _ amorphous
_ silica ( a - sio@xmath1 ) matrix . identifying this
fact , tu and tersoff studied si / a - sio2 interface using a bond order dependent empirical potential and proposed that si - o - si bridges are the main blocks at the interface lowering the surface energy @xcite . using the same model potential hadjisavvas _
studied si - ncs embedded in a - sio@xmath1 and they also reported bridge bonds as the mechanism for lowering surface strain energy @xcite . in our previous work on the formation and structure of si ncs , we also observed that this model potential falls short to characterize the system accurately , especially the structure and chemistry at the interface @xcite . as a rigorous approach to interface structure and dynamics , pasquarello _ et al . _
used first - principles molecular dynamics ( md ) technique to investigate planar si / sio@xmath1 interface @xcite .
they observed that the oxygen atoms momentarily are bonded to three silicon atoms during oxidation process .
the net effect of these threefold coordinated oxygen atoms during oxidation process was to expel silicon atoms out of interface .
this is interpreted as a balancing process to decrease the increased density of si / sio@xmath1 interface due to oxidation @xcite .
however , no substantial breakthrough was made over the last decade on this issue .
in this work , we employ the reactive force field ( reaxff ) model developed by van duin _ et al . _ which improves brenner s reactive bond order model @xcite to a level of accuracy and validity allowing molecular dynamics simulations of the full reaction pathways in bulk @xcite .
the parameters for this force field were obtained from fitting to the results of ab initio calculations on relevant species as well as periodic boundary condition dft - based calculations of various crystalline polymorphs of relevant materials .
the reaxff calculates bond orders which is the measure of bond strength from local geometry .
this allows realistic chemical environment such as over / under coordination and bond breaking / formation for large - scale ( about 5000 atoms ) md simulations . to facilitate our discussion regarding the surface - bonded oxygen complexes , we distinguish among three different types of silicon atoms .
we label those silicon atoms with all silicon neighbors each with zero oxidation state as _ c _ , to denote core silicon atom . among the _ remaining _
( non-_c _ ) silicon atoms , those with at least one bond to _ c _ are labeled as _ s _ , denoting as a surface silicon atom . for further investigation of nc ,
we seperate core si atoms into two sub - categories as inner - core and outer - core atoms : among core si atoms which have at least one surface si neighbour categorized as outer - core si atoms and rest of core si atoms categorized as inner - core si atoms .
finally , any other silicon atom is labeled as _ m _ , denoting matrix silicon atom .
hence , a complex consisting of an oxygen atom bonded with two surface silicon atoms is labeled as _ ss_. the other oxygen complexes are _ sm _ , _ ssm _ , _ sss _ , _ mms _ as sketched in fig .
[ fig : abbrevations ] where the last three correspond to three - coordinated oxygen ( 3co ) atoms .
we use reaxff to represent the interactions in the model system .
we start with a large simulation cell ( box length @xmath2 ) of silica glass formed through a melting and quenching process used by one of us @xcite earlier to study silica glasses earlier .
next , similar to hadjisavvas _ et al . _
@xcite , we delete all atoms within a predetermined radius to insert crystalline silicon to form nc . in this way
, we create ncs with radii ranging from @xmath3 to @xmath4 . for the largest nc we insert @xmath5 si atoms into a spherical hole with radius @xmath4 created in amorphous matrix . even for this case , the minimum distance between nc surface to simulation box face is about @xmath6 which can still accommodate the interface layer .
we also pay special attention in the removal of spherical region so that the correct stoichiometry for the amorphous matrix is met .
thus , our amorphous matrix has two o atoms for every si atom with a density of @xmath7 g/@xmath8 which is the density of glass at room temperature and atmospheric pressure .
we set periodic boundary conditions in all directions and while keeping nc at 100 k , we employ simulated annealing process to sio@xmath1 region to end up with an amorphous matrix free of artificial strain around the nc .
then we set whole systems temperature to room temperature ( 300 k ) and continue performing md simulation for 75 ps to have thermal equilibrium between the two regions .
we set md simulation time step to 0.25 fs for all simulations .
for every 62.5 fs time interval , we have record the configurations to analyze the transitions taking place between different bond topologies .
the numbers attached to each arrow in fig . [ fig : abbrevations ] indicate the total number of registered transitions during the simulation in that direction between the complexes for a representative nc of radius @xmath9 .
almost balanced rates in opposite directions is an assurance of the attainment of the steady state in our simulation .
note that we do not observe any direct transition other than the paths indicated in fig .
[ fig : abbrevations ] . for instance , a direct transition of the complex _ ss _ to _ sm _ does not take place , but it is possible through an intermediate transition over the _ ssm _ which is a 3co .
we should also remark that the balanced transitions continue to take place after the steady state is attained which indicates that the interface bond topology is dynamics , i.e. not frozen .
. ] to analyze si - nc / a - sio@xmath1 interface , we construct nc surface using delaunay triangulation scheme ( fig . [
fig : rdfncsisi ] inset ) @xcite . in two dimensions
a delaunay triangulation of a set of points corresponds to defining triangles such that no point in the set is inside the circumcircle of any triangle . in three dimensions triangles extend to tetrahedra and
circumcircles become circumspheres .
since our ncs are nearly spherical in shape , we triangulate projection points of surface si atoms onto the unit sphere .
hence , we apply delaunay triangulation over the two dimensional @xmath10-@xmath11 plane . we can then create surfaces using this triangulation .
this surface enables us to calculate every atoms distance to nc surface . by this
means we plot various data such as charge , bond order etc . with respect to distance to the surface , to extract information about surface chemistry of si - nc embedded in amorphous matrix .
for different nc radii we observe similar trends in bond order distribution , average charges etc .
, therefore , we present only the figures of the system for a typical nc of radius @xmath9 .
a useful data to elucidate the structure of these systems is the radial distribution function ( rdf ) . in fig .
[ fig : rdfncsisi ] we present rdf of nc atoms only , where the first broad peak centered around 2.34 with a 0.3 full width at half maximum value ( fwhm ) , represents si - si bond length distribution in nc ( fig . [ fig : rdfncsisi ] inset ) .
the maximum extent of the nc can also be read from same plot at about 27 , where the rdf goes to zero .
observation of a broad first peak at fig [ fig : rdfncsisi ] demands further investigation of si - si rdf of nc atoms . for this purpose , in fig .
[ fig : detailedrdf ] we present bond length probability distributions ( akin to rdf ) for three categories of si atoms : inner core ( with no bonds to surface atoms ) , outer core ( bonded to surface ) and surface nc atoms .
we observe from fig .
[ fig : detailedrdf ] that si - si bond lengths in the inner core are centered around the equilibrium value and have a narrow width due mainly to thermal vibrations , whereas the bond length distributions of outer core and surface atoms have increasing shift for the most probable bond length values and broader widths .
these shifts and particularly the increase in distribution widths can not be attributed to thermal broadening .
taken together these two observations is a clear indication of increasing strain as a function of distance from the center of the nc . to further investigate this deviation of si - si bonds from crystalline si , in fig .
[ fig : strain ] we present bond length distribution with respect to distance to nc surface averaged over 2 ps of simulation time after the steady state is reached .
this figure illustrates the gradual development of radial strain from the center to nc surface .
these observations show clearly that oxidation at the surface of nc results in a tensile strain at si - si bonds which becomes significant only around the interface , while keeping the inner core almost unstrained .
this tensile strain in the nc agrees with previous measurement of hofmeister _
_ @xcite .
another consequence of this tensile strain is that the total bond orders of core - nc atoms are somewhat smaller than those of oxide - si s as seen in fig .
[ fig : panel3 ] . in the same figure
we also show the calculated net charges using electron equilibration method @xcite .
nearly zero net charges of the core - si atoms reflects the covalent type of bonding well within the nc .
the bonding becomes increasingly ionic away from the nc core as observed by the charges of si atoms which reach the value of 1.3@xmath12 at the oxide region ( fig .
[ fig : panel3 ] ) . as a result ,
the positive charges of surface - si atoms form a shell at the surface of nc .
this observation is similar with those obtained with dft calculations @xcite . on the other hand , negative charges of oxygen atoms bonded to surface form another shell that enclose nc and finally total average charges approach to zero within the oxide region . in fig . [
fig : panel3 ] we also observe that the magnitude of average charges of oxygen atoms which are bonded to surface are greater than those in the matrix .
but , the bond orders are nearly the same .
this is due to existence of 3co atoms bonded to surface .
note that the average bond order of oxygen atoms which are bonded to surface is about two ( cf .
[ fig : panel3 ] ) .
thus , those oxygen atoms form three partial bonds , two strong and one weak bond . finally , we would like to note that unlike many others @xcite , we do not observe any double bonds .
cccc ccc cc ccc cc cc c @xmath13 & @xmath14 & @xmath15&@xmath16 & & & & & & + ( )&&&&@xmath17 & charge&@xmath18&@xmath17 & charge & @xmath19 & charge&@xmath18&@xmath20 & charge&@xmath21 & charge & + 11.0 & 10 & 25 & 32 & 2 & -0.88 & 169.7 & 24 & -0.76 & 5 & -0.74 & 97.0 & 0 & & 1 & -0.81 & + 15.4 & 42 & 42 & 59 & 5 & -0.77 & 136.1 & 51 & -0.74 & 1 & -0.73 & 82.7 & 0 & & 2 & -0.77 & + 18.0 & 83 & 62 & 77 & 11 & -0.82 & 120.5 & 56 & -0.74 & 1 & -0.83 & 141.2 & 0 & & 9 & -0.79 & + 19.8 & 114 & 76 & 82 & 14 & -0.82 & 139.3 & 49 & -0.73 & 4 & -0.81 & 123.1 & 1 & -0.85 & 14 & -0.79 & + 26.8 & 353 & 143 & 170 & 20 & -0.81 & 118.0 & 123 & -0.74 & 11 & -0.79 & 120.9 & 0 & & 15 & -0.76 & + 30.8 & 558 & 203 & 243 & 35 & -0.83 & 123.9 & 159 & -0.74 & 10 & -0.80 & 115.9 & 2 & -0.80 & 34 & -0.79 & + 33.4 & 718 & 238 & 268 & 44 & -0.83 & 123.0 & 179 & -0.75 & 18 & -0.80 & 125.7 & 4 & -0.84 & 23 & -0.78 & + [ table : statatom ] the occurrence of 3co has been noted by a number of groups .
pasquarello showed that the bistable @xmath22 defect of @xmath23-quartz structure may lead to 3co as a metastable state as well as si - si dimer bond and calculated the energy of the former to be higher than the latter .
pasquarello proposed that 3co acts as an _ intermediate metastable state _ during structural relaxations at the interface @xcite .
similarly boero _ et al .
_ observed 3co atoms in their _ ab initio _
calculations @xcite and reported this feature as a metastable state . in table
[ table : statatom ] we present the collected statistical data at the end of the simulation of 75 ps . for all oxygen complexes ,
the number of bridges , average charges of bridge oxygens , and the average bridge angles for _ s_-o-_s _ are tabulated .
we observe in table [ table : statatom ] that the number of _ sss _ complexes is very small due to narrow bond angle requirement of this configuration .
for the remaining 3co complexes , _ ssm _ and _ mms _ , their percentages are seen to increase with curvature .
this can explain the fact that other studies @xcite which have identified the 3co complexes as metastable were based on the _ planar _ si / sio@xmath1 interfaces .
so , this is an indication of the importance of curvature in the stability of 3co complexes .
hence , as one would expect , there is a linear relation between the total number of bridges with surface area as indicated in fig .
[ fig : nobridgesr2 ] .
this finding is supported by kroll _
et al . _ who reported 3 and 33 such bridges for si - nc with radii @xmath24 and @xmath25 , respectively @xcite .
in conclusion , the realistic chemical environment provided by reactive force field model enables us to understand the bond topology of si - nc / a - sio@xmath1 interface and its internal dynamics .
particularly , it reveals that there are different types of oxygen complexes at the si - nc surface some of which contain 3co complexes whereas there are no double bonds .
the curvature has a positive effect on the occurence of 3co .
the relative abundance of different complexes and their charge and geometrical characteristics are extracted .
the inner core is observed to be almost unstrained while the outer core and the interface region of the nc are under increasing strain up to about a few per cents . in general , our work clearly shows that the si nc - oxide interface is more complicated than the previously proposed schemes which were based on solely simple bridge and double bonds .
the provided information here paves the way to construct realistic monte carlo moves for the simulation of large - scale silicon nanostructures embedded in oxide matrix .
this work has been supported by the european fp6 project seminano with the contract no .
nmp4 ct2004 505285 , by the turkish scientific and technical council tbitak with the project no .
the visit of tahir ain to bilkent university was facilitated by the tbitak bideb-2221 programme .
the computational resources are supplied in part by tbitak through tr - grid e - infrastructure project .
99 t. canham , appl .
phys . lett . *
57 * , 1046 ( 1990 ) .
m. v. wolkin , j. jorne , p. m. fauchet , g. allan , and c. delerue , phys .
* 82 * , 197 ( 1999 ) .
a. puzder , a.j .
williamson , j. c. grossman , and g. galli , phys .
lett . * 88 * , 097401 ( 2002 ) .
e. luppi , f. iori , r. magri , o. pulci , s. ossicini , e. degoli , and v. olevano , phys . rev .
b * 75 * , 033303 ( 2007 ) .
m. gatti and g. onida , phys .
b * 72 * , 045442 ( 2005 )
. i. vasiliev , j. r. chelikovsky and r. martin , phys .
b * 65 * , 121302 ( 2002 ) . y. tu and j. tersoff , phys . rev . lett . * 89 * , 086102 ( 2002 ) . g. hadjissavvas and p. c. kelires , phys . rev
. lett . * 93 * , 226104 ( 2004 ) .
d. e. yilmaz , c. bulutay , t. cagin , j. nanosci .
_ in press_. a. pasquarello , m. s. hybertsen , and r. car , nature ( london ) * 395 * , 58 ( 1998 ) .
d. w. brenner , phys .
b * 42 * , 9458 ( 1990 ) ; d. w. brenner , o. a. shenderova , j. a. harrison , s. j. stuart , b. ni , and s. b. sinnot , j. phys . condens . matter * 14 * , 783 ( 2002 ) .
a. c. t. van duin , a. strachan , s. stewman , q. zhang , x. xu , and w. a. goddard , iii , j. phys .
a * 107 * , 3803 ( 2003 ) .
e. demiralp , t. ain , and w. a. goddard , phys .
* 82 * , 1708 ( 1999 ) .
n. t. huff , e. demiralp , t. ain , and w. a. goddard , j. noncrystal . solids , * 253 * , 133 ( 1999 ) .
r. j. renka , acm trans .
math . soft .
( toms ) , * 22 * 1 ( 1996 ) h. hofmeister , f. huisken , and b. kohn , eur
j. d. * 9 * , 137 ( 1999 ) .
a. k. rappe and w. a. goddard , j. phys . chem . *
95 * , 3358 ( 1991 ) p. kroll and h. j. schulte , phys .
( b ) 243 r47-r49 ( 2006 ) .
p. kroll and h. j. schulte , mater .
958 0958l07 - 16 ( 2007 ) .
a. pasquarello , app . surf . sci . * 166 * , 451 ( 2000 ) .
m. boero , a. oshiyama , p .
l. silvestrelli , and k. murakami , appl .
phys . lett .
* 86 * , 201910 ( 2005 ) . | the interface chemistry of silicon nanocrystals ( ncs ) embedded in amorphous oxide matrix is studied through molecular dynamics simulations with the chemical environment described by the reactive force field model .
our results indicate that the si nc - oxide interface is more involved than the previously proposed schemes which were based on solely simple bridge or double bonds .
we identify different types of three - coordinated oxygen complexes , previously not noted . the abundance and the charge distribution of each oxygen complex is determined as a function of the nc size as well as the transitions among them .
the oxidation at the surface of nc induces tensile strain to si - si bonds which become significant only around the interface , while the inner core remains unstrained . unlike many earlier reports on the interface structure
, we do not observe any double bonds .
furthermore , our simulations and analysis reveal that the interface bond topology evolves among different oxygen bridges through these three - coordinated oxygen complexes . |
graphene , as a novel gapless two dimensional ( 2d ) chiral electron - hole system , has attracted great interest in recent years , both experimentally and theoretically@xcite .
its transport properties have been at the center of key fundamental and technological efforts with vast potential for applications in future nanotechnology@xcite . for monolayer graphene ( mlg ) ,
the fundamental interest arises from its unique linear chiral dirac carrier dispersion with a zero energy gap between conduction and valence band@xcite .
the bilayer graphene ( blg ) is also intriguing as its physical properties lie between mlg and 2d semiconductor - based electron gas ( 2deg ) systems which are gapped and non - chiral with a quadratic band dispersion .
much of the early work on graphene transport focused on the density - dependent ( i.e. , gate voltage tuned)@xcite and temperature - dependent @xcite conductivity in homogeneous mlg and blg systems .
the basic graphene transport properties , particularly at high densities far from the charge neutral dirac point , are now reasonably well - understood@xcite .
however , unintended charged impurities , which are invariably present in the graphene environment , ( e.g. , the substrate - graphene interface ) , lead to the formation of inhomogeneous electron - hole puddles in the system @xcite , which have been confirmed by experiments@xcite using the techniques of scanning potential and tunneling microscopies . although mlg samples show a metallic behavior at high densities a weak insulating " temperature - dependent conductivity @xmath0 has been measured at low carrier density and at the charge neutrality point ( cnp ) @xcite .
( we define insulating / metallic temperature dependence of conductivity @xmath0 as @xmath1 being positive / negative at fixed gate voltage . ) in addition , a recent experiment@xcite on low mobility mlg grown by chemical vapor deposition ( cvd ) shows a strong insulating " behavior at low temperatures and a metallic feature at high temperatures manifesting a non - monotonic temperature dependence in the measured electrical conductivity . in blg samples @xcite
the strong insulating behavior in the temperature dependent conductivity has been observed not only near cnp but also at carrier densities as high as @xmath2@xmath3 or higher . to be more specific , in ref . [ ] , @xmath0 in blg increases by @xmath4 as temperature @xmath5 increases from @xmath6 k for carrier density in the range @xmath7 @xmath3 . to understand this anomalous temperature dependence in @xmath0 , both mlg and blg , it is essential to know the role of disorder in graphene transport .
we note that phonon scattering ( ref . [ ] ) , although being weak in graphene , always contributes an increasing resistivity with increasing temperature and thus always leads to metallic behavior , and thus can not be the mechanism for the intriguing insulating temperature dependence often observed in graphene transport at lower carrier densities in fact , at very high temperatures ( @xmath8 k ) graphene should always manifest metallic temperature dependence in its conductivity due to phonon scattering effects which we would ignore in the current work .
our goal here is to theoretically study in a comprehensive manner the temperature dependence of graphene transport properties arising entirely from the disorder effects .
the experimentally measured anomalous temperature dependent conductivity of blg has been theoretically investigated by applying the analytic statistical theory to the inhomogeneous potential fluctuation and it is found that the anomalous blg @xmath0 is likely to be caused by the electron - hole puddles induced by randomly distributed disorder in the graphene environment @xcite . in this paper , we extend this work and apply the same analytic statistical theory to mlg systems and explain the intriguing coexistence of both metallic and insulating features of mlg @xmath0 . in the presence of large fluctuating potentials @xmath9 associated with microscopic configurations of coulomb disorder in the system , the local fermi level , @xmath10 , would necessarily have large spatial fluctuations .
we carry out an analytical theory implementing this physical idea by assuming that the value of the potential at any given point follows a gaussian distribution , parametrized by @xmath11 ( the root - mean - square fluctuations or the standard deviation in @xmath9 about the average potential ) .
this distribution can then be used to average the local density of states to obtain effective carrier densities , which can then be used to compute the physical quantities of interest@xcite .
the observed anomalous temperature dependent @xmath0 is then understood as the competition between the thermal activation of carrier density and temperature - dependent screening effects .
our theory explains the suppression of the insulating behavior in higher mobility samples with lower disorder , which is consistent with experimental observations .
we also provide the similarity and the quantitative difference between monolayer and bilayer graphene transport in the presence of puddles .
the motivation of our theory comes from the observation that the electron - hole puddles , which dominate the low - density graphene landscape , allow for a 2-component semiclassical transport behavior , where the usual metallic diffusive carrier transport is accompanied by transport by activated carriers which have been locally thermally excited above the potential fluctuations imposed by the static disorder .
this naturally allows for both insulating and metallic transport behavior occurring preferentially respectively at lower and higher carrier densities since the puddles disappear with increasing carrier density due to screening . at zero temperature ( where no activation is allowed ) or at very high carrier density ( where puddles are suppressed )
, only diffusive transport is possible .
but at any finite temperatures and at not too high densities , there would always be a 2-component transport with both activated and diffusive carriers contributing to conductivity .
our theory develops this idea into a concrete description .
we emphasize that our theory explicitly takes into account the inhomogeneous nature of the graphene landscape and is non - mean - field as a matter of principle .
this paper is organized as follows . in sec .
[ sec : mden ] , we introduce the analytical statistical theory to describe random electronic potential fluctuations created by charged impurities in the environment .
we also calculate the modified density of states and the corresponding temperature - dependent effective carrier density in monolayer graphene .
then , in sec . [ sec : mono ]
, we describe the calculations and the main features of the temperature - dependent conductivity of mlg in the presence of density inhomogeneity . in sec .
[ sec : bden ] and [ sec : bcon ] , we elaborate and extend our earlier results for the interplay between density inhomogeneity and temperature in bilayer graphene ( blg ) transport . we further discuss the connection of our theory to earlier theories in sec .
[ sec : connection ] .
we discuss the similarities and quantitative differences among the effects of inhomogeneity ( i.e. , the puddles ) on mlg and blg transport and summarize our results in sec .
[ sec : conclu ] . in appendix
[ sec : appen ] , we discuss a microscopic theory to calculate the effects of potential fluctuation on graphene systems , providing a self - consistent formulation of graphene density of states in the presence of random charged impurities near graphene / substrate interface , showing in the process that this microscopically calculated density of states agrees well with the model density of states obtained from the gaussian fluctuations .
it is well known that mlg breaks up into an inhomogeneous landscape of electron - hole puddles , especially around the charge neutral point ( cnp ) @xcite .
below we derive an analytic statistical theory taking account of the effects of inhomogeneous density in monolayer graphene ( mlg ) to explain the nonmonotonic temperature dependent transport observed in mlg @xcite .
we start by assuming that charged impurities , located in the substrate or near the graphene , create a local electrostatic potential , which fluctuates randomly about its average value across the surface of the graphene sheet .
the potential fluctuations themselves are assumed to be described by a statistical distribution function @xmath12 where @xmath13 is the fluctuating potential energy at the point @xmath14 in the 2d mlg plane .
we approximate the probability @xmath15 of finding the local electronic potential energy within a range @xmath16 about @xmath17 to be a gaussian form , i.e. , @xmath18 where @xmath19 is the standard deviation ( or equivalently , the strength of the potential fluctuation ) , which is used as an adjustable parameter to tune the tail - width@xcite . in the appendix , we provide a microscopic approach to self - consistently solve the strength of potential fluctuations in the presence of charged impurities .
due to the electron - hole symmetry in the problem , we only provide the formalism and equations for electron like carriers and the hole part can be obtained simply by changing @xmath20 to @xmath21 .
the potential fluctuations given by eq .
( [ eq : fluc ] ) affect the overall electronic density of states ( dos ) in mlg .
in our model we do not assume that the size of the puddles to be identical , but we take the puddle sizes to be completely random controlled by the distribution function given in eq .
[ eq : fluc ] .
we emphasize that our assumption of a gaussian distribution for the potential fluctuations , equivalently implying a gaussian distribution for the density fluctuations associated with the puddles , is known to be an excellent quantitative approximation to the actual numerically calculated puddle structures in graphene@xcite .
the characteristics of the puddles are determined by both the sign and the magnitude of @xmath22 , i.e. , a negative ( positive ) @xmath22 indicates an electron ( hole ) region . a different approach utilizing equal size puddles with a certain potential
@xmath17 has been used to calculate transport coefficients using a numerical transfer matrix technique @xcite .
then in the presence of electron - hole puddles the density of states is increased by the allowed electron region fraction and given by @xcite @xmath23 , \label{eq : mdos } \end{array}\end{aligned}\ ] ] where erfc@xmath24 is the complementary error function , @xmath25 and @xmath26 , where @xmath27 is the graphene fermi velocity , @xmath28 and @xmath29 are the spin and valley degeneracies , respectively .
we have @xmath30 @xmath3/mev@xmath31 with the fermi velocity @xmath32 m / s .
note that the tail of the dos is determined by the potential fluctuation strength @xmath19 .
for the case @xmath33 , the system becomes homogeneous and @xmath34 . in this case
there is no carrier density at dirac point ( @xmath35 ) at zero temperature .
it is apparent that in the presence of potential fluctuations , the @xmath36 starts at finite value @xmath37 at @xmath35 and approaches @xmath38 in high energy limit . for high - energy limit
, the carrier is essentially free since nearly every point of the system is accessible . in fig .
[ fig : mdos ] , we show the normalized density of states as a function of energy for both electrons and holes in mlg .
we mention that the self - consistent microscopic theory gives the same structure for the density of states of graphene systems ( see appendix [ sec : appen ] ) .
( @xmath39 ) , which gives rise to electron ( hole ) puddles at @xmath40 and @xmath39 . ]
since monolayer graphene is a semi - metal or zero - gap semiconductor , the electron density at finite temperatures increases due to the direct thermal excitation from valence band to conduction band , which is one of the important sources of temperature dependent transport at low carrier densities .
therefore , we first consider the temperature dependence of thermally excited electron density . the total electron density is given by @xmath41 where @xmath42 and @xmath43 is the chemical potential . at @xmath44
, @xmath43 becomes the fermi energy @xmath45 .
when the fermi energy is zero ( or at cnp ) all electrons are located in the band tail at @xmath44 and the electron and hole densities in the band tail are given by @xmath48 note that the electron ( or hole ) density in the band tails increases quadratically with the standard deviation @xmath19 . at finite temperatures the behavior of @xmath49 at cnp becomes @xmath50 .
\label{eq : den_0}\ ] ] the leading order temperature dependence in @xmath49 is quadratic . for homogeneous mlg ( @xmath33 ) with the linear - in - energy behavior of the dos , the electron density is given by @xmath51 .
in particular , in the ballistic regime the number of propagating channels increases due to the thermal smearing of the fermi surface , which leads to the observation of an insulating behavior in @xmath0 at cnp for high mobility suspended graphene samples@xcite .
the presence of the band tail does not change the quadratic temperature dependence in the thermal excitation when the system is at the charge neutral point ( @xmath52 ) .
but the inhomogeneous mlg has @xmath53 electrons in the band tails . in fig .
[ fig : den](a ) we show the temperature dependent electron density at cnp for different values of standard deviation @xmath19 .
= 1.8 in = 1.8 in = 1.8 in in the case of finite doping ( or gate voltage ) , i.e. , @xmath55 , the electron density of the homogeneous mlg ( i.e. , @xmath33 ) is given by @xmath56 where @xmath57 , and @xmath58 is the chemical potential of homogeneous mlg and is determined by the conservation of the total electron density .
then the chemical potential is given by the following relation , @xmath59 .
using the asymptotic forms @xcite of the function @xmath60 for @xmath61 and @xmath62 , i.e. , @xmath63\theta(x)+x \text{ln}(1+e^{-|x|})\ \text{for}\ |x|\gg 1 , \end{array}\ ] ] we have the asymptotic formula for the chemical potential in both low- and high - temperature limits for homogeneous mlg @xmath64 \ \ \ \text{for}\ \ t \ll t_f \\ \\ \mu_0 ( t ) \simeq
\dfrac{e_f}{4 \ln2}\dfrac{t_f}{t } \ \ \ \text{for}\ \ t \gg t_f .
\label{eq : mhomochem } \end{array}\ ] ] then the corresponding asymptotic formula of the electron density ( eq . [ eq : hf ] ) are given by @xmath65 since the direct thermal excitation is suppressed due to the finite fermi energy , the excited electron density at low temperatures ( @xmath66 ) increases quartically rather than quadratically .
but at high temperatures ( @xmath67 ) , the total electron density becomes a quadratic function of temperature as shown for an undoped mlg .
next , we derive the temperature dependence of thermally excited electron density in the presence of electron - hole puddles ( @xmath68 ) at finite doping ( @xmath55 ) . at zero temperature
the electron density for the inhomogeneous mlg can be written as : @xmath69 \\
\\ n_h ( 0 ) = \dfrac{d_1 e_f^2}{4 } \big[(1 + \tilde{s}^2)\text{erfc } ( \dfrac{1}{\sqrt{2}\tilde{s}})- \sqrt{\dfrac{2}{\pi } } \tilde{s } \exp(-\dfrac{1}{2\tilde{s}^2})\big ] \end{array}\end{aligned}\ ] ] where @xmath70 .
the presence of electron - hole puddles does not induce any additional charge in the mlg system and the net carrier density @xmath71 should be conserved .
then , the finite temperature chemical potential @xmath72 changes as a function of both temperature and the strength of potential fluctuation @xmath19 , and it should satisfy the following relation : @xmath73 where @xmath36 is the electronic density of states given by eq .
[ eq : mdos ] and @xmath74 is the density of states for holes . the asymptotic analytical formula of the chemical potential @xmath75 for inhomogeneous mlg is obtained as : @xmath76 \ \ \ \text{for}\ \ t \ll t_f \\ \\ \mu ( t ) \simeq e_f b ( \tilde{s},t)\ \ \ \text{for}\ \ \ \ t \gg t_f
\label{eq : mchem } \end{array}\ ] ] where functions @xmath77 and @xmath78 are given as follows : @xmath79\big/\big(\sqrt{\dfrac{2}{\pi } } \tilde{s}+e^{\frac{1}{2 \tilde{s}^2 } } \text{erf}\big[\dfrac{1}{\sqrt{2 } \tilde{s}}\big]\big ) \\ \\ b ( \tilde{s } , t)=\big(\dfrac{e^{-\frac{1}{2 \tilde{s}^2 } } \tilde{s } } { \sqrt{2 \pi } } + \dfrac{1}{2 } ( \tilde{s}^2 + 1 ) \text{erf}\big[\dfrac{1}{\sqrt{2 } \tilde{s}}\big]\big)\big/\big(2 \ln2 \ t + \dfrac{\tilde{s}^2}{4 t}\big ) \end{array}\ ] ] where @xmath80 and @xmath81 is the error function .
combining eqs .
[ eq : mdos ] , [ eq : mdenint ] and [ eq : mchem ] , we obtain the asymptotic analytical formula of the electron density for inhomogeneous mlg at low- and high - temperature limits as : @xmath82 in the low temperature limit ( @xmath66 ) , the leading order term for the electron density has the same quadratic behavior as in undoped homogeneous mlg ( @xmath83 ) , but the coefficient is strongly suppressed by fluctuation for the case of @xmath84 , i.e. , the high carrier density sample .
while in the case of @xmath85 , i.e. , the low carrier density sample , the existence of electron - hole puddles gives rise to a notable quadratic behavior for electron density @xmath86 [ see fig . [
fig : den](b ) ] .
in this section , we calculate the finite temperature conductivity for inhomogeneous mlg with the temperature - dependent effective carrier density derived above .
the existence of electron - hole puddles allows that the current flows through percolation channels " and the transport properties of the inhomogeneous mlg system can be derived using the self - consistent effective medium theory of conductance in composite mixtures@xcite , where the number of electrons per puddle is not an important issue for our theory .
the percolation assumption is valid as long as the potential fluctuation is larger than the thermal energy of the carriers .
otherwise transport due to disorder scattering dominates .
we emphasize that in our formalism the crossover from the percolation transport to ordinary scattering - dominated diffusive transport is guaranteed as the temperature is increased since we are explicitly taking into account both diffusive transport of free carriers and activated transport of the classically - localized carriers in our theory .
the only effects we neglect are quantum tunneling through the potential barriers and quantum interference since ours is a semiclassical theory .
we also do not consider klein tunneling explicitly in this paper because the klein tunneling occurs at zero temperature for normal incident carriers at the electron - hole puddle boundary .
we also apply the boltzmann transport theory , where we include the scattering mechanism with screened coulomb impurities and short - range disorder@xcite .
note that the application of boltzmann transport theory is justifiable because the quantum interference effects are not experimentally observed in the temperature regime of interest to us in this work .
it is conceivable that quantum interference and localization play some roles in graphene transport at very low temperatures , which is beyond the scope of this paper .
we also neglect all phonon effects in this work since electron - phonon coupling is weak in graphene .
phonon effects are relevant at high temperatures ( @xmath87 k ) and have been considered in the literature@xcite . at cnp ( @xmath83 )
electrons and holes are equally occupied . as the fermi energy increases ,
more electrons occupy increasingly larger proportion of space . as the fermi energy increases to @xmath88 ,
nearly all space is populated by the electrons [ see fig . [
fig : den](c ) ] and the conductivity of the system approaches the characteristic of the homogeneous material .
thus , there is a possible coexistence of metallic and thermally - activated transport in the presence of electron - hole puddles .
when electron puddles occupy more space than hole puddles , most electrons follow the continuous metallic paths extended throughout the system , but it is possible at finite temperatures that the thermally activated transport of electrons persists above the hole puddles . on the other hand , holes in hole puddles propagate freely , but when they meet electron puddles , activated holes conduct over the electron puddles . carrier transport in each puddle is characterized by propagation of weak scattering transport theory@xcite . the activated carrier transport of prohibited regions , where the local potential energy @xmath17 is less ( greater ) than fermi energy for electrons ( holes ) , is proportional to the fermi factor . if @xmath89 and @xmath90 are the average conductivity of electron and hole puddles , respectively , then the activated conductivities are given by @xmath91 , \\
\sigma_h^{(a)}(v ) & = & \sigma_h \exp[\beta ( v - e_f)],\end{aligned}\ ] ] where the density and temperature dependent average conductivities ( @xmath89 and @xmath90 ) are given within the boltzmann transport theory @xcite by @xmath92 and @xmath93 , where @xmath94 and @xmath95 are average electron and hole densities , respectively , and @xmath96 is the average transport relaxation time which includes the thermal smearing effects and depends explicitly on the scattering mechanism @xcite and it is given by , @xmath97 where @xmath98 and @xmath99 are , respectively , the energy - dependent transport scattering time and the finite temperature fermi distribution function .
because the density inhomogeneity effects already been considered in the variation of effective carrier density , we use the dos of homogeneous mlg @xmath100 in eq .
[ eq : avtau ] to avoid double counting . and @xmath98 is given by @xmath101\delta(\epsilon_{p\mathbf{k'}}-\epsilon_{p\mathbf{k } } ) \label{eq : mscatt}\end{aligned}\ ] ] where @xmath102 is the carrier energy for the pseudospin state @xmath103 " and @xmath104 is the 2d wave vector , @xmath105 is the matrix element of the impurity disorder potential in the system environment
, @xmath106 is the scattering angle between in- and out- wave vectors @xmath104 and @xmath107 , @xmath108/2 $ ] is a wave function form factor associated with the chiral nature of mlg ( and is determined by its band structure ) .
@xmath109 is the appropriate 2d areal concentration of the impurity centers giving rise to the random disorder potential@xcite .
we consider two different kinds of disorder scattering mechanisms : ( i ) randomly distributed screened coulomb disorder for which @xmath110 , where @xmath111 is the fourier transform of the 2d coulomb potential in an effective background lattice dielectric constant @xmath112 and @xmath113 is the 2d finite temperature static rpa dielectric function@xcite ( note that we use @xmath114 to denote the charged impurity density ) ; ( ii ) short - range disorder for which @xmath115 where @xmath116 is the 2d impurity density and @xmath117 is a constant short - range ( i.e. a @xmath118-function in real space ) potential strength .
note that the use of born approximation for short - range disorder requires weak scattering condition@xcite , which is verified by the disorder parameters we use in our calculation .
now we denote the electron ( hole ) puddle as region ` 1 ' ( ` 2 ' ) . in region 1 electrons are occupied more space than holes when @xmath119 .
the fraction of the total area occupied by electrons with fermi energy @xmath120 is given by @xmath121 .
then the total conductivity of region 1 can be calculated , @xmath122 at the same time the holes occupy the area with a fraction @xmath123 and the total conductivity of region 2 becomes @xmath124 the @xmath125 and @xmath126 are distributed according to the binary distribution .
the conductivity of binary system can be calculated by using the effective medium theory of conductance in mixtures@xcite .
the result for a 2d binary mixture of components with conductivity @xmath125 and @xmath126 is given by @xcite @xmath127 .
\label{eq : sig_tot}\ ] ] this result can be applied for all fermi energy . for a large doping case ,
in which the hole puddles disappear , we have @xmath128 and @xmath129 , then eq .
( [ eq : sig_tot ] ) becomes @xmath130 , i.e. , the conductivity of electrons in the homogeneous system .
of mlg at charge neutral point for different @xmath19 ( eq . [ eq : msig1 ] and @xmath94 as given in eq .
[ eq : den_0 ] ) .
inset shows the thermally activated conductivity of mlg as a function of temperature , where @xmath131 .
[ fig : sig_mu0 ] , width=245 ] we first consider the conductivity at cnp ( @xmath83 ) .
the conductivities in each region are given by @xmath132 , \label{eq : msig1 } \\
\sigma_2 & = & \sigma_{h } \left [ 1 + \frac{1}{2q\eta } e^{\beta^2 s^2/2 } { \rm erfc } ( \beta s/\sqrt{2 } ) \right ] , \end{aligned}\ ] ] where @xmath133 is the ratio of the hole density to the electron density .
since the electrons and holes are equally populated , we have @xmath134 and @xmath135 , then the total conductivity becomes @xmath136 .
the asymptotic behavior of the conductivity at low temperatures ( @xmath137 ) becomes @xmath138 .
\label{eq : siglast}\ ] ] the activated conductivity increases linearly with a slope @xmath139 as temperature increases .
typically @xmath19 is smaller in higher mobility samples , which gives rise to stronger insulating behavior at low temperatures .
the next order temperature correction to conductivity arises from the thermal excitation given in eq .
( [ eq : den_0 ] ) which gives quadratic ( @xmath140 ) temperature corrections .
thus , in the low temperature limit the total conductivity at the cnp is given by : @xmath141.\ ] ] at high temperatures ( @xmath142 ) we have @xmath143.\ ] ] where the temperature dependence of @xmath89 has been given in eq .
( [ eq : den_0 ] ) . the total conductivity due to the activation behavior
approaches a limiting value and all temperature dependence comes from the thermal excitation through the change of the effective carrier density in the presence of the inhomogeneity given in eq .
( [ eq : den_0 ] ) .
thus at very high temperatures ( @xmath144 ) the mlg conductivity at the charge neutral point increases quadratically regardless of the sample quality . in fig .
[ fig : sig_mu0 ] the temperature dependent conductivity has been calculated at charge neutral point , where the temperature dependent scattering mechanism can be neglected . in ref .
[ ] , about 60% increase of conductivity is observed as the temperature increases from 4 k to 300 k. we estimate the potential fluctuation parameter @xmath145 mev for this sample based on our theoretical analysis as compared with the data . at finite doping ( @xmath54 )
the temperature dependent conductivities are very complex because three energies ( @xmath120 , @xmath19 , and @xmath146 ) are competing among them .
especially when @xmath147 , regardless of @xmath120 , we have the asymptotic behavior of conductivities in region 1 and 2 from eqs .
( [ eq : sig1 ] ) and ( [ eq : sig2 ] ) , respectively , @xmath148 , \\ \sigma_2 & = & \sigma_{h } \left [ 1 + \frac{1}{2q \eta } e^{-1/2\tilde{s}^2 } \sqrt{\frac{2}{\pi } } \frac{1}{\tilde{s}/t+1/\tilde{s } } \right ] , \end{aligned}\ ] ] [ eq : msigt ] where @xmath70 and @xmath149 .
the leading order correction is linear but the coefficient is exponentially suppressed by the term @xmath150 .
this fact indicates that in the high mobility sample with small @xmath19 , the activated conductivity is weakly temperature dependent except at low density regimes , i.e. @xmath151 . since the density increase by thermal excitation is also suppressed exponentially by the same factor [ see eq .
( [ eq : den_mu ] ) ] , the dominant temperature dependent conductivity arises from the scattering time @xcite , which manifests the metallic behavior . on the other hand , for a low mobility sample with a large @xmath19 ,
the linear temperature dependence due to thermal activation can be observed even at high carrier densities @xmath152 .
= 2.0 in = 2.0 in in fig .
[ fig : sig_act ] we present the total conductivities of inhomogeneous mlg as a function of temperature ( a ) for a fixed fermi energy and several @xmath19 and ( b ) for a fixed @xmath19 and several fermi energies .
the calculations for fig .
[ fig : sig_act ] are all carried out for mlg on sio@xmath153 substrate ( corresponding to dielectric constant @xmath154 ) , charged impurity density @xmath155 @xmath3 and short - ranged disorder strength @xmath156 ( ev )@xmath157 . for total conductivity ,
the thermally activated insulting behavior competes with the temperature - dependent screening effects , where the latter always give the metallic behavior in conductivity for mlg samples .
when @xmath19 is small , the activated behavior is suppressed and the total conductivity shows the metallic behavior . while for large value of @xmath19 , i.e. , the low mobility sample , the thermal activation overwhelms the metallic temperature dependence and the system manifests insulating behavior .
for @xmath158 the situation becomes much complex . at low temperatures , the leading order of the temperature dependence is linear ( the second term in eq .
( [ eq : msigt ] ) ) and the total conductivity starts at weakly insulating behavior . as the temperature increases , the screening effects begin to dominant leading to the metallic behavior . as a result ,
the temperature evolution of the conductivity becomes non - monotonic and for large @xmath19 ( or low mobility samples ) the nonmonotonic behavior can be more pronounced as shown in experiments @xcite .
in the following of this paper , we extend our previous study@xcite on the insulating behavior in metallic bilayer graphene and compare it with mlg situation .
the most important difference between mlg and blg comes from the fact that , in the blg , the two layers are weakly coupled by interlayer tunneling , leading to an approximately parabolic band dispersion with an effective mass about @xmath159 ( @xmath160 corresponds to the bare electron mass ) contrast to linear - dispersion dirac carrier system for mlg . as done for mlg , we assume the electronic potential fluctuations in blg system to be a gaussian form given in eq . [ eq : fluc ] and this potential is felt equally by both layers@xcite .
( @xmath39 ) , which gives rise to electron ( hole ) puddles at @xmath40 and @xmath39 . ] in the presence of potential fluctuations the density of states ( dos ) for disordered blg is given by @xmath161 , where @xmath162 is the dos in a homogeneous blg system , where @xmath28 and @xmath29 are the spin and valley degeneracies , respectively .
we have @xmath163 @xmath3/mev assuming @xmath164 .
the dos of hole can be calculated from the following relation : @xmath165 . in fig .
[ fig : bdos ] , the density of states of both electron and hole are shown for the inhomogeneous blg system . in the presence of potential fluctuations ,
the electron and hole coexist for certain amount of regions near cnp and their dos approach to the homogeneous case as the carrier energy further increases .
because blg is also a gapless semiconductor like mlg , the direct thermal excitation from valence band to conduction band at finite temperatures composes an important source of temperature dependent transport in blg .
thus , the temperature dependence of thermally excited electron density is first to be considered . with the help of eq .
[ eq : mdenint ] , we could get the total electron density for blg in the presence of electron - hole puddles .
we first consider the situation at cnp , where all electrons are located in the band tail at @xmath44 and the electron density in the band tail is given by @xmath166@xcite .
contrast to the quadratic dependence of @xmath19 in mlg , the electron density in the band tail for blg is linearly proportional to the standard deviation @xmath19 . unlike mlg , which has the exact formula for @xmath167 ( i.e. , eq .
( [ eq : den_0 ] ) ) , we could only find the asymptotic behavior of @xmath167 at finite temperatures for blg .
the low temperature ( @xmath168 ) behavior of electron density at cnp becomes @xmath169 .
\label{eq : bden_0}\ ] ] thus , the electron density increases quadratically at low temperature limit . for homogeneous blg with the constant dos the electron density at finite temperatures is given by @xmath170 , which has the universal slope @xmath171 .
the presence of the band tail suppresses the thermal excitation of electrons and gives rise to the quadratic behavior .
however , at high temperature limit , the density increases linearly with the same slope approaching to the homogeneous system , i.e. , @xmath172 .
\label{eq : bden_0h}\ ] ] in fig .
[ fig : bden](a ) we show the temperature dependent electron density at cnp for different standard deviations . compared with the inset of fig .
[ fig : sig_mu0 ] , it is apparent that , even for the same strength of potential fluctuation @xmath19 , the effects of thermal excitation of carrier density are much stronger in blg than in mlg sample , which leads to more easily observed insulating behavior in blg samples .
= 1.8 in = 1.8 in = 1.8 in in this subsection , we derived the total electron density at finite temperatures for inhomogeneous blg away from cnp .
contrary to mlg , we need to calculate the finite temperature chemical potential ( i.e. , eqs .
[ eq : mhomochem ] and [ eq : mchem ] ) .
the charge conservation relation in both homogeneous and inhomogeneous blg gives the temperature independent chemical potential @xmath173 , allowing us to directly calculate the total effective electron ( hole ) density . in the case of finite gate voltage ,
i.e. , @xmath55 , the electron density of the homogeneous blg for @xmath33 is given by @xmath174,\ ] ] where @xmath149 and @xmath175 .
the thermal excitation is exponentially suppressed due to the fermi function at low temperatures ( @xmath66 ) . while at high temperatures ( @xmath67 ) it increases linearly . in the presence of finite potential fluctuations ( @xmath68 ) ,
the electron and hole density at zero temperature for the inhomogeneous system are given by : @xmath176,\\ \\
n_h(0 ) = { d_0 e_f } \left [ - \dfrac{1}{2}{\rm erfc } \left ( \dfrac{1}{\sqrt{2}\tilde{s } } \right ) + \dfrac{\tilde{s}}{\sqrt{2\pi } } e^{-1/2\tilde{s}^2 } \right ] \end{array}\ ] ] where @xmath70 and the difference of electron and hole density ( @xmath177 ) is independent of the strength of potential fluctuation @xmath19 ( see fig . [
fig : bden](c ) ) . at low temperatures ( @xmath66 )
the asymptotic behavior of the electron density is given by @xmath178 the leading order quadratic behavior of @xmath49 as in undoped blg ( @xmath83 ) is strongly suppressed by potential fluctuation .
for the situation @xmath85 , the existence of electron - hole puddles gives rise to a notable quadratic behavior [ see fig .
[ fig : bden](b ) ] . at high temperatures ( @xmath67 )
we find @xmath179 where the linear temperature dependence of electron density is dominant as the homogeneous system .
with the help of total electron and hole density calculated above , we will derive the temperature dependent conductivity for blg in the presence of electron - hole puddles .
we will apply both boltzmann theory@xcite and effective medium theory@xcite to interpret the intriguingly insulating behavior observed in blg samples@xcite .
the density and temperature dependent average conductivities in blg , denote as @xmath89 and @xmath90 , are given within the boltzmann transport theory : @xmath180 where @xmath94 and @xmath95 are average electron and hole densities , respectively .
@xmath96 is the transport relaxation time for bilayer graphene : @xmath181 and @xmath98 is calculated with eq .
[ eq : mscatt ] .
but for blg systems , one needs to use the parabolic dispersion relation @xmath182 for the pseudo - spin state @xmath103 " and the static dielectric screening function derived in ref .
the wave function form factor associated with the chiral nature of blg is also different from the case in mlg , which is given by @xmath183/2 $ ] . to determine the average scattering time in blg
, we take into account the long - range charged impurity scattering and short - range defect scattering , which has been established that both contribute significantly to bilayer graphene transport properties@xcite .
the activated conductivities should also be included in the presence of density inhomogeneity in the blg , which follow the same relation as given for mlg : @xmath91 , \\
\sigma_h^{(a)}(v ) & = & \sigma_h \exp[\beta ( v - e_f)],\end{aligned}\ ] ] when electron - hole puddles form in the blg samples ( denote the electron ( hole ) puddle as region ` 1 ' ( ` 2 ' ) ) , the transport properties can be treated with effective medium theory as described in sec .
[ sec : mono ] . and
[ eq : sig1]-[eq : siglast ] for the inhomogeneous mlg also apply to the inhomogeneous blg system .
we will first discuss the total conductivity of blg at cnp ( @xmath184 ) .
in this case , the electron and hole are equally occupied and the total conductivity @xmath185 ( see eq .
[ eq : msig1 ] and [ eq : siglast ] ) . at low temperature limit ( @xmath186 ) , the activated conductivities increase linearly with a slope @xmath139 as the temperature increases .
the next order temperature correction to the conductivity is quadratic @xmath140 , which arises from the thermal activation ( see eq .
[ eq : bden_0 ] ) .
thus , at low temperature limit the total conductivity at cnp is given by @xmath187.\ ] ] at high temperatures ( @xmath142 ) , the total conductivity is given by : @xmath143.\ ] ] it is apparent that the activation behavior approaches a limiting value at high temperature limit ( @xmath188 ) while the thermally activated electron density becomes dominant , which increases linearly with a universal slope @xmath189 regardless of the sample quality .
thus , all temperature dependence of the total conductivity comes from the thermal excitation through the change of the carrier density given in eq .
( [ eq : bden_0h ] ) . in fig .
[ fig : bsig_mu0 ] we show the calculated temperature dependent conductivity at charge neutral point .
the inset present the activated conductivity versus the temperature . in ref .
[ ] , @xmath190 at the cnp of the blg sample increases almost two times as temperature @xmath5 varies from 4 k to 300 k. our theoretical analysis using a potential fluctuation parameter @xmath191 mev gives reasonable agreement with the experimental data .
of blg at charge neutral point for different @xmath19 ( eq .
[ eq : msig1 ] with @xmath94 for blg ) .
inset shows the thermally activated conductivity in blg as a function of temperature , where @xmath131 , the same as for the mlg case .
[ fig : bsig_mu0 ] , width=245 ] the temperature dependent conductivities at finite doping ( @xmath54 ) are very complex because three energies ( @xmath120 , @xmath19 , and @xmath146 ) are competing . regardless of @xmath120 ,
when @xmath147 , we have the asymptotic behavior of conductivities in region 1 and 2 the same as mlg situation , given in eq .
( [ eq : msigt ] ) .
but the average electron and hole conductivities ( @xmath89 and @xmath90 ) are quite different from mlg case , which is determined by the specific band dispersion relation and also the dielectric function @xmath192 .
thus , the leading order correction to @xmath193 in blg is also linear , which comes from the activated conductivity , but the coefficient is exponentially suppressed by the term @xmath150 . in the high mobility sample with small @xmath19 , the activated conductivity is weakly temperature dependent except around cnp , i.e. @xmath151 . since the density increase by thermal excitation is also suppressed exponentially by the same factor [ see eq .
( [ eq : bden_mu ] ) ] the dominant temperature dependent conductivity arises from the scattering mechanism@xcite . on the other hand , in the low mobility sample with large value of @xmath19 , the linear temperature dependence due to thermal activation can be observed even at high carrier densities @xmath152 .
of blg with the following parameters : @xmath194 @xmath3 and @xmath156 ( ev )@xmath157 . ( a ) @xmath195 for @xmath196 mev and for different @xmath19 .
( b ) @xmath195 for @xmath197 mev and for several @xmath198 , 40 , 60 , 80 mev , which correspond to the net carrier densities @xmath199 , @xmath200 , @xmath201 , and @xmath202 @xmath3 . , title="fig:",width=245 ] of blg with the following parameters : @xmath194 @xmath3 and @xmath156 ( ev )@xmath157 .
( a ) @xmath195 for @xmath196 mev and for different @xmath19 .
( b ) @xmath195 for @xmath197 mev and for several @xmath198 , 40 , 60 , 80 mev , which correspond to the net carrier densities @xmath199 , @xmath200 , @xmath201 , and @xmath202 @xmath3 . , title="fig:",width=245 ] in fig . [ fig : bsig_tot ] we calculate the total conductivities ( a ) for a fixed @xmath120 and several @xmath19 and ( b ) for a fixed @xmath19 and several @xmath120 . even for homogeneous blg
, there are two scattering mechanism competing with each other .
the short - range disorder in blg contributes to a strong insulating transport behavior for all temperature , whereas screened coulomb scattering always leads to a metallic behavior for @xmath203@xcite . at low temperature limit ,
the total conductivity @xmath195 decreases with increasing temperature , but at higher temperatures , the short - range disorder contribution becomes quite big and leads to a @xmath195 increasing with @xmath5 .
therefore , when @xmath19 is small , the scattering mechanism is dominant and the total conductivity manifests a non - monotonic temperature dependence ( see fig . [
fig : bsig_tot](a ) ) .
however , for large @xmath19 the activated temperature dependence behavior overwhelms the metallic temperature dependence , and the system shows insulating behavior ( see fig .
[ fig : bsig_tot](b ) ) .
it clearly shows that the insulating behavior in blg sample appeared at carrier densities as high as @xmath2 @xmath3 or higher .
we have demonstrated theoretically that the observed insulating behavior in temperature - dependent monolayer and bilayer graphene conductivity can be explained by the thermal activation between puddles .
there are also other theories which have been elaborated to explain low carrier density graphene transport @xcite . in this section ,
we establish the bridge to connect our current theory and earlier theories on graphene transport due to the formation of inhomogeneous electron - hole puddles near the charge neutrality point .
the key qualitative difference between our theory and all earlier graphene transport theories is the introduction of the 2-component transport model where regular diffusive metallic carrier transport coexists with local activated transport due to activation across potential fluctuations in the puddles .
our theory just explicitly accounts for the inhomogeneous landscape in the system , which earlier theories ignored .
this 2-component nature of graphene transport , where both metallic and insulating behavior coexist because of the existence of puddles , produces the experimentally observed complex temperature dependence with the low - density behavior being primarily insulating - like and the high - density behavior being primarily metallic - like .
two different theories have been developed to study the low - density transport in graphene , where the strong density inhomogeneity is dominated . in ref .
[ ] adam _ et al .
_ qualitatively explained the plateau - like approximate nonuniversal minimum conductivity at low carrier density observed in monolayer graphene samples .
the basic idea is to introduce an approximate pinning of the carrier density at @xmath204 at low carrier density limits @xmath205 , where @xmath114 is an impurity density .
the constant minimum conductivity is then given by @xmath206 for @xmath207 .
this simple theory for monolayer graphene transport qualitatively explained the existence of conductivity minimum plateau and the extent to which the minimum conductivity is not universal , which was in good agreement with the observed density - dependent conductivity over a wide range of charged impurity densities@xcite .
however , this theory did not take account of the highly heterogeneous structure near charge neutrality point and the thermally activated conductivity at finite temperatures , which then can not explain the observed non - monotonic temperature dependent transport in low mobility graphene samples@xcite .
a more elaborate thomas - fermi - dirac ( tfd ) theory and an effective medium approximation ( emt ) have been introduced in refs . [ ] and [ ] to study the electrical transport properties of disordered monolayer graphene .
the ground state - density landscape @xmath208 can be obtained within this tfd approach and the resultant electrical transport can be calculated by averaging over disorder realizations and the effective medium theory .
this theory gives a finite minimum conductivity and is able to explain the crossover of the density - dependent conductivity from the minimum value at the dirac point to its linear behavior at higher doping .
later , this tfd - emt theory is also applied to calculate the conductivity of disordered bilayer graphene in ref . [ ] .
the tfd - emt technique successfully explains the graphene , both mlg and blg , transport properties in the theoretically difficult inhomogeneity - dominant regime near the charge neutral point , but this approach fails to explain the temperature dependence of the conductivity for a wide range of temperatures
. in our current model discussed above , we include three effects , the electron - hole structure formation , the thermal activated conductivities and the temperature dependence of screening effects , to explain the temperature - dependent conductivity in both monolayer and bilayer graphene systems .
the nonmonotonic temperature - dependent conductivity in graphene systems is then naturally understood from the competition between the thermal activation of charge carriers and the temperature - dependent screening effects .
our transport theory qualitatively explains the observed coexisting metallic and insulating transport behavior in both mlg and blg systems . for low mobility mlg samples , the dominant role on graphene conductivity switches from the thermally activated transport of inhomogeneous electron - hole puddles to metallic temperature - dependent screening effects , which gives rise to a nonmonotonic behavior from the strong insulating behavior at low temperatures to metallic behavior at high temperatures . on the other hand
, another nonmonotonic temperature - dependent transport can be observed in very high mobility bilayer graphene devices , i.e. , from metallic behavior at low temperatures due to the screening effects of coulomb scattering to insulating behavior at high temperatures due to the short - range disorder .
the merit of our model is that it is so simple that we could get the asymptotic behavior at low and high temperature limits analytically .
moreover , it provides a clear physical picture of the dominant mechanisms at different regimes as discussed above .
we first discuss the similarity and the difference between mlg and blg transport from the perspective of our transport - theory considerations . we find that both manifest an insulating behavior in @xmath195 for low mobility samples .
we also find that both systems could exhibit a non - monotonic temperature dependent conductivity for low mobility samples .
however , the physical origin for the non - monotonic temperature dependence is quite different in the two systems : in the mlg the non - monotonic feature comes from the competition between thermal activation and the metallic screening effects , which leads to @xmath195 first increasing and then decreasing with increasing temperature ( see fig .
[ fig : sig_act](a ) ) . while for blg , the competition between short - range insulating scattering and metallic coulomb screening effects leads to @xmath195 first decreasing and then increasing as temperature increases ( see fig .
[ fig : bsig_tot](a ) ) .
most important quantitative difference between mlg and blg transport comes from their band dispersions , which leads to much weaker effects of density inhomogeneity in mlg so that the anomalous insulating temperature dependence of @xmath0 is typically not observed in mlg away from the cnp although the gate voltage dependence of mlg and blg conductivities are similar@xcite .
the linear dirac carrier system for mlg leads to linear dos , which goes to zero at cnp , but the parabolic band dispersion relation in blg leads to a constant dos .
due to the difference in the density of states between homogeneous mlg and blg , the modified dos in inhomogeneous mlg is increased ( see fig .
[ fig : mdos ] ) rather than decreased in inhomogeneous blg ( see fig .
[ fig : bdos ] ) .
the dimensionless potential fluctuation strength @xmath209 ( @xmath210 ) is much weaker in mlg than in blg from simple estimates : @xmath211 where @xmath212 , and @xmath213 upto @xmath214 @xmath3 .
direct calculations @xcite show that the self - consistent values of @xmath19 tend to be much larger in blg than in mlg for identical impurity disorder .
in addition , the qualitatively different dos leads to much stronger effective short - range scattering in blg compared with mlg even for the same bare scattering strength .
thus , the insulating behavior in @xmath195 will show up at high temperatures even for relatively higher mobility blg samples ( i.e. , small @xmath19 ) .
in contrast , only in very low mobility mlg samples , where @xmath19 is very large , can the insulating behavior of temperature dependent resistivity be observed@xcite .
no simple picture would apply to a gapped ( @xmath215 ) blg system , since four distinct energy scales ( @xmath216 , and @xmath215 ) will compete and the conceivable temperature dependence depends on their relative values@xcite .
our assumption of blg quadratic band dispersion is valid only at low ( @xmath217 @xmath3 ) carrier densities , where most of the current transport experiments are carried out . at higher densities
the band dispersion is effectively linear and the disorder effects on @xmath195 are weaker . before concluding , we emphasize that our theory is physically motivated since puddles are experimental facts in all graphene samples . puddles automatically imply a 2-component nature of transport since both diffusive carriers and activated carriers can , in principle , contribute to transport in the presence of puddles
of course , the effect of puddles is much stronger at low carrier densities , explaining why insulating ( metallic ) temperature dependence is more generic at low ( high ) graphene carrier densities .
we emphasize that local carrier activation in puddles is just one of ( at least ) four different independent transport mechanisms contributing to the temperature dependent conductivity .
the other three are temperature dependent screening ( ref .
[ ] ) , phonons ( refs . [ ] ) , and fermi surface thermal averaging ( refs . [ ] ) .
our theory presented here includes the three electronic mechanisms for temperature dependence : screening , fermi surface averaging , and puddle activation .
we leave out phonons , which have been considered elsewhere ( ref . [ ] ) and will simply add to the temperature dependent resistivity . the weak phonon contribution to graphene resistivity
makes it possible for the electronic mechanisms to dominate even at room temperatures , but obviously at high enough temperatures , the system will , except perhaps at the lowest densities around the cnp , manifest metallic temperature dependence with the resistivity increasing with temperature because of phonon scattering . similarly , the puddle effects dominate low densities and therefore , the insulating behavior will persist to very high temperatures around the zero - density cnp since activation across potential fluctuations are dominant at the cnp .
it is gratifying to note that these are precisely the experimental observations .
we note that in general the temperature dependent conductivity of graphene could be very complex since many distinct mechanisms could in principle contribute to the temperature dependence depending on the carrier density , temperature range , and disorder in the system .
inclusion of phonons ( at high temperatures ) and quantum localization ( at low temperatures ) effects , which are both neglected in our theory , can only complicate things further .
what we have shown in this work is that the low - density conductivity near the cnp is preferentially dominated by density inhomogeneity and thermal carrier activation effects leading to an insulating temperature dependence in the conductivity whereas the high - density conductivity , where the puddles are screened out , is dominated by a metallic conductivity due to temperature - dependent screening effects .
this general conclusion is consistent with all experimental observations in both mlg and blg systems to the best of our knowledge except at very high temperatures where phonon effects would eventually lead to metallic behavior at all densities . to conclude
, we have investigated both mlg and blg transport in the presence of electron - hole puddles within an analytic statistical theory .
our theory explains the experimentally measured insulating behavior at low temperatures and the consequent nonmonotonic behavior for low mobility samples @xcite .
a reasonable quantitative agreement with the experimental data can be obtained by choosing appropriate disorder parameters in our theory ( i.e. potential fluctuation and impurity strength ) for different samples .
we find that the puddle parameter @xmath19 , defining typical potential fluctuations , to be around @xmath218 mev in typical graphene samples as extracted by fitting our theory to existing experimental transport data near the charge neutrality point .
these values of potential fluctuations characterizing the graphene charge neutrality point are very consistent with direct numerical calculations of graphene electronic structure in the presence of quenched charged impurities@xcite .
we also relate our current model to earlier theories using the picture of diffusive transport through disorder - induced electron - hole puddles . finally , we show the similarity and the quantitative difference between mlg and blg transport in the presence of puddles .
ql acknowledges helpful discussions with d. s. l. abergel .
the work is supported by onr - muri , nri - nsf - swan .
below we provide a microscopic theory to calculate self - consistently the electronic density of states in the presence of the potential fluctuations caused by random charged impurities located near graphene / substrate interface , which has been applied to two dimensional semiconductor based electron gas systems@xcite .
this self - consistent approach mainly addresses two problems with the presence of random charged impurities .
one is the screening of the long - range coulomb interactions between the carriers and the charged impurities .
the other is the real - space potential fluctuations produced by the random array of charged impurities .
the motivation for this appendix is two - fold : ( 1 ) providing a microscopic self - consistent theory of graphene density of states in the presence of puddles ; ( 2 ) showing that our approximate physically - motivated density of states ( eq . [ eq : mdos ] ) is an excellent approximation to the self - consistent density of states . versus the screening constant @xmath219 ( loglog plot ) in mlg by varying the fermi level .
the dotted blue line is for @xmath220 @xmath3 , @xmath221 nm and @xmath222 nm .
the solid red line is for @xmath223 @xmath3 , @xmath221 nm and @xmath224 nm . the dashed green line is for @xmath223 @xmath3 , @xmath221 nm and @xmath222 nm . the dotdashed black line is for @xmath223 @xmath3 , @xmath225 nm and @xmath222 nm . ] in mlg versus electron density using the following parameters : the insulator thickness @xmath222 nm , the impurity distance from the interface @xmath226 nm . the solid red line is for unperturbed density of states .
the dashed green , dotted blue and dotdashed black lines are corresponding to @xmath227 , @xmath228 and @xmath229 @xmath3 , respectively . ] the simple theory of linear screening gives@xcite : @xmath230 where @xmath219 is the thomas - fermi screening wavevector , @xmath231 is the density of states at the fermi level and @xmath112 is the dielectric constant ( @xmath232 for graphene on sio@xmath153 substrate ) .
the screening constant shown in eq .
[ eq : s ] enters poisson s equation for the potential change @xmath233 produced by a charge density @xmath234 ( associated with the charged impurities in the graphene / substrate environment ) . for a charge @xmath235 ( we use @xmath236 in the calculation ) located at @xmath237 and @xmath238 ( on top of the graphene layer ) ,
the additional coulomb potential satisfies : @xmath239 where @xmath240 in the vacuum ( @xmath241 ) , @xmath242 in sio@xmath153 ( @xmath243 ) and @xmath244 . for graphene
, @xmath245 is the carrier density distribution normal to the interface and @xmath246 . to solve eq .
[ eq : poisson ] we take advantage of the cylindrical symmetry to write@xcite : @xmath247 the potential will satisfy eq .
[ eq : poisson ] if @xmath248 at the interface @xmath249 , @xmath250 must be continuous and satisfy @xmath251 .
@xmath250 should also satisfies the boundary condition @xmath252 as @xmath253 .
in addition , the impurity potential @xmath254 will go to zero at the metallic contact below the sio@xmath153 ( i.e. , @xmath255 and @xmath256 is the thickness of the sio@xmath153 layer ) .
such screening effects are absent in the sio@xmath153 .
after some algebra , the explicit expression of @xmath257 for insulator thickness @xmath256 and the impurity distance from graphene / substrate interface @xmath258 is given by : @xmath259 for the thickness of insulator in the limit @xmath260 , we have @xmath261 , which has been given in the appendix b of ref . [ ] .
the potential fluctuations with an array of point charges at random positions in the plane @xmath262 have a mean - square variation about the average potential @xcite : @xmath263 ^ 2 k dk \label{eq : flucv}\ ] ] of mlg versus energy @xmath20 for different impurity configurations and carrier densities @xmath264 .
the solid red lines are for the non - interacting mlg system .
( a ) calculated @xmath36 in mlg for the insulator thickness @xmath222 nm , the impurity distance from the interface @xmath226 nm and carrier density @xmath265 @xmath3 . the dashed green , dotted blue and dotdashed black lines are corresponding to @xmath227 , @xmath228 , @xmath266 @xmath3 , respectively .
( b ) calculated @xmath36 in mlg for @xmath222 nm , @xmath267 @xmath3 and @xmath265 @xmath3 .
the dashed green , dotted blue and dotdashed black lines are corresponding to @xmath221 , @xmath268 , @xmath269 nm , respectively . ( c ) calculated @xmath36 in mlg for @xmath222 nm , @xmath226 nm and @xmath223 @xmath3 .
the dashed green , dotted blue and dotdashed black lines are corresponding to @xmath270 , @xmath228 , @xmath266 @xmath3 , respectively . ] to obtain specific results for the electronic density of states and the screening constant we use the simple gaussian broadening approximation for the density of states@xcite .
the disorder - induced potential energy fluctuations is described by @xmath271 ( eq .
[ eq : flucv ] ) . then the density of states becomes @xmath272 \end{array}\ ] ] where erfc@xmath24 is the complementary error function , @xmath273 , @xmath274 , @xmath27 is the graphene ( fermi ) velocity , @xmath28 and @xmath29 are the spin and valley degeneracies , respectively . by choosing the chemical potential @xmath120 as a tuning parameter we have the following coupled equations : @xmath275 ^ 2 k dk \end{array}\ ] ] for fixed values of @xmath120 , @xmath276 , @xmath256 and @xmath258
, we get the self consistent results for @xmath19 , @xmath219 by solving the above two coupled equations .
the electron density could be gotten from the formula : @xmath277 where @xmath278 is the fermi - dirac distribution function .
the electron density in the presence of disorder - induced electron - hole puddles has been discussed in sec .
[ sec : mden ] , where we use the potential fluctuation @xmath19 as a fixed parameter . and here we self - consistently solve the parameter @xmath19 from a microscopic point of view , which is in good agreement with the results shown in sec .
[ sec : mden ] . the potential fluctuation in eq .
[ eq : flucv ] affects the electronic density of states .
but the fluctuations depend on the screening via eq .
[ eq : ak ] while the screening depends on the density of states via eq .
[ eq : s ] .
therefore , we have a coupled problem which must be solved self - consistently . in fig .
[ fig : parameter ] , the standard deviation of the potential fluctuation @xmath19 and the screening constant @xmath219 are plotted for different values of the fermi level .
the self - consistently solved parameters @xmath279 depend on the fixed charged impurity density @xmath276 , the sio@xmath153 thickness @xmath256 , the location of the fixed charged impurity @xmath258 , and the fermi level @xmath120 ( i.e. the carrier density @xmath264 ) .
all these four effects can be understood from physical intuition .
the reduction of the sio@xmath153 thickness weakens the potential fluctuations when the screening length is small even though there is a little effect for strong screening . as the charged impurities go away from the graphene layer the potential fluctuations
is also reduced , while the potential fluctuations becomes stronger with the higher impurity density . increasing the carrier density @xmath264
gives rise to the stronger screening effects , and leads to weaker potential fluctuations . in fig .
[ fig : mdosden ] , the density of states of monolayer graphene is given with the parameters of sio@xmath153 thickness @xmath222 nm and the distance of fixed charged impurities @xmath221 nm for different impurity densities @xmath276 . in fig .
[ fig : mdose ] , we present the electronic density of states for different carrier densities and impurity configurations . the self - consistent calculation of the density of states verifies the results presented in sec.[sec : mden ] as shown in fig . [
fig : mdos ] , where we choose the potential fluctuation @xmath19 as an adjustable parameter . for monolayer graphene ,
the presence of spatially random charged impurities increases the electronic density of states in the whole range of energy .
the corresponding hole density of states can be obtained by changing the sign of energy @xmath280 . versus the screening constant @xmath219 ( loglog plot ) in blg by varying the fermi level .
the dotted blue line is for @xmath220 @xmath3 , @xmath221 nm and @xmath222 nm . the solid red line is for @xmath223 @xmath3 , @xmath221 nm and @xmath224 nm . the dashed green line is for @xmath223 @xmath3 , @xmath221 nm and @xmath222 nm .
the dotdashed black line is for @xmath223 @xmath3 , @xmath225 nm and @xmath222 nm . ] of blg versus electron density using the following parameters : the insulator thickness @xmath222 nm , the impurity distance from the interface @xmath226 nm . the solid red line is for unperturbed density of states .
the dashed green , dotted blue and dotdashed black lines are corresponding to @xmath227 , @xmath228 and @xmath229 @xmath3 , respectively . ] in this subsection , we provide the density of states in bilayer graphene in the presence of potential fluctuations . as shown for monolayer graphene , we use the linear screening written as@xcite : @xmath281 where @xmath231 is the density of states of blg at the fermi level and @xmath112 is the dielectric constant and for blg on sio@xmath153 , @xmath282 . of blg versus energy @xmath20 for different impurity configuration and carrier densities @xmath264 .
the solid red lines are for the non - interacting blg system .
( a ) calculated @xmath36 in blg for the insulator thickness @xmath222 nm , the impurity distance from the interface @xmath226 nm and carrier density @xmath265 @xmath3 . the dashed green , dotted blue and dotdashed black lines are corresponding to @xmath227 , @xmath228 , @xmath266 @xmath3 , respectively .
( b ) calculated @xmath36 in blg for @xmath222 nm , @xmath267 @xmath3 and @xmath265 @xmath3 .
the dashed green , dotted blue and dotdashed black lines are corresponding to @xmath221 , @xmath268 , @xmath269 nm , respectively .
( c ) calculated @xmath36 in blg for @xmath222 nm , @xmath226 nm and @xmath223 @xmath3 .
the dashed green , dotted blue and dotdashed black lines are corresponding to @xmath270 , @xmath228 , @xmath266 @xmath3 , respectively . ]
following the same procedure discussed for mlg , the disorder - induced potential fluctuation is described by the gaussian form @xmath283 and the corresponding density of states can be written as ( also see sec . [
sec : bden ] ) : @xmath284 where erfc@xmath24 is the complementary error function , @xmath273 ( as given in eq .
[ eq : flucv ] ) , @xmath285 , @xmath28 and @xmath29 are the spin and valley degeneracies , respectively .
the main difference between mlg and blg is in their density of states of non - interacting systems .
the homogeneous mlg system has the linear energy - dependent density of states while the density of states of the homogeneous blg is independent of energy , which leads to different thomas - fermi screening wavevectors .
the potential fluctuation in eq .
[ eq : flucv ] affects the electronic density of states in blg .
but the fluctuations depend on the screening via eq .
[ eq : ak ] while the screening depends on the density of states via eq .
[ eq : bs ] .
therefore , we have a coupled problem which must be solved self - consistently . in fig .
[ fig : bparameter ] , the broadening parameter @xmath19 and the screening constant @xmath219 are plotted for various fermi levels ( i.e. the carrier density @xmath264 ) .
as shown for mlg , the blg parameters @xmath279 are also non - trivial function of the fixed charge density @xmath276 , the sio@xmath153 thickness @xmath256 , the location of the fixed charged impurity @xmath258 , and the fermi level .
the different charged impurity configurations and carrier densities have similar effects on potential fluctuations of bilayer graphene as we discussed for monolayer graphene .
the results for @xmath286 are also quite similar to that of mlg ( in fig .
[ fig : parameter ] ) only with small numerical difference . in fig .
[ fig : bdosden ] , the self - consistent electronic density of states of blg has been calculated using sio@xmath153 thickness @xmath222 nm and distance of fixed charged impurities @xmath221 nm for different impurity densities .
the higher impurity density changes the density of states more dramatically . in fig .
[ fig : bdose ] , we show the electronic density of states of blg for different charged impurity configurations and carrier densities .
the existence of random charged impurities reduces the electronic density of states for @xmath39 but create a band tail for @xmath40 . | we theoretically study the transport properties of both monolayer and bilayer graphene in the presence of electron - hole puddles induced by charged impurities which are invariably present in the graphene environment .
we calculate the graphene conductivity by taking into account the non - mean - field two - component nature of transport in the highly inhomogeneous density and potential landscape , where activated transport across the potential fluctuations in the puddle regimes coexists with regular metallic diffusive transport .
the existence of puddles allows the local activation at low carrier densities , giving rise to an insulating temperature dependence in the conductivity of both monolayer and bilayer graphene systems .
we also critically study the qualitative similarity and the quantitative difference between monolayer and bilayer graphene transport in the presence of puddles .
our theoretical calculation explains the non - monotonic feature of the temperature dependent transport , which is experimentally generically observed in low mobility graphene samples .
we establish the 2-component nature ( i.e. , both activated and diffusive ) of graphene transport arising from the existence of potential fluctuation induced inhomogeneous density puddles .
the temperature dependence of the graphene conductivity arises from many competing mechanisms , even without considering any phonon effects , such as thermal excitation of carriers from the valence band to the conduction band , temperature dependent screening , thermal activation across the potential fluctuations associated with the electron - hole puddles induced by the random charged impurities in the environment , leading to very complex temperature dependence which depends both on the carrier density and the temperature range of interest . |
there is a very strong scientific case that most of the matter in the universe consists of non - baryonic stable particles .
since the standard model of particle physics does not contain any heavy stable non - baryonic particles new particle physics is required .
but what is this new physics
? it is widely assumed that the particles comprising non - baryonic dark matter are weakly interacting in the sense that they interact with ordinary matter via exchange of w , z gauge bosons , higgs bosons or more exotic heavy particles . from collider bounds
( e.g. lack of new particles in decays of the w and z gauge bosons ) , the masses of any new weakly interacting particles should be ( typically ) greater than about 30 - 45 gev , unless indicated otherwise . ] depending on the model .
however , such heavy weakly interacting particles should decay with a lifetime of order @xmath1 seconds ( for @xmath2 ) about 41 orders of magnitude too short - lived to be suitable as dark matter candidates seconds . ] .
thus , one must make additional ad hoc assumptions in order to stabalize such hypothetical particles . in the end , theories with such particles require multiple unrelated assumptions and become very ugly from a particle physics point of view .
a well known example is the popular neutralino model which requires at least three independent hypothesis : a ) broken low energy supersymmetry exists which provides wimp candidates , b ) an exact unbroken r - parity symmetry is proposed to prevent the lightest superpartner from decaying and c ) the lightest superpartner is hypothesised to be neutral to make it suitable for dark matter .
it seems to me that a more plausible candidate for this new physics arises from the hypothesis of exact unbroken mirror symmetry [ @xmath3 .
it is more plausible because it involves only a single well motivated hypothesis .
improper space - time symmetries , such as parity and time reversal symmetries , stand out as the only obvious symmetries which are not respected by the interactions of the known elementary particles .
it is an interesting and non - trivial fact that these symmetries can be exact , unbroken symmetries of nature if a set of mirror particles exist .
even more interesting is that the mirror particles have the right broad properties to be identified with the non - baryonic dark matter in the universe .
the ordinary and mirror particles form parallel sectors each with gauge symmetry @xmath4 ( where @xmath5 in the simplest case ) so that the full gauge group is @xmath6 .
mathematically , mirror symmetry has the form:@xcite @xmath7 where @xmath8 are the standard @xmath9 gauge particles , @xmath10 are the standard leptons and quarks ( @xmath11 is the generation index ) and the primes denote the mirror particles .
there is also a standard higgs doublet @xmath12 with a mirror higgs doublet partner , @xmath13 .
importantly , there is a large range of parameters of the higgs potential for which mirror symmetry is _ not _ spontaneously broken by the vacuum ( i.e. @xmath14 ) so that it is an exact , unbroken symmetry of the theory@xcite . ) , but the simplest models of this type are disfavoured for a variety of reasons@xcite .
[ more complicated models with broken mirror symmetry are still possible and have been studied in the literature , see e.g. ref.@xcite ] . ]
interestingly , despite doubling the number of particle types the number of free parameters have not ( yet ! ) been increased ; mirror symmetry implies that the masses and couplings of the particles in the mirror sector are exactly the same as the corresponding ones in the ordinary sector .
ordinary and mirror particles couple with each other via gravity and possibly by new interactions connecting ordinary and mirror particles together .
constraints from gauge invariance , mirror symmetry and renormalizability , suggest only two types of new interactions@xcite : a ) higgs - mirror higgs quartic coupling ( @xmath15 ) , and b ) via photon - mirror photon kinetic mixing : ) .
however , compared to other ideas beyond the standard model , many of which have literally hundreds of new parameters , mirror symmetry _ is _ a fairly minimal extension of the standard model .
also note , if the neutrinos have mass , mass mixing between ordinary and mirror neutrinos is also possible@xcite and might be implicated by the observed atmospheric , solar and lsnd neutrino anomalies .
however , the experimental situation is still not clear@xcite . ]
@xmath16 where @xmath17 ( @xmath18 ) is the field strength tensor for electromagnetism ( mirror electromagnetism ) .
gauge bosons , since only for abelian @xmath19 symmetry is the mixing term , @xmath20 , gauge invariant@xcite .
therefore there is both @xmath21 and @xmath22 kinetic mixing .
[ however , experiments are much more sensitive to @xmath23 kinetic mixing which is why it is more important ] . in the case of theories without @xmath19 gauge symmetries , such as guts , the @xmath21 mixing can arise radiatively provided that there exists a mixed form of matter carrying both ordinary and mirror electric charges@xcite .
interestingly , there is a class of such models@xcite where @xmath24 vanishes at one and two loop level , and therefore naturally of the order of @xmath25 . ]
the effect of the higgs - mirror higgs quartic coupling is to modify the properties of the standard higgs boson@xcite .
this interaction will be tested if / when scalar particles are discovered .
one effect of photon - mirror photon kinetic mixing is to cause mirror charged particles ( such as the mirror proton and mirror electron ) to couple to ordinary photons with effective electric charge @xmath26@xcite .
as we will see , this non - gravitational interaction between ordinary and mirror particles provides a key way to experimentally test the theory . to summarize : the only obvious space - time symmetries that are not respected by the interactions of the known elementary particles are the improper lorentz symmetries ( such as parity and time reversal ) .
these symmetries can be unbroken symmetries of nature provided that the universe contains both ordinary and mirror particles .
the mirror particles have identical masses to the corresponding ordinary particles and have identical , but separate gauge interactions ( the gauge group is @xmath27 ) .
the mirror particles couple to the ordinary ones via gravity and possibly via higgs - mirror higgs interactions and photon - mirror photon kinetic mixing .
it turns out that the mirror particles lead to an elegant explanation for the inferred non - baryonic dark matter component of the universe , as we will explain in more detail in the following sections
. 0.3 cm
0.3 cm there is a substantial range of evidence for non - baryonic dark matter in the universe .
we have already emphasised in the introduction that a basic requirement is that the massive particles comprising dark matter need to be stable and have no or small coupling to ordinary photons .
there are other desirable features that are also required , including ( in random order ) : * an explanation for @xmath28 .
* it should be capable of explaining the large scale structure of the universe .
* asymptotically flat rotation curves in spiral galaxies suggest that dark matter is ( roughly ) spherically distributed in a ` halo ' in spiral galaxies .
this is in contrast to ordinary matter which is distributed in the disk and bulge . * a substantial fraction ( @xmath29 ) of the mass of the halo appears to be in the form of massive ( @xmath30 ) compact invisible objects ( machos ) .
what are the machos and why are they invisible ? * the direct experimental detection of halo dark matter particles has been achieved by the dama / nai collaboration .
other experiments report only negative results .
we now examine each of these items from a mirror matter perspective .
precision cosmic microwave background measurements ( culminating with the recent wmap results@xcite ) have established that the universe is spatially flat , i.e. @xmath31 .
furthermore , the wmap results , together with observations of high redshift type 1a supernovae@xcite , and other measurements , suggest that the universe consists predominately of three components : ordinary matter ( @xmath32 ) , non - baryonic dark matter ( @xmath33 ) and dark energy ( @xmath34 ) .
it is striking that each of these three different components should have energy densities of the same order of magnitude .
since @xmath35 scales differently in time with @xmath36 and @xmath37 , the similarity between @xmath35 and @xmath38 might simply be a coincidence .
however , the similarity in magnitude of the ordinary and dark matter densities : @xmath39 is expected to be constant in time until a very early epoch .
this means that the amount of dark matter produced in the early universe is of the same order of magnitude as the ordinary matter , despite their apparent disparate properties .
the similarity in the abundances of ordinary and dark matter hints at an underlying similarity between the microscopic properties of the elementary particles comprising the ordinary matter and the dark matter .
clearly , the standard exotic weakly interacting dark matter scenarios offer no hope in explaining this cosmic coincidence because these particles have completely different properties ( different masses and interactions ) from the ordinary baryons . a priori
, a dark matter / ordinary matter ratio of , say , @xmath40 would appear to be equally likely in these scenarios .
however if dark matter is identified with mirror baryons , then it seems to be possible to explain the similarity of @xmath36 and @xmath37 because the microscopic properties of the mirror particles mirror those of the ordinary particles .
in fact , @xmath41 would occur if the initial conditions of the universe were also mirror symmetric and no macroscopic asymmetry ( such as a temperature difference ) was produced during the early evolution of the universe . however , the success of standard big bang nucleosynthesis ( bbn)@xcite does suggest that @xmath42 was somewhat less than @xmath43 during the bbn epoch , @xmath44 in order for the expansion rate of the universe to have been within an acceptable range then the energy density at the bbn epoch would be double the standard value significantly increasing the expansion rate of the universe at that time .
the equilibration of the three mirror neutrinos , mirror electron / positron and mirror photons is equivalent to an extra @xmath45 neutrino species .
more generally , @xmath46 , so that demanding @xmath47 would imply @xmath48 . ] .
if the temperatures are different , then this means that either the initial conditions of the universe were asymmetric or that the asymmetry was induced during the early evolution of the universe .
actually , within the inflation paradigm it is quite easy to generate the required temperature asymmetry@xcite . in particular , it is natural to have an ` ordinary inflaton ' coupling to ordinary matter , and a ` mirror inflaton ' coupling to mirror matter .
if inflation is triggered by some random fluctuation , then it can occur in the two sectors at different times , leading to @xmath49 after reheating in the two sectors .
in such a scenario , one expects the baryon number and mirror baryon number to be unequal ( since baryogenesis or leptogenesis depends on the temperature and expansion rate ) .
provided that the temperatures of the two sectors are not too different this might explain the fact that @xmath37 is within an order of magnitude of @xmath36 .
clearly , the details will depend on the precise model for baryogenesis used by nature , which is of course not known ( see the first paper of ref.@xcite for a couple of examples ) .
even if there is a large hierarchy in temperatures for the two sectors , similar abundances of ordinary and mirror particles can be achieved if there are interactions which can transfer lepton or baryon asymmetry between the two sectors@xcite .
the simplest such interaction is given by the dimension 5 operator : @xmath50 where @xmath51 is a left - handed ordinary lepton doublet , @xmath52 is its mirror partner , @xmath12 is the ordinary higgs doublet and @xmath13 is its mirror partner ( @xmath53 hermitian conjugate ) . with such operators and some plausible assumptions about the physics governing the early evolution of the universe ,
it is even possible@xcite to _ quantitatively _ explain the inferred non - baryonic dark matter proportion , @xmath54 .
we will not discuss much more about the very early universe ( the era prior to bbn ) .
these were the cosmic dark ages where much is speculated but little is known . a more recent development ( in the history of the universe ) was the formation of large scale structure , which is also the next topic . we know from measurements of the cosmic microwave background that the universe was extraordinarily homogeneous in the past .
at the present epoch , however , the universe is no longer particularly homogeneous : it contains galaxies , clusters of galaxies , superclusters etc .
this large scale structure is believed to arise from small primordial inhomogeneities that grow via gravitational instability .
however ordinary baryonic density perturbations can not begin to grow until photon decoupling occurs at a temperature of around @xmath55 ev , corresponding to a red shift of @xmath56 .
[ prior to photon decoupling , the radiation pressure prevents the growth of perturbations ] .
but this is too late : perturbations which have amplitude of order @xmath57 ( as inferred from the anisotropies of the cosmic microwave background ) do not have enough time to grow into galaxies , where @xmath58 .
this suggests that inhomogeneities begin to grow prior to photon decoupling .
this is one role that non - baryonic dark matter is expected to fill : it should be weakly coupled to the ordinary particles in the plasma so that density perturbation growth can begin prior to photon decoupling .
mirror particles are weakly coupled to the ordinary ones
. however , mirror baryonic density perturbations can only begin to grow after mirror photon decoupling occurs ( roughly when @xmath59 ev ) .
the key point is that if @xmath60 [ which we infer from bbn , see eq.([4 ] ) ] then mirror photon decoupling necessarily occurs _ earlier _ than ordinary photon decoupling .
thus , we expect that mirror baryonic structure formation should begin earlier than ordinary baryonic structure . according to refs.@xcite
, they find that for @xmath61 , large scale structure formation with mirror matter - type dark matter closely resembles the standard cold dark matter scenario ( but with some intriguing differences ) and would provide a successful framework to understand the large scale structure of the universe@xcite .
a consequence of @xmath60 , required for successful big bang nucleosynthesis and large scale structure formation , is that mirror bbn occurs earlier than ordinary bbn .
this will mean that the proportion of mirror helium ( @xmath62 ) to mirror hydrogen ( @xmath63 ) synthesised in the early universe will be different to their ordinary matter counterparts .
in fact , since the expansion rate of the universe is faster at earlier times the mirror neutron / mirror proton ratio should be closer to unity c.f . ordinary bbn .
this means that the @xmath64 ratio is expected to be significantly greater than the corresponding ordinary @xmath65 ratio@xcite .
this chemical imbalance between the ordinary and mirror sectors will no doubt have important effects .
for example , the initial mass function for mirror stars can be quite different than for ordinary stars .
ultimately , this chemical imbalance might even be responsible for the different distribution of ordinary and mirror matter in galaxies , as we will now discuss .
0.3 cm 0.3 cm we have briefly mentioned cosmological evidence for non - baryonic dark matter in sections 2.1 and 2.2 .
there is also strong astrophysical evidence for a large amount of dark matter in galaxies and galaxy clusters .
asymptotically flat rotation curves in spiral galaxies , illustrated in figure 1 , imply that there must exist invisible ` halos ' in galaxies such as our own milky way .
these halos are , roughly , spherical distributions of invisible matter which dominate the mass of the galaxy . for example , the mass of the invisible halo of the milky way galaxy is estimated to be @xmath66 , which is about an order of magnitude more than the estimated mass of the galactic disk component@xcite .
0.2 cm 0.8 cm figure 1 : the observed rotation curve for the the spiral galaxy m33 superimposed on its optical image .
[ figure from ref.@xcite ] .
1.2 cm there is strong evidence that this galactic halo dark matter must be something exotic ; ordinary baryons simply can not account for it@xcite . taking the case of white dwarfs as an example
, these would be dim enough to escape detection ( unless they are very young ) .
however , in the collapse process where they are formed the outer layers of the star are ejected into space .
this ejected material is rich in heavy elements such as oxygen and nitrogen .
if such material were present in the halo it would have been revealed from characteristic absorption / emission lines .
alternatively , if the ejected material were to collapse onto the galactic disk due to collisional processes , its estimated abundance would be greater than the entire mass of the disk .
thus , old ( ordinary ) white dwarfs can not provide a consistent picture for halo dark matter .
all other conventional candidates for galactic dark matter run into similar severe difficulties .
obviously a ( roughly ) spherical halo containing mirror stars , mirror planets , mirror dust and mirror gas would be much less problematic since any absorption / emission lines would be absent ( and @xmath67 given the fit of the dama / nai experiment , see later discussion in section 2.5 ) . ] .
of course , there is still the important problem of explaining the roughly spherical mirror matter distribution in the galaxy , with ordinary matter collapsed onto the disk . a priori this is possible : although ordinary and mirror matter have identical microscopic interactions , _ there is no macroscopic mirror symmetry_. recall this macroscopic asymmetry is necessary to explain a ) different abundance of ordinary and mirror matter in the universe ( @xmath68 , but @xmath69 ) and b ) the different temperatures of the ordinary and mirror sector ( in the early universe ) required by successful bbn and large scale structure ( as discussed earlier ) .
because of this macroscopic asymmetry , the evolution of the ordinary and mirror sectors can be significantly different . assuming that the halo is dominated by a mirror gas component
which is approximately spherical and isothermal , its distribution can be obtained from the condition of hydrostatic equilibrium : @xmath70 where @xmath71 is the pressure and @xmath72 is the local acceleration at a radius @xmath73 . for a dilute gas ,
the pressure is related to the mass density , @xmath74 , via @xmath75 , where @xmath76 is the average mass of the particles in the gas ( @xmath77 is the proton mass ) .
the local acceleration can be simply expressed in terms of the energy density , via : @xmath78 where @xmath4 is newton s constant .
the solution of eq.([pr],[fin ] ) is @xmath79 : @xmath80 the rotational velocity at a radial location @xmath81 , @xmath82 , can be obtained from @xmath83 which implies : @xmath84 clearly , @xmath85 , implied by a spherically symmetric self gravitating gas in hydrostatic equilibrium , gives the required asymptotically flat rotation curve ( a well - known result ) .
furthermore , from the above equation , @xmath86 , which means that we can express @xmath74 and @xmath43 in terms of @xmath87 : @xmath88 henceforth we focus on the milky way galaxy , for which @xmath89 km / s .
since @xmath43 is much greater than the ionization energy for the light mirror elements ( @xmath90 ) , these elements should be ionized .
it follows that bremsstrahlung and other processes will radiate off energy at a rate per unit volume of@xcite : @xmath91 where @xmath92 is the ( free ) mirror electron number density and @xmath93 is a calculable function ( which depends on cross section , temperature , composition etc ) . for a temperature of @xmath94 ev , @xmath95 ( see ref.@xcite for details ) .
note that @xmath96 ( for @xmath62 mass dominated halo ) , which implies [ using eq.([wed ] ) ] @xmath97 because @xmath98 , the total halo luminosity , @xmath99 is divergent as @xmath100 .
however , the inner region of the galaxy should contain a high density of mirror dust , stars , supernova , blackholes etc which will make things very complicated .
energy sources ( such as supernova ) can heat the inner region .
this effect , as well as the effect of mirror dust particles ( which can potentially make the inner region opaque to mirror radiation ) could potentially lead to an increasing temperature towards the galactic centre breaking the isothermal approximation .
this would be consistent with observations which imply that the rotation curves in spiral galaxies fall in the inner region ( as shown in the example of figure 1 ) , suggesting that the mass density is not increasing , but roughly constant there@xcite .
in other words , the observations themselves imply that the halo density appears to be heated up"@xcite , inexplicable in the standard cold dark matter scenario , but possible for mirror matter - type dark matter . in view of the above discussion , we introduce a phenomenological cutoff , @xmath101 , and consider only the energy produced at @xmath102 . in this case
, the energy radiated from the halo is roughly @xmath103 the above calculation assumes that the halo contains only a gas component . from general considerations , as well as specific evidence from microlensing studies ( as will be discussed in the following subsection ) , a significant component of the halo will be in the form of compact mirror objects : mirror stars , planets etc .
furthermore compact mirror objects can potentially dominate the mass in the inner regions of the galaxy which would alleviate the cooling problem to some extent .
still , a heat source of order @xmath104 erg / s seems to be required to compensate for the energy lost due to radiative cooling .
supernova offer promising candidate heat sources .
an obvious possibility is that mirror supernova can heat the halo ; during an explosion the outer layers of the star are ejected into interstellar space , with energy of order @xmath105 erg per explosion . in order to achieve a rate of around @xmath106 erg / s would require a mirror supernova rate in our galaxy of around one every few years , which is about an order of magnitude greater than the rate of ordinary supernova .
presumably this is possible given the uncertainties in the mirror sector .
for example , as discussed at the end of the previous subsection , it is possible that the initial mass function for mirror stars is very different to ordinary stars because of the different chemical composition ( different light element ratios etc ) .
this would mean that the rate of ordinary supernova could be quite different to mirror supernova .
a more subtle , but equally promising possible energy source might come from ordinary supernova explosions . due to the effects of photon - mirror photon kinetic mixing , eq.([km ] ) , in the core of the supernova a significant fraction , @xmath107 , of an ordinary supernova s total energy , @xmath108 , can be converted into mirror photons and mirror electrons / positrons , which can provide a significant heat source for the halo .
the amount of energy going into the halo from ordinary supernova explosions , is of order@xcite : particles which are potentially observable .
in fact , they may have already been observed in the form of gamma ray bursts and positron annihilation radiation from the galactic bulge@xcite ( this will be briefly reviewed in section 3.1 ) . ]
@xmath109 evidently , ordinary supernova s can potentially supply about the right amount of energy to replace the energy lost in radiative cooling , if ordinary supernova s occur at a rate , @xmath110 , of order once per hundred years and of order @xmath111 of the supernova s energy is converted into mirror @xmath112 .
presumably the heating of the mirror sector and ordinary sector needs to be different in order to explain why ordinary matter has collapsed onto the disk and mirror matter has not .
this is not impossible given the lack of macroscopic mirror symmetry , leading to e.g. asymmetric ordinary and mirror supernova rates .
it seems therefore that asymmetric heating of the ordinary and mirror sectors is feasible and we conclude that mirror matter - type dark matter is capable of explaining the dark matter halo in spiral galaxies . in one sense the existence of a dark halo is more directly explained within the standard wimp paradigm .
wimps being collisionless are non - dissipative and could not collapse onto the disk . however this cure has serious side - effects which are potentially fatal .
wimps being collisionless particles are relatively simple , macroscopically , and their distribution can be predicted .
the result is a dark matter density profile that goes like@xcite @xmath113 , with @xmath114 .
this prediction is in clear disagreement with the observations ( see e.g. ref.@xcite ) . in other words , the standard wimp paradigm can simply explain the existence of dark matter in galaxies , but fails to explain the detailed distribution of dark matter within the halo .
this ` fact ' seems to support the idea that the dark matter is , macroscopically , more complicated than collisionless wimps . more evidence for the complexity of the dark halo is provided by microlensing surveys of stars in nearby galaxies , which brings us to the next topic .
if non - baryonic dark matter is identified with mirror matter , then a substantial fraction of the non - baryonic dark matter should be in the form of compact bodies such as mirror stars .
this leads naturally to an explanation@xcite for the mysterious massive astrophysical compact halo objects ( or macho s ) discovered by the macho collaboration .
the macho collaboration@xcite has been studying the nature of halo dark matter with the gravitational microlensing technique@xcite , using source stars in the large magellanic cloud .
this australian - american experiment has collected 5.7 years of data and provided statistically strong evidence for dark matter in the form of invisible star sized objects which is what you would expect if there is a significant amount of mirror matter in our galaxy .
the macho collaboration@xcite has done a maximum likelihood analysis which implies a macho halo fraction of @xmath115 for a typical halo model with a @xmath116 confidence interval of @xmath117 their most likely macho mass is between @xmath118 and @xmath119 depending on the halo model . on the other hand
, the eros team@xcite studying microlensing towards the small magellanic cloud did not find evidence for a significant population of compact halo objects .
this yielded a constraint which was however consistent with the @xmath115 halo mass fraction reported by the macho collaboration .
more recently , a new survey@xcite has begun examining stars across the face of m31 .
they find significant evidence for a population of halo microlensing dark matter objects , inferring a halo mass fraction of @xmath120 .
this result is consistent with the positive results of the macho collaboration and provides important independent confirmation of their positive signal .
furthermore , they find significant evidence for an asymmetry in the distribution of microlensing events across the face of m31 , which is expected if their events are correctly interpreted as a large population of invisible massive compact halo objects .
it is important to realize that the inferred macho halo fraction , @xmath115 , is consistent with a mirror matter halo ; the entire halo need not be in the form of mirror stars .
mirror gas and dust would also be expected as they are a necessary consequence of stellar evolution and can significantly populate the halo .
as we have just seen , the results of the microlensing surveys suggest a macho halo fraction of @xmath115 . within the mirror matter theory
, this gets a natural interpretation in terms of mirror stars , mirror white dwarfs etc . the remaining fraction
, @xmath121 should presumably be in the form of mirror gas . assuming a roughly spherical and isothermal distribution for this ionized gas
, it would have a mass density ( at our location , @xmath122 kpc ) and temperature [ using eq.([wed ] ] : @xmath123 considering a particular chemical element , @xmath124 ( e.g. @xmath125 etc ) , the velocity distribution for these halo mirror particles is then : @xmath126 \nonumber \\ & \equiv & exp[-v^2/v_0 ^ 2]\ , \label{fr5}\end{aligned}\ ] ] where @xmath127 . using eq.([wed2 ] ) , we have : @xmath128 recall , @xmath76 is the mean mass of the particles comprising the mirror ( gas ) component of the halo and @xmath129 km / s is the local rotational velocity .
evidently the characteristic velocity , @xmath130 , for a particular halo component element , @xmath124 , depends on the chemical composition of the halo ( through the dependence on @xmath76 ) .
mirror bbn will generate @xmath90 and , possibly , heavier mirror elements as well , quite unlike the ordinary matter case . and also
if @xmath131 , then the number density of mirror nucleons present during the mirror bbn epoch can be several orders of magnitude greater than the number density of ordinary nucleons at the time of ordinary bbn .
the greater mirror nucleon number density can dramatically increase the rate of three - body processes such as the triple alpha process .
thus , it seems to be an interesting possibility that a significant abundance of heavy mirror elements ( such as @xmath132 ) could be generated in the early universe . ]
heavy mirror elements can also be generated in mirror stars . in any case , we consider two representative possibilities : first that the mass of the halo is dominated by @xmath62 and the second is that the halo is dominated by @xmath63 .
the mean mass of the particles in the halo are then ( taking into account that the light halo mirror atoms should be fully ionized ) : @xmath133 the @xmath134 values can then be easily obtained from eq.([z3 ] ) : @xmath135 it is important to realize that halo atoms can potentially be detected in conventional dark matter experiments via the nuclear recoil signature@xcite .
the reason is that the photon - mirror photon kinetic mixing interaction , eq.([km ] ) , gives the mirror nucleus , with ( mirror ) atomic number @xmath136 , a small effective ordinary electric charge of @xmath137 .
this means that ordinary and mirror nuclei can elastically scatter off each other ( essentially rutherford scattering ) .
the basic feynman diagram for this process is given in the figure 2 ( on the following page ) . for a mirror atom of mass @xmath138 and ( mirror ) atomic number @xmath136 scattering on an ordinary target atom of mass @xmath139 and atomic number @xmath140 ,
the differential cross section is given by : @xmath141 where @xmath142 ( the @xmath143 are the nuclear form factors ) .
in eq.([cs ] ) @xmath144 is the velocity of the mirror nucleus relative to the earth and @xmath145 is the recoil energy of the ordinary ( target ) nucleus .
0.7 cm 0.3 cm figure 2 : ordinary and mirror nuclei can elastically scatter via the photon - mirror photon kinetic mixing interaction ( indicated by a ` cross ' in this feynman diagram ) . 0.9 cm in dark matter direct detection experiments ( such as dama / nai@xcite ) , the measured quantity is the recoil energy , @xmath145 , of the target nucleus .
the differential interaction rate is @xmath146 where @xmath147 is the number of target atoms per kg of detector ( for detectors with more than one target element we must work out the interaction rate for each element separately and add them up to get the total interaction rate ) . in the above equation @xmath148
is the velocity distribution of the mirror element ( @xmath149 is the normalization factor ) , @xmath124 , and @xmath150 is the earth velocity relative to the galaxy .
since @xmath151 is the velocity of the mirror particles relative to the galaxy , it follows from eq.([fr5 ] ) , that @xmath152 $ ] . in eq.([55 ] ) the lower velocity limit , @xmath153 , is the minimum velocity for which a mirror atom of mass @xmath138 impacting on a target atom of mass @xmath154 can produce a recoil energy of @xmath145 for the target atom .
this minimum velocity satisfies the kinematic relation : @xmath155 interestingly , most of the existing dark matter experiments are not very sensitive to mirror matter - type dark matter because @xmath156 [ eq.([v ] ) ] turns out to be too high .
this is because they either use target elements which are too heavy ( i.e. large @xmath154 ) and/or have a @xmath145 threshold which is too high .
the dark matter experiment most sensitive to halo mirror matter - type dark matter is the dama / nai experiment@xcite .
the aim of the dama / nai experiment is to measure the nuclear recoils of @xmath157 atoms due to the interactions of halo dark matter particles . because of the dependence of the interaction rate , eq.([55 ] ) , on @xmath150 , the interaction rate of halo dark matter with a detector depends on the earth s velocity relative to the halo . because of the earth s annual motion , its velocity satisfies : @xmath158 where @xmath159 km / s is the sun s velocity with respect to the galaxy and @xmath160 km / s is the earth s orbital velocity around the sun ( @xmath161 , with @xmath162 year ) . the inclination of the earth s orbital plane relative to the galactic plane is @xmath163 , which means that @xmath164 km / s . thus , the differential interaction rate in an experiment will thus contain an annual modulation term : @xmath165 where @xmath166 according to the dama analysis@xcite , they indeed find an annual modulation over 7 annual cycles at more than @xmath167 c.l .
their data fit gives @xmath168 year and @xmath169 , consistent with the expected values .
[ the expected value for @xmath170 is 152 ( 2 june ) , where the earth s velocity , @xmath150 , reaches a maximum with respect to the galaxy ] .
their signal occurs in the energy range @xmath171 kevee , which is related to the actual energy , @xmath145 , by @xmath172 , where @xmath173 is the quenching factor corresponding to a given target element , @xmath174 . for the dama / nai experiment , @xmath175.@xcite ] and
the amplitude of their signal is @xmath176 cpd / kg / kevee [ cpd @xmath177 counts per day ] .
these are extremely impressive results .
the self consistency of their signal is highly non - trivial : there is simply no reason why their data should contain a periodic modulation or why it should peak near june 2 .
in fact , the known systematic errors are several orders of magnitude too small to account for the signal@xcite
. it therefore seems probable that dama has indeed discovered dark matter . interestingly
the interpretation of the dama / nai signal in terms of standard wimps appears to be disfavoured by a number of experiments@xcite , the most impressive of which is the recent null cdmsii / ge results@xcite .
however , if we interpret the dama / nai signal in terms of mirror matter - type dark matter then the positive dama / nai signal and the negative results of the other experiments can be reconciled@xcite .
the dama experiment is not particularly sensitive to very light dark matter particles such as mirror hydrogen and mirror helium .
impacts of these elements ( typically ) do not transfer enough energy to give a signal above the detection threshold@xcite .
if stellar nucleosynthesis in the mirror sector is sufficiently similar to the ordinary sector , then the next most abundant element should be mirror oxygen . in the analysis of ref.@xcite
the spectrum of heavy mirror elements were approximated by just two components , @xmath178 .
naturally this is just a crude approximation : in general there will be a distribution of mirror elements which is very difficult to theoretically predict ( because of e.g. unknown initial mass function for mirror stars etc ) .
of course , it may turn out that approximating the spectrum in terms of the two components , @xmath178 , will be insufficient in the future as more detailed data is obtained . anyway , interpreting the dama / nai annual modulation signal in terms of @xmath178 , it was found that@xcite : @xmath179 where the errors denote a 3 sigma allowed range and @xmath180 is the @xmath124 proportion ( by mass ) of the halo dark matter ( at the earth s location ) .
suggested by the dama experiment , eq.([dama55 ] ) , would also have important implications for the orthopositronium system@xcite .
the current experimental situation , summarized in ref.@xcite , implies that @xmath181 , which is easily consistent with eq.([dama55 ] ) .
importantly , a new orthopositronium experiment has been proposed@xcite which can potentially cover much of the @xmath24 parameter space suggested by the dama experiment .
such an experiment is very important not just as a check of the mirror matter explanation but also because dark matter experiments are sensitive to @xmath182 and an independent measurement of @xmath24 would allow the extraction of @xmath183 values . ]
this fit to the data is shown in figure 3 .
-0.1 cm 0.2 cm figure 3 : dama / nai annual modulation signal ( taking data from the second paper of ref.@xcite ) together with the mirror matter prediction ( initial time is august @xmath184 ) .
1.1 cm in ref.@xcite it was found that a dama / nai annual modulation signal dominated by an @xmath185 component , is experimentally disfavoured for three independent reasons : a ) it predicts a mean differential energy spectrum rate larger than the measured dama / nai rate b ) potentially leads to a significant diurnal effect ( sidereal daily variation ) and c ) should have been observed in the cdmsi experiment .
thus it is probable that lighter mirror elements , such as a mirror oxygen component , dominates the dama annual modulation signal . from eq.([dama55 ] ) this means that @xmath186 .
recently , a more stringent limit of @xmath187 was obtained in ref.@xcite using the recent cdmsii / ge result@xcite .
if the dama signal is dominated by @xmath188 , then things depend on only one parameter , @xmath189 .
this parameter is fixed from the annual modulation signal , eq.([dama55 ] ) , which means that the event rate ( due to @xmath188 interactions ) can be predicted for other experiments .
it turns out that , with the exception of one experiment ( cressti@xcite ) , all of the other experiments are not sensitive to @xmath188 interactions .
for example , the predicted rate for cdmsii / ge due to @xmath188 interactions is given in figure 4 ( taken from ref.@xcite ) .
0.3 cm 0.4 cm figure 4 : predicted differential event rate , @xmath190 , ( binned into 10 kev bins ) due to @xmath188 dark matter with @xmath191 ( dama / nai annual modulation best fit ) for the cdmsii / ge experiment .
the solid line corresponds to a standard halo model with @xmath62 dominated halo while the dashed line assumes a @xmath63 dominated halo .
1.4 cm as the figure shows , the event rate is predicted to be very low . for the cdmsii / ge experiment
the predicted event rate is just 1 event per @xmath192 kg - days for @xmath62 dominant halo and 1 event per @xmath193 kg - days if @xmath63 dominates the halo . given that cdmsii has only 52.6 kg - day raw exposure in ge , this implies a predicted number of events of just @xmath194 ( assuming @xmath62 dominant halo ) and even less if @xmath63 dominates the mass of the halo .
clearly this prediction is nicely consistent with the null result of cdmsii / ge , it is much more sensitive to interactions of heavier mirror elements such as @xmath185 .
future cdms data may well find a positive signal due to these heavier elements because they should be there at some level . ] . in the case of standard spin independent wimps , the cdmsii experiment is more sensitive than the dama / nai experiment .
however , as we have discussed above , this is clearly not the case for @xmath188-type dark matter ( with dominant @xmath64 component ) .
the diverse behaviour of the two types of dark matter candidate has to do with their basic differences : * the mass of @xmath188 is only 15 gev , while standard wimps are typically assumed to have masses which are greater than @xmath195 gev ( depending on the model ) . * for @xmath188-type
dark matter in an @xmath64 dominated halo , @xmath196 km / s [ eq.([19y ] ) ] , while the characteristic velocity of standard wimps are assumed to be approximately 220 km / s .
* the differential cross section for mirror matter - type dark matter is inversely proportional to the square of the recoil energy , while that for standard wimps is energy independent ( excepting the energy dependence of the form factors ) .
these three key differences mean that experiments with low threshold energy and light target elements are much more sensitive ( to @xmath188-type dark matter ) than experiments with higher threshold energy and/or heavy target elements . in the case of dama / nai ,
the event rate for mirror matter - type dark matter is dominated by interactions with the light @xmath197 component .
the actual threshold energy of 6.7 kev ( for @xmath197 ) , implies a threshold impact velocity , obtained from eq.([v ] ) , of 290 km / s for @xmath188 impacts . in the case of cdmsii
/ ge , the threshold energy of 10 kev and heavy ge target gives a threshold impact velocity of 450 km / s ( see ref.@xcite for a table of threshold velocities for the various experiments ) .
given the low value of @xmath198 [ using eq.([19y ] ) , @xmath199 km / s ( @xmath200 km / s ) for @xmath62 ( @xmath63 ) dominated halo ] the number of @xmath188 atoms with impact velocity above threshold is clearly much lower for cdmsii / ge compared with dama / nai ( in fact it is exponentially suppressed ) .
note that the edelweiss i / ge ( ref.@xcite ) and zeplin i / xe ( ref.@xcite ) experiments are even less sensitive than cdmsii / ge because the threshold impact velocity of those experiments is even higher@xcite .
there is one experiment , besides dama / nai , which was potentially sensitive to mirror matter interactions , namely the cresst i experiment@xcite .
that experiment had a target consisting of sapphire crystals ( @xmath201 ) , with a low recoil energy threshold of @xmath202 kev
. however , the results of that experiment turned out to be roughly consistent with the mirror matter prediction ( i.e. with parameters fixed by the dama / nai annual modulation signal)@xcite , providing tentative support for the mirror matter interpretation of the dama / nai experiment .
unfortunately , this experiment did not collect enough data to do an annual modulation analysis ( it has now been discontinued , replaced by a new experiment , cresst ii , which will use a @xmath203 target , and has an expected threshold energy of 10 kev , which will be less sensitive than cresst i , but should still be useful )
. 0.3 cm
0.3 cm in section 2 we have examined the conventional cosmological and astrophysical implications of mirror matter - type dark matter including direct experimental evidence from the dama / nai experiment . however mirror matter - type dark matter is an unconventional dark matter candidate with numerous unconventional implications .
included among these is the possibility of binary ordinary / mirror systems , possible manifestations of mirror matter in our solar system , implications for supernova etc .
we now briefly examine some of these applications .
photon - mirror photon kinetic mixing , eq.([km ] ) , of magnitude @xmath204 ( as suggested by the dama annual modulation signal ) will lead to important implications for core collapse supernova both ordinary and mirror types .
recall , that in the core of ordinary supernova , the temperature reaches , @xmath205 mev , leading to a plasma of @xmath206 ( @xmath207 ) . because of the photon - mirror photon kinetic mixing , mirror @xmath208 can also be produced via a variety of processes ( the most obvious being @xmath209 ) .
actually the main production process for mirror particles in the core of a mirror supernova is expected to be the plasmon decay process ( see e.g. ref.@xcite for a review ) .
the energy loss rate for production of minicharged particles has been calculated in ref.@xcite : @xmath210 where @xmath211 is a factor of order unity .
@xmath212 is comparable to the energy loss rate due to neutrino emission for@xcite @xmath213 thus , the production of mirror particles in the core of ordinary supernova must lead to important effects as a significant part of the emission of ordinary supernova s will be in the form of @xmath214
. one of these effects is that the @xmath215 produced in the core will help supernova s to explode as we will now explain .
supernova explosions of massive stars are believed to be driven by the convectivelly supported neutrino - heating mechanism@xcite .
but refined simulations have shown@xcite that there is insufficient neutrino energy transfer behind the stalled supernova shock to produce the explosion .
this is actually a long standing problem in supernova dynamics .
it suggests some missing piece of physics , which might well be photon - mirror photon kinetic mixing : the @xmath215 produced in the core will interact and heat the matter behind the shock ( adding to the effect of neutrino - heating ) thereby producing the explosion . for this to be possible
we require that the cross section for mev @xmath216 ( and/or large angle @xmath208 ) scattering with ordinary electrons ( i.e. @xmath217 and @xmath218 ) to be of roughly the same magnitude as the neutrino nucleon cross section .
the mirror particle cross section is : @xmath219 where @xmath220 is the classical radius of the electron .
the neutrino nucleon cross section is @xmath221 where @xmath222 is the energy of the neutrino .
evidently the cross sections for the two completely different processes are indeed comparable !
importantly , the energy dependence is different : compared with neutrino interactions , the mirror particle interactions with ordinary matter are larger at lower energies .
it follows that the heating effect of the mirror particle interactions on the ordinary matter just behind the shock is expected to be comparable to or may even exceed the neutrino effect .
a significant portion of the @xmath215 will escape the supernova , however direct detection of these particles seems to be very difficult for ordinary matter observers .
even if we can not directly detect this emission it does not mean that it is unimportant ; as we discussed earlier in section 2.3 , these mirror particles may have an important role in heating the galactic halo to compensate for the energy lost due to radiative cooling . in the case of a _ mirror _
type ii supernova is also very interesting . in this case
, the core of the mirror supernova would be a source of ordinary electrons , positrons and gamma rays making such an event easily detectable for ordinary matter observers .
in fact , they may have already been detected !
provided that the number of ordinary baryons is sufficiently low the @xmath223 ` fireball ' will lead to a gamma ray burst ( grb)@xcite .
of course , grb s have been observed for some time , and their origin has been a long standing puzzle .
it is certainly interesting that the mirror supernova , with photon - mirror photon kinetic mixing interaction has roughly the right characteristics ( energy release , time scale , and potentially small baryon load ) to be identified with the observed gamma ray bursts .
in addition to being a source of photons , grb will also eject electrons and positrons into the interstellar medium .
this might explain@xcite the 511 kev photon emission from the galactic bulge .
this emission was first detected more than 30 years ago@xcite and studied in a number of experiments culminating in the recent integral - spi measurements@xcite . while grbs and galactic positron emission are certainly rather spectacular possible manifestations of the mirror world , something even more tantalizing would be the discovery of a mirror world itself .
if mirror matter exists in our galaxy , then binary systems consisting of ordinary and mirror matter should also exist .
while systems containing approximately equal amounts of ordinary and mirror matter are very unlikely due to e.g. differing rates of collapse for ordinary and mirror matter ( due to different initial conditions such as chemical composition , temperature distribution etc ) , systems containing predominately ordinary matter with a small amount of mirror matter ( and vice versa ) should exist .
remarkably , there is interesting evidence for the existence of such systems coming from extra - solar planet astronomy . in 1995 , the first planet orbiting another star was discovered@xcite . since that time the field of extra - solar planet astronomy has been moving at a rapid pace . to - date
, more than 100 extra - solar planets have been discovered orbiting nearby stars@xcite .
they reveal their presence because their gravity tugs periodically on their parent stars leading to observable doppler shifts .
several transiting planets have been observed allowing for an accurate determination of the planet s size and mass in those systems .
one of the surprising characteristics of the extrasolar planets is that there are a class of large ( @xmath224 ) close - in planets with a typical orbital radius of @xmath225 ( which is about eight times closer than the orbital radius of mercury ) .
ordinary ( gas giant ) planets are not expected to form close to stars because the high temperatures do not allow them to form .
theories have been invented where they form far from the star where the temperature is much lower , and migrate towards the star@xcite .
a fascinating alternative possibility presents itself in the mirror world hypothesis .
the close - in planets may be mirror worlds composed predominately of mirror matter@xcite .
they do not migrate significantly , but actually formed close to the star which is not a problem for mirror worlds because they are not significantly heated by the radiation from the star .
this hypothesis can potentially explain the opacity of transiting planets because mirror worlds would accrete ordinary matter from the solar wind which accumulates in the gravitational potential well of the mirror world .
it turns out that the effective radius , @xmath226 at which the planet becomes opaque to ordinary radiation depends sensitively on the mass of the planet , with ref.@xcite providing a prediction : @xmath227 where @xmath228 is the surface temperature of the planet and @xmath229 is the mass of the planet .
this was only a rough prediction ( especially the dependence on @xmath228 ) but a prediction nevertheless .
heuristically it is very easy to understand : increasing the planet s mass increases the force of gravity which causes the gas of ordinary matter to become more tightly bound to the mirror planet ( thereby decreasing the effective size , @xmath226 ) , while increasing the temperature of the gas increases the volume that the gas occupies ( thereby increasing @xmath226 ) .
of these two effects we expect that the dependence on @xmath229 should be the more robust prediction . because the size of ordinary gas giant planets ( i.e. planets made mostly of ordinary matter ) depends quite weakly on their mass , the dependence on @xmath229 which is significant according to eq.([1new ] ) should allow a decisive test of the mirror planet hypothesis .
there are currently four extrasolar planets for which measurements of @xmath226 and @xmath229 are available : hd209458b@xcite , ogle - tr-56b@xcite , ogle - tr-113b@xcite and ogle - tr-132b@xcite .
we summarize their properties in the following table : -0.1 cm -0.15 cm these measurements ( ignoring ogle - tr-132b because of its huge uncertainty in @xmath226 ) , together with the 2001 prediction , eq.([1new ] ) , are shown in figure 5 ( from ref.@xcite ) .
the solid line is the prediction , eq.([1new ] ) , where we have used hd209458b to fix the proportionality constant .
0.4 cm ( 1500,900)(0,0 ) = cmr10 at 10pt ( 181.0,123.0 ) ' '' '' ( 161,123)(0,0)[r ] 0 ( 1419.0,123.0 ) ' '' '' ( 181.0,215.0 ) ' '' '' ( 161,215)(0,0)[r ] 0.2 ( 1419.0,215.0 ) ' '' '' ( 181.0,307.0 ) ' '' '' ( 161,307)(0,0)[r ] 0.4 ( 1419.0,307.0 ) ' '' '' ( 181.0,399.0 ) ' '' '' ( 161,399)(0,0)[r ] 0.6 ( 1419.0,399.0 ) ' '' '' ( 181.0,492.0 ) ' '' '' ( 161,492)(0,0)[r ] 0.8 ( 1419.0,492.0 ) ' '' '' ( 181.0,584.0 ) ' '' '' ( 161,584)(0,0)[r ] 1 ( 1419.0,584.0 ) ' '' '' ( 181.0,676.0 ) ' '' '' ( 161,676)(0,0)[r ] 1.2 ( 1419.0,676.0 ) ' '' '' ( 181.0,768.0 ) ' '' '' ( 161,768)(0,0)[r ] 1.4 ( 1419.0,768.0 ) ' '' '' ( 181.0,860.0 ) ' '' '' ( 161,860)(0,0)[r ] 1.6 ( 1419.0,860.0 ) ' '' '' ( 181.0,123.0 ) ' '' '' ( 181,82)(0,0 ) 0 ( 181.0,840.0 ) ' '' '' ( 410.0,123.0 ) ' '' '' ( 410,82)(0,0 ) 0.2 ( 410.0,840.0 ) ' '' '' ( 638.0,123.0 ) ' '' '' ( 638,82)(0,0 ) 0.4 ( 638.0,840.0 ) ' '' '' ( 867.0,123.0 ) ' '' '' ( 867,82)(0,0 ) 0.6 ( 867.0,840.0 ) ' '' '' ( 1096.0,123.0 ) ' '' '' ( 1096,82)(0,0 ) 0.8 ( 1096.0,840.0 ) ' '' '' ( 1325.0,123.0 ) ' '' '' ( 1325,82)(0,0 ) 1 ( 1325.0,840.0 ) ' '' '' ( 181.0,123.0 ) ' '' '' ( 1439.0,123.0 ) ' '' '' ( 181.0,860.0 ) ' '' '' ( 29,595)(0,0)@xmath230 $ ] ( 810,21)(0,0)@xmath231 ( 181.0,123.0 ) ' '' '' ( 1325.0,759.0 ) ' '' '' ( 1315.0,759.0 ) ' '' '' ( 1315.0,805.0 ) ' '' '' ( 1090.0,616.0 ) ' '' '' ( 1080.0,616.0 ) ' '' '' ( 1080.0,763.0 ) ' '' '' ( 950.0,588.0 ) ' '' '' ( 940.0,588.0 ) ' '' '' ( 940.0,653.0 ) ' '' '' ( 1285.0,782.0 ) ' '' '' ( 1285.0,772.0 ) ' '' '' ( 1365.0,772.0 ) ' '' '' ( 1033.0,690.0 ) ' '' '' ( 1033.0,680.0 ) ' '' '' ( 1147.0,680.0 ) ' '' '' ( 866.0,620.0 ) ' '' '' ( 866.0,610.0 ) ' '' '' ( 1033.0,610.0 ) ' '' '' ( 181.0,123.0 ) ' '' '' ( 161,123)(0,0)[r ] 0 ( 1419.0,123.0 ) ' '' '' ( 181.0,215.0 ) ' '' '' ( 161,215)(0,0)[r ] 0.2 ( 1419.0,215.0 ) ' '' '' ( 181.0,307.0 ) ' '' '' ( 161,307)(0,0)[r ] 0.4 ( 1419.0,307.0 ) ' '' '' ( 181.0,399.0 ) ' '' '' ( 161,399)(0,0)[r ] 0.6 ( 1419.0,399.0 ) ' '' '' ( 181.0,492.0 ) ' '' '' ( 161,492)(0,0)[r ] 0.8 ( 1419.0,492.0 ) ' '' '' ( 181.0,584.0 ) ' '' '' ( 161,584)(0,0)[r ] 1 ( 1419.0,584.0 ) ' '' '' ( 181.0,676.0 ) ' '' '' ( 161,676)(0,0)[r ] 1.2 ( 1419.0,676.0 ) ' '' '' ( 181.0,768.0 ) ' '' '' ( 161,768)(0,0)[r ] 1.4 ( 1419.0,768.0 ) ' '' '' ( 181.0,860.0 ) ' '' '' ( 161,860)(0,0)[r ] 1.6 ( 1419.0,860.0 ) ' '' '' ( 181.0,123.0 ) ' '' '' ( 181,82)(0,0 ) 0 ( 181.0,840.0 ) ' '' '' ( 410.0,123.0 ) ' '' '' ( 410,82)(0,0 ) 0.2 ( 410.0,840.0 ) ' '' '' ( 638.0,123.0 ) ' '' '' ( 638,82)(0,0 ) 0.4 ( 638.0,840.0 ) ' '' '' ( 867.0,123.0 ) ' '' '' ( 867,82)(0,0 ) 0.6 ( 867.0,840.0 ) ' '' '' ( 1096.0,123.0 ) ' '' '' ( 1096,82)(0,0 ) 0.8 ( 1096.0,840.0 ) ' '' '' ( 1325.0,123.0 ) ' '' '' ( 1325,82)(0,0 ) 1 ( 1325.0,840.0 ) ' '' '' ( 181.0,123.0 ) ' '' '' ( 1439.0,123.0 ) ' '' '' ( 181.0,860.0 ) ' '' '' ( 29,595)(0,0)@xmath230 $ ] ( 810,21)(0,0)@xmath231 ( 181.0,123.0 ) ' '' '' ( 181,123 ) ( 181.00,123.59)(0.950,0.485)11 ' '' '' ( 181.00,122.17)(11.251,7.000)2 ' '' '' ( 194.00,130.59)(0.758,0.488)13 ' '' '' ( 194.00,129.17)(10.547,8.000)2 ' '' '' ( 206.00,138.59)(0.950,0.485)11 ' '' ''
( 206.00,137.17)(11.251,7.000)2 ' '' '' ( 219.00,145.59)(0.950,0.485)11 ' '' '' ( 219.00,144.17)(11.251,7.000)2 ' '' '' ( 232.00,152.59)(0.824,0.488)13 ' '' '' ( 232.00,151.17)(11.443,8.000)2 ' '' '' ( 245.00,160.59)(0.874,0.485)11 ' '' '' ( 245.00,159.17)(10.369,7.000)2 ' '' ''
( 257.00,167.59)(0.950,0.485)11 ' '' '' ( 257.00,166.17)(11.251,7.000)2 ' '' '' ( 270.00,174.59)(0.824,0.488)13 ' '' '' ( 270.00,173.17)(11.443,8.000)2 ' '' '' ( 283.00,182.59)(0.874,0.485)11 ' '' ''
( 283.00,181.17)(10.369,7.000)2 ' '' '' ( 295.00,189.59)(0.950,0.485)11 ' '' '' ( 295.00,188.17)(11.251,7.000)2 ' '' '' ( 308.00,196.59)(0.824,0.488)13 ' '' '' ( 308.00,195.17)(11.443,8.000)2 ' '' '' ( 321.00,204.59)(0.874,0.485)11 ' '' '' ( 321.00,203.17)(10.369,7.000)2 ' '' '' ( 333.00,211.59)(0.950,0.485)11 ' '' '' ( 333.00,210.17)(11.251,7.000)2 ' '' '' ( 346.00,218.59)(0.950,0.485)11 ' '' '' ( 346.00,217.17)(11.251,7.000)2 ' '' '' ( 359.00,225.59)(0.824,0.488)13 ' '' '' ( 359.00,224.17)(11.443,8.000)2 ' '' '' ( 372.00,233.59)(0.874,0.485)11 ' '' '' ( 372.00,232.17)(10.369,7.000)2 ' '' '' ( 384.00,240.59)(0.950,0.485)11 ' '' '' ( 384.00,239.17)(11.251,7.000)2 ' '' '' ( 397.00,247.59)(0.824,0.488)13 ' '' '' ( 397.00,246.17)(11.443,8.000)2 ' '' '' ( 410.00,255.59)(0.874,0.485)11 ' '' '' ( 410.00,254.17)(10.369,7.000)2 ' '' '' ( 422.00,262.59)(0.950,0.485)11 ' '' '' ( 422.00,261.17)(11.251,7.000)2 ' '' '' ( 435.00,269.59)(0.824,0.488)13 ' '' '' ( 435.00,268.17)(11.443,8.000)2 ' '' '' ( 448.00,277.59)(0.950,0.485)11 ' '' '' ( 448.00,276.17)(11.251,7.000)2 ' '' '' ( 461.00,284.59)(0.874,0.485)11 ' '' '' ( 461.00,283.17)(10.369,7.000)2 ' '' '' ( 473.00,291.59)(0.824,0.488)13 ' '' '' ( 473.00,290.17)(11.443,8.000)2 ' '' '' ( 486.00,299.59)(0.950,0.485)11 ' '' '' ( 486.00,298.17)(11.251,7.000)2 ' '' '' ( 499.00,306.59)(0.874,0.485)11 ' '' '' ( 499.00,305.17)(10.369,7.000)2 ' '' '' ( 511.00,313.59)(0.824,0.488)13 ' '' '' ( 511.00,312.17)(11.443,8.000)2 ' '' '' ( 524.00,321.59)(0.950,0.485)11 ' '' '' ( 524.00,320.17)(11.251,7.000)2 ' '' '' ( 537.00,328.59)(0.950,0.485)11 ' '' '' ( 537.00,327.17)(11.251,7.000)2 ' '' '' ( 550.00,335.59)(0.758,0.488)13 ' '' ''
( 550.00,334.17)(10.547,8.000)2 ' '' '' ( 562.00,343.59)(0.950,0.485)11 ' '' '' ( 562.00,342.17)(11.251,7.000)2 ' '' '' ( 575.00,350.59)(0.950,0.485)11 ' '' '' ( 575.00,349.17)(11.251,7.000)2 ' '' '' ( 588.00,357.59)(0.758,0.488)13 ' '' '' ( 588.00,356.17)(10.547,8.000)2 ' '' '' ( 600.00,365.59)(0.950,0.485)11 ' '' '' ( 600.00,364.17)(11.251,7.000)2 ' '' '' ( 613.00,372.59)(0.950,0.485)11 ' '' '' ( 613.00,371.17)(11.251,7.000)2 ' '' '' ( 626.00,379.59)(0.874,0.485)11 ' '' '' ( 626.00,378.17)(10.369,7.000)2 ' '' '' ( 638.00,386.59)(0.824,0.488)13 ' '' '' ( 638.00,385.17)(11.443,8.000)2 ' '' '' ( 651.00,394.59)(0.950,0.485)11 ' '' '' ( 651.00,393.17)(11.251,7.000)2 ' '' '' ( 664.00,401.59)(0.950,0.485)11 ' '' '' ( 664.00,400.17)(11.251,7.000)2 ' '' '' ( 677.00,408.59)(0.758,0.488)13 ' '' '' ( 677.00,407.17)(10.547,8.000)2 ' '' '' ( 689.00,416.59)(0.950,0.485)11 ' '' '' ( 689.00,415.17)(11.251,7.000)2 ' '' '' ( 702.00,423.59)(0.950,0.485)11 ' '' '' ( 702.00,422.17)(11.251,7.000)2 ' '' '' ( 715.00,430.59)(0.758,0.488)13 ' '' '' ( 715.00,429.17)(10.547,8.000)2 ' '' '' ( 727.00,438.59)(0.950,0.485)11 ' '' '' ( 727.00,437.17)(11.251,7.000)2 ' '' ''
( 740.00,445.59)(0.950,0.485)11 ' '' '' ( 740.00,444.17)(11.251,7.000)2 ' '' '' ( 753.00,452.59)(0.824,0.488)13 ' '' '' ( 753.00,451.17)(11.443,8.000)2 ' '' '' ( 766.00,460.59)(0.874,0.485)11 ' '' '' ( 766.00,459.17)(10.369,7.000)2 ' '' '' ( 778.00,467.59)(0.950,0.485)11 ' '' '' ( 778.00,466.17)(11.251,7.000)2 ' '' '' ( 791.00,474.59)(0.824,0.488)13 ' '' '' ( 791.00,473.17)(11.443,8.000)2 ' '' '' ( 804.00,482.59)(0.874,0.485)11 ' '' '' ( 804.00,481.17)(10.369,7.000)2 ' '' '' ( 816.00,489.59)(0.950,0.485)11 ' '' '' ( 816.00,488.17)(11.251,7.000)2 ' '' '' ( 829.00,496.59)(0.824,0.488)13 ' '' '' ( 829.00,495.17)(11.443,8.000)2 ' '' '' ( 842.00,504.59)(0.874,0.485)11 ' '' '' ( 842.00,503.17)(10.369,7.000)2 ' '' '' ( 854.00,511.59)(0.950,0.485)11 ' '' '' ( 854.00,510.17)(11.251,7.000)2 ' '' ''
( 867.00,518.59)(0.824,0.488)13 ' '' '' ( 867.00,517.17)(11.443,8.000)2 ' '' '' ( 880.00,526.59)(0.950,0.485)11 ' '' '' ( 880.00,525.17)(11.251,7.000)2 ' '' '' ( 893.00,533.59)(0.874,0.485)11 ' '' '' ( 893.00,532.17)(10.369,7.000)2 ' '' '' ( 905.00,540.59)(0.950,0.485)11 ' '' '' ( 905.00,539.17)(11.251,7.000)2 ' '' '' ( 918.00,547.59)(0.824,0.488)13 ' '' '' ( 918.00,546.17)(11.443,8.000)2 ' '' '' ( 931.00,555.59)(0.874,0.485)11 ' '' '' ( 931.00,554.17)(10.369,7.000)2 ' '' '' ( 943.00,562.59)(0.950,0.485)11 ' '' '' ( 943.00,561.17)(11.251,7.000)2 ' '' '' ( 956.00,569.59)(0.824,0.488)13 ' '' '' ( 956.00,568.17)(11.443,8.000)2 ' '' '' ( 969.00,577.59)(0.950,0.485)11 ' '' '' ( 969.00,576.17)(11.251,7.000)2 ' '' '' ( 982.00,584.59)(0.874,0.485)11 ' '' '' ( 982.00,583.17)(10.369,7.000)2 ' '' '' ( 994.00,591.59)(0.824,0.488)13 ' '' '' ( 994.00,590.17)(11.443,8.000)2 ' '' '' ( 1007.00,599.59)(0.950,0.485)11 ' '' '' ( 1007.00,598.17)(11.251,7.000)2 ' '' '' ( 1020.00,606.59)(0.874,0.485)11 ' '' '' ( 1020.00,605.17)(10.369,7.000)2 ' '' '' ( 1032.00,613.59)(0.824,0.488)13 ' '' '' ( 1032.00,612.17)(11.443,8.000)2 ' '' '' ( 1045.00,621.59)(0.950,0.485)11 ' '' '' ( 1045.00,620.17)(11.251,7.000)2 ' '' '' ( 1058.00,628.59)(0.874,0.485)11 ' '' '' ( 1058.00,627.17)(10.369,7.000)2 ' '' '' ( 1070.00,635.59)(0.824,0.488)13 ' '' '' ( 1070.00,634.17)(11.443,8.000)2 ' '' '' ( 1083.00,643.59)(0.950,0.485)11 ' '' '' ( 1083.00,642.17)(11.251,7.000)2 ' '' '' ( 1096.00,650.59)(0.950,0.485)11 ' '' '' ( 1096.00,649.17)(11.251,7.000)2 ' '' '' ( 1109.00,657.59)(0.758,0.488)13 ' '' '' ( 1109.00,656.17)(10.547,8.000)2 ' '' '' ( 1121.00,665.59)(0.950,0.485)11 ' '' '' ( 1121.00,664.17)(11.251,7.000)2 ' '' '' ( 1134.00,672.59)(0.950,0.485)11 ' '' '' ( 1134.00,671.17)(11.251,7.000)2 ' '' '' ( 1147.00,679.59)(0.758,0.488)13 ' '' '' ( 1147.00,678.17)(10.547,8.000)2 ' '' '' ( 1159.00,687.59)(0.950,0.485)11 ' '' '' ( 1159.00,686.17)(11.251,7.000)2 ' '' '' ( 1172.00,694.59)(0.950,0.485)11 ' '' '' ( 1172.00,693.17)(11.251,7.000)2 ' '' '' ( 1185.00,701.59)(0.824,0.488)13 ' '' '' ( 1185.00,700.17)(11.443,8.000)2 ' '' '' ( 1198.00,709.59)(0.874,0.485)11 ' '' '' ( 1198.00,708.17)(10.369,7.000)2 ' '' '' ( 1210.00,716.59)(0.950,0.485)11 ' '' '' ( 1210.00,715.17)(11.251,7.000)2 ' '' '' ( 1223.00,723.59)(0.950,0.485)11 ' '' '' ( 1223.00,722.17)(11.251,7.000)2 ' '' '' ( 1236.00,730.59)(0.758,0.488)13 ' '' '' ( 1236.00,729.17)(10.547,8.000)2 ' '' '' ( 1248.00,738.59)(0.950,0.485)11 ' '' '' ( 1248.00,737.17)(11.251,7.000)2 ' '' '' ( 1261.00,745.59)(0.950,0.485)11 ' '' '' ( 1261.00,744.17)(11.251,7.000)2 ' '' '' ( 1274.00,752.59)(0.824,0.488)13 ' '' '' ( 1274.00,751.17)(11.443,8.000)2 ' '' '' ( 1287.00,760.59)(0.874,0.485)11 ' '' '' ( 1287.00,759.17)(10.369,7.000)2 ' '' '' ( 1299.00,767.59)(0.950,0.485)11 ' '' '' ( 1299.00,766.17)(11.251,7.000)2 ' '' '' ( 1312.00,774.59)(0.824,0.488)13 ' '' '' ( 1312.00,773.17)(11.443,8.000)2 ' '' '' ( 1325.00,782.59)(0.874,0.485)11 ' '' ''
( 1325.00,781.17)(10.369,7.000)2 ' '' '' ( 1337.00,789.59)(0.950,0.485)11 ' '' '' ( 1337.00,788.17)(11.251,7.000)2 ' '' '' ( 1350.00,796.59)(0.824,0.488)13 ' '' '' ( 1350.00,795.17)(11.443,8.000)2 ' '' '' ( 1363.00,804.59)(0.874,0.485)11 ' '' '' ( 1363.00,803.17)(10.369,7.000)2 ' '' '' ( 1375.00,811.59)(0.950,0.485)11 ' '' '' ( 1375.00,810.17)(11.251,7.000)2 ' '' '' ( 1388.00,818.59)(0.824,0.488)13 ' '' '' ( 1388.00,817.17)(11.443,8.000)2 ' '' '' ( 1401.00,826.59)(0.950,0.485)11 ' '' '' ( 1401.00,825.17)(11.251,7.000)2 ' '' '' ( 1414.00,833.59)(0.874,0.485)11 ' '' '' ( 1414.00,832.17)(10.369,7.000)2 ' '' '' ( 1426.00,840.59)(0.824,0.488)13 ' '' '' ( 1426.00,839.17)(11.443,8.000)2 ' '' '' 0.2 cm figure 5 : the measured effective size , @xmath226 , of the transiting planets ( from top to bottom ) hd209458b , ogle - tr-56b and ogle - tr-113b versus @xmath231 ( in units where @xmath232 for hd209458b ) .
the solid line is the prediction , eq.([1new ] ) , which assumes that the planets are composed predominately of mirror matter .
1.5 cm evidently the 2001 prediction , eq.([1new ] ) , is in reasonable agreement with the observations .
this appears to be non - trivial : in the case of ordinary matter planets , increasing the mass does not significantly affect the radius , and does not generally lead to a decreasing radius ( for example , jupiter is three times heavier than saturn , but is 15% _ larger _ ) .
however , it is possible that the apparent agreement with the rough prediction , eq.([1new ] ) is coincidental
so more data would be welcome .
especially decisive would be the discovery of a much heavier transiting planet , @xmath233 , which should have a radius less than @xmath234 if it is a mirror world .
0.2 cm 0.2 cm if this mirror world interpretation of the close - in planets is correct then it is very natural that the dynamical mirror image system of a mirror star with an ordinary planet could also exist .
such a system would appear to ordinary observers as an isolated " ordinary planet . remarkably , such
isolated " planets have been identified in the @xmath235 orionis star cluster@xcite .
these planets have estimated mass of @xmath236 and appear to be gas giants which do not seem to be associated with any visible star .
given that the @xmath235 orionis cluster is estimated to be less than 5 million years old , the formation of these isolated " planets must have occurred within this time ( which means they ca nt orbit faint stellar bodies such as old white dwarfs ) .
zapatero osorio et al@xcite argue that these findings pose a challenge to conventional theories of planet formation which are unable to explain the existence of numerous isolated planetary mass objects .
thus , the existence of these planets is surprising if they are made of ordinary matter , however there existence is natural from the mirror world perspective since they can be interpreted as ordinary planets orbiting mirror stars@xcite .
furthermore , if the isolated planets are not isolated but orbit mirror stars then there must exist a periodic doppler shift detectable on the spectral lines from these planets .
this represents a simple way of testing this hypothesis@xcite .
0.2 cm 0.2 cm perhaps the most fascinating possible implication of mirror matter - type dark matter is that our solar system contains mirror matter space - bodies ( sb)@xcite .
there is not much room for a large amount of mirror matter in our solar system .
for example , the amount of mirror matter within the earth has been constrained to be less than @xmath237@xcite .
however , we do nt know enough about the formation of the solar system to be able to exclude the existence of a large number of space bodies made of mirror matter if they are small like comets and asteroids .
the total mass of asteroids in the asteroid belt is estimated to be only about 0.05% of the mass of the earth .
a similar or even greater number of mirror bodies , perhaps orbiting in a different plane or even spherically distributed like the oort cloud is a fascinating possibility .
in fact , the comets themselves and hence the oort cloud itself
might actually be composed of mirror matter ( as we will discuss in the following subsection ) .
anyway , collisions of such bodies with themselves and ordinary bodies would generate a solar system population of mirror gas and dust particles and larger bodies .
the impact velocity of such solar system objects ( relative to the earth ) would be in the range : @xmath238 if such small mirror matter bodies do in fact exist and happen to collide with the earth , what would be the consequences ? if the only force connecting mirror matter with ordinary matter is gravity , then the consequences would be minimal .
the mirror matter space body would simply pass through the earth and nobody would know about it unless the body was so heavy as to gravitationally affect the motion of the earth . however , if there is photon - mirror photon kinetic mixing of magnitude , @xmath239 , as indicated by the dama / nai experiment , then the mirror nuclei of the space body can interact with the ordinary nuclei in the earth via elastic rutherford scattering ( see figure 2 ) .
small dust particles could thereby be detectable in simple surface experiments .
in particular , experiments such as the st .
petersburg experiment@xcite are sensitive to solar system mirror dust particles@xcite .
such particles can produce a burst of photons in a scintillator due to elastic collisions between the mirror atoms of the dust particle and the ordinary scintillator atoms . not only can these photons be detected via a photomultiplier ( pm ) tube , but the velocity of the mirror dust particle can be determined if the pm tubes are appropriately arranged .
this is important because ordinary cosmic rays should be travelling close to the speed of light , and can thereby be distinguished from relatively slow moving mirror dust particles .
petersburg experiment finds a positive signal consistent with a flux of about 1 mirror dust particle per square meter per day .
impacts of larger bodies should be less frequent , nevertheless there is a fascinating range of evidence for their existence .
the largest recorded impact event was the 1908 tunguska event .
remarkably no significant asteroid or cometary remnants were recovered from the tunguska site@xcite .
people have _ assumed _ that the impacting body was made of ordinary matter , however there is ( literally ! ) no solid evidence to support this claim .
the tunguska body may have been made out of dark matter which is a logical possibility if mirror matter is identified with the dark matter of the universe .
in fact , this hypothesis seems to provide a better explanation for the known features of the tunguska event@xcite .
there are also many other ` anomalous ' impact events , on smaller scales@xcite , and evidence for anomalous impact events on larger scales .
included among the latter are the impact events responsible for strange glass fields such as edeowie glass@xcite , libyan desert glass , tektites etc .
all of these impact related phenomena share a common feature which is the remarkable lack of clearly defined extraterrestrial material or even chemical traces ( such as iridium excess ) .
this fact is obviously explicable if the events were due to the impact of a mirror matter space body .
other solar system evidence for mirror matter also exists coming from the lack of small craters on the asteroid eros@xcite and also from the anomalous slow - down of _ both _
pioneer spacecraft@xcite .
the overall situation is summarized in figure 6 .
0.4 cm 0.3 cm figure 6 : favoured range of @xmath24 from various experiments / observations . also shown
are the current direct experimental bound , @xmath240 which comes from orthopositronium lifetime studies@xcite and also the limit , @xmath241 , suggested by early universe cosmology ( successful bbn ) @xcite .
1.3 cm direct detection of mirror matter fragments in the ground is also possible at these impact sites .
the photon - mirror photon kinetic mixing interaction will lead to a static force which can keep small mirror matter fragments ( of size @xmath242 ) near the earth s surface , provided that@xcite @xmath243 such fragments can be experimentally detected via the centrifuge technique@xcite and through the thermal effects of the embedded mirror matter on the surrounding ordinary matter@xcite .
note however that impacts of galactic halo mirror ions / electrons will vaporize these small fragments over time .
the flux of halo mirror electrons is roughly @xmath244 and defining @xmath245 to be the mean number of mirror atoms evaporated from the impact of each halo mirror electron ( @xmath246 ) , then the rate at which a mirror matter fragment would evaporate would be of order @xmath247 ( where @xmath248 is the atomic number density of mirror atoms in the mirror fragment ) .
this suggests that mirror matter fragments probably could not be recovered from the remnants of old impact events , such as edeowie glass , libyan desert glass , tektites etc ( which are of order 1 myr old or older ) but might be recovered from relatively recent anomalous impact events such as the tunguska event@xcite and small anomalous impact events@xcite .
if mirror matter space - bodes do exist in our solar system , then one might expect other unconventional scientific implications .
below we mention just a few more of these things .
comets are believed to originate from an approximately spherically symmetric cloud extending out about half way to the nearest star .
this comet cloud , called the oort cloud , is reminisant of the dark halo of our galaxy .
both are largely invisible , are distributed differently to the ` visible ' matter , and are also hypothetical .
of course , this analogy is very simplistic and should not be taken very seriously .
nevertheless , it is also true that comets seem to have a number of puzzling features and are not altogether well understood .
one interesting feature of comets is that they seem to contain a very dark nucleus .
for example , the nucleus of halley s comet has an albedo of only 0.03 making it one of the darkest objects in the solar system darker even than coal !
this has led to the suggestion@xcite that the nucleus could be composed predominately of mirror matter .
of course , pure mirror matter would be transparent , but if it contained a small admixture of ordinary matter embedded within , it might appear opaque and dark .
if the ordinary matter had a volatile component such as water ice , then this would explain the large head and tail observed when the comet passed close to the sun .
furthermore , such a picture would simply explain the long standing comet fading problem : that many comets lose a large factor ( 100 - 1000 ) in average brightness after approaching the sun for the first time .
if this interpretation is correct , then comets may simply become dimmer and dimmer over time rapidly losing all of their volatile ordinary matter component .
they may effectively become invisible .
of course , the rate that this occurs will depend on many things such as the proportion of ordinary to mirror matter , the chemical composition , details of the orbit etc .
interestingly , a recent study@xcite has concluded that many old comets must have either become invisible or have somehow disintegrated .
the number of cometary remnants ( assumed to be asteroid - like objects ) is 100 times less abundant than theoretically expected!@xcite .
clearly , this seems to support ( or at least , encourage ) the mirror matter interpretation of the comets .
of course , if comets are predominately made of mirror matter then this fits - in nicely with the mirror matter interpretation of the anomalous small impact events ( and tunguska event ) , which was discussed in the previous subsection .
it might also be connected with atmospheric anomalies . to explain the anomalous small impact events we require that some of the mirror matter space - bodies to survive and hit the ground without completely melting and vaporizing in the atmosphere .
detailed studies@xcite have shown that this is possible , especially for non - volatile mirror matter ( such as mirror iron ) .
sometimes , it could happen that a mirror space - body would heat up enough to completely vaporize in the atmosphere .
after vaporizing , the mirror atoms interact with the air atoms by rutherford scattering .
although initially the mirror matter will heat up the ordinary matter because of its large kinetic energy ( since its initial velocity is at least 11 km / s ) , after a short time , the mirror matter will cool the atmosphere .
the mirror atoms will draw in heat from the surrounding ordinary atoms and radiate it away into mirror photons .
since the mirror atoms are not absorbing mirror photons from the environment , heat will be lost from the system .
the net effect is a localized rapid cooling of the atmosphere which might lead to the formation of unusual clouds and other strange atmospheric phenomena .
this might explain the remarkable observations of falling ice blocks@xcite and maybe even the observations of atmospheric ` holes'@xcite .
it seems that the answer may indeed be ` blowing in the wind ' but only for a sort time !
historically , imposing symmetries of particle interactions has led to the prediction and subsequent discovery of a variety of ` new ' fundamental particles including : * antiparticles predicted to exist by imposing proper lorentz symmetry ; * neutrino predicted to exist by imposing time translational symmetry ( energy conservation ) ; * top quark predicted to exist from @xmath249 electroweak gauge symmetry ( to partner the bottom quark ) ; * the @xmath250 baryon predicted from @xmath251 flavour symmetry in the quark model .
mirror matter is also an offspring of this methodology ; it is an attempt to follow this historically successful approach .
in fact it appears to be theoretically unique , arising from a single well motivated hypothesis : the improper lorentz symmetries ( such as parity and time reversal invariance ) stand out as the only space - time symmetries which are not respected by the interactions of the known elementary particles , but can be exact unbroken symmetries of nature if a set of mirror particles exist .
mirror matter is thus very well motivated from a particle physics point of view .
furthermore it seems to have the right properties to be identified with the inferred non - baryonic dark matter in the universe .
specifically , mirror dark matter seems to provide a consistent explanation for : a ) the basic dark matter particle properties ( mass , stability , darkness ) ; b ) the similarity in cosmic abundance between ordinary and non - baryonic dark matter , @xmath0 ; c ) large scale structure formation ; d ) microlensing ( macho ) events ; e ) asymptotically flat rotation curves in spiral galaxies and f ) the impressive dama / nai annual modulation signal . of course ,
any theory of dark matter should also be measured against the standard paradigm that non - baryonic dark matter consists of hypothetical weakly interacting particles i.e. essentially collisionless particles .
however this comparison is actually favourable . in the standard wimp hypothesis : the basic dark matter properties ( stability , darkness ) require ad hoc hypothesis ; macho events can not be explained ; @xmath252 is mysterious ; dama / nai annual modulation signal is difficult to understand consistently with other experiments such as cdmsii .
perhaps the only thing that wimps might explain better is the existence of the halo in galaxies .
this is because wimps are non - dissipative .
however this success is significantly eroded by the facts ; standard collisionless dark matter predicts@xcite overly dense cores in galaxies and over abundance of small scale structures within halos which are _ not _ consistent with the observations@xcite .
thus , by either comparing mirror matter - type dark matter with experiments and observations or with the standard wimp paradigm , it is clear that it is a strong candidate for the non - baryonic dark matter in the universe , deserving of serious consideration and further study .
0.4 cm * acknowledgements * 0.1 cm it is a pleasure to thank my collaborators , sergei gninenko , sasha ignatiev , henry lew , saibal mitra , zurab silagadze , ray volkas and t. l. yoon .
i would also like to thank s. mitra , z. silagadze and r. volkas for their comments on a draft of this article .
this work was supported by the australian research council .
r. foot , h lew and r. r. volkas , phys .
b272 , 67 ( 1991 ) .
the idea that exact unbroken parity symmetry can be respected by the fundamental interactions if a set of mirror particles exist was earlier discussed , prior to the advent of the standard model of particle physics , in t. d. lee and c. n. yang , phys . rev .
104 , 256 ( 1956 ) ; i. kobzarev , l. okun and i. pomeranchuk , sov . j. nucl .
phys . 3 , 837 ( 1966 ) ; m. pavsic , int . j. theor .
9 , 229 ( 1974 ) .
the first application to non - baryonic dark matter was given in : s. i. blinnikov and m. yu .
khlopov , sov .
36 , 472 ( 1982 ) ; sov .
27 , 371 ( 1983 ) .
r. foot , phys .
b496 , 169 ( 2000 ) [ hep - ph/0007065 ] ; r. foot and r. r. volkas , phys .
b543 , 38 ( 2002 ) [ hep - ph/0204265 ] ; r. foot , mod .
a18 , 2079 ( 2003 ) [ hep - ph/0210393 ] ; mod .
a18 , 2071 ( 2003 ) [ hep - ph/0303005 ] .
e. w. kolb , d. seckel and m. s. turner , nature 314 , 415 ( 1985 ) ; h. m. hodges , phys .
d47 , 456 ( 1993 ) ; z. g. berezhiani , a. d. dolgov and r. n. mohapatra , phys .
b375 , 26 ( 1996 ) [ hep - ph/9511221 ] ; v. berezinsky and a. vilenkin , phys .
d62 , 083512 ( 2000 ) [ hep - ph/9908257 ] .
r. foot and r. r. volkas , phys . rev .
d68 , 021304 ( 2003 ) [ hep - ph/0304261 ] ; phys .
d69 , 123510 ( 2004 ) [ hep - ph/0402267 ] see also : l. bento and z. berezhiani , phys .
87 , 231304 ( 2001 ) [ hep - ph/0107281 ] ; hep - ph/0111116 .
z. k. silagadze , acta phys .
b32 , 99 ( 2001 ) [ hep - ph/0002255 ] ; r. foot and z. k. silagadze , acta phys .
b32 , 2271 ( 2001 ) [ astro - ph/0104251 ] ; z. k. silagadze , acta phys . pol .
b33 , 1325 ( 2002 ) [ astro - ph/0110161 ] .
r. foot , acta phys .
b32 , 3133 ( 2001 ) [ hep - ph/0107132 ] ; r. foot and t. l. yoon , acta phys .
33 , 1979 ( 2002 ) [ astro - ph/0203152 ] ; r. foot and s. mitra , astropart .
19 , 739 ( 2003 ) [ astro - ph/0211067 ] ; z. k. silagadze , astro - ph/0311337 .
m. morawska - horawska and a. manechi , przeglad geofizyczny 40 , 335 ( 1995 ) ; j. a. docobo _
et al_. , meteoritics & planetary science 33 , 57 ( 1998 ) ; a. olkhovatov , http://olkhov.narod.ru/gr1997.htm .
the 2001 jordan event , http://www.jas.org.jo/mett.html . http://tierra.rediris.es/megacryometeors/index2.html and j. martinez - frias _ et al .
_ , compositional heterogeneity of hailstones : atmospheric conditions and possible environmental implications , ambio vol .
30 ( 2001 ) . | there are six main things which any non - baryonic dark matter theory should endeavour to explain : ( 1 ) the basic dark matter particle properties [ mass , stability , darkness ] ; ( 2 ) the similarity in cosmic abundance between ordinary and non - baryonic dark matter , @xmath0 ; ( 3 ) large scale structure formation ; ( 4 ) microlensing ( macho ) events ; ( 5 ) asymptotically flat rotation curves in spiral galaxies ; ( 6 ) the impressive dama / nai annual modulation signal .
only mirror matter - type dark matter is capable of explaining all six of these desirable features .
the purpose of this article is to provide an up - to - date and pedagogical review of this dark matter candidate . |
quantum game theory started from a seminal paper by meyer @xcite .
later eisert et .
@xcite studied the important bimatrix game of prisoner s dilemma ( pd ) while focussing on the concept of nash equilibrium ( ne ) from noncooperative game theory .
this concept attracted much attention in other recent works in quantum game theory @xcite .
in fact cournot ( 1838 ) @xcite anticipated nash s definition of equilibrium by over a century but only in the context of a particular model of duopoly . in economics ,
an oligopoly is a form of market in which a number @xmath0 of producers , say , @xmath1 , and no others , provide the market with a certain commodity . in the special case where @xmath2 , it is called a duopoly .
cournot s work @xcite is one of the classics of game theory and also a cornerstone of the theory of industrial organization @xcite . in cournot model of duopoly two - firms
simultaneously put certain quantities of a homogeneous product in the market .
cournot obtained an equilibrium value for the quantities both firms will decide to put in the market .
this equilibrium value was based on a rule of behavior which says that if all the players except one abide by it , the remaining player can not do better than to abide by it too .
nash gave a general concept of an equilibrium point in a noncooperative game but existence of an equilibrium in duopoly game was known much earlier .
the `` cournot equilibrium '' refers to ne in noncooperative form of duopoly that cournot considered . in an interesting later development stackelberg ( 1934 )
@xcite proposed a dynamic model of duopoly in which , contrary to cournot s assumption of simultaneous moves , a leader ( or dominant ) firm moves first and a follower ( or subordinate ) firm moves second .
a well known example is the general motors playing this leadership role in the early history of u.s .
automobile industry when more than one firms like ford and chrysler acted as followers . in this sequential game
a `` stackelberg equilibrium '' is obtained using the _ backwards - induction outcome _ of the game .
in fact stackelberg equilibrium refers to the sequential - move nature of the game and this is a stronger solution concept than ne because sequential move games sometimes have multiple ne , only one of which is associated with the backwards - induction outcome of the game @xcite . the seminal paper by meyer @xcite has initiated the new field of quantum games and has motivated many people to look at games from quantum perspectives .
eisert et .
@xcite quantized the famous bimatrix game of prisoner s dilemma ( pd ) and showed that dilemma disappears in quantum world .
they allowed players an access to a maximally entangled state that can be generated from a system of two qubits .
each player then unitarily manipulates a qubit in his possession .
players apply a unitary operator from a particular subset of the general set of unitary operators that also forms a group .
benjamin @xcite later showed that if the players have access to the set of general unitary operators then there is no ne in pure strategies in the pd game .
benjamin implied that eisert s set of unitary operators was carefully chosen to generate a ne that has no classical counterpart .
eisert et .
@xcite used an unentangling gate in their scheme to be put before the quantum state is forwarded to measuring apparatus that collapses the wave function and gives the payoffs .
the unentangling gate in eisert s scheme ensured that classical game could be reproducible but it motivated marinatto @xcite to question the necessity of its presence in the scheme to play a quantum game .
marinatto and weber @xcite came up with a solution where an initial quantum state , they called it an `` initial strategy '' , is made available to the players .
this initial strategy is then unitarily manipulated by the players in the `` tactics '' phase of the game that consisted of applying two unitary and hermitian operators ( the identity and an inversion operator ) with classical probabilities on the initial quantum strategy .
this inversion operator reverts the quantum state just like the pauli s spin flip operator does @xcite .
marinatto and weber @xcite showed that for an initial strategy that is a maximally entangled state a unique ne for the game of battle of sexes can be obtained .
later in an interesting comment benjamin @xcite considered the players access to apply only two unitary and hermitian operators with classical probabilities on a quantum state a severe restriction on all quantum mechanically possible manipulations .
marinatto and weber replied @xcite that the only restriction on a quantum form of a game is that corresponding classical game must be reproducible as a special case of the quantum form . agreeing with this reply we studied the concept of evolutionary stability in asymmetric as well as symmetric quantum games @xcite .
our prime motivation was an important and interesting element in marinatto and weber s scheme .
it was the fact that a switch - over between a classical and a non - classical form of a game could be achieved by having a control over the parameters of the initial quantum state or initial strategy . in the scheme of eisert , wilkens , and lewenstein @xcite
such a switch - over takes place when the players assign specific values to the parameters of the unitary operators in their possession ; while the initial two - qubit quantum state always remains to be maximally entangled .
starting a quantum game from a general pure state of two qubits that have interacted previously we were able to show the possibility that an unstable symmetric ne of a bimatrix classical game becomes stable in a quantum form of the same game @xcite .
our assumption in this approach was that a replacement of the maximally entangled two qubit quantum state , that marinatto and weber used to get a unique ne , with a general two qubit pure quantum state also results in another form of the same game .
this assumption led us to find the conditions on the constants of a general three qubit pure state that made it possible to counter against coalition formation in a three player symmetric game @xcite .
the person responsible for preparing quantum states can thus make vanish the motivation to make a coalition in the cooperative game .
motivated by these recent developments in quantum games as well as the notion of backwards - induction outcome of a dynamic game of complete information @xcite we present a quantum perspective on the interesting game of stackelberg duopoly . in present paper
we start with the same assumption that a game is decided only by players unitary manipulations , payoff operators , and the measuring apparatus deciding payoffs .
when these are same a different input quantum initial state gives only a different form of the same game .
this was our assumption when we studied evolutionary stability of a mixed ne in rock - scissors - paper game @xcite .
therefore all the games that can be obtained by using a general two qubit pure state are only different forms of the same game if the rest of the procedures in playing the quantum game are same .
for example in marinatto and weber s battle of sexes @xcite the game remains battle of sexes for all two qubit pure quantum states . with this assumption
we start an analysis of stackelberg duopoly by asking a fundamental question : is it possible to find a two qubit pure quantum state that generates the classical cournot equilibrium as a backwards - induction outcome of the quantum form of stackelberg duopoly ?
why this question can be of interest ?
for us it is interesting because in case the answer is yes then the very important resource in quantum game theory i.e. entanglement can potentially be a particularly useful element for ` follower ' in the leader - follower model of stackelberg duopoly @xcite .
this is because in classical settings when static duopoly changes itself into a dynamic form the follower becomes worse - off compared to leader who becomes better - off .
we find that under certain restrictions it is possible to find the needed two qubit quantum states .
therefore a quantum form of a dynamic game of complete information gives out an equilibrium that corresponds to classical static form of the same game . in our analysis
the equilibrium can be obtained for a certain range of the constant of the duopoly game .
this restriction appears only due to an assumption that we introduce to simplify calculations .
this fact , however , does not rule out the possibility of getting a quantum form of stackelberg duopoly game with no such restriction .
to make this paper self - contained we give an introduction of the backwards - induction outcome of a sequential game .
for this we find very useful the ref .
consider a simple three step game 1 .
player 1 chooses an action @xmath3 from the feasible set @xmath4 2 .
player 2 observes @xmath3 and then chooses an action @xmath5 from the feasible set @xmath6 3 .
payoffs are @xmath7 and @xmath8 this game is an example of the dynamic games of complete and perfect information
. key features of such games are 1 .
the moves occur in sequence 2 .
all previous moves are known before next move is chosen , and 3 .
the players payoffs are common knowledge . at the second stage of the game
when player @xmath9 gets the move he or she faces the following problem , given the action @xmath3previously chosen by the player @xmath10 assume that for each @xmath3 in @xmath4 , player @xmath9 s optimization problem has a unique solution @xmath11 also called the best response of player @xmath9 .
now player @xmath12 can also solve player @xmath9 s optimization problem by anticipating player @xmath9 s response to each action @xmath3 that player @xmath12 might take .
so that player @xmath12 faces the following problem @xmath13 suppose this optimization problem also has a unique solution for player @xmath12 and is denoted by @xmath14 . the solution @xmath15 is the backwards - induction outcome of this game . in a simple version of the cournot s model two firms
simultaneously decide the quantities @xmath16 and @xmath17 respectively of a homogeneous product they want to put into the market
. suppose @xmath18 is the aggregate quantity i.e. @xmath19 and @xmath20 be the market - clearing price , the price at which all products or services available in a market will find buyers . assume the total cost to a firm producing quantity @xmath21 is @xmath22 i.e. there are no fixed costs and the marginal cost is a constant @xmath23 with @xmath24 .
in a two - player game theoretical model of this situation a firm s payoff or profit can be written as @xcite @xmath25 = q_{i}\left [ a - c-(q_{i}% + q_{j})\right ] = q_{i}\left [ k-(q_{i}+q_{j})\right ] \label{payoffeq}%\ ] ] solving for the ne easily gives the cournot equilibrium @xmath26 at this equilibrium the payoffs to both firms from eq .
( [ payoffeq ] ) are @xmath27 we now come to consider the classical form of duopoly game when it becomes dynamic . the game becomes dynamic but the payoffs to players are given by the same eq .
( [ payoffeq ] ) as for the case of the cournot s game .
we find backwards - induction outcome in classical and a quantum form of stackelberg s duopoly .
taking advantage from bigger picture given to this dynamic game by hilbert space structure we then find two qubit pure quantum states that give classical cournot s equilibrium as the backwards - induction outcome of the quantum game of stackelberg s duopoly .
a leader ( or dominant ) firm moves first and a follower ( or subordinate ) firm moves second in stackelberg model of duopoly @xcite .
the sequence of events is 1 .
firm @xmath28 chooses a quantity @xmath29 2 .
firm @xmath30 observes @xmath16 and then chooses a quantity @xmath31 3 .
the payoffs to firms @xmath28 and @xmath30 are given by their respective profit functions as @xmath32 \nonumber\\ p_{b}(q_{1},q_{2 } ) & = q_{2}\left [ k-(q_{1}+q_{2})\right ] \label{proffunc}%\end{aligned}\ ] ] the backwards - induction outcome is found by first finding firm @xmath30 s reaction to an arbitrary quantity by firm @xmath33 denoting this quantity as @xmath34 we find @xmath35 with @xmath36 .
now the interesting aspect of this game is that firm @xmath28 can solve the firm @xmath30 s problem as well .
therefore firm @xmath28 can anticipate that a choice of the quantity @xmath16 will meet a reaction @xmath34 . in the first stage of the game firm @xmath28
can then compute a solution to his / her optimization problem as @xmath37 = \underset{q_{1}\geq0}{max}\frac{q_{1}(k - q_{1})}{2}%\ ] ] it gives @xmath38 it is the classical backwards - induction outcome of dynamic form of the duopoly game . at this equilibrium payoffs or profits to players @xmath28 and @xmath30 are given by eqs .
( [ proffunc ] ) and ( [ stkeq ] ) @xmath39 _ { stackelberg}% = \frac{k^{2}}{8},\text { \ \ \ \ } p_{b}\left [ q_{1}^{\star},r_{2}(q_{1}% ^{\star})\right ] _ { stackelberg}=\frac{k^{2}}{16 } \label{stkpayffs}%\ ] ] from eq .
( [ stkpayffs ] ) find the following ratio @xmath40 _ { stackelberg}% } { p_{b}\left [ q_{1}^{\star},r_{2}(q_{1}^{\star})\right ] _ { stackelberg}}=2 \label{ratio}%\ ] ] showing that with comparison to cournot game in stackelberg game the leader firm becomes better - off and the follower firm becomes worse - off .
this aspect also hints an important difference between single and multi - person decision problems . in single - person decision theory having more information can never make the decision maker worse - off . in game theory , however , having more information ( or , more precisely , having it made public that one has more information ) can make a player worse - off @xcite .
now we look at backwards - induction outcome in a quantum perspective .
our motivation is an interesting aspect that quantum form can bring into backwards - induction outcome .
it is the possibility that the ` extra information ' that firm @xmath30 has does not make firm @xmath30 worse - off .
stackelberg duopoly is a two player sequential game .
meyer @xcite considered quantum form of sequential game of pq penny flip by unitary operations on single qubit .
important difference between meyer s pq penny flip and stackelberg duopoly is that at the second stage player in pq penny flip does nt know the previous move but in stackelberg duopoly he knows that .
marinatto and weber @xcite used two qubits to play the noncooperative game of battle of sexes .
players apply unitary operators @xmath41 and @xmath42 with classical probabilities on a two - qubit pure quantum state .
@xmath41 is identity and @xmath42 is inversion or pauli s spin - flip operator .
we prefer this scheme to play the sequential game of stackelberg duopoly for two reasons : 1 .
interesting feature in this scheme that classical game is reproducible on making initial state unentangled .
2 . we assumed that other games obtained from every pure two qubit initial state are only quantum forms of the classical game provided players actions and payoff generating measurement are exactly same @xcite . this assumption reduces the problem of finding a quantum form of stackelberg duopoly with property that its equilibrium is same as in cournot s duopoly to the problem of finding conditions on parameters of two - qubit pure quantum state .
if the conditions are realistic then the corresponding quantum game will give cournot s equilibrium as a backwards - induction outcome .
stackelberg.eps + fig .
1 . playing a quantum form of stackelberg duopoly .
1 ] stackelberg duopoly is a dynamic game of complete information .
its quantum form in marinatto and weber s scheme @xcite starts by preparing a pure two - qubit initial quantum state .
both these qubits are then forwarded to two players in the game called alice and bob .
marinatto and weber expanded on the earlier scheme proposed by eisert , wilkens , and lewenstein @xcite using a two - qubit system in a maximally entangled quantum state .
the fundamental idea remains the same i.e. to exploit entanglement to get new results for a simultaneous - move game .
suppose first move is to be played by alice and she plays her strategy by applying two operators in her possession on her qubit with classical probabilities .
she also announces her move immediately so that bob knows alice s strategy before playing his move .
bob plays his strategy on his qubit and both alice and bob forward their qubits to a setup where measurement can be done to decide payoffs to both .
the concept of information about the previous moves is crucial for the understanding of the present paper .
a comparison between the sequential game of stackelberg duopoly with the simultaneous moves of the battle of sexes makes clear the different information structure between the two games .
for example in the battle of sexes , when played sequentially , alice does not announce her first move to bob before he makes his move . in this way
the game becomes sequential but the information structure of the game is still the same as in its static form .
consequently , the sequential battle of sexes in the above form has the same ne as in its static form .
an unobserved - action form of a game has the same ne as its simultaneous - move form .
this observation led us to play a quantum form of stackelberg duopoly while keeping intact the original structure of a scheme designed for simultaneous moves .
a consideration of playing a sequential game in a quantum way brings to mind the meyer s pq penny flip @xcite where only one qubit is used in the game .
contrary to this , in present paper we use the two - qubit system of a simultaneous moves game to play a sequential game .
one can ask what is the point of taking extra pains by using two qubits when a quantum form of this sequential game can also be played by only one qubit in similar way as meyer s pq penny flip .
we prefer two qubits for a reason that appeared to us quite important .
_ in two - qubit case a comparison between classical and a quantum form of the game translates itself into comparing two games resulting from using unentangled and entangled initial quantum states_. we do not rule out the possibility that a consideration of the dynamic game using only single qubit gives equally or even more interesting results .
the quantum form of marinatto and weber s @xcite battle of sexes game reduces to its classical form for the initial state @xmath43 where @xmath44 represents the pure classical strategy called ` opera ' . in the same spirit
we let classical payoffs in stackelberg duopoly given by eq .
( [ proffunc ] ) reproduced when the initial state @xmath45 is used to play the game .
the state @xmath46 means that both qubits are in lower state .
we represent the upper state of a qubit by number @xmath9 .
for @xmath45 the corresponding density matrix is @xmath47 when players apply the unitary operators @xmath41 and @xmath42 such that alice and bob apply @xmath41 with probabilities @xmath48 and @xmath49 respectively the density matrix ( [ inidenmat ] ) changes to @xcite @xmath50 where the inversion operator @xmath42 interchanges the vectors @xmath51 and @xmath52 i.e. @xmath53 @xmath54 and @xmath55 .
we now assume that in stackelberg duopoly also players moves are again given by probabilities lying in the range @xmath56 $ ] .
the moves by alice and bob in classical duopoly game are given by quantities @xmath16 and @xmath17 where @xmath57 .
we assume that alice and bob agree on a function that can uniquely define a real positive number in the range @xmath58 $ ] for every quantity @xmath59 in @xmath60 .
a simple such function is @xmath61 , so that alice and bob find @xmath48 and @xmath49 respectively as @xmath62 and use these real numbers as probabilities with which they apply the identity operator @xmath41 on the quantum state at their disposal . with a substitution from eqs .
( [ inidenmat],[deffuns ] ) the final density matrix ( [ findenmat ] ) can be written as @xmath63\ ] ]
we now also suppose that in the measurement and payoffs finding phase the quantities @xmath16 and @xmath17 are known to the ` agent ' doing this action .
the agent applies the payoff operators @xmath64 , @xmath65 given as follows @xmath66 \nonumber\\ ( p_{b})_{oper } & = ( 1+q_{1})(1+q_{2})q_{2}\left [ k\left|
11\right\rangle \left\langle 11\right| -\left| 21\right\rangle \left\langle 21\right| -\left| 12\right\rangle \left\langle 12\right| \right ] \label{payopers}%\end{aligned}\ ] ] note that the classical payoffs of eq .
( [ proffunc ] ) are reproduced with the initial state @xmath45 with the following trace operations @xmath67 \nonumber\\ p_{b}(q_{1},q_{2 } ) & = trace\left [ ( p_{b})_{oper}\rho_{fin}\right ] \label{playerspayoffs}%\end{aligned}\ ] ] a more general form of quantum duopoly can now be played by keeping the payoff operators of eq .
( [ payopers ] ) in the agent s possession and preparing a general initial pure state of the following form @xmath68 where @xmath69,for @xmath70 or @xmath71 are complex numbers and @xmath72 are orthonormal basis vectors in @xmath73 dimensional hilbert space .
the payoffs to alice and bob can now be obtained in this more general quantum game from eqs .
( [ playerspayoffs ] ) that use the same payoff operators of eqs .
( [ payopers ] ) .
the payoffs to alice and bob can now be written as @xcite @xmath74 _ { qtm } & = \frac{(\omega_{11}+\omega _ { 12}q_{2})+q_{1}(\omega_{21}+\omega_{22}q_{2})}{(1+q_{1})(1+q_{2}% ) } \nonumber\\ \left [ p_{b}(q_{1},q_{2})\right ] _ { qtm } & = \frac{(\chi_{11}+\chi_{12}% q_{2})+q_{1}(\chi_{21}+\chi_{22}q_{2})}{(1+q_{1})(1+q_{2 } ) } \label{qpayoffs}%\end{aligned}\ ] ] where the subscript @xmath75 is for ` quantum ' and @xmath76{c}% \omega_{11}\\ \omega_{12}\\ \omega_{21}\\ \omega_{22}% \end{array } \right ] & = \left [ \begin{array } [ c]{cccc}% \left| c_{11}\right| ^{2 } & \left| c_{12}\right| ^{2 } & \left| c_{21}\right| ^{2 } & \left| c_{22}\right| ^{2}\\ \left| c_{12}\right| ^{2 } & \left| c_{11}\right| ^{2 } & \left| c_{22}\right| ^{2 } & \left| c_{21}\right| ^{2}\\ \left| c_{21}\right| ^{2 } & \left| c_{22}\right| ^{2 } & \left| c_{11}\right| ^{2 } & \left| c_{12}\right| ^{2}\\ \left| c_{22}\right| ^{2 } & \left| c_{21}\right| ^{2 } & \left| c_{12}\right| ^{2 } & \left| c_{11}\right| ^{2}% \end{array } \right ] \left [ \begin{array } [ c]{c}% kq_{1}(1+q_{1})(1+q_{2})\\ -q_{1}(1+q_{1})(1+q_{2})\\ -q_{1}(1+q_{1})(1+q_{2})\\ 0 \end{array } \right ] \nonumber\\ \left [ \begin{array } [ c]{c}% \chi_{11}\\ \chi_{12}\\ \chi_{21}\\ \chi_{22}% \end{array } \right ] & = \left [ \begin{array } [ c]{cccc}% \left| c_{11}\right| ^{2 } & \left| c_{12}\right| ^{2 } & \left| c_{21}\right| ^{2 } & \left| c_{22}\right| ^{2}\\ \left| c_{12}\right| ^{2 } & \left| c_{11}\right| ^{2 } & \left| c_{22}\right| ^{2 } & \left| c_{21}\right| ^{2}\\ \left| c_{21}\right| ^{2 } & \left| c_{22}\right| ^{2 } & \left| c_{11}\right| ^{2 } & \left| c_{12}\right| ^{2}\\ \left| c_{22}\right| ^{2 } & \left| c_{21}\right| ^{2 } & \left| c_{12}\right| ^{2 } & \left| c_{11}\right| ^{2}% \end{array } \right ] \left [ \begin{array } [ c]{c}% kq_{2}(1+q_{1})(1+q_{2})\\ -q_{2}(1+q_{1})(1+q_{2})\\ -q_{2}(1+q_{1})(1+q_{2})\\ 0 \end{array } \right ] \nonumber\\ & \label{consts}%\end{aligned}\ ] ] the classical payoffs of duopoly game given in eqs .
( [ proffunc ] ) are recovered from the eqs .
( [ qpayoffs ] ) when the initial quantum state becomes unentangled and given by @xmath77 .
classical duopoly is , therefore , a subset of its quantum version .
we now find the backwards - induction outcome in this quantum form of stackelberg duopoly .
we proceed in exactly the same way as it is done in the classical game except that players payoffs are now given by eqs .
( [ qpayoffs ] ) and not by eqs .
( [ proffunc ] ) .
the first step in backwards - induction in quantum game is to find bob s reaction to an arbitrary quantity @xmath16 chosen by alice .
denoting this quantity as @xmath78 _ { qtm}$ ] we find @xmath79 _ { qtm}=\underset{q_{2}\geq0}{max}\left [ p_{b}(q_{1},q_{2})\right ] _ { qtm}=\frac{q_{1}\triangle_{1}+\triangle_{2}% } { -2\left\ { q_{1}\triangle_{3}+\triangle_{4}\right\ } } \text { \ \ \ \
where}% \nonumber\\ \left| c_{11}\right| ^{2}+\left| c_{22}\right| ^{2}-k\left| c_{21}\right| ^{2}=\triangle_{1}\text { , \ \ \ \ \ \ \ \ } \left| c_{12}\right| ^{2}+\left| c_{21}\right| ^{2}-k\left| c_{11}\right| ^{2}=\triangle_{2}\nonumber\\ \left| c_{12}\right| ^{2}+\left| c_{21}\right| ^{2}-k\left| c_{22}\right| ^{2}=\triangle_{3}\text { , \ \ \ \ \ \ \ \ } \left| c_{11}\right| ^{2}+\left| c_{22}\right| ^{2}-k\left| c_{12}\right| ^{2}=\triangle_{4 } \label{qbestres}%\end{gathered}\ ] ] this reaction reduces to its classical value of eq .
( [ bestres ] ) when @xmath80 .
similar to classical game alice can now solve bob s problem as well .
therefore alice can anticipate that a choice of the quantity @xmath16 will meet a reaction @xmath81 _ { qtm}$ ] . in the first stage of the game like
its classical version alice can compute a solution to her optimization problem as @xmath82 _ { qtm}%\ ] ] to find it alice calculates the following quantity @xmath83 _ { qtm}}{dq_{1 } } & = \frac{(\left| c_{11}\right| ^{2}+\left| c_{22}\right| ^{2}-\left| c_{12}\right| ^{2}-\left| c_{21}\right| ^{2})}{(1+q_{1})}\left\ { -2q_{1}^{2}% + q_{1}(k-2)+k\right\ } \nonumber\\ & + ( 1 + 2q_{1})\left\ { ( k-1)\left| c_{21}\right| ^{2}-\left| c_{12}\right| ^{2}\right\ } + k(\left| c_{12}\right| ^{2}-\left| c_{22}\right| ^{2})\nonumber\\ & -q_{1}\frac{dq_{2}}{dq_{1}}\left\ { \triangle_{4}+q_{1}\triangle _ { 3}\right\ } -q_{2}\left\ { 2q_{1}\triangle_{3}+\triangle_{4}\right\ } \nonumber\\ & \label{derivative}%\end{aligned}\ ] ] and replaces @xmath17 in eq .
( [ derivative ] ) with @xmath84 _ { qtm}$ ] given by eq .
( [ qbestres ] ) and then equates eq .
( [ derivative ] ) to zero to find a @xmath85 that maximizes her payoff @xmath86 _ { qtm}$ ] . for a maxima
she would ensure that the second derivative of @xmath87 with respect to @xmath16 at @xmath88 is a negative quantity .
the @xmath85 together with @xmath89 _ { qtm}$ ] will form the backwards - induction outcome of the quantum game .
an interesting situation is when the backwards - induction outcome in quantum version of stackelberg duopoly becomes same as the classical cournot equilibrium of duopoly .
the classical situation of leader becoming better - off while the follower becoming worse - off is then avoided in the quantum form of stackelberg duopoly . to look for this possibility
we need such an initial state @xmath90 that at @xmath91 we should have following relations holding true also with the normalization condition given in eq .
( [ normn ] ) @xmath92 _ { qtm}}{dq_{1}}\mid_{q_{1}=q_{1}^{\star}% } = 0\label{condn1}\\ \left [ \frac{d^{2}\left [ p_{a}\left\ { q_{1},\left\ { r_{2}(q_{1})\right\ } _ { qtm}\right\ } \right ] _ { qtm}}{dq_{1}^{2}}\mid_{q_{1}=q_{1}^{\star}}\right ] < 0\label{condn2}\\ q_{2}^{\star}=\left\ { r_{2}(q_{1}^{\star})\right\ } _ { qtm } \label{condn3}%\end{gathered}\ ] ] the conditions ( [ condn1],[condn2 ] ) simply say that the backwards - induction outcome of the quantum game is the same as cournot equilibrium in classical game .
the condition ( [ condn3 ] ) says that bob s reaction to alice s choice of @xmath93 is @xmath94 . to show that such quantum states can exist for which the conditions ( [ condn1],[condn2],[condn3 ] ) with the normalization ( [ normn ] ) hold true we give an example where @xmath95 and @xmath96 are written as functions of @xmath97 with our assumption that @xmath98 .
though this assumption puts its own restriction on the possible range of @xmath97 for which the above conditions hold for these functions but still it shows clearly the possibility of finding the required initial quantum states .
the functions are found as @xmath99 also @xmath100 now , interestingly , given that allowed range of @xmath97 is @xmath101 , all of the conditions ( [ normn],[condn1],[condn2],[condn3 ] ) hold at @xmath91 .
so that in this range of @xmath97 a quantum form of stackelberg duopoly exists that gives the classical cournot equilibrium as backwards - induction outcome .
the restriction on allowed range of @xmath97 is result of our assumption that @xmath102 . in fig.@xmath9 below @xmath103and @xmath104
are plotted against @xmath97 in the above range .
plot1.eps + fig .
2 . @xmath105 @xmath104against @xmath97 when @xmath98 [ fig .
what can possibly be a relevance of considering a quantum form of a game that models a competition between two firms in macroscopic world of economics ?
quantum mechanics was developed to understand phenomena in the regime of atomic and subatomic interactions and is still mostly used in that domain .
what is of interest in extending a game theoretical model of interaction between firms towards quantum domain ?
these questions naturally arise not only with reference to stackelberg duopoly considered in this paper but also other related works in quantum games . apart from exciting new directions that quantum mechanics brings to game theory
there is also a fundamental interest in quantum games from the view of quantum information theory .
the fact that quantum algorithms may be thought of as games between classical and quantum agents was pointed out by meyer @xcite .
meyer indicated a strong motivation for the study of quantum games by considering zero - sum quantum games in order to have a new starting point at hand to find quantum algorithms that outperform its classical analogue .
many quantum information exchange protocols have been modelled like games .
eavesdropping and optimal cloning are two such often cited examples where objective before a player is to gain as much information as possible .
we believe that like other notions of game theory finding some relevance in quantum information a consideration of backwards - induction can be of interest for exactly the same reasons .
it does not seem hard to imagine situations in quantum information where moves occur in sequence , all previous moves are observed before the next move is chosen , players payoffs from each feasible combination of moves are common knowledge .
interesting questions then arise about how a quantum version of dynamic game of complete information can influence the outcome .
our primary motivation , however , to study backwards - induction in quantum games is from the view of dynamic stability , especially of symmetric ne , that has important relevance in evolutionary games @xcite that we found interesting in quantum settings @xcite .
the duopoly game models economic competition between firms and applied economics is the area where it is studied in detail .
we considered this game in a scheme that tells how to play a quantum game and gives a hilbert structure to the strategy space to which players have access to .
the fact that quantum game theory can give entirely new views on games important in economics is apparent in recent interesting papers by piotrowski and sladkowski @xcite proposing a quantum - like description of markets and economies where players strategies belong to hilbert space .
it shows that quantum games certainly have features of interest to applied economists . reciprocating with it we showed that games played by firms in economic competition can give counter - intuitive solutions when played in a quantum world .
we conclude our results as follows .
a comparison between the ne in cournot game with backwards - induction outcome in classical stackelberg duopoly shows that having alice ( or firm @xmath28 who acts first ) know that bob ( or firm @xmath30 who acts second ) knows @xmath16 ( alice s move ) hurts bob .
in fact in classical stackelberg game bob should not believe that alice has chosen its stackelberg quantity i.e. @xmath106 .
we have shown that there can be a quantum version of stackelberg duopoly where bob is not hurt even if he knows the quantity @xmath16 chosen by alice .
the backwards - induction outcome of this quantum game is the same as the ne in classical cournot game where decisions are made simultaneously and there is no such information that hurts a player .
though this outcome in quantum game is obtained for a restricted range of the variable @xmath97 but it is only because of a simplification in calculations .
this work is supported by pakistan institute of lasers and optics , islamabad .
j. eisert , m. wilkens , m. lewenstein .
quantum games and quantum strategies .
83 , 3077 ( 1999 ) .
quant - ph/9806088 .
also j. eisert , m. wilkens .
47 ( 2000 ) 2543 .
quant - ph/0004076 | in economics duopoly is a market dominated by two firms large enough to influence the market price .
stackelberg presented a dynamic form of duopoly that is also called ` leader - follower ' model .
we give a quantum perspective on stackelberg duopoly that gives a backwards - induction outcome same as the nash equilibrium in static form of duopoly also known as cournot s duopoly .
we find two qubit quantum pure states required for this purpose .
pacs : 03.67.lx 02.50.le 87.23.kg keywords : quantum game theory , nash equilibrium , backwards - induction outcome , stackelberg and cournot models of duopoly game . |
now generally accepted that about half of abundances of chemical elements heavier than iron are produced by slow neutron captures ( s process ) in the deep layers of asymptotic giant branch ( agb ) stars .
mixing brings freshly synthesized heavy elements to the stellar surface ( the third dredge - up process ) and then stellar wind carries them away in the stellar enviroments .
this is a simplified description of the processes , the sequence of which enriches both circumstellar and interstellar medium with heavy metals . for more detail
see , e.g. , @xcite and references there .
agb stars are therefore the principal suppliers of heavy metals and important suppliers of carbon and nitrogen to the interstellar medium , thereby participating in the chemical evolution of galaxies . in this article , we shall focus on the first observational evidences on the presence of circumstellar spectral features of heavy metals in the optical spectra of nearest descendants of agb - stars protoplanetary nebulae ( ppne ) .
ppne are objects at the post asymptotic branch ( post agb ) stage of evolution . these descendants of agb stars are low mass cores with typical masses of 0.6m@xmath0 surrounded by an extended and often structured gaseous dusty envelope , which formed as a result of substantial mass loss by the star during the preceding evolutionary stages .
central stars are usually surrounded by envelopes in the form of extended haloes , arcs , lobes , and tori .
some stars exhibit various combinations of the above features as well as bipolar and quadrupolar nebulae with dust bars .
examples of the latter two types include the egg nebula = rafgl 2688 and _ iras_19475 + 3119 nebulae , for which high - resolution images were taken by the hubble space telescope @xcite .
all ppne and about 80% of planetary nebulae are asymmetric @xcite .
much remains to be understood about condensation of dust particles and formation of the dust fraction in agb star envelopes ( see @xcite and references therein ) .
a circumstellar gas and dust envelope shows up in the form of peculiarities in the ir , radio , and optical spectra of post - agb supergiants .
the optical spectra of ppne differ from those of classical massive supergiants by the presence of molecular bands superimposed onto the spectrum of an f g supergiant and by the anomalous behavior of the profiles of selected spectral features .
these may include complex emission and absorption profiles of hi , nai , and hei lines , and metal emission features ( see for detail @xcite and references therein ) .
furthermore , all these peculiarities are variable . in whole
we see that the known types of spectral features in the optical spectra of post agb stars are : 1 ) low or moderate intensity symmetric metal absorptions without apparent distortions ; 2 ) complex profiles of neutral hydrogen lines , which vary with time and include absorption and emission components ; 3 ) absorption or emission bands of mostly carbon containing molecules ; 4 ) envelope components of the nai and ki resonance lines , and also 5 ) narrow permitted or forbidden metal emission lines that form in envelopes .
the presence of type 25 features is the key difference of the spectra of ppne from those of massive supergiants .
here we analyze the manifestations of circumstellar envelopes in the optical spectra of ppne focusing on the homogeneous subsample of stars whose atmospheres , according to previous studies , underwent evolutionary variations of the chemical composition .
section 2 briefly describes the employed observational data and lists the studied stars and their basic parameters . in section3
we analyze the available data on the peculiarities of the profiles of metal lines , found in high - resolution spectra , as well as data on the presence of molecular bands and outflow velocities for objects with different envelope structures . in sections 4 and 5
we discuss the obtained results and summarize the main conclusions .
over the past two decades more than 40 post - agb candidates supergiants with ir excesses and several related luminous stars with unclear evolutionary status have been spectroscopically monitored with the 6m telescope of the special astrophysical observatory . as a result
, a collection of high - quality spectra has been acquired with the primary purpose of searching for anomalies of stellar chemical composition due to the nucleosynthesis of chemical elements in the interiors of low- and intermediate - mass stars and the subsequent dredge - up of the synthesis products to the surface layers of stellar atmospheres .
these observational data are also used to search for peculiarities in the ppne spectra , to analyze the velocity fields in the atmospheres and envelopes of these stars with mass loss , and to search for the likely long - term spectral and radial velocity pattern variations .
here we use the data acquired in the nasmyth focus with the nes @xcite echelle spectrograph .
the nes spectrograph , equipped with a 2048@xmath12048 ccd and an image slicer , produces a spectroscopic resolution of r@xmath2 . since 2011
the nes spectrograph has been equipped with a 2048@xmath14096 ccd which made it possible to significantly extend the wavelength coverage .
the spectra of the faintest program objects ( the optical component of the ir sources _ iras_04296 + 3429 and 20000 + 3239 ) were acquired with the pfes echelle spectrograph mounted in the primary focus of the 6-m telescope @xcite .
this spectrograph , equipped with a 1k@xmath11k ccd , produces a spectroscopic resolution of r@xmath315000 .
we described the details of spectrophotometric and position measurements of the spectra in our earlier papers , the corresponding references can be found in the original papers listed in the paper by @xcite .
> c| > l| > l| > l| >
l| > l object & morphology & type of & + & of the & the c@xmath4 & & & + & envelope@xmath5 & bands & co & c@xmath4 & baii + 04296 + 3429 & bipolar+ & abs & 10.8@xmath6 & 7.7@xmath7 & + [ -10pt ] & halo+ bar & emis & & 12@xmath8 & + 07134 + 1005 & elongated & abs & 10.2@xmath6 & 8.3@xmath7 & + [ -10pt ] & halo & abs & & 11@xmath9 & + 08005@xmath102356 & bipolar & uncertain & 100:@xmath11 & 43.7@xmath7 & + [ -10pt ] & & abs & & 42@xmath12 & + 19500@xmath101709 & bipolar & no & 17.2 , 29.5@xmath6 & & 20 & 30@xmath13 + [ -10pt ] & & & 10 , 3040@xmath14 & & + 20000 + 3239 & elongated & abs & 12.0@xmath15 & 12.8@xmath7 & + [ -10pt ] & halo & abs & & 11.1@xmath16 & + rafgl2688 & multipolar+ & abs & 17.9 , 19.7@xmath17 & 17.3@xmath7 & + [ -10pt ] & halo+arcs & emis & & 60@xmath18 & + 22223 + 4327 & halo+ & abs .
& 1415@xmath17 & 15.0@xmath7 & + [ -10pt ] & small lobes & emis & & 15.2@xmath19 & + 22272 + 5435&elongated+ & abs & 9.19.2@xmath6 & 9.1@xmath7 & + [ -10pt ] & halo+arcs & abs & & 10.8@xmath20&10@xmath20 + 23304 + 6147 & quadrupole+ & abs & 9.210.3@xmath6 & 13.9@xmath7 & + [ -10pt ] & halo+arcs & emis & & 15.5@xmath21&15.1@xmath21 + + [ -10pt ] + [ -10pt ] + [ -10pt ] + [ ppn ]
our comprehensive study of the program stars allowed us to determine ( or refine ) their evolutionary status .
one of the results of our analysis is that the studied sample of luminous stars with ir excesses is not homogeneous @xcite . in this paper
we consider the peculiarities of the optical spectra of post agb stars paying special attention to the subsample of objects listed in table[ppn ] .
it contains _ iras _ objects with central stars whose atmospheres are overabundant in carbon and heavy metals .
their circumstellar envelopes have a complex morphology and are usually rich in carbon , as evidenced by the presence of c@xmath4 , c@xmath22 , cn , co , etc .
molecular bands in their ir , radio , and optical spectra . presence and type of c@xmath4 bands for stars in table[ppn ]
are published by @xcite ; @xcite . furthermore ,
the objects from table[ppn ] are among those few ppne whose ir spectra exhibit the so far unidentified emission band at 21@xmath23 m @xcite . despite an extensive search for appropriate chemical agents , so far no conclusive identification has been proposed for this rarely observed feature .
however , its very presence in the spectra of ppne with carbon enriched envelopes suggests that this emission may be due to the presence of a complex carbon - containing molecule in the envelope ( see @xcite for details and references ) .
overabundances of carbon and heavy - metal [ s@xmath24fe ] overabundance ( or lack thereof ) in the atmosphere of the central star were published earlier by @xcite for _ iras_04296 + 3429 , 07134 + 1005 , 08005@xmath102356 , and 23304 + 6147 respectively ; @xcite for _ iras_20000 + 3239 ; @xcite and @xcite for rafgl 2688 ; @xcite for v448 lac and @xcite for v354 lac .
chemical abundances for five stars from table[ppn ] were also published by @xcite . as follows from table[split ] , all stars with excess carbon and heavy metals are objects with moderate iron deficiency and may belong to the thick disk of the galaxy .
the spectra of protoplanetary nebulae with f k - type supergiant as central stars that have carbon - enriched atmospheres show features of carbon - containing molecules c@xmath4 ,
c@xmath22 , cn , and ch@xmath25 .
position measurements of molecular features in the spectra indicate that they form in expanding circumstellar envelopes .
authors @xcite and @xcite used high - resolution optical spectra to analyze molecular bands for several post agb stars including some objects from table[ppn ] .
it appears that the emission in the swan bands or nai d lines is observed in the spectra of ppne with bright and conspicuously asymmetric circumstellar nebulae .
the results of spectroscopic observations of several ppne confirm this hypothesis .
an analysis of the spectra taken with the 6m telescope revealed c@xmath4 swan emission bands of different intensities ( relative to the continuum ) in the spectra of the central stars of the following sources : _ iras _ 04296 + 3429 @xcite , 08005@xmath102356 @xcite , @xmath26 2688 @xcite , _ iras _ 22223 + 4327 @xcite , and _ iras _ 23304 + 6147 @xcite . according to hst images @xcite ,
these objects have structured ( and often bipolar ) envelopes .
in addition to a sample of related c
rich objects with the above features , table[ppn ] also includes the infrared source _ iras _ 08005@xmath102356 .
this object has so far been poorly studied , and no data are available either on the peculiarities of the chemical composition of its atmosphere or on the presence of the 21-@xmath23 m band .
however , _ iras _ 08005@xmath102356 can be viewed as related to the objects of this sample because its optical spectrum exhibits c@xmath4 swan bands , the hydrogen and metal lines in its spectrum have emission - absorption profiles @xcite , and the circumstellar envelope is observed in co emission @xcite .
note that our subsample of ppne , thoroughly studied by high - resolution spectra ( table[ppn ] ) , practically coincides with the list of c rich and 21@xmath23 m protoplanetary nebulae the photometric and spectral properties of which were extensively studied in papers @xcite . of fundamental importance to us is the conclusion in @xcite about the rare occurrence of binaries among post agb stars of the type considered , which is based on the long - term investigation of the velocity field in ppn atmospheres conducted by the above authors .
thus , the results of @xcite provide further evidence for the homogeneity of the ppne subsample considered here .
the systematic monitoring of ppn candidates , which we performed with a high spectral resolution , has released a new result an unknown earlier peculiarity : the splitting ( or asymmetry ) of strongest metal absorptions due to distortion by envelope features .
now such peculiar profiles of the strongest absorptions were revealed in the spectra of five stars listed in table[split ] : cy cmi @xcite , v354 lac @xcite , v448 lac @xcite , v5112 sgr @xcite and cgcs 6918 ( _ iras _ 23304 + 6147 ) @xcite .
let us illustrate this effect using the example of the spectrum of the high - latitude supergiant v5112 sgr , where it is most pronounced .
figure[v5112sgr - ba4934fr ] shows a fragment of the spectrum of v5112 sgr taken on july 7 , 2001 with a split baii 4934 line . at the same time , the profiles of strong absorptions of iron - group metals in the spectrum of this star are neither asymmetric nor split , as is immediately apparent from the same fig.[v5112sgr - ba4934fr ] , where we see a very strong but unsplit feii 4924 line .
besides , a comparison of the line profiles in the spectra of v5112 sgr taken during different nights reveals substantial variability of the profile shapes and of the positions of the components of the split lines . to illustrate the variability effect
, we show in fig.[v5112sgr - ba4934 ] the baii 4934 line profile , which is most asymmetric and most variable .
the different widths of the components are immediately apparent in both figs.[v5112sgr - ba4934fr ] and [ v5112sgr - ba4934 ] : the red component is about twice broader than the blue components , which are offset substantially relative to the systemic velocity .
this difference between the component widths indicates that the red and blue components form under different physical conditions .
it follows also from fig.[v5112sgr - ba4934fr ] that the position of the photospheric ( red ) component of the complex profile is variable , whereas the blue components , which , as shown in @xcite , form in the envelope , are stable .
nai the bold line and baii6141 the the thin line , width=241 ] > r > c > c > c > c > c star & _ iras _ & teff , _ k _ & [ fe / h]@xmath0 & [ s / fe]@xmath0 & ref + cy cmi & 07134 + 1005 & 7000 & @xmath27 & + 1.4 & a + [ -10pt ] v5112 sgr & 19500@xmath101709 & 8000 & @xmath28 & + 1.1 & b + [ -10pt ] v448 lac & 22223 + 4327 & 6500 & @xmath29 & + 0.9 & b + [ -10pt ] v354 lac & 22272 + 5435 & 5650 & @xmath30 & + 1.2 & c + [ -10pt ] cgcs 6918 & 23304 + 6147 & 5900 & @xmath28 & + 1.2 & d + + [ split ] the semiregular variable v354 lac is the closest analog to v5112 sgr among the objects listed in tables[ppn ] and [ split ] in terms of the structured envelope and chemical abundances .
the spectroscopic monitoring of this star @xcite carried out at the sao 6m telescope with a resolution of @xmath31 also revealed the splitting of the strongest absorptions with the low - level excitation potential of @xmath32 ev .
an analysis of the kinematic pattern showed that the blue component of the split line forms in the powerful gas and dust envelope of v354 lac .
this splitting shows up most conspicuously in the profile of the strong baii 6141 line .
the shift of the blue component of the baii line coincides with that of the circumstellar component of nai d lines , which forms in the same layers as the circumstellar c@xmath4 swan bands .
this coincidence indicates that the complex profile of the baii 6141 line contains , in addition to the atmospheric component , a component that forms in the circumstellar envelope . such splitting ( or the profile asymmetry due to the more shallow slope of the blue wing )
is also observed for other baii ( @xmath334554 , 5853 , 6496
) lines as well as for the strong yii 5402 , laii 6390 , and ndii 5234 , 5293 lines .
the lines of these ions in the spectrum of v354lac are enhanced to the extent that their intensities are comparable to those of neutral - hydrogen lines .
for example , the equivalent width of the baii 6141 line reaches @xmath34 , and that of h@xmath35 reaches @xmath36 . as follows from figs.[v448lac - ba6141 ] and [ profiles-23304 ] , in the spectra of two stars
v448 lac and cgcs 6918 a central star of _
iras_23304 , the baii 6141 line profile is or asymmetric or split for different moments of observations .
note that the variations are displayed only by the short - wavelength profile wings and the position of the core , whereas the positions and intensities of the long - wavelength wings of these strong baii , laii , and yii absorption lines do not vary in time .
asymmetric and variable profiles of strongest lines ( yii , baii , feii and other strong absorption lines ) formed in the expanding stellar atmosphere were also found in @xcite in the spectra of cy cmi . however , as follows from fig.5 in the paper by these authors , in the spectra of cy cmi profiles are just only asymmetric , but not split .
thus , we see a difference between the peculiarity types of the profiles of three stars
v5112 sgr , v354 lac , and cgcs 6918 with split profiles of the strongest absorptions of selected elements , and two stars
cy cmi and v448 lac with asymmetric but unsplit profiles .
such a difference permit us to suggest that the morphology of the circumstellar envelope may be the main factor that causes the peculiarity and variability of the profiles of the strongest lines . as is evident from table[ppn ] , the two stars v5112 sgr , v354 lac and cgcs 6918 with split absorptions have bipolar ( or quadrupolar ) envelopes , whereas the absorptions are unsplit in the spectra of cy cmi and v448 lac the envelope of these stars have a less structured environment .
this hypothesis is further corroborated by the three - component structure of the strong absorption profiles observed in the spectrum of v5112 sgr , where co observations show both the slow ( v@xmath37=10kms@xmath38 ) and the fast ( @xmath39@xmath40kms@xmath38 ) expansion @xcite . as follows from data in table[ppn ] , the profiles of the split lines include a photospheric component and two envelope components , one of which , like in the case of the co profile , arises in the envelope that formed at the agb stage and expands at a velocity of v@xmath41 kms@xmath38 , and the other one arises in the envelope that moves at a velocity of v@xmath42 kms@xmath38 and formed later .
worth mentioning here is another interesting phenomenon which was revealed in the spectrum of v5112 sgr .
klochkova @xcite showed that the optical spectrum of this high latitude post
agb supergiant contains weak absorptions whose positions are indicative of their formation in the circumstellar envelope .
the velocities v@xmath43(db ) were measured in the 57806379 wavelength interval from the positions of reliably identified 5780 , 5797 , 6196 , 6234 , and 6379 features .
the mean v@xmath43(dib ) averaged over several spectra and determined with an accuracy better than @xmath44 kms@xmath38 agrees excellently with the velocity determined from the blue component of the nai d lines .
the resulting agreement leads us to conclude that the weak bands found in the spectrum of v5112sgr origin in the circumstellar envelope .
kipper @xcite independently came to similar conclusions for the same object based on the spectra of v5112 sgr taken with a different instrument .
this dibs identification at the present time is a unique result in the task of searching for these spectral features in the circumstellar environment .
note that the splitting of strong absorptions were observed in the spectra only those post
agb stars that are listed in the table[split ] , in atmospheres which were found significant excess carbon and heavy metals .
the second precondition for the splitting of the strong absorptions in the spectrum of these post agb stars is presence of a complex structured envelope .
we can suggest that the process of the formation of a circumstellar envelope may contribute to the enrichment of this envelope by products of stellar nucleosynthesis .
the profiles of the split lines contain a photospheric and one two envelope components , one of which , like in the case of the co profile , arises in the envelope that formed at the agb stage , and the other one in the envelope that formed later .
the circumstellar components of strong heavy - metal absorptions have been conclusively identified in the spectra of v5112 sgr @xcite , v354 lac @xcite and cgcs 6918 @xcite .
the structure of the circumstellar nebulae of v354lac and cgcs6918 may be more complex than it appears in the hst observations ( see table[ppn ] ) .
polarimetric observations of v354lac @xcite indicate the presence of a ring structure embedded in an extended nebula .
nakashima et al .
@xcite point out that the axes of the optical and infrared images of the nebula are almost perpendicular to each other . based on the kinematic pattern of the nebula as determined from the mapping of the co emission ,
authors @xcite concluded that the structure of the nebula includes not only a torus and a spherical component but also another element ( possibly a jet ) .
currently , no consensus has been reached concerning the development of deviations from spherical symmetry in ppne .
authors @xcite analyzed high spatial resolution optical images of a sample of ppne taken by the hubble space telescope and concluded that the optical depth of the circumstellar matter is the crucial factor that determines the formation of a particular morphology of stellar envelopes .
the dense and often spherical envelope that formed during the agb stage is believed to expand slowly , whereas the rapidly expanding feature is the axisymmetric part of the envelope that formed later at the post
agb stage @xcite . the sequence of these processes results in the development of an optical depth gradient in the direction from the equator to the polar axis of the system .
the presence of a companion and/or a magnetic field in the system may also prove to be the physical factor that causes the loss of the spherical symmetry of the stellar envelope during the short evolutionary interval between the agb and post agb stages ( see @xcite and references therein ) . in their recent paper @xcite proposed a simple ppn model based on a pair of evacuated cavities inside a dense spherical halo .
the above authors demonstrated that all the morphological features observed in real bipolar ppne can be reproduced by varying the available parameters ( mass density inside the cavity , its size and orientation ) of this model .
so far , the discovery of the heavy - metal enrichment of the circumstellar envelopes of the post
agb supergiants v5112 sgr , v354 lac and cgcs 6918 remain the only results . here
it should be recalled that mauron & huggins @xcite detected atomic metals ( ca , fe and upper limits for al , ti , mn and sr ) in the c rich circumstellar envelope of the agb
star cw leo , which is a central star of the ir source _ irc_+10216 .
the effect of splitting of metallic lines found requires to continue the high resolution spectroscopy of very related post agb stars .
the most promising objects could be _ iras _ 04296 + 3429 and _ rafgl _ 2688 .
it follows from table[ppn ] that the totality of properties of their envelopes coincide with those of the objects listed in table[split ] : they are c enriched and have a very complex structure .
besides , the atmospheres of the weak central stars of both sources are enriched in heavy metals @xcite . evidently , the next step in study of the discovered splitting of the profiles should be theoretical modeling of the spectra envelopes and calculation chemical abundances , which could be consider with regard to selective depletion of chemical elements in a dusty environment .
we used the results of high spectral resolution observations made with the 6m telescope to analyze the peculiarities of the optical spectra of a sample of post agb stars with atmospheres enriched in carbon and heavy s
process metals and with carbon - enriched circumstellar envelopes .
we showed that presence of the peculiarities of the line profiles ( the asymmetry and splitting of the profiles of strong absorptions ) is associated with the kinematic and chemical properties of the circumstellar envelope and the type of its morphology .
the splitting of the profiles of the strongest heavy - metal absorptions in the spectra of the v5112 sgr , v354 lac and cgcs 6918 supergiants found as a result of our observations suggests that the formation of a structured circumstellar envelope is accompanied by the enrichment of this envelope with the products of stellar nucleosynthesis .
attempts to find a definite link between the peculiarities of the optical spectrum and the morphology of the circumstellar environment are complicated by the fact that the observed structure of the envelope depends strongly on the inclination of the symmetry axis to the line of sight and on the angular resolution of the spectroscopic and direct - imaging instruments . in the spectrum of the high latitude post
agb supergiant v5112sgr were revealed weak absorptional dibs 5780 , 5797 , 6196 , 6234 , and 6379 .
their mean radial velocity v@xmath43(dibs ) determined with an accuracy better than @xmath44 kms@xmath38 agrees excellently with the velocity determined from the blue circumstellar component of the nai d lines .
this work was supported by the russian foundation for basic research ( project no.140200291a ) .
this research has made use of the simbad database , operated at cds , strasbourg , france , and nasa s astrophysics data system . | a new type of peculiarity a splitting or asymmetry of strong absorption lines , is found in the optical spectra of selected post agb stars with c rich circumstellar envelopes .
the effect is maximal in baii lines whose profile is split into two three components .
the particular components of the split absorption lines are shown to be formed in a structured circumstellar envelope , suggesting an efficient dredge up of the heavy metals produced during the preceding evolution of this star into the envelope .
we suspect that the splitting ( or asymmetry ) of the profiles of strongest absorptions with low excitation potential of the low level can be associated with the kinematic and chemical properties of the circumstellar environment and with type of its morphology . |
even though the existence of hydrogen fine structure lines and their explanation using dirac s atomic theory has been known for close to a century , a hydrogen fine structure line has never been detected from an astrophysical object .
an interesting hydrogen fine structure line is the @xmath4 transition .
the main difficulty in detecting this line is that the line strength is proportional to the population of either the @xmath5 or @xmath6 states which , being excited states , are not so readily populated in most astrophysical circumstances .
moreover , the line width of the excited @xmath5 state , which is determined by its decay time , is large ( 99.8 mhz ) , making the detection of the fine structure line a difficult observation .
one astrophysical setting where the feasibility of detecting such a line has been studied is the interstellar h ii regions ( see , _ e.g. _ , dennison , turner , & minter 2005 and references therein ; ershov 1987 ) ; in h ii regions , the excited levels are populated by recombination .
here we consider two cosmological settings in which the excited levels are populated by either recombination or pumping by lyman-@xmath7 photons from an external source : ( a ) _ the recombination epoch _ : the universe makes a transition from a fully ionized to an almost fully neutral medium at @xmath8 ( spergel et al .
2006 ; for details see , _ e.g. _ , peebles 1993 and references therein ) . during this era , as the density and temperature of the universe drops , recombination is stalled owing to a high lyman-@xmath7 radiation density and progresses either by the depopulation of the @xmath5 state owing to redshifting of the photons out of the line width or the 2-photon decay of the @xmath6 state .
this results in a significant @xmath6 and @xmath5 level population during the recombination era .
( b ) _ the reionization epoch _ : recent observations suggest that the universe made a transition from nearly fully neutral to fully ionized within the redshift range @xmath9 ( page et al .
2006 ; white et al . 2003
; fan et al . 2002 ; djorgovski et al .
2001 ; becker et al .
it is widely believed that this ` reionization ' was achieved by the percolation of individual h ii regions around the sources of reionization .
the nature of these sources is not well understood : they might be pop iii stars , active galactic nuclei or star - forming galaxies . during this epoch ,
a signal from the @xmath10 fine - structure transition might originate from either within the cosmological h ii regions or from the almost fully neutral medium surrounding the h ii region .
the level population of the first excited state in the former case would be largely determined by recombinations and in the latter case by lyman-@xmath7 photons from the central source .
we shall show below that for most cases of interest the fine - structure line from within the cosmological h ii region might be negligible as compared to the signal from the regions immediately surrounding the h ii region . throughout this work
we adopt the currently - favoured @xmath11cdm model : spatially flat with @xmath12 and @xmath13 ( spergel et al .
2006 ; riess et al . 2004 ; perlmutter et al .
1999 ) with @xmath14 ( spergel et al . 2006 ; tytler et al .
2000 ) and @xmath15 ( freedman et al . 2001 ) .
subsequent to the recombination of the primeval baryon gas at redshift @xmath16 ( spergel et al .
2006 ) and the transformation of the gas to an almost completely neutral state , it is believed that the gas was reionized during epochs corresponding to the redshift range @xmath9 .
wmap measurements of cosmic microwave background radiation ( cmbr ) anisotropy in total intensity and polarization have been used to infer that the baryons were likely neutral at redshifts @xmath17 ; however , the detection of cmb polarization anisotropy requires substantial ionization by about @xmath18 ( page et al .
observationally , the gunn - peterson ( gp ) test shows that the universe is highly ionized at redshifts lower than @xmath19 ; the detection of gp absorption at greater redshifts suggests that the neutral fraction of the intergalactic hydrogen gas rises to at least @xmath20 in the redshift range @xmath21 , and that reionization was not complete till about @xmath22 ( white et al .
2003 ; fan et al . 2002 ; djorgovski et al . 2001
; becker et al .
however , from the gp test alone it is not possible to infer the neutral fraction of the medium ; it only gives a rather weak lower limit of @xmath23 on the neutral fraction of the universe for @xmath24 .
from other considerations it is possible to put more stringent bounds on the neutral fraction ; for example , wyithe & loeb ( 2004 ) obtain a lower limit of @xmath25 on the neutral fraction of the universe at @xmath0 ( see also mesinger & haiman 2004 ) .
our understanding of the nature of the sources that caused the reionization is far from complete .
the transition from an almost completely neutral gas to a highly ionized gas during redshifts @xmath9 is a key problem in modern cosmology and considerable theoretical and experimental efforts are currently directed at this unsolved problem .
here we propose a new method , based on the @xmath10 fine - structure transition , to determine the evolution of the neutral fraction of the intergalactic medium within this epoch . owing to fine structure splitting ,
the two possible transitions between the @xmath6 and the @xmath5 states are : @xmath26 at a frequency @xmath27 that has an einstein @xmath28 coefficient @xmath29 and @xmath30 at a frequency @xmath31 that has an einstein @xmath28 coefficient @xmath32 .
the einstein @xmath28 coefficient for the latter transition is more than an order of magnitude greater than the former ; therefore , in this work we consider only the @xmath30 transition and hereinafter we refer to this specific transition simply as the @xmath4 transition .
the ionizing uv photons from sources in the reionization era create ` stromgren spheres ' .
whereas the gas in the cosmological h ii regions are highly ionized by the photons , the ionization level of the gas beyond the stromgren spheres is determined by the history of the gas , the density , and the mean specific intensity of the background ionizing photons , which includes both the photons diffusing out of the stromgren spheres as well as the background radiation field . in this work
we assume that the ionizing sources at these high redshifts are agns and in illustrative examples adopt parameters of a few qsos that have been observed at @xmath33 .
the photons at the lyman-@xmath7 transition frequency ( here and throughout , unless otherwise specified , we shall continue to refer to frequencies between lyman-@xmath7 and lyman - limit as lyman-@xmath7 ) from a high redshift qso escape the mostly - ionized stromgren sphere and are strongly scattered and absorbed in the medium beyond .
the population of the @xmath5 level in this region is determined by ( a ) the intensity of lyman-@xmath7 photons from the central source , ( b ) recombination rate of free electrons , ( c ) absorption of cmbr photons by electrons in the @xmath6 state ( it is assumed here and throughout this work that the only radio source at high redshifts is the cmbr ) and ( d ) collisional transition from the @xmath6 state .
the @xmath6 state is populated via ( a ) the recombination rate of free electrons , ( b ) collisional transfer of atoms from the @xmath5 state ,
( c ) the spontaneous decay of the @xmath5 state , and ( d ) transition from the @xmath5 state stimulated by cmbr photons . additionally , the absorption of photons from the central sources , with energy equal to or in excess of the lyman-@xmath34 transition , would result in electronic transitions to the second excited state , which could be followed by spontaneous decay to the @xmath6 state .
( it might be pointed out here that both the @xmath6 and @xmath5 states could also be populated by atoms cascading from excited states with @xmath35 . in particular , all photons absorbed from @xmath36 states to any excited state can directly de - excite to the @xmath6 level .
however , the rate of transition from @xmath36 to any excited state is roughly @xmath37 ( e.g. rybicki & lightman 1979 ) and , therefore , we include only the most dominant transition in each case . ) the population of the ground state is denoted by @xmath38 .
we denote the level populations of the the two states@xmath39 and @xmath40by the number densities @xmath41 and @xmath42 ; these may be solved for , respectively , from the following two equations of detailed balance : @xmath43 @xmath44 here @xmath45 is the fraction of all the atoms that recombine to the @xmath6 state . in equilibrium @xmath46 as the @xmath47 state splits into three doublets : @xmath48 , @xmath40 and @xmath39 .
@xmath49 is the recombination coefficient and @xmath50 is the density of the ionized gas .
@xmath51 is the einstein @xmath52 coefficient for the @xmath30 transition in terms of the corresponding einstein @xmath28 coefficient @xmath53 ( note that the two @xmath52 coefficients are equal as the two states have the same degeneracy ) .
@xmath54 is the rate coefficient of transition owing to collisions with electrons .
@xmath55 is the number density of cmbr photons within the transition line width .
@xmath56 is the number density ( per unit frequency ) of photons with frequency equal to or larger than the lyman-@xmath7 frequency ( and smaller than the lyman - limit frequency ) at any location ; @xmath57 and @xmath58 are , respectively , the line profiles corresponding to the lyman-@xmath34 and lyman-@xmath7 transitions , and @xmath59 is the probability for the electron transition to the @xmath6 state following excitation to @xmath60 via absorption of a lyman-@xmath34 photon . we have not included the induced lyman-@xmath7 transition because the number density of atoms in the @xmath5 state is negligible as compared to that in the @xmath36 state .
@xmath59 is the probability that an atom in the third excited state ( @xmath61 ) will decay to the @xmath6 state .
@xmath62 is the einstein @xmath28 coefficient corresponding to the 2-photon decay of the @xmath6 state .
other symbols have their usual meanings . owing to the fact that the mostly neutral medium in the vicinity of the cosmological stromgren spheres is optically thick to lyman-@xmath7 scattering , the lyman-@xmath7 photons from the decay of the @xmath5 state are strongly scattered by the gas .
therefore , @xmath56 will contain contributions from both the lyman-@xmath7 photons from the central source as well as the lyman-@xmath7 photons that arise from recombinations outside the stromgren sphere and are multiply - scattered therein : @xmath63 .
we neglect the multiply - scattered lyman-@xmath7 photons from the stromgren sphere that have been reprocessed via recombination within the stromgren sphere because these would be redshifted redward of the lyman-@xmath7 line before encountering the boundary of the stromgren sphere .
the scattering of recombination photons in an optically thick , expanding medium is a complex problem ( field 1959 ; rybicki & dellantonio 1994 ) .
one of its first applications was to study the recombination of primeval plasma ( peebles 1968 ; zeldovich , kurt & sunyaev 1969 ) . in these
analyzes it was implicitly assumed that apart from 2-photon decay , in an expanding universe the dominant effect that results in resonant photons ceasing interaction with the gas , and leaving the system , is its redshifting out of the line profile .
the effect of scattering off the moving atoms was deemed to be either negligible or at best comparable .
this assumption has been borne out by more recent detailed analysis that have taken into account the effect of scattering on the photon escape ( krolik 1990 ) .
taking only the redshift as the main agent of photon escape , it can be shown that the net effect of the scattering of a resonance photon before it drops out of consideration is to reduce the decay time of the @xmath5 state from @xmath64 to @xmath65 ( zeldovich et al .
1969 ; for more recent work see , _
e.g. _ , chluba , rubino - martin & sunyaev 2007 , seager , sasselov & scott 1999 and references therein ; we give a concise derivation in appendix a ) . in this expression , the gunn - peterson optical depth @xmath66a_{2p1s}\lambda_\alpha^3 n_{1s}$ ] .
similar complications exist in computing @xmath59 , the probability that an atom in the third excited state will decay spontaneously to the @xmath6 state , in an optically thick medium . in optically thin media , @xmath67 ;
on the other hand , in an optically thick medium , we would need to take into account the ` trapping ' of the lyman-@xmath34 photon owing to resonant scattering .
the effect of this scattering in an optically thick medium would be that the fraction of photons that decay directly to the ground state are reabsorbed ` locally ' to the third excited state and , therefore , all photons absorbed to the third excited state result in an h@xmath7 photon and an atom in the @xmath6 state .
this means that the appropriate value of @xmath59 is close to unity in an optically thick medium : in this work we assume @xmath68 .
the astrophysical setting in which we seek solutions to the algebraic equations above is cosmological h ii regions at high redshift . in particular , we are interested in the signal from the neutral region surrounding the cosmological h ii region . for a fully neutral inter - galactic medium ( igm ) at @xmath69 , @xmath70 . if we adopt spectral luminosities corresponding to qsos observed at these high redshifts , it may be shown that the populating of the @xmath5 state via direct recombinations from the free - free state , collisional transfer from the @xmath6 state , and upward transitions from the @xmath6 to @xmath5 state arising from absorption of background cmbr photons may all be neglected .
( the relevant parameters are : @xmath71 in the h ii region surrounding the sources , with @xmath50 expected to be much smaller in the neighboring mostly neutral medium ; the number density of cmbr photons that might cause a @xmath4 transition is @xmath72 ; the number density of lyman-@xmath7 photons from the central source , assuming luminosities typical of sdss quasars at @xmath73 ( more details in 4 ) , is @xmath74 .
first , for these parameters , the dominant process that populates the excited state is the pumping by lyman-@xmath7 photons .
second , it may be readily verified that for these plausible values for the parameters the signal expected from the h ii region surrounding the central source is negligible as compared to the signal from the surrounding neutral region . )
given that the lyman-@xmath7 flux from the central qso is the dominant causative factor for populating the @xmath5 state , the two equations ( [ levpops ] and [ levpopp ] ) that determine the level populations is simplified .
the number density of atoms in the @xmath5 state is given approximately by : @xmath75 where @xmath76 , with @xmath77 denoting the neutral fraction and @xmath78 @xmath79 is the number density of baryons in the igm .
@xmath80 ( in units of @xmath81 ) is the transition rate to the @xmath5 state owing to the lyman-@xmath7 photons from the central source , where @xmath82 is the number density of lyman-@xmath7 photons from the central source alone .
similarly , the dominant process that determines the population of the @xmath6 state is the absorption of lyman-@xmath34 photons ( for details , see the discussion above and 4 ) and the subsequent decay to the @xmath6 state .
the number density of atoms in the @xmath6 state is : @xmath83 where @xmath84 ( in units of @xmath81 ) is the transition rate to the @xmath6 state owing to the lyman-@xmath34 photons from the central source .
the universe made a transition from fully ionized to nearly fully neutral at @xmath8 ( spergel et al .
2006 ; peebles 1968 ; zeldovich et al . 1969 ) .
this transition is mainly accomplished by the 2-photon decay of the @xmath6 state and the slow redshifting of the lyman-@xmath7 photons which deplete the @xmath5 state ( see peebles 1993 and references therein for a detailed discussion ) .
the saha ionization formula , valid for thermodynamic equilibrium conditions between the hydrogen level populations and free electrons , is a poor approximation for studying the epoch of recombination .
a good approximation for studying this transition is to assume that all states , excepting @xmath36 , are in equilibrium with the cmbr ( matter temperature to a very good approximation remains equal to the cmbr temperature throughout this transition ) ( seager et al .
1999 ; peebles 1968 ) .
however , in this approximation , where the matter temperature and all transitions , excepting the lyman-@xmath7 line , are in thermal equilibrium , the @xmath4 signal is unobservable because the excitation temperature for this transition equals the background radiation temperature . even though this second approximation might be useful for studying the evolution in ionization fraction
, it will be strictly true only if the dominant mechanisms that determine the level populations of the @xmath5 and the @xmath6 states are either interaction with the cmbr photons or collisions between atoms .
as there are a variety of other processes relevant to the determination of the level populations for example , the free decay of either of the two states a small deviation from equilibrium is expected in the @xmath6 and @xmath5 level populations and it is our aim here to compute it .
one approach to this problem is to simultaneously solve for the level populations of the @xmath5 and the @xmath6 states as well as the change in the ionization fraction .
however , assuming thermal equilibrium between these two states is a good approximation for solving the evolution of ionization .
therefore , the approach we have adopted is to solve for the evolution of ionization using the method of peebles ( 1968 ) , and use the resulting ionized / neutral fraction to solve for the populations of the @xmath6 and the @xmath5 states using detailed balance .
the resulting equations are : @xmath85 and @xmath86 where @xmath87 and @xmath88 are , respectively , the ionization potentials of the ground and the first excited states , @xmath89 is the boltzmann s constant , and @xmath90 is the temperature of the cmbr .
@xmath91 is the rate at which the cmbr photons cause a bound - free transition of electrons from the @xmath47 state ( @xmath6 or @xmath5 states ) .
the various other terms in these equations have the same meanings as in the previous case .
the main difference is that the cmbr photons and baryonic matter at the time of recombination are hot and dense enough to directly affect the level populations of the excited states by ionizing the excited state , and the two - photon capture to the excited state is not completely negligible ( see peebles 1993 and 1968 for details of the different physical processes that are relevant at this epoch ) .
the brightness temperature in the @xmath4 transition is : @xmath92 here @xmath93 and @xmath94 is the rest wavelength of the @xmath4 transition . @xmath95 $ ] is the excitation temperature corresponding to the transition , where @xmath96 is the planck constant .
@xmath97 is the optical depth of the source in the @xmath4 signal .
@xmath98 takes into account the selection rules for transitions between the @xmath40 and @xmath39 levels .
given the selection rules there are three allowed transitions between these two states ; they occur at frequencies @xmath99 , 9875 , and 10029 mhz ( see , _
e.g. _ , ershov 1987 ) .
the first two transitions are blended by the natural width of the line , which is approximately 100 mhz , but the third should be observable as a distinct line .
this implies that @xmath100 if the observing frequency is @xmath101 mhz . for computing the strength of this signal in a typical case , we adopt observed parameters of qso sdss j1030 + 0524 ( see , _
e.g. _ , wyithe & loeb 2004 and references therein ) , which is at redshift @xmath102 and shows no detectable flux beyond the qso stromgren sphere : the radius of the stromgren sphere has been estimated to be @xmath103 ( mesinger & haiman ( 2004 ) argue that the size of the stromgren sphere could be roughly 30% higher ; this makes no essential difference to our results ) .
it may be noted here that the qso sdss j1148 + 5251 also has similar parameters .
we arrive at an estimate of @xmath82 by assuming that @xmath104 photons per second , with wavelengths corresponding to the lyman-@xmath7 transition , are emitted by the central source in an effective frequency range @xmath105 , and that the photons are absorbed at a radial distance @xmath106 .
this leads to : @xmath107 we assume typical values : @xmath108 , @xmath109 and @xmath110 .
these lead to the following estimate for the transition rate , @xmath111 , in the gas at the boundary of the stromgren sphere arising due to the photons from the central qso : @xmath112 it should be noted that the mean specific intensity of lyman-@xmath7 photons in the igm would also give a non - zero signal .
assuming a mean specific intensity of @xmath113 ( this might be needed to couple the hi spin temperature to the matter temperature ; see e.g. madau , meiksin , & rees 1997 ) , the expected signal is many orders of magnitude smaller than we have computed from the outskirt of bright sources .
if we assume that the central sources are continuum emitters in the frequency range between lyman-@xmath7 and lyman - limit frequencies , we may assume similar parameter values for computing the expectations for @xmath114 .
in that case , @xmath115 where @xmath116 and @xmath117 are the oscillator strengths of the lyman-@xmath34 and lyman-@xmath7 transitions , respectively . from eqs .
( [ num2p ] ) , ( [ num2s ] ) and ( [ pho_ratio ] ) : @xmath118 for a completely neutral medium at @xmath119 , @xmath120 , which may be used to show that @xmath121 for the parameters of igm in the redshift range of interest : @xmath122 .
this implies that in the outskirts of the stromgren sphere surrounding the qso , the transition is expected to be observable as an absorption feature against the cmbr .
using eqs .
( [ num2p ] ) and ( [ num2s ] ) in eq .
( [ bri_gen ] ) , the observable brightness temperature is estimated to be : @xmath123 where @xmath77 is the neutral fraction of hydrogen outside the stromgren sphere , and might be close to unity . we have adopted parameters typical of qsos observed at the edge of the reionization epoch in estimating the above temperature decrement .
the main uncertainty above is in the estimation of the ` lyman-@xmath7 ' flux from the central source , and as the observed temperature decrement is directly proportional to this flux from the central source , this constitutes a major uncertainty in reliably computing the expectations for the signal . for qsos that have strong lyman-@xmath7 and lyman-@xmath34 lines
, the blueward side of the lines will be strongly absorbed in the medium just beyond the stromgren sphere ( and this has been observed to happen in many cases ) , provided that the blueward side photons have not been redshifted to frequencies smaller than the lyman-@xmath7 frequency while transversing the stromgren sphere . in the case of qsos that have large line fluxes and small stromgren spheres ,
the lyman-@xmath7 luminosity @xmath104 might be underestimated . using equations ( [ levpop0 ] ) and ( [ levpop1 ] ) the level populations of the @xmath5 and the @xmath6 states
may be computed .
solving for the level populations and using equation ( [ bri_gen ] ) , we have computed the expected signal from the recombination epoch ; the expected signal is shown in fig . 1 .
the fine structure line transition is expected to be an absorption feature , with a maximum temperature decrement of order 100 @xmath124k at an observing frequency of about 7 mhz .
the width of the decrement in frequency space corresponds to a redshift span @xmath125 , which is roughly the width of the visibility function at recombination .
as eq . ( [ bri_tem ] ) shows , a detection of the fine structure line in the outer regions of cosmological h ii regions is potentially a probe of the cosmological neutral hydrogen density in the vicinity of qsos in the reionization epoch , and might be a tool for the investigation of the evolution in the neutral fraction with cosmic epoch through the reionization era . given the cosmological importance of such a measurement , and the fact that there does not exist many reliable methods for the detection of h i at high redshifts ( see , _
e.g. _ , barkana & loeb 2001 ) , the detection of the @xmath4 line transition in the cosmological context assumes additional importance . in deriving equation ( [ bri_gen ] ) the hubble expansion
was assumed to be the only cause for the velocity width in the observed line .
however , an important contribution to the velocity dispersion in the line in this case is the natural width of the fine structure line : owing to the rapid decay of the @xmath5 state , the natural lorentzian width in the rest frame of the gas is @xmath126 ( see , _
e.g. _ , dennison et al . 2005 ) .
therefore , the peak brightness in the observed line profile might be suppressed by a factor @xmath127 , where @xmath128 is the line - of - sight peculiar velocity dispersion owing to the hubble flow across the region being observed .
however , in the case @xmath129 this suppression has a negligible effect . for qsos at redshift @xmath130 , the natural lorentzian width of the fine structure line is equivalent to the peculiar hubble flow across a proper line - of - sight distance of @xmath131 .
this distance is approximately the size of the stromgren spheres around qsos at that redshift ; therefore , the natural width of the line does not significantly diminish the expectations , given by equation ( [ bri_gen ] ) , for the peak brightness temperature .
a second inference is that in the case of a qso that has a smaller stromgren sphere , the increase in mean brightness temperature is roughly proportion to the inverse of the radius of the stromgren sphere : @xmath132 , and not @xmath133 .
another assumption that was made while deriving eq .
( [ bri_gen ] ) is that the only radio frequency radiation that needs to be considered for the determination of the level populations and brightness temperature decrement is the cmbr .
it may be that the central ionizing source , which may be a qso , is radio loud .
there may also exist radio sources behind the observed stromgren sphere and within the angular region over which the lyman-@xmath7 flux from the qso is appreciable .
equation ( [ bri_gen ] ) may be modified , to account for this , by replacing @xmath90 with @xmath134 , where @xmath135 is the brightness temperature of the radio source at wavelengths corresponding to the rest frequency of the transition , which is @xmath136 . the result would be to enhance the brightness temperature of the line signal by @xmath137 .
if the radio source is unresolved , it is appropriate to estimate the expected signal in terms of the optical depth .
the optical depth corresponding to equation ( [ bri_tem ] ) is @xmath138 .
we now discuss the feasibility of the detection of the fine structure line towards sdss j1030 + 0524 .
we shall assume that the neutral fraction , @xmath77 , outside the stromgren sphere of this qso is unity , consistent with the measured gp trough .
the redshifted fine structure line would be expected at @xmath139 . the observed line
` width ' , considering natural broadening and the hubble flow across the stromgren sphere , is expected to be approximately @xmath140 ; using @xmath102 , we obtain @xmath141 .
the fine structure line would be expected to originate in a shell that is roughly the size of the stromgren sphere , and fall of as @xmath142 beyond this shell , where @xmath143 is the distance from the qso . in the case of sdss j1030 + 0524 , the angular size of the stromgren sphere
is expected to be 15 arcmin .
the frequency of the redshifted line is in the observing bands , and the line width is within the spectral line capabilities , of several currently operational telescopes .
however , large - collecting - area arrays like the giant metrewave radio telescope ( gmrt ) have large aperture antennas of 45-m diameter and , therefore , poor surface brightness sensitivity for such extended structures . the australia telescope
compact array ( atca ) , with 22-m antennas , has reasonable brightness sensitivity for this problem and , additionally , operates with 128-mhz bandwidths in spectral line mode . at @xmath144 ,
the atca has a system temperature @xmath145 and antenna efficient @xmath146 .
the five movable 22-m diameter antennas may be configured into an ultra - compact 2-d close - packed h75 ( 75-m maximum baseline ) array , and this would yield a number of baselines sensitive to the 15-arcmin scale fine - structure line signal .
the brightness sensitivity of this array , for a 8-arcmin scale structure , is 400 @xmath124k in 6 hr integration time .
the brightness sensitivity in fourier synthesis images could be enhanced somewhat , by factors of a few , by appropriately weighting the baselines to match the synthetic beam to the expected structure scale ; however , the required integration times for a detection are still in the ball park of @xmath147 hr . the detection of the signal from cosmological h ii regions , however , might be feasible using facilities under construction or planned for the near future , like the xntd in australia or the square kilometer array ( ska ) .
these arrays would have smaller antenna sizes making the detection of these large - angular - scale structures detectable in interferometers and significantly more numbers of antennas , giving more numbers of short baselines that would usefully respond to the large - angular - scale fine - structure line .
however , the array configuration designs would have to factor in the extraordinary brightness temperature sensitivity requirements for this demanding observation .
as an example , the ska might consist of about @xmath148 1215 m class reflector antennas , with aperture efficiency of 60% , and the system temperature at 1.4 ghz might be about 20 k. assuming that the visibilities are optimally weighted so that the synthetic beam of the fourier synthesis array is matched to the source size of about 15 arcmin fwhm , the line strength would have a peak of about 20 @xmath124jy .
a 5-@xmath149 detection of the fine structure line towards a typical qso , in a reasonable integration time of about 1 hr , will require that about 250 baselines ( just 0.001% of the total ) be within about 50 m. the signal from the recombination epoch is observable in the frequency band 68 mhz as a broad decrement in the brightness temperature of the extragalactic background sky , and would be extremely small compared to the orders of magnitude more intense galactic non - thermal emission as well as the average low - frequency background brightness temperature arising from the numerous extragalactic radio sources .
this decrement may be considered to be a distortion to the cmbr spectrum at long wavelengths , and would be an all - sky cosmological signal . however , owing to the ionosphere , the frequency range in which this feature is expected to appear is too low a radio frequency to be easily accessible using ground - based observatories .
therefore , even though the observation of this signal might be yet another tool to probe the epoch of recombination , new custom - made instruments , which should presumably operate from space and above the ionosphere , will require to be built if a detection of this signal is to be attempted .
an additional cause for concern is that the galactic and extragalactic background radiations might have low - frequency spectral turnovers at these frequencies , as a result of free - free absorption as well as synchrotron self - absorption , and these would result in significant spectral features in the band that would require a careful modelling in order to detect any cmbr decrement feature arising from fine structure transition absorption .
interference from terrestrial man - made transmitters , as well as from auroral phenomena and solar system objects would also be an issue .
we have discussed the possibility of detecting the hitherto undetected fine structure line of @xmath4 transition in two cosmological settings : the epoch of recombination at @xmath16 and the epoch of reionization at @xmath150615 .
the expectations for the line from the environments of ionization sources in the epoch of reionization are interesting and worthy of attention as a novel tool for the investigation of the reionization process and the cosmological evolution of the gas .
the signal is expected to be observable as an extended and weakly absorbing source , which causes a decrement in the brightness of the background cmbr , and may be detected by interferometers as a negative source akin to the sunyaev - zeldovich decrements observed along the lines of sight through hot gas in clusters of galaxies .
detection of the @xmath4 line signal from the outskirts of cosmological h ii regions at different redshifts within the reionization era may serve to determine the neutral fraction of the medium during the epoch of reionization , which is a quantity of significant interest in modern cosmology .
in particular , we have computed a representative signal strength by adopting parameters typical of a qso observed at @xmath33 .
these qsos show gp troughs in their spectra ; however , owing to the weakness of the gp test , the spectra have only been useful in setting weak limits on the neutral fraction ( constraining the neutral fraction to be @xmath151 ) outside the observed h ii region .
we have shown that for a fully neutral medium the line peak may reach @xmath152 , which is potentially observable by radio interferometers that are being designed today .
other interesting probes of the reionization epoch include detecting oi line from this epoch ( hernandez - monteagudo et al .
2006 ) .
it is of interest to compare the relative difficulty associated with detecting the neutral hydrogen in this indirect way , using the redshifted fine structure line , with direct imaging of the redshifted 21-cm line from neutral hydrogen during the epoch of reionization ( see , _ e.g. _ , sethi 2005 ; zaldarriaga , furlanetto , & hernquist 2004 ) .
the ` all - sky ' h i signal might be detectable with a peak strength of @xmath153 ; however , it could be very difficult to detect owing to calibration and foreground contamination issues ( _ e.g. _ , zaldarriaga et al . 2004 ; shaver et al . 1999 , and references therein ) .
a better approach might be to attempt to detect the fluctuating component of the sky signal , which could have peak intensities of @xmath154 at observing frequencies of @xmath155 ( zaldarriga et al .
2004 ; shaver et al .
this translates to roughly the same signal strength ( specific intensity ) as we have obtained in eq .
( [ bri_tem ] ) for the fine structure line .
this is not entirely unexpected : even though the level populations of the excited states are much smaller as compared to the ground level population needed in computing the hi signal , the a coefficient of the fine structure transition we consider here is roughly 8 orders of magnitude larger than the hi hyperfine transition a coefficient . currently , there are many ongoing and planned radio interferometer experiments for the detection of the redshifted h i emission / absorption from the epoch of reionization ( _ e.g. _ , pen , wu , & peterson 2004 ; the lofar project at www.lofar.org ; the mwa project at www.haystack.mit.edu/ast/arrays/mwa/site/index.html ) .
the detection of the h i signal is firstly more difficult requiring greater sensitivity because the typical frequency width of the signal @xmath156 ( zaldarriga et al .
2004 ) , which is far smaller than the typical width expected for the fine structure line ( @xmath157 ) .
second , the fine structure line is expected at 1.4 ghz , where the sky background temperatures are significantly lower than in the @xmath158 band , making the telescope system temperatures lower .
an additional advantage of the the indirect detection is that the redshifted fine structure line appears at higher frequencies ( @xmath159 ) , which are relativity free of interference as compared to the low frequency band of @xmath158 .
the main advantage of the direct detection is that unlike the method we suggest here , it is independent of the existence of strong lyman-@xmath7 emitters at high redshifts .
another advantage of direct detection is that it may be detected ` statistically ' and such a detection might be achieved with greater ease than direct ` imaging ' ( e.g zaldarriaga et al .
2004 ; bharadwaj & sethi 2001 ) ; however , the foreground subtraction problem becomes a severe constraint for a statistical detection .
to summarize : if strong lyman-@xmath7 emitting sources are present at high redshifts , they would facilitate the indirect detection of the neutral hydrogen via enabling the detectability of the fine structure line .
the imaging issues and problems associated with the detection of this signal appears to be less of a challenge as comparison to the direct detection of redshifted h i from those epochs .
another astrophysical context in which the fine structure line might be detectable is the environments of high redshift galaxies , which are strong lyman-@xmath7 emitters .
as equation ( [ bri_tem ] ) shows , the observed signal depends on the neutral fraction outside the stromgren sphere .
therefore , detection of this signal would constitute an alternate probe of the neutral fraction of the igm at large redshifts .
the aim of this work has been to examine the detectability of the fine structure line in cosmological contexts , to point out the cosmological significance of detections , and spawn work that may refine the modelling presented herein and improve the case for appropriate design of future telescopes , which might enable the detection of the fine structure line towards multiple sources in the reionization era .
the evolution of the photon distribution function , neglecting the effect of scattering off moving atoms , is ( see _ e.g. _ rybicki & dellantonio 1994 ) : @xmath160 here @xmath161 and @xmath162 .
the equation may be written as : @xmath163 here @xmath164 and @xmath165 .
the equation above lends itself to a ready interpretation .
if the second term on the left hand side ( which is owing to the expansion of the universe ) was absent , the distribution function will approach @xmath166 on a time scale @xmath167 , where @xmath168 ( @xmath38 here has units @xmath169 ) .
it may be readily verified that for the recombination epoch and also the epoch of reionization , @xmath170 , excepting when the neutral fraction of the medium is very small .
we work here with the assumption that the neutral fraction is always large enough so that @xmath171 . in this case
the solution to eq .
( [ disfun ] ) may be simplified : to leading order the distribution function approaches @xmath166 and the first order term ( of the order of @xmath172 ) represents the slow time variation of the distribution function owing to the expansion of the universe . in this approximation , one may write the solution to eq .
( [ disfun ] ) as : @xmath173 using these equations we may proceed to prove the contention that the net effect of the ` trapping ' of lyman-@xmath7 photons is to reduce the decay time of the @xmath5 state by a factor @xmath174 . the @xmath175 transition rate , which is given by eq .
( [ levpop1 ] ) , may be solved using eq .
( [ dissol ] ) : @xmath176 the first term on the right hand side cancels with the decay term of the @xmath5 state on the right hand side of eq .
( [ levpop1 ] ) , and , therefore , the net effect of the scattering of recombination photons is to reduce the decay time of the @xmath5 state by a factor @xmath174 .
it may be pointed out that the condition needed to derive the above expression roughly translates to the condition that @xmath177 . for the reionization case
, this requires that the neutral fraction @xmath178 . in the case of primordial recombination
, it leads to an even weaker condition that the neutral fraction is @xmath179 .
one of us ( sks ) would like to thank jens chluba for many useful discussions and to zoltan haiman for many useful comments on the manuscript .
barkana , r. and loeb , a. 2001 , phys .
, 349 , 125 becker , r. h. et al.2001 , aj , 122 , 2850 bharadwaj , s. & sethi , s. k. 2001 , j. astrophys .
astr . , 22 , 293 chluba , j. , rubino - martin , j. a. , & sunyaev , r. a. 2007 , mnras , 374 , 1310 dennison , b. , turner , b. e. , & minter , a. h. 2005 , apj , 633 , 309 djorgovski , s. g. , castro , s. , stern , d. , & mahabal , a. a. 2001 , , 560 , l5 ershov , a. a. 1987 , sov .
13(2 ) , 115 fan , x. , narayanan , v. k. , strauss , m. a. , white , r. l. , becker , r. h. , pentericci , l. , & rix , h.2002 , aj , 123 , 1247 field , g. 1959 , apj , 129 , 551 freedman , w. l. et al .
2001 , , 553 , 47 hernandez - monteagudo , c. , haiman , z. jimenez , r. , & verde , l. 2006 , astro - ph/0612363 krolik , j. h. 1990 , apj , 353 , 21 madau , p. , meiksin , a. , & rees , m. j. 1997 , 475 , 429 mesinger , a. and haiman , z. 2004 , apjl , 611 , 69 page , l. et al .
2006 , astro - ph/0603450 peebles , p. j. e. 1993 , principles of physical cosmology , princeton university press peebles , p. j. e. 1968 , apj , 153 , 1 pen , u .- l . , wu , x .- p . ,
peterson , j. 2004 , astro - ph/0404083 perlmutter , s. et al . 1999 , , 517 , 565 riess , a. g. et al .
2004 , apj , 607 , 665 rybicki , g. b. & dellantonio , i. p. 1994 , apj , 427 , 603 rybicki , g. b. & lightman , a. 1979 , radiative processes in astrophysics , wiley - interscience seager , s. , sasselov , d. d. , & scott , d. 1999 , apjl , 523 , l1 sethi , s. k. 2005 , mnras , 363 , 818 shaver , p. a. , windhorst , r. a. , madau , p. & de bruyn , a. g. 1999 , a&a , 345 , 380 spergel , d. n. et al .
2006 , astro - ph/0603449 tytler , d. , omeara , j. m. , suzuki , n. , & lubin , d. 2000 , , 333 , 409 , r. l. , becker , r. h. , fan , x. & strauss , m. a. 2003 , apj , 126 , 1r wyithe , j. s. b. & loeb , a. 2004 , nature , 432 , 194 zaldarriaga , m. , furlanetto , s. r. & hernquist , l. 2004 , apj , 608 , 622 zeldovich , ya . b. , kurt , v. g. & sunyaev , r. a. 1969 , jetp lett , 28 , 146 | we propose a method to study the epoch of reionization based on the possible observation of 2p2s fine structure lines from the neutral hydrogen outside the cosmological h ii regions enveloping qsos and other ionizing sources in the reionization era .
we show that for parameters typical of luminous sources observed at @xmath0 the strength of this signal , which is proportional to the h i fraction , has a brightness temperature @xmath1 for a fully neutral medium .
the fine structure line from this redshift is observable at @xmath2 and we discuss prospects for the detection with several operational and future radio telescopes .
we also compute the characteristics of this signal from the epoch of recombination : the peak brightness is expected to be @xmath3 ; this signal appears in the frequency range 5 - 10 mhz .
the signal from the recombination era is nearly impossible to detect owing to the extreme brightness of the galactic emission at these frequencies . |
the prospect of a semiconductor laser in the infrared and thz region has been one of the key reasons for the development and study of semiconductor heterostructure elements , since the first proposal of semiconductor superlattices in 1970 @xcite .
a possible gain mechanism may be based on two different ideas : ( i ) at certain electrical fields resonant tunneling between different subbands can lead to population inversion associated with gain at the transition energy @xcite .
this idea was realized in the quantum cascade laser @xcite , which has become an important device in the infrared region .
lasing in the thz region has been demonstrated very recently as well @xcite .
for further details , see the review @xcite .
( ii ) the occurrence of negative differential conductivity in superlattices gives raise to gain in the low frequency range extending up to frequencies of the order of the bloch frequency @xcite . despite a strong effort by several groups all over the world , this idea is still not realized .
the difficulty in its realization is attributed to the instability of the operating state leading to domain formation @xcite , which may be circumvented by different strategies @xcite . for further references and a detailed discussion
see ref . .
these concepts for gain are commonly described in different ways : ( i ) an external electromagnetic field causes transitions between different levels , where the transition rate @xmath0 is evaluated by fermi s golden rule .
if stimulated emission dominates stimulated absorption ( i.e. , for population inversion ) , gain occurs at the transition frequency . for a pair of states
@xmath1 with energies @xmath2 , occupation probabilities @xmath3 , and a dipole matrix element @xmath4 , the contribution to the material gain ( i.e. , increase of light intensity per length ) is given by @xmath5 where @xmath6 is the normalization volume , @xmath7 is the electron charge , @xmath8 is the vacuum speed of light , and @xmath9 the background dielectric constant .
\(ii ) transport theory in the presence of alternating electric fields provides the complex dynamical conductance @xmath10 .
standard electrodynamics ( see , e.g. , section 7.5 of ref .
@xcite ) gives the material gain ( the negative absorption coefficient @xmath11 , which is assumed to be small compared to the wave vector here ) @xmath12 thus , gain is equivalent to a negative differential conductivity in the respective frequency range .
this concept has been frequently applied to superlattice transport .
( for superlattices eq .
( [ eqgainsimple ] ) gives zero gain in the basis of wannier stark states , as the translational symmetry gives identical @xmath13 for all wannier - stark levels . ) here the current is either evaluated from a semiclassical miniband transport model @xcite or by sequential tunneling between different layers @xcite . while these approaches use a macroscopic current density in eq .
( [ eqgainsigma ] ) , the consideration of polarization currents allows for a derivation of eq .
( [ eqgainsimple ] ) , see , e.g. , ref . . for complicated semiconductor heterostructures , with a variety of different tunneling and optical transitions
, gain may result both from optical transitions in the spirit of ( i ) and macroscopic currents in the spirit of ( ii ) .
the aim of this paper is to show the feasibility of a general approach based on a quantum transport theory which treats both mechanisms on a equal footing .
the method of nonequilibrium green functions @xcite is used here , which has been frequently applied to light emission in semiconductors ( see , e.g. , refs . and references given therein ) .
while these works consider transitions between the conduction and valence band , the focus in this paper is on transitions within the conduction band in current - driven heterostructure devices .
the calculations are performed in two steps : first the transport problem is solved by evaluating a self - consistent stationary solution of the quantum kinetic equations for a given applied bias .
this provides us with the green functions describing the electronic state far from equilibrium . in a second step , an additional weak radiation field
is taken into account .
the time - dependent kinetic equations are linearized around the stationary nonequilibrium state in oder to study the linear response of the current - driven system .
the formulation used here employs the full two - time structure of the theory without any sort of ( generalized ) kadanoff - baym ansatz @xcite .
the capability of the approach is demonstrated by calculations of the gain spectra in a quantum cascade laser and a superlattice .
we describe a general system by a set of orthonormal states labeled @xmath11 with energies @xmath14 and write the hamiltonian in the form : @xmath15 where @xmath16 and @xmath17 are electron annihilation and creation operators in the state @xmath11 .
@xmath18 represents the matrix elements of the kinetic energy and the potential part of the hamiltonian @xmath19 . in the following ,
@xmath20 will be split into a constant part @xmath21 and a small time - dependent perturbation @xmath22 describing the interaction with a radiation field .
finally , @xmath23 contains the part of the hamiltonian which will be solved perturbatively ( such as impurity , phonon , or electron - electron scattering matrix elements ) . within the theory of nonequilibrium green functions @xcite the key quantities
are the correlation function ( or lesser green function ) @xmath24 and the retarded / advanced green functions @xmath25\\ & \times \big\langle \hat{a}_{\alpha_1}(t_1)\hat{a}^{\dag}_{\alpha_2}(t_2 ) + \hat{a}^{\dag}_{\alpha_2}(t_2)\hat{a}_{\alpha_1}(t_1 ) \big\rangle\ , , \end{split}\ ] ] respectively , where the heisenberg picture is used .
first consider a stationary state with a time independent matrix @xmath26 , neglecting the radiation field . in this case
all functions depend only on the time difference @xmath27 [ these stationary state functions are labeled by a tilde , e.g. , @xmath28 , in the following ] , and it is convenient to work in fourier space defined by @xmath29 then eqs .
( [ eqgret - time - tilde],[eqkeldysh - time - tilde ] ) of the appendix become the matrix equations @xmath30 these have to be solved self - consistently together with the equations for the self - energies which are functionals of the green functions @xmath31 this standard approach allows for a self - consistent evaluation of the green functions @xmath32 in a nonequilibrium situation caused by an applied bias ( contained in @xmath21 ) .
details , such as the specific form of the functionals in the self - consistent born approximation , can be found in , e.g. , @xcite . the current density ( in the @xmath33-direction )
can be evaluated from the expectation value of the momentum operator divided by the mass : @xmath34\rangle\\ = & \frac{e}{\hbar v}\sum_{\alpha\beta}w_{\alpha\beta } g^<_{\beta\alpha}(t , t)+\mbox{scattering currents } \label{eqcurrent } \end{split}\ ] ] where @xmath35 .
the scattering currents result from the scattering part of the hamiltonian and can be expressed in terms of self - energies .
details will be given elsewhere .
now we consider the influence of an additional time - dependent potential @xmath36 in the hamiltonian .
( bold capital symbols denote matrices in the state indices @xmath1 ) .
we treat the change of the system in linear response , and set @xmath37 the chance in the self - energies is described within the linearization of the functional ( here in the time domain ) @xmath38 such a decomposition has been used in ref . for hartree - fock self - energies , e.g .. in this case , the time - dependence of the self - energies allows for a reduction to density matrix equations by setting @xmath39 .
in contrast , in our case , where scattering effects are considered , @xmath40 exhibits an explicit time dependence in both arguments .
this two time structure is fully taken into account here .
therefore we apply the fourier decomposition in both times via @xmath41 the same decomposition is used for @xmath42 .
it is shown in appendix [ appder ] that @xmath40 is determined within linear response by the following equations : @xmath43 \tilde{\bf g}^{\text{ret / adv}}(e ) \label{eqgretlinresp}\\ \delta { \bf g}^{<}(\omega , e)&= & \tilde{\bf g}^{\text{ret}}(e+\hbar \omega)\delta { \bf u}(\omega ) \tilde{\bf g}^{<}(e ) + \tilde{\bf g}^{<}(e+\hbar \omega)\delta{\bf u}(\omega ) \tilde{\bf g}^{\text{adv}}(e)\nonumber \\ & + & \tilde{\bf g}^{\text{ret}}(e+\hbar\omega ) \delta\bm{\sigma}^{\text{ret}}(\omega , e)\tilde{\bf g}^{<}(e ) + \tilde{\bf g}^{\text{ret}}(e+\hbar\omega ) \delta\bm{\sigma}^{<}(\omega , e)\tilde{\bf g}^{\text{adv}}(e)\nonumber \\
& + & \tilde{\bf g}^{<}(e+\hbar \omega ) \delta\bm{\sigma}^{\text{adv}}(\omega , e)\tilde{\bf g}^{\text{adv}}(e ) \label{eqglesslinresp}\end{aligned}\ ] ] the changes in self - energy @xmath42 are functionals of @xmath40 , which have to be evaluated self - consistently with @xmath40 .
the derivation of the functionals from the linearization of eq .
( [ eqfuncsigmatime ] ) is straightforward and can be performed for arbitrary self - energies .
the situation is particularly simple for the self - consistent born approximation , where the functional @xmath44 is linear in @xmath45 , and one finds : @xmath46 with the same functional as eq .
( [ eqfuncsigmae ] ) .
( note , that @xmath47^{\dag}$ ] does _ not hold _ for the time - dependent quantities .
thus , the advanced quantities should be calculated explicitely ) the terms related to @xmath48 correspond to the ladder corrections in the evaluation of diagrams for the kubo formula in equilibrium .
it is a general advantage of linear response within nonequilibrium green functions that these terms are obtained directly , see also @xcite .
now we consider the linear response to an external radiation field where the electric field points in the @xmath33 direction .
this gives additional terms in the hamiltonian @xmath49 , where @xmath50 are the electromagnetic potentials .
neglecting quadratic terms in the radiation field and terms containing @xmath51 ( i.e. , assuming that the wavelength is large compared to the size of the active region ) , the following perturbation potentials are obtained ( see appendix [ appgauge ] ) : in the _ coulomb gauge _ @xmath52 neglecting the scattering part of @xmath19 ; in the _ lorentz gauge _ @xmath53 finally , the change in current density is given by the change in eq .
( [ eqcurrent ] ) @xmath54\\ & \cdot\tilde{\bf g}^<(e)+{\bf w}\cdot\delta { \bf g}^<(\omega , e ) \big\}\ , .
\end{split}\ ] ] the scattering current in eq .
( [ eqcurrent ] ) has not been included here . with these ingredients ,
one obtains the complex conductivity @xmath55 , as well as the gain coefficient via eq .
( [ eqgainsigma ] ) .
in the following we consider two different structures : ( i ) the quantum - cascade - laser structure of ref . .
( ii ) the superlattice used in ref . with 2
nm al@xmath56ga@xmath57as barriers and 8 nm gaas wells and a doping of @xmath58 . in both cases
, we use a basis set consisting of products of wannier functions ( in the growth direction @xmath33 ) and plane waves with wave vector @xmath59 ( in the @xmath60-plane ) , and obtain @xmath61 from nominal sample parameters @xcite . the self - consistent solution for the electrostatic potential
is included as well .
for the scattering hamiltonian we include interface - roughness scattering and phonon scattering ( including both optical phonons and a second phonon with low energy to mimic acoustic phonons ) .
the self - energies are evaluated in the self - consistent born approximation .
furthermore , we use momentum - independent scattering matrix elements ( evaluated for a typical momentum transfer ) and diagonal self - energies . in this approximation
the self - energies are not @xmath59-dependent , which significantly reduces the numerical effort .
results and further details for superlattices @xcite and quantum cascade lasers @xcite have been given previously .
the current - field characteristics are displayed in fig .
[ figkenn](a , b ) for both structures , which agree reasonably well with the respective experimental data .
note that domain formation in the region of negative differential conductivity modifies the behavior of the experimental curve for the superlattice .
such effects are beyond the scope of the general approach discussed here , but can be easily studied using simpler transport models @xcite .
similar results for the current - field relation of this quantum cascade laser have been reported in ref . .
for the superlattice structure , the simple 2-time miniband transport model @xcite gives @xmath62 for a non - degenerate electron gas .
here @xmath63 is the electrical field , @xmath64 the miniband width and @xmath65 the period of the superlattice .
@xmath66 ps and @xmath67 ps represent fitted , phenomenological energy and momentum scattering times .
@xmath68 are the modified bessel functions . in fig .
[ figkenn](b ) this model shows reasonable agreement with the full quantum transport calculation for @xmath69 mev , the field range where miniband transport holds true @xcite .
( for lower temperatures the shape of the current - field relation becomes more complicated both in the semiclassical boltzmann and the quantum transport model @xcite so that the simple 2-time model can not fit the data . ) ) . ] in the following the response to an external radiation field is studied : first let us use the coulomb gauge and neglect the terms with @xmath42 in eq .
( [ eqglesslinresp ] ) . the gain spectrum for the quantum cascade laser of ref .
is displayed in fig .
[ figgainqcl ] .
gain sets in for current densities above 0.8 ka/@xmath70 and increases with current .
the peak gain coefficient is 57 @xmath71 at 6.5 ka/@xmath70 and a photon energy of 130 mev , in excellent agreement with the findings of ref .
( estimated losses divided by confinement factor @xmath72 @xmath71 and threshold current @xmath73=7.2 ka/@xmath70 , lasing at @xmath74 mev ) .
the width of the gain spectrum agrees well with the findings of ref . .
results for the superlattice are given in fig .
[ figsigmasl ] .
the real part of the conductivity essentially follows the result from the simple 2-time miniband model @xcite @xmath75 with @xmath76 .
thus , this simple model gives , at least for wide minibands and high temperatures , reliable results .
mv and 20 mv , respectively .
dotted and dash - dotted line : corresponding results from the 2-time miniband model @xcite .
] now we want to study the relevance of the terms @xmath42 in eq .
( [ eqglesslinresp ] ) , which have been neglected so far .
[ figgaincomp ] shows that their inclusion barely changes the result in the coulomb gauge ( the full and dashed lines are almost indistinguishable ) .
therefore these self - energy corrections may be neglected ( at least for diagonal self - energies assumed here ) allowing for a significant reduction of the numerical effort .
in contrast , in the case of the lorentz gauge , the terms evaluated with self - energy corrections ( dash - dotted line ) differ substantially from the simpler version without these terms ( dotted line ) .
the difference is particularly strong in the case of the superlattice , where the diagonal terms in the perturbation potential ( [ eqdeltau - lorentz ] ) are essential for the transition .
furthermore , the results of coulomb gauge and lorentz gauge are almost identical if the self - energy corrections are included .
this demonstrates that the inclusion of these terms is essential to obtain a consistent theory , which must be independent of the choice of the gauge . for ( a ) the superlattice at @xmath77 and ( b ) the quantum cascade laser at @xmath78 mv .
full line : coulomb gauge neglecting @xmath79 ; dotted line : lorentz gauge neglecting @xmath79 ; dashed line : coulomb gauge with @xmath79 ( falls together with the full line ) ; dash - dotted line : lorentz gauge with @xmath79 . ]
a formulation of the theory of nonequilibrium green functions has been given which allows for a quantitative description of current - induced gain in semiconductor heterostructures .
quantitative agreement with the experiment for the gain spectrum of a quantum cascade laser was obtained . the standard 2-time miniband model for the superlattice
could be verified for a large miniband width and under room temperature operation .
the numerical results are not sensitive to the choice of the gauge if the variations of self - energy terms are taken into account .
these terms are of particular importance for the lorentz gauge , while their contribution is small in the coulomb gauge in the calculations presented here .
helpful discussions with a. knorr , a .-
jauho , s .- c .
lee , m. pereira and e. schll as well as financial support by dfg within for394 are gratefully acknowledged .
the green functions are determined by the dyson equation ( see , e.g. , chapter 5 of ref . ) @xmath80 and the keldysh relation , see also appendix [ appremark ] .
@xmath81 now we consider the stationary state , where @xmath82 .
( [ eqgret - time],[eqkeldysh - time ] ) become : @xmath83 and their fourier - transformation ( [ eqfourier - tilde ] ) provides us with eqs .
( [ eqdysone - tilde],[eqkeldyshe - tilde ] ) .
eq . ( [ eqkeldysh - time ] ) is a particular solution of the inhomogeneous differential equations ( 5.3 ) of ref . , which are given more explicitly in eqs .
( 106 - 107 ) of ref . in the notation used here .
the keldysh relation ( 5.11 ) derived in ref . contains an additional term : @xmath84\\ \times{\bf g}_0^<(t_b , t_c ) \left[\hbar\delta(t_c - t_2){\bf 1 } + { \bf u}(t_c){\bf g}^{\mathrm{adv}}(t_c , t_2 ) + \int \frac{{{\rm d}}t_d}{\hbar}\bm{\sigma}^{\mathrm{ret}}(t_c , t_d ) { \bf g}^{\mathrm{adv}}(t_d , t_2)\right]\end{gathered}\ ] ] this term constitutes a homogeneous solution of the inhomogeneous differential equations ( for fixed @xmath85 and @xmath86 ) which guarantees that @xmath87 for @xmath88 .
this satisfies a general assumption underlying the perturbation expansion of the @xmath89-matrix . in the following
we show that @xmath90 vanishes for finite times @xmath91 . using eqs .
( 108,109 ) of ref .
one obtains @xmath92_{t_a=-\infty}^{t_a=\infty}\ , , \end{split}\ ] ] where @xmath93 means that the derivative operates to the left . in the second line ,
partial integration has been used . as @xmath94 ,
the first term vanishes .
furthermore , at the upper bound @xmath95 holds for finite times @xmath96 . applying the same manipulations for @xmath97 , we find @xmath98 for a real system the retarded and advanced green functions decay for large time differences and we find @xmath99 for finite @xmath91
. therefore this additional term is spurious . a second
, more physical argument relates to the idea , that the stationary state of a real physical system should not depend on the initial conditions used in a formal theory . except for specific situations exhibiting a hysteresis
, scattering processes will install a stationary state which is independent of initial conditions .
therefore all additional homogeneous solutions should reflect transient effects , which vanish if the initial state is chosen at @xmath100 and @xmath101 .
setting @xmath102 , @xmath103 and @xmath104 in eqs .
( [ eqgret - time],[eqkeldysh - time ] ) gives in linear order @xmath105 : @xmath106 this constitutes an inhomogeneous differential equation , where the homogeneous part ( left - hand side ) is identical with the defining equation ( [ eqgret - time - tilde ] ) for @xmath107 .
thus @xmath107 is the corresponding green function and the solution is given by : @xmath108 \tilde{\bf g}^{\text{ret}}(t'-t_2)\ ] ] and fourier transformation using eq .
( [ eqfourier ] ) gives eq .
( [ eqgretlinresp ] ) after some algebra .
@xmath109 is evaluated in the same way with ret replaced by adv .
the linearization of the keldysh relation ( [ eqkeldysh - time ] ) provides us with @xmath110\end{gathered}\ ] ] fourier transformation gives eq .
( [ eqglesslinresp ] ) after some algebra .
we study the linear response to an external radiation field propagating in @xmath111-direction @xmath112 thus we have two reasonable choices of the electromagnetic potentials @xcite : @xmath113 in the spirit of a dipole approximation ( i.e. , assuming that the wavelength is large compared to the size of the active region ) the terms containing @xmath51 are neglected , providing the expressions ( [ eqdeltau - coulomb],[eqdeltau - lorentz ] ) in the perturbation hamiltonian , where quadratic terms @xmath114 are neglected . | gain in current - driven semiconductor heterostructure devices is calculated within the theory of nonequilibrium green functions . in order to treat the nonequilibrium distribution self - consistently the full two - time structure of the theory
is employed without relying on any sort of kadanoff - baym ansatz .
the results are independent of the choice of the electromagnetic field if the variation of the self - energy is taken into account .
excellent quantitative agreement is obtained with the experimental gain spectrum of a quantum cascade laser .
calculations for semiconductor superlattices show that the simple 2-time miniband transport model gives reliable results for large miniband widths at room temperature . |
heavy - ion collisions at the relativistc heavy ion collider ( rhic ) indicate the production of an opaque ( i.e. strongly jet - suppressing ) @xcite , strongly - coupled , fast - thermalizing medium that possibly needs to be described using methods derived from ads / cft string theory @xcite .
however , so far neither the initial conditions of the collisions nor the microscopic dynamics of the jet - energy loss are conclusively understood .
two models are commonly used to characterize the initial conditions . the glauber model @xcite , describing incoherent superpositions of proton - proton collisions , and the `` color glass condensate '' ( cgc ) @xcite , given e.g. by the kln model @xcite , where saturation effects are taken into account .
they differ by their initial temperature gradients , their initial high-@xmath0 parton distribution , and the distance travelled by each parton , leading to a different opacity estimate .
in addition , both models exhibit large event - by - event fluctuations @xcite .
the jet - energy loss can either be described as multiple scatterings of the parton @xcite , specific for a weakly - coupled pqcd medium , or using the ads / cft correspondence where the problem of a parton stopped in a thermal medium is related to the problem of a string falling into a @xmath3-dimensional black hole @xcite .
experimentally , jet - energy loss is parametrized by the suppression factor @xmath4 , defined as the ratio of jets produced in @xmath5 collisions to the expectation for jets being produced in @xmath6 collisions @xmath7 where @xmath8 is the number of collisions , a theoretical parameter ( calculated within the glauber model @xcite ) depending on centrality ( i.e. , on the number of participants @xmath9 ) . the first attempt to disentagle the initial conditions ( glauber vs. cgc ) and the energy - loss mechanism ( pqcd vs. ads / cft ) while simultaneously describing the nuclear modification factor @xmath10 and the elliptic flow @xmath11 for high-@xmath0 particles was given in refs .
@xcite , favoring cgc initial conditions and a strongly - coupled energy loss at rhic .
a similar ansatz was used in ref .
@xcite , although the fourier components were not shown explicitly .
to further study the differences of a pqcd and an ads / cft - like energy loss , we investigate the role of the energy dependence in the energy - loss prescription , examine the power of the path - length dependence , and calculate higher - jet harmonics .
we want to examine if a generic energy - loss ansatz that includes both a path - length and an energy dependence confirms the above conclusion that only cgc initial conditions and an ads / cft energy - loss can describe both the @xmath12 and the @xmath13 appropriately . in the high - temperature limit , all dependences on the intrinsic scales of the system ( @xmath14 , etc . )
because of this , a generic energy - loss rate @xmath15 is given by an arbitrary combination of dimensionful parameters constrained by the total dimension of the observable and the requirement that faster particles and hotter media result in a bigger suppression .
we choose @xmath16 , \label{genericeloss}\end{aligned}\ ] ] for an energy loss as a function of time @xmath17 considering a jet starting at @xmath18 and propagating in direction @xmath19 with the coupling @xmath20 .
@xmath21 is the momentum of the jet(s ) considered , @xmath22 is the temperature , and @xmath23 are parameters controlling the jet energy ( momentum ) and path - length dependence , respectively . in the bethe - heitler limit @xmath24 and @xmath25 , while in the deep lpm pqcd limit @xmath26 and @xmath27 . if @xmath28 and @xmath29 , our model coincides with the model referred to as `` ads / cft '' in refs .
however , on - shell ads / cft calculations @xcite show that @xmath30 and @xmath29 , thus we are going to consider @xmath30 throughout the whole paper .
however , one should keep in mind that @xmath30 is a weak lower bound for falling strings used for illustration .
we note that , as it is clearly shown in refs.@xcite , @xmath29 is only a _
lower _ limit corresponding to an `` on - shell '' quark whose stopping distance is @xmath31 and whose initial energy is @xmath32 . in a realistic medium ,
the first assumption is likely to be violated , resulting in values possibly of @xmath33 .
please note that in contrast to refs .
@xcite , @xmath20 is a dimensionless parameter . in a particular case of radiative - dominated scattering in the lpm regime , @xmath34 @xcite . in general
, @xmath20 measures the interaction cross - section , while @xmath35 is related to the density of scattering centers .
this ( like soft observables in general ) can be used as a constraint for @xmath20 , but additionally one has to assume nearly complete thermalization on the timescale of jet propagation throughout the system as well as a straight - forward relation between entropy density and multiplicity . while these assumptions are reasonable , they are not easily falsifiable , and , in particular , in the @xmath36-limit , one finds that @xmath20 can _ not _ describe jet quenching and a realistic gluon density at the same time @xcite .
therefore , we limit ourselves to fitting @xmath20 to the most central data point of @xmath4 measured at rhic @xcite , disregarding any interpretation in terms of the density of soft degrees of freedom , comparing to the rhic data on @xmath37-averaged @xmath4 versus centrality for a range of @xmath38 gev @xcite .
in a static medium , @xmath39 , while in a dynamic medium , @xmath15 will aquire additional powers of @xmath17 due to the dependence of temperature on @xmath17 , implicitly included in eq .
( [ genericeloss ] ) . here , we assume a @xmath40d bjorken expansion @xcite @xmath41 with different values of @xmath42 . surprisingly enough , we find that even the high-@xmath0 second fourier moment ( @xmath43 ) is most sensitive to the poorly known early - time evolution during the first fm / c . considering that @xmath4 is given by the ratio of jets in the qgp to jets in vacuum , one obtains the following formula for the nuclear modification factor from eq .
( [ genericeloss ] ) for a jet starting at @xmath18 and propagating in direction @xmath19 @xmath44\,,\ ] ] with @xmath45,\ ] ] where @xmath46 is the spectral index ( taken to be @xmath47 ) , @xmath48 , @xmath49 is the jet s initial momentum and the line integral is @xmath50d\tau\ , .
\label{ourlineint}\ ] ] in case of @xmath24 ( the bethe - heitler limit ) , @xmath51 diverges , but since @xmath52 , @xmath53 from the above calculated @xmath54 , the @xmath10 can be obtained by averaging over all possible @xmath18 and @xmath19 @xmath55 where the nuclear overlap function is used in case of the glauber model and @xmath56 for the cgc prescription . after having fixed @xmath20 ,
the @xmath1 can be computed via @xmath57\right\ } \,r_{aa}(n_{part},\phi ) } { \int d\phi \,r_{aa}(n_{part},\phi ) } \,.\end{aligned}\ ] ] the fourier density components @xmath58 and the reaction - plane axis @xmath59 are determined according to the initial density distribution used in ref .
@xcite @xmath60 and @xmath61 by definition , the impact parameter points into the @xmath62-direction , but event - by - event fluctuations introduce non - trivial and harmonic dependent @xmath59 s .
we checked that we reproduce the results of ref .
@xcite when our approach is simplified to the one used in this reference , where @xmath63 and the line integral is @xmath64 here , @xmath65 and @xmath66 describe the path - length ( @xmath67 ) dependence for pqcd and ads / cft - like energy - loss , repsectively
. however , one of the main differences to our approach is that we also consider the energy dependence in the jet - energy loss . in other words , @xmath28 in refs .
@xcite while we assume @xmath30 [ see eq .
( [ genericeloss ] ) ] throughout the whole study .
the above analysis can be extended to dijets , analogously to ref .
dijet suppression is parametrized by the factor @xmath68 , the ratio of the dijet suppression to the suppression of jets , defined experimentally as in @xcite and related to the parameters we defined earlier as in @xcite @xmath69 with the line integrals for the trigger ( t ) and away - side ( a ) jet .
please note that in contrast to ref .
@xcite , our coupling @xmath20 is equal for the trigger and the away - side jet . clearly , a higher @xmath68 then @xmath4 , as observed in ref .
@xcite , implies that if one jet survives , the other one has a higher probability to survive as well . in turn
, this indicates that jets emitted from less dense periphery regions have a larger impact .
thus , as remarked in ref .
@xcite , comparing jet abundance to dijet abundance can be a sensitive medium probe .
( top panels ) and @xmath4 ( bottom panels ) of high - momentum particles as a function of the number of participants , @xmath9 , for @xmath70 ( left panel ) and @xmath29 [ right panel , see eq .
( [ genericeloss ] ) for definition ] and @xmath71 fm .
the rhic data are taken from ref .
[ v2raa ] ] ( top panels ) and @xmath4 ( bottom panels ) of high - momentum particles as in fig . [ v2raa ] for @xmath72 fm . ] vs. @xmath43 ( a ) and @xmath73 vs. @xmath74 ( b ) for the glauber model ( reddish lines ) and the kln model ( bluish lines ) for @xmath75 and @xmath71 fm .
a clear shift between the glauber and the kln model as well as a saturation effect for larger @xmath76 can be seen .
[ vnvniaa ] ] we have carried out the procedure described in the previous section for a variety of impact parameters and initial conditions at rhic energies .
lhc energies have been mentioned in ref .
@xcite and will be explored in detail in ref .
@xcite . in this subsection ,
we focus on the mean correlations .
the width of these correlations will be discussed in the following subsection .
please note that this width is the actual physical geometry fluctuation . throughout the paper ,
we distinguish four different cases .
the glauber model @xcite and the cgc / kln prescription @xcite , both event - by - event and averaged over many events , i.e. in the latter case the initial conditions are smoothed over many events and the energy loss is calculated subsequently . in fig .
[ v2raa ] we choose @xmath71 fm , in line with recent hydrodynamic calculations @xcite .
the figure shows that for both glauber ( red ) and kln ( blue ) initial conditions as well as for both pqcd - like [ fig .
[ v2raa ] ( a ) ] and ads / cft - like [ fig . [ v2raa](c ) ] energy loss the @xmath10 can be described choosing an appropriate value for @xmath20 .
however , surprisingly enough , the results for @xmath13 get close to the rhic data when using kln initial conditions for both pqcd - like [ fig .
[ v2raa](b ) ] and ads / cft - like [ fig . [ v2raa](d ) ] energy loss , while glauber initial conditions underpredict the data .
moreover , the difference between pqcd - like and ads / cft - like energy loss is rather weak .
this result is a clear contradiction to the one shown in ref .
@xcite and questions the conclusion that only cgc / kln initial conditions and an ads / cft energy - loss can describe both the @xmath12 and the @xmath13 appropriately .
however , choosing a much smaller @xmath42 , as done in fig .
[ v2t0 ] , reduces the absolute value of @xmath13 for both pqcd and ads / cft - like energy loss scenarios , while it increases the difference between the pqcd and ads / cft results as seen in refs .
@xcite . here , the difference to the fit by jia et al .
@xcite ( black long dashed - dotted line ) , mainly seen for the pqcd - like energy - loss , is due to the additional energy loss dependence , parametrized by @xmath30 .
additionally , both plots show that there is a small discrepancy between the event - by - event and the averaged scenario @xcite .
we would like to mention here that in refs .
@xcite different values of @xmath42 were analyzed , nevertheless only the @xmath77 fm case was shown in ref .
@xcite , leaving out a discussion about the physical meaning of @xmath42 .
setting @xmath71 fm means to assume that there is no energy loss within the first fm .
pqcd does not give any excuse for this assumption and thus @xmath77 fm would be a natural assumption .
however , @xmath42 also describes the formation time of hydrodynamics which seems to be @xmath78 fm @xcite . on the other hand ,
setting @xmath71 fm is also equivalent to the ads / cft result that the energy loss is suppressed at early times ( due to the @xmath79 dependence ) .
thus , it is important to note that the @xmath43 of high-@xmath0 particles is sensitive to short - distance properties , suggesting that there is either weak coupling with a @xmath78 fm or strong coupling which in itself features the suppression of energy loss at early times . extending the analysis to dijets and calculating the @xmath68
, our results meet the datapoint of ref .
@xcite for both glauber and kln initial conditions when assuming that @xmath20 is the same for both parts of the dijet , in contrast to ref .
@xcite .
since the energy loss calculated by ads / cft @xcite clearly states that @xmath29 is only a lower limit , this raises the question which power in the jet path - length dependece has to be included when an ads / cft - like energy loss is considered .
a detailed discussion of this issue will follow in the next subsection ( see the fig . [ expfig ] ) . before that
, we will focus on correlations of higher fourier coefficients of dijet observables , @xmath80 vs. @xmath43 and @xmath73 vs. @xmath74 .
[ vnvniaa ] displays those correlations for the glauber ( reddish lines ) and the kln model ( bluish lines ) for @xmath75 .
a clear saturation effect occurs for larger @xmath76 as well as a shift between the glauber and the kln model .
please note that while for the @xmath43 correlation larger values of @xmath76 of the glauber model coincide with lower values of the kln model , this is no longer true for the @xmath74 correlations .
therefore , the different means of the correlations between @xmath81 vs. @xmath82 remain a most sensitive probe to differentiate between cgc / kln and glauber initial state geometries . after having focussed on the mean correlations in the last subsection ,
we now want to investigate the width of these correlations which is extremely important to conclude about the significance of the previously shown differences in the path - length dependence of the energy loss and the initial states considered .
( top panels ) and @xmath4 ( bottom panels ) of high - momentum particles as a function of the number of participants , @xmath9 , for @xmath70 ( left panel ) and @xmath29 ( right panel ) and @xmath71 fm .
the rhic data are taken from ref .
[ v2fig ] ] vs. @xmath43 ( a ) and @xmath73 vs. @xmath74 ( b ) for the glauber model ( reddish lines ) and the kln model ( bluish lines ) for @xmath83 and @xmath71 fm .
the blue area displays the region covered by the error bars of the kln scenario for @xmath29 .
the other cases show similar widths .
[ iaafig ] ] figs .
[ v2fig ] and [ iaafig ] are a repetition of figs .
[ v2raa ] and [ vnvniaa ] , including the fluctuations for the event - by - event analysis . as can be seen , those fluctuations are small for @xmath10 , but the width for the @xmath11 is rather large .
thus , in an experimental analysis it is less straighforward to distinguish between the different initial conditions than originally hoped for .
nevertheless , the cgc / kln initial conditions seem to be favorable . in fig .
[ iaafig ] it becomes obvious that the while larger fourier coefficients become smaller , their width becomes larger , making it experimentally more difficult to disentangle the different scenarios . please note that the scale of fig .
[ iaafig](b ) is larger than the one of fig .
[ vnvniaa](b ) , emphasizing the very broad event - by - event fluctuations of the higher - moment correlations .
therefore , in order to draw an experimentally testable conclusion , it is necessary to always determine the mean and the width of the correlations considered .
while the 2nd fourier harmonics of jet quenching have been thoroughly investigated @xcite , the sensitivity of higher harmonics has remained relatively unexplored . in ref .
@xcite higher harmonics were mentioned but not shown explicitly . in a first step ,
we examine the differences in the eccentricities between the glauber and the kln model , see figure [ efig ] . here
it becomes clear again that while on average differences certainly exist ( that are bigger for @xmath84 than for @xmath85 ) , the width of the distributions is again rather large .
this is true both for harmonics present on average as well as event - by - event , such as @xmath84 , and harmonics only present once fluctuations are taken into account , such as @xmath85 , suggesting that higher harmonics of jet observables are rather insensitive to initial conditions .
[ v3v4fig ] displays the mean and the width of @xmath86 and @xmath87 for glauber and kln initial conditions . while @xmath88 is zero unless event - by - event fluctuations are taken into account
, @xmath74 is not too different between average and event - by - event initial conditions .
please note that the absolute amount of our @xmath88 seems to be a little larger than reported but not shown in ref .
@xcite . in both cases , however , even a hypothetical ideal experiment capable of determining the impact parameter precisely would be unable to distinguish between glauber and cgc / kln initial conditions using only jet harmonics , since event - by - event physical fluctuations are enough to drawn out model sensitivity . while higher - order coefficients are more sensitive to local gradients , they are also more susceptible to event - by - event fluctuations in initial conditions ( hotspots , etc . ) , resulting in a larger @xmath89 event - by - event fluctuation .
is only present when fluctuations are taken into account , it has to vanish for the averaged analysis .
[ efig ] ] fig .
[ expfig ] shows the sensitivity of @xmath4 and @xmath90 to the microscopic mechanism of energy loss , in particular to the power of the path - length dependence @xmath76 .
here , we consider an impact parameter of @xmath91 fm that maximizes @xmath90 . as can be seen , once @xmath4
is fixed via the coefficient @xmath20 , a residual sensitivity remains mostly in the fourier components @xmath43 and @xmath88 .
a saturation effect seems to occur for larger values of @xmath76 .
( top panels ) and @xmath74 ( bottom panels ) of high - momentum particles as a function of the number of participants , @xmath9 , for @xmath70 ( left panel ) and @xmath29 ( right panel ) and @xmath71 fm . [ v3v4fig ] ] ( a ) and @xmath90 ( b - d ) of high - momentum particles at an impact parameter of @xmath91 fm as a function of the path - length exponent @xmath76 , defined in eq .
( [ genericeloss ] ) for @xmath71 fm .
pqcd energy loss assumes that @xmath70 @xcite , for ads / cft on - shell partons @xmath29 , and ads / cft off - shell partons are @xmath33 @xcite .
a saturation effect can be seen for all @xmath90 at large @xmath76 .
the rhic data ( black solid and dashed lines representing the datapoint at @xmath92 and its errorbar ) are taken from ref .
[ expfig ] ] comparing the values of @xmath43 obtained by the glauber and the kln model in fig .
[ expfig ] to the rhic data obtained by the phenix experiment ( black solid and dashed lines representing the datapoint at @xmath92 and its errorbar ) @xcite show that kln initial conditions seem to favor lower and glauber intial conditions higher exponents of the path - length dependence .
note that @xmath33 is allowed by ads / cft @xcite , combined with a significant parton virtuality ( which is reasonable since the typical stopping length is _ not _
@xmath31 ) .
since the virtuality can be directly measured in jet - photon collisions , the dependence of the exponent on virtuality could become a quantitative signature of ads / cft dynamics .
thus , a simultaneus measurement of @xmath43 , @xmath88 , and @xmath74 could elucidate the microscopic mechanism of jet - energy loss .
however , a question that naturally rises in this context is : why are @xmath90 s at high-@xmath0 so insensitive to initial conditions when hydrodynamics shows that the @xmath90 s of soft particles are extremely sensitive to initial conditions , leading to a systematic error of @xmath93 in the viscosity @xcite ?
in fact , the difference between these two regimes is _ not _ so surprising and can be readily understood physically . while viscous forces are driven by
_ local gradients _ in flow , jet absorption is driven by _
global differences _ in the integrated @xmath94 .
the two effects are generally not the same and can indeed be very different if the distributions ( like initial distributions of energy density in a lorentz - contracted nucleus ) are not smooth .
both glauber and kln initial conditions are tuned to reproduce the observed multiplicity distributions , and hence their @xmath95 is similar , even if the _
local gradients _ of @xmath22 might be very different .
therefore , if follows that hydrodynamics and tomography lead to very different results . in conclusion , we investigated different fourier harmonics of jet quenching at rhic energies and showed that the second fourier coefficients [ @xmath11 ] are remarkably sensitive to the initial time @xmath42 . if this @xmath71 fm , as suggested by recent hydrodynamic calculations @xcite , then the conclusion drawn in ref .
@xcite that only cgc / kln initial conditions and an ads / cft - like energy loss can simultaneously describe the @xmath10 and @xmath11 measured at rhic can no longer be sustained .
in contrast , for @xmath71 fm both pqcd - like and ads / cft - like energy loss seem to reproduce the rhic data well if cgc / kln initial conditions are taken into account .
moreover , we studied the microscopic mechanism of jet - energy loss by including an energy dependence and exploring the exponent @xmath76 of the jet path - length dependence .
we found that higher fourier harmonics of jet quenching are remarkably insensitive to the differences between the glauber and cgc / kln model initial conditions .
the different @xmath2 vs. @xmath1 correlations between the moments of monojet and dijet nuclear modifications factors remain a very sensitive probe to differentiate between glauber and cgc / kln initial conditions .
b.b . is supported by the alexander von humboldt foundation via a feodor lynen fellowship .
m.g . and b.b .
acknowledge support from doe under grant no .
de - fg02 - 93er40764 .
g.t . acknowledges the financial support received from the helmholtz international center for fair within the framework of the loewe program ( landesoffensive zur entwicklung wissenschaftlich - konomischer exzellenz ) launched by the state of hesse .
the authors thank a. dumitru for providing his kln code to simulate the cgc initial conditions .
d. kharzeev and e. levin , phys .
b * 523 * , 79 ( 2001 ) .
h. j. drescher , a. dumitru , c. gombeaud and j. y. ollitrault , phys .
c * 76 * , 024905 ( 2007 ) .
h. j. drescher , a. dumitru , a. hayashigaki and y. nara , phys .
c * 74 * , 044905 ( 2006 ) .
p. sorensen , j. phys .
g * 37 * , 094011 ( 2010 ) .
b. alver , m. baker , c. loizides and p. steinberg , arxiv:0805.4411 [ nucl - ex ] .
p. m. chesler , k. jensen and a. karch , phys .
d * 79 * , 025021 ( 2009 ) .
p. arnold and d. vaman , arxiv:1101.2689 [ hep - th ] .
a. drees , h. feng and j. jia , phys .
c * 71 * , 034909 ( 2005 ) ; j. jia and r. wei , phys . rev .
c * 82 * , 024902 ( 2010 ) .
a. adare _ et al .
_ [ phenix collaboration ] , phys .
* 105 * , 142301 ( 2010 ) .
j. jia , w. a. horowitz and j. liao , arxiv:1101.0290 [ nucl - th ] .
r. j. fries and r. rodriguez , nucl .
phys . a * 855 * , 424 ( 2011 ) . j. casalderrey - solana and c. a. salgado , acta phys . polon .
b * 38 * , 3731 ( 2007 ) .
j. d. bjorken , phys .
d * 27 * , 140 ( 1983 ) .
a. adare _ et al . _
[ phenix collaboration ] , phys .
c * 78 * , 014901 ( 2008 ) .
w. a. horowitz and m. gyulassy , arxiv:1104.4958 [ hep - ph ] . | second fourier harmonics of jet quenching have been thoroughly explored in the literature and shown to be sensitive to the underlying jet path - length dependence of energy loss and the differences between the mean eccentricity predicted by glauber and cgc / kln models of initial conditions .
we compute the jet path - length dependence of energy - loss for higher azimuthal harmonics of jet - fragments in a generalized model of energy - loss for rhic energies and find , however , that even the high-@xmath0 second moment is most sensitive to the poorly known early - time evolution during the first fm / c .
moreover , we demonstrate that higher - jet harmonics are remarkably insensitive to the initial conditions , while the different @xmath1 vs. @xmath2 correlations between the moments of monojet and dijet nuclear modifications factors remain a most sensitive probe to differentiate between glauber and cgc / kln initial state sqgp geometries . |
these conference proceedings are the condensed result of the 2nd meeting on asymmetrical pne , held at mit during august of 1999 .
being the last such meeting of this millennium , and a successor of the successful 1st asymmetrical meeting held at oranim in israel during august 1994 , and published as anns.isr.phys.soc.,11 in 1995 , it was especially good to see that the subject has made significant progress over the last lustrum .
the main impact , as you can see immediately by just a quick skim through this volume , has come from the imaging done by the hubble space telescope .
the wealth of complex detail revealed by these unprecedented images has left the theorists reeling , the observers gasping , and both groups with plenty ground based work ahead of them .
a special highlight of the meeting was the free hand out to participants of a cd with superb wfpc2 images by hajian & terzian !
note that references to papers in this volume are from here on indicated by an asterisk next to the authors name .
my overall impression from this meeting is that the observed morphologies of pne are shown to be more and more complex , mainly due to the images from hst .
the circle with a dot in the middle of 15 years ago is no more !
the impact of the increased spatial resolution data available from the space observations is extraordinary ! archival research with hst data is now becoming a important tool for pne research , especially in combination with ground based high resolution spectroscopy .
to date there are about 125 hst images of pne available .
building on the basis created by the work of josef solf in the early eighties ( e.g. solf 1983 , 1984 ) , the spatial resolution and spectroscopic separation at the sub
pixel level of small features is now coming of age , and is teaching us a lot about the formation of pne .
all this detail , however , is clearly posing a problem for the theorists : the complexity is enormous and at present theory understandably lags the observations .
+ having said this , theory has also made progress ; several models are getting better at producing the observed gross shapes of nebulae ( frank , garca
segura , livio , soker and others ( * ) ) , and attempts at explaining small scale structure ( dwarkadas ( * ) ) are being made .
the overall theoretical picture , however , is still quite elusive .
mass loading obviously plays an important role in some objects ( e.g. globules in the ring nebula ( dyson ( * ) ) , and perhaps also in ngc6369 , and ngc6751 just from their complex morphology , and the theory for this phenomenon is well developed .
symmetry has become very popular and it seems that there are hardly any asymmetrical pne left that do not have some kind of point symmetry .
i recall that at the meeting in la serena ( mass loss on the agb and beyond ) held in 1992 , the concept was hardly known .
+ the use of movies to show time dependent behaviour of data is also on the increase with several presentations both of observational and theoretical material at the meeting , giving some interesting new ways of visualising data .
the movie of m29 s inner nebula ( balick ( * ) ) especially struck me , since it showed the nebula s rotation , which would have got lost had the data been presented any other way .
to `` see ''
the ism penetrating a pn ( villaver ( * ) ) and thus creating an asymmetry was also spectacular .
internet access to this kind of movie is also an excellent way of showing the community what is happening .
as a result of the increasing complexity of the observed morphologies , the terminology has recently also proliferated .
table 1 lists the terms i heard at this meeting to describe features in pne .
they range from the practical to the sublime , and even beyond ... + c|c|c|c|c|c|c|c jets & fliers & rings & cocoons & toroidal cocoons + disks & ansae & shells & trails & parallelograms + haloes & real haloes & blobs & lobes & bullets + tori & structures & knots & filaments & bubbles + forks & waists & hoops & loops & double bubbles + tips & cylinders & envelopes & funnels & tuning forks + stripes & dark lanes & arcs & caps & capped bicones + tails & remnants & globules & cometary ... & nipples + rays & corkscrews & rolls & ellipsoids & shadow columns + lis & lattices & petals & slobs & smoking guns + `` paradigm '' is definitely in , the previously much over - used `` scenario '' is out ; `` weather '' ( which you ignore ... ) is in ; `` physics '' ( ... which you do nt ) is out .
other much used key words were : precession , point
symmetry , brets , episodic or periodic outflows , irregular , mass loading .
finally , a nice piece of nomenclature was reportedly found in the address of a famous observatory , the `` navel absorbatory '' .
+ it is clear that astronomers are also influenced by `` fashions '' or `` trends '' .
please start a trend and use the word `` polarigenic '' ( coined by yours truly about 15 yrs ago ) , as in `` polarigenic mechanism '' for something that produces polarisation .
figure 1 shows the possible appearances of bipolar and point
symmetric nebulae as projected onto the plane of the sky . under certain circumstances
blue and red shifted components can be located on the same side of the object , even though this is , at first sight , counter intuitive . for objects with a precession cone cut by the line of sight or plane of the sky this is observed ( e.g. ic4634 ) and such objects are point symmetric .
m29 is a bipolar that shows plane symmetry in it s inner nebula , and point symmetry in it s outer lobes which are also both red shifted since they are reflecting light from the central object .
so far this is unique among pne .
nebulae are put into the following classes : bipolars : no ( near , or not too far from plane of sky ) , eo ( nearly end on ) , rr ( both lobes red shifted ) point
symmetricals : p1 ( 0 @xmath0 i @xmath1 0.5@xmath2 ) , p2 ( 0.5@xmath2 @xmath0 i @xmath1 90 - 0.5@xmath2 ) , p3 ( 90 - 0.5@xmath2 @xmath0 i @xmath1 90 ) , where @xmath2 the precession cone opening angle , and i is the inclination angle to the plane of the sky of the nebula .
some lines of research are promising and should be continued in the near future . here
i mention a few , based on what i heard and saw at the meeting , that i find of particular potential interest .
the whole issue of what kind of mechanism drives the production of the most asymmetrical and physically different class of pne , the bipolar nebulae ( corradi & schwarz , 1995 ) , depends critically on finding the binary central stars .
there is strong circumstantial evidence for the presence of binaries in several bipolars , but it has so far been very difficult to obtain direct determinations of the numbers of bipolars and other pne that contain binary central stars .
a promising method is that of radial velocity determinations , especially in the infra red , where obscuration is less of a problem .
bond ( * ) has done pioneering imaging work with hst , which is rightly being continued . + from the theorist s point of view the single versus binary star model for bipolar nebulae is split into two parts : the making of a bipolar nebula starting with an asymmetrical density distribution , and the mechanism to produce that initial density distribution in the first place .
to make a bipolar nebula with some outflow in an asymmetrical density distribution , is relatively easy and many models can do this .
particularly successful are the models of garca segura ( * ) , a two wind interaction combined with magnetic field and near break up rotation , and the work by frank et al .
( * ) based originally on work by icke ( 1988 ) using the kompaneets theory .
the nebulae with the highest aspect ratios , such as m29 , pose more of a problem as the required density gradient and other parameters become un physically high . here
the binaries form a more natural system to generate a strong pole to equator density gradient , and their accretion disks can blow disk winds that are highly collimated and fast , explaining the high observed velocities , the strong collimation , and the large wings of the observed hydrogen line profiles .
excretion disks , as proposed by morris ( 1987 ) , can shape the slow wind from the primary star into two lobes , through which the fast wind then blows out hh like objects as in he2 - 104 . in single stars
there is no identified mechanism to produce the necessary high underlying density gradient that shapes the wind , and a particular combination of strong magnetic fields , rotation , and equatorial mass loss has to be invoked to give asymmetrical nebulae at all .
my overall impression is that binaries are needed for the most extreme bipolars and perhaps for all bipolars .
symmetry is most easily explained by precession ; again something that occurs naturally in binaries .
also the combination in m29 of both point and plane symmetry is difficult to explain without a binary central object , as is the presence of [ oiii ] lines and a low luminosity central object .
i think that an object like he31475 ( see the hst image on the cd handed out at the meeting ) can not be explained by any single star model ! we must go and find those binaries !
there were several papers on the interactions of pne with the ism ( knill degani ; theory , kerber ; observations , villaver ; modeling , with movie ( * ) ) .
this is clearly a very interesting asymmetry producing mechanism , especially affecting the outer parts of the pne , and in some cases even resulting in stars being outside their own nebula .
the ratio of density between a typical pn and the ism is about 50 in the galactic plane and 300 in the halo . under certain circumstances
the ism can penetrate the pn , giving spectacular interactions , as shown in the presented movie .
the morphologies derived from hst images of a sample of pne taken and to be taken in the magellanic clouds , can yield important relationships between morphology and the central star evolutionary stage .
this work was started by using galactic pne , extracting statistical information about a sample of which narrow band images had been taken , and making a morphological classification of the nebulae .
the result was a plot of their central stars in the hr diagram as a function of morphological class , yielding some intriguing correlations , especially for the bipolars ( stanghellini et al.1993 ) .
the large uncertainty in the individual distances to the pne in this sample make this result only statistically interesting . by going to the clouds ,
the distance uncertainty is gone and much better results can be obtained , again using hst .
stanghellini is undertaking this important work .
polarimetry was meagrely represented at this conference , as at most meetings that do not specifically deal with the subject .
only one out of 55 talks and 2 out of 38 posters contained some item about polarimetry .
since polarimetry often can provide a missing and vital piece of information , this is a plea for all observers to use and be more aware of polarimetry .
the fact that the title of this meeting contains the word `` asymmetrical '' should already make people think about polarimetry , as it is uniquely suitable to detect asymmetries . + two examples of the importance of polarimetry are : trammell et al .
( 1994 ) found that 77% of 31 agb objects are intrinsically polarised , while johnson and jones ( 1991 ) found 74% of their sample to be polarised .
there clearly is a high fraction of agb stars with asymmetries .
if there are no features in the polarisation spectrum , there has to be a global asymmetry in the object , and probably dust scattering ; if there are spectral features , local asymmetries such as convection cells or other atmospheric features must be present .
the whole critical question of the onset of asymmetry during late stellar evolution , whereby essentially spherical stars produce pne of which many are asymmetrical , can be uniquely studied with polarimetry .
the mechanism of dust scattering in the faint loops in m2 - 9 ( schwarz et al . 1997 ) was a supposition until the loops were found to be 60% polarised at right angles to the plane formed by the source , scatterer and observer
. such high degree of polarisation is not only easy to observe it also firms up the proposed dust scattering mechanism to a certainty .
this then allowed the distance to m29 to be determined accurately from it s expansion parallax .
polarimetry provided the key piece of information that allowed the physical parameters of this object to be found .
+ continued and increased work using polarisation measurements on pne and agb stars such as that in the press ( apj ) by weintraub , kastner ( who showed the data at this meeting ) , sahai , & hines , who used nicmos to image afgl2688will surely yield important results ! + in summary , this was a meeting clearly dominated by the observers using hst imagery making theorist s lives difficult .
the old chinese curse `` may you live in interesting times ! ''
is applicable here , and times will become even more interesting in the near future .
thanks for inviting me , joel and noam , and see you at the next asymmetrical meeting !
corradi , r.l.m . &
schwarz , h.e .
1995 , aa , 293 , 871 icke , v. 1988 , aa , 202 , 1771 johnson , j.j . &
jones , t.j .
1991 , aj , 101 , 1735 morris , m. 1987 , pasp , 95 , 1115 schwarz , h.e .
1997 , aa , 319 , 267 solf , j. 1983 , apjl , 266 , l113 solf , j. 1984 , aa , 139 , 296 stanghellini , l. , corradi , r.l.m . ,
schwarz , h.e .
1993 , aa , 279 , 521 trammell , s.r . ,
dinerstein , h.l . ,
goodrich , r.w .
1994 aj , 108 , 984 | this paper gives a personal impression of the conference `` asymmetrical pne ii : from origins to microstructures '' , flags some of the highlights , gathers together some facts and terminology , and indicates some promising future lines of work in this field .
# 1_#1 _ # 1_#1 _ = # 1 1.25 in .125 in .25 in |
the purpose of this paper is to introduce and analyze two exactly solvable discrete time versions of @xmath0-tasep . before introducing them we recall the continuous time poisson @xmath0-tasep which was previously studied in @xcite .
let us fix some notation used throughout . for @xmath2
we denote the state of @xmath0-tasep with @xmath3 particles as @xmath4 , where we have fixed a virtual particle at infinity by setting @xmath5 .
the gap between particle @xmath6 and @xmath7 is denoted @xmath8 .
we define and provide some background for the continuous time poisson @xmath0-tasep .
all of the results of this paper will be stated simultaneously for this process as well as its two discrete time variants .
the @xmath3-particle continuous time poisson @xmath0-tasep with particle rate parameters @xmath9 is an interacting particle system @xmath10 in which for each @xmath11 , @xmath12 moves to @xmath13 at exponential rate @xmath14 ( which vanishes when @xmath15 ) , where @xmath0 is a parameter in @xmath16 .
here we assume that these exponentially distributed jumping events are all independent , and we note that since they occur in continuous time , no two jumps occur simultaneously , almost surely . also note that the evolution of @xmath17 only depends on those particles with lower indices than @xmath6 .
step initial condition is defined as setting @xmath18 for @xmath11 .
[ zerorem ] the continuous time poisson @xmath0-tasep and the recognition of its exact solvability comes from the work of @xcite on macdonald processes .
it was soon after studied in its own right in @xcite via the type of many body system approach methods which we develop herein for the two discrete time variants of the process .
recently @xcite showed how @xmath0-tasep arises from considerations involving a @xmath0-deformed generalization of the dual rsk correspondence and @xcite showed how it arises from a certain class of nearest - neighbor stochastic dynamics on gelfand - tsetlin patterns .
it would be very interesting to extend those works or the method of @xcite to include the ( more general ) discrete time @xmath0-taseps of the present paper as well .
the evolution of @xmath19 for @xmath11 ( for the continuous time poisson @xmath0-tasep ) is given by a totally asymmetric zero range process with site dependent jump rate @xmath20 and infinite sink and source at the boundary .
a variant of this zero range process can be seen as corresponding to a particular representation of the @xmath0-boson hamiltonian considered in @xcite . in @xcite , integrability ( in the form of @xmath21 and @xmath22 matrices satisfying yang - baxter relations ) for this @xmath0-boson hamiltonian was shown , though the connection of that to the exact solvability of @xmath0-tasep discussed below remains to be understood . a stationary ( infinite lattice )
variant of the above gap zero range process was also discussed immediately after theorem 2.9 of @xcite and an @xmath23 upper and lower bound on the variance of the stationary current is established therein .
poisson @xmath0-tasep has an interpretation as a model for traffic on a one - lane ( discrete ) road @xmath24 in which the rate at which cars jump forward is modulated by the distance to the next car ( as well as a car dependent rate parameter @xmath25 ) . as
the distance goes to zero , the jump rate goes to zero , and as the distance grows , the jump rate approaches @xmath25 .
step initial condition corresponds to an initially jammed configuration . when @xmath26 , continuous time poisson @xmath0-tasep becomes the well - studied model of continuous time tasep ( and likewise the discrete time versions of @xmath0-tasep we introduce become known discrete time version of tasep ) .
the configuration of particles in tasep for a fixed time @xmath27 can be described as a determinantal point process in which all correlation functions are given by minors of a single correlation kernel ( cf . @xcite and references therein ) . in other terminology , tasep is free - fermion , or related to schur processes , or non - interacting line ensembles .
given that structure there exists a clean path to fredholm determinant formulas for marginal distributions , which in turn allow readily for asymptotic analysis ( cf .
@xcite and references therein ) . the present work , where @xmath28 , does not appear to be determinantal , hence new ideas are necessary to study and extract asymptotic distributional information about the processes considered . in recent years there has been a flurry of activity surrounding the analysis of non - determinantal yet still exactly solvable stochastic interacting particle systems @xcite . in @xcite
it was explained how , instead of the determinantal structure of correlation functions , these non - determinantal systems are exactly solvable due to the existence of concise and exact formulas for expectations of large classes of observables of the particle systems ( in fact , in many cases large enough classes to uniquely characterize the fixed time distribution of the particle systems ) .
a natural question about @xmath0-tasep is to compute the distribution of @xmath29 ( which is the location of particle @xmath30 at time @xmath27 ) .
there are ( presently ) two approaches to compute this distribution .
the first is through the theory of macdonald processes @xcite , and the second is through the many body system ( or duality ) approach @xcite . the first approach is entirely based on the integrable properties of macdonald polynomials ( see section [ macsec ] for a brief review on the relationship of the two approaches ) , and these properties naturally lead to the discovery of @xmath0-tasep as well as the computation of nested contour integral formulas for expectations of certain observables of the process .
in particular , for step initial condition this shows that ( see also @xcite or theorem [ momthm ] below ) for @xmath31 , @xmath32= \frac{(-1)^k q^{k(k-1)/2}}{(2\pi \iota)^k } \int \cdots \int \prod_{1\leq a < b\leq k } \frac{z_a - z_b}{z_a - qz_b } \prod_{j=1}^{k } \left(\prod_{i=1}^{n_j } \frac{a_i}{a_i - z_j}\right ) e^{(q-1)tz_j } \frac{dz_j}{z_j},\ ] ] where the contour of integration for @xmath33 contains @xmath34 as well as @xmath35 , but not zero ( see figure [ circontours ] for an example of such a contour when all @xmath36 ) . the above formula is given in theorem [ momthm ] along with parallel formulas for the discrete time versions . since the observables @xmath37 are all in @xmath16 , their moment problem is well - posed .
this means that the above formulas uniquely characterize the joint distribution of @xmath0-tasep at time @xmath27 ( provided it is started from step initial condition at time zero ) .
it remains a challenge to extract manageable ( from the perspective of large @xmath27 and @xmath30 asymptotics ) formulas for joint distributions , though it is understood how to derive a fredholm determinant formula which characterizes the one - point distribution of @xmath29 .
this is described in theorem [ distthm ] .
the second approach to solving the continuous time poisson @xmath0-tasep is based solely on the observation that the dynamics of the particle system implies that the expectations on the left - hand side of ( [ exp1 ] ) satisfy certain closed systems of coupled odes , which we call _ many body systems_. these systems have unique solutions and it is easy to check that the nested contour integral formulas on the right - hand side of ( [ exp1 ] ) satisfy these systems ( and hence are equal to the expectations on the left - hand side of ( [ exp1 ] ) ) .
this second approach is more direct ( it avoids explicit mention of macdonald polynomials ) and appears to be more general ( it applies to asep @xcite , which is not known to fit into the theory of macdonald processes ) . on the other hand ,
it requires non - trivial input of a suitable interacting particle system , the correct observables to study and some inspiration in guessing a closed form solution to the associated many body system . for all of the versions of @xmath0-tasep
the fact that these expectations satisfy the many body system can be seen immediately from the macdonald process perspective as a consequence of certain commutation relations involving macdonald difference operators ( see section [ macsec ] ) .
it is this second approach which we employ in this paper to study the two exactly solvable discrete time @xmath0-taseps we now introduce .
before we proceed , let us fix some additional notation .
we use @xmath38 to denote particle rate parameters , and @xmath39 and @xmath40 to denote ( discrete ) time dependent jump parameters .
the indicator function of an event @xmath41 is written @xmath42 .
the @xmath0-pochhammer symbol is defined as @xmath43 finally , we define a variant of the weyl chamber as @xmath44 the two discrete time versions of @xmath0-tasep which we introduce and study were not fabricated out of thin air , but rather came from further investigations into dynamics related to macdonald processes which may be detailed in a future work ( cf .
@xcite for some related developments ) .
all three versions of @xmath0-tasep are illustrated in figure [ qtaseps ] .
-tasep , discrete time geometric @xmath0-tasep with parallel update , and discrete time bernoulli @xmath0-tasep with right - to - left sequential update . ]
we will make use of the following @xmath0-deformation of the truncated geometric distribution .
for @xmath45 and @xmath46 the distribution @xmath47 is defined as @xmath48 for @xmath49 we extend the definition so that @xmath50 [ sumone ] for @xmath51 and @xmath46 , @xmath52 for all @xmath53 and @xmath54 the positivity is immediate as @xmath55 .
that these @xmath47 sum to one can be seen inductively ( in @xmath56 ) by using the recursion ( see ( 10.0.3 ) in @xcite ) for the @xmath0-binomial coefficients ( which occur in the definition of @xmath47 as the fraction ) .
we may now introduce the first discrete time version of @xmath0-tasep .
we work in the greatest generality for which this is exactly solvable . for simplicity
, a reader may want to think of all parameters @xmath36 and all @xmath57 .
the @xmath3-particle discrete time geometric @xmath0-tasep with particle rate parameters @xmath38 and time dependent jump parameters @xmath58 is an interacting particle system @xmath10 in discrete time @xmath59 . at time @xmath60
the system stochastically evolves from its state at time @xmath27 according to the following _ parallel _ update rule : for each @xmath11 , @xmath61 where the update is independent for each @xmath6 and @xmath27 . since we observe the convention of @xmath62 , in the above update rule @xmath63 .
we assume throughout the paper at the @xmath25 and @xmath64 are such that @xmath65 for all @xmath6 and @xmath27 , otherwise the process is not well - defined .
these dynamics mean that at time @xmath60 each particle @xmath66 hops to a random location in the set @xmath67 with hop length distributed according to independent random variables with distribution @xmath68 .
these dynamics clearly preserve the order of particles .
when @xmath26 the jump probabilities become @xmath69 this corresponds with a geometric jump of rate @xmath70 in which all of the weight given to jumps @xmath71 are collapsed onto the weight of @xmath72 . in other words ,
in parallel each particle attempts a geometric jump to the right subject to blocking by the previous location of the next particle .
this process was studied previously in @xcite as a marginal of dynamics on gelfand - tsetlin patterns which preserve the class of schur processes .
the @xmath3-particle discrete time bernoulli @xmath0-tasep with particle rate parameters @xmath38 and time dependent jump parameters @xmath73 is an interacting particle system @xmath10 in discrete time @xmath59 . at time @xmath60
the system stochastically evolves from its state at time @xmath27 according to the following _ sequential _ update rules : first update @xmath74 according to @xmath75 \frac{a_1\beta_{t+1}}{1+a_1\beta_{t+1 } } & \textrm{for } b=1,\\[.5em ] 0 & \textrm{otherwise}.\end{cases}\ ] ] then , for @xmath76 through @xmath3 ( in that order ) if @xmath77 ( i.e. , the previous jump occurred ) then update @xmath78 \frac{a_i\beta_{t+1}}{1+a_i\beta_{t+1 } } & \textrm{for } b=1,\\[.5em ] 0 & \textrm{otherwise}.\end{cases}\ ] ] if @xmath79 ( i.e. , the previous jump did not occur ) then update @xmath80 ( 1-q^{{\ensuremath{\mathrm{gap}_{i}}}(t)})\frac{a_i\beta_{t+1}}{1+a_i\beta_{t+1 } } & \textrm{for } b=1,\\[.5em ] 0 & \textrm{otherwise}.\end{cases}\ ] ] the updates are all independently distributed .
these dynamics mean that @xmath81 jumps right by one with probability @xmath82 and otherwise stays put .
sequentially , if @xmath83 jumped , then @xmath66 jumps right by one with probability @xmath84 , otherwise if @xmath83 stayed put , then @xmath66 jumps right by one with probability @xmath85 .
these dynamics clearly preserve the order of particles .
when @xmath26 the update rule becomes the following : at time @xmath60 , particle @xmath81 jumps to the right by 1 with probability @xmath82 and stays put otherwise ; then sequentially for @xmath86 through @xmath87 , the particle @xmath66 jumps to the right by 1 with probability @xmath84 if the destination site is unoccupied , otherwise it stays put .
to be clear , if @xmath88 and @xmath89 ( neighbors ) and @xmath81 jumps right so that @xmath90 , then the rightward jump for @xmath86 is available and @xmath91 .
this sequential discrete time tasep was studied previously in @xcite as a marginal of dynamics on gelfand - tsetlin patterns which preserve the class of schur processes . as @xmath92 and @xmath93
go to zero and time is suitably rescaled , the discrete time @xmath0-taseps both converge to the continuous time poisson version .
that is to say ( taking all @xmath94 and @xmath95 ) that with @xmath96 and @xmath97 geometric @xmath0-tasep @xmath10 converges as a process to continuous time poisson @xmath0-tasep with @xmath98 representing time .
the same result holds for bernoulli @xmath0-tasep with @xmath99 and @xmath100 . as @xmath101 ,
the dynamics of the continuous time poisson @xmath0-tasep converge ( under appropriate scaling ) to the solution of the semi - discrete stochastic heat equation ( equivalently the oconnell - yor directed polymer @xcite ) see theorem 4.1.26 of @xcite or proposition 6.2 of @xcite .
this semi - discrete stochastic heat equation converges under diffusive space / time scaling and weak noise scaling to the continuum stochastic heat equation @xcite .
it is also shown in @xcite that the continuous time poisson @xmath0-tasep has a direct limit to the solution to the continuum stochastic heat equation .
presumably a similar sequence of limits for the discrete time @xmath0-tasep dynamics should exist .
it would be interesting to investigate whether there exist other degenerations of these discrete time @xmath0-taseps besides those just mentioned .
one could , for instance , keep the @xmath92 or @xmath93 parameters fixed and scaling @xmath101 .
at least for the geometric case , one expects to make some contact with the work of @xcite .
the three versions of @xmath0-tasep share the following three surprising properties , which amount to their exact solvability : 1 .
the expectations of observables @xmath102 for @xmath103 evolve according to closed systems of coupled odes ( in the continuous time case ) or coupled difference equations ( in the discrete time case ) .
we call these systems _ true evolution equations_. 2 .
the true evolution equations are almost constant coefficient or separable , except for effects which arise when some of the @xmath104 are close together . rather than trying to solve the true evolution equation directly
, one can look for solutions to the free evolution equation ( i.e. , the constant coefficient and separable system neglecting the boundary effects ) on the larger space @xmath105 which have the right initial data when restricted to @xmath106 and which satisfy certain boundary conditions when some of the @xmath104 are all equal .
the restriction of such a solution to @xmath106 will coincide with the solution to the true evolution equation . generally , for every possible combination of clusters of @xmath107 (
i.e. , strings of equal coordinates ) there will be additional boundary conditions which must be satisfied .
it turns out that for the three versions of @xmath0-tasep we consider , it suffices to consider only @xmath1 two - body boundary conditions ( which ends up being the same for all three systems ) corresponding to when @xmath108 .
that the two - body boundary conditions imply all many body boundary conditions is the hallmark of _ integrability _ in the language of ( quantum ) many body systems .
this type of reduction goes back to the 1931 work of on diagonalizing the heisenberg spin chain @xcite .
our reduction , under the scaling limit to the continuum stochastic heat equation mentioned above , cincides with that for the quantum delta bose gas ( also known as the lieb - liniger model @xcite ) .
3 . a general class of solutions to the free evolution equation exists since it is separable and constant coefficient .
it is not immediately clear how to combine solutions from this class so as to additionally satisfy the @xmath1 two - body boundary conditions and the initial data .
however , for the initial data corresponding to step initial condition , it is also possible to ( easily ) check that _ nested contour integral formulas _
( such as in ( [ exp1 ] ) ) solve the free evolution equation and satisfy the @xmath1 two - body boundary conditions .
one may speculate that these formulas should be related to the _ bethe ansatz _ for diagonalizing this system ( cf .
@xcite for such a relationship for a limiting continuum version of these many body problems ) . from this
it would be possible to produce solutions corresponding to general initial conditions .
to our knowledge the bethe ansatz has not been worked out sufficiently in this context ( see however related work of @xcite ) . in section [ results ]
we record the consequences of the above mentioned properties .
the rest of the paper is devoted to proving these results along the lines of the three steps given above . in particular , in section [ truesec ] we prove that the expectations of the above mentioned observables do satisfy explicit true evolution equations . in section [ freesec ]
we demonstrate the integrability of these systems by reducing consideration to free evolution equations with @xmath1 boundary conditions . in section [ checkingformulas ]
we check that for initial data corresponding to step initial conditions this system is indeed solvable via a nested contour integral formula . in section [ endsec ]
we also briefly remark on the relationship between the true evolution equations and markov process duality as well as certain commutation relations involving macdonald difference operators .
ab was partially supported by the nsf grant dms-1056390 .
ic was partially supported by the nsf through dms-1208998 as well as by microsoft research through the schramm memorial fellowship , and by a clay mathematics institute research fellowship .
we state in parallel our main results for all three variants of @xmath0-tasep . to facilitate this we use @xmath10 for each process and instead use different expectation symbols to denote running the different dynamics .
[ momthm ] fix @xmath28 .
consider : 1 .
continuous time @xmath0-tasep with particle rate parameters @xmath38 .
let @xmath109 represent the expectation operator for this process started from step initial condition .
define @xmath110 2 .
discrete time geometric @xmath0-tasep with particle rate parameters @xmath38 and time dependent jump parameters @xmath58 .
let @xmath111 represent the expectation operator for this process started from step initial condition .
define @xmath112 3 .
discrete time bernoulli @xmath0-tasep with particle rate parameters @xmath38 and time dependent jump parameters @xmath73 .
let @xmath113 represent the expectation operator for this process started from step initial condition .
define @xmath114 fix @xmath115 , then for all @xmath116 ( i.e. , @xmath117 ) and @xmath118 , @xmath119 = \frac{(-1)^k q^{\frac{k(k-1)}{2}}}{(2\pi \iota)^k }
\int \cdots \int \prod_{1\leq a < b\leq k } \frac{z_a - z_b}{z_a - qz_b } \prod_{j=1}^{k } \left(\prod_{i=1}^{n_j}\frac{a_i}{a_i - z_j}\right ) \frac{f^{\ell}(qz_j)}{f^{\ell}(z_j ) } \frac{dz_j}{z_j},\ ] ] where the contour of integration for @xmath33 contains @xmath34 , and @xmath35 but not 0 or any poles of @xmath120 . and @xmath36 .
] the remaining sections of this article are concerned with proving the above result ( see in particular section [ checkingformulas ] ) .
[ arborder ] consider running an arbitrary combination of the three types of @xmath0-tasep with general parameters : run continuous time poisson @xmath0-tasep for time @xmath121 and discrete time geometric @xmath0-tasep with @xmath122 and discrete time bernoulli @xmath0-tasep with @xmath123 .
the methods we develop herein imply the following results .
the terminal state of the particle system after running combinations of the three dynamics does not depend on the order in which they were run ( in fact , one can take turns between running the three different dynamics ) .
moreover , letting @xmath10 represent the state of the process after having run these three dynamics we find that for all @xmath115 and @xmath116 , @xmath124 & = & \frac{(-1)^k q^{\frac{k(k-1)}{2}}}{(2\pi \iota)^k } \int \cdots \int \prod_{1\leq a < b\leq k } \frac{z_a - z_b}{z_a - qz_b } \\ & & \times\,\prod_{j=1}^{k } \left(\prod_{i=1}^{n_j } \frac{a_i}{a_i - z_j}\right ) e^{(q-1)\gamma z_j } \prod_{m\geq 1 } ( 1-\alpha_m z_j)\frac{1+q\beta_m z_j}{1+\beta_m z_j } \frac{dz_j}{z_j},\end{aligned}\ ] ] where the contour of integration for @xmath33 contains @xmath34 , and @xmath35 but not 0 or any other poles of the integrand .
the above expression is just a product of @xmath120 over @xmath118 .
the above nested contour integral formulas arise naturally in the study of macdonald processes ( cf .
proposition 3.1.5 of @xcite and corollary 4.7 of @xcite ) and the existence of the associated integrable quantum many body systems is actually a consequence of this connection ( see section [ macsec ] for more details ) .
[ distrem ] the expectations considered above in theorem [ momthm ] uniquely characterize the fixed time joint distribution of @xmath125 , and hence also that of @xmath10 .
this is because the @xmath126 are deterministically in @xmath16 , and thus their moments uniquely identify their distributions @xcite . in principle
it should be possible to extract formulas for any fixed time expectations of @xmath10 from the explicit formulas in theorem [ momthm ]
. it would be desirable to find moment generating functions ( which characterize the joint distributions of @xmath10 ) for which there are sufficiently compact formulas to allow for asymptotics .
presently it is only known how to do this for the one - point distribution of @xmath29 . as an application of the above moment formulas it is possible to prove the following fredholm determinant formula for the @xmath0-laplace transform of @xmath37 which in turn characterize the distribution of @xmath29 via a simple inversion formula ( see remark [ qinversion ] ) . the procedure for going from the nested contour integral formulas above to the fredholm determinant formula
is developed in @xcite and the below result follows immediately from this procedure . for simplicity of contours
, we assume here that all @xmath36 ( see remark [ notasrem ] ) .
[ distthm ] fix @xmath28 and particle rate parameters @xmath36 .
consider the three versions of @xmath0-tasep corresponding with @xmath118 ( in the notation of theorem [ momthm ] ) .
then for all @xmath127 @xmath128 = \det(i + k^{\ell}_{\zeta})\ ] ] where @xmath129 is the fredholm determinant of @xmath130 for @xmath131 a small positively oriented circle containing 1 .
the operator @xmath132 is defined in terms of its integral kernel @xmath133 with @xmath134 where the function @xmath135 is defined as in theorem [ momthm ] .
the following second formula also holds : @xmath136 = \frac{\det(i + \zeta \tilde{k}^{\ell})}{(\zeta;q)_{\infty}}\ ] ] where @xmath137 is the fredholm determinant of @xmath138 times the operator @xmath139 for @xmath140 a positively oriented circle containing 0 and 1 ( and no poles of @xmath135 ) . the operator @xmath141 is defined in terms of its integral kernel @xmath142 where the function @xmath135 is defined as in theorem [ momthm ] .
we refer to the first fredholm determinant ( [ mellinbarnes ] ) as mellin - barnes type , and the second fredholm determinant ( [ cauchy ] ) as cauchy type .
these mellin - barnes type and cauchy type fredholm determinant formulas were previously proved in the form of theorems 3.2.10 and 3.2.16 ( respectively ) of @xcite .
however , those results are phrased in terms of macdonald processes ( in which @xmath143 corresponds with @xmath144 $ ] ) . the general scheme which leads to these results
is recorded in sections 3.1 and 3.2 of @xcite and shows how to go from the nested contour integral formulas for @xmath145 $ ] to the ( respective ) mellin - barnes and cauchy type fredholm determinant formulas in the theorem . in the application of propositions 3.6 and 3.10 of @xcite
it is necessary to check certain technical conditions on the function @xmath146 to make sure that the formal manipulations which lead to the fredholm determinants are , in fact , numerical identities .
these conditions are checked in the proofs of theorems 3.2.10 and 3.2.16 ( respectively ) of @xcite so we do not repeat them .
[ notasrem ] the only modification of theorem [ distthm ] for general @xmath25 is that in the first fredholm determinant , @xmath147 and consequently some care must be taken in specifying the appropriate choice of contours for the @xmath148 and @xmath149 integration ( so that the @xmath148 contour contains the @xmath150 while avoiding other poles ) .
see for example , theorem 3.2.11 of @xcite , or theorem 4.13 of @xcite .
likewise , in the second fredholm determinant formula , the term @xmath151 in defining @xmath152 is replaced by @xmath153 and the contour @xmath140 is replaced by a suitable one which contains 0 and all @xmath150 .
[ qinversion ] the transform from the probability distribution of @xmath37 to the expectation on the left - hand sides of ( [ mellinbarnes ] ) and ( [ cauchy ] ) is called a @xmath0-laplace transform . just as the usual laplace transform
can be inverted via a single contour integral in the spectral variable , so too can we recover the distribution of @xmath154 or @xmath155 from its transform . to state this inversion ,
write @xmath156.\ ] ] then it follows from proposition 3.1.1 of @xcite ( see also @xcite ) that @xmath157 where @xmath158 is a positively oriented circle which encircles @xmath159 .
there are a number of interesting limit theorem results which should be accessible via asymptotic analysis of the fredholm determinants in theorem [ distthm ] .
the mellin - barnes type fredholm determinant seems to be most easily analyzed asymptotically .
presently , the only asymptotics related to @xmath0-tasep which have been worked out correspond with first taking the limit for @xmath30 fixed and @xmath101 in which @xmath0-tasep converges @xcite to the semi - discrete stochastic heat equation @xcite . at that level
@xcite have proved gue tracy - widom limit theorems ( see also @xcite for a related discrete stochastic heat equation @xcite ) .
there remain a number of limit theorems which should be provable from the fredholm determinant formulas above , such as the gue tracy - widom limit theorems for the three variants of @xmath0-tasep .
it is possible that in proving such results slightly different choices of contours ( as in @xcite ) will be necessary .
the first surprising property of q - tasep is that the expectations of the observables @xmath102 evolve according to closed systems of coupled odes ( in the continuous time case ) or coupled difference equations ( in the discrete time case ) .
we call these systems the _ true evolution equations_. this property is not a generic property of interacting particle systems , but rather something quite special to these systems and these choices of observables . in order to state the true evolution equations it will be useful to have an alternative notation for @xmath116 wherein @xmath160 counts the number of @xmath161 ( i.e. , the size of the cluster of @xmath162 s equal to @xmath6 ) .
more precisely , define @xmath163 to each @xmath116 associate @xmath164 via @xmath165 , and to each @xmath166 associate @xmath167 as the unique @xmath107 such that @xmath168 .
thus @xmath169 lists the multiplicities of each number in @xmath170 in @xmath107 and @xmath171 associates to such multiplicities an ordered list in @xmath106 . for instance if @xmath172 then @xmath173 .
we say that @xmath107 has @xmath174 clusters if @xmath175 .
for the present example @xmath172 has three clusters . for a vector @xmath176 and a vector @xmath177 with @xmath178
define @xmath179 both of these map @xmath180 into another element of @xmath181 .
the following combinatorial coefficients are used in defining the true evolution equation . for @xmath182 , @xmath28 , @xmath183 and integers
@xmath184 define @xmath185 with the convention that @xmath186 for @xmath187 or @xmath188 .
we record the following easily checkable properties of the coefficients @xmath189 which will prove useful : @xmath190 for @xmath191 , @xmath11 and @xmath192 define the following difference operators @xmath193 [ zeroremark ] if @xmath194 for @xmath180 with @xmath195 , then it follows that @xmath196 moreover , for such an @xmath146 , we have that @xmath197 with @xmath198 .
[ truedef ] we say that @xmath199 solves the ( rate parameter @xmath34 ) continuous time poisson @xmath0-tasep / discrete time geometric @xmath0-tasep ( with time dependent jump parameters @xmath200 ) / discrete time bernoulli @xmath0-tasep ( with time dependent jump parameters @xmath201 ) _ true evolution equation _ with initial data @xmath202 if 1 . for all @xmath192 and @xmath203 , @xmath204 2 . for all @xmath192 such that @xmath195 , @xmath205 for all @xmath203 ; 3 . for all @xmath192 , @xmath206 .
[ existuniq ] the above respective true evolution equations have unique solutions .
the operators @xmath207 and @xmath208 map the space of functions @xmath209 onto itself .
in fact , these operators restrict to mapping the space of functions @xmath210 onto itself . on account of this ,
the above true evolution equations restrict to a collection of closed systems of linear odes ( for the continuous time case ) or difference equations ( for the two discrete time cases ) , indexed by @xmath211 .
it suffices , therefore , to prove existence and uniqueness when restricted to the system corresponding to each @xmath115 . for fixed @xmath211 ,
our system of linear differential or difference equations is also triangular in the following sense .
define a partial ordering on @xmath181 so that @xmath212 if for all @xmath213 , @xmath214 .
then triangularity means for the continuous time case that @xmath215 depends only upon @xmath216 for @xmath217 , and for the discrete time cases that @xmath218 depends only upon @xmath216 for @xmath217 .
this is easily seen from the definition of the true evolution equations . finally , for each @xmath115 , the associated closed , triangular system of linear odes / difference equations is also finite .
one account of this , one can apply standard methods ( such as in @xcite for the continuous time case or linear algebra for the discrete time cases ) to conclude the existence and uniqueness of solutions .
we may now state the main result of this section , which is that expectations of certain observables of the three variants of @xmath0-tasep solve the above ( respective ) true evolution equations .
[ truethm ] consider the ( rate parameter @xmath34 ) continuous time poisson @xmath0-tasep / discrete time geometric @xmath0-tasep ( with time dependent jump parameters @xmath39 ) / discrete time bernoulli @xmath0-tasep ( with time dependent jump parameters @xmath40 ) started from an arbitrary initial condition @xmath219 .
then for any @xmath115 and @xmath116 , @xmath220 = h(t;\vec{y}(\vec{n}))\ ] ] where @xmath199 solves the ( respective ) true evolution equation with initial data @xmath221.\ ] ] due to lemma [ existuniq ] it suffices to show that the left - hand side of ( [ truelhs ] ) satisfies the true evolution equation ( in each of the three cases ) .
in what follows let @xmath222 .
it is then convenient to rewrite the left - hand side of ( [ truelhs ] ) as @xmath223= { \ensuremath{\mathbb{e}}}\left[\prod_{i=0}^{n } q^{(x_i(t)+i)y_i}\right]\ ] ] with the convention that when @xmath195 , the above is zero , and when @xmath224 , the above product starts at @xmath225 . from this
one readily sees that condition ( 2 ) of the true evolution equation is satisfied
. condition ( 3 ) is immediate as well .
we prove condition ( 1 ) for each of the three cases .
in fact , the poisson case was previously proved in @xcite , though we include it here as well for completeness . * poisson case * : let @xmath21 denote the generator of continuous time poisson @xmath0-tasep
. then it follows that @xmath226 & = & l { \ensuremath{\mathbb{e}}}\left [ \prod_{i=0}^{n } q^{(x_i(t)+i)y_i}\right]= { \ensuremath{\mathbb{e}}}\left[l \prod_{i=0}^{n } q^{(x_i(t)+i)y_i}\right]\\ & = & { \ensuremath{\mathbb{e}}}\left[\sum_{j=1}^{n } a_j(1-q^{x_{j-1}(t)-x_{j}(t)-1})\big(q^{(x_j(t)+1+j)y_j}-q^{(x_j(t)+j)y_j}\big ) \prod_{\substack{i=0\\i\neq j}}^{n } q^{(x_i(t)+i)y_i}\right]\\ & = & { \ensuremath{\mathbb{e}}}\left[\sum_{j=1}^{n } a_j(1-q^{x_{j-1}(t)-x_{j}(t)-1})(q^{y_j}-1 ) \prod_{i=0}^{n } q^{(x_i(t)+i)y_i}\right]\\ & = & { \ensuremath{\mathbb{e}}}\left[\sum_{j=1}^{n } { \ensuremath{\mathcal{l}^{(a_j)}_{j } } } \prod_{i=0}^{n } q^{(x_i(t)+i)y_i}\right ] = \sum_{j=1}^{n } { \ensuremath{\mathcal{l}^{(a_j)}_{j } } } { \ensuremath{\mathbb{e}}}\left[\prod_{i=0}^{n } q^{(x_i(t)+i)y_i}\right].\end{aligned}\ ] ] this shows that the expectation in question does satisfy the true evolution equation .
* geometric case * : we show the following stronger statement .
let @xmath227 be the sigma - field generated by the random variables @xmath228 . then , as @xmath227 measurable random variables @xmath229 = { \ensuremath{\mathcal{a}^{(-a_1 \alpha_{t+1})}_{1 } } } \cdots { \ensuremath{\mathcal{a}^{(-a_n \alpha_{t+1})}_{n } } } \prod_{i=0}^{n } q^{(x_i(t)+i)y_i}.\ ] ] taking expectations of both sides ( with respect to the random variables @xmath10 ) recovers the desired result . in what follows we write @xmath92 instead of @xmath230 .
in order to show the above result , observe that it follows immediately from the geometric dynamics that @xmath231 = \prod_{i=1}^{n } \left(\sum_{j=0}^{{\ensuremath{\mathrm{gap}_{i}}}(t ) } { \ensuremath{\mathbf{p}_{{\ensuremath{\mathrm{gap}_{i}}}(t ) , a_i \alpha}}}(j ) q^{jy_i}\right ) \prod_{i=0}^{n } q^{(x_i(t)+i)y_i}.\ ] ] [ cqm ] for all @xmath232 and @xmath191 @xmath233 and likewise @xmath234 applying this lemma we find @xmath235 with the convention @xmath236 .
the second line follows from the definition of @xmath237 .
thus , we have reached a formula for the left - hand side of ( [ geolhs ] ) which , in light of equation ( [ rhsexp ] ) matches the right - hand side of ( [ geolhs ] )
. we will prove equation ( [ lemmalhs ] ) .
the @xmath49 case is in the same spirit so we leave it out .
taking @xmath56 fixed , denote the left - hand side of ( [ lemmalhs ] ) as @xmath238 .
we may apply lemma [ sumone ] with the parameter @xmath92 from its statement taken to be @xmath239 .
this yields @xmath240 from the definition of @xmath241 this equality can be rewritten as @xmath242 we may expand the product @xmath243 as @xmath244 where @xmath245 is the degree @xmath246 elementary symmetric polynomial ( see e.g. section i.2 of @xcite ) . plugging this into ( [ secondset ] ) and rearranging the summations of @xmath53 and @xmath246 yields @xmath247 rewriting in terms of the @xmath238
, we find @xmath248 note also that ( see e.g. section i.3 , exercise 1 of @xcite ) @xmath249 the relation ( [ aboverelation ] ) on the @xmath238 uniquely characterizes them , hence it suffices to check that the right - hand side of ( [ lemmalhs ] ) also satisfies the relation when substituted for the @xmath238 .
this amounts to @xmath250 in order to check that the above relation holds we can gather all coefficients associated with @xmath251 and check that for @xmath252 the coefficients combine to equal 1 , and for @xmath253 they combine to 0 . letting @xmath254 , we must therefore check that @xmath255 using the definition of the @xmath256 coefficients the above relation reduces to @xmath257 from corollary 10.2.2(c ) of @xcite we find that @xmath258 from which ( [ aboverelation2 ] ) immediately follows .
* bernoulli case * : we will show the following stronger statement .
for @xmath259 and @xmath203 , let @xmath260 the sigma - field generated by the random variables @xmath261 .
then , as @xmath227 measurable random variables @xmath262 = { \ensuremath{\mathcal{a}^{(qa_1\beta_{t+1})}_{1 } } } \cdots { \ensuremath{\mathcal{a}^{(qa_n\beta_{t+1})}_{n } } } \prod_{i=0}^{n } q^{(x_i(t+1)+i)y_i}.\ ] ] taking expectations of both sides ( with respect to the random variables @xmath10 ) recovers the desired result .
in what follows we write @xmath93 instead of @xmath263 . the following lemma will quickly yield a proof of ( [ berncond ] ) .
[ bernlem ] for @xmath264 , @xmath265\qquad\qquad\qquad\qquad&\\ = \sum_{s_i=0}^{y_i } c_{qa_i\beta } ( y_i , s_i ) q^{(x_i(t)+i)(y_i - s_i)}q^{(x_{i-1}(t)+i-1)s_i},\qquad\qquad\qquad\qquad\qquad&\end{aligned}\ ] ] and for @xmath225 @xmath266 = c_{qa_1\beta } ( y_1,0 ) q^{(x_1(t)+1)y_1}.\ ] ] let us first consider the @xmath225 case .
given the knowledge of @xmath267 , the dynamics of bernoulli @xmath0-tasep implies that with probability @xmath268 , we have that @xmath269 and with probability @xmath270 , we have that @xmath271 .
this implies that @xmath272 & = & c_{a_1\beta}(y_1,0 ) q^{(x_1(t)+1)y_1}\left(\frac{a_1\beta}{1+a_1\beta } q^{y_1 } + \frac{1}{1+a_1\beta}\right ) \\ & = & c_{qa_1\beta } ( y_1,0 ) q^{(x_1(t)+1)y_1},\end{aligned}\ ] ] as desired .
here we used the third relation of ( [ cid ] ) with @xmath273 to reach the above conclusion . now consider the case when @xmath264 . let @xmath274 represent the indicator function for the event that @xmath275 ( i.e. , particle @xmath7 jumped at time @xmath60 ) .
note that this is measurable with respect to @xmath276 .
thus , by virtue of the bernoulli dynamics , we have that for @xmath277 @xmath278\\ & = i_{i-1 } q^{(x_i(t)+i)r } q^{(x_{i-1}(t)+1+i-1)s } \left(\frac{a_i\beta}{1+a_i\beta } q^r + \frac{1}{1+a_i\beta}\right)\\ & \quad + ( 1-i_{i-1 } ) q^{(x_i(t)+i)r } q^{(x_{i-1}(t)+i-1)s } \left((1-q^{x_{i-1}(t)-x_i(t)-1})\frac{a_i\beta}{1+a_i\beta } q^r + \frac{1+a_i\beta q^{x_{i-1}(t)-x_i(t)-1}}{1+a_i\beta}\right)\\ & = q^{(x_i(t)+i)r } q^{(x_{i-1}(t)+i-1)s } \frac{1}{1+a_i\beta } \left(i_{i-1 } a_i\beta q^{r+s } + i_{i-1 } q^s + ( 1-i_{i-1 } ) q^r + ( 1-i_{i-1})\right)\\ & \quad + q^{(x_i(t)+i)(r-1 ) } q^{(x_{i-1}(t)+i-1)(s+1 ) } \frac{1}{1+a_i\beta } \left(a_i\beta(1-i_{i-1})(1-q^r)\right).\end{aligned}\ ] ] using the above , we may now compute @xmath279 \\ & = \sum_{s_i=0}^{y_i } q^{(x_i(t)+i)(y_i - s_i)}q^{(x_{i-1}(t)+i-1)s_i } \\ & \qquad\qquad\times \ , \bigg(c_{a_i\beta}(y_i , s_i)\frac{1}{1+a_i\beta}\left(i_{i-1}a_i\beta q^{y_i } + i_{i-1 } q^{s_i } + ( 1-i_{i-1})a_i\beta q^{y_i - s_i } + ( 1-i_{i-1})\right ) \\ & \qquad\qquad\qquad\qquad\qquad\qquad\qquad\qquad\qquad+ c_{a_i\beta}(y_i , s_i-1 ) \frac{a_i\beta}{1+a_i\beta}(1-i_{i-1})(1-q^{y_i - s_i+1})\bigg).\end{aligned}\ ] ] observe that we may rewrite the above factor ( in large parentheses ) as @xmath280 this last equality came from two facts .
the left - hand side has two terms .
the first is equal to @xmath281 by the third relation of ( [ cid ] ) .
the second term ( with @xmath274 ) is zero by the first relation of ( [ cid ] ) . as a consequence of the above calculation
we find that @xmath282\\ & = \sum_{s_i=0}^{y_i } c_{qa_i\beta}(y_i , s_i ) q^{(x_i(t)+i)(y_i - s_i)}q^{(x_{i-1}(t)+i-1)s_i},\end{aligned}\ ] ] as desired to complete the proof of the lemma . in order to conclude the proof of the bernoulli case we use conditional expectations to rewrite the left - hand side of ( [ berncond ] ) with @xmath224 as @xmath283 \big\vert \sigma^n_t,\sigma^{n-2}_{t+1}\big]\cdots \big\vert \sigma^{n}_t , \sigma^{1}_{t+1 } \big]\big\vert \sigma^{n}_t\big]\\ & = { \ensuremath{\mathcal{a}^{(a_1\beta)}_{1 } } } \cdots { \ensuremath{\mathcal{a}^{(a_{n-1}\beta)}_{n-1 } } } { \ensuremath{\mathbb{e}}}\big[q^{(x_1(t+1)+1)y_1 } { \ensuremath{\mathbb{e}}}\big [ q^{(x_2(t+1)+2)y_2 } \cdots { \ensuremath{\mathbb{e}}}\big [ q^{(x_{n-1}(t+1)+n-1)y_{n-1 } } \\ & \quad\times\,\sum_{s_n=0}^{y_n } c_{a_n\beta}(y_n , s_n ) { \ensuremath{\mathbb{e}}}\big [ q^{(x_n(t+1)+n)(y_n - s_n)}q^{(x_{n-1}(t+1)+n-1)s_n } \big\vert \sigma^n_t,\sigma^{n-1}_{t+1}\big]\big\vert \sigma^n_t,\sigma^{n-2}_{t+1}\big]\cdots \big\vert \sigma^{n}_t , \sigma^{1}_{t+1 } \big]\big\vert \sigma^{n}_t\big]\\ & = { \ensuremath{\mathcal{a}^{(a_1\beta)}_{1 } } } \cdots { \ensuremath{\mathcal{a}^{(a_{n-1}\beta)}_{n-1 } } } { \ensuremath{\mathbb{e}}}\big[q^{(x_1(t+1)+1)y_1 } { \ensuremath{\mathbb{e}}}\big [ q^{(x_2(t+1)+2)y_2 } \cdots { \ensuremath{\mathbb{e}}}\big [ q^{(x_{n-1}(t+1)+n-1)y_{n-1 } } \big\vert \sigma^n_t,\sigma^{n-2}_{t+1}\big]\cdots \big\vert \sigma^{n}_t , \sigma^{1}_{t+1 } \big]\big\vert \sigma^{n}_t\big ] \\ & \quad\times\,\sum_{s_n=0}^{y_n } c_{qa_n\beta } ( y_n , s_n ) q^{(x_n(t)+n)(y_n - s_n)}q^{(x_{n-1}(t)+n-1)s_n}\end{aligned}\ ] ] where the first equality was from definition and the second equality followed from applying lemma [ bernlem ] with @xmath284 .
we may continue to use lemma [ bernlem ] to reduce the expression above , ultimately leading to @xmath285 with the convention @xmath236 . on account of remark [ zeroremark ]
we may now recognize that the right - hand side above , is equal to the right - hand side of ( [ berncond ] ) , as desired to complete the proof of the bernoulli case .
it is not a priori clear how to approach the problem of solving the true evolution equations of theorem [ truethm ] in closed form . away from the boundary of @xmath106 ( i.e. , when all of the elements of @xmath107 are different , or equivalently all @xmath286 ) the true evolution equation is constant coefficient and separable .
this constant coefficient , separable equation can be extended to all of @xmath287 and is what we will call the _ free evolution equation_. the true and free evolution equations do not match near the boundary of @xmath106 when there is clustering in @xmath107 .
the following idea can be traced to bethe s 1931 solution to the heisenburg spin chain @xcite and can be thought of as a generalization of the method of images or reflection principle .
bethe s idea is to try to rewrite the true evolution equation as the restriction of the free evolution equation subject to certain boundary conditions which , if satisfied , would imply that the restriction matched the true evolution equation .
generally the free evolution equation may be defined on a larger space ( here @xmath287 ) than is physically meaningful ( here @xmath106 ) and the initial data need only be imposed on the physically relevant portion of the space . if such a solution satisfying the free evolution equation , boundary conditions and initial data exists , then its restriction to the physically relevant space will necessarily solve the true evolution equation for the right initial data .
this existence of solutions is not assured .
if the system in consideration is @xmath211-dimensional , then generically there could be boundary conditions corresponding to all possible compositions of @xmath288 .
if in fact only the @xmath1 nearest neighbor boundary conditions are necessary , then the system is called _ integrable _ in the language of quantum many body systems .
this is the case for the true evolution equations of poisson , geometric and bernoulli @xmath0-tasep . for the poisson @xmath0-tasep
this reduction was observed in proposition 2.7 of @xcite .
one should note that it is not a priori clear that the boundary conditions for the discrete time @xmath0-taseps with time varying parameters @xmath64 or @xmath289 would not depend on @xmath27 .
in fact , it turns out that this @xmath27 dependence of the true evolution equation plays no role in the form of the boundary condition .
in fact , between all three versions of @xmath0-tasep , the exact same boundary condition arise . for
a function @xmath290 define the operators @xmath291 and @xmath292 via @xmath293 for a function @xmath294 , let @xmath295_i$ ] and @xmath296_i$ ] act on the @xmath297 coordinate of @xmath146 .
the following theorem relates the true evolution equations for @xmath0-tasep to free evolution equations with @xmath1 boundary conditions .
[ freethm ] if @xmath298 solves : 1 . for all @xmath299 and @xmath203 , @xmath300_i(t;\vec{n } )
\qquad\qquad\qquad\quad\ , \textrm{(poisson)},\\ u(t+1;\vec{n } ) & = & [ { \ensuremath{\nabla_{-a_{n_1 } \alpha_{t+1 } } } } ] _ 1 \cdots [ { \ensuremath{\nabla_{-a_{n_k } \alpha_{t+1}}}}]_k u(t;\vec{n } ) \qquad\textrm{(geometric)},\\[.2em ] \,[{\ensuremath{\nabla _ { a_{n_1}\beta_{t+1 } } } } ] _ 1 \cdots [ { \ensuremath{\nabla_{a_{n_k } \beta_{t+1}}}}]_k u(t+1;\vec{n } ) & = & [ { \ensuremath{\nabla_{q a_{n_1}\beta_{t+1}}}}]_1 \cdots [ { \ensuremath{\nabla_{q a_{n_k}\beta_{t+1}}}}]_k u(t;\vec{n})\qquad\ , \textrm { ( bernoulli)};\end{aligned}\ ] ] 2 . for all @xmath301 such that for some @xmath302 , @xmath303 , @xmath304_i - q [ \nabla]_{i+1}\big ) u(t;\vec{n})=0;\ ] ] 3 . for all @xmath299 such that @xmath305 , @xmath306 for all @xmath203 ; 4
. for all @xmath116 , @xmath307
; then for all @xmath176 , @xmath308 where @xmath309 is the solution to the true evolution equation for the ( rate parameter @xmath34 ) continuous time poisson @xmath0-tasep / discrete time geometric @xmath0-tasep ( with time dependent jump parameters @xmath39 ) / discrete time bernoulli @xmath0-tasep ( with time dependent jump parameters @xmath73 ) , started from initial data @xmath202 .
it is immediate from the third and fourth hypotheses of theorem [ freethm ] that conditions ( 2 ) and ( 3 ) of definition [ truedef ] are satisfied .
it suffices to check that condition ( 1 ) of the true evolution equation is satisfied by @xmath310 .
we show this first for the poisson case ( which previously appeared in the proof of proposition 2.7 of @xcite ) .
we then deal simultaneously with the geometric and bernoulli cases by using lemma [ dblem ] below . in what follows
we assume that @xmath311 and hence @xmath312 .
* poisson case : * recall that the size of the cluster of elements of @xmath107 equal to @xmath6 is given by @xmath160 . consider the cluster of elements equal to @xmath3 : @xmath313 .
every other cluster works similarly to what we now describe and though it is possible that the cluster we study is empty , we may repeat the below calculation for any other cluster . in order to prove the true evolution equation it suffices ( by summing over all clusters ) to show that @xmath314_i u(t;\vec{n } ) = a_n ( 1-q^{y_n } ) \nabla_{y_n } u(t;\vec{n } ) = { \ensuremath{\mathcal{l}^{(a_n)}_{n}}}u(t;\vec{y}),\ ] ] where the second equality follows immediately from the definition of @xmath315 . to see the first equality we use the boundary condition for the free evolution equation which implies that for @xmath316 , @xmath296_i u(t;\vec{n } ) = q^{y_n - i}[\nabla]_{y_n } u(t;\vec{n})$ ] .
summing over @xmath6 gives the desired equality .
* geometric and bernoulli cases : * we use lemma [ dblem ] below to prove the equivalence of the free and true evolution equations for both the geometric and bernoulli cases .
however , we first provide a lemma which quickly leads to the proof of lemma [ dblem ] .
[ prevlem ] fix @xmath317 and @xmath191 .
assume that a function @xmath318 is such that if @xmath108 for any @xmath319 then @xmath320_i - q[\nabla]_{i+1})f(\vec{n } ) = 0 $ ]
. then @xmath321_1 \cdots [ { \ensuremath{\nabla_{a}}}]_y f(n,\ldots , n ) = \sum_{s=0}^{y } c_{a}(y , s ) f(\underbrace{n,\ldots , n}_{y - s } , \underbrace{n-1,\ldots , n-1}_{s}).\ ] ] we prove this by induction in @xmath322 . for @xmath323 observe that by definition of @xmath324 $ ] , @xmath325_1 f(n ) & = & ( 1+a)f(n ) - a f(n-1)\\ & = & c_a(1,0 ) f(n ) + c_a(1,1 ) f(n-1),\end{aligned}\ ] ] since @xmath326 and @xmath327 . for @xmath328
assume that we have proved the lemma for @xmath329 .
thus we have @xmath330_1 [ { \ensuremath{\nabla_{a}}}]_2 \cdots [ { \ensuremath{\nabla_{a}}}]_y f(n,\ldots , n ) & = & [ { \ensuremath{\nabla_{a}}}]_1 \sum_{s=0}^{y-1 } c_{a}(y-1,s ) f(n,\underbrace{n,\ldots , n}_{y-1-s } , \underbrace{n-1,\ldots , n-1}_{s})\\ & = & \sum_{s=0}^{y-1 } c_{a}(y-1,s ) ( 1+a ) f(n,\underbrace{n,\ldots , n}_{y-1-s } , \underbrace{n-1,\ldots , n-1}_{s})\\ \nonumber & & + \sum_{s=0}^{y-1 } c_{a}(y-1,s ) ( -a ) f(n-1,\underbrace{n,\ldots , n}_{y-1-s } , \underbrace{n-1,\ldots , n-1}_{s}).\end{aligned}\ ] ] using the relation @xmath320_i - q[\nabla]_{i+1})f(\vec{n } ) = 0 $ ] for @xmath107 such that @xmath108 , we see that @xmath331 plugging this into ( [ aboveplug ] ) we arrive at @xmath332 grouping the coefficients of @xmath333 in the above summations and using the second relationship in ( [ cid ] ) we recover ( [ desiredf ] ) and hence complete the inductive step .
[ dblem ] fix @xmath334 and @xmath38 .
assume that a function @xmath318 is such that if @xmath108 for any @xmath319 then @xmath320_i - q[\nabla]_{i+1})f(\vec{n } ) = 0 $ ] .
then for all @xmath31 , @xmath335_1 \cdots [ { \ensuremath{\nabla_{a_{n_k}}}}]_k f(n_1,\ldots , n_k ) = \sum_{s_1=0}^{y_1 } c_{a_1}(y_1,s_1 ) \cdots \sum_{s_n=0}^{y_n } c_{a_n}(y_n , s_n ) f(\vec{n}(\vec{y}^{\,\vec{s}})).\ ] ] observe that @xmath107 can be split into clusters of equal variables so that @xmath336 and @xmath337 and so on ( recall that @xmath222 gives the sizes of these clusters ) . due to the ordering of @xmath107 , the result we seek to prove follows readily from applying lemma [ prevlem ] to each cluster of @xmath107 starting with the cluster of @xmath104 which equal 1 , and ending with the cluster of @xmath104 equal to @xmath3 .
let us now complete the proof of theorem [ freethm ] .
we may apply lemma [ dblem ] ( given below ) to the solution @xmath298 of the geometric / bernoulli free evolution equation ( the boundary condition on @xmath338 implies the hypotheses of the lemma are met ) .
since @xmath306 for @xmath305 , it follows that the summation over @xmath339 in the outcome of the application of the lemma should be removed and only the @xmath198 term remain ( that is , because all terms involving @xmath340 are necessarily zero ) . comparing the outcome to the ( respective ) true evolution equation of definition [ truedef ]
one finds that they match .
the following is an immediate corollary of the combination of theorems [ freethm ] and [ truethm ] .
[ freethmcor ] consider the ( rate parameter @xmath34 ) continuous time poisson @xmath0-tasep / discrete time geometric @xmath0-tasep ( with time dependent jump parameters @xmath39 ) / discrete time bernoulli @xmath0-tasep ( with time dependent jump parameters @xmath40 ) started from an arbitrary initial condition @xmath219 .
then for any @xmath115 and @xmath31 , @xmath341 = u(t;\vec{n})\ ] ] where @xmath298 solves the ( respective ) free evolution equation with @xmath1 boundary conditions started from initial data @xmath342.\ ] ]
we now conclude the proof of theorem [ momthm ] by showing that the right - hand side of ( [ momthmeq ] ) solves the free evolution equations with @xmath1 boundary conditions which is given in theorem [ freethm ] .
corollary [ freethmcor ] then immediately implies theorem [ momthm ] .
for ease of readability let us recall the right - hand side of ( [ momthmeq ] ) and denote it as @xmath343 there are four hypotheses to check in theorem [ freethm ] .
the first hypothesis to check is that @xmath344 solves the free evolution equation corresponding with the choice of @xmath118 .
since the free evolution equation is linear and separable , it suffices to check that for arbitrary @xmath345 and @xmath30 , @xmath346 satisfies the @xmath347 version of the respective free evolution equations .
* poisson case : * we must check that @xmath348 the @xmath349 brings out a factor @xmath350 on the right - hand side , while the @xmath292 brings out a factor @xmath351 . comparing both sides we find equality . * geometric case : * we must check that @xmath352 the @xmath353 brings out a factor @xmath354 which simplifies to @xmath355 , and hence we find equality .
* bernoulli case : * we must check that @xmath356 the @xmath357 on the left - hand side brings out a factor @xmath358 which simplifies to @xmath359 , while the @xmath360 on the right - hand side brings out a factor @xmath361 which simplifies to @xmath362 . comparing both sides we find equality . the second hypothesis to check
is that the boundary condition is satisfied . without loss of generality
assume @xmath363 ( general @xmath108 is identical ) .
we wish to show that @xmath364_1-q[\nabla]_2\right ) m(t;\vec{n } ) = 0.\ ] ] applying the operator @xmath296_1-q[\nabla]_2 $ ] to the integrand of @xmath344 brings out a factor of @xmath365 .
this new factor cancels the denominator @xmath366 . on account of this
we may freely ( without encountering any poles ) deform the contours for @xmath367 and @xmath368 to coincide .
hence we may write @xmath364_1-q[\nabla]_2\right ) m(t;\vec{n } ) = \int \int ( z_1-z_2 ) g(z_1)g(z_2)dz_1 dz_2\ ] ] where the function @xmath369 involves the integrals in @xmath370 .
since the two contours are identical , this integral is clearly zero , as desired .
the third hypothesis to check is that for @xmath305 , @xmath371 .
this follows from simple residue calculus since when @xmath305 there is no pole at @xmath372 and since this was the only pole contained by the @xmath373 contour , by cauchy s theorem the @xmath373 integral equals 0 .
the final hypothesis to check is the initial data .
since we are dealing with step initial condition @xmath18 , it follows that @xmath374 and hence we must show that @xmath375 .
this follows from residue calculus as well .
expand the @xmath367 contour to infinity .
since @xmath376 , @xmath377 and so the only pole in @xmath367 is encountered at @xmath378 ( @xmath379 is not a pole because of the decay coming from @xmath380 ) . because we pass the pole at 0 from the outside , the contribution of the residue is @xmath381 times the same integral , but with @xmath1 variables . repeating this procedure yields the desired result .
this completes the proof of theorem [ momthm ] .
at least for the poisson and geometric @xmath0-tasep , the true evolution equations are closely related to markov process dualities .
this was understood for poisson @xmath0-tasep in @xcite .
two markov processes @xmath382 and @xmath383 ( with state spaces @xmath384 and @xmath385 respectively ) are said to be dual with respect to a function @xmath386 if for all @xmath387 , @xmath388 and @xmath203 , @xmath389 = { \ensuremath{\mathbb{e}}}^{y } \left[h(x;y(t))\right]\ ] ] where @xmath390 and @xmath391 refer to the expectations of the respective markov chains @xmath382 and @xmath383 started from @xmath392 and @xmath393 .
let us recall the duality for the poisson @xmath0-tasep .
let @xmath394 be the totally asymmetric zero range process with state space @xmath395 from ( [ yn ] ) in which the rate at which a particle moves from site @xmath6 to @xmath7 ( for @xmath396 ) is given by @xmath397 .
notice that this is the same jump rate as in remark [ zerorem ] though presently there are no sources or sinks .
let @xmath10 be the continuous time poisson @xmath0-tasep with @xmath3 particles .
then it is easily checked that @xmath10 and @xmath394 are dual with respect to @xmath398 where for @xmath224 the product is over @xmath399 and for @xmath195 , the product is taken to be zero ( due to the virtual particle @xmath400 ) .
the true evolution equation follows immediately from this duality since @xmath401 = \frac{d}{dt } { \ensuremath{\mathbb{e}}}^y\left[h(\vec{x};\vec{y}(t))\right ] = \sum_{i=1}^{n } { \ensuremath{\mathcal{l}^{(a)}_{i } } } \left[h(\vec{x};\vec{y}(t))\right].\ ] ] here , the first equality is from duality and the second one is from the kolmogorov forward equation and the fact that @xmath402 is the generator of the dual zero range process .
it is also possible to relate the discrete time geometric @xmath0-tasep to a discrete time totally asymmetric zero range process @xmath394 with state space @xmath395 . in parallel
, the state of @xmath394 is updated to that of @xmath403 via the following procedure which occurs at each site @xmath396 . at time @xmath27
there are @xmath404 particles lined up above site @xmath6 .
start with the bottom - most particle and with probability @xmath70 move the particle to site @xmath7 at time @xmath60 .
then proceed sequentially by considering the second particle from bottom , and then the third and so on . for
the @xmath405 particle from bottom , with probability @xmath406 move it to site @xmath7 at time @xmath60 . here
@xmath407 counts the number of particles in site @xmath6 which are below the @xmath405 particle and which have been tagged to move to site @xmath7 at time @xmath60 .
the update rule is sequential within each site @xmath6 , but parallel amongst different sites .
this duality can be proved using similar considerations as in the above proof of the geometric true evolution equation . since it is not necessary for the purposes of this paper
, we leave this just as a remark and do not present this proof .
( ascending ) macdonald processes ( see section 2 and 3 of @xcite for more details ) are measures on triangular arrays of interlacing nonnegative integers @xmath408 we write @xmath409 for the @xmath405 level of the triangle , and note that @xmath409 is a partition of length at most @xmath53 .
the measure on @xmath410 is specified by @xmath38 , @xmath411 , @xmath412 , @xmath121 and two parameters @xmath413 as follows : @xmath414 and the marginal of this measure on a single level @xmath415 is given by @xmath416 in the above , @xmath417 and @xmath418 are macdonald symmetric functions ( indexed by partitions or skew partitions ) and @xmath419 is a macdonald positive specialization ( i.e. , a homomorphism from the algebra of symmetric functions to @xmath420 that sends macdonald symmetric functions to elements of @xmath421 ) specified via the generating function @xmath422 for such a specialization and positive @xmath25 , the numerator in the definition of @xmath423 above is non - negative for all triangular arrays @xmath410 , and as long as the normalizing constant @xmath424 is finite , the above expression describes a probability measure . for a partition @xmath425 of length @xmath3 ,
the macdonald polynomial @xmath426 is an eigenfunction for the first macdonald difference operator @xmath427 ( where @xmath428 ) with eigenvalue @xmath429 notice that when @xmath376 the eigenvalue is simply @xmath430 .
we will assume henceforth that @xmath376 ( though everything except the commutation relations holds for general @xmath27 ) .
the normalizing constant is given by @xmath431 it was observed in @xcite that by linearity of @xmath432 and by the above eigenvalue relation , for any @xmath433 @xmath434.\ ] ] a generalization of this was provided in @xcite , showing that if @xmath435 represents the first difference operator acting on @xmath436 , then for any @xmath437 @xmath438.\ ] ] since the @xmath439 and @xmath440 are explicit and suitably nice , it is possible to encode the above repeated application of difference operators onto @xmath440 in terms of a nested contour integral formula ( one recovers the difference operators by expansion of the formula into residues ) .
it is shown in @xcite that for @xmath441 with @xmath442 , and @xmath443 , @xmath444 = \frac{(-1)^k q^{\frac{k(k-1)}{2}}}{(2\pi \iota)^k } \int \cdots \int \prod_{1\leq a < b\leq k } \frac{z_a - z_b}{z_a - qz_b } \prod_{j=1}^{k } \left(\prod_{i=1}^{n_j } \frac{a_i}{a_i - z_j}\right ) \frac{\pi(qz_j;\rho)}{\pi(z_j;\rho ) } \frac{dz_j}{z_j}.\ ] ] where the contour of integration for @xmath33 contains @xmath34 , and @xmath35 but not 0 or any other poles of the integrand .
plugging in the formula for @xmath440 ( given above in terms of @xmath92 s , @xmath93 s and @xmath445 ) it is immediately clear that @xmath446 = { \ensuremath{\mathbb{e}}}\left[\prod_{i=1}^{n } q^{(x_i(t)+i)y_i } \right],\ ] ] where the expression on the right - hand side corresponds to the version of @xmath0-tasep considered in remark [ arborder ] .
therefore , since these moments characterize the joint law of @xmath447 and @xmath448 , it follows that these two sets are equal in distribution .
there are two immediate questions this equality prompts .
first is whether there is a way to embed the various @xmath0-tasep dynamics into a markov chain on interlacing triangular arrays so that this equality becomes natural .
second is how the true evolution equation ( which the right - hand side of the above equality satisfies ) can be seen directly from the left - hand side as a consequence of macdonald difference operator relations ( without any reference to @xmath0-taseps ) .
we answer both of these questions for continuous time poisson @xmath0-tasep first , and then indicate some extensions of the answer to the second question for the two discrete time versions as well .
there exist markov dynamics on the space of interlacing triangular arrays which preserve the class of macdonald processes .
one example of such a dynamic was introduced and studied in sections 2 and 3 of @xcite ( see @xcite for other such dynamics ) .
when the parameter @xmath376 the continuous time dynamics on the triangular arrays becomes quite simple ( cf . section 3.3 of @xcite ) .
each @xmath449 attempts to increase its value by one according to independent exponentially distributed jumping times with rate given by @xmath450 where terms involving indices outside of the triangular array are omitted . the simplest macdonald process with which to initiate @xmath410 is specified by taking all @xmath451 and corresponds with taking @xmath452 .
then if the above dynamics are run for time @xmath27 ( not to be confused with the macdonald parameter @xmath27 which has been fixed to be zero ) then @xmath453 is distributed according to a macdonald process specified by taking all @xmath454 but @xmath455 .
the above dynamics have the property that the evolution of @xmath456 is marginally markovian ( with respect to its own filtration ) .
in fact , calling @xmath457 we readily observe that @xmath12 evolves according to continuous time poisson @xmath0-tasep with particle jump parameters @xmath34 .
it was through this line of reasoning that poisson @xmath0-tasep was first introduced and through the above macdonald difference operator considerations that the formulas for moments of @xmath458 were initially calculated . in light of equation ( [ lambdax ] )
we are led to the second question , of how the true evolution equation for poisson @xmath0-tasep can be seen directly from the language of macdonald difference operators . in particular
, we would like to show that @xmath459\ ] ] solves @xmath460 we claim that the fact that @xmath199 solves the true evolution equation can be seen as a consequence of a commutation relation for macdonald first difference operators , as well as the fact that by the above discussion @xmath461 1 . for any @xmath464 and any @xmath465 , @xmath466 = ( 1-q^k ) x_n ( d_{n-1}-d_n ) ( d_n)^{k-1}.\ ] ] 2 . for any @xmath464 , @xmath465 and @xmath46 , @xmath467 3 .
for any @xmath464 , @xmath465 and @xmath468 , @xmath469 we first remark that in the statement of the lemma , @xmath470 , @xmath471 and @xmath472 should all be treated as operators which act by multiplication .
thus , we seek to prove the above identities as operators applied to general functions @xmath473 . by polarization
it suffices to prove these identities when applied to functions of the form @xmath474 .
we prove part 1 of the lemma and simply note that analogous considerations ( or comparison to the treatment of the geometric and bernoulli many - body systems earlier in this paper ) suffice to prove parts 2 and 3 .
we can rewrite the desired identity we wish to prove as ( adding harmless @xmath475 factors to both sides ) @xmath476 for @xmath474 .
this may be expressed in terms of nested contour integrals .
the left - hand side becomes @xmath477 while the right - hand side becomes @xmath478 in both sets of integrals we assume that the @xmath33 contour can be chosen so as to contain @xmath479 for @xmath480 as well as @xmath481 , but not 0 or any poles of @xmath482 .
observe that due to the choice of nested contours , it follows that for @xmath483 and any symmetric function @xmath484 ( which does not introduce new poles into the below integrand ) @xmath485 using this , we find that the term @xmath486 in the integrand of ( [ exp1s ] ) can be turned into @xmath487 which exactly matches the term in the integrand of ( [ exp2s ] ) and thus proves the lemma . to conclude ,
let us explain how the commutation relations imply the true evolution equation .
observe that @xmath488 where @xmath489 , and on the right - hand side we assign @xmath490 after applying all of the operators ( we suppress writing this assignment above and in what follows ) .
the commutation relation allows us to pull the @xmath491 in the second term from the right side of the expression to the left side . in the first application of ( [ commute ] )
we replace @xmath492 and find @xmath493 repeating this @xmath3 times we find @xmath494 the right - hand side of which is readily matched to that of the true evolution equation . | we introduce two new exactly solvable ( stochastic ) interacting particle systems which are discrete time versions of @xmath0-tasep .
we call these geometric and bernoulli discrete time @xmath0-tasep .
we obtain concise formulas for expectations of a large enough class of observables of the systems to completely characterize their fixed time distributions when started from step initial condition .
we then extract fredholm determinant formulas for the marginal distribution of the location of any given particle
. underlying this work is the fact that these expectations solve closed systems of difference equations which can be rewritten as free evolution equations with @xmath1 two - body boundary conditions discrete @xmath0-deformed versions of the quantum delta bose gas .
these can be solved via a nested contour integral ansatz .
the same solutions also arise in the study of macdonald processes , and we show how the systems of equations our expectations solve are equivalent to certain commutation relations involving the macdonald first difference operator .
[ section ] [ section ] [ theorem]conjecture [ theorem]lemma [ theorem]proposition [ theorem]corollary [ theorem]claim [ theorem]critical point derivation [ theorem]experimental result # 1 # 1#1 # 1 [ theorem]remark [ theorem]example [ theorem]definition [ theorem]definitions |
the developments in electronics , combining the magnetic and semiconducting materials ( so - called magnetoelectronis or spintronics ) @xcite , have brought half - metallic magnets , initially predicted by de groot and collaborators in 1983 @xcite , to the center of scientific research . in these materials
the two spin bands show a completely different behavior . while the spin - up electronic band structure is metallic , in the spin - down band the fermi level falls within an energy gap as in semiconductors @xcite .
such half - metallic compounds exhibit , ideally , a 100% spin polarization at the fermi level and therefore they should have a fully spin - polarized current and be ideal spin injectors into a semiconductor , thus maximizing the efficiency of spintronic devices @xcite .
the interest on half - metallic magnets has been mainly focused on the heusler compounds .
the are two distinct families of heuslers .
the so - called full - heuslers , with the chemical formula x@xmath0yz , crystallize in the l2@xmath1 structure which consists of four fcc sublattices , where x is a high valent transition or noble metal atom , y is a low - valent transition metal atom and z is a sp element @xcite . in fig .
[ fig1 ] we present the lattice structures of all the compounds revised in this short review .
the second class of heuslers encompass the so - called semi- ( or half- ) heusler compounds of the chemical form xyz , crystallizing in the c1@xmath2 structure .
the structure as shown in fig .
[ fig1 ] is similar to the l2@xmath1 one but now one of the four sites along the diagonal is empty .
nimnsb , a semi - heusler , was the first one predicted to be a half - metal in 1983 by de groot and collaborators @xcite .
the main advantages of heusler alloys with respect to other half - metallic systems ( e.g. some oxides like cro@xmath0 and fe@xmath3o@xmath4 and some manganites like la@xmath5sr@xmath6mno@xmath3 ) @xcite ) are their relatively high curie temperatures @xmath7 @xcite ( while for the other compounds @xmath7 is near the room temperature , _
e.g. _ for nimnsb it is 730 k and for co@xmath0mnsi it reaches the 985 k @xcite ) as well as their structural similarity to the zincblende structure , adopted by binary semiconductors widely used in industry ( such as gaas on zns ) . in 2000 , akinaga and collaborators @xcite reported growth of a few layers of cras in the metastable zincblende phase ( zb ) for the first time ; this was achieved by molecular beam epitaxy on a gaas substrate ( the ground state structure of cras is the mnp structure ) . in the zb phase , cras was reported to be ferromagnetic , with a curie point higher than 400 k. in the same publication , first - principles calculations of bulk zinc - blende cras showed the material to be a half - metallic ferromagnet ; the moment per formula unit was calculated to be 3 @xmath8 , in agreement with the experimental result .
these findings initiated a strong activity on the transition metal ( tm ) pnictides and chalcogenides , because several merits were combined as in heusler compounds mentioned above : the half - metallic property , coherent growth on a semiconductor , and @xmath7 higher than room temperature .
the activity was extended beyond zb cras , encompassing a variety of tetrahedrally bonded tm compounds with @xmath9 atoms of the iv , v and vi groups of the periodic table @xcite . except the tm pnictides and chalcogenides ,
the half - metallic property was also predicted in another class of binary materials the so - called sp - electron ferromagnets or d@xmath10-ferromagnets @xcite .
in two pioneering papers published by geshi et al @xcite and kusakabe et al @xcite , it was shown using first - principles calculations that cap , caas and casb compounds present half - metallic ferromagnetism when grown in the zincblende structure .
several first - principles calculations have been carried in several i / ii - iv / v compounds and it was shown that they are half - metallic magnets in all three zincblende , wurtzite and rocksalt ( rs ) metastable structures ; the ground state in all cases being the rs lattice @xcite . in fig .
[ fig1 ] we present both the zb and rs structures which are similar to the cubic lattice of the heusler compounds with two voids per unit cell .
evidence of the growth of such nanosctructures has been provided by liu et al who have reported successful self - assembly growth of ultrathin can in the rocksalt structure on top of cu(001 ) @xcite .
finally we have to note that materials containing c or n seem to be more promising for applications since the hund energy for the light atoms in the second row of the periodic table is similar to the hund energy of the 3d transition metal atoms .
cubic structure adopted by the semi - heusler compounds , ( ii ) the l2@xmath1 structure adopted by the full - heusler compounds , ( iii ) the zinc - blende structure adopted by the transition - metal pnictides and chalcogenides , and ( iv ) the rocksalt structure of the d@xmath10-ferromagnets .
the cube contains exactly four unit cells . ]
slater and pauling had shown in two pioneering papers that in the case of binary magnetic compounds when we add one valence electron in the compound this occupies spin - down states only and the total spin magnetic moment decreases by about 1 @xmath8 @xcite .
interestingly a similar behavior can be also found in half - metallic magnets as confirmed by first - principles ( ab - initio ) electronic structure calculations . in the half - metallic alloys
the spin - down band - structure is fixed ; the number of spin - down occupied bands and their character does not change among the half - metallic members of the same family of compounds .
the extra valence electron now occupies exclusively spin - up states increasing the total spin magnetic moment by about 1 @xmath8 .
it was shown that in the case of the semi - heusler compounds like nimnsb the total spin magnetic in the unit cell , @xmath11 scales , as a function of the total number of valence electrons , @xmath12 , following the relation @xmath13 @xcite , while in the case of the l@xmath14 full - heuslers this relation becomes @xmath15 @xcite .
moreover for the tm pnictides and chalcogenides as well as the d@xmath10-ferromagnets the corresponding rule takes the form @xmath16 although its origin is different in the two families of compounds @xcite these slater - pauling ( sp ) rules connect the electronic properties ( appearance of the half - metallic behavior ) directly to the magnetic properties ( total spin magnetic moments ) and thus offer a powerful tool to the study of half - metallic compounds since ( i ) magnetic measurements can be used to confirm the half - metallic character of a compound , and ( ii ) simple valence electrons counting can predefine the magnetic properties of a half - metal .
the aim of the present short review is to provide an overview of the origin of the sp rule in all mentioned families of half - metallic magnets as derived from electronic structure calculations based on the density functional theory .
the most studied half - metallic magnets are the heusler compounds .
we will start our discussion from the case of semi - heuslers , like nimnsb , since it was the first to be predicted to be a half - metal @xcite .
the role of the _ sp _ element is to provide in the spin - down electronic band structure a single _ s _ and a triple - degenerated _ p _ band deep in energy ; they are located below the _ d_-states and accommodate _
d_-charge from the transition metal atoms . in fig .
[ fig2 ] we present schematically the character of the bands and their degeneracy in the spin - down band structure .
the bands below the fermi level are occupied while the ones above it are empty .
the _ d_-orbitals of the two transition metal atoms hybridize strongly creating five occupied bonding and five unoccupied antibonding _
d_-states in the spin - down band structure @xcite .
each set of five occupied(unoccupied ) @xmath17-hybrids contains the double degenerated @xmath18 and the triple degenerated @xmath19 states .
note that the notations @xmath18 and @xmath19 are , strictly speaking , valid only for states at the center of the brillouin zone ; however , the energy bands formed by them are energetically rather separated , and therefore this notation can be used to describe the different bands . as a result
there are in total exactly nine occupied spin - down states and the sp relation is @xmath13 , where @xmath11 is the total spin magnetic moment in @xmath8 and @xmath12 the total number of valence electrons @xcite .
the number of the occupied spin - down states is always kept equal to nine for all half - metallic semi - heusler compounds ( the fermi level is fixed within the spin - down energy gap ) and does not depend on the chemical species of the constituent atoms and thus the total spin magnetic moment depends only on the number of valence electrons and not on the specific chemical type of the half - metal , _
e.g. _ femnsb and cocrsb have both 20 valence electrons and a total spin magnetic moment of 2 @xmath8 . when the total spin magnetic moment is positive and thus are more than 18 valence electrons in the unit cell , in the spin - up band structure all nine bonding _
s_- , _ p_- and _ d_-states are occupied , as in the spin - down band , and the extra charge occupies the antibonding spin - up states @xcite . in fig .
[ fig3 ] we have gathered the calculated total spin magnetic moments versus the total number of valence electrons for some selected representative semi heusler compounds .
a special case is mncrsb , which has a zero total spin magnetic moment but mn and cr atoms posses large antiparallel spin magnetic moments @xcite .
these compounds are known as half - metallic antiferromagnets and are of technological importance since they create in - principle zero external fields and thus minimal energy losses in devices . in the case of the half - metallic l@xmath14 full - heuslers the origin of the sp rule
is more complicated due to the more complicated hybridization effects between the _
d_-orbitals @xcite .
the _ sp _ element provides in the spin - down electronic band structure a single _ s _ and a triple - degenerated _ p _ band deep in energy as for the semi - heusler compounds . with respect to the @xmath17-orbtials
, one has first to consider the interaction between the x elements .
although the symmetry of the l@xmath14 lattice is the tetrahedral one , the x elements themselves , if we neglect the y and z atoms , form a simple cubic lattice and sit at sites of octahedral symmetry @xcite .
the _ d_-orbitals of the neighboring x atoms hybridize creating five bonding _ d_-states , which after hybridize with the _ d_-orbitals of the y atoms creating five occupied and five unoccupied _ d_-hybrids , and five non - bonding _ d_-hybrids of octahedral symmetry ( the triple - degenerated @xmath20 and double - degenerated @xmath21 states ) .
these non - bonding hybrids can not couple with the orbitals of the neighboring atoms , since they do not obey the tetrahedral symmetry , and only the @xmath20 are occupied leading to a total of 12 occupied spin - down states and the sp relation is now @xmath15 @xcite . in the case of full - heusler compounds when @xmath22 the spin - up non - bonding @xmath21 states are the first to be occupied followed by the antibonding states , while when @xmath23 the fermi level crosses either the spin - up non - bonding @xmath20 states or the spin - up bonding
d_-states @xcite . in fig .
[ fig3 ] we have gathered some representative cases of calculated half - metallic full - heusler compounds .
the total spin magnetic moment can be : ( i ) zero as for cr@xmath3se , which is a half - metallic antiferromagnet @xcite , ( ii ) positive when @xmath22 going up to a maximum value of 5 @xmath8 for co@xmath0mnsi ( it has been reported that electronic correlation restore half - metallicity in co@xmath0fesi which has a total spin magnetic moment of 6 @xmath8 @xcite ) , or ( iii ) negative as for the half - metallic ferrimagnetic mn@xmath0vge and mn@xmath0val compounds where the v spin moment is antiparallel to the spin moment of the mn atoms @xcite . , in @xmath8 as a function of the total number of valence electrons , @xmath12 , in the unit cell for selected representative semi- and full - heusler compounds .
the two lines represent the two variants of the slater - pauling rule followed by the two families of heuslers . ] for the binary half - metallic magnets under study . both transition - metal pnictides ( chalcogenides ) and d@xmath10-ferromagnets
follow the same slater - pauling rule despite the different origin of the rule ( see fig .
[ fig2 ] ) . ]
the principal mechanism leading to the appearance of the half - metallic gap in zincblende compounds of transition metals ( tm ) with @xmath9 elements is the hybridization of the @xmath17 wavefunctions of the tm atom with the @xmath24 wavefunctions of the @xmath9 atom @xcite .
this _ @xmath24-@xmath17 _
hybridization is dictated by the tetrahedral environment , where each atom is surrounded by four atoms of the other species as shown in fig .
in particular , in the presence of tetrahedral symmetry the @xmath17-states split in two irreducible subspaces : the @xmath19 threefold - degenerate subspace , consisting of the @xmath25 , @xmath26 and @xmath27 states , and the twofold - degenerate @xmath28 subspace , which includes the @xmath29 and @xmath30 states .
only states of the former subspace can hybridize with the @xmath24 states of the @xmath9-atom neighbors , forming bonding and antibonding hybrids . in the spin - down band structure
the bonding states are mainly of @xmath24 character while the antibonding states of @xmath19 character .
the @xmath28 states of the tm atom , on the contrary , remain rather non - bonding .
the situation is shown schematically in fig .
[ fig2 ] for the spin - down band structure . below the fermi level
there is a single - degenerated @xmath31-band due to the @xmath9 atom followed by the bonding @xmath32 hybrids which are mainly of @xmath24 character with a small @xmath19 admixture . above the fermi level
are the non - bonding tm @xmath18 states followed by the antibonding @xmath32 hybrids which are mainly of @xmath19 character .
thus for the transition - metal pncitides and chalcogenides there are in total four occupied states in the spin - down band - structure and the sp rule becomes @xmath16 @xcite . in fig .
[ fig4 ] we have gathered the calculated total spin magnetic moment for some transition metal pnictides and chalcogenides versus the total number of valence electrons .
the total spin magnetic moment can vary from 1 @xmath8 for vge which has 9 valence electrons per unit cell up to 4 @xmath8 for mnas and crse which have 12 valence electrons .
compounds with 13 valence electrons like mnse prefer to be antiferromagnets that half - metallic ferromagnets since in the latter case all antibonding @xmath32 hybrids in the spin - up band - structure should be occupied which is energetically unfavorable @xcite . in the case of d@xmath10-ferromagnets like caas
the sp rule is identical to the tm pnictides and chalcogenides if we consider that the energy gap is located in the spin - down band structure ; in most references like @xcite the energy gap is considered to be located in the spin - up band structure and the sp rule is @xmath33 .
its origin is different than the tm pnictides , discussed above since the d@xmath10-ferromagnets are made up of two @xmath9 atoms and thus no @xmath32 hybridization is present . in the spin - down band - structure the @xmath31 and @xmath24 orbitals of the two @xmath9 atoms hyrbidize creating bonding and antibonding hybrids . in the spin - down band - structure , as shown in fig .
[ fig2 ] , the bonding @xmath31 and @xmath24 states are occupied while the antibonding ones are empty leading to a total of 4 occupied spin - down bands . in the spin - up band structure
the bonding @xmath31 state is occupied and the fermi level crosses the bonding @xmath24-states since we have less than 8 valence electrons in the unit cell ( compounds with 8 valence electrons like cas are semiconductors ) . in fig .
[ fig4 ] we have gathered the calculated total spin magnetic moments for some representative d@xmath10-ferromagnets and the total moment ranges from -3 @xmath8 for kc which has just 5 valence electrons per unit cell up to -1 @xmath8 for caas which has 7 valence electrons in the unit cell . finally we should note that the arguments just presented are valid for both the rocksalt ( which is the most stable structure ) and the zincblende structure since as presented in fig .
[ fig1 ] both obey the tetrahedral symmetry and similar symmetry arguments are valid .
half - metallic magnets are a special class of materials which has attracted considerable attention in spintronics / magnetoelectronics due the predicted perfect spin - polarization of the electrons at the fermi level . in the case of heusler and binary half - metallic compounds , it is possible to formulate slater - pauling rules which connect directly the electronic properties ( number of valence electrons in the unit cell ) to the magnetic properties ( total spin magnetic moment in the unit cell ) .
the fermi level is fixed within the spin - down energy gap and extra electrons accommodate exclusively spin - up states .
these rules are a strong tool in the study of the half - metallic magnets since : ( i ) magnetic measurements can be employed to confirm the half - metallic character of samples , and ( ii ) the chemical formula of a half - metallic compound predefines its total spin magnetic moment .
the origin of the slater - pauling rule differs in the various families of compounds under study and for each case the possible hybridization of the orbitals with respect to both the local and the crystal s symmetry should be taken into account . in semi - heusler compounds , like nimnsb , the hybridization of the @xmath17-orbitals leads to a @xmath13 relation , where @xmath11 the total spin magnetic moment in @xmath8 and @xmath12 the total number of valence electrons in the unit cell . in the case of full - heuslers , like co@xmath0mnsi
, the existence of @xmath17-hybrids of octahedral symmetry located exclusively at the co sites leads to a more complicated hybridization effect and a @xmath15 rule .
finally for both type of binary compounds the slater - pauling rule is @xmath16 but its origin is the @xmath32 hybridization in the transition - metal pnictides and chalcogenides and the @xmath34 hybridization in the d@xmath10-ferromagnets .
p. j. webster and k. r. a. ziebeck , in _ alloys and compounds of d - elements with main group elements .
_ , edited by h. r. j. wijn , landolt - bornstein , new series , group iii , vol .
19,pt.c ( springer - verlag , berlin 1988 ) , pp .
75 - 184 .. | we review the appearance of slater - pauling rules in half - metallic magnets .
these rules have been derived using ab - initio electronic structure calculations and directly connect the electronic properties ( existence of minority - spin energy gap ) to the magnetic properties ( total spin magnetic moment ) in these compounds .
their exact formulation depends on the half - metallic family under study and they can be easily derived if the hybridization of the orbitals at various sites is taken into account .
+ + * keywords : * electronic structure calculations , magnetism , half - metals , slater - pauling + + * corresponding author :* _ postal address : department of materials science , university of patras , gr-26504 patra , tel.:+30 - 2610 - 969925 , fax:+30 - 2610 - 969925 , e - mail address : [email protected]_ |
during the past decades the development of the epitaxial crystal growth techniques such as molecular beam epitaxy and metal - organic chemical vapor deposition has made the growth of the quasi - two - dimensional ( quantum well ) or quasi - one - dimensional ( quantum wire)@xcite systems with controllable well thickness or wire radius became possible .
these quantum structures have been applied to many semiconductor devices , such as high - electron - mobility transistors .
recent progresses in growth and fabrication techniques have been able to fabricate the quantum wires with radii less than 100 @xmath0 .
theoretically , the electronic properties of a hydrogenic impurity in the quantum well5,6,7,8 and the quantum wire @xcite have been studied by many authors .
the impurity binding energies of a quantum wire with infinite or finite potential barrier @xcite and with different shapes of the cross - section@xcite have been discussed .
the effect of location10,11 of impurities with respect to the wire axis was also studied previously .
the emission line for quantum wires was observed@xcite to be two to three times broader than that of quantum wells and with 6 - 10 mev higher binding energy .
it is expected that the same properties in quantum wells were further improved by the reduction of dimensionality to quasi - one - dimensional quantum wires .
the physics of impurity states in quantum wire is very interesting because specific properties can be easily achieved by varying the wire radius .
an electron bound to an impurity on the axis of the quantum wire behaves like a bounded three - dimensional electron when the boundary is far away .
however , as the wire radius is reduced , the electron confinement due to the potential barrier becomes very important . especially in the quantum wire with infinitely high potential wall ,
the total energy of the electron may change from negative to positive at a certain radius and finally diverges to infinity as the radius approaches zero .
furthermore , it is well known that the reduction of dimensionality increases the effective strength of the coulomb interaction .
the binding energy @xmath1 of the ground state of a hydrogenic impurity in n - dimension is given by @xmath1=@xmath2 ^{2}r_{y}^{*}$ ] , where @xmath3 is the effective rydberg .
hence the dramatic change in the binding energy may serve as a clear signal for variation in the effective dimension of the quantum wire .
it is known an electron weakly bound to a hydrogen impurity in a polar semiconductor will interact with the phonons of the host semiconductor . in the past decade , many authors have studied the polaron effect on the binding energy of impurity or exciton in quantum well16,17,18,19,20,21,22,23,24 . recently ,
the electron - phonon effect on the binding energy of the donor impurity in a quantum wire with rectangular cross - section was reported@xcite .
it was found the polaron effect on the binding energy becomes sizeable as the electron gets more deeply bound .
the polaron shifts in donor energy levels are found to be of the order of 10% in a weakly polar system . in studying the polaron effect on the impurity binding energy
, most of the previous works considered the interaction of the electron and bulk optical(bo ) phonon only .
however , in ionic crystal , the motion of an electron near the surface may be affected very much by the surface longitudinal optical ( so ) phonon@xcite .
an electron may be trapped at the surface by the electron - so phonon interaction . besides , the electron phonon interaction hamiltonian in the previous works was valid only for the bulk .
therefore , we will choose the hamiltonian derived by li and chen@xcite , who considered the confined phonon modes in the cylindrical quantum dot . most of the previous approaches concentrating on the polaron effect on the ground state of an impurity in a quasi - one - dimensional wire employ the variational method or perturbation method . since the construction of variational trial wave functions bases entirely on physical intuition , and the estimation of the accuracy of the result obtained from variational approach is very difficult .
furthermore , the perturbation method is only a good access to those systems with very small perturbation in most cases .
therefore , it would be most desirable to have an alternative approach which is not only simple but also efficient to the quantum wire problem . in this work
, we employ a simple approximation treatment which combines the spirit of both variational principle and perturbational approach to study the effect of electron - phonon interactions on the ground state binding energy of a hydrogenic impurity located inside a quantum wire .
consider now a hydrogenic impurity located on the axis of a rigid wall cylindrical quantum wire with a radius @xmath4 .
the hamiltonian of the impurity electron interacting with the phonon can be expressed as:@xmath5 here , @xmath6 is the electronic part of the hamiltonian @xmath7 @xmath8 is the confining potential which is assumed as : @xmath9 and @xmath10 and @xmath11 are the dielectric constant of the well and the effective mass of the electron .
recently , li and chen@xcite has derived the confined the longitudinal - optical phonon and surface phonon modes of a free - standing cylindrical quantum dot of radius @xmath4 and height @xmath12 .
we will follow their hamiltonian and let @xmath13 approach infinity , such that the dot system can become a quantum wire .
therefore , @xmath14 is the bulk phonon hamiltonian which can be expressed as : @xmath15 where @xmath16 is the dispersionless bulk optical ( bo ) phonon energy , @xmath17 is the creation ( annihilation ) operator for bo phonon .
@xmath18 is the interaction between the electron and bo phonon which can be expressed as : @xmath19\end{aligned}\ ] ] with @xmath20 } \left ( \frac{1}{\varepsilon _ { \infty } } -% \frac{1}{\varepsilon _ { 0}}\right ) , \ ] ] where @xmath21 is the mth - order bessel function , @xmath22 is the nth - root of @xmath23 , and @xmath24 is the crystal volume .
@xmath25 is the surface optical phonon ( so ) phonon hamiltonian which can be expressed as : @xmath26 where @xmath27 is the surface optical ( so ) phonon energy , @xmath28 is the creation ( annihilation ) operator for so phonon .
@xmath29 is the interaction between electron and so phonon : @xmath30 with @xmath31 } \notag \\ & & \cdot \left ( \frac{1}{\varepsilon \left ( \omega _ { sp}\right ) -\varepsilon _ { 0}}-\frac{1}{\varepsilon \left ( \omega _ { sp}\right ) -\varepsilon _ { \infty } % } \right ) , \end{aligned}\ ] ] @xmath32 , \ ] ] @xmath33 where @xmath34 , and @xmath35 .
@xmath36 and @xmath37 are , respectively , the mth - order modified bessel function of the first and second kind . following landau and pekar s variational approach@xcite
, the trial wavefunction can be written as : @xmath38 where @xmath39 depends only on the electron coordinate , and @xmath40 is the phonon vacuum state defined by @xmath41 , @xmath42 , and u is a unitary transformation given by : @xmath43 where @xmath44 and @xmath45 are the variational function and the unitary operators @xmath46 and @xmath47 transform the bulk phonon and surface phonon operators as follows : @xmath48 the parameters @xmath49 , @xmath50 , @xmath51 , @xmath52 can be obtained by minimizing the @xmath53 with respect to the parameters @xmath44 , @xmath54 , @xmath45 , @xmath55 .
then the expectation value @xmath56 turns out to be @xmath57 the axis of the wire is assumed to be along the z direction . to solve the electronic part
, one can employ the perturbative - variational approach as we did in the above subsection .
two variational parameters @xmath58 and @xmath59 are introduced by adding and subtracting two terms @xmath60 and @xmath61 into the original hamiltonian @xmath6 and then regroup @xmath6 into three groups:@xmath62 where @xmath63 in the above equations , @xmath64 is treated as a perturbation , and @xmath58 and @xmath65 are treated as variational parameters which can be determined by requiring the perturbation term to be as small as possible .
decomposing @xmath6 into two terms @xmath66 and @xmath67 is equivalent to dividing the space into a two - dimensional ( in xy plane ) and a one - dimensional ( in z - axis ) subspace .
the unperturbed part of the hamiltonian @xmath6 contains two terms , i.e. @xmath66 and @xmath68 , where @xmath66 represents the one dimensional harmonic oscillator , and @xmath69 represents a two dimensional hydrogen atom located inside a quantum disk@xcite .
both can be solved exactly . for illustration ,
the ground state energy and wavefunction of the unperturbed part can be expressed as : @xmath70 respectively , where @xmath71 is the ground state wavefunction of the 1d harmonic oscillator , and @xmath72 is the ground state wavefunction of the 2d hydrogen atom located at the center of an infinite circular well .
the ground state eigenvalue and eigenfunction of the 1d harmonic oscillator can be expressed as : @xmath73 the ground state eigenvalue and eigenfunction of the 2d hydrogenic impurity located at the center of an infinite circular well can be obtained as14:(1 ) for @xmath74 , @xmath75 where @xmath76 , @xmath77 , @xmath78 , @xmath79 is the confluent hypergeometric function , and @xmath80 is the normalization constant.(2 ) for @xmath81 , @xmath82 where @xmath83 , @xmath84 , @xmath85 , @xmath86 is the irregular coulomb wave function , and @xmath87 is the normalization constant.(3 ) the turning point for energy changing from @xmath81 to @xmath74 in the quantum circle system may be determined by setting @xmath88 = 0\;\;for\;\;m=0,\ ] ] and@xmath89 = 0\;\;for\;\;m=1,\ ] ] the requirement of the continuity of the wavefunctions and its first derivative at boundary yields:(1 ) for @xmath74 , @xmath90 ( 2 ) for @xmath81 @xmath91 the eigenvalues are then given as : @xmath92 the first order energy correction can thus obtained as : @xmath93 the second term of the above equation can be integrated analytically and the result is : @xmath94 then the total energy up to the first order perturbation correction can then be obtained as : @xmath95 the variational parameters are then chosen by requiring the total energy @xmath96 to be minimized with respect to the variation of @xmath97 and @xmath65 .
this is equivalent to requiring : @xmath98 @xmath99 for the excited states , the eigenvalues and eigenfunctions can be treated in the same way .
we have calculated the effect of the confined the longitudinal - optical phonon and surface phonon interactions on the hydrogenic impurity located in a quantum wire . and the well potential is considered as infinite .
figure 1 shows the ground state energy as a function of the wire radius .
the binding energy @xmath1 of the hydrogenic impurity is defined as the energy difference between the ground state energy of the cylindrical wire system with and without the impurity , i.e.@xmath100where @xmath101 is the ground state energy of the quantum wire system without the impurity , while @xmath102 is the ground state energy of the quantum wire system with the impurity located on the axis of the cylindrical wire .
one can see from fig.1 that the energy of the 1@xmath103 state becomes negative when the wire radius is larger than 1.65@xmath104 .
it means that the confining energy is larger than the coulomb energy as the wire radius is smaller than 1.65@xmath104 .
and one can also note that as the radius of the quantum wire is decreased , the ground state energy increases .
as the wire radius @xmath4 becomes smaller , the electron is pushed toward the axis of the cylindrical wire .
this makes the electron get close to the nucleus . as the electron gets close to the nucleus , both the ground state energy and the binding energy increase rapidly
this is because the coulomb potential , which varies with @xmath105 ( @xmath4 is the wire radius ) , becomes more negative , while the kinetic energy of the electron , which varies with @xmath106 ( by the uncertainty relation ) , increases more rapidly . as a result
, the ground state energy is increased as the electron gets close to the nucleus . the binding energy defined in eq .
( 39 ) is effectively the negative sign of the of the coulomb interaction energy between the electron and the nucleus , i.e. @xmath107 , therefore , the binding energy of the electron is also increased as the electron gets near to the nucleus .. as a result , the ground state energy is increased as the electron gets close to the nucleus . the binding energy defined in eq .
( 39 ) is effectively the negative sign of the of the coulomb interaction energy between the electron and the nucleus , i.e. @xmath107 , therefore , the binding energy of the electron is also increased as the electron gets near to the nucleus .
our results show that for small wire radius , the binding energies are in good agreement with previous results@xcite . as the radius becomes very large , our result approaches the correct limit 1@xmath108 while the previous work @xcite can only yield a value of 0.22@xmath108 .
the large discrepancy of the previous work may be due to the artificial dividing of the variational trial wavefunction into a one - dimensional hydrogen atom and a two - dimensional hydrogen atom and thus forces the creation of an additional node of the wavefunction at z=0 . in this work
, the trial wavefunction is adopted to be the form of 1d harmonic oscillator wavefunction instead of one dimensional hydrogen atom .
this prevents our wavefunction from introducing any additional node at z=0.figure 2 presents the 2s excited state binding energies as the functions of wire radius .
one can note from the figure that as the wire radius increases , the binding energy approaches 0.25@xmath108 which gives correctly the limiting value of 3d hydrogen atom .
figure 3 presents the confined bo phonon and so phonon effect as a function of wire radius . with increasing the wire radius , the magnitude of the confined bo phonon effect decreases from large value and then approaches to the bulk value .
when the wire radius is less than 1.5 a * , the polaron effect increases rapidly .
one might think as the radius becomes very small , the confined bo phonon effect should approach zero , like the case in quantum well @xcite .
in fact , similar results were obtained by oshiro in a spherical quantum dot@xcite .
they found the polaron energy shift is enhanced as the dot radius becomes small .
this is due to the fact that the electron becomes complete localized ( e@xmath109 approaches infinity ) in small wire ( or dot ) radius while the binding energy approaches 4@xmath108 in small well width . in the case of quantum
well , the confined so phonon effect plays the dominant role for small well width@xcite .
but in quantum wire , the confined so phonon is less important , just like that in quantum dot system @xcite .
this is because the surface area of a quantum wire ( or quantum dot ) decreases with the radius .
thus the number of vibration modes of confined so phonon becomes fewer . in fig.4 ,
three curves are presented .
the dotted curve represents the binding energy of the impurity without considering the interactions between the electron and phonon .
the dashed curve represents the binding energy of the impurity with only confined bo phonon effect being taken into account . while the solid curve is the binding energy of the impurity including both confined bo phonon and so phonon effects in the calculation . comparing to
the impurity binding energy , the confined so phonon is negligible in quantum wire .
we then conclude that because of the similarity in geometry , the behavior of the polaron effect on the quantum wire system is like that on the quantum dot system .
in this work , analytical solutions for the effects of the electron - phonon interaction on the binding energies of an impurity located inside a quantum wire are obtained by a simple but efficient perturbation - variation method . as the radius becomes very large
, the correct limiting value can be obtained .
we have also discussed both the confined bo and so phonon effects .
we found the confined bo phonon effect is prominently for a quantum wire with small radius .
we also found that the energy corrections of the polaron effects on the impurity binding energies increase rapidly when the wire radius is less than 1.5 a*. | the effect of electron - optical phonon interaction on the hydrogenic impurity binding energy in a cylindrical quantum wire is studied . by using landau and
pekar variational method , the hamiltonian is separated into two parts which contain phonon variable and electron variable respectively . a perturbative - variational technique is then employed to construct the trial wavefunction for the electron part .
the effect of confined electron - optical phonon interaction on the binding energies of the ground state and an excited state are calculated as a function of wire radius .
both the electron - bulk optical phonon and electron - surface optical phonon coupling are considered .
it is found that the energy corrections of the polaron effects on the impurity binding energies increase rapidliy as the wire radius is shrunk , and the bulk type optical phonon plays the dominant role for the polaron effects .
pacs : 71.38+i;73.20.dx;63.20.kr |
in ref . @xcite we have shown through the example of the extraction of the cabibbo - kobayashi - maskawa ( ckm ) angle @xmath0 from @xmath1 decays that the bayesian treatment as done , e.g. , by the utfit collaboration @xcite , suffers from major difficulties .
the problems we have found are related to the presence of unknown free parameters which should actually be constrained by the data .
we have shown that the results of the bayesian analysis depend on the priors and the chosen parameterization in an uncontrollable manner , and may even diverge in some cases .
furthermore we have demonstrated that the priors can not always be specified in a consistent way with respect to the symmetries of the problem , which results in the present example in an incorrect description of the @xmath2 limit .
the authors of ref .
@xcite replied to our criticism . in the first part of their paper
they agree with us on the dependence of the analysis with respect to the choice of priors and parameterization , even when the data are sufficiently precise so that the posterior pdfs are contained in the corresponding prior ranges .
@xcite claims that this should not be viewed as a fundamental drawback of the bayesian treatment , because in the absence of additional information on the parameters one should use uniform priors on the quantities that are directly measured ( what we have called the `` explicit solution '' parameterization in @xcite ) .
we emphasize in the following that this recipe can only be ( approximately ) justified within the framework of classical statistics and is thus not actually bayesian .
moreover , it is far from being general . in the second part of ref . @xcite it is argued that in addition to the isospin analysis one may use bounds on the magnitude of the fit parameters related to hadronic matrix elements , and when taken into account the bayesian treatment behaves more reliably .
while this new recommendation may appear physically sound , the problem reported in @xcite uses the original isospin analysis @xcite , where the hadronic parameters are unknown and only isospin symmetry is used . changing the hypotheses can not be considered as a satisfying answer to our findings : indeed the validity of the statistical approach must be proven independently of the problem at hand . despite this inconsistency ,
we accept the new proposal and show below that even with the use of this additional information one finds important quantitative and qualitative differences between the frequentist and the bayesian results . in any case , the new recommendations of the utfit collaboration for treating problems with _ a priori _ free unknown parameters differ quite importantly from their previous publications , which implies that the results published therein are obsolete .
we are therefore expecting new analysis results for the angles @xmath0 from @xmath3 and @xmath4 from charmful modes , for which similar difficulties can be apprehended .
we conclude that ref .
@xcite does not abrogate our criticism , because the only known examples where the bayesian treatment leads to a numerically reasonable result do not provide any general argument in favor of the underlying hypotheses .
in contrast , the frequentist approach can handle many types of situations smoothly , independently of whether there are free parameters and/or theoretical uncertainties .
the term `` isospin analysis '' ( which in ref .
@xcite is replaced by `` minimal assumptions '' ) refers to the original paper by gronau and london @xcite . within this framework ,
only isospin symmetry is used to parametrize the decay amplitudes .
although it is not an exact symmetry , its accuracy in @xmath5 is expected to be of the order of 1% @xcite ( once the known dominant contribution from electroweak penguins has been taken into account ) and is thus of phenomenological relevance .
neither the standard model nor isospin symmetry imply anything on the `` natural '' scale of the hadronic amplitudes . in ref .
@xcite we showed that the pivk - le diberder ( pld ) and explicit solution ( es ) parameterizations give the most reasonable answer for the bayesian treatment , because the corresponding parameters are close or identical to the measured quantities .
this is confirmed ( for the es parameterization ) by the authors of ref .
@xcite , who thus agree on the fact that other parameterizations lead to unstable results . the better behavior of the es description ( which follows from our detailed analysis and should be acknowledged as such ) can be easily understood as follows .
let us stick to the gaussian case and first assume that there is no degeneracy so that the angle @xmath0 is given by a known single - valued function of the observables .
one can construct a sampling distribution for @xmath0 as defined by its analytical expression by randomizing the observables according to gaussian pdfs where the true ( unknown ) central values and standard deviations have been replaced by their corresponding measured estimators .
it is then straightforward to show that the position of the peak and its width are _ consistent estimators _ of , respectively , @xmath0 and its uncertainty .
this procedure does not need to define priors nor integration on the true values of the parameters , and does not rely on bayes theorem : it just follows from standard rules of classical statistics .
technically this procedure can be reproduced by applying the `` bayesian '' treatment of ref .
@xcite to the es parameterization @xcite with uniform priors .
because for gaussian measurements the standard deviation of a quantity is equal to the 68% cl frequentist error , it is not a surprise to have similar numbers in the es parameterization between the `` bayesian '' and frequentist treatments .
the remaining differences , namely the fact that one finds exactly degenerate mirror solutions only in the frequentist approach , mainly come from the integration ( logical and ) over mirror solutions in the randomization treatment , which results from an unacceptable interpretation of physical constants that could simultaneously take different values @xcite . besides , the explicit solution trick is not general enough to treat with `` minimal assumptions '' all the interesting physical problems .
even when there is the same number of parameters as the number of independent observables , there is no guarantee that the theoretical equations can be inverted to express the parameters as functions of the observables .
the gronau - london isospin analysis is quite special in this respect .
more importantly , most of the time one has overconstrained systems , that is one has more observables than parameters . in this case
no `` explicit solution '' parameterization exists , and no general probability argument would help to find what is the most `` natural '' parameterization for a bayesian treatment with uniform priors . to close this section
we point out two misconceptions in ref .
the first one is related to the claim that the 68% or 95% confidence level or credible intervals are more meaningful than the full curve .
the fact that experimental measurements are often summarized as one- or two standard deviation intervals is just a matter of convention and if the curve has a complicated shape , it can not be reduced to a one or two - number description without loss of information .
we emphasize that the full answer is given by the posterior distribution , and reporting only these summaries in the case of complex distributions ( e.g. multimodal and/or asymmetrical pdfs ) can be misleading , because people tend to think of a gaussian model if no further information is provided . _ '' ( g. dagostini , 2003 @xcite ) . ] in particular for the present case , we emphasize the importance of having exactly degenerate peaks in the angle @xmath0 ( as are reproduced by the constrained frequentist fit , see below ) to be in agreement with isospin symmetry no matter the values of the 68% or 95% error intervals .
the second misconception is related to the above misunderstanding of the meaning of the degeneracy that is intrinsic to the isospin analysis .
the fact that with the new data from summer 2006 @xcite one sees only four peaks instead of eight as a function of @xmath0 is not related to the improved measurements , but to the fact that now the experimental central values are slightly outside the physical region that is defined by the isospin symmetry ( which is related to the property that both the three amplitudes and their @xmath6 conjugates must form a triangle in the complex plane ) .
most presumably the data will evolve in such a way that the eightfold ambiguity reappears , which is the general case in the mathematical sense , and not a `` fortuitous accident '' as claimed in @xcite .
would a different ambiguity pattern be confirmed by better data , then either the standard model or the isospin symmetry should be questioned .
the authors of ref .
@xcite recommend that the `` default '' plot summarizing the present constraints on the angle @xmath0 coming from @xmath5 decays should take into account additional theoretical and phenomenological information .
the physical arguments that are presented in favor of such an approach are perfectly legitimate and we do not object them .
however the problem that was studied in detail in @xcite is not this one , but rather the original isospin analysis .
let us recall the advantages of performing the analysis assuming only isospin symmetry .
* neither isospin symmetry nor @xmath5 experimental measurements give any insight on the typical scale of the hadronic amplitudes ; thus it is a natural choice for the experimental collaborations to present their results in a way that is independent of possible assumptions on the hadronic amplitudes .
* the original paper @xcite only assumed isospin symmetry .
* one can think of many relevant physical problems containing completely free unknown parameters , the typical scale of which is not controlled even as an order of magnitude .
this is the case , _
e.g. _ , of general beyond the standard model scenarios where one parametrizes new arbitrary contributions with a few quantities that _ a priori _ can take any value .
the original isospin analysis is thus a pedagogical example that exhibit many features that would appear in more general situations . having this in mind we can now examine the new proposal of ref .
first we remark that the statement made in @xcite according to which one can get new constraints on @xmath0 , and eventually lift the degeneracies and suppress the solution at @xmath7 , provided one uses additional theoretical and phenomenological knowledge , is obviously trivial and was the guideline of , e.g. , section vi.1 of ref .
@xcite and many similar studies in the past .
second we stress that additional information is ill - defined .
the assumptions made in @xcite are equivalent to an analysis with a finite theoretical error on each parameter ( except the one that is scanned , here the angle @xmath0 ) , such as the `` historical '' determination of the unitarity triangle @xcite .
it is well known that in general there is no unambiguous definition of the meaning of a theoretical error , and when such uncertainties are present the bayesian and frequentist methods can not be compared in a rigorous way .
the authors of ref .
@xcite take as a strong argument in favor of their approach the fact that their results are weakly dependent of the assumed order of magnitude for the amplitudes .
we do not find this convincing , but rather would like to take the following two limits , which to our knowledge are the only ones where the concept of a theoretical error becomes perfectly well defined .
* theory errors go to zero , _
i.e. _ the corresponding parameters are fixed constants . in this case
it is trivial to show that both frequentist and bayesian approaches lead to equivalent numerical results .
* theory errors go to infinity , _
i.e. _ the corresponding parameters are completely free unknowns ; this is the case that is discussed in the previous section and where the bayesian treatment in a generic parameterization simply fails .
hence , only in the very specific case where both experimental and theoretical uncertainties are `` sufficiently '' reduced , and there is no free parameter , the numerical comparison of the bayesian treatment with the frequentist classical approach may become meaningful , and in the bayesian case one might expect a `` reasonable '' stability with respect to priors and parameterization .
this stability , however , must be extensively checked case - by - case , which has not been done in the previous publications by the utfit collaboration @xcite . , taking into account the following constraints @xcite : @xmath8 , @xmath9 and @xmath10 ( natural units ) , for @xmath5 decays .
upper plot : data that correspond to the plots in ref .
@xcite , lower plot : new data that correspond to fig . 5 in @xcite .
[ constrainedfit],title="fig:",scaledwidth=40.0% ] , taking into account the following constraints @xcite : @xmath8 , @xmath9 and @xmath10 ( natural units ) , for @xmath5 decays .
upper plot : data that correspond to the plots in ref .
@xcite , lower plot : new data that correspond to fig . 5 in @xcite . [ constrainedfit],title="fig:",scaledwidth=40.0% ]
[ constrainedfit ] shows the results of a constrained frequentist fit that takes into account the same information on the hadronic parameters as in fig . 5 of @xcite , namely that the amplitudes verify the following bounds : @xmath8 , @xmath9 and @xmath10 ( in natural units ) .
in particular these bounds suppress the parameter configurations with @xmath2 , as emphasized in @xcite .
have a good confidence level because large , but finite , values for the hadronic amplitudes can generate the observed @xmath6 violation
. however the @xmath6-conserving value @xmath11 is strongly disfavored in the standard model , so that the confidence level is discontinuous at this point , as explained in @xcite .
contrary to the claims in @xcite , these results are in perfect agreement with quantum field theory and do not violate any fundamental principle .
] indeed this curve is not too different from the one that corresponds to the bayesian treatment , however it appears clearly that the frequentist analysis , in contrast to the bayesian one , respects the exact degeneracy between the mirror solutions that survive the phenomenological bounds , as requested by the symmetries of the problem .
table [ param ] shows the parameter values that correspond to each of the peaks in the plots of fig .
[ constrainedfit ] , as obtained analytically from the exact formulas or numerically from the fit . ) , the exact degeneracy of the constrained fit is only threefold because the central values are outside the physical region ; again the parameter values that correspond to the peaks exactly reproduce the same observable values , and all satisfy the phenomenological bounds on the amplitudes , so that there is no possibility to tell which solution is to be preferred . ] since these parameter configurations all lead to the very same observables and all verify the phenomenological bounds advocated in ref .
@xcite , there is no way to tell which configuration is `` more probable '' . in turn
, this means that the bayesian treatment lifts the degeneracy only because of the unphysical marginalization over the hadronic parameters , and that reducing the output information to the 68% or 95% probability intervals is very misleading . from table
[ param ] one sees that to lift the degeneracy between the remaining solutions , even partially , one would need to know the amplitudes with a relative uncertainty of better than 100% , which is feasible using more dynamical approaches to hadronic elements , but which would request far more than just isospin symmetry and orders of magnitude for the hadronic amplitudes . to close this section
we emphasize that there are many interesting problems depending on completely unknown parameters on which we have no clue , not even a rough order of magnitude .
this is for example the case in generic new physics scenarios , such as the one of the second paper of ref .
@xcite . there
arbitrary new physics contributions to @xmath12 mixing is parametrized by a quantity @xmath13 ; by definition , since no assumption is made on specific new physics models , the analysis does not give any insight to the `` natural '' scale of the parameter @xmath14 .
thus one again may expect the bayesian treatment to suffer from instabilities when moving from , e.g , the @xmath15 to the @xmath16 parameterization ( such instabilities are expected to increase with the number of new physics parameters ) . in other words
the bayesian approach as recommended by ref .
@xcite can not be applied ( although in practice it is ) to the full generality of problems at hand .
centre de physique thorique is umr 6207 du cnrs associe aux universits daix - marseille i et ii et universit du sud toulon - var ; laboratoire affili la frumam - fr2291 .
this work was supported in part by the eu contract no .
mrtn - ct-2006 - 035482 , `` flavianet '' .
9 m. bona _ et al . _
[ utfit collaboration ] , arxiv : hep - ph/0701204 j. charles _ et al .
_ , arxiv : hep - ph/0607246 m. bona _ et al .
_ [ utfit collaboration ] , jhep * 0507 * ( 2005 ) 028 ; jhep * 0603 * ( 2006 ) 080 ; jhep * 0610 * ( 2006 ) 081 m. gronau and d. london , phys . rev . lett .
* 65 * ( 1990 ) 3381 m. gronau and j. zupan , phys .
d * 71 * ( 2005 ) 074017 m. ciuchini _ et al .
_ , jhep * 0107 * ( 2001 ) 013 g. dagostini , rep .
* 66 * ( 2003 ) 1383 the heavy flavor averaging group , http://www.slac.stanford.edu/xorg/hfag/ j. charles _ et al . _
[ ckmfitter group ] , eur .
phys . j. c * 41 * ( 2005 ) 1 | in reply to ref . @xcite we demonstrate why the arguments made therein do not address the criticism exposed in ref .
@xcite on the fundamental shortcomings of the bayesian approach when it comes to the extraction of parameters of nature from experimental data . as for the isospin analysis and the ckm angle
@xmath0 it is shown that the use of uniform priors for the observed quantities in the explicit solution parametrization is equivalent to a frequentist construction resulting from a change of variables , and thus relies neither on prior pdfs nor on bayes theorem .
this procedure provides in this particular case results that are similar to the confidence level approach , but the treatment of mirror solutions remains incorrect and it is far from being general . in a second part it is shown that important differences subsist between the bayesian and frequentist approaches , when following the proposal of ref .
@xcite and inserting additional information on the hadronic amplitudes beyond isospin invariance . in particular the frequentist
result preserves the exact degeneracy that is expected from the remaining symmetries of the problem while the bayesian procedure does not .
moreover , in the bayesian approach reducing inference to the 68% or 95% credible interval is a misconception of the meaning of the posterior pdf , which in turn implies that the significant dependence of the latter to the chosen parametrization can not be viewed as a minor effect , contrary to the claim in ref . @xcite . |
the fermi coupling constant , @xmath1 , constitutes together with the electromagnetic coupling constant and the mass of the @xmath2 boson the most precise input parameters of the standard model of elementary particle physics .
@xmath1 is defined through the muon lifetime , and the decay of the muon , as a purely leptonic process , is rather clean both experimentally and theoretically .
the one - loop corrections of order @xmath3 were computed more than 40 years ago @xcite whereas only recently the two - loop corrections of order @xmath4 have been evaluated @xcite .
the large gap in time shows that this calculation is highly non - trivial .
the inclusion of the two - loop terms removed the relative theoretical error of @xmath5 which was an estimate on the size of the missing corrections .
the remaining error on @xmath1 now reads @xmath6 and is of pure experimental nature .
upcoming experiments will further improve the accuracy of the muon lifetime measurement and therefore the @xmath7 corrections to the muon decay are very important and constitute a crucial ingredient from the theoretical side .
these facts make it desirable to have an independent check on the correctness of the @xmath8 result .
we also want to mention that in a recent article @xcite optimization methods have been used in order to estimate the coefficient of order @xmath9 . in view of the upcoming @xmath10 physics experiments
the evaluation of quantum corrections to @xmath10 meson properties have become topical . in particular
it is possible to use the semileptonic decay rate of the bottom quark in order to determine the cabibbo - kobayashi - maskawa ( ckm ) matrix elements with quite some accuracy . in
@xcite an approximate expression for the @xmath11 corrections of @xmath12 has been obtained where a non - vanishing charm quark mass has been included . in the decay @xmath13
the mass of the @xmath14 quark can be neglected which reduces the calculation to the situation given in the muon decay .
the only additional diagrams are those which arise due to the non - abelian structure of qcd .
the results of order @xmath0 have been obtained in @xcite .
the corrections proportional to the number of light quarks have already been computed in @xcite . in this letter
we confirm through an independent calculation the results of order @xmath4 to the muon decay @xcite and the one of order @xmath0 to the semileptonic bottom quark decay @xcite . in the next section our method
is discussed , in section [ sec : res ] the results are presented .
in @xcite the imaginary part of the muon propagator is computed up to the four - loop level .
recurrence relations based on the integration - by - parts technique @xcite are used in order to reduce the integrals to be evaluated to a minimal set so - called master integrals .
they are finally evaluated by computing expansions in the ratio of external momentum and internal muon mass .
it is possible to take the on - shell limit and perform the infinite sum which finally leads to an exact result for the integrals . for concise reviews of expansion methods
see e.g. @xcite .
in contrast to @xcite our approach is based on an expansion of the full fermion propagator in the limit @xmath15 where @xmath16 is the external momentum and @xmath17 is the propagator mass of the muon and bottom quark , respectively .
these diagrams are computed without any expansion . ] . throughout the whole paper
we will neglect effects induced by the non - vanishing electron and up quark mass .
the on - shell limit @xmath18 will be performed afterwards with the help of pad approximations .
this , of course , only provides an approximation to the exact result .
however , the integrals to be evaluated are simplified considerably .
we will demonstrate that the accuracy obtained with our method is sufficient to check the existing results and enables the same reduction of the theoretical error on @xmath1 .
the notation is essentially adopted from @xcite , where corrections of @xmath11 to the decay @xmath19 have been computed . for completeness
we briefly repeat in this paper the main formulae . the decay rate both for the muon and the bottom quark can be written in the form @xmath20\bigg|_{z=1 } , \label{eq : gam}\end{aligned}\ ] ] where @xmath21 are functions of the variable @xmath22 @xmath17 is the on - shell mass . @xmath23 and @xmath24 represent the scalar and vector part of the corresponding fermion propagator , respectively .
they are functions of the external momentum @xmath16 and the bare mass @xmath25 of the fermion under consideration . in our case
they further depend on the bare electromagnetic coupling @xmath26 and the strong coupling constant @xmath27 , respectively , and are proportional to the square of the fermi coupling constant , @xmath28 .
the mass renormalization constant , @xmath29 , entering in ( [ eq : sssv ] ) can be extracted from @xcite .
in contrast , @xmath30 has to be evaluated in the limit @xmath31 since the handling of @xmath30 is determined by the computation of the fermion propagator . as we are only interested in the imaginary part and furthermore consider only qed or qcd corrections to the leading order term the quantity @xmath30 has to be known up to two loops only .
the result can be taken from @xcite .
the renormalization of @xmath3 and @xmath32 proceeds in the usual way where we have chosen to renormalize also the electromagnetic coupling in a first step in the @xmath33 scheme . in order to get reliable results
it is necessary to compute as many terms as possible in the expansion parameter @xmath34 .
subsequently a pad approximation is applied which is at length described in @xcite . we just want to mention that before the pad procedure a conformal mapping can be used which maps the complex @xmath34-plane into the interiour of the unit circle . following ref .
@xcite we denote those results by @xmath35-pads and the ones obtained without conformal mapping by @xmath34-pads .
some pad approximants develop poles inside the unit circle ( @xmath36 and @xmath37 , respectively ) in conflict with the analycity of the exact result . in general
we will discard such numbers in the following . in some cases
, however , the pole coincides with a zero of the numerator up to several digits accuracy , and these pad approximations will be included in our sample . to be precise : in addition to the pad results without any poles inside the unit circle
, we will use the ones where the poles are accompanied by zeros within a circle of radius 0.01 , and the distance between the pole and the physically relevant point @xmath38 is larger than 0.1 .
the central values and the estimated uncertainty will be extracted from pad results @xmath39 $ ] with @xmath40 not too small and @xmath41 .
the central value is obtained by averaging the pad results and the uncertainty is given by the maximal deviation from the central value .
it is convenient to parameterize the radiative corrections for the semileptonic bottom quark decay in the following form : @xmath42 \ , , \nonumber\\ a_b^{(2 ) } & = & c_f^2 a_{b , a}^{(2 ) } + c_a c_f a_{b , na}^{(2 ) } + c_f t n_l a_{b , l}^{(2 ) } + c_f t a_{b , f}^{(2 ) } \ , , \label{eq : gamparb}\end{aligned}\ ] ] with @xmath43 . for qcd
the colour factors are given by @xmath44 , @xmath45 , and @xmath46 .
@xmath47 is the number of massless quark flavours and will be set to @xmath48 at the end .
@xmath49 corresponds to the abelian part already present in qed , @xmath50 represents the non - abelian contribution , and @xmath51 and @xmath52 denote the corrections involving a second fermion loop with massless and heavy quarks , respectively . in principle
there is also a contribution involving a virtual top quark loop .
it is , however , suppressed by @xmath53 and will thus be neglected . in fig .
[ fig : bdec ] a representative diagram for each one of these four colour structures is pictured . in eq .
( [ eq : gamparb ] ) @xmath54 is defined with five active flavours .
the analytic result for @xmath55 can be found in @xcite .
an analogous expression to ( [ eq : gamparb ] ) can also be defined in the case of the muon decay @xmath56 \ , , \nonumber\\ a_\mu^{(2 ) } & = & a_{\mu,\gamma\gamma}^{(2 ) } + a_{\mu , e}^{(2 ) } + a_{\mu,\mu}^{(2 ) } \ , , \label{eq : gamparmu}\end{aligned}\ ] ] with @xmath57 .
@xmath58 represents the electromagnetic coupling in the @xmath33 scheme .
@xmath59 represents the purely photonic corrections whereas @xmath60 and @xmath61 contain an additional electron and muon loop , respectively .
the contribution involving a virtual @xmath62 loop is not listed in ( [ eq : gamparmu ] ) as it is suppressed by @xmath63 and almost four orders of magnitudes smaller than the other terms @xcite .
let us in the following describe our method used for the practical calculation in the case of the @xmath64 decay .
the difference to the quark decay consists only in the transition from qed to qcd which increases the number of diagrams and makes it necessary to include the colour factors ; the idea is , however , applicable in the very same way .
following common practice , we investigate the effective theory where the @xmath65 boson is integrated out . the qed corrections to the resulting fermi contact interaction
were shown to be finite to all orders @xcite .
it is quite advantageous to perform a fierz transformation which for a pure @xmath66 theory has the consequence that afterwards the two neutrino lines appear in the same fermion trace .
thus the qed corrections only affect the fermion trace involving the muon and the electron .
this also provides some simplifications in the treatment to @xmath67 since in the case of vanishing electron mass a fully anticommuting prescription can be used .
as described above we consider the fermion two - point functions and compute the imaginary part arising from the intermediate states with two neutrinos and the electron . as a consequence already for the born result a two - loop diagram has to be considered .
however , it turns out that the loop integration connected to the two neutrino lines can be performed immediately as it constitutes a massless two - point function .
this is also the case after allowing for additional photonic corrections .
as a result one encounters in the resulting diagram a propagator with one of the momenta raised to power @xmath68 where @xmath69 is the space - time dimension .
this slightly increases the difficulty of the computation of the resulting diagrams . especially for the order @xmath4 corrections , where the original four - loop diagrams are reduced to three - loop ones with non - integer powers of denominators , it is a priori not clear that these integrals can be solved analytically .
however , it turns out that for the topologies needed in our case this is indeed possible . for the computation of the massless two - point functions we have used the package mincer @xcite .
slight modifications enabled us to use this package also for the computation of the new type of integrals .
the calculation is performed with the help of the package geficom @xcite .
it uses qgraf @xcite for the generation of the diagrams and exp @xcite for the application of the asymptotic expansion procedures . for more technical details we refer to a recent review concerned with the automatic computation of feynman diagrams @xcite .
the dispersive and absorbtive part of the fermion self energies are gauge dependent for @xmath70 . hence the dependence on the gauge parameter @xmath71 in eq .
( [ eq : gam ] ) only drops out after summing infinitely many terms in the expansion around @xmath72 and setting @xmath73 .
since we are only dealing with a limited number of terms , our approximate results will still depend on the choice of @xmath71 even after taking @xmath74 .
it is clear that with extreme values of @xmath71 almost any number could be produced .
thus the question arises which value of @xmath71 should be taken in order to arrive at a reliable prediction for the decay rates . as the one - loop corrections can be evaluated for an arbitrary gauge parameter an extensive study can be performed and we can gain some hints for the choice of @xmath71 at order @xmath4 .
let us in a first step present the results for the muon decay and afterwards discuss the additional diagrams necessary for the qcd corrections to the bottom quark decay .
the lowest order ( born ) diagram can be computed directly . in this case
the electron mass can be chosen different from zero and an expansion in @xmath75 can be performed reproducing the exact result @xmath76 we want to demonstrate the power of our method for the order @xmath3 corrections where four three - loop diagrams contribute . at this order
we are able to evaluate a large number of moments which gives us a suggestion how many terms are necessary at @xmath7 in order to obtain reliable results .
furthermore the computation can be performed for arbitrary gauge parameter which also provides some hints for the two - loop qed corrections . applying the asymptotic expansion in the limit @xmath77 to the four three - loop diagrams contributing to the @xmath78 correction leads to the following result for the first nine expansion terms @xmath79 .[tab : oal]@xmath34-pad results for the @xmath78 corrections for different choices of @xmath71 . [ cols="<,<,^,^,^,^,^,^,^",options="header " , ] in order to get predictions for @xmath55 we again add the moments in a first step and perform the pad approximations afterwards . from the results listed in tab .
[ tab : na_2 ] we deduce @xmath80 this number is in good agreement with the one stated in @xcite .
the error is quite small and amounts only to roughly 3% .
the semileptonic decay rate of the bottom quark finally reads @xmath81 \,.\end{aligned}\ ] ]
in this paper the two - loop qed corrections to the decay rate of the muon have been evaluated .
a new method has been used in order to confirm via an independent calculation the result of ref .
@xcite . from the muon decay the fermi coupling constant , @xmath1 , is determined which constitutes one of the basic input parameters of the standard model .
thus it is very important to have independent checks on such highly non - trivial computations .
the inclusion of the new correction terms removes the theoretical error on @xmath1 .
after the additional computation of the non - abelian diagrams the qcd corrections to the semileptonic decay rate of the bottom quark @xmath82 are obtained . again agreement with the literature @xcite is found .
we would like to thank k.g .
chetyrkin and j.h .
khn for valuable discussions and careful reading of the manuscript .
useful discussions with t. van ritbergen are gratefully acknowledged .
this work was supported by the _ graduiertenkolleg `` elementarteilchenphysik an beschleunigern '' _ , the _ dfg - forschergruppe `` quantenfeldtheorie , computeralgebra und monte - carlo - simulationen '' _ under contract ku 502/8 - 1 and the _
schweizer nationalfonds_. # 1#2#3_act .
phys . pol .
_ * b # 1 * ( # 2 ) # 3 # 1#2#3_act
. phys .
austr . _ * # 1 * ( # 2 ) # 3 # 1#2#3_comm .
phys . _ * # 1 * ( # 2 ) # 3 # 1#2#3_comp .
commun . _ * # 1 * ( # 2 ) # 3 # 1#2#3_eur .
j. _ * c # 1 * ( # 2 ) # 3 # 1#2#3_fortschr .
phys . _ * # 1 * ( # 2 ) # 3 # 1#2#3_int .
j. mod .
_ * c # 1 * ( # 2 ) # 3 # 1#2#3_int .
j. mod .
phys . _ * a # 1 * ( # 2 ) # 3 # 1#2#3_j . comp .
_ * # 1 * ( # 2 ) # 3 # 1#2#3_jetp lett . _ * # 1 * ( # 2 ) # 3 # 1#2#3_mod . phys
. lett . _ * a # 1 * ( # 2 ) # 3 # 1#2#3_nucl .
_ * a # 1 * ( # 2 ) # 3 # 1#2#3_nucl .
* b # 1 * ( # 2 ) # 3 # 1#2#3_nuovo cim .
_ * # 1a * ( # 2 ) # 3 # 1#2#3_phys .
* b # 1 * ( # 2 ) # 3 # 1#2#3_phys
. reports _ * # 1 * ( # 2 ) # 3 # 1#2#3_phys . rev . _ * d # 1 * ( # 2 ) # 3 # 1#2#3_phys .
rev . _ * # 1 * ( # 2 ) # 3 # 1#2#3_phys .
lett . _ * # 1 * ( # 2 ) # 3 # 1#2#3_phys .
reports _ * # 1 * ( # 2 ) # 3 # 1#2#3_prog . theor .
phys . _ * # 1 * ( # 2 ) # 3 # 1#2#3_sov .
j. nucl .
_ * # 1 * ( # 2 ) # 3 # 1#2#3_teor . mat .
fiz . _ * # 1 * ( # 2 ) # 3 # 1#2#3_yad .
fiz . _ * # 1 * ( # 2 ) # 3 # 1#2#3_z .
phys . _ * c # 1 * ( # 2 ) # 3 # 1#2#3_prog . part .
_ * # 1 * ( # 2 ) # 3 # 1#2#3ibid . * # 1 * ( # 2 ) # 3 | in this article two - loop qed corrections to the muon decay and corrections of order @xmath0 to the semileptonic decay of the bottom quark are considered .
we compute the imaginary part of the four - loop diagrams contributing to the corresponding fermion propagator in the limit of small external momentum .
the on - shell condition is obtained with the help of a conformal mapping and pad approximation . via this method
we confirm the existing results by an independent calculation . |
high - mass stars ( m@xmath14m@xmath9 ) are predominantly observed in young ( @xmath15myr ) , massive clusters ( m@xmath16m@xmath9 ) located within giant molecular clouds ( gmcs ) . despite their rarity and short life times , high - mass stars are responsible for injecting a significant amount of energy into the interstellar medium ( ism ) by means of stellar winds , ionising radiation and supernovae .
the impact of such feedback mechanisms from high - mass stars on the ism and star formation is still a topic of debate but is thought to play an important role in driving the evolution of gmcs .
feedback can be both constructive and destructive by sweeping up and clearing out parsec - scale cavities ( or bubbles ) in the ism @xcite resulting in dense gas shells or destroying and dispersing the molecular environment .
this means that feedback from high - mass stars should have the potential to both inhibit or enhance subsequent star formation and profoundly influence the star formation history ( sfh ) of gmcs @xcite .
the g305 gmc complex is located within the scutum - crux arm of the milky way at @xmath17 , @xmath18 at a distance of @xmath19kpc and is one of the most massive and luminous star - forming regions in the galaxy @xcite .
the g305 complex consists of a number of molecular clouds that surround a cavity centered on the open clusters danks1 and 2 ( fig .
[ im : obsmopra ] , triangles ) and the wolf rayet star wr48a . the feedback from the high - mass stars in and around danks1 and 2 appears to be responsible for driving the expansion of a large ( @xmath20pc ) cavity into the surrounding molecular gas .
previous observations of g305 , summarised by @xcite , have revealed the presence of ongoing star formation surrounding the central cavity ( see fig.2 in @xcite ) , which suggests that the overall morphology of the complex is consistent with numerous sites of ongoing high - mass star formation that may have been triggered by interaction between the winds and ionising radiation of danks1 and 2 and the natal molecular clouds .
these conditions make g305 an excellent laboratory in which to study high - mass star formation and the impact of feedback from high - mass stars on the evolution of gmcs . in this work we present single dish observations of @xmath0co , @xmath1co and c@xmath2o in the rotational transition to explore the global distribution and properties of molecular gas in g305 .
co is the most abundant molecule in the ism , after h@xmath3 , and the low - energy rotational transition is an excellent tracer of the cool ( @xmath21k ) low density gas ( @xmath22@xmath23 ) within molecular clouds @xcite . to explore the ionised environment we present wide - area interferometric observations of the large - scale 5.5 and 8.8ghz radio continuum which is generated by the thermal bremsstrahlung mechanism in the ionised gas that surrounds high - mass stars .
previous observations of the radio continuum emission towards g305 have been limited to low resolution ( 45 ) single dish observations @xcite that are unable to resolve the radio substructure on the scale of compact hii regions .
higher resolution ( 1 - 20 ) interferometric radio continuum observations exist but suffer from low sensitivity and significant spatial filtering @xcite and so are unable to identify extended features .
the data set presented in the following section allows the investigation of the molecular environment in which high - mass star formation is ongoing and allows us to trace the distribution and investigate the impact of ionising radiation from high - mass stars . by searching the literature for tracers of ongoing star formation ,
we are able to investigate the relationship between the ionised and molecular environment with star formation and thus broadly trace the sfh of the region .
this paper is organised as follows ; in section [ sect : obs ] we present the observations , data reduction and ancillary data utilised in this study , section [ sect : results ] presents the results and analysis of the molecular and ionised observations , in section [ sect : discussion ] we present our discussion of the results including the relationship between the ionised and molecular gas and star formation , the role of triggering and sfh of the region .
we end this paper with a summary in section [ sect : summary ] .
co observations of the g305 complex were obtained during two observing runs in september of 2010 and 2011 using the 3 mm receiver on the mopra 22-m millimetre - wave telescope operated by the australia telescope national facility ( atnf ) .
the back end of mopra is the university of new south wales mopra spectrometer ( mops ) which consists of four 2.2ghz bands that overlap to provide 8ghz of continuous bandwidth .
the narrow or `` zoom '' mode of mops was used which allows 16 intermediate frequencies ( ifs ) to be observed simultaneously .
each of the 16 zoom `` windows '' has a bandwidth of 137.5mhz over 4096 channels giving a velocity resolution at 110ghz of @xmath24kms@xmath13 per channel .
the velocities presented in this work are in the kinematic local standard of rest ( lsr ) frame centred on @xmath25kms@xmath13 , the systematic velocity of g305 @xcite .
we centred the 8ghz band at 112ghz and placed the zoom windows of mops to cover three co isotopologues ; @xmath0co , @xmath1co and c@xmath2o in the rotational transition ( table [ tab : mopslines2 ] ) . the remaining zoom windows were set to the rest frequency of lines of interest such as hc@xmath26n ( ) , ch@xmath26cn ( ) and c@xmath27o ( ) but no emission @xmath28 , where @xmath29 is the rms background noise , was detected . | we present 109115ghz ( 3 mm ) wide - field spectral line observations of @xmath0co , @xmath1co and c@xmath2o molecular emission and 5.5 and 8.8ghz ( 6 and 3 cm ) radio continuum emission towards the high - mass star forming complex known as g305 .
the morphology of g305 is dominated by a large evacuated cavity at the centre of the complex driven by clusters of o stars surrounded by molecular gas .
our goals are to determine the physical properties of the molecular environment and reveal the relationship between the molecular and ionised gas and star formation in g305 .
this is in an effort to characterise the star - forming environment and constrain the star formation history in an attempt to evaluate the impact of high - mass stars on the evolution of the g305 complex .
analysis of co emission in g305 reveals 156 molecular clumps with the following physical characteristics ; excitation temperatures ranging from 725k , optical depths of 0.20.9 , h@xmath3 column densities of @xmath4@xmath5@xmath6 , clump masses ranging from @xmath7@xmath8m@xmath9 and a total molecular mass of @xmath10m@xmath9 .
the 5.5 and 8.8ghz radio continuum emission reveals an extended low surface brightness ionised environment within which we identify 15 large - scale features with a further eight smaller sources projected within these features . by comparing to mid infrared emission and archival data , we identify nine hii regions , seven compact hii regions , one region , four extended regions .
the total integrated flux of the radio continuum emission at 5.5ghz is @xmath11jy corresponding to a lyman continuum output of @xmath12photonss@xmath13 .
we compare the ionised and molecular environment with optically identified high - mass stars and ongoing star formation , identified from the literature .
analysis of this dataset reveals a star formation rate of 0.0080.016m@xmath9yr@xmath13 and efficiency of 712% , allows us to probe the star formation history of the region and discuss the impact of high - mass stars on the evolution of g305 .
[ firstpage ] radio hii stars : formation ism : clouds |
tidal interactions between galaxies are suggested , both from observations and theory , to dramatically change the formation rates of field stars and globular clusters due to the tidal compression of gas clouds and their efficient conversion into stars ( kennicutt 1998 ; ashman & zepf 1992 ; noguchi & ishibashi 1986 ; bekki & couch 2001 ; bekki et al .
the large and small magellanic clouds ( lmc and smc ) , which are the nearest pair of interacting galaxies in our vicinity , have served as unique laboratories for studying the interplay between galactic dynamical evolution and star formation activity in galaxies ( westerlund 1997 ; van den bergh 2000 ) .
a major curiosity in this context is that globular cluster formation which is considered to be a special mode of star formation ( harris & pudritz 1994 ; elmegreen & efremov 1997 ) is observed to be ongoing in the clouds but not in the galaxy .
clearly the physical conditions required for cluster formation currently exist in the clouds but not in the galaxy ( westerlund 1997 ; van den bergh 2000 ) .
moreover , the age distribution of the lmc clusters shows a gap extending from 3 to 13 gyr with only one cluster in this age range suggesting a second epoch of cluster formation started abruptly in the lmc about 3 gyr ago ( e.g. , da costa 1991 ; geisler et al .
1997 ; rich et al . 2001 ; piatti et al . 2002 ) .
the origin of the `` age gap '' , which has not been observed for the smc , has remained elusive so far ( e.g. , da costa 1991 ) .
the purpose of this letter is to provide a new explanation for the origin of the mysterious age gap of the lmc s globular cluster system .
the present study also tries to explain the latest observational results of the lmc s structural and kinematical properties ( e.g. , alves & nelson 2000 ; van der marel et al .
we present here a new numerical model of the dynamical evolution and the star formation history of the lmc as it orbits the galaxy and is gravitationally influenced both by the galaxy and the smc , in an attempt to explain the origin of the age gap .
we first determine the most plausible and realistic orbits of the clouds and then investigate the evolution of the lmc using fully self - consistent n - body models . in determining the orbits , we adopt the same numerical method as those in previous studies ( murai & fujimoto 1980 ; gardiner et al .
1994 ; gardiner & noguchi 1996 ) , in which the equations of motion of the clouds are integrated backwards in time , from the present epoch until @xmath0 9 gyrs ago .
figure 1 shows the best orbit model for the clouds calculated in this way , where it can be seen they experience a very close tidal encounter ( at a pericenter distance of 6.4 kpc ) around 3.6 gyr ago ( t@xmath1 ) and then become dynamically coupled . if dynamical friction between the clouds and the mass of the lmc ( @xmath2 @xmath3 ) are considered something which has not been done in previous studies ( e.g. , gardiner et al .
1994 ) , then the present - day binary nature of the clouds can not have existed for a hubble time : they must have been separate objects in the past .
using gas - dynamical n - body simulations ( bekki et al . 2004 ) , we investigate dynamical evolution and star formation histories of the lmc for the best orbit model .
the lmc is modeled here as a bulgeless , gas - rich disk embedded in a massive dark matter halo with a mass 2.3 times larger than the disk mass , whereas the smc is treated as a point mass with the mass of @xmath4 @xmath3 .
the exponential old disk of the lmc is taken to have a scale length of 2.6kpc ( kinman et al .
1991 ) and a maximum circular velocity ( @xmath5 ) of 71kms@xmath6 , consistent with recent observations ( van der marel et al .
the gas disk is represented by a collection of discrete gas clouds that follow the observed mass - size relationship ( larson 1981 ) .
each pair of two overlapping gas clouds collides with the same restitution coefficient of 0.5 ( hausman & roberts 1984 ) .
globular cluster formation in the present model is discriminated from field star formation as follows : the gas is converted into field stars according to the schmidt law with the observed threshold gas density ( kennicutt 1998 ) .
we use the cluster formation criteria derived by previous analytical works ( e.g. , kumai et al .
1993 ) and hydrodynamical simulations with variously different parameters of cloud - cloud collisions on a 1 - 100pc scale ( bekki et al .
2004 ) in order to model globular cluster formation .
a gas particle is converted into a cluster if it collides with other high velocity gas ( ranging from 30kms@xmath6 to 100kms@xmath6 ) and having an impact parameter ( normalized to the cloud radius ) less than 0.25 .
this model is strongly supported by recent observations ( e.g. , zhang et al .
2001 ) that have revealed that there is a tendency for young clusters to be found in gaseous regions with higher velocity dispersion , where cloud - cloud collisions are highly likely .
chemical evolution of the gas with an assumed effective chemical yield of 0.005 is also incorporated in the simulations .
our numerical results on field star formation histories in the lmc do not depend strongly on the above parameters of gas dynamics ( e.g. , gas mass fraction ) and star formation ( the exponent of the schmidt law ) for a reasonable set of parameters . in the present paper ,
we show only the results of the best model in which the clouds first came together about 4 gyr ago . it should be here stressed that this conjunction is not a unique characteristic of the best model : about 95 % of the orbital models have a disintegration of the lmc - smc binary recall the integrations are done backwards in time occurring within the last 4 gyr for models with the orbital parameters consistent with observations .
we have investigated orbital evolution of the clouds for @xmath7 models with the current space velocities ( @xmath8,@xmath9,@xmath10 ) ranging from @xmath11 km s@xmath6 to @xmath12 km s@xmath6 for the clouds and thereby derived the epoch when the clouds become separated . we have confirmed that ( 1 ) it is highly unlikely that the clouds were a binary system more than 6 gyr ago , and ( 2 ) they most likely came together around 4 gyr ago .
thus our main conclusions on the epoch of the commencement of cluster formation in the lmc do not depend so strongly on orbital parameters .
as seen in figure 2 , the combined tidal effects of the smc and the galaxy distorts the lmc disk ( t=@xmath13 and @xmath14 gyr ) to form compact stellar and gaseous bars in its center after a few pericenter passages ( t=@xmath14gyr ) .
consequently , young stars are formed repeatedly in the tidally compressed high - density gaseous regions , in particular the central bar , where the star formation rate is at a maximum at 0.38 @xmath3 yr@xmath6 .
the tidal interaction significantly increase the degree of random motion of the gas clouds so that they collide much more frequently with one another .
however , the cloud - cloud collisions with moderately high speed ( between 30 and 100 km s@xmath6 ) and small impact parameter ( @xmath15 0.25 ) , required for cluster formation , do not occur until the lmc begins to interact violently with the smc when the two are less than 10kpc apart ( t=@xmath16gyr ) . the dynamical evolution over the last @xmath0 3gyr results in about 0.5% of the initial gas cloud mass being converted into clusters .
the new globular cluster system has a flattened , ` disky ' spatial distribution with nearly all clusters within the central @xmath0 3 kpc and confined within @xmath0 1 kpc from the lmc disk plane ( i.e. , an indication of a structure similar to the galactic thick disk ) .
they are supported mainly by rotation with the central velocity dispersion of only 20 km s@xmath6 and with a fairly asymmetric radial dispersion profile caused by recent tidal interaction with the galaxy and the smc .
the derived structure and kinematics of clusters are broadly consistent with the observed ones for intermediate - age clusters ( schommer et al .
1992 ) . due to the rapid chemical enrichment associated with this star formation activity
, the mean metallicity of clusters increase from [ fe / h ] = @xmath17 to @xmath18 over this period .
such a recently formed cluster population consequently has a narrower age distribution than the new field stars ( figure 3 ) .
thus the origin of the age gap and in particular the trigger for the most recent episode of globular cluster formation in the lmc may well be related to the commencement of strong tidal interactions between the lmc , the smc and the galaxy .
the tidal interaction between the three may also play an important role in changing the structural and kinematical properties of the lmc .
because of the strong tidal perturbation , the initial thin , non - barred disk is dynamically heated to form a thick , barred disk with an outer stellar warp .
furthermore , @xmath19% of the stars in the disk are spatially redistributed to form an outer stellar halo with a velocity dispersion of @xmath20kms@xmath6 at a distance of 7.5kpc from the lmc center .
the kinematically hot stellar halo is dominated by old stars with the fraction of new field stars equal to only 2% , because most halo stars originate from the outer part of the initial thin disk .
in contrast , @xmath21% of the stars within the central disk are new field stars formed from the triggered star formation .
for example , our simulations indicate that the half mass radius is @xmath22kpc for the old field populations and @xmath23kpc for the new populations ( i.e. , the different structure of the different aged populations ) .
the stellar kinematics along the major axis shows the central velocity dispersion , @xmath24 , to be @xmath25kms@xmath6 with @xmath26 of @xmath0 2.3 .
these results are broadly consistent with observations by van der marel et al .
these also suggest that the lmc stellar halo has a significantly larger fraction of young , relatively metal - rich ( [ fe / h ] @xmath15 @xmath27 ) stellar populations compared with the galactic stellar halo dominated by very old , metal - poor ( [ fe / h ] @xmath0 @xmath28 ) stars .
the origin of the field stars formed in recent star formation activity in the lmc disk is one of the most important problems associated with the formation of the lmc ( e.g. , butcher 1997 ; olszewski et al .
1996 ; gallagher et al .
our evolutionary model strongly suggests that such recent star formation is due to the repetitive tidal interaction between the lmc / smc and the galaxy .
field star formation is more sensitive to tidal perturbation than cluster formation so that enhancement in field star formation can occur from the early evolutionary stage of the lmc ( i.e. , @xmath29 gyr ago ) .
this can clearly explain why `` the age gap '' can be more clearly observed in cluster populations than in field ones .
chemical enrichment associated with the tidally triggered star formation provides a natural explanation for the observed rapid increase of metallicity both in clusters and in field stars ( da costa 1991 ; dopita et al .
1997 ) . what then are the fossil records of any past dynamical interaction ?
the tidal radius of the lmc , within which its stars are gravitationally bound , is linearly proportional to the pericenter of the orbit for the fixed masses of the lmc and the galaxy ( gardiner & noguchi 1996 ) .
therefore , as the pericenter of the lmc orbit decreases as a result of dynamical friction , it loses an increasingly larger number of stars .
figure 4 shows that the stripped stars in the simulation form a great circle - a relic stream in which the stars are diffusely and inhomogeneously distributed along the lmc s orbit in the galactic halo region . such a stream has been predicted in previous studies ( weinberg 2000 ) , but for the first time our simulations provide quite clear predictions on the spatial distribution and radial velocities of stars within the stream .
ongoing and future surveys for stellar substructure in the galactic halo ( majewski et al .
2000 ) will allow these predictions to be tested .
indeed preliminary observational results have already suggested that the coherent radial velocity structures amongst giant stars in fields encircling the clouds are due to tidal debris that originated from the lmc ( majewski et al .
how then can the model explain the observed different age distributions of clusters between the lmc and the smc ( piatti et al .
we propose that such a difference can be understood in terms of the differences in the birthplaces and initial masses between the two .
the lmc was formed as a relatively low surface brightness galaxy , being more distant ( @xmath30kpc , corresponding to the apocenter of its early orbit ) from the galaxy so that the galactic tidal field alone could not trigger cluster formation efficiently until it first encounters with the smc . in contrast
, the less massive smc , which is therefore more susceptible to the galactic tide , was born less distant ( @xmath31kpc ) from the galaxy , and thus influenced by the galaxy strongly enough to form globular clusters from the early evolutionary stage ( several to 10gyrs ago ) .
thus the difference in cluster formation histories could reflect the fact that the lmc / smc were formed as different entities rather than as a binary protogalaxy .
tidal interaction between the clouds and the galaxy has long been considered to be a physical mechanism responsible for the formation of the magellanic stream and the evolution of the smc ( murai & fujimoto 1980 ; lin & lynden - bell 1982 ) .
we have demonstrated that such interaction can also cause dramatic changes not only in the formation history of field stars and globular clusters but also in the lmc s structure and kinematics .
our model predicts that if globular cluster formation is associated with a dramatic increase in the random motions in the lmc s interstellar medium , then the age- , metallicity- and spatial - distributions of these systems will be very much dependent on the orbital evolution of the clouds .
this can be inferred observationally from radial velocity and proper motion measurements of their stars .
future proper motion measurements with @xmath32micro - arcsec accuracy ( perryman et al .
2001 ) , coupled with numerical simulations for different yet reasonable orbits of the clouds , will therefore enable us to obtain an integrated and systematic understanding of the formation of the lmc clusters .
we are grateful to the anonymous referee for valuable comments , which contribute to improve the present paper . k.b .
, w.j.c . , and
g. da c. acknowledge financial support from the australian research council ( arc ) throughout the course of this work .
the numerical simulations reported here were carried out on grape systems kindly made available by the astronomical data analysis center ( adac ) at national astronomical observatory of japan ( naoj ) . | the large magellanic cloud ( lmc ) has a unique cluster formation history in that nearly all of its globular clusters were formed either @xmath0 13 gyr ago or less than @xmath0 3 gyr ago .
it is not clear what physical mechanism is responsible for the most recent cluster formation episode and thus the mysterious age gap between the lmc clusters .
we first present results of gas dynamical n - body simulations of the evolution of the lmc in the context of its galactic orbit and interactions with the smc , paying special attention to the effect of tidal forces .
we find that the first close encounter between the lmc and the small magellanic cloud ( smc ) about 4 gyr ago was the beginning of a period of strong tidal interaction which likely induced dramatic gas cloud collisions , leading to an enhancement of the formation of globular clusters which has been sustained by strong tidal interactions to the present day .
the tidal interaction results in the formation of a barred , elliptical , thick disk in the lmc .
the model also predicts the presence of a large , diffuse stellar stream circling the galaxy , which originated from the lmc . |
type ia supernovae ( sne ia ) are mature standardised candles .
intrinsic dispersion of the peak absolute magnitude can be minimised to the level of @xmath6 magnitude or even smaller , and can be accurately used as a cosmological distance indicator @xcite .
however , this standardisation generally requires a non - standard extinction law , i.e. , @xmath7 , as opposed to the typical galactic value of @xmath8 ( e.g. , * ? ? ?
this indicates that the properties of dust in sn ia host galaxies may be systematically different from those in our galaxy ( e.g. , * ? ? ?
alternatively , it has been suggested that multiple _
scatterings _ of sn photons on circumstellar ( cs ) dust may explain the non - standard nature of the reddening law @xcite . despite the implications that have been discussed ( e.g. , * ? ? ?
* ) , researchers have yet to find a smoking gun to test this hypothesis , apart from the various model predictions for optical wavelengths @xcite . this interpretation is also associated with as yet unclarified progenitor evolution , through the dense circumstellar medium ( csm ) and high mass loss rate that would be required .
qualitatively , the scenario favours a so - called single - degenerate model ( i.e. , a pair consisting of a c+o white dwarf and a non - degenerate star ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , which should create a relatively dense ( or ` dirty ' ) cs environment . in this paper , we examine the effects of _ re - emission _ of sn photons for infrared ( ir ) wavelengths . while the _ ` scattering ' _ must be accompanied by _ `
absorption and re - emission ' _ ( a cs dust echo ) , this process has not been examined with respect to the cs environment around sne ia ( with the exception of recent work that focused on the mid- and far - ir ; * ? ? ?
the remainder of this paper is organised as follows . in 2 , we develop a simple echo model for cs dust heated by sn photons . in 3 ,
we apply the method to observed nir sn ia samples .
we place both conservative and deeper limits for the amount of cs dust under different assumptions . in 4 , we discuss the effects of different types of cs dust species , as well as future observational strategies for a better constraint on the amount of cs dust and the environment .
the paper is closed in 5 with conclusions .
in this section , we present our model of cs dust echoes .
we restrict the model to absorption and re - emission processes , not including scattering .
this is justified because , ( 1 ) given that the spectral peak of the sn light lies at optical wavelengths , the contribution of the scattered sn light at nir - ir wavelengths is very small , compared to that of thermal emission from cs dust , and ( 2 ) in the distribution of cs dust of interest in this paper , further scattering of the re - emitted nir / ir photons is negligible . while computational methods for an echo from spherically distributed cs dust ( e.g. , * ? ? ?
* ; * ? ? ?
* ) and further from aspherically distributed cs dust ( especially for modelling the light echo from sn 1987a ; * ? ? ?
* ; * ? ? ?
* ) have been developed , we assume for the sake of simplicity that the distribution of cs dust is represented by an infinitely thin shell located at radius @xmath9 from the centre of an explosion .
the thin - shell approximation further simplifies the computation as shown below .
furthermore , in providing an upper limit for the amount of cs dust _ as a function of _ @xmath9 , the thin - shell approximation basically provides the most conservative limit , as any spherically symmetric distribution can be described as a convolution of different shells .
namely , if one adds another shell at @xmath10 , that is different from @xmath9 for which the constraint is considered , then this additional shell will increase the predicted flux without contributing to the cs dust mass at @xmath9 ; therefore , the upper limit on this cs dust mass should be reduced ( i.e. , the predicted flux must be below observations ) .
let us first consider a situation in which we observe a dust shell with temperature @xmath11 and mass @xmath12 , where @xmath13 is the angle between the observer s line of sight and a direction vector pointing from the shell centre ( i.e. , the explosion centre ) to the position of the volume element within the shell under consideration .
then , the luminosity of the thermal emission from dust is given as follows : @xmath14 here , @xmath15 is the absorbing opacity of the dust at frequency @xmath16 .
it is specified by the dust properties , which are assumed to be uniform within the shell .
furthermore , the shell is assumed to have uniform density ( @xmath17 ) .
the above expression can be converted into time - dependent luminosity by introducing @xmath18 ( the time since the explosion , measured in the observer s frame ) and @xmath19 ( observer - frame time when the light was emitted from a given volume element at @xmath13 as observed at time @xmath18 ) .
the relationship between @xmath18 and @xmath19 is given by the following : @xmath20 combining eqs .
( 1 ) and ( 2 ) , we obtain the following expression for the echo luminosity at time @xmath18 : @xmath21 the temperature evolution of the cs dust at time @xmath18 , under radiative equilibrium with an incoming sn flux , @xmath22 , is given by @xmath23 this equation provides @xmath24 once the sn flux evolution is specified . under the condition considered in this paper ,
the photon - dust collision time scale is shorter than the cooling time scale of the dust by thermal radiation .
therefore , stochastic heating is negligible , and we can assume that virtually all of the dust particles have a temperature determined by radiative equilibrium .
we first compute @xmath24 for a given @xmath9 using the above equation . for the sn flux
, we use the hsiao template spectra as an input @xcite . for the dust opacity
, we use the lmc - like dust properties as our fiducial model ; specifically , we adopt the opacity from @xmath25 to @xmath26 from @xcite so as to be consistent with the proposed cs scattering model .
we extend the opacity toward the blue ( beyond @xmath25 ) and red ( beyond @xmath27 ) , assuming slopes of @xmath28 and @xmath29 , respectively .
then the evolution of @xmath24 is used as an input in the computations of echo luminosity and multi - band light curves .
because the incoming sn light peaks in optical wavelengths , and because the dust temperature under the situation examined in this paper is determined by the opacity in the nir wavelength , the treatment of the opacity below the @xmath25-band and above the @xmath26-band does not substantially affect our results . to check this , we performed two additional test calculations : one having a slope of @xmath30 below the @xmath25-band and the other having a slope of @xmath31 above the @xmath27-band for our reference lmc dust model ; these models are unrealistically steep and flat , respectively , and thus provide a conservative measure of the uncertainty for the treatment of different slopes .
the difference between the reference model and the one using the different slopes ( @xmath32 , @xmath33 , @xmath34 ) is larger for a model with a larger @xmath9 .
for the typical maximum distance for which a meaningful constraint is placed at different bands ( see 3 ) , the difference in the echo magnitude is given by @xmath35 mag ( for 0.03 pc ) , @xmath36 mag ( for 0.3 pc ) , and @xmath37 mag ( 1 pc ) for the different slope below the @xmath25-band .
for the different slope above the @xmath26-band , @xmath38 mag ( for 0.03 pc ) , @xmath39 mag ( for 0.3 pc ) , and @xmath40 mag ( for 1 pc ) .
therefore , in the worst - case scenario , our echo luminosity constraint ( and , thus , the dust mass ) could have an uncertainty factor of 5 ; however , in most cases the uncertainty would be much smaller than this .
a further discussion of the dust properties is given in 4.1 .
the resulting echo luminosity , as computed above , is for unobscured luminosity .
we now consider the optical depth within the shell .
the average optical depth is computed as @xmath41 ; then the nir echo luminosity is dimmed by this amount of absorption .
the same procedure is adopted for the model with a different geometry ( see below ) .
the csm around an sn ia may be confined within the equatorial plane ( e.g. , * ? ? ?
* ) , therefore we also consider a torus / ring - like structure , again under the infinitely thin assumption in the radial direction .
we denote @xmath42 as the opening angle of the ring , as measured from the equatorial direction .
while it is possible to derive general expressions for an arbitrary viewing direction , hereafter , we focus on two extreme cases : pole - on and edge - on viewing . for an observer sitting at the pole - on position ,
the echo luminosity is expressed as follows : @xmath43 if @xmath44 , then the integral is set to zero . to be an explanation for the non - standard extinction law requires @xmath45 cm , to introduce the effect of scattered photons around the maximum light
an observer in the equatorial direction does not have this constraint .
we note that both situations should be considered simultaneously , because there should be roughly an equal number of edge - on and pole - on counterparts ( depending on @xmath42 ) , if we adopt the ring - like geometry . finally , the echo luminosity arising from the same configuration , but viewed edge - on , is given as follows .
@xmath46 \ .\end{aligned}\ ] ] another interesting possibility for the cs dust distribution is a bipolar morphology , i.e. , a pair of blobs / jets , as inferred for some novae ( e.g. , * ? ? ?
therefore , we also consider a pair of bipolar blobs for the distribution of the cs dust , again under the infinitely thin assumption in the radial direction .
we denote @xmath42 as the opening angle subtended by each blob , as measured from the polar direction . for an observer sitting at the pole - on position ,
the echo luminosity is expressed as follows : @xmath47 \ .\end{aligned}\ ] ] the echo luminosity arising from the same configuration , but viewed edge - on , is given as @xmath48 similar to the case of the pole - on observer for the ring - like cs dust distribution , a constraint is placed on the cs dust radius if the multiple scattering model is to account for the non - standard extinction law . for the edge - on observer of the bipolar blob geometry , the cs dust radius must satisfy @xmath45 cm to introduce the scattered photon effect around the maximum light .
this constraint does not apply to the pole - on observer ; however , both situations ( pole - on and edge - on ) should be considered simultaneously , following the same argument given for the ring - like geometry .
figure 1 shows the evolution of @xmath49 and the predicted nir ( @xmath50 , @xmath51 , @xmath26 ) light curves for different values of @xmath9 and for a fixed value of @xmath12 ( @xmath52 ) ; for the light curves , only the shell model is shown .
the evolution of @xmath49 closely follows the optical light curve of sn ia . for @xmath53 pc
, the peak temperature exceeds @xmath54k ; therefore , the dust may well have evaporated , depending on the nature of the cs dust . in such a case , essentially there should be neither scattering nor thermal emission from the cs dust .
therefore , no constraint can be obtained regarding the amount of pre - existing cs dust , _ and _ the cs dust can not be an origin of the non - standard extinction law toward sne ia .
the effects of different geometries are demonstrated in fig .
2 ; light curves are shown for the shell csm , pole - on and edge - on views for a ring - like csm with @xmath55 , and pole - on and edge - on views for bipolar blobs with @xmath56 . throughout this paper , we consider these geometry types for the cs dust distribution , while adopting the shell model as our fiducial model .
the light curve evolution is similar in various nir bands ; its time scale is basically determined by the travelling time of light , and thus the cs dust at larger @xmath9 results in a longer - lived , fainter nir echo . for small @xmath9 ,
the travelling time of light is shorter than the time scale of the sn light curve evolution ; thus in this case , the light curve reflects the sn light curve evolution , and the echo light curve evolution traces the decreasing temperature of the cs dust as a function of time . on the other hand , for large @xmath9 , the travelling time of light far exceeds the sn light curve evolution time scale ; therefore , the temperature evolution has little influence on the shape of the light curve .
emission at longer wavelengths is stronger for larger @xmath9 , which stems from smaller @xmath49 for larger @xmath9 . as such , the peak wavelength in the thermal emission is shorter than the @xmath50-band for smaller @xmath9 , while it is even longer than the @xmath26-band pass for larger @xmath9 .
figure 3 shows examples of the constraint on the amount of cs dust using the shell model . without the dust echo , nir luminosities of sne
are powered by radioactive decay of @xmath57co and subsequent thermalisation ; the nir spectral energy distribution ( sed ) roughly follows a blackbody of @xmath58 k in the early photospheric phase ( @xmath59 days ) , while the sed reflects the forbidden lines of fe - peak elements in late nebular phases ( @xmath60 days ) . in any case , a conservative upper limit on the cs dust mass is obtained by the condition that the resulting echo luminosity can not exceed the observation at any epochs . as the echo luminosity is proportional to @xmath12 in the optically thin regime , for a given @xmath9 , one can derive a maximally allowed value for @xmath12 . note that this is not a fit to the observed light curve
the echo light curve model for small @xmath9 resembles the observed light curve to some extent , but this is because the echo evolution follows the sn evolution for small @xmath9 ( see above ) .
the left panel of fig .
3 shows how an early - phase light curve ( up to @xmath61 days since the @xmath62-band maximum ) can be used to constrain the amount of cs dust .
generally , the early phase data provide a strong constraint on @xmath12 for small @xmath9 , but lose diagnostic power for cs dust at large @xmath9 .
the model prediction for small @xmath9 is similar for different bands ; typically the @xmath50-band provides the deepest limit , thanks to its extensive coverage within the nir .
we note that the @xmath63-band , which typically has much better coverage than the @xmath50-band , is not so useful for this purpose , as the wavelength of the thermal emission peak is longer than that of the @xmath63-band .
the right panel of fig .
3 shows the other extreme within the nir wavelengths , namely a constraint obtained through late - time photometry in the @xmath26-band . such an observation can provide a constraint on the cs dust at large @xmath9 , up to @xmath64 pc . by applying different models ( i.e. , with a different @xmath9 and geometry ) to the available , observed , multi - band light curves , a constraint can be obtained for @xmath12 as a function of @xmath9 for each sn
the light curves for different bands provide mutually independent upper limits ; we adopted the strongest limit among @xmath50 , @xmath51 , and @xmath26 . the result is shown in fig .
4 , where the method was applied to a ` typical nir observation ' using the csp sample , sn 2006x , as an example of very intensively observed early - phase light curves , including the @xmath26-band and extending to a relatively late epoch ; also shown are sne 1998bu , 2000cx , 2001el , 2003du , and 2003hv , for which nir late - phase ( @xmath64 year ) magnitudes are available .
it is seen that generally the strongest constraints are obtained through the @xmath50 or @xmath51-band data for the cs dust at @xmath65 pc , and through the @xmath26-band at @xmath66 pc .
figure 4 also contains lines indicating various physical situations .
the thick solid lines indicate the absorptive optical depth in the @xmath62-band , which gives indication of the likelihood that the cs dust is the origin of the non - standard extinction law ( i.e. , an optical depth of an order unity is required ) .
the thin solid lines show the equivalent mass loss rates of the progenitor systems ; we adopted the dust - to - gas ratio of 0.01 ( i.e. , @xmath67 , where @xmath68 is the velocity of the mass loss outflow ) .
vertical dashed lines show the temperature of the cs dust as a function of @xmath9 .
figure 4 shows that the upper limits for @xmath12 , as obtained in the argument using the nir echo ( with the shell model ) , are @xmath69 for the cs dust at @xmath70 pc , @xmath71 at @xmath72 pc , and @xmath73 at @xmath74 pc . beyond this radius ,
the results are especially sensitive to the quality of the observational data , as essentially one needs still - rare late - time photometry for the @xmath26-band .
the strongest constraints for @xmath66 pc are obtained for sne 2000cx , 2001el , and 2003hv , which are only three examples to date having the late time @xmath26-band photometry at @xmath64 year . for these sne , @xmath75 is derived for @xmath76 pc . at @xmath77 pc ,
even the nir bands are too blue ; thus , we can only provide a very loose constraint on the cs dust mass as @xmath78 , even for the best - case scenario ( sn 2003hv ) . to probe the cs dust beyond @xmath79 pc , mid- or far - ir observations
should be more effective @xcite .
the derived upper limits are sufficiently tight to constrain the origin of the non - standard extinction as a scattered echo , as proposed by @xcite and @xcite .
the resulting upper limit for the optical depth in the @xmath62-band is generally @xmath80 for the cs dust at @xmath65 pc . at @xmath66 pc , the constraint is weaker , but for a few sne , we place the constraint as @xmath81 at @xmath76 pc . above @xmath79 pc ,
the upper limit is consistent with @xmath82 .
these values should be compared to the requirement that the optical depth must be at an order of unity for the scattered echo to have a substantial effect on the extinction law . from this analysis
, we conclude that it is difficult for the scattered echo to be a general origin of the non - standard extinction law toward sne , as long as the cs dust is placed at @xmath65 pc . a possibility that the cs dust at @xmath77 pc creates the non - standard extinction law
is not rejected , although categorising it as ` cs ' dust is questionable ( see below ) .
a dust shell at @xmath83 pc is rejected for some sne ia when the late - time @xmath26-band data are available , but not generally constrained due to a lack of such data .
the above constraints are for the shell model .
one might guess that the constraint would be weakened if the cs dust is confined in a specific direction , e.g. , in an equatorial ring or in bipolar blobs , as the cs dust mass required to create @xmath82 should be reduced . while this is true ,
several additional effects must be considered .
first , the shell model does indeed provide the most conservative upper limit on the cs dust mass .
if one confines the same amount of cs dust in a specific direction , then the local density increases .
this generally results in larger luminosity at a specific epoch than in the corresponding shell model , and thus a tighter limit on @xmath12 is obtained using the observational data at the corresponding epoch ( fig .
2 ) . for example
, the edge - on ring model predicts a larger luminosity than the shell model with the same cs dust mass at the beginning and the end of the nir echo .
in contrast , the same model viewed pole - on has a shorter duration and a larger luminosity .
this large luminosity in the ring model is achieved when the emission at the ring - edge reaches the observer .
a similar argument applies to the blob model . as for the required cs dust mass within the scattering echo model for the extinction law , it is true that the required mass is reduced for an edge - on observer to the ring - like cs dust structure ( fig .
4 ) . however , for the pole - on case , this effect is compensated for by an increase in the predicted luminosity ( fig .
thus , the resulting upper limit for @xmath84 in the pole - on ring model is similar to that of the shell model ( fig .
4 ) . statistically , there must be edge - on and pole - on counterparts ; the nir echo constraint in this paper , using a number of sne , is sufficiently strong to reject this possibility . if one introduces a _ more confined _
configuration , then this increases the likelihood that it will not be viewed from that direction .
in addition , for a pole - on observer to have a substantial scattering effect , the time required for light to travel to the cs dust must be shorter than the rising time of sne ia ( i.e. , @xmath85 days ) , and therefore the cs dust beyond @xmath86 pc never affects the extinction law if viewed pole - on .
the above argument should be slightly modified for a very confined case , such as when considering the bipolar blob geometry .
similar to the ring - like geometry , the upper limits on the cs dust mass for a bipolar blob are comparable to or deeper than the corresponding shell model , depending on the viewing direction .
if this is converted into the optical depth , the derived upper limits are indeed not extremely strong to readily reject the multiple scattering model for the extinction law , i.e. , @xmath82 ( fig .
however , in this case , the amount of scattered photons should already be limited , due to a small opening angle subtended by the distribution of cs dust . in summary , introducing the confined cs geometry does not help interpret the non - standard extinction law as _ generally _ having come from cs scattered light .
sne suffering a substantial ` cs ' scattering echo effect must be limited to at most only a fraction of sn ia samples . for the cs dust to have originated from the mass loss of a progenitor ,
the cs gas mass should be @xmath87 a few @xmath88 , i.e. , @xmath89 a few @xmath90 .
this amount of cs dust can create a substantial scattering effect , only if @xmath91 pc , even if we use the conservative requirement @xmath92 ( or @xmath93 pc if we adopt @xmath82 ) .
we generally reject such a large amount of cs dust at @xmath91 pc ( see above ) . therefore ,
if the echo scenario is to be a general origin of the non - standard extinction law , then the dust must indeed be ism in origin , located at @xmath94 pc .
one may question what implications this mass loss rate may have for the progenitor scenario .
indeed , the cs scattering model requires a much larger amount of csm , compared to that predicted by typical progenitor scenarios .
our limit is @xmath95 yr@xmath1 for the mass - loss wind velocity of @xmath2 km s@xmath1 below 0.01 pc , or , in the final @xmath96 years before the explosion .
this is not as strong as the constraint from radio and x - rays ( e.g. , @xmath97 yr@xmath1 for the same velocity , for sn 2011fe : * ? ? ?
* ; * ? ? ?
however , we emphasise that radio and x - rays probe the materials inside @xmath98 pc ; our proposed method using the nir echo can extend the constraint by an order of magnitude in the spatial dimension ( or to the past before the explosion ) , making our limit independent and unique . from the present analysis
( see below for possible improvement ) , the only scenario that can be marginally rejected , for _ normal _ sne examined in this paper , is a symbiotic system .
however , we do note that sne ia resulting from this path may appear as an atypical sn ia @xcite .
thus far , our limit has been rather conservative , and is based on the condition that the echo luminosity can not exceed the observed one .
however , most of the light should have originated from a radioactive decay chain and subsequent thermalisation , as demonstrated by nir spectroscopy @xcite .
given this observational fact , we can refine our limit as follows .
first , attribute the _ template _ light curves in each band to the sn light without any contribution of the possible nir echo ; subtract the templates from the light curves from individual sne ; and then use the residual light curve ( here we take the absolute value if the subtraction results in a negative value ) to place an upper limit on the cs dust mass .
namely , we assume that the echo luminosity does not exceed the _ variation _ in the nir light curves , seen in different sne . as a demonstration , in fig .
3 ( left panel ) we show the case of sn 2006x , for which a good light curve is available .
the figure shows that the magnitude of this residual light curve is @xmath99 less than that of the original , highlighting the standard nature of the nir light curves . using this argument ,
much tighter constraints can be obtained ( as shown by the thin solid lines in fig .
3 ) . figure 5 shows the results for sn 2006x . because @xmath50- and @xmath51-band data are of the best quality , following the template closely , the upper limit for the cs dust mass , especially at @xmath65 pc , improved significantly . for small @xmath9 ,
the upper limit was reduced by two orders of magnitude , which reflects a residual light curve that was fainter than the original by @xmath100 magnitudes . with this method ,
all of the models with different geometric types result in @xmath101 ( @xmath102 for most situations ) , robustly rejecting the cs scattering model for the extinction law for this particular sn .
the upper limit for the mass loss rate is now @xmath103 yr@xmath1 for @xmath70 pc , much deeper than the conservative limit , as described earlier in this section .
in this paper , our analysis has been based on a particular dust model , i.e. , the lmc - type dust opacity . in this section ,
we examine the effects of different types of cs dust .
figure 6 shows the ` conservative ' limits on the cs dust mass derived for ( a ) astronomical silicate and ( b ) amorphous carbon or graphite , both for the shell model .
figure 7 shows the same but for the ` deeper limit ' for sn 2006x ( i.e. , figure 5 but for astronomical silicate or carbon ) .
the opacities are calculated using optical constants by @xcite and @xcite . to the first approximation ,
the behaviour can be understood by the difference in a typical opacity for different dust models , e.g. , the opacity in the @xmath62-band .
the opacity is roughly a factor of two lower for silicate than the reference lmc - like dust model ; thus , the upper limit for the cs dust mass is weakened by the same factor .
the opacity is roughly an order of magnitude higher for carbon than the lmc - like dust model , and therefore the upper limit on the cs dust mass is deeper for the carbon dust model by an order of magnitude than our fiducial model .
however , these different constraints on the cs dust mass would not substantially alter the constraints on the ability of the cs dust to create the non - standard extinction law .
the optical depth itself is scaled by the same factor ; therefore , the upper limits on @xmath84 would not change substantially .
our results should be compared to sne , in which a large amount of dust is inferred either by the detection of an optical ( scattered ) echo or other probes ( e.g. , narrow absorption systems ) .
for example , we have obtained a tight limit for the amount of cs dust for sn 2006x ( fig .
5 ) , while this sn was detected in a scattered echo @xcite ; its optical spectrum at @xmath64 year after the explosion showed a blue echo component , indicating @xmath82 . in contrast
, we obtained the constraint that @xmath104 , as long as the echo source is at @xmath105pc ( or even @xmath102 at @xmath65 pc ) .
taken together , the detected echo in sn 2006x must be created by ism , as suggested by @xcite using hubble space telescope ( _ hst _ ) data . to date , even for cases where an ( optical ) echo is detected , evidence of a cs echo remains weak .
another example is highly extinct sn ia 2014j in m82 .
no variability has been found in the narrow absorption lines ( e.g. , * ? ? ?
@xcite have claimed temporal evolution in @xmath106 , while no apparent variability has been detected in polarisation measurements @xcite .
while a consistent interpretation for all of these features has yet to be reached , it seems that together they suggest that the cs scattering scenario is unlikely , but perhaps a combination of multiple scattering and true absorption at the ism scale could provide a consistent solution . indeed , detection of an echo has been reported for sn 2014j through high - resolution _ hst _ images at @xmath107 days after the explosion @xcite .
this has been interpreted as echoes originating at the interstellar scale . at this epoch
, the sn still dominates the light in the vicinity of the sn and no strong signature of the echo from the cs dust has been obtained ( see also * ? ? ?
* ) . in fig .
8 , we show the results of simulated @xmath26-band images of an echo from the cs dust located at @xmath108 pc for different geometries , for an sn at @xmath109 mpc . for the ring and blob geometries ,
we have adopted @xmath110 and the viewing direction of @xmath111 for the purpose of demonstration .
the model images are convolved with a gaussian kernel of @xmath112 arcsec to roughly simulate observations by _ hst _ ; the narrow field channel of _ hst / nicmos _ has a spatial resolution of @xmath113 arcsec at @xmath114 .
thus our simulation could overestimate the spatial resolution up to @xmath115% , but it could anyway be compensated by increasing the assumed physical scale of the cs dust by the same factor . note that these images are for a ` pure ' echo component if the projected echo size is smaller than or comparable to the instrument s spatial resolution , then the sn must fade away below the level of the echo luminosity to examine the geometry of the echo signal .
note that there is a better chance to apply this exercise to later epochs .
figure 8 shows the expected evolution covering @xmath100 years , to investigate the possibility of distinguishing different cs dust geometries if a cs dust echo is detected . even for nearby sn 2014j and using _ hst _ , probing the cs dust echo using spatially resolved imagery can be difficult . for a ring - like structure ,
an elongated structure is visible when @xmath116 . for this particular bipolar blob model ,
the echo ` disappears ' when @xmath116 . for both models
, there is a ( non - monotonic ) shift in the centre of the echo light .
the size of this shift for these particular models is about half of the spatial resolution , and could be detectable with precision astrometry techniques
. these diagnostics could apply to sn 2014j if the cs dust echo is detected in the future following the sn luminosity decrease .
continuous _ hst _
high - resolution observations are strongly encouraged . while investigation of the cs echo through direct imaging observations is currently limited by the spatial resolution of @xmath5 arcsec at nir wavelengths even with
_ hst _ ( 4.2 ) , the situation is expected to improve dramatically in the coming decade .
the planned james webb space telescope ( _ jwst _ ) will reach to a spatial resolution by a factor of @xmath117 better than _
hst_. furthermore , with the availability of nir adaptive optics ( ao ) attached to @xmath118 m class telescopes such as the thirty meter telescope ( tmt ) , european extremely large telescope ( e - elt ) , and giant magellan telescope ( gmt ) , spatial resolutions of @xmath3 arcsec are expected .
figures 9 and 10 show synthesised images at different epochs for an sn at a distance of @xmath109 mpc , surrounded by cs dust shell / ring / blobs , as viewed with a spatial resolution of @xmath119 arcsec ( for the viewing direction of @xmath111 as a specific example ) . with this resolution
, one can distinguish different cs dust geometries , even for the cs dust located at @xmath120 pc . for a spherically symmetric distribution ,
investigation of the radial distribution is within reach , as seen by the hole for the shell model .
the ring and blob models show asymmetric image profiles that are clearly different from the shell model ( but depend on the viewing direction ) .
the shift of the centre in these models should be easily resolved beyond the spatial resolution for example , using this shift , the blob model is distinguishable from any spherically symmetric distribution .
an sn at a distance of @xmath109 mpc is a rare event , at an observed rate of one per decade or two .
figures 11 and 12 show the same images , but for an sn at a distance of @xmath121 mpc . at this distance
, we expect an event rate of roughly one per year ; thus , such observations are , in principle , annual occurrences . for cs dust located at @xmath120 pc around an sn at a distance of @xmath121 mpc , as observed by _ tmt / e - elt / gmt _ ( fig .
11 ) , the situation is similar to the case shown in fig .
an elongated profile of the ring distribution could be marginally resolved , and the shift in images for asymmetrically distributed cs dust ( e.g. , ring or blobs ) could be detectable by a sub - spatial resolution astrometry .
for the cs dust at @xmath108 pc , observations with a spatial resolution of @xmath119 arcsec can distinguish various cs dust geometries , given that such echo signals are detected .
thus , we conclude that the proposed future high - resolution nir ao can open up a new window to examine the cs environment around nearby sne .
we note that a target does not have to be one discovered after the new instruments become available .
for example , sn 2014j can already be a target assuming an observation at year 2024 , cs dust located at @xmath122 pc can still be investigated .
pilot work on this measurement may already be possible with @xmath123 m class telescopes approaching the diffraction limit ( @xmath93 arcsec ) .
we investigated the effects of _ re - emission _ of sn photons by cs dust for ir wavelengths .
we have shown that this effect allows observed ir light curves to be used to place constraints on the position / size and the amount of cs dust .
we applied the method to observed nir sn ia samples , and found that meaningful upper limits on the cs dust mass can be derived , even under conservative assumptions .
our results are summarised as follows : 1 .
@xmath69 for the cs dust at @xmath70 pc , @xmath71 at @xmath72 pc , and @xmath73 at @xmath74 pc .
the mass loss rate must be generally @xmath95 yr@xmath1 up to @xmath124 pc , rejecting a symbiotic path as a main population of the _ normal _ sne ia .
3 . for a particular case of sn 2006x ,
a deeper limit is obtained ( see below ) : @xmath125 for the cs dust at @xmath70 pc , @xmath69 at @xmath72 pc , and @xmath126 at @xmath74 pc .
the corresponding mass loss rate ( upper limit ) is @xmath127 yr@xmath1 up to @xmath128 pc and @xmath103 yr@xmath1 up to @xmath129 pc .
the above upper limits become tighter if one considers a non - spherical cs distribution .
the above upper limits become even tighter by an order of magnitude for carbon dust . from these results
, we conclude the following regarding the possibility of the dust - scattering scenario for the non - standard extinction law toward sne ia : 1 . the ` cs ' dust ( at @xmath130 pc ) echo can not be a general explanation of the non - standard extinction law with regard to sne ia .
the scenario would work for at most a fraction of sne ia , presumably at most , the highly reddened ones .
if the echo scenario is responsible for the extinction law in general , then the dust must be located at @xmath77 pc .
the required mass is , however , @xmath131 ; therefore , this must be attributed to ism rather than csm .
we have also demonstrated an approach that allows deeper limits to be applied .
this involves first using the standard nature of the nir light curves .
we demonstrated this for sn 2006x , showing that the upper limit can be deepened by two orders of magnitude from the conservative limit , as given above .
we suggest that further improvement is possible by refining the analysis method along this line via standardisation of the nir light curves for a given sample and minimisation of the dispersion as a function of time .
this would allow a deeper limit to be placed on the mass loss rate , potentially to the level to limit various models other than the particular symbiotic scenario .
again , one of the main advantages of the proposed method , compared to radio and x - ray constraints ( which are indeed complementary to the proposed method here ) , is the following : the limit to the amount of dust placed by the nir echo extends an order of magnitude larger in the spatial dimension ( or an order of magnitude to the past in the time dimension ) , reflecting the fact that the light speed ( for the echo ) is by an order of magnitude faster than the shock velocity ( for radio and x - rays ) . as we have shown in this paper , late - time photometry , even sparse , especially in the @xmath26-band ,
is essential to constrain dust ( and gas ) mass in the environment at @xmath66 pc .
our method underlines the importance and usefulness of such data .
our analysis is also highlighted by the possibilities of ( 1 ) constraining the origin of the non - standard extinction law , and ( 2 ) constraining the csm mass up to @xmath64 pc , namely the regimes that the other methods like radio and x - ray diagnostics can not probe . because the data are still rare , we propose increasing the sample of nir / mid - ir sne with continuous observational efforts .
of course , if one wants to go further toward the ism regime , mid- and far - ir data are important @xcite , and a combination of the early radio / x , early and late - time nir , and late - time mid-/far - ir , perhaps complemented by the absorption system study ( e.g. , * ? ? ?
* ) , will hopefully provide a complete view of these issues .
we also investigated the possibility of studying the cs dust distribution through high spatial resolution observations .
our results are summarised below . 1
. cs dust around sn 2014j could potentially be resolved by future _ hst _ observations , given that dense cs dust shell / ring / blobs exist at @xmath77 pc from the sn site .
future nir ao instruments attached to @xmath118 m class telescopes are capable of distinguishing various distributions of cs dust , thus providing a powerful means for investigating the cs environment around sne ia .
3 . such observations will be possible for nearby sne at a rate of roughly once a year .
the authors thank ariel goobar for fruitful discussion . k.m . acknowledges financial support by grant - in - aid for scientific research ( no . 23740141 and 26800100 ) from the japanese ministry of education , culture , sports , science and technology ( mext ) .
the work by k.m . is partly supported by world premier international research center initiative ( wpi initiative ) , mext , japan .
the authors also thank the anonymous referee for her / his many constructive comments . | the circumstellar ( cs ) environment is key to understanding progenitors of type ia supernovae ( sne ia ) , as well as the origin of a peculiar extinction property toward sne ia for cosmological application . it has been suggested that multiple _
scatterings _ of sn photons by cs dust may explain the non - standard reddening law . in this paper , we examine the effect of _ re - emission _ of sn photons by cs dust in the infrared ( ir ) wavelength regime .
this effect allows the observed ir light curves to be used as a constraint on the position / size and the amount of cs dust .
the method was applied to observed near - infrared ( nir ) sn ia samples ; meaningful upper limits on the cs dust mass were derived even under conservative assumptions .
we thereby clarify a difficulty associated with the cs dust scattering model as a general explanation for the peculiar reddening law , while it may still apply to a sub - sample of highly reddened sne ia .
for sne ia in general , the environment at the interstellar scale appears to be responsible for the non - standard extinction law .
furthermore , deeper limits can be obtained using the standard nature of sn ia nir light curves . in this application , an upper limit of @xmath0 yr@xmath1 ( for the wind velocity of @xmath2 km s@xmath1 )
is obtained for a mass loss rate from a progenitor up to @xmath3 pc , and @xmath4 yr@xmath1 up to @xmath5 pc .
[ firstpage ] circumstellar matter stars : mass - loss dust , extinction supernovae : general . |
the blast shell of a supernova remnant ( snr ) can accelerate charged particles to cosmic ray energies through diffusive shock acceleration by the snr shocks @xcite . particles are scattered upstream and downstream of a shock and some particles cross the shock repeatedly , gaining energy each time @xcite .
diffusive shock acceleration requires a seed population of particles with energies well in excess of the thermal ones , since only such particles can cross the shock repeatedly and be scattered efficiently by the magnetohydrodynamic waves on either side of the shock .
the presence of shock - accelerated electrons is evidenced by the radio emissions of snrs .
it is not clear yet how the electrons can reach the energy threshold for diffusive shock acceleration ; that requires that their kinetic energies are comparable to those of the ions .
some electrons must be accelerated locally , that is close to the snr shock or close to one of its precursors @xcite .
this injection of electrons is thought to be accomplished by the interplay of energetic ions with electrons and dust @xcite by means of plasma instabilities @xcite .
the instabilities occur in the upstream plasma , just ahead of the snr shock .
this region is called the foreshock .
one source of the ion beams is the downstream plasma .
ions can leak through the shock boundary and outrun it @xcite .
the shock can also reflect a substantial fraction of the upstream ions .
if the reflection is specular , then the ions can cross the foreshock at twice the shock speed . reflected and leaked ion beams can be observed in situ at the earth bow shock @xcite .
particle - in - cell ( pic ) simulations of fast plasma shocks suggest , that they may also occur at snr shocks when the upstream magnetic field is orthogonal to the shock normal @xcite .
precursor shocks move ahead of the main snr shock .
they can reach mildly relativistic speeds for the most extreme supernova explosions @xcite .
such shocks can also reflect ions in the presence of a quasi - parallel magnetic field @xcite .
we focus in this work on ion beam instabilities in the foreshock , but we do not model the shock itself .
we rely on a foreshock model that has been widely used , e.g. in the works by @xcite and by @xcite .
this model considers an ion beam , which moves through the upstream plasma , consisting of cool electrons and protons .
the system is spatially uniform . often a second
, counter - propagating ion beam is introduced , which cancels the current of the first one .
this second beam can be due to ions that have been reflected by a perpendicular shock and which return to the shock , after they have been rotated by the upstream magnetic field @xcite . here
a beam of co - moving negatively charged oxygen ions cancels the current of a proton beam .
this foreshock model can be used if all processes develop on scales well below an ion larmor radius and an inverse ion cyclotron frequency in the case of a perpendicular shock .
the foreshock model may be representative over larger scales otherwise , because ions can expand much farther if they move along the magnetic field , as it is observed at the earth s bow shock .
this foreshock model is not self - consistent , because it decouples the foreshock physics from that of the shock .
it also assumes a plasma that is spatially uniform , while the shock reflected ion beams vary in their mean speed and temperature .
however , we can examine with this model the acceleration of the upstream electrons as a function of the bulk parameters , such as the beam temperatures and mean speeds and the background magnetic field .
we can also separate individual processes from the global shock dynamics , that involves coupling over many scales @xcite .
this simplifies their identification and interpretation .
the periodic boundary conditions also allow us to reduce the box size , making pic simulations computationally efficient even for realistic ion - to - electron mass ratios .
the insight gained from such simulations can then guide us in the selection of plasma parameters for the more realistic and demanding pic simulations of full shocks .
simulation studies employing this foreshock model have revealed that proton beam - driven instabilites are not efficient electron accelerators .
an obstacle is the mass difference between the electrons and the protons .
most plasma instabilities saturate by accelerating the electrons .
the fields are too weak to accelerate the ions and the heated electrons quench most of the known plasma instabilities , e.g. through landau damping .
often only a few per cent of the ion energy is transfered to the electrons @xcite .
initially , this is also the case for ion beams that expand into the upstream plasma @xcite prior to the formation of a shock .
certain mechanisms @xcite can , in principle , accelerate the electrons to relativistic speeds , but this remains yet to be demonstrated with self - consistent pic simulations . however , the electrons , that have been accelerated by an instability like the buneman instability ( bi ) @xcite , are often spatially bunched , for example in form of phase space holes @xcite .
electron bunches or plasmons have a much higher inertia than electrons and they modulate the ions to give ion acoustic waves .
the ion acoustic waves can trigger further processes , one of which we consider in this work .
we investigate a plasma consisting of two spatially uniform beams .
one beam is composed of protons and electrons and the second one of protons and oxygen carrying a single negative charge . each of the beams is charge and current neutral .
both beams move at a relative speed of 0.4c .
all species have the temperature 9 kev .
these initial conditions are the simplest ones that take into account three particle species and they permit a first assessment of the impact of heavier ions .
section 2 discusses in more detail these initial conditions and their potential relevance for a foreshock plasma .
section 3 describes the simulation results , which can be summarized as follows .
a bi develops between the electrons and the protons of the second beam .
the bi saturates by forming electron phase space holes and these plasmons feed wave energy into the ion acoustic wave @xcite .
proton phase space holes develop , once the ion acoustic waves are strong enough to trap the beam protons .
the electrons eventually thermalize and they reach a mildly relativistic thermal speed , which is stabilizing the system for some time .
the bi plays no further role after the electron thermalization .
we examine in detail a proton phase space hole that grows into being a proton density pulse with a negative density @xcite .
the growth of this proton phase space hole may be caused by its interaction with the negative oxygen @xcite or by the disruption of the electron current @xcite .
the pulse triggers a strong ion acoustic double layer and it evolves subsequently into an electrostatic shock .
the electron phase space structure and evolution of this double layer resembles that , found by a recent simulation study by @xcite .
that study addressed the electron acceleration in the earth s auroral ionosphere .
however , our simulation yields the much stronger electron acceleration that is necessary for injecting the electrons into the diffusive shock acceleration at snr shocks .
the proton phase space distribution also differs from the standard picture of double layers outlined , for example , in the review papers by @xcite and by @xcite .
the trapped and free streaming protons in our simulation are well - separated in velocity , which is not generally the case for double layers .
the substantial free energy stored in the difference of the mean velocities of both beams causes the extreme electron acceleration up to 50 mev .
we discuss our simulation results in section 4 , we relate it to previous investigations of double layers and we discuss the necessary future research .
the bulk plasma of a supernova blast shell can expand into the ambient medium at a speed of up to 0.1c-0.2c , at least during its early phase @xcite .
a fraction of the blast shell plasma may expand even faster .
a plasma flow speed of 0.9c has , for example , been observed for a subshell ejected by a particularly violent supernova explosion @xcite .
shocks form where the expanding plasma impacts on the ambient plasma , the interstellar medium .
the typical shock speeds should be comparable to the few per cent of @xmath0 the main blast shell plasma has , but some precursor shocks may be considerably faster .
the detailed structure in particular of the fastest snr shocks is unclear .
we assume here , that the foreshock region is resembling at least qualitatively that of the well - known earth bow shock @xcite and we focus on the ion foreshock .
the ion foreshock is characterized by the presence of ion beams that can propagate far into the upstream .
however , we expect that a hot background of electrons is present in the foreshock , as pic simulations of mildly relativistic shocks demonstrate @xcite .
we examine the interaction between two spatially uniform and unbounded plasma beams , consisting of one negatively and one positively charged species each .
one beam consists of electrons and protons and the second one of protons and oxygen carrying a single negative charge .
the simulation box frame is the reference frame , in which both plasma beams move in opposite x - directions and with equal speed moduli @xmath1 .
the simulation models only the x - direction , which allows us to resolve simultaneously the ion scales and the electron scales at the correct mass ratio .
the beam 1 consists of electrons with the plasma frequency @xmath2 , where @xmath3 and @xmath4 are the elementary charge and the electron number density and @xmath5 is the dielectric constant .
the equally dense protons have the plasma frequency @xmath6 , where @xmath7 .
both species move at the mean speed @xmath8 with @xmath9 .
the beam 2 is composed of protons with @xmath10 and oxygen , which carries a single negative charge . the oxygen plasma frequency is @xmath11 .
the mean speed of both species of beam 2 is @xmath12 .
all species have the temperature @xmath13 kev , which gives the electron thermal speed @xmath14 .
by neglecting the @xmath15 and the heavy species of each beam , the dispersion relation for the beam - aligned electrostatic modes can be cast into the form @xmath16 the wavenumber and frequency in the reference frame of beam 1 of the most unstable wave are @xmath17 and @xmath18 .
the doppler - shift will reduce the frequency of the most unstable wave to @xmath19 . the wavelength @xmath20 .
the phase speed @xmath21 in the reference frame of beam 1 or @xmath1 in the box frame . replacing the oxygen with electrons gives symmetric beams , which lets the phase speed of the waves vanish in the box frame .
an exact resonance between the plasmons and the proton beam would not be possible .
the estimates for @xmath22 and @xmath23 , as well as the growth rate @xmath24 of the cold bi @xcite are fairly accurate .
figure [ fig1 ] shows the numerically obtained growth rate @xmath25 for our plasma parameters .
this growth rate map has been obtained by using a warm fluid model @xcite .
the growth rate map is symmetric around @xmath1 , so we do not need to consider @xmath26 .
the growth rate in the warm plasma model is lower than in the cold one by about 10% .
the corresponding wave number is increased by 20% .
the growth rate decreases for increasing @xmath27 .
the pic simulation method @xcite solves the vlasov - maxwell set of equations with the method of characteristics .
it approximates the phase space density @xmath28 by an ensemble of computational particles ( cps ) .
the cps correspond to phase space volume elements and are not physical particles .
the charge @xmath29 and mass @xmath30 of a cp can thus differ from those of the particles it represents .
the charge - to - mass ratio @xmath31 must , however , equal that of the physical particles and the ensemble properties of the cps approximate well those of the physical plasma .
the trajectories of the cps are prescribed by the lorentz equation of motion .
the maxwell - lorentz set of equations can be normalized , which provides us with results that are independent of the plasma density .
let @xmath3 and @xmath32 be the elementary charge and the electron mass .
our code is based on the numerical scheme outlined by @xcite and it solves the maxwell s equations for all components of the electromagnetic fields together with the lorentz equation , which updates the three - dimensional vector of the relativistic momentum for an ensemble of computational particles ( cps ) . the relevance of double layers for solar and astrophysical particle acceleration has been proposed first by @xcite .
@xcite have found double layers with pic simulations of oblique shocks , which dissipated the electron energy . in our pic simulation
the double layer emerges out of the bi and prior to the formation of the shock .
a relativistic and monoenergetic electron beam is ejected by the double layer .
the time evolution of the proton pulse implies , that the mean energy of this beam rises in time to about 50 mev , which is a much stronger acceleration than that reported , for example , for previous pic simulations of double layers @xcite .
the nonrelativistic beam speed and the high temperature ensure that the bi grows fastest .
the approximation of this system by a 1d pic simulation , which enforces a monodirectional ( plane ) wave spectrum , may thus initially be justified .
this geometry and the idealized initial conditions facilitate the interpretation of the physical processes and they permit us to resolve the relevant spatio - temporal scales .
however , the growth rate of the oblique modes is not negligible and the wave power is isotropized in time through weak turbulence @xcite .
the pic simulation may also become unrealistic when the electrons have been accelerated to relativistic speeds .
this is , because of the unresolved higher - dimensional instabilities , the generation of energetic photons through synchrotron- and bremsstrahlung and the back - reaction of these processes onto the plasma @xcite . charged dust and heavy ions exist in the supernova ejecta and in the interstellar medium @xcite .
their ionization state and their relative densities are unclear .
it is typically only the dust , which has a negative time - dependent charge @xcite , while the oxygen is positively charged @xcite .
our initial conditions are thus not necessarily representative for snr blast shell plasmas . however , the purpose of the oxygen is to introduce a slower time scale @xcite and also to obtain a resonance between plasmons and the ion acoustic wave , which facilitates their coupling @xcite .
both conditions can probably also be achieved by heavier ions , by a mix of ions or by dust .
the beam speed of 0.4c is too high for typical snr bulk flows , which can reach about 10 - 20% of the speed of light @xmath0 @xcite .
however , the shock - reflected ion beams can outrun the shock .
some components of the expanding plasma may also move faster than the mean speed of the snr ejecta .
precursor shocks of particularly strong snrs may reach speeds of up to 0.9c @xcite .
we select this beam speed , because it facilitates the processes we consider and because it allows us to estimate the maximum energy , which the electrons could gain in typical snr flows through the plasma instabilities discussed here .
the beam speed is also low enough to favor the growth of the bi rather than that of the multi - dimensional filamentation instability @xcite .
the simulation box length is @xmath33 .
it is resolved by @xmath34 grid cells , each with the length @xmath35 .
the electron debye length is @xmath36 .
each of the four species is represented by 250 cps per grid cell , giving a total of @xmath37 per species .
the system is followed in time for @xmath38 , which is subdivided into @xmath39 time steps @xmath40 .
the beams propagate the distance @xmath41 and even a light pulse can cross @xmath42 just once .
effects due to the periodic boundary conditions of the simulation box can not occur .
the electrostatic bi develops between the electrons of beam 1 and the protons of beam 2 . however , the secondary instabilities give rise to electric fields , which will let all particle species interact .
figure [ fig2 ] displays the time evolution of the relevant energy densities , which are the energy density @xmath43 of the electrostatic field and the @xmath44 $ ] , where the summation is over all particles @xmath45 of the species @xmath46 with the mass @xmath47 and a speed @xmath48 giving @xmath49 .
the @xmath50 grows in the interval @xmath51 at an exponential rate @xmath52 , which is below @xmath53 .
the growth rate @xmath54 is calculated for a sine wave .
the @xmath55 and the broad @xmath56 spectrum of unstable waves in fig .
[ fig1 ] implies , however , that we integrate over a mixture of waves , all of which have a growth rate @xmath57 and the initial growth of @xmath58 can not be compared directly to @xmath54 .
the electron trapping lets @xmath59 saturate at @xmath60 and it causes the growth of @xmath61 .
thereafter @xmath59 decreases and reaches a meta - equilibrium at @xmath62 .
the @xmath61 is almost constant until @xmath63 , when it and @xmath59 start to grow again .
all ion beams have lost less than 1% of their energy prior to @xmath63 , confirming that the bi is not an efficient electron accelerator .
then @xmath64 and @xmath65 start to change and even @xmath66 has decreased by 1% at @xmath67 .
figure [ fig3 ] shows the phase space distributions of the trapped electrons and of the protons of beam 2 at @xmath68 in a subinterval of the simulation box .
the trapped electrons gyrate around a potential that is moving with the mean speed @xmath1 of beam 2 , which is also the phase speed of the buneman wave .
only a fraction of the electrons is trapped .
the electrons of the untrapped bulk population move on oscillatory phase space paths .
the electron phase space hole in the interval @xmath69 has the expected size , while the electron phase space hole in the interval @xmath70 is larger .
electron phase space holes increase their size by coalescence and their mutual interaction will deform them @xcite .
the mean momentum of the proton beam is weakly modulated .
the initially periodic train of electron phase space holes will coalesce , until only a dissipative equilibrium and solitary electron phase space holes remain @xcite .
figure [ fig4 ] confirms this .
a simulation box interval is selected at the time @xmath71 , in which an electron depletion at @xmath72 separates the turbulent region @xmath73 from a smooth plateau at larger @xmath74 .
the latter constitutes a dissipative equilibrium .
the supplementary movie 1 follows the time - evolution of the electron phase space distribution at the position of the electron depletion .
the color scale is the 10-logarithm of the number of cps .
note that the absolute position in the box is given in movie 1 , while fig .
[ fig4 ] uses a relative position .
the movie 1 demonstrates that it is this density depletion , which is responsible for the turbulent phase space region .
a spatial correlation is visible in the fig .
[ fig4 ] between the electron depletion and a phase space hole in the protons of beam 2 .
this explains why the electron depletion is stationary in the reference frame of movie 1 , which moves with @xmath1 in the box frame .
the electron temperature is now relativistically high , by which the linear damping of the ion acoustic waves is reduced and these modes can accumulate wave energy .
the thermal anisotropy of the electrons causes the growth of the weibel instability @xcite , but it is suppressed here by the 1d geometry .
figure [ fig5 ] plots the @xmath75 , the density of the proton beam 2 and that of the electrons at the same location and time as fig .
the electric field at @xmath76 is attractive for the protons and repels the electrons , which is in agreement with the observed phase space structure in fig .
the electric field amplitude at this position is not unusually large , while the plasma density depletion is .
the oxygen density is unaffected at this time , as we show below .
the component energies in fig .
[ fig2 ] demonstrate that the equilibrium between @xmath59 and @xmath61 breaks at @xmath63 and that a secondary instability sets in after that time .
an examination of the full phase space data shows , that this secondary instability is spatially localized .
two similar , but spatially well - separated structures develop in the phase space distribution of the protons of beam 2 in the box .
that two structures rather than one grow suggests , that they develop with a reasonable probability and that their growth is not accidental .
figure [ fig6 ] displays the larger of these two structures at @xmath77 . a different sub - interval of the box than in the fig [ fig4 ]
is investigated , since the proton phase space structures convect with @xmath1 .
we refer to this proton structure as a pulse .
this pulse constitutes a density depletion @xcite .
the supplementary movie 2 animates in time the development of the pulse out of the proton phase space hole .
the color scale shows the 10-logarithm of the number of cps .
the movie 2 displays absolute box coordinates and its reference frame moves with @xmath1 through the box .
it demonstrates that the proton phase space hole is stationary for a long time , before it develops into a rapidly growing pulse .
it is this pulse that is responsible for the energy changes in the fig [ fig2 ] .
the electron phase space distribution reveals in fig .
[ fig6 ] a double layer that is driven by the pulse .
this double layer accelerates on @xmath78 a beam of electrons out of the bulk to about 7.5 mev .
the movie 1 demonstrates that the mean energy of the electron beam grows in time , while the beam temperature remains unchanged .
the electron beam then interacts with the bulk electrons through a two - stream instability , which is responsible for the large beam momentum oscillations .
the electric field of the double layer is strong enough to decrease the mean speed modulus of the proton beam 1 , which supplies the energy for the electron acceleration .
the movie 2 is stopped , when the protons of beam 1 and 2 start to mix by the formation of an electrostatic shock .
the movie 1 is stopped , when the electron beam energy exceeds the one resolved by the movie window . the electron distribution at @xmath79 is illustrated by fig [ fig7 ] .
the double layer has not spatially expanded but the electrons are now accelerated to 50 mev in the box frame , exceeding even the 20 mev energy of the protons in the same reference frame .
the total density of the ultrarelativistic beam in fig .
[ fig7]b ) is about 40% of the density of the bulk electrons in fig .
[ fig7]c ) .
the momentum distribution of the bulk electrons is practically identical to that in the fig.[fig4 ] .
this ultrarelativistic electron beam would drive the unresolved and faster - growing oblique mode instability @xcite , rather than the two - stream instability evidenced by the movie 1 . the type of instability should , however , not be important for the evolution of the electron double layer because it forms well behind it .
figure [ fig8 ] displays the phase space distributions of both proton beams in the same spatial interval as fig .
the proton distributions of both beams have merged and a large electrostatic shock forms at @xmath80 .
a second electrostatic shock involving only the protons of beam 2 occurs at @xmath81 .
the movie 2 illustrates that the second shock is driven by the rapid expansion of the proton density pulse .
it is remarkable that some protons of beam 2 have increased their @xmath82 momentum by a factor 2.6 through the electrostatic field , which is also shown by fig.[fig8 ] .
the electron double layer is thus also a proton accelerator .
the electrostatic field of the double layer does not change its sign , which contrasts the bipolar fields of phase space holes , like in fig [ fig5 ] .
solitary phase space holes are typically time - stationary in one dimension @xcite and the question arises , why the proton phase space holes grow here into such powerful rarefaction pulses .
the proton phase space hole is stable over a long time and it then suddenly turns into a pulse , which suggests an effect that develops on a slower time scale .
the electric field of the proton phase space hole will gradually displace the oxygen ions .
figure [ fig9 ] displays the density distribution of the oxygen ions at the three times @xmath83 , @xmath84 and @xmath85 .
the frame is moved with the beam 2 and the proton phase space hole is located at @xmath86 in the chosen frame .
the oxygen depletion at @xmath71 is comparable to the statistical fluctuations of the number of cps .
the integration of the density over 3 grid cells reduces these statistical fluctuations by the factor @xmath87 , but it keeps the density depletion value unchanged .
the proton phase space hole grows into the pulse at @xmath88 .
figure [ fig9]b ) reveals that the oxygen density modulation is significant at around this time .
it is thus likely that the sudden growth of the proton phase space hole is a consequence of this oxygen depletion .
as the proton pulse grows , the structure and amplitude of the oxygen depletion remain qualitatively unchanged , but the depletion expands in space .
this may reflect the observation from the movie 2 that the shape of the proton pulse is practically unchanged until @xmath89 .
the proton pulse and the oxygen depletion are intertwined .
the proton beam 1 and the electrons in our simulation have a significant mean speed in the reference frame of the proton phase space hole and they could provide the energy for the growth of the proton pulse .
the movie 1 reveals that the proton phase space hole gives rise to vortices in the electron phase space distribution and the electron flow is turbulent , which excerts a drag on the proton phase space hole .
a coupling between the proton phase space hole and the streaming electrons is established , which can also trigger or accelerate the growth of the phase space hole @xcite .
the simulation further evidences that the proton pulse grows at the expense of the kinetic energy of the proton beam 1 .
this work has examined secondary instabilities triggered by the buneman instability ( bi ) @xcite with a pic code @xcite .
one spatial x - direction has been resolved , which implies that the wave spectrum driven by the bi is monodirectional .
this limits the realism of the simulation but it firstly allows us to resolve all relevant spatio - temporal scales and , secondly , it facilitates the identification of the relevant physical processes .
the large box has ensured that the periodic boundary conditions have not affected the results .
we have considered two unmagnetized , interpenetrating , uniform and unbounded beams that have been composed of three particle species .
the beam 1 has contained protons and electrons and the beam 2 has been formed by protons and by oxygen with a single negative charge .
each beam has been charge- and current neutral .
the relative speed between both beams has been set to 0.4c and the initial temperature of all species has been selected to be 9.1 kev .
the bi has saturated by the trapping of electrons @xcite , which groups the electrons into phase space holes that give rise to charge density modulations .
the latter can behave as quasi - particles and are thus referred to as plasmons .
we could match the mean ( group ) speed of these plasmons with that of the protons of beam 2 by selecting negatively charged oxygen rather than electrons .
such a resonance facilitates the coupling of the plasmon energy into the ion acoustic wave mode @xcite .
initially the equilibrium between the electric fields and the electrons has been preserved @xcite and the ion beams had lost only a small fraction of their energy ; the electrons moved consequently only at mildly relativistic speeds .
the thermal anisotropy of the electron distribution would trigger an instability similar to that proposed by @xcite , resulting in the growth of magnetic fields and multi - dimensional field structures that can not be represented by a 1d pic simulation .
the weibel instability may , however , not be important .
the proton phase space holes form simultaneously with the collapse of the electron phase space holes that gives rise to the thermal anisotropy .
the electromagnetic fields of the weibel instability may not be strong enough to affect the interplay between the already present proton phase space holes and the oxygen and might be negligible . as the time progressed , two large density rarefaction pulses @xcite have developed in the protons of the beam 2 .
we have examined the larger one in more detail .
the long time between the formation of the proton phase space hole and the pulse has suggested that it has been caused by the oxygen response .
we have found that the proton density pulse grows , once the electric field of the proton phase space hole has displaced a substantial fraction of the oxygen ions .
a related effect has been observed by @xcite .
there it was shown that an electron phase space hole was ejected by a density depletion in the background ions , which its electric fields excavated . here , the consequence of the oxygen depletion is not the acceleration of the phase space hole but its growth .
the electrons may also contribute to the growth of the proton phase space hole @xcite .
the growth of the proton density pulse has resulted in an electron double layer that accelerated the electrons to ultrarelativistic speeds . to the best of our knowledge
this is the first time that such an electron acceleration could be observed in pic simulations of nonrelativistic beam instabilities .
the electrons have reached a peak energy of 50 mev .
this immense electron acceleration implies , that the 1d pic simulation geometry is not appropriate any more .
firstly , the electrons moving at such speeds drive the oblique mode instability @xcite , rather than the two - stream instability .
the oblique modes are excluded by the simulation geometry .
secondly , the extreme electron temperature anisotropy resulting from the thermalization of the ultrarelativistic electron beam would drive a very strong weibel - type instability .
the subsequent magnetic field growth would imply that electrons emit synchrotron radiation on top of the bremsstrahlung @xcite .
the balance of these processes @xcite requires a three - dimensional geometry and a method to deal with the radiation , which is not possible with our pic code . at the same time
, this electron double layer constitutes a plasma structure that can easily accelerate the electrons to a speed , which allows them to enter the diffusive shock acceleration mechanism .
this could be achieved with beam speeds that are not much higher than the flow speeds of supernova blast shells @xcite .
the oxygen ions we have modelled can also be found in such plasmas , although not in the high density and in the ionisation number we have assumed here @xcite .
our negative charge is more representative of dust , which we also find in supernova blast shells @xcite .
our findings can thus not be applied directly to a supernova scenario .
however , the formation of the electron double layer in the presence of heavier ions clearly demonstrates the need to introduce these into pic simulations , which so far only consider one ion species @xcite .
future simulation work must address the minimum speed , at which the double layer can form .
this has to be done in form of parametric simulation studies , because it is not clear for which combinations of the beam speeds and temperatures the double layers can form .
existing analytic studies @xcite do not cover the case we consider here .
the effects of the obliquely propagating modes driven by the initial beam instability have to be examined , as well as those due to the weibel instability driven by the thermal anisotropy of the electron distribution after the initial bi has saturated .
this requires at least two - dimensional pic simulations .
finally , other ion species and plasma compositions have to be examined to assess how easily double layers can form .
this is important , because the ion composition is not easily obtained @xcite and not unique .
the modelling of double layers and their consequences is also of interest for other astrophysical scenarios .
ion double layers , which are characterized by ion distributions that resemble the electron distribution here , may regulate the mass inflow onto neutron stars @xcite and accelerate particles in accretion discs @xcite .
they may also contribute to solar coronal ejections @xcite .
this work has been financially supported by the visiting scientist scheme of the queen s university belfast , by the swedish research council vetenskapsrdet and by the projects ftn 2003 - 00721 of the spanish ministerio de educacin y ciencia and pai-05 - 045 of the consejera de educacin y ciencia de la junta de comunidades de castilla - la mancha .
the swedish high - performance computer center north ( hpc2n ) has provided the computer time and support . | two charge- and current neutral plasma beams are modelled with a one - dimensional pic simulation .
the beams are uniform and unbounded .
the relative speed between both beams is 0.4c .
one beam is composed of electrons and protons and one out of protons and negatively charged oxygen ( dust ) . all species have the temperature 9 kev .
a buneman instability develops between the electrons of the first beam and the protons of the second beam .
the wave traps the electrons , which form plasmons .
the plasmons couple energy into the ion acoustic waves , which trap the protons of the second beam .
a proton phase space hole grows , which develops through its interaction with the oxygen and the heated electrons into a rarefaction pulse .
this pulse drives a strong ion acoustic double layer , which accelerates a beam of electrons to about 50 mev , which is comparable to the proton kinetic energy .
the proton distribution eventually evolves into an electrostatic shock .
beams of charged particles moving at such speeds may occur in the foreshock of supernova remnant shocks .
this double layer is thus potentially relevant for the electron acceleration ( injection ) into the diffusive shock acceleration by supernova remnants shocks . |
dimitri mihalas and co - workers developed the formalism and the numerical algorithms to model the atmospheres of hot stars in non - lte .
after having accomplished this task for the static , plane - parallel case , the problem of spherically expanding atmospheres was attacked in a series of seminal papers by mihalas , kunasz , and hummer in the years 1975/76 .
one important step was to realize that the radiative transfer in highly supersonic flows , where the doppler shifts are much larger than the width of the line absorption profile , are treated most conveniently in the co - moving frame of reference ( cmf ) .
for the equations of statistical equilibrium the `` radiative transition rates '' are required , and these have to be evaluated with the ( angle - averaged ) radiation intensity in the cmf as it is `` seen '' by the matter . as a drawback , the equation of radiative transfer becomes a _ partial _ differential equation in the cmf : because the frequency @xmath0 is measured with respect to the co - moving frame , a photon is changing its frequency when propagating through the differentially moving medium . for a ray with coordinate @xmath1 , a velocity @xmath2 in radial direction , and angle @xmath3 between the radial direction and the ray ,
the cmf transfer is @xmath4 in spherically extended atmospheres , the radiation field becomes very anisotropic , and hence many rays of different angle ( or , impact parameter ) are needed for an accurate numerical representation .
however , for isotropic opacity the radiative rates depend only on the angle - averaged radiation intensity @xmath5 .
therefore it is sufficient , and much more economic , to solve the _ moment _ equations of the radiative transfer .
most of the iterations can be performed by solving the moment equations with given eddington factors . only from time to time
the latter must be updated . for this purpose
the moments of the radiation field must be newly integrated ( eq.([eq : momdef ] ) ) from the angle - dependent intensity , which is obtained from solving the angle - dependent transfer equation for many rays . as a closure to the moment equations , auer & mihalas ( 1970 ) introduced the `` method of variable eddington factors '' for the static case . in mihalas ,
kunasz & hummer ( 1976 ) they extended this concept to spherically expanding atmospheres . _ four _ moments of the radiation intensity are needed in this case : @xmath6 = \int { \rm d}\mu\ i_\nu~ [ 1 , \mu,\mu^2 , \mu^3]\ .
\label{eq : momdef}\ ] ] eddington factors are defined as @xmath7 ( like in the static case ) and @xmath8 . in the numerical application of this method , we encountered problems with both of these definitions . on physical grounds
, the intensity - like moment @xmath5 should be positive .
solving the angle - dependent cmf transfer eq.([eq : rt ] ) by means of a feautrier - like differencing scheme , however , sometimes yields slightly negative values at some frequency and radius points , just due to the limited numerical accuracy .
this can make the definition of the eddington factor @xmath7 singular .
we overcame this problem by solving the angle - dependent radiative transfer with an _ integration _
method along short characteristics .
such method can guarantee positive intensities .
the problem with the other eddington factor , @xmath8 , is even more severe .
there is actually no reason why the eddington flux @xmath9 may not become negative in some situations .
therefore we replaced the originally proposed definition of @xmath10 by @xmath11 .
although now relating a flux - like to an intensity - like moment , this eddington factor behaves well in the iteration process . for the simultaneous solution of the ( discretized ) radiative transfer equation and the constraint equations , mihalas and co - workers composed a large system of algebraic equations and solved them by linearization and elimination techniques .
such method is very much restricted to only a few atomic levels and transitions .
present - day non - lte codes instead simply solve the radiative transfer and the statistical equilibrium in turn .
such an iteration scheme , a so - called lambda iteration , does not converge when large optical depths are involved .
however , in the 1980s it has been discovered that convergence can be achieved when only the local feedback of the radiative transfer is consistently incorporated into the statistical equations .
this method , termed `` accelerated lambda iteration ( ali ) '' or `` approximate lambda operator ( alo ) '' technique , allows to calculate model atmospheres with many hundred non - lte levels , thousands of line transitions , and even millions of further lines in an approximate way as needed for `` iron line blanketing '' ( see fig.[fig : wr6 ] for an example ) .
mass - loss from o stars is driven by the momentum transfer from uv photons to the ions of heavy elements .
this has been worked out in hydrodynamically self - consistent models for o stars , based on the fundamental concept of castor , abbott & klein ( 1975 ) .
the photospheric radiation is absorbed in a large number of lines , some of them being optically thick . a line
which remains optically thick throughout the wind sweeps up all photospheric flux over a bandwidth @xmath12 , as illustrated in fig.[fig : linedriving ] .
the intercepted momentum is then @xmath13 where @xmath14 is the specific luminosity at the line frequency @xmath15 .
comparison with the wind momentum per time , @xmath16 , reveals that the mass - loss rate driven by a single optically thick line is @xmath17 remarkably just the same rate by which mass is converted to energy in the stellar interior for sustaining the luminosity .
the actual mass - loss rate in a star can be conveniently expressed in terms of a @xmath18 , the effective number of lines : @xmath19 .
as hydrogen burning consumes about 120 times more fuel than the mass that is coverted into energy , wind mass loss becomes a signifant driver of stellar evolution when @xmath18 is comparable to this factor . within this model approach
, the maximum mass - loss rate is obviously reached when the whole spectrum is covered by lines and each photospheric photon delivers its momentum once to the wind , i.e. when @xmath20 or @xmath21 .
the mass - loss rates from o and b stars fall below this `` single - scattering limit '' , albeit for supergiants like @xmath22pup not by much . for many of the wolf - rayet stars , however , empirical mass - loss rates exceed this limit considerably .
the cak models fall short to explain such dense winds .
mass - loss exceeding the single - scattering limit requires that photons deliver their momentum more than once .
such multiple - scattering effects are not included in cak - type models . to overcome this shortfall , we combined the potsdam wolf - rayet powr model atmosphere code with the hydrodynamic equations . as the whole radiative transfer is treated consistently , all possible multi - line effects are automatically included .
indeed we obtained self - consistent wr models , as has been demonstrated in grfener & hamann ( 2005 ) for the wc5 prototype wr111 .
figure[fig : combined - acc ] shows for this model how the inward and outward forces are perfectly balanced in the consistent , stationary hydrodynamic solution .
hence we are sure now that the winds of wr stars can be driven by radiation pressure .
the previous shortfalls in explaining wr winds were obviously due to deficiencies in the conventional radiative - transfer treatment in cak - type models . specifically in the dense winds from wr stars the multiple - scattering effects from the numerous lines of iron - group elements become important .
but why wr stars develop so much denser winds than o stars ? we have studied this question by means of a model grid for luminous wn stars of late subtype ( wnl ) in grfener & hamann ( 2008 ) . according to our self - consistent hd models , the key point is the @xmath23 ratio ( see fig.[fig : wnl - z ] ) .
wr stars are just very close to the eddington limit where the radiation pressure on free electrons balances gravity ( @xmath24 ) .
mass - loss rates of about @xmath25 , as typical for wr stars , are obtained for an eddington @xmath26 of 0.5 if the metallicity is solar .
but even at very low metallicity , strong winds are predicted if @xmath26 is sufficiently close to unity . and pressure gradient , red / dotted curve ) .
for a model with prescribed velocity field ( `` beta law '' ) and mass - loss rate , the equality of both terms is not given ( left panel ) . by incorporating the hydrodynamical equations into the powr model atmosphere code , a solution for @xmath27 and @xmath28 is found that is hydrodynamically consistent everywhere in the wind ( right panel ) . ] ) , but different @xmath23 ratio ( eddington factor @xmath29 ) and different metallicities @xmath30 ( compared to the solar metallicity @xmath31 ) .
hydrodynamically consistent powr models are represented by a black dot , blue lines connect models with the same metallicity .
the empirical parameters of the wn7 star wr22 are indicated by a red square .
( from grfener et al . 2008 ) ] the hydrodynamic models described in the previous two subsections are stationary solutions of the problem .
however , these solutions are expected to be unstable because of the de - shadowing effect : if one volume element would experience a positive perturbation of its velocity , it intercepts more radiation in optically thick lines and is therefore even more accelerated .
the effect of this instability can be seen in the time - dependent hydrodynamical simulations presented by feldmeier et al .
these calculations are for spherical symmetry , but account for heating and cooling processes .
a snapshot for an o star wind is shown in fig.[fig : shocks ] .
the instability leads to the formation of dense shells , which contain almost all of the wind material and have a radial separation of about one stellar radius .
these structures are created by strong reverse shocks with velocity jumps of several hundred km / s .
the shocked gas reaches temperatures of a few million kelvin . corresponding multidimensional simulations which account for the radiative force on lines in non - radial directions and for the the heating and cooling of the shocks are not yet available .
as discussed in sect.[sect : hydroshocks ] , hot - star winds are expected to be structured by the hydrodynamic instability that is inherent to the line - driving mechanism . this prediction is supported by various observational evidences for wind variability and inhomogeneity . an important spectroscopic fingerprint of wind `` clumping '' had been pointed out already by hillier ( 1984 ) , namely the electron - scattering wings of strong emission lines .
the `` thomson scattering '' of photons by free electrons is often considered as being coherent . however , this is only correct in the rest frame of the electron ( when neglecting the compton shift due to the recoil ) , but not when considering the electron s thermal motion .
in fact , the thermal speed of the electrons is very high , about 500km / s at 10kk , and therefore the corresponding doppler shift causes a redistribution in frequency that is comparable with the width of wind - broadened emission lines .
the emission lines that dominate the spectra of wolf - rayet stars are mainly fed by recombination cascades .
the recombination process scales with the square of the density .
the thomson scattering , in contrast , scales linearly with the ( electron ) density .
considering an inhomogeneous medium , the ratio between the mean of the density - squared and the square of the mean density , @xmath32 , becomes larger than unity .
therefore , the fit of emission lines and their electron - scattering winds can provide a tool to estimate the degree of stellar - wind clumping . in a simple approach
, one can assume that the whole stellar wind is concentrated in clumps , while the interclump medium is void .
when the clumps fill a fraction @xmath33 of the volume , the density in the clumps is by a factor @xmath34 enhanced over a smooth wind .
a further simplifying assumption is that the individual clumps are optically thin at all frequencies .
this approximation is termed `` microclumping '' .
it can be easily implemented in model atmosphere codes , as it leaves the radiation field homogeneous . for wolf - rayet stars ,
the electron - scattering wings of strong emission lines are usually well reproduced with this formalism when the clumping contrast @xmath35 is set to values between 4 and 10 ( e.g. hamann & koesterke 1998 ) .
the spectra of o and b stars , however , do not show strong enough emission lines for which this method could be applied .
there are empirical indications that clumping develops already deep in the wind where the expansion velocity is still small , and decreases at larger distance from the star ( puls et al .
2006 ) .
a major consequence of microclumping is a reduction of empirical mass - loss rates that are derived from @xmath36-depending processes , such as emission lines fed by recombination cascades , as well as the free - free radio emission .
compared to `` unclumped '' diagnostic , this reduction is by a factor of @xmath37 . in a recent work , zsarg et al .
( 2008 ) introduced a second , diluted wind component that fills the space between the clumps .
this thin interclump gas becomes more easily ionized , especially from hard radiation produced in wind shocks , and therefore contributes to the spectral lines from high ions , e.g. , from ovi .
the _ microclumping _ assumption that the individual clumps have only small optical depth can not be justified for all frequencies .
the wind - compressed shells in the hydrodynamical simulations ( cf.fig.[fig:shocks ] ) have a thickness of the order of 0.1@xmath38 .
estimates from variability lead to similar clump diameters . in the winds from wr and o - type stars , clumps of such size are expected to be optically thick in the cores of strong lines , and also in the far - uv and soft x - ray continua from bound - free and k - shell absorption .
therefore we introduce a more general concept which accounts for the finite optical thickness of clumps
. however , drastic approximations are necessary to keep the problem manageable .
early - type stars can be often observed as x - ray sources .
many such sources are found to be binaries , where the x - rays may be produced by the collision of two stellar winds . however , there are also many putatively single early - type stars with x - ray detections .
x - ray emission is believed to originate from hot , shocked gas embedded in the stellar winds .
there are two arguments why the x - rays are production must be located in the inner wind region , firstly because the line - driving instability should be largest where the acceleration is strongest , and secondly because the x - ray line profiles are typically wind - broadened with only about half of the terminal wind velocity . in order to be observable , this x - ray emission has to escape from the expanding atmosphere through the absorbing `` cool '' wind component which contains the bulk of the matter .
this wind attenuation may account for the observed decrease of the emergent x - ray flux with increasing density of the stellar wind illustrated in fig.[fig:3xrays ] .
checking this wind attenuation quantitatively reveals a severe problem , as demonstrated in fig.[fig : tau1 ] . for x - ray wavelengths where
emission lines are observed , the plot shows the radius where the radial optical depth reaches unity . according to these predictions from typical models ,
x - rays should not be able to escape from lower wind regions , at least not for an o supergiant like @xmath22pup , and definitely not for wolf - rayet stars .
an escape from this dilemma can be searched by reducing the stellar mass - loss rates by adopting very high clumping factors @xmath35 . at least in the case of wn stars
this runs in various contradictions , e.g. with the electron - scattering line wings ( see above ) , or with the mass - loss rate of the binary v444cyg from its period change .
alternatively , whe have shown that the x - ray observations can be explained when the macroscopic size of the clumps is taken into account .
the x - rays lines are emitted from hot shocks , while the absorption takes place in the cool wind component by continuum processes .
this decoupling of emission and absorption greatly facilitates a semi - empirical modeling .
monte - carlo calculations have been applied to model the emergent x - ray line profiles with randomly distributed line emitters and absorbing clumps ( oskinova et al .
2004 , 2006 ) .
basically the same results are obtained from an analytical treatment that holds in the limit that the emitting spots and the absorbing clumps are numerous enough for a statistical description ( feldmeier et al .
2003 , owocki & cohen 2006 ) .
the consequence of the wind clumping is a reduction of the effective opacity .
some atomic absorbers are hiding in the shadow of others , while the gaps between the shell fragments open channels where photons can travel far . the potsdam work ( feldmeier et al . 2003 , oskinova et al .
2004 , 2006 ) found that the observed x - ray line profiles can be explained best if the `` clumps '' are assumed to have the shape of `` pancakes '' or `` shell fragments '' ( see fig.[fig : xmodel ] ) .
this idea is also supported from the theoretical consideration that the forces which compress the clumps act mainly in radial direction .
if clumps are anisotropic , their _ projected _ cross section leads to a `` venetian blind effect '' , where in the limit of entirely flat fragments the opacity scales proportional to @xmath39 . with the corresponding effective
opacity becoming angle - dependent as @xmath40 , the optical depth increment for a ray of any impact parameter grows with the change of radius : @xmath41 , in contrast to the isotropic case where @xmath42 . in oskinova
et al . ( 2007 ) we considered specifically the case of a spectral _ line _ formation .
we take the statistical approach as already used for the x - ray absorption , but in the line case we restrict ourselves to isotropic clumps which have the same optical diameter @xmath43 in all directions . at a given location ,
all clumps have uniform size . under these assumptions ,
the _ effective _ opacity becomes @xmath44 where @xmath45 is the well - known `` porosity length '' ( e.g. owocki , gayley & shaviv 2004 ) .
note that for optically thin clumps the smooth - wind opacity @xmath46 is recovered , while in the limit of large @xmath43 the effective opacity becomes @xmath47 .
in the case of line formation we have to address the issue of doppler shifts . in our one - component model _
the clumps are the wind_. hence clumps as such move with the wind velocity field @xmath28 . due to the supersonic expansion
, rays of a given ( observer s frame ) frequency can only interact with clumps near the _ surface of constant radial velocity _
the line opacity of all other clumps is doppler - shifted out of the resonance .
hence the porosity effect for lines is very pronounced , as illustrated in fig.[fig : clumps ] .
an important issue is the doppler broadening _ within _ a clump , which widens the crvs to a resonance _
zone_. we describe the clump opacity with a gaussian distribution corresponding to a doppler - broadening velocity @xmath48 , which may schematically account for the velocity distribution inside the clump caused by stochastic motion ( thermal , microturbulent ) , as well as by velocity gradients .
note that the degree of porosity for line radiation depends on this parameter : for smaller @xmath48 , the opacity profile of a clump peaks to a higher maximum , resulting in a smaller effective opacity of the atmosphere .
the 1d hydrodynamic calculations shown in fig.[fig : shocks ] predict that within the dense shells the radial velocity gradient is negative , and span typically over a velocity range of 100km / s or less .
when the motion is strictly radial , the transverse velocity gradient is positive . assuming that the clumps cover a cone angle of a few degrees ( seen from the center of the star ) , a wind speed of @xmath492000km / s leads to a transverse expansion of a clump with less than 100km / s .
empirically it is known that the narrowest spectral features from stellar winds , like the dacs ( discrete absorption components ) , still have typical widths corresponding to 50 ... 100km / s .
therefore values of @xmath48 in this range seem to be a reasonable choice .
a further important parameter is of course the clump size , which is related to the mean separation of the clumps , and to their total number at one instant of time .
several observational and theoretical arguments help to restrict this parameter , as discussed in oskinova et al .
there we implemented the _ macroclumping _ formalism only in the `` formal integral '' of the powr code , thus making the first approximation that the non - lte source function is not affected .
a motivation of this study was the so - called pv problem .
bouret et al .
( 2005 ) and fullerton et al .
( 2006 ) had pointed out that in o star spectra the pv resonance line in the extreme uv is observed much weaker than predicted . they suggest to reduce drastically the adopted mass - loss rates , and compensate for this by higher density - enhancement factors @xmath35 in the microclumping treatment .
resonance lines depend , like the electron - scattering wings discussed in sect.[sect : microclumping ] , only linearly on density .
pup as observed by copernicus is shown for comparison ( black , ragged line ) .
we adopt a microturbulence velocity of 50km / s to account for the velocity dispersion inside the clumps .
numerous weak spectral features in this range are due to iron . the usual modeling yields much too strong pcygni features ( blue , continuous line ) . with our macroclumping formalism , the line features are reduced to the observed strength ( red , dotted curve ) . from oskinova
et al . ( 2007 ) ] as an unwanted side effect , macroclumping slightly de - saturates the black absorptions of strong uv resonance lines like the doublets from civ and nv .
however , according to zsarg et al . ( 2008 ) these lines are anyhow not mainly formed in the dense clumps , but in the diluted interclump medium which is neglected in our modeling . | analyzing the spectra from wolf - rayet stars requires adequate non - lte modeling of their expanding atmosphere .
the numerical schemes for solving the radiative transfer in the co - moving frame of reference have been developed by mihalas and co - workers 30 years ago .
the most elaborate codes can cope today with many hundred explicit non - lte levels or super - levels and account for metal - line blanketing .
the limited agreement with observed spectra indicates that the model simplifications are still severe .
one approximation that has to be blamed is homogeneity .
stellar - wind clumping on small scales was easily implemented , while `` macro - clumping '' is still a big challenge .
first studies showed that macro - clumping can reduce the strength of predicted p - cygni line profiles in o - star spectra , and largely affects the x - ray line spectra from stellar winds .
the classical model for radiation - driven winds by castor , abbot and klein fails to explain the very dense winds from wolf - rayet stars .
only when we solved the detailed non - lte radiative transfer consistently with the hydrodynamic equations , mass - loss rates above the single - scattering limit have been obtained .
address = universitt potsdam , germany address = universitt potsdam , germany , , altaddress = armagh observatory , northern ireland address = universitt potsdam , germany address = universitt potsdam , germany |
two - component bose - einstein condensates ( becs ) are increasingly appreciated as a rich and versatile source of intricate non - equilibrium pattern dynamics phenomena .
in addition to experimental observations @xcite , pattern dynamics in two - component becs has attracted significant theoretical interest ( see , e.g. , @xcite and citations in @xcite ) . in a number of two - component bec experiments
reported over more than a decade , a standard technique has been to start from the equilibrium state of a single - component bec , e.g. , populating a single hyperfine state of @xmath0rb , and then using a @xmath1 pulse to switch half the atoms to a different hyperfine state @xcite .
this results in a binary condensate where the two intra - species interactions ( @xmath2 and @xmath3 ) and one inter - species interaction ( @xmath4 ) are all slightly different from each other , but the starting state is the ground state determined by @xmath2 alone .
since it has been realized several times in several different laboratory setups , this is a paradigm non - equilibrium initial state for binary condensate dynamics .
a thorough and general analysis of the dynamics subsequent to such a @xmath1 pulse is thus clearly important . in this article
we present such an analysis , clarifying the combined role of the inter - species interaction ( @xmath4 ) and the strength @xmath5 of the trapping potential .
we provide a stability analysis mapping out regions of the @xmath5@xmath4 parameter space hosting different types of dynamics .
since it is now routine to monitor real - time dynamics in such experiments ( e.g. @xcite ) , we also directly analyze the real - time evolution after a @xmath1 pulse .
it is widely known that the ground state of a uniform two - species bec is phase separated or miscible depending on whether or not the inter - species repulsion dominates over the self - repulsions of the two species , i.e. , if @xmath6 then the ground state is phase separated @xcite .
this criterion is also a key ingredient in understanding dynamical features such as pattern dynamics in the density difference between the two species such `` spin patterns '' emerge when the phase separation condition is satisfied .
this can be understood as the onset of a modulation instability @xcite , identified by the appearance of an unstable mode in the excitation spectrum around a reference stationary state . for
a _ homogeneous _ situation , linear stability analysis shows that modulation instability sets in when the condition of eq .
( [ eq : cond ] ) is satisfied @xcite .
the situation is different in the presence of a trapping potential .
phase separation in the ground state , as well as the appearance of modulation instability when starting from a mixed state , now requires larger inter - species repulsion @xcite .
this suggests that the region of parameter space where pattern dynamics occurs also depends on the trap .
a trap is almost always present in cold - atom experiments , and it is easy to imagine experiments where the trapping potential is not extremely shallow but varies between tight and shallow limits .
it is thus necessary to examine the relevance of eq .
( [ eq : cond ] ) for trapped binary becs . to this end
, we explore different trap strengths spanning several orders of magnitude , and identify the appropriate extensions of eq .
( [ eq : cond ] ) for the type of spin dynamics resulting from the @xmath1 protocol described above .
we focus on the effects of two parameters .
first , we study effects of changing cross - species interaction @xmath4 , thus generalizing eq . for trapped situations .
second , we explore the role of the relative strength of the trap with respect to the interactions .
our analysis , performed for a one - dimensional ( 1d ) geometry , sheds light on the situation where @xmath2 and @xmath3 are close but unequal : ( a ) the stability analysis is performed for @xmath7 and their difference serves only to select appropriate instability modes ; ( b ) the simulations are performed with @xmath8 . in section [ sec_model ]
, we introduce the formalism and geometry . in section [ sec_stability_analysis ] , we show results from a linear stability analysis for a sequence of trap strengths , and identify and analyze relevant modulation instabilities . through an analysis of unstable modes , we present a classification of the parameter space into dynamically distinct regions , in relation to the prototypical initial state explained above .
this may be regarded as a dynamical `` phase diagram '' .
a remarkable aspect is that the `` phase transition '' line most relevant to spin pattern dynamics does not arise from the first modulation instability ( studied in ref .
this first instability mode is antisymmetric in space , and as a result is not naturally excited in a symmetric trap with symmetric initial conditions .
complex dynamics ( not due to collective modes but rather due to modulation instability ) is generated only when the first _ spatially symmetric _ mode becomes unstable , which occurs at a higher value of @xmath4 . in section [ sec_timeevolution ]
we provide a relatively detailed account of the time evolution .
for each trap strength @xmath5 , for values of @xmath4 not much larger than @xmath2 , we observe simple collective modes . above a threshold value of @xmath4 ,
the oscillation amplitude becomes sharply stronger , and at the same time the motion becomes notably aperiodic , signaling that the dynamics is more complex than a combination of a few modes .
dynamical spin patterns start appearing at this stage and become more pronounced as @xmath4 is increased further .
the threshold value at which the dynamics changes sharply corresponds to the second modulation instability line rather than the first , as we demonstrate through careful choice of parameters in each region of the phase diagram derived from stability analysis .
some further connections between the stability analysis and dynamical features , relating to the length scale of generated patterns , appear in section [ sec_lengthscales ] . in the concluding section [ sec_conclusions ] we place our results in context and
point out open questions .
the relevant time - resolved experiments have been performed in both quasi-1d geometries ( highly elongated traps with strong radial trapping ) @xcite and in a 3d bec of cylindrical symmetry with the radial variable playing analogous role as the 1d coordinate @xcite .
since the basic phenomena are very similar , we expect the same theoretical framework to describe the essential features of each case . for definiteness , in this work
we show results for 1d geometry .
we expect the general picture emerging from this work to be qualitatively true also for other geometries exhibiting the same type of spin dynamics .
we describe the dynamics in the mean field framework at zero temperature , i.e. , by two coupled gross - pitaevskii equations @xcite : @xmath9 condensate wave functions @xmath10 and @xmath11 are normalized to unity , and @xmath5 is the strength of the harmonic trap .
factors of particle number and radial trapping frequency are absorbed as appropriate into the effective 1d interaction parameters @xmath12 @xcite .
we consider purely non - dissipative dynamics , i.e. , we do not attempt to model experimental loss rates with a phenomenological dissipative term as done in , e.g. , refs . @xcite .
the equations above are in dimensionless form because we measure lengths in units of trap oscillator length and time in units of inverse trapping frequency , for a hypothetical trap of unit strength ( @xmath13 ) .
the scale for trap strengths is itself fixed by imposing @xmath14 . with this convention ,
small values of @xmath5 correspond to a bec in the thomas - fermi limit . for comparison
, we note that the parameters of the experiment of ref .
@xcite corresponds to @xmath5 of order @xmath15 in these units .
of course one can switch between different units via the transformation : @xmath16 , @xmath17 , @xmath18 , @xmath19 , and @xmath20 , where @xmath21 is an appropriately chosen scale .
the initial state after a @xmath1 pulse involves both components occupying the ground state of a single - component system of interaction @xmath22 , because the atoms were all in the first hyperfine state before the pulse .
we model this initial situation as a two - component bec with @xmath23 .
the @xmath1 pulse may then be regarded as a sudden change ( a _ quantum quench _
@xcite ) of the interaction parameters @xmath3 and @xmath4 .
we use @xmath24 for the stability analysis of section [ sec_stability_analysis ] . for the explicit time evolution
reported in section [ sec_timeevolution ] , we use @xmath2 and @xmath3 values close but unequal : @xmath14 , @xmath25 .
this choice of close values is convenient for illustrating the structure of the phase diagram , especially for shallow traps . in rubidium experiments
the difference between @xmath2 and @xmath3 is somewhat larger ( in the common case using @xmath0rb hyperfine states @xmath26 and @xmath27 ) ; however our insights should be relevant to a broad regime of possible experiments .
a full exploration of the regime of arbitrary differences ( @xmath28 ) remains an open task , beyond the scope of the present manuscript .
numerical simulations presented in section [ sec_timeevolution ] were performed using a semi - implicit crank - nicolson method @xcite .
we provide in this section a stability analysis for @xmath7 that maps out the regions of @xmath5-@xmath4 parameter space which support pattern formation instabilities .
ideally , one might like to perform a stability analysis around the initial state .
however , in contrast to the homogeneous case @xcite , we are faced with the situation that the initial state is not a stationary state of the final hamiltonian .
the choice of reference state is therefore a somewhat subtle aspect of the present analysis .
we use as reference state @xmath29 the lowest - energy spatially symmetric stationary state of the case @xmath30 , with parameter @xmath4 set to its final value .
( for large @xmath4 , this is not the ground state for these parameters , which is phase - separated . )
this reference state has the advantage of looking relatively similar to our actual initial state ( two components totally overlapping in space ) , and of being a stationary state of the hamiltonian for which we analyze linear stability .
our reference state can be regarded as placing both components in the single - component ground state for interaction @xmath31 .
we are not aware of a suitable stationary state even more similar to the actual initial state .
we will see that our stability analysis around this reference state will predict remarkably well the main observed time - evolution features described in section [ sec_timeevolution ] .
note that it is not natural to use @xmath32 , because stationary states for such a case typically do not overlap completely in space . instead
, in our approach the difference between @xmath2 and @xmath3 will play the important role of selecting certain instability modes over others .
for this reason , inferences from the present analysis apply only to small relative differences between @xmath2 and @xmath3 .
we linearize eqs .
( [ eq : gp1 ] ) and ( [ eq : gp2 ] ) around the reference stationary state @xmath29 : @xmath33 \exp(-i \mu t),\nonumber\\ \psi_2(x , t ) = \left [ \psi_{0}(x ) + \delta\psi_2(x , t ) \right ] \exp(-i \mu t),\end{aligned}\ ] ] where @xmath34 is the chemical potential corresponding to the reference state . by keeping only terms of the first order in @xmath35 and @xmath36
, we obtain a system of linear equations which can be cast in the form : @xmath37 here @xmath38 is a matrix differential operator which , upon discretization or upon expansion in a set of orthogonal functions , becomes the so - called stability matrix ( e.g. , @xcite ) .
we analyze below the eigenmodes of the stability matrix , which we have obtained by numerically calculating the reference stationary state @xmath29 and expanding in the basis of harmonic trap ( non - interacting ) eigenstates .
since we use @xmath30 for the stability analysis , eigenmodes will have well - defined `` species parity '' , i.e. will all be either even [ @xmath39 or odd [ @xmath40 with respect to the interchange of species .
even modes describe in - phase motion of the two components and simply correspond to the excitation spectrum of a single - component bec with interaction constant @xmath31 .
odd modes are more interesting they describe out - of - phase motion of two components and are therefore reflected in the spin dynamics .
additionally , due to the spatial inversion symmetry @xmath41 , the solutions will also have well - defined spatial parity , and we can distinguish spatially symmetric and antisymmetric modes .
typical eigenspectra are presented in fig .
[ fig : fig3 ] . in the case of a tight trap @xmath42
, we notice two modes whose frequencies are nearly constant .
these are even modes encoding single - component or in - phase physics .
the lower one is the dipole ( kohn ) mode with frequency equal to the trap frequency @xmath5 .
the second nearly - constant mode is the breathing mode , which for elongated traps takes value close to @xmath43 .
the breathing mode ( oscillations of cloud size ) is visible in the plots of fig .
[ fig : fig1 ] ( section [ sec_timeevolution ] ) as a fast oscillation of the total condensate widths .
( color online . )
the dynamical `` phase diagram '' showing the critical values of @xmath4 for the onset of the two types of modulation instability versus the trap strength @xmath5 .
the instability lines are shown as straight lines joining numerically determined values .
the oscillation schematics on the right ( and corresponding arrows ) indicate that left - right - left oscillation modes are persistent everywhere above the @xmath44 line , while in - out - in modes are persistent only above the higher @xmath45 line .
the spatially symmetric instability ( @xmath45 line ) is the one relevant for experimental situations with symmetric traps .
squares mark values used in the dynamical simulations of figs .
[ fig : fig1 ] and [ fig : fig2 ] ( table 1 ) . ]
the two lowest - lying eigenmodes are odd modes encoding out - of - phase physics . for @xmath46 , their frequencies are significantly below the breathing mode , and therefore lead to relatively slow oscillations in the spin density .
this will also be visible in the real - time dynamics presented in section [ sec_timeevolution ] ( first two columns of figs .
[ fig : fig1 ] and [ fig : fig2 ] ) .
the forms of the corresponding eigenvectors are shown in the lower right of fig .
[ fig : fig3 ] .
the nature of the eigenvectors shows that the motion related to the lowest mode corresponds to the left - right oscillations of the two species , while the next odd mode corresponds to spatially symmetric spin motion .
the frequencies of these two modes become imaginary at certain values of @xmath4 , thus leading to the onset of modulation instabilities .
the antisymmetric mode becomes unstable at smaller value of @xmath4 ( @xmath47 for @xmath42 ) in comparison to the symmetric mode ( @xmath48 for @xmath42 ) . in a spatially symmetric trap
, there is no natural mechanism for exciting the spatially antisymmetric mode . on the other hand
, any difference between @xmath2 and @xmath3 naturally excites the second ( spatially symmetric ) mode .
thus , the second mode , occurring at larger @xmath4 , is the relevant instability for understanding the dynamics observed in experiments and explored numerically in section [ sec_timeevolution ] .
we find similar excitation spectra for trap strengths @xmath5 spanning several orders of magnitude .
the spatially antisymmetric mode becomes unstable before the spatially symmetric mode , and both instabilities get closer to 1 as the trap gets shallower . for example , for @xmath49 ( also shown in fig .
[ fig : fig3 ] ) the lowest instability sets in for @xmath50 , while the next one appears at @xmath51 .
the distinction between two instabilities becomes ever smaller as we go toward a uniform system @xmath52 , where the phase - separation condition eq .
( [ eq : cond ] ) becomes exact .
nevertheless , even for shallow traps , the issue is not purely academic as the precision in experimental measurement and control of scattering lengths continues to improve @xcite . in fig .
[ fig : fig4 ] ( main panel ) , the results of the stability analyses are combined to present a dynamical `` phase diagram '' .
the two lines show the two instabilities ( @xmath44 and @xmath45 ) as a function of trap strength @xmath5 .
for very shallow traps , the two transition lines merge as @xmath53 . the lower transition line ( @xmath44 )
was previously introduced in ref .
however , for a trap and initial state with left - right spatial symmetry , this is not the relevant dynamical transition line , because the first even mode only becomes unstable at some higher @xmath4 value , given by the @xmath45 line . in the next section
, we will see that spin pattern dynamics is indeed only generated when the inter - component repulsion @xmath4 exceeds the second instability line ( @xmath54 ) , and that crossing the first instability ( @xmath55 ) is not enough for pattern formation in a spatially symmetric trap .
in this section we present and analyze the dynamics obtained from direct numerical simulation of the gross - pitaevskii equations ( [ eq : gp1 ] ) and ( [ eq : gp2 ] ) , after the system is initially prepared in the ground state of the situation @xmath56 .
the subsequent dynamics is performed with @xmath57 , @xmath58 , and several different values of @xmath4 for each trap strength @xmath5 .
.parameters from the first five columns are used for the plots in figs .
[ fig : fig1 ] and figs . [ fig : fig2 ] .
the instability values @xmath59 and @xmath60 ( introduced in figs .
[ fig : fig3 ] and [ fig : fig4 ] and discussed in section [ sec_stability_analysis ] ) are also given for each trap strength . [
cols="^,^,^,^,^,^,^",options="header " , ] it is difficult to show the full richness of pattern dynamics through plots of a few quantities .
we choose to show the dynamics through two types of plots ( figs .
[ fig : fig1 ] and [ fig : fig2 ] ) . fig .
[ fig : fig1 ] shows the time dependence of the root mean square widths of the two components @xmath61 while fig .
[ fig : fig2 ] shows density plots of the density difference ( spin density ) , @xmath62 . in both figures
, each row corresponds to a different trap strength ( @xmath5 ) , and we approach the shallow trap ( thomas - fermi ) limit going from top to bottom . for each @xmath5 the four values of @xmath4 from table [ tab : tab ] are used for figs .
[ fig : fig1 ] and [ fig : fig2 ] .
we have chosen @xmath4 values such that the first panel in each row is in the parameter region where there are no instabilities , the second one is in the region where the only instability is the antisymmetric one , and the third on each row is at @xmath4 values just above the second , relevant , instability .
the fourth panel on each row is at higher @xmath4 values .
the choice of @xmath4 values with respect to instability lines is clear in the tighter traps of the top three rows , as also shown by squares in fig .
[ fig : fig4 ] . for shallow traps ( lower rows ) ,
the instability lines are too close together and too close to @xmath63 , so making such choices is not meaningful . in the following ,
as we compare features of the different columns , we implicitly exclude the lowest row ( smallest @xmath5 ) .
this is also indicated by the fact that the schematic instability lines in figs .
[ fig : fig1 ] and [ fig : fig2 ] are not extended to the lowest row .
broadly speaking , we note that there is only regular ( collective - mode ) dynamics in the second - column figures ( @xmath55 ) even though an instability is present .
there is generally a sharp difference between the second and third figure in each row , indicating that the second instability ( @xmath45 ) is the relevant one .
the fourth panel on each row is at higher @xmath4 values , showing more rich dynamics . in fig .
[ fig : fig1 ] , we show time - dependence of the individual widths ( @xmath64 , @xmath65 ) and also of the total root mean square width , @xmath66 .
consistent with our observation that spatially symmetric modes ( and not the antisymmetric ones ) are naturally excited in the current setup , the dynamics shows signatures of the two most prominent spatially symmetric modes noted in fig .
[ fig : fig3 ] .
the breathing mode is the easiest to notice and most ubiquitous it shows up in almost every parameter choice as oscillations in the total density ( in - phase in the two components ) , with a typical period given by @xmath67 .
this follows from the frequency of this mode being almost constant near @xmath68 .
we also see out - of - phase motion of the two components , associated with the lower spatially - symmetric mode in fig .
[ fig : fig3 ] , which has odd species parity . in the first two columns of fig .
[ fig : fig1 ] , corresponding to smaller values of @xmath4 such that this mode has small real frequencies , this is excited as a regular ` spin ' mode . for example , at @xmath49 and @xmath69 , we observe an out - of - phase oscillation with the period of approximately @xmath70 , much slower than the breathing mode .
in addition , the oscillation period of the out - of - phase motion is slower in the second than in the first column of each row , corresponding to the decreasing frequency of the mode , as seen in stability analysis ( fig .
[ fig : fig3 ] ) .
once @xmath4 becomes large enough that the instability threshold for this mode is crossed , the oscillation amplitudes increase sharply and the width dynamics becomes strongly aperiodic and irregular ( third column of fig .
[ fig : fig1 ] ) .
this signifies the onset of pattern dynamics , as opposed to the excitation of a regular collective mode around a stable state .
irregularity of the width dynamics at stronger @xmath4 is even more apparent in the fourth column of fig .
[ fig : fig1 ] .
it is noteworthy that the spatially antisymmetric modes play no role and do not show up in these dynamical simulations .
we see no signature of the kohn mode . nor do we see any sharp change associated with the instability of the antisymmetric mode , i.e. , there is no sharp difference between the first two columns of fig . [
fig : fig1 ] . in fig .
[ fig : fig2 ] we show the dynamics of the `` spin density '' @xmath62 .
the case of very shallow traps ( last row ) , resembles the data in refs .
as in fig .
[ fig : fig1 ] , the first two columns show regular oscillations , corresponding to collective modes without instability .
a sharp change occurs , not across the first instability line ( between 1st and 2nd column ) , but instead across the second instability line ( 2nd and 3rd columns ) , especially for tighter traps ( top three rows ) where comparison with instability lines is meaningful .
the sharp change can be noted through the color scales , which is dramatically different between second and third columns in the upper rows .
in homogeneous stability analysis , the length scale of patterns is inferred from the wavevector ( momentum ) at which an instability first occurs .
since we perform our stability analysis specifically for trapped systems , we do not have a momentum quantum number .
nevertheless , the eigenvectors of the unstable modes contain information about the form of patterns generated in the dynamics of the trapped system .
this is illustrated in fig .
[ fig : fig5 ] , where the eigenvectors of the lowest unstable even - parity modes are shown for several values of @xmath4 , together with the spin patterns generated in the non - equilibrium dynamics .
there is a close match between the distance between nodes of the eigenvectors ( rough analog of ` wavevector ' ) and the length scales involved in the patterns . in fig .
[ fig : fig2 ] we see that the patterns contain more spatial structure in shallow traps .
the top two rows ( tight traps ) only show in - out - in type of patterns .
this can be understood from the idea that the interactions induce length scales ( ` healing lengths ' ) in the problem , which are smaller for larger interactions , and which set the length scale of spatial structures . for tight traps ,
the healing length set by the interactions is large or comparable to the cloud size , so that only global dynamical patterns are generated . in such traps ,
generation of complex patterns with many spatial oscillations would require much higher values of @xmath71 . for shallow traps ,
the healing length becomes much smaller than the cloud size ; as a result one can have a multitude of dynamical spin structures in the system , of the type seen in experiments and prior simulations @xcite .
this heuristic explanation can be made quantitative by counting the number of nodes appearing in the eigenmodes ( as in fig .
[ fig : fig5 ] ) .
in this article we have analyzed a widely used dynamical protocol for two - component becs , which involves starting from the ground state of one component and switching half the atoms to a different component through a @xmath1 pulse .
we have presented a stability analysis suitable to the trapped situation , and also presented results from extensive dynamical simulations . through an analysis of unstable modes , we have presented a classification of the parameter space into a number of dynamically distinct regions , in relation to the prototypical initial state .
this may be regarded as a dynamical `` phase diagram '' . in the ` stable ' regime of parameter space ( no modulation instabilities )
, our stability analysis explains the observed slow spin oscillations compared to the fast breathing mode oscillations of the total density .
we demonstrate that the important `` phase transition '' line for spatially symmetric situations relevant to most experiments is not the first instability ( studied in ref .
@xcite ) , but a second transition line .
the first instability is antisymmetric in space , and as a result is not naturally excited in a symmetric trap .
our stability analysis is performed relative to a stationary state of the situation @xmath7 .
the @xmath1 pulse of the experiments ( in the cases where @xmath72 ) can be considered as turning on a nonzero @xmath73 , i.e. , turning on ` buoyancy ' such that one component gains more energy by being in the interior of the trap compared to the other .
this helps to select instability modes which are symmetric in space .
since we have used a stability analysis with @xmath7 to analyze dynamics with @xmath72 , an obvious question is how the ratio @xmath74 affects the regime of applicability of this scheme .
we expect that features of this ( @xmath7 ) stability analysis are useful for dynamical predictions as long as @xmath75 is roughly more than @xmath76 .
for example , for shallow traps ( small @xmath5 ) , the instabilities occur at @xmath75 values comparable to 0.01 , which is why the placement of parameters in the three dynamical regions of the ` phase diagram ' ( fig .
[ fig : fig4 ] ) is not meaningful for the smallest @xmath5 values ( lowest rows of figs .
[ fig : fig1 ] and [ fig : fig2 ] ) . for the stability analysis we used a reference stationary state which is of course
not the initial state : the initial state is the ground state for @xmath77 , while the reference state is the lowest - energy spatially symmetric stationary state corresponding to the final value of @xmath4 .
the instability lines found in this stability analysis would describe even better a situation where the dynamics is triggered by a small quench of @xmath4 , as opposed to the changes of @xmath4 that we consider here , which can be relatively large .
we have looked at some examples of this type of dynamics and indeed find instabilities matching the stability analysis extremely well .
however , although the initial state in the @xmath1 dynamics is somewhat different from the reference state of our stability analysis , our results show that this stability analysis does provide an excellent overall picture of the dynamics generated by the @xmath1 protocol .
( color online . )
top : eigenvectors of the most unstable even eigenmodes , from the stability analysis of section [ sec_stability_analysis ] , for @xmath78 , and @xmath79 , @xmath80 , and @xmath81 from left to right . below each eigenvector , the corresponding spin dynamics after the @xmath1 protocol ( parameters of section [ sec_timeevolution ] ) is shown through the time evolution of @xmath62 . ] our work opens up a number of questions deserving further study .
first of all , we have thoroughly explored the @xmath5@xmath4 parameter space , while assuming that the intra - component interactions @xmath2 and @xmath3 are unequal but close in value .
the regime of large difference ( @xmath28 ) clearly might have other interesting dynamical features which are yet to be explored .
second , in this work we have restricted ourselves to the mean field regime . while the mean field description captures well the richness of pattern formation phenomena ( c.f . refs .
@xcite in addition to present work ) , it may be worth asking whether quantum effects beyond mean field might have interesting consequences for the patter dynamics generated by a @xmath1 pulse . for bosons in elongated traps , regimes other than mean field ( such as lieb - liniger or tonks regimes ) may occur naturally in experiments @xcite .
dynamics subsequent to a @xmath1 pulse in strongly interacting 1d gases outside the mean field regime is an open area of investigation .
third , we have assumed a spatially symmetric trap and an initial condition with spatial symmetry , and this plays a crucial role in the selection of instability channels . in a real - life experiment
, the trap will have some left - right asymmetry .
also , thermal and quantum fluctuations can initiate spatially antisymmetric excitations . the extent to which a small spatial asymmetry affects spin dynamics remains unexplored
; in such a case we would have some type of competition between two types of instabilities .
@xcite has studied dynamical effects of fluctuations ( noise ) , but the effects of thermal and quantum fluctuations is yet to be studied in the context of a @xmath1 protocol .
finally , one could consider time evolution and spatiotemporal patterns generated by a @xmath1 pulse in the presence of an optical lattice , described by the dynamics of a two - component bose - hubbard model .
this is a situation easy to imagine realizing experimentally .
one could speculate complex interplay between spin dynamics and the spatial arrangement of mott and superfluid regions .
. acknowledges discussions with a. bala and support by the ministry of education and science of the republic of serbia , under project no .
numerical simulations were run on the aegrandomis e - infrastructure , supported in part by fp7 projects egi - inspire , prace-1ip , prace-2ip , and hp - see .
njvd acknowledges support from fom and nwo .
m. r. matthews , d. s. hall , d. s. jin , j. r. ensher , c. e. wieman , e. a. cornell , f. dalfovo , c. minniti , and s. stringari , phys .
81 * , 243 ( 1998 ) .
d. s. hall , m. r. matthews , j. r. ensher , c. e. wieman , and e. a. cornell , phys .
lett . * 81 * , 1539 ( 1998 ) .
h. j. lewandowski , d. m. harber , d. l. whitaker , and e. a. cornell , phys .
88 * , 070403 ( 2002 ) k. m. mertes , j. w. merrill , r. carretero - gonzlez , d. j. frantzeskakis , p. g. kevrekidis , and d. s. hall , phys .
99 * , 190402 ( 2007 ) .
p. wicke , s. whitlock , and n. j. van - druten , arxiv:1010.4545 .
r. p. anderson , c. ticknor , a. i. sidorov , and b. v. hall , phys .
a * 80 * , 023603 ( 2009 ) p. boehi , m. f. riedel , j. hoffrogge , j. reichel , t. w. haensch , and p. treutlein , nat .
* 5 * , 592 ( 2009 ) .
k. kasamatsu and m. tsubota , phys .
lett . * 93 * , 100402 ( 2004 ) k. kasamatsu and m. tsubota , phys .
a * 74 * , 013617 ( 2006 ) .
l. wen , w. m. liu , y. cai , j. m. zhang , and j. hu , phys .
a * 85 * , 043602 ( 2012 ) . c. k. law , h. pu , n. p. bigelow , and j. h. eberly , phys .
* 79 * , 3105 ( 1997 ) .
busch , j. i. cirac , v. m. perez - garcia , and p. zoller , phys .
a * 56 * , 2978 ( 1997 ) .
h. pu and n. p. bigelow , phys .
lett . * 80 * , 1134 ( 1998 ) .
r. graham and d. walls , phys .
a * 57 * , 484 ( 1998 ) . | in the study of trapped two - component bose gases , a widely used dynamical protocol is to start from the ground state of a one - component condensate and then switch half the atoms into another hyperfine state . the slightly different intra - component and inter - component interactions can then lead to highly nontrivial dynamics .
we study and classify the possible subsequent dynamics , over a wide variety of parameters spanned by the trap strength and by the inter- to intra - component interaction ratio . a stability analysis suited to the trapped situation
provides us with a framework to explain the various types of dynamics in different regimes . |
neutron stars are objects of extremely rich internal structure .
although their interior structure is still very uncertain , it seems that observations and theoretical studies of neutron stars are quite in agreement concerning the structure of their exterior parts .
more specifically , there is agreement that neutron stars , 1 - 2 minutes after their formation , are cold enough to solidify their exteriors and form a crystal crust thanks to coulomb forces between the various atomic nuclei .
the crust is covered by a very thin fluid ocean , while the interior is formed by a super - fluid mantle ( up to 5 km in size ) .
the composition of the core is highly uncertain .
the crystal crust extends from the neutron star s atmosphere @xmath3 km down where densities reach nuclear densities around @xmath4 gr/@xmath5 .
the coulomb forces of the crystal ions forming the crust can be described via the shear modulus @xmath6 which is inversely proportional to the 4th power of the ion spacing . since the restoring force is the coulomb force
, the periods of the torsional modes will strongly depend on the shear modulus , and its values will characterize the spectrum .
up to now , there are only two detailed calculations of the crust equation of state @xcite .
both were based on an approximate formulation @xcite leading to a typical value for the shear modulus of @xmath7 @xcite . in the newtonian limit and in the absence of strong magnetic fields
, @xcite found that the period of the fundamental torsional modes @xmath8 depends mainly on the radius of the star @xmath9 , the speed of the shear waves @xmath10 and the angular index @xmath11 via the following relation @xmath12^{1/2 } v_s / r \label{eq : sigma}\ ] ] where @xmath13 , and @xmath6 , @xmath14 are the shear modulus and the density , respectively .
torsional modes , which are axial - type oscillations , are thought to be more easily excited during a fracturing of the crust since they only involve oscillations of the velocity .
the velocity field of torsional oscillations is actually divergence - free without a radial component .
torsional modes ( _ t_-modes ) are labeled as @xmath15 , where @xmath11 is the angular index , while the index @xmath16 corresponds to the number of radial nodes in the eigenfunctions of the overtones for a specific @xmath11 .
the shear and torsional modes are well studied in newtonian theory , see e.g. @xcite , while there are only a few studies in general relativity @xcite .
relativistic effects have been found to increase significantly the fundamental @xmath17 torsional mode period by roughly 30% for a typical @xmath18 , @xmath19 model .
the aim of this work is the study of the effect of rotation on torsional modes . up to now
this is studied only in newtonian theory in a single study by @xcite , see also @xcite .
that study suggested that the frequency of a torsional mode in a rotating star is given by @xmath20 where @xmath21 is the fundamental frequency of the torsional mode of a non - rotating star and @xmath22 is its rotational frequency .
since the typical frequency of the torsional modes of non - rotating stars @xmath21 is of the order of 25 - 40 hz , it is obvious that they will be subject to the so called chandrasekhar - friedman - schutz ( cfs ) instability , even for small rotational frequencies @xcite .
this is a quite interesting result for rotating neutron stars since the instability ( together with that of the r - modes ) might lead to further fracturing and/or melting of the crust .
recently , there is increased interest in the study of torsional modes because of the belief that soft gamma repeaters ( sgrs ) could be magnetars experiencing star - quakes that are connected ( through the intense magnetic field ) to gamma ray flare activity .
magnetar star - quakes may be driven by the evolving intense magnetic field which accumulates stress and eventually leads to crust fracturing .
there are three sgr events detected up to now .
the first event was detected already in 1979 from the source sgr 0526 - 66 @xcite , the second in 1998 from sgr 1900 + 14 @xcite , while the third and most energetic one was observed in december 2004 from the source sgr 1806 - 20 @xcite .
analysis of the tail oscillations , for a full discussion see @xcite , revealed the presence of oscillations from a few tenths of hz up to about 2 khz . in an attempt to fit the observed frequencies to the torsional modes of various eos for the core and the crust , @xcite suggested that neutron star models should have stiff equations of state and a mass between 1.6 - 2 @xmath23 .
still these results depend critically on the interpretation of the order of each mode . for this , see the discussion in @xcite . in the present work ,
we derive the perturbation equations for torsional oscillations of rotating stars in general relativity following the approach by @xcite in the cowling approximation which neglects perturbations of the space - time .
the article is organized as follows . in the next section
, we describe the general - relativistic equations that have been used to describe the background stellar configuration , and we derive the perturbation equations for torsional oscillations of rotating stars in the cowling approximation . in the third section
, we describe the numerical techniques used to calculate the modes together with a toy problem .
the last is meant to explain the numerical results .
the article closes with a summary and discussion .
we consider a slowly rotating relativistic star described by the metric @xmath24 where @xmath25 , @xmath26 and @xmath27 ( the frequency of the local frame ) are functions of the radial coordinate @xmath28 . up
to first order in @xmath22 , the background 4-velocity of the star is given by @xmath29 \label{bvelocity}\ ] ] where @xmath22 is the angular velocity of the star .
the background stellar models are solutions of the tov equations and an equation describing the dragging of inertial frames @xmath30 where @xmath31 .
we assume that the star consists of a perfect fluid described by the energy - momentum tensor @xmath32 we also assume that the star is isotropic , therefore , the background shear tensor vanishes . due to the spherical symmetry of the background
, the perturbations of the background configuration can be decomposed into spherical harmonics .
this leads to a large system of partial differential equations @xcite . here
the space - time perturbations are omitted ( cowling approximation ) and also the coupling between spheroidal ( polar ) and toroidal ( axial ) perturbations since they only marginally affect the eigenfrequencies of the torsional modes .
under these approximations , the radial component of the perturbed velocity field and the variations of pressure and density will remain unaffected since they are polar perturbations .
the perturbation of the 4-velocity @xmath33 is related to the displacement vector @xmath34 through the relation @xmath35 .
for an observer co - rotating with the star , this is translated to : @xmath36 in other words , the velocity perturbations are described by the time derivative a function @xmath37 related to the displacement vector @xmath34 , here the dot stands for the temporal derivative .
the perturbed energy - momentum tensor , which includes the contribution from shear , is given by @xmath38 where @xmath39 is the shear tensor , @xmath39 is defined by @xmath40 . here @xmath41 is the rate of shear given by the lie derivative of the shear along the world lines @xcite @xmath42 and @xmath43 is the projection tensor @xmath44 the speed of shear waves on the crust depends on the shear modulus @xmath6 , the density @xmath45 and the pressure @xmath46 of the star according to @xmath47 .
a typical value of the speed of shear waves is @xmath48 .
the perturbation equation for the energy - momentum tensor @xmath49 provides a single equation for the axial perturbations @xmath50 + 2 { \mbox{\rm i}}m \varpi \left[{1 \over \lambda } + v_s^2 \left ( 1 - { 2 \over \lambda } \right ) \right ] \dot{z}\end{aligned}\ ] ] where @xmath51 . in the absence of shear @xmath52 ,
an inertial observer obtains @xmath53 this gives the well known relation for the r - mode frequency in the newtonian limit ( @xmath54 ) , while in the relativistic case it leads to a continuous spectrum @xcite .
in the last case couplings between higher @xmath55 and polar perturbations need to be included for a proper study of the spectrum .
in this section , we present an approximate solution of the boundary value problem and describe the numerical techniques used in the calculation of the frequencies of torsional modes of rotating stars . for the numerical estimation of the frequencies , we use two different techniques .
the first approach assumes a harmonic time - dependence of the perturbations .
this leads on a boundary value problem .
the second approach is based on a direct time evolution of equation ( [ eq : t_master ] ) followed by a fourier transform in time of the obtained values at a fixed radial position .
we will describe only the first approach .
the second approach has only been used for the verification of the results .
the fourier transform of equation ( [ eq : t_master ] ) , i.e. , the assumption that @xmath56 , leads to the following differential equation @xmath57 - e^{2\nu } { \lambda - 2\over r^2 } \right\ } z = 0 \ , \ , .\end{aligned}\ ] ] the boundary conditions in the center ( or at the lower end of the crust ) and on the stellar surface are @xmath58 where @xmath59 is the distance of the lower end of the crust from the center .
the above system of equations defines an eigenvalue problem for the frequencies of the torsional modes @xmath60 . for its solution
, we use two approaches .
the first is an approximate analytic solution , and the second is a numerical solution . here
we provide an approximate analytic solution to the eigenvalue problem based on bessel functions . in order to be able to use bessel functions and to treat the problem semi -
analytically , we need to simplify equation ( [ eque0 ] ) .
for this we make the following simplifying assumptions , @xmath61 , @xmath62 , @xmath63 and @xmath64 .
this reduces equation ( [ eque0 ] ) to a bessel equation @xmath65 z = 0 \label{eq : toy4}\ ] ] where @xmath66 together with the boundary conditions ( [ eq : bc1 ] ) , this equation leads on a simpler boundary value problem . actually , for high frequencies , the eigenvalues could easily be estimated using the wkb method . on the other hand ,
since the frequency of the fundamental torsional mode is relatively small , it is unclear whether the wkb approximation is applicable , in particular in the discussion of cfs instability .
the last involves the investigation of the limit @xmath67 .
the general solution of equation ( [ eq : toy4 ] ) can be given in the form of bessel functions @xmath68 where @xmath69 and @xmath70 are arbitrary constants .
since the condition for regularity in the center demands that @xmath71 , the contribution of the bessel functions @xmath72 is excluded since it is divergent for @xmath73 .
the second boundary condition demands that @xmath74 which leads to the following transcendental equation @xmath75 where @xmath76 .
the roots of this equation can be found by elementary numerical methods .
this leads to the following values for @xmath77 @xmath78 the eigenfrequencies can be obtained from the roots of the equation @xmath79 by setting @xmath80 ( the frequency in the absence of rotation ) . in this way
, we get an approximate form of the torsional mode frequency @xmath81 in the rotating frame @xmath82 ^ 2 } + \frac{m \omega}{\ell \left(\ell+1\right ) } \ , \label{eq : toy3c}\ ] ] which for an inertial observer ( @xmath83 ) has the form @xmath82 ^ 2 } - \frac{m \omega \left(\ell^2+\ell-1\right)}{\ell \left(\ell+1\right ) } \ , \ , .
\label{eq : toy3d}\ ] ] both relations agree with the newtonian results of @xcite , equation ( [ eq : toy3c ] ) and of @xcite equation ( [ eq : toy3d ] ) .
note that @xmath84 leads to a well known form of the frequency of the fundamental torsional mode for non - rotating newtonian stars , see equation ( [ eq : sigma ] ) .
as function of the rotation ( both in hz ) .
the dashed line corresponds to the frequency of the toy model and the continuous line to the frequency of the relativistic model achieved by a direct numerical solution of the eigenvalue problem .
the data are for a stellar model with a realistic eos a @xcite for the fluid core and an eos for the crust given by @xcite ( model @xmath85 ) .
this stellar model has radius @xmath86 km @xmath87 and crust thickness @xmath88% .
, width=377 ] the numerical solution of the eigenvalue problem defined by equations ( [ eque0 ] ) - ( [ eq : bc1 ] ) is solved by a shooting method .
its results have been tested by direct numerical evolution of the time dependent equation ( [ eq : t_master ] ) .
the numerical results verify the suggestion by @xcite that the torsional modes _ are cfs unstable_. in figure 1 , we plot the frequency of the fundamental torsional mode ( @xmath89 for @xmath90 ) as a function of the stellar rotation frequency @xmath22 . we also show the results derived by the approximate analytic method described earlier .
it is clear that there is a difference in the results of the order of 30% .
mainly , this is due to the fact that in one case we used the approximate newtonian form of the equation , while in the other case we used its exact relativistic form .
it is obvious that the torsional modes become secularly ( cfs ) unstable even for very slowly rotating relativistic stars . in a newtonian study @xcite torsional mode frequencies , measured in a co - rotating reference frame ,
can be described approximately by the relation ( [ eq : sigma ] ) , where @xmath91 . moreover
, an inertial observer will measure @xmath92 the cfs instability sets in when the frequency of the torsional mode for the inertial observer @xmath93 or , equivalently , when the phase velocity of the mode is equal to the rotational frequency @xmath94 , i.e. , when the critical rotation frequency of the star is given by @xmath95 the approximate results derived earlier ( [ eq : toy3c])-([eq : toy3d ] ) verify the newtonian results .
also the numerical results agree extremely well on the effect of rotation on the frequencies of the torsional modes , i.e. , @xmath96 with a typical error of the order of 2 - 5% , depending on the compactness of the star .
the perturbation equations have been solved for a number of different equations of state for the fluid core and the crust which are listed in table 1 of @xcite .
the overall picture is the same , i.e. , shows a 20 - 30% difference of the frequencies of the relativistic and the newtonian equations , and a 2 - 5% difference in the rate of the frequency as function of rotation ( @xmath97 ) .
the frequencies plotted in figure 1 are for a stellar model with a realistic eos a @xcite for the fluid core together with an eos for the crust given by @xcite ( model @xmath85 ) .
this stellar model has radius @xmath86 km @xmath87 and crust thickness @xmath88% .
in this article , we showed by means of numerical and semi - analytic methods that the torsional modes of rotating relativistic stars are subject to the cfs instability as suggested by @xcite for newtonian stars .
this instability might be only of academic interest since viscosity works against it , and therefore it will probably never prevail .
we would like to emphasize that the cfs instability does not operate in the up to now observed sgrs because they are very slowly rotating stars with periods of the order of seconds .
still it is possible that the torsional modes of the newly born neutron stars ( soon after they form a crust ) will be cfs unstable because rotation periods of the order of tenths or hundreds of hz are expected .
the instability of torsional modes in the crust and the unstable r - modes in the fluid core might work together in the direction of breaking or melting the crust .
if strong magnetic fields are present , then the accumulated stress might enhance the above scenario suggesting that the young neutron stars with strong magnetic fields will probably have a more frequent flare activity .
cfs type rotational instabilities of magnetic field modes will also be present in rotating neutron stars , and their effect in the stellar flare activity is part of an extension of this work . finally , rotation ( as well as the presence of strong magnetic fields ) will produce frequency shifts towards both lower and higher frequencies which poses extra difficulties for the identification of the various observed frequencies from sgrs .
we are grateful to h. sotani , n. stergioulas , s. yoshida and j.l .
friedman for helpful discussions .
this work is supported by the greek gsrt programs heracleitus and pythagoras ii and by the german science foundation ( dfg ) , via a sfb / tr7 . | we study the effects of rotation on the torsional modes of oscillating relativistic stars with a solid crust .
earlier works in newtonian theory provided estimates of the rotational corrections for the torsional modes and suggested that they should become cfs unstable , even for quite low rotation rates . in this work ,
we study the effect of rotation in the context of general relativity using elasticity theory and in the slow - rotation approximation .
we find that the newtonian picture does not change considerably .
the inclusion of relativistic effects leads only to quantitative corrections .
the degeneracy of modes for different values of @xmath0 is removed , and modes with @xmath1 are shifted towards zero frequencies and become secularly unstable at stellar rotational frequencies @xmath2 20 - 30 hz .
[ firstpage ] relativity methods : numerical stars : neutron stars : oscillations stars : rotation |
the first proposal of experimental search for @xmath1 violation effects in atoms was made almost 40 years ago @xcite .
the modern advanced experimental techniques realized in the experiments on electron s electric dipole moment ( edm ) are based on that idea and have the sensitivity of @xmath2 @xcite .
apart from the electron edm , experimental searches for the edm of the neutron are on - going @xcite .
current sensitivity allows for detection of electric dipole moment ( edm ) of the neutron at the level of @xmath3 @xmath4 . from theoretical point of view
, a non - zero edm could imply non - zero values for the qcd @xmath5-term as the latter can induce an edm @xcite . an attractive idea to enhance the experimental sensitivity to the electron edm
was proposed in @xcite . in an atom
, the electrons are moving in the electromagnetic field of the nucleus .
if the electron possess an edm , the electric coulomb field of the nucleus would induce a magnetic dipole moment , proportional to the electron edm and the electromagnetic field strength , leading to the magnetization of the sample .
since the electromagnetic field strength inside the nucleus can be several orders of magnitude larger than those achievable in the experiment , the effect of a non - zero electron edm is expected to be magnified .
due to chaotic relative orientation of atoms in a sample , these elementary magnetisations sum up to zero , therefore one would need a polarizing electric field to be applied to the sample , in order to observe the sample magnetization . at the same time , the authors of @xcite indicate that the effect of these atomic @xmath0-odd polarizabilities might interfer with the effect of nuclear edm , and has to be taken into account . if a non - zero neutron edm is discovered in the near future , further tests of our understanding of qcd and fundamental interactions will bring us to the study of the microscopic structure of the @xmath1-violating content of the nucleon , which can be described in terms of @xmath0-odd polarizabilities of the nucleon .
in this paper i address the following questions .
can the presence of these @xmath1-violating structure constants lead to a substancial interference with the measurement of the edm ?
how and whether they can be measured ?
the concept of the polarizabilities was first introduced in classical electrodynamics and characterizes the ability of the elementary charges within a given system to be displaced from their positions in the presence of an external electric field .
the electric dipole moment resulting from such a displacement is proportional to the strength of the applied field , and the coefficient of proportionality is called the ( electric dipole ) polarizability .
this constant quantitatively describes the forces that put the system together .
for instance , the atomic or molecular electric dipole polarizability is known to be of the order of the atom s or molecule s volume @xcite . instead , the nucleon electric dipole polarizability @xmath6 is only about 1/1000 of its volume , @xmath7 , which characterizes the strong forces that hold the proton together considerably stronger than the electromagnetic forces holding the electron in the atom .
the @xmath0-violating polarizabilities of the nucleon result from two pieces : the short range @xmath0-violating physics ( @xmath0-violation is generated well above the electroweak scale by unknown new physics ) responsible , for instance , for the qcd @xmath5-term , and the ( mostly ) long range pion physics which is quite analogous to the usual compton scattering case .
thus , the natural size of the nucleonic @xmath0-violating polarizabilities is expected to be ^t&~&10 ^ -3g_0fm^3 , where @xmath8 is the strength of the @xmath5-term and @xmath9 denotes a @xmath0-violating nucleon polarizability .
one can compare this to the estimates for the atomic @xmath1-violating polarizability @xmath10 @xcite , & ~&10 ^ -2a.u .
, + & ~&10 ^ -12a.u . , where atomic units were adopted , and @xmath11 is the edm of the neutron .
real compton scattering can be described , under the assumption of invariance under parity , charge conjugation and time reversal , by means of 6 structure dependent amplitudes @xmath12 , @xmath13 , with @xmath14 the c.m .
energy of the initial and outgoing photons and @xmath5 being the c.m .
scattering angle : & & t=^\ { a_1(,)()+ a_2 ( , ) ( ) ( ) . + & & + a_3(,)i ( ) + a_4(,)i ( ) ( ) + & & + a_5(,)i + & & + .a_6(,)i},[eq : comptonampl ] where @xmath15 stand for the polarization vector , direction of the initial ( final ) photon , and @xmath16 is the spin polarization of the nucleon . following refs .
@xcite , the functions @xmath12 can be expanded into a series in powers of ( small ) photon energy @xmath17 : @xmath18 \frac{e^2}{2m_n^2}\omega -\frac{e^2(e_n+2\kappa_n)}{8m_n^4}\cos\theta\omega^3 + 4\pi\left[\gamma_1^n-\left(\gamma_2^n+2\gamma_4^n\right)\right ] \cos\theta\omega^3 + o(\omega^4)\nn\\ a_4^{c.m.}(\omega,\theta)&=&-\frac{e^2\mu_n^2}{2m_n^2}\omega + 4\pi\gamma_2^n\omega^3+o(\omega^4)\nn\\ a_5^{c.m.}(\omega,\theta)&=&\frac{e^2\mu_n^2}{2m_n^2}\omega + 4\pi\gamma_4^n\omega^3+o(\omega^4)\nn\\ a_6^{c.m.}(\omega,\theta)&=&-\frac{e^2\mu_n}{2m_n^2}\omega + 4\pi\gamma_3^n\omega^3+o(\omega^4 ) \label{eq : comptonlex}\end{aligned}\ ] ] for each compton amplitude , the leading terms in the @xmath17 expansion are given by model - independent born contributions which are completely defined by the static properties of the nucleon as a spin-@xmath19 particle with the electric charge @xmath20 , anomalous magnetic moment @xmath21 ( magnetic moment @xmath22 ) and mass @xmath23 @xcite .
the higher order terms describe internal structure - dependent effects and are parametrized in terms of 6 polarizabilities .
two of them , @xmath24 and @xmath25 , are spin - independent electric and magnetic polarizabilities which enter the amplitude at @xmath26 and measure the deformation of the charge and magnetisation distributions in the presence of quasi - static electric @xmath27 and magnetizing @xmath28 external fields , = 4_e , = 4_e , with @xmath29 denoting the induced electric ( magnetic ) dipole moment .
the other four polarizabilities @xmath30 , @xmath31 describe the response of the spin - dependent distributions inside the nucleon to the quasi - static external field .
for example , the polarizabilities @xmath32 quantify the induced spin - dependent electric dipole moment in the external magnetic field , _
s^b&=&4_1 , + _ s^b&=&4_3 ( ) .
similarly , the polarizabilities @xmath33 quantify the magnetic dipole moment induced in the external electric field : _
s^e&=&4_2 ( ) . + _
quite in the spirit of the previous section , we will construct the compton amplitude which violates both parity and time - reversal ( and thus @xmath1 ) . & & t^=^\ { a_1^ ( , ) ( ) ( - ) .
+ & & + a_2^ ( , ) ( ) ( ) ( - ) + & & + a_3^(,)i(- ) + & & + a_4^(,)i + & & + a_5^(,)i + & & + .a_6^(,)i}[eq : comptoncpv ] in the above formula , i use the notation as close to the @xmath1-even compton scattering as possible .
all the structures we are interested in should contain at most one spin vector since one can always reduce the number of the @xmath35 s in a product through @xmath36 and the minimal power of photon momenta .
i do not change the dependence of the structures in eq.([eq : comptonampl ] ) on the photon polarization vectors , if possible , but modify the dependence on the photons momenta and nucleon spin such that the result is @xmath1-odd .
the first two structures in eq.([eq : comptoncpv ] ) are obtained by multiplying the corresponding @xmath37 , @xmath34 and @xmath0 conserving structures of eq.([eq : comptonampl ] ) by the explicitly @xmath34 and @xmath0 violating scalar @xmath38 . the third structure results from substituting the @xmath34-even , @xmath0-odd spin vector @xmath35 by @xmath34-odd , @xmath0-even vector @xmath39 . the fourth structure in eq.([eq : comptonampl ] )
does not allow for a modification to obtain a @xmath1-odd structure distinct from that coming with @xmath40 , therefore i use another structure instead .
the last two structures are obtained from the corresponding compton structures by changing the relative sign between two terms . in this way
, these combinations become @xmath0-odd but are @xmath34-even . for completeness ,
i also give here the amplitudes which are @xmath34-odd and @xmath0-even , t^&=&^\ { a_1^ ( , ) ( ) ( + ) .
+ & + & a_2^ ( , ) ( ) ( ) ( + ) + & + & a_3^(,)i(+ ) + & + & .a_4^(,)i}.[eq : comptonteven ] together with the @xmath37 , @xmath34 , and @xmath0 even compton amplitudes of eq.([eq : comptonampl ] ) and @xmath0-odd amplitudes of eq.([eq : comptoncpv ] ) , these amplitudes form the complete basis for real compton scattering with @xmath41 possible polarization states .
the next step is to calculate the born part of the @xmath0-violating compton scattering amplitude .
this born contribution corresponds to an elastic absorption ( emission ) of the initial ( final ) photon by either the initial or final nucleon with the single nucleon propagating in the intermediate state .
for such an amplitude to violate time - reversal one needs @xmath0-violating photon coupling to the nucleon .
this can be arranged by including a non - zero edm term into the electromagnetic vertex of the nucleon , ^(q)=ewhere the dimensionless @xmath42 is the electric dipole moment ( edm ) of the nucleon measured in units of the nuclear magneton @xmath43 , and the index @xmath44 indicates whether the nucleon is the proton or the neutron , respectively . in the following , i will use the edm in the usual units , d_n=_n .
the direct calculation of the diagrams in fig.[fig : born ] leads to the following results : & & a_1^=(2e_n+_n)[eq : toddlex ] + & & + 4 ^ 3+o(^4 ) + & & a_2^=-(2e_n+_n)^2 + & & -^3[e_n-_n-2(2e_n+_n ) ] + & & -4_2^t^3+o(^4 ) + & & a_3^=_n^2 + 4_3^t^2 + & & -^3 [ e_n-_n+(2_n+_n)]+o(^4 ) + & & a_4^=(e_n-_n ) + & & -[_n+(3e_n-_n)]^2 + 4_4^t^3 + & & + [ _ n - e_n+3_n-2e_n^2]^3 + o(^4),where the four constants were introduced , the @xmath0-odd , @xmath34-odd polarizabilities of the nucleon @xmath45 , @xmath31 . as in the case of @xmath34-even @xmath0-even compton scattering , the spin independent polarizability parametrizes the term which is quadratic in photon energy , while the spin dependent ones come at order @xmath46 .
one should note that the fact that the amplitudes do not possess definite crossing symmetry ( @xmath47 ) is due to the use of the c.m .
frame which makes the direct and crossed channels asymmetric by imposing that the nucleon in the intermediate state is at rest in the @xmath48-channel , @xmath49 , while it is not the case for the @xmath50-channel , @xmath51 .
the polarizabilities introduced above can be interpreted as the deformation of the system under the influence of the external electromagnetic field , as it was done for the usual compton scattering .
the polarizability @xmath52 quantifies the electric dipole moment induced by the gradient of the external electric field in the direction of the spin , and similarly for the polarizability @xmath53 _ s^e&=&4_1^t ( ) , + _ s^b&=&4_2^t ( ) .
furthermore , there is one polarizability which characterizes the electric dipole moment induced by the external magnetic field ( without spin ) , ^e&=&4_3^t , and finally , another polarizability which quantifies the electric dipole moment induced by the gradient of the projection of the external electric field onto the nucleon spin , _ s^e&=&4_4^t ( ) .
the qcd @xmath5-term leads to a @xmath34-odd @xmath0-odd coupling of the pion to the nucleon , l_&=&-g_0|nn .
the contribution to the polarizability @xmath54 arises due to neutral pion exchange in the @xmath55-channel , as shown in fig .
[ fig : pi0pole ] , for which one needs the anomalous @xmath56 vertex provided by wess - zumino - witten lagrangian , l_^0^wzw&=&- _ f^f^^0 , with @xmath57 the pion decay constant . furthermore , we will need the usual @xmath1-conserving pion - nucleon lagrangian .
the lowest order chpt lagrangian in the heavy baryon formalism is well known and we refer the reader to ref .
@xcite for the details .
the representative diagrams at one loop are shown in fig .
[ fig : pion ] . keeping the leading terms in @xmath58
only , we obtain the following results for the @xmath0-odd polarizabilities : _ 1^t&=&-g_0 + _ 2^t&=&g_0 + _ 3^t&=&-g_0 +
_ 4^t&=&0 + o(/m_n ) , [ eq : polhbchpt ] where @xmath59 stands for the axial coupling of the nucleon .
we note that that the polarizability @xmath60 arises due to nucleon recoil effects and is thus of order @xmath58 .
in principle , one might try to measure these new structure constants of the nucleon using the well established experimental techniques used in the edm - type experiments . in such experiments ,
one measures the difference in the precession frequency of the spin in the external magnetic and electric fields depending on the field orientation .
however , these experiments use static fields . under these conditions ,
compton effect is undetectable , as the corresponding compton frequency shift is @xmath61 . as one can see from eqs.([eq : toddlex],[eq : polhbchpt ] ) , the polarizabilities contributions arise as corrections in powers of @xmath62 to the leading born contributions . to see these corrections
one has to go to photon energies ( frequencies ) comparable to the pion mass .
such experimental conditions may be accomplished in a compton scattering experiment .
one of the possibilities would be to scatter circularly polarized photons off unpolarized nucleon target .
one can flip the polarization of the photon and without detecting the polarization state of the photon in the final state , measure the difference in the signal .
the general expression of such a single spin asymmetry in terms of the amplitudes defined above is somewhat lengthy . making use of the lex of the pctc compton amplitude for the proton
, one has = - , [ eq : ssa ] where the lex of the amplitudes @xmath63 and @xmath64 are given in eqs .
( [ eq : comptonlex ] ) and ( [ eq : toddlex ] ) , while the lex of the @xmath34-odd amplitude @xmath65 can be found in ref .
@xcite .
one notices that this asymmetry obtains contributions both from pvtc and pvtv amplitudes .
the contribution of the former originates mainly from the @xmath34-odd combination of compton helicity amplitudes @xmath66 which does not violate time reversal .
the main source of the latter of the two is the combination of the backward - dominant helicity amplitudes @xmath67 which is non - zero in the presence of both @xmath34- and @xmath0-violation .
the parity - violating asymmetry was first considered in @xcite and it is expected to be of order @xmath68 at forward angles . if @xmath1-violation is due to a non - zero qcd @xmath5-term that is very tightly constrained by the experimental limit on edm , the expected value of the single spin asymmetry in eq .
( [ eq : ssa ] ) is a^cp10 ^ -11 [ eq : estimate ] for @xmath69 mev photons and very backward angles .
the effect is quadratic in the photon energy @xmath17 .
although this effect is tiny as compared to the parity - violating contribution , by going to very backward angles this latter is highly suppressed due to the @xmath70 factor in front .
it furthermore has a cubic dependence on the photon energy @xcite .
another important background is represented by the analyzing power of compton scattering .
experimentally , it is impossible to achieve a 100% circular polarization .
the remaining linear polarization component can lead to a similar @xmath0-odd observable that does not require @xmath0-violation , but arises from the final state interaction .
one has for the compton cross section @xcite =( ) ( 1+p2 ) with @xmath71 the usual unpolarized differential cross section and @xmath72 the angle between the linear photon polarization direction and the reaction plane .
finally , @xmath73 represents the degree of linear polarization , and @xmath34 the analyzing power .
the analyzing power arises as an interference between the purely real leading order compton amplitude ( below the pion threshold ) , and the imaginary part of the next - to - leading order compton amplitude , as shown in fig .
[ fig : compton_an ] . inserting the leading in lex thomson term in place of the blobs in the figure ,
we obtain : = _ em+o(^2 ) where the energy dependent factor is due to the phase space .
the analyzing power @xmath34 is then obtained from interference terms like p ~ ~_em + o(^3 ) + , the above formula gives an adequate result for the leading contribution at low energies , and the corrections due to imaginary parts of other amplitudes arise at the order @xmath46 .
assuming the degree of linear polarization in the photon beam to be of order 1% , and going to very backward angles , so that @xmath74 , we see that such qed final state interactions can lead to asymmetries of the order a^fsi&~&0.010.1 2 + & ~&10 ^ -72 as one can see , the analyzing power represents a very substantial background for a measurement of a @xmath1-violating compton scattering , as well as that of the parity - violating compton scattering as proposed in @xcite .
this background can be distinguished experimentally by measuring the complete azimuthal angle dependence of the considered single spin asymmetry .
the final result is given by a cosine - modulated fsi contribution plus a constant term in @xmath72 that is the @xmath34- and @xmath1-odd pieces .
these estimates indicate that whenever @xmath1-violation in compton scattering is due to mechanisms that can directly contribute to the edm , as well , the corresponding asymmetries can not exceed roughly @xmath75 at energies just below the pion threshold .
i will next examine the upper limit constraint on @xmath1-violation in spin - independent compton scattering . as a possible ,
although exotic scenario , one can assume that an unknown source of @xmath1-violation may exist , that generates the @xmath0-odd compton amplitude but is for some reason forbidden to show up in just one photon vertex ( for instance , the spin - independent @xmath1-odd term is not present in the one photon coupling , but arises in compton scattering ) .
an example of such a lagrangian is a dimension-7 operator l = c ( ) _ f^f^ |nn , where @xmath76 represents the scale of the unknown @xmath1-violating new physics that is integrated out and is replaced by the above effective vertex , and @xmath77 the corresponding wilson coefficient .
this operator generates the non - born part of the amplitude @xmath78 only , since the edm is supposed not to obtain contribution from this mechanism , 4_3^= [ eq : a3np ] where the wilson coefficient should be taken at low energy .
the edm then arises at one - loop level as a qed radiative correction , as shown in fig .
[ fig : npedm ] ( d_n)~ such that we obtain a limit on this new physics contributions from the edm , c()10 ^ -10 the loop calculation contains a quadratic divergence due to the neglection of the vertex structure within an effective field theory treatment . to obtain
an order - of - magnitude estimate of this loop contribution , i use the `` naive dimensional analysis '' method @xcite that is based on the dimensional regularization approach that ensures that the only mass scales arising in such a calculation are physical particle masses . the quadratic divergence
itself should cancel exactly , once one specifies the underlying theory that is renormalizable .
unless there exists a symmetry that enforces the exact cancellation , at low photon energies this cancellation should not occur at 100% level .
then , the naive dimensional analysis estimate is still adequate . combining this limit with eqs .
( [ eq : a3np ] ) and ( [ eq : ssa ] ) we arrive to the following upper limit for the @xmath1-violating asymmetry generated by the new physics : a^np10 ^ -11 to arrive to this result i neglected the running of the wilson coefficient @xmath77 .
this running may indeed be substantial , but one would rather expect a logarithmical , and not quadratic running , therefore the @xmath79 will be dominant for the dependence of the above limit on the new physics scale .
an accurate treatment can indeed change the estimate somewhat , and i leave this investigation for a future work , as this calculation goes beyond the scope of the present article . since the pion is the lightest hadronic state
, this result means that the calculation of the @xmath1-odd polarizabilities of the nucleon in the chpt with pions of eq .
( [ eq : polhbchpt ] ) represents the dominant contribution due to the @xmath79 suppression of heavier particles contributions .
in other words , if a substantial @xmath1-violating mechanism in nucleon spin - independent compton scattering exists , it should involve some light particles , like pions or eta s for instance , to be observable at the same level as the @xmath1-violating @xmath80 coupling contribution considered here in greater detail .
in summary , i considered compton scattering in the presence of @xmath1-violation .
i defined the compton amplitudes that possess this feature and applied the low energy theorem to this amplitude .
after the separation of the born contribution that is defined by static properties of the nucleon and its edm , i parametrized the unknown non - born part by introducing the model - independent structure constants , the nucleon @xmath0-odd polarizabilities .
i calculated these polarizabilities in the assumption that the @xmath1-violation is due to non - zero qcd @xmath5-term that generates a @xmath1-violating @xmath80 coupling .
i furthermore proposed an observable in compton scattering that is directly sensitive to @xmath1-violation that is a single - spin asymmetry with circularly polarized photon beam . within the model for the @xmath0-odd polarizabilities used in this work
, i estimated such an asymmetry to be of order @xmath81 , @xmath17 being the photon energy , and found that it is picked at backward angles .
the above limit originates from the experimentaql limits on edm translated into the strength of the qcd @xmath5-term .
experimentally , the @xmath1-odd asymmetry is always accompanied by the @xmath34-violation contribution to compton scattering which was estimated in the literature as @xmath82 and is suppressed at backward angles as @xmath70 .
i also considered a background process that involves the linear polarization component in the photon beam which can lead to azimuthal angle - dependent asymmetry due to final state interactions that generate an imaginary part of compton scattering . at very backward angles ,
such fsi can lead to asymmetries of order @xmath83 , with @xmath72 the angle between the direction of the linear polarization and the scattering plane .
the author is grateful to michael j. ramsey - musolf , petr vogel , brad filippone , brad plaster , charles horowitz and michael snow for enlightening discussions .
the work was supported under doe - fg02 - 05er 41361 and under us nsf grant phy 0555232 .
regan , e.d .
commins , c.j .
schmidt , d. de mille , phys .
. lett . * 88 * , 071805 ( 2002 ) ; j.j .
hudson , b.e .
sauer , m.r .
tarbutt , phys .
lett . * 89 * , 023003 ( 2002 ) ; d. de mille et al , phys . rev . * a 61 * , 052507 ( 2002 ) ; | i consider compton scattering off the nucleon in the presence of @xmath0 violation .
i construct the compton tensor which possesses these features and consider low energy expansion ( lex ) of the corresponding amplitudes .
it allows to separate out the born contribution which only depends on the static properties of the nucleon , such as the electric charge , the mass , the magnetic moment , and the electric dipole moment ( edm ) .
i introduce new structure constants , the @xmath0-odd nucleon polarizabilities which parametrize the unknown non - born part .
these constants describe the response of the @xmath0-violating content of the nucleon to the external quasistatic electromagnetic field . as an estimate
, i provide a hbchpt calculation for these new polarizabilities and discuss the implications for the experiment . |
the prompt emission detected from gamma - ray bursts ( grbs ) is believed to originate at large distances from the central engine , from within an ultrarelativistic outflow @xcite .
this ultra - relativistic motion is necessary to avoid strong @xmath0 annihilation , a signature that is not observed ( see , e.g. , * ? ? ?
thermal emission is naturally expected in such a scenario . indeed , since the densities at the base of the relativistic flow are very large , the medium is optically thick to radiation owing to thomson scattering by entrained electrons .
the optical depth decreases during the relativistic expansion and the outflow eventually becomes transparent for its own radiation , at the photospheric radius .
any internal energy that is still carried out by the flow can be radiated at the photosphere and will be observed as a thermal component in the prompt spectrum . this expected photospheric emission in grb spectra
was early suggested on such theoretical grounds by @xcite , @xcite , and @xcite , among others .
the non - thermal component observed in the spectrum has to be produced by another mechanism in the optically thin region , i.e. , well above the photosphere . due to the ultra - relativistic motion
, this difference in the radius of the emission implies a delay between the observation of the two components that is usually small compared to the typical duration of a long grb , and is also small compared to the typical duration of time intervals used for time - dependent spectroscopic analysis .
the thermal and non - thermal components should then appear superimposed for the observer ( e.g. * ? ? ?
@xcite pointed out that in the standard fireball model , the photospheric component can easily be dominant in the spectrum if the efficiency @xmath1 of the mechanism responsible for the non - thermal emission is only moderate ( @xmath2 ) .
observationally , @xcite , @xcite , and @xcite argued that a photospheric component is present in cgro batse data . the limited energy range provided by batse ( 20 - 2000 kev ) , however , hampered the possibility of unambiguously identifying the emission process .
since the launch of fermi in 2008 , the combination of the gamma ray burst monitor ( gbm ) and the large area telescope ( lat ) provides an unprecedented energy range for grb spectroscopy , and the identification of the emission processes responsible for the gamma - ray prompt emission may become a reality .
gbm alone covers a wider energy range than its predecessor batse , and the design of its data enables finer resolution spectroscopy .
this allows better constraints on spectral fits , with increasingly complex models .
grb energy spectra in the kev - mev energy range are usually well represented by the band function @xcite , two power - laws , smoothly joined and parameterized by e@xmath3 which represents the energy at which peak power is radiated @xcite .
the value of the low - energy power - law index , @xmath4 , is higher than the value of the high - energy power - law index , @xmath5 , and the parameter e@xmath3 of the band function for grbs generally appears to follow predictable trends with time and flux level @xcite .
it is , however , an empirical function rather than a physically - motivated model .
the meaning of the parameters in the context of emission and transport mechanisms taking place in grbs is not well understood , but is generally believed to represent the non - thermal emission from accelerated charged particles . in section 1 ,
we describe our observational results consisting of a grb prompt - emission spectrum best fit with the combination of a thermal component and a standard band function . in section 2
, we use these results to constrain the origin of the energy released in the grb jet .
gbm is composed of 12 sodium iodide ( nai ) detectors covering an energy range from 8 kev to 1 mev and two bismuth germanate ( bgo ) detectors sensitive between 200 kev and 40 mev @xcite .
the instrument triggered on 24 july 2010 , at t@xmath6=00:42:05.992 ut on the very bright grb _ 100724_b @xcite .
the event was also seen at higher energies in the fermi large area telescope ( lat ) @xcite .
the most precise position for the direction of the burst is the intersection of the interplanetary network annulus obtained using gbm , konus - wind @xcite , and messenger data with the 90% lat confidence level location error box , and is a strip of sky centered on ra = 118.8@xmath7 and dec = 75.8@xmath7 which is 1.2@xmath7 long and 0.2@xmath7 wide ( k.hurley and v.palshin , private communication ) .
figure [ fig : figure1 ] ( top two panels ) shows the gbm light curve of grb _ 100724_b in two energy bands .
multiple peaks of varying intensity are superimposed on a pre - trigger plateau , with a decaying tail that is detected over 200 s from t@xmath8 .
( in blue ) and the bb temperature kt ( in red ) over the duration of the burst .
the vertical dashed lines indicate the period used in the time - integrated analysis .
[ fig : figure1 ] ] we simultaneously fit the spectral data of the nai detectors with a source angle less than 60@xmath7 ( nais 0 , 1 , 2 , 3 , and 5 ) and the data of the brightest bgo detector ( bgo 0 ) using the analysis package rmfit 3.3rc8 . an effective area correction is applied between each of the nais and bgo 0 during the fit process .
this correction is used to handle possible discrepancies between the flux in the detectors due to the choice of the model to generate the instrument responses for instance .
we performed a time - integrated spectral analysis over the main part of the burst ( t@xmath6 - 1.024s to t@xmath6 + 83.969s ) using the band function .
the band parameters are in part fairly typical of the ensemble of grbs , with @xmath9 and e@xmath3 = @xmath10 kev @xcite . however , with an index @xmath11 , the high - energy power law systematically overshoots the observed flux above 1 mev in bgo , as can be seen by the fit residuals in figure [ fig : figure2 ] ( top two panels ) , which also indicate systematic patterns at low energy .
this suggests a simple band function does not adequately represent the spectrum of this burst .
we identify the best shape to fit the above - mentioned spectral deviation by fitting the same data simultaneously with a band function combined with each of the following models : single power- law ( pl ) , black body ( bb ) , band function , power law with exponential cut off ( comp " for comptonized model ) , and gaussian .
we select the best model by choosing the fit with the lowest castor c - stat value ( later c - stat ) .
c - stat differs from the poisson likelihood statistic by an offset which is a constant for a particular dataset .
table [ table : spectraldeviationfit ] shows the results of these fits .
the effective area correction described above is on the order of a few percent and does not change the c - stat for each fit more than a few units , nor does it change the value of the parameters resulting from the fit .
[ cols= " < , < , < , < , < , < , < , < , < , < , < , < , < , < " , ] while spectral deviations from the standard band function were previously identified in the form of an additional pl to the band function sometimes extending from the lower energy in the gbm to the higher energy in the lat @xcite , in the case of grb _ 100724_b a pl spectral component does not improve on the band - only fit and an additional bb component to the band function is the best model to fit the spectral deviation . an equal c - stat is obtained for band+band and band+compt , but with @xmath4 close to + 1 for the additional band and compt functions , and a very low value for @xmath5 ( only constrained as an upper limit , below -5 ) for the extra band function , the band and compt functions can be interpreted as a planck function . even with more parameters ,
the additional band and compt functions resemble a bb component , reinforcing band+bb as the best combination .
additional models were tried , such as a log - parabola function @xcite , but the results were highly disfavored , and we exclude them from table 1 .
figure [ fig : figure2 ] ( bottom two panels ) shows the bb contribution below e@xmath3 .
compared to the band - only fit , e@xmath3 is shifted towards higher energy to @xmath12 kev and @xmath5 is lower with a value of @xmath13 .
this index is consistent with the flux detected above 1 mev , and the spectrum seen in the lat @xcite at higher energies .
@xmath4 is also significantly lowered to @xmath14 .
while the simultaneous fit of all the selected detectors provides the best constraints on the two spectral components , fits with band+bb to combinations of individual nai detectors with bgo 0 result in similar parameter values and offer significant improvement over the band - only fit .
this provides a check that the bb component is real and not introduced by effects such as detector deadtime or spectral distortions that would affect each detector in a different way depending on the angle of the detector to the source . to verify that the improvement in the fit obtained by adding a bb component to the band function is not a statistical fluctuation , we generated 20,000 synthetic spectra for each selected detector . for the simulations we used the parameters from the fit performed with the band - only function , which we take as the null hypothesis
. to create the simulated spectra , for each detector the real background is added to the source spectrum model and poissonian fluctuations are applied to the sum
all the detectors are then fit simultaneously with both band and band+bb , and their c - stat are compared .
none of the 20,000 simulated spectra give a difference larger than 45 units of c - stat ( [ band][band+bb ] ) , while in the real data , this difference is 95 units , corresponding to a probability lower than @xmath15 that the bb excess is due to statistical fluctuations . to study the evolution of the spectral components , 22 time intervals were devised by requiring that each interval produce a band+bb spectral fit with well - constrained band function parameters , while attempting to separate the peaks and valleys of the light curve so that the spectral fit parameters can be tracked with burst flux as well as with time .
the bottom panel of figure [ fig : figure1 ] exhibits the evolution of e@xmath3 and bb temperature , kt , through these intervals .
we notice that e@xmath3 tracks the lightcurve and globally decreases over time .
the bb component is detected throughout the burst , and its temperature shows weak correlation with e@xmath3 .
the significance of this correlation is difficult to assess , mostly because the variation in temperature is small .
overall , it appears the temperature is quite stable , with figure [ fig : figure3 ] showing more clearly the small scatter in kt .
can be fit by a gaussian distribution with a mean of @xmath16 , and a @xmath17 standard deviation of @xmath18 .
[ fig : figure3 ] ]
with a fluence of @xmath19 erg @xmath20 measured in 85 sec from [email protected] between 8 kev and 40 mev , grb _ 100724_b is the most intense grb detected by gbm over this energy range through 2010 september .
combined with the broad energy range of the gbm , this allows for accurate modeling of its energy spectrum even with this complex model .
previous observational results regarding thermal components in grbs were ambiguous and some were limited to individual fine time - slices rather than a spectral fit over the entire emission period .
some studies showing bb fits did not demonstrate that the bb fit was statistically preferred to a simple non - thermal component @xcite .
other analyses found bb+pl spectra for isolated portions of selected grbs , raising the possibility that these spectra are actually adequately fit with a standard band function but that due to a weak signal in small time slices and extreme parameters for the band function , a bb shape is competitive with the band function @xcite .
the non - thermal component fit with a single power law suggested a break well beyond the common e@xmath3 values , and the bb temperature and its variations intriguingly matched those of a typical e@xmath3 . despite the broader energy range of rhessi grb observations ,
one analysis found difficulties in fitting combined thermal plus non - thermal models @xcite .
we find here that the joint bb plus non - thermal ( band ) fit is highly statistically preferred so that in simultaneously detecting both components we are confident in their correct identification .
time - resolved spectroscopy of grb _ 100724_b reveals that this bb component is seen throughout the burst and does nt evolve much over time , while the non - thermal component follows the typical variations @xcite .
the consistency of the mean kt value with the temperature obtained in the time - integrated spectral fit , combined with the detection of the bb component throughout the burst , strengthen the case for an underlying thermal component in the gamma - ray emission seen from grb _ 100724_b and show that the presence of the bb in the time - integrated spectrum can not be attributed to spectral evolution of the band function during the burst .
e@xmath3 varies substantially , from @xmath21 to @xmath22 kev . at the same time
, the thermal component remains relatively steady with the temperature varying only modestly between 30 and 50 kev as suggested by a @xmath23 dependence expected from a bb component .
time - averaged values of the temperature and the flux of the thermal component , and of the ratio of this flux over the total gamma - ray flux are @xmath24 kev , @xmath25 erg / s/@xmath26 and @xmath27 . in the standard fireball model
these observables allow determination of the physical properties of the outflow and its photosphere .
owing to the imprecise and delayed localization of grb _ 100724_b , optical follow up to determine the distance to the source was impossible .
for this reason the temperature of the bb can be translated into a real source temperature only as a function of source distance .
we assume in the following argument a typical redshift @xmath28 .
we find that the lorentz factor is @xmath29 , the photospheric radius is @xmath30 and the radius at the base of the flow is @xmath31 @xcite . here
@xmath32 is a geometrical factor of order unity and @xmath1 is the efficiency of the mechanism responsible for the non - thermal emission . with an extreme efficiency @xmath33 ,
these estimates are in good agreement with the typical values expected in the fireball model .
the dependence on redshift is not strong : at z=3 ( resp .
8) , @xmath34 ( resp .
1290 ) , @xmath35 ( resp .
@xmath36 ) , and @xmath37 ( resp .
@xmath38 ) . using more realistic values for the efficiency , the radius @xmath39 is the most altered , with @xmath40 for @xmath41 .
such small values are puzzling .
if the central engine is a rotating black hole , as in the popular collapsar model for long grbs @xcite , with a minimal mass in the range 5 10 @xmath42 , such radii are smaller than the typical value expected for the innermost stable orbit , from 44 - 89 km for a non - rotating black hole to 22 - 43 km for a highly - rotating black hole having a spin @xmath43 .
these results for the time - integrated spectrum imply a small @xmath44 or a very large efficiency and the constraint is even stronger in some time bins . we conclude that observations of grb _ 100724_b require either a very high efficiency for the non - thermal process , or a very small size of the region at the base of the flow , both of which are quite challenging for the standard fireball model , if not excluding it . a simple solution to this discrepancy between the standard fireball model and the observations is to assume that the initial energy release by the central engine is not purely thermal , but that the flow is highly magnetized close to the source @xcite .
the magnetization @xmath45 is the ratio of the poynting flux over the power ( thermal + kinetic ) carried by the baryons .
if no magnetic dissipation occurs below the photosphere , the efficiency @xmath1 in the estimates of @xmath46 , @xmath47 , and @xmath39 above should be replaced by @xmath48 .
a magnetization @xmath49 will therefore reconcile the observations with physically acceptable values for the radius at the base of the flow and the efficiency of the mechanism responsible for the non - thermal emission .
a similar conclusion is reached for scenarios where magnetic dissipation occurs early and contributes efficiently to the acceleration of the jet .
however , the appearance of a low intensity thermal component in the spectrum probably excludes the most extreme version of the magnetized outflow scenario , where the energy is released by the central engine as a pure poynting flux ( @xmath50 ) .
we have shown that the simultaneous presence of thermal and non - thermal components to the spectra of grb _ 100724_b is statistically preferred .
although the non - thermal component is dominant , the black body flux is well within the gbm sensitivity .
deviations from the band function may be measurable in less fluent bursts or in bursts where the thermal component is less prominent , providing that the black body component lies in the band pass of the instrument and its peak in energy is distinguishable from @xmath51 . if the presence of an unresolved thermal component in other bursts modifies the band function parameters in the same sense as the band - only fit for grb _ 100724_b , then we might expect a systematic bias yielding values of @xmath4 and @xmath5 that are higher ( harder ) than in the true non - thermal component .
two important consequences of this bias are that the perceived violation of the synchrotron limit that disallows values @xmath52 ( for slow - cooling electrons ) and @xmath53 ( for fast - cooling electrons ) may not be as common as suggested by @xcite and @xcite , and that the relatively low rate of bursts detected by the lat compared to the predictions of @xcite and the observations in abdo et al .
( in preparation ) based on extrapolations of @xmath5 from lower energies might be explained by this bias in @xmath5 , a possibility suggested also by @xcite .
our observations provide strong evidence for the presence of a photospheric spectral component , long suspected to exist in the standard fireball model .
in addition , our results require implausible parameters for the standard baryonic fireball model and therefore favor a substantial magnetic component to the outflow .
the gbm project is supported by the german bundesministerium fr wirtschaft und technologie ( bmwi ) via the deutsches zentrum fr luft- und raumfahrt ( dlr ) under the contract numbers 50 qv 0301 and 50 og 0502 . , f. , axelsson , m. , zhang , b. b. , mcglynn , s. , peer , a. , lundman , c. , larsson , s. , battelino , m. , zhang , b. , bissaldi , e. , bregeon , j. , briggs , m. s. , chiang , j. , de palma , f. , guiriec , s. , larsson , j. , longo , f. , mcbreen , s. , omodei , n. , petrosian , v. , preece , r. , and van der horst , a. j. : 2010 , , l172 | observations of grb _ 100724b _ with the fermi gamma - ray burst monitor ( gbm ) find that the spectrum is dominated by the typical band functional form , which is usually taken to represent a non - thermal emission component , but also includes a statistically highly significant thermal spectral contribution . the simultaneous observation of the thermal and non - thermal components allows us to confidently identify the two emission components .
the fact that these seem to vary independently favors the idea that the thermal component is of photospheric origin while the dominant non - thermal emission occurs at larger radii .
our results imply either a very high efficiency for the non - thermal process , or a very small size of the region at the base of the flow , both quite challenging for the standard fireball model .
these problems are resolved if the jet is initially highly magnetized and has a substantial poynting flux . |
various observational and theoretical studies have been carried out which aimed at the investigation of possible ionization mechanisms of the diffuse ionized gas ( dig , also called interstellar froth @xcite ) in the milky way ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) as well as in external galaxies ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . as a main result of these studies , photoionization by ob - stars turned out to be the main ionization mechanism of this gas .
it soon became obvious that existing photoionization models can not predict the observed dig spectrum correctly . in order to reconcile theory with observations ,
the current strategy is to include additional heating and ionization mechanisms , such as shock ionization , turbulence or magnetic reconnection ( to name only a few ) .
even this did nt solve the inconsistencies completely .
therefore , we decided to analyze the existing models critically according to the radiative transfer and the geometry they consider
. the most commonly used codes to model dig are cloudy @xcite and the one written by mathis @xcite with its two modifications presented by @xcite and @xcite .
in addition , an implementation of ferland s cloudy has been published by @xcite for the edge - on galaxy ngc891 .
there are two severe shortcomings common to all these codes : ( 1 ) the problem of the radiative transfer is solved analytically which requires unrealistic approximations ( ots , owo , etc . ) and underestimate the _ diffuse _ ionizing radiation field .
( 2 ) the geometry assumed by these codes is spherically symmetric which is not realistic to model galaxies with three dimensional stellar disks and extended gaseous halos .
hence , photoionization by ob - stars as the main heat source of the dig should not be abandoned , unless realistic pure photoionization models have proven to be unsuccessful in reproducing the observed line ratios .
since the radiative transfer can only be solved correctly by means of statistical methods , a new self - consistent 3d monte carlo ( mc ) code , called _ soap _ ( simulations of astrophysical plasmas ) , has been developed .
mc techniques are a proven means of calculating the radiative transfer in an inhomogeneous medium . as an example
, @xcite and @xcite developed a self - consistent mc code which successfully models the multi - line transfer in the wind of wolf - rayet and other early - type stars .
this model was modified by @xcite in order to self - consistently simulate spherical steady state gaseous nebulae with arbitrary densities .
a comparison of their results with those produced by independent analytical photoionization codes for the standard _
meudon h@xmath0-region _
@xcite evidenced the correctness of the mc model code . based on this work
, we extended the 1d model grid to 3d , introduced a 3d stellar distribution , assigned individual gas densities to every grid point , and adapted the radiative transfer , correspondingly . as a result , new comprehensive data , such as the ionization structure of h
, he , and heavier ions at high distances above the disk plane as well as first line ratios , could be obtained .
we are aware that the interaction of the radiation field with dust particles is currently neglected .
however , contributions from dust will be modeled as soon as the effects of pure photoionization for different parameter settings are investigated and can be disentangled from a dusty plasma .
before the modeling of gaseous galaxy halos was started , we made sure that the standard _
meudon h@xmath0-region _ results obtained with the och - model could be fully reproduced by _
soap_. a detailed description of the code is given in @xcite . since we want to investigate the ionization conditions mainly along the @xmath1 direction , cartesian coordinates are used rather than spherical or cylindrical ones .
our model galaxy extends from the center 15kpc in each direction .
the disk is assumed to have a radial extend of 9kpc and a height of 1.5kpc .
the whole geometry is subdivided into 1400 cells , with 10 cells in @xmath2 and @xmath3-direction , respectively , and 14 cells along the @xmath1-axis . except for the disk region
all cells have constant lengths of 1.5kpc and identical volumes . within the disk , there are 5 vertical layers with heights of 100 , 100 , 200 , 400 , and 700pc .
stars are assumed to be present only in a thin disk which occupy the innermost 100pc .
the remaining layers are filled up with gas of pre - selected density and metal abundance . above the disk ,
gas densities are significantly lower and drop exponentially as a function of @xmath1 , the distance above the disk .
although the selected geometry underestimates the radial dimensions of a `` real '' spiral galaxy by a factor of 2 and is only sensitive to changes of the radiation field on scales of 1.5kpc the chosen setup represents the best compromise between spatial resolution and computing time .
_ soap _ has been developed to primarily investigate dig properties in edge - on galaxies .
this would be similar to spectroscopic observations where the slit is aligned along the minor axis of a galaxy . here ,
a typical offset of @xmath4kpc from the galactic center was assumed . at this position , all relevant emission coefficients from recombination and forbidden
lines are summed up and volume averaged along the @xmath3-axis for each of the 14 @xmath1-layers , separately .
a model halo with exponentially decreasing densities is simulated assuming the relative elemental abundances of the _ meudon h@xmath0-region _ ( h=1.0 , he=0.1 , c=2.2@xmath5 , n=4.0@xmath6 , o=3.3@xmath5 , ne=5.0@xmath6 , and s=9.0@xmath7 ) .
densities in the lowermost layer start at @xmath8 and drop to @xmath9@xmath10 at @xmath11kpc .
given the symmetry of the problem , it is sufficient to model only one octant of the galaxy . within the volume occupied by the disk , approximately 640 o6 stars ( blackbody ) are distributed in a thin ( 100pc ) stellar layer with a radius of 9.0kpc .
a typical model run consists of @xmath12150 iterations with 10@xmath13 photons per iteration .
we made sure that the solution remained stable even after 450 model runs . in figures
[ f1 ] and [ f2 ] ionization fractions are plotted for the most important elements ( h , he , o , n , and s ) as a function of @xmath3 and @xmath1 . for the sake of clarity ,
only data of 4 disk and 4 halo layers is shown , where each symbol / line addresses a certain distance above the galactic plane .
the ionization structures for h@xmath14 and o@xmath14 are nearly identical due to very similar ionization potentials . for @xmath15kpc ( radial direction )
low fractions indicate that most of the hydrogen and oxygen is ionized . at distances
@xmath16kpc from the center the disk ends and the ionization structures change , implying that h and o are now almost completely neutral .
similar effects are well known in normal hii - regions ( e.g. , * ? ? ?
* ; * ? ? ?
the steep increase between 7 and 9kpc indicates that all photons are absorbed and can not ionize the gas at larger @xmath3-distances .
this is the equivalent case to the strmgren radius which determines the outer boundary of a classical hii - region .
between @xmath17kpc there is a local minimum in both ionization fractions which demonstrates that h and o are partly ionized .
the reason is most likely due to ionizing photons which are created in the halo and are re - emitted at angles that allow them to reach the galactic plane .
in other words , the diffuse radiation field is responsible for the ionization of the shadowed outer - disk region .
if one moves along the minor axis the corresponding ionization fractions of @xmath18 and @xmath19 become more and more independent from the disk . beyond 9.0kpc the ratios
drop continuously , implying that _ with decreasing densities of the dig , the fraction of ionized gas gets larger_. the corresponding curves for higher ionization stages of o , s , and n also support this finding .
this surprising result is explainable as follows : stellar and diffuse ionizing photons that traversed the disk get dilute as they expand into the halo .
decreasing halo densities lead to decreasing opacities and enable photons to reach the outer halo border .
this in turn implies low recombination rates and higher temperatures in the halo ( cf .
lower left panel of fig .
typical recombination times for hydrogen at gas temperatures of 8000k and densities of about @xmath20@xmath10 yield : @xmath21yr . apparently , the geometric dilution of the radiation field is not a severe obstruction for the photons to ionize nearly all atoms within a given cell volume .
once the gas is ionized , it remains ionized for about 100 million years .
if one considers the constant stream of ionizing photons from the disk it appears likely that recombination time scales are even longer .
nevertheless , dig seems to be far from ionization equilibrium .
the same qualitative results and arguments also apply to heavier ions ( for a discussion see * ? ? ?
high ionization fractions of @xmath22 and @xmath23 close to 1 at large distances from the disk might suggest that the adopted radiation field produced by o6 stars is too hard . changing the spectral type of the ionizing sources to o7 or o8 ,
would restrict the energy distribution function to less energetic photons and lead to a reduced amount of he@xmath24 in the halo without substantially altering the ionization structure of the lower ionization species .
line ratios presented in fig .
[ f3 ] are normalized to h@xmath25 or h@xmath26 and have been computed by summing up the volume averaged line intensities of all cells along the line of sight for each @xmath1-layer .
it turns out that our simple model is well suited to reproduce observed line ratios for ngc891 , especially those of [ oiii]/h@xmath26 and [ oii]/h@xmath26 .
deviations between predicted and observed ratios , as e.g. , visualized in hei@xmath275876/h@xmath25 and [ oi]@xmath276300/h@xmath25 , suggest that the adopted incident radiation field is too hard .
however , pure photoionization by ob - stars seems to be appropriate to explain all observed line ratios consistently .
finally , the photon statistics shown in fig . [ f4 ] allows to extract the artifical spectrum which leaves the disk and the halo .
this plot also reveals a critical average density of about 8 @xmath28 at which no ionizing radiation escapes the galaxy . below this value ,
a significant fraction of ionizing photons leave the halo and contributes to the intergalactic radiation field .
a quantification of this fraction can be achieved after dust has been modeled in . although estimates @xcite indicate that only a few percent of the ionizing photons contribute to the ionization of the igm , this represents a significant contribution in view of the large number densities of normal spiral galaxies .
first results from a 3d self - consistent monte carlo model that accurately simulates dig in galactic halos with respect to the geometry and the radiative transfer have been presented .
the main conclusions can be summarized as follows : \(2 ) all ionization fractions in the disk at @xmath29kpc are in good agreement with predictions from model hii - regions @xcite .
we conclude that the disk can be approximated by a single huge hii - region .
\(6 ) dig seems to be in an extreme non - equilibrium condition .
+ forthcoming work will involve a comprehensive parameter study in order to investigate the dependence between line ratios , gas densities , stellar photon luminosities , and gas phase abundances .
after dusty plasmas can be simulated , detailed modeling of individual dig properties of well observed galaxies , such as ngc5775 or ngc891 , can be started . in case theory and observations
should reconcile , final proof would be given that photoionization by o stars can serve as the only ionization mechanism of the dig if the physics of radiative transfer is calculated accurately and a realistic geometry is considered .
abbott , d. c. , & lucy , l. b. 1985 , , 288 , 679 bland - hawthorn , j. , & maloney , p. r. 1999 , , 501 , l33 bland - hawthorn , j. , & maloney , p. r. 2001 , , 550 , l231 collins , j. a. , & rand , r. j. 2001 , , 551 , 57 domgrgen , h. , & mathis , j. s. 1994 , , 428 , 647 ferland , g. j. , binette , m. , contini , m. , harrington , j. , kallmann , t. , netzer , h. , pequignot , d. , raymond , j. , rubin , r. , shields , g. , sutherland , r. , & viegas , s. 1995 , in : `` the analysis of emission lines '' , stsci symp .
series 8 , eds .
williams , m. livio , cambridge university press , p. 83
ferland , g. j. , korista , k. t. , verner , d. a. , ferguson , j. w. , kingdon , j. b. , & verner , e. m. 1998 , , 110 , 761 hunter , d. a. , & gallagher , j. s. , iii . 1990 , , 362 , 480 lucy , l. b. , & abbott , d. c. 1993 , , 405 , 738 mathis , j. s. 1986 , , 301 , 423 mathis , j. s. 2000 , , 544 , 347 miller , s. t. , & veilleux , s. 2003 , , 592 , 79 och , s. r. , lucy , l. b. , & rosa , m. r. 1998 , , 336 , 301 osterbrock , d. e. 1989 , `` astrophysics of gaseous nebulae and active galactic nuclei '' , university science books otte , b. , gallagher , j. s. , iii .
, & reynolds , r. j. 2002 , , 572 , 823 rand , r. j. 1997 , , 474 , 129 reynolds , r. j. 1984 , , 282 , 191 reynolds , r. j. , haffner , l. m. , & tufte , s. l. 1999 , , 525 , l21 sokolowski , j. , & bland - hawthorn , j. 1991 , , 103 , 911 tllmann , r. 2002 , `` observations and modeling of diffuse ionized gas in edge - on galaxies '' , ph.d . thesis , ruhr - university of bochum , http://www-brs.ub.rub.de/netahtml/hss/diss/tuellmannralph/ tllmann , r. , & dettmar , r.j . 2000 , , 362 , 119 tllmann , r. , dettmar , r.j . , soida , m. , urbanik , m. , & rossa , j. 2000 , , 364 , l36 tllmann , r. , rosa , m. r. , & dettmar , r.j .
2004 , , submitted wood , k. , mathis , j. s. , & ercolano , b. 2004 , , 348 , 1337 wood , k. , mathis , j. s. 2004 , , in press | we investigate , whether the alleged failure of standard photoionization codes of the diffuse ionized gas ( dig ) is simply caused by geometrical effects and the insufficient treatment of the radiative transfer .
standard photoionization models are applicable only to homogeneous and spherical symmetric nebulae with central ionizing stars , whereas the geometry of disk galaxies requires a 3d distribution of ionizing sources in the disk which illuminate the halo .
this change in geometry together with a proper radiative transfer model is expected to substantially influence ionization conditions .
therefore , we developed a new and sophisticated 3d monte carlo photoionization code , called _ soap _
( simulations of astrophysical plasmas ) , by adapting an existing 1d code for hii - regions @xcite such , that it self - consistently models a 3d disk galaxy with a gaseous dig halo .
first results from a simple ( dust - free ) model with exponentially decreasing gas densities are presented and the predicted ionization structure of disk and halo is discussed .
theoretical line ratios agree well with observed ones , e.g , . for the halo of ngc891 .
moreover , the fraction of ionizing photons leaving the halo of the galaxy is plotted as a function of varying gas densities .
this quantity will be of particular importance for forthcoming studies , because rough estimates indicate that about 7% ionizing photons escape from the halo and contribute to the ionization of the igm . given the relatively large number density of normal spiral galaxies , ob - stars could have a much stronger impact on the ionization of the igm than agn or qsos . |
shear alfvn waves ( saw ) are electromagnetic plasma waves propagating as transverse waves along the magnetic field , with the characteristic alfvn velocity @xmath1 ( @xmath2 is the magnetic field and @xmath3 the mass density of the plasma ) . in fusion plasmas , fast ions in the mev energy range have velocities comparable with the typical alfvn speed . in addition , saw group velocity is directed along the magnetic field line and , therefore , fast ions can stay in resonance and effectively exchange energy with the wave @xcite . saw in a nonuniform equilibrium experience collisionless dissipation ( _ continuum damping _
@xcite ) , due to singular structures that are formed where the saw continuum is resonantly excited . in tokamaks ,
the magnetic field intensity varies along the field line .
this creates gaps in the saw continuous spectrum @xcite due to translational symmetry breaking , analogous to electrons traveling in a periodic lattice @xcite .
two types of collective shear alfvn instabilities exist in tokamak plasmas : energetic - particle continuum modes ( epm ) @xcite , with frequency determined by fast particle characteristic motions , and discrete alfvn eigenmodes ( ae ) , with a frequency inside saw continuum gaps @xcite .
the former can become unstable provided the drive exceeds a threshold determined by the continuum damping absorption ; the latter , have a generally lower instability threshold , being practically unaffected by continuum damping @xcite .
for this reason , the importance of understanding the continuum radial structure is clear , if one faces the problem of reaching the ignition condition with magnetically confined fusion plasmas , due to the potential impact of ae stability on the plasma confinement .
the nonlinear dynamics of saw is a research topic still rich of open issues . in the special case of uniform ideal incompressible plasmas a peculiar state exists , called the _ alfvnic state _
@xcite , where the nonlinear effects due to maxwell and reynolds stresses cancel and give a self - consistent nonlinear state .
on the other hand , in tokamak plasmas , nonuniformities and non - ideal effects such as resistivity or finite plasma compressibility are present .
this breaks the alfvnic state conditions and therefore saw are characterized by a rich nonlinear dynamics . as a consequence , the saw continuous spectrum can be modified by the interaction with low - frequency mhd fluctuations @xcite , such as magnetic islands .
magnetic islands are the result of magnetic - field reconnection processes , which occur in tokamak plasmas because of non ideal effects , such as finite resistivity @xcite .
the typical island oscillation frequency and growth rate are much lower than the saw frequency . due to this time scale separation between island and the saw dynamics ,
the island is modelled here as a static helical distortion of the equilibrium .
we derive our fluid theoretical description of the saw continuum structure inside a finite size magnetic island in finite-@xmath4 tokamak plasmas .
the magnetic island is modelled here as a straight flux tube with a noncircular cross section .
therefore , we neglect toroidicity effects on coupling modes with different poloidal mode numbers . however , we do take account the low - frequency gap in the saw continuum due to the effects of geodesic curvature and finite compressibility @xcite . in our model ,
the coupling of saw and the island magnetic perturbation is a linear problem of saw dynamics in a modified static equilibrium .
the local differential equation yielding the saw continuum structure is solved numerically with a shooting method code , in the spatial range of interest inside the magnetic island @xcite .
we are interested here in modes with different helicity from that of the magnetic island ; in other words , we assume @xmath5 , where @xmath6 is the coordinate along the flux tube . the solution of this problem inside and outside the magnetic island , for modes with the same helicity of the magnetic island , is given in a different paper @xcite .
as suggested by the magnetic field line helicity behavior @xcite , in an equilibrium with a magnetic island a generalized safety factor @xmath0 can be defined for each flux surface , as an appropriate average of the ratio of the magnetic field components along and perpendicular to the magnetic axis .
this safety factor is a monotone function of the distance from the o point , and grows towards the separatrix , where it has a singularity .
this is due to the vanishing of the magnetic field intensity near the x point of the magnetic island . within the separatrix
, we find a continuous spectrum very similar to that of tokamak plasmas , where the coupling of modes with different mode numbers creates frequency gaps .
in particular , a wide frequency gap is formed @xcite , analogous to the ellipticity induced alfvn eigenmode @xcite ( eae ) gap in tokamaks
. the presence of this gap is due to the strong eccentricity of the island cross section . in this gap ,
a magnetic - island induced alfvn eigenmode ( miae ) could exist as a bound state , essentially free of continuum damping , and interact nonlinearly with the magnetic island , if the thermal or energetic component of the plasma provide sufficient free energy for driving the mode .
due to the frequency dependence of the shear alfvn wave continuum on the magnetic - island size , the possibility of utilizing miae frequency scalings as a novel magnetic island diagnostic is also discussed .
the scheme of the paper is the following . in sec .
[ sec : equilibrium - coordinates ] we describe the equilibrium magnetic field , starting with the tokamak coordinates and introducing the cylinderlike coordinates for the magnetic island flux tube .
the generalized safety factor defined inside the magnetic island , is shown to have similar behavior of the safety factor in diverted tokamaks . in sec .
[ sec : continuous - spectrum ] , the saw model equations are introduced and the equation for general shear alfvn modes is provided .
the equation for continuum modes is then derived in the cylinderlike coordinate set defined inside the island and the eigenvalue problem is established .
the resulting continuous spectrum , obtained by solving numerically the eigenvalue problem with a shooting method code , is shown to be very similar to the continuous spectrum of tokamaks .
the result is provided for typical tokamak parameters , and typical size magnetic islands .
the analogy with the tokamak saw continuum is shown in the limit of very large magnetic islands , when the section of the magnetic island flux tube is nearly circular .
finally sec .
[ sec : discussion ] is devoted to a discussion about the applications of our results , in gaining a better knowledge on the stability of tokamak plasmas .
the differential operators calculated in the cylinderlike coordinate set are provided in appendix [ appendix : differential - operators ] .
we consider a tokamak geometry where @xmath7 is the major radius of the torus .
the equilibrium is made of an axisymmetric tokamak magnetic field , plus a helical perturbation generating the magnetic island .
we start our analysis by describing the axisymmetric field with an orthogonal coordinate system @xmath8 where @xmath9 is the radial coordinate , @xmath10 is the poloidal angle , corresponding to the field @xmath11 , and @xmath12 the toroidal angle , corresponding to the toroidal field @xmath13 , considered uniform for simplicity .
the subscript @xmath14 denotes here the tokamak coordinates .
the gradients associated with these coordinates are : @xmath15 , @xmath16 and @xmath17 .
we want to describe with a slab model , the region of the tokamak plasma where a magnetic island with poloidal and toroidal numbers @xmath18 is located .
this is the region around the magnetic island s rational flux surface @xmath19 , corresponding to @xmath20 , where the axisymmetric field has value @xmath21 , with @xmath22 . here
@xmath23 is the tokamak safety factor @xmath24 , and @xmath25 is the inverse aspect ratio , assumed to be small : @xmath26 .
therefore , we introduce an orthogonal coordinate system @xmath27 through a rotation in the ( @xmath28 ) plane , defining @xmath29 and @xmath30 .
this is appropriate to describe a sheared field problem , in the proximity of the rational surface , because the axisymmetric magnetic field is directed only along @xmath31 at @xmath32 .
the gradients associated with these coordinates are : @xmath33 , @xmath34 and @xmath35 .
the parameter @xmath36 is the magnetic shear calculated at the rational surface .
it is useful to rewrite the gradient of the coordinates @xmath37 and @xmath6 in the form @xmath38 and @xmath39 , where @xmath40 , and @xmath41 .
in fact , we see that @xmath42 is the length of the magnetic island in the @xmath37 direction and @xmath43 is the length of a magnetic field line of the axisymmetric equilibrium at @xmath44 . in other words , @xmath43 is the periodicity length of the magnetic flux tube defined as the region inside the magnetic - island separatrix .
the axisymmetric magnetic field components can be written now by considering that @xmath31 forms an angle @xmath45 with @xmath46 . therefore in these variables
the magnetic field of an equilibrium with a magnetic island is @xmath47 , with : @xmath48 where we dub the components rescaled with the length of the corresponding basis vector as _ physical _ components ( see appendix [ appendix : differential - operators ] ) .
we have adopted the _ _
constant-__@xmath49 approximation for the magnetic island , assuming a constant magnetic field amplitude of magnetic island perturbation , @xmath50 .
the x points of the magnetic island are at @xmath51 and @xmath52 and the o point at @xmath53 .
the flux surfaces of this equilibrium are labeled by @xmath54 .
m is a constant determined by the condition @xmath55 @xmath56 .
we obtain @xmath57 , where @xmath58 is the poloidal magnetic field calculated at the rational surface .
the safety factor defined inside a magnetic island , plotted versus the radial position @xmath59 , and normalized to @xmath60 . for typical magnetic islands , @xmath61 and
we have @xmath62.,scaledwidth=47.0% ] here , we model the plasma inside the magnetic island as a straight flux tube , where @xmath6 is the translational symmetry coordinate , and @xmath49 labels the nested flux surfaces .
the magnetic axis and island o point are at @xmath63 , while the separatrix is labeled by @xmath64 .
we define a complete set of cylinderlike coordinates @xmath65 , with radial - like and anglelike coordinates given by : @xmath66 with these definitions , the magnetic axis is at @xmath67 and the separatrix radius is @xmath68 , which corresponds to the magnetic island half - width @xmath69 , given by the rutherford formula @xcite : @xmath70 with @xmath71 defined below in this paragraph .
the angle @xmath72 is defined in the domain @xmath73 and extended to @xmath53 by reflection symmetry with respect to @xmath74 , with values @xmath75 at the rational surface @xmath76 .
further extension to @xmath77 is obtained by reflection symmetry for @xmath78 .
we also note that the flux surface s cross section in the @xmath79 plane and in the proximity of the o point is an ellipse , and the value of its eccentricity is @xmath80 .
typical magnetic islands in tokamak experiments have eccentricities close to @xmath81 .
the case of zero eccentricity corresponds to a value of magnetic island width , which is never encountered in tokamak plasmas because a saturation or a disruption occurs long before the island grows up to these sizes .
nevertheless , in the following sections we calculate the saw continuous spectrum for the case @xmath82 too .
in fact , this case shows clearly the analogy with the well - known results of continuous spectrum in a tokamak , whose cross section has typically small values of eccentricity .
the differential operators in this coordinate set are provided in appendix [ appendix : differential - operators ] .
these are necessary to give an explicit form to the equation for the saw dynamics , that will be introduced in sec .
[ subsec : model ] . the equilibrium magnetic field in the region inside the island , can be described in the cylinderlike coordinates @xmath65 defined in sec .
[ subsec : coordinates - inside ] . the contravariant _ physical _ components are : @xmath83 here @xmath84 is the value of @xmath85 at @xmath86 , @xmath87 , where all parameters refer to the original axisymmetric equilibrium .
the function @xmath88 and the definition of contravariant _ physical _ components are provided in appendix [ appendix : differential - operators ] .
we define the safety factor @xmath0 inside the magnetic - island flux tube as the average of @xmath89 over @xmath72 , with @xmath90 ; i.e. , @xmath91 where we have introduced the complete elliptic integral of the first kind @xmath92 , with @xmath93 and @xmath94 .
a similar definition for the safety factor was given in ref .
@xcite , but our result is a factor @xmath95 smaller . the basic difference of our derivation from that of ref .
@xcite , is that we take into account that the length of a magnetic island flux tube is @xmath96 , whereas in ref .
@xcite a flux tube with length @xmath97 is considered , which is the major circumference of the tokamak ( see also ref .
@xcite ) . the safety factor inside a magnetic island , normalized to the minimum value @xmath98 , is shown in fig .
[ fig : q(x ) ] .
we see that @xmath99 has a singular behavior near the separatrix @xmath100 , since @xmath101 vanishes at @xmath102 , @xmath103 .
this feature is due to the x points and is analogous to safety factor profiles in diverted tokamak plasmas . continuous spectrum of saw , @xmath104 , evaluated from the safety factor profile as in eq .
[ eq : cont - approx ] , for the case @xmath105 ( corresponding to @xmath106 ) .
modes with @xmath107 , and @xmath108 to @xmath109 ( where @xmath110 is the poloidal mode number ) are considered , and @xmath111 is chosen . here
the approximation of cylindrical symmetry is adopted , and therefore the continuum branches of the modes with different poloidal mode number @xmath110 intersect without interacting .
compare with fig .
[ fig : e - zero ] and fig .
[ fig : m - small ] , where the continuous spectrum is calculated retaining all geometrical effects due to the noncircularity of the flux surface s cross section.,scaledwidth=47.0% ]
here , we study the shear alfvn wave dynamics in a general equilibrium , adopting the ideal mhd model . in next sections
we introduce the geometry defined with the coordinates ( @xmath112 ) , and give the explicit solution of the problem .
the starting linear equations of our model are : @xmath113 where @xmath3 is the mass density , @xmath114 the fluid velocity , @xmath115 the pressure , @xmath116 the plasma current , @xmath117 and @xmath118 the magnetic and electric fields , @xmath119 the ratio of the specific heats and @xmath120 the speed of light .
this set of equations is solved for the perturbed velocity , the perturbed magnetic field , and the perturbed pressure , expressed respectively as @xmath121 , @xmath122 and @xmath123 , where @xmath124 is the perturbed scalar potential and @xmath125 is the perturbed vector potential .
the set of equations [ eq : momentum ] , [ eq : ohm ] , [ eq : state ] for shear alfvn waves can be casted in the following linearized equation for the scalar potential : @xmath126_\perp + \nonumber\\ & & + 4\pi \big[\big(\bm{b } \times\frac{2\bm\kappa_s}{b^2 } \big ) \cdot \bm\nabla_\perp \big ] \big [ \big(\bm{b } \times\frac{\bm\nabla p}{b^2 }
\big ) \cdot \bm\nabla_\perp\big ] \phi + \nonumber \\ & & + \nabla_\parallel \nabla_\perp^2 \nabla_\parallel \phi - \frac{4\pi \gamma p}{b^2 } \big [ \big ( \bm{b}\times\frac{2\bm\kappa_s}{b}\big ) \cdot \bm\nabla_\perp \big]^2 \phi = 0 \label{eq : general - modes}\end{aligned}\ ] ] where all quantities but the scalar potential @xmath124 , are referred to the equilibrium .
all curvature effects are retained here : @xmath127 is a function of space , and the geodesic curvature @xmath128 allows the pressure to affect the saw dynamics . the geodesic curvature is defined as the component of the curvature @xmath129 tangent to the magnetic flux surfaces . equation [ eq : general - modes ] is valid for both global modes and radially localized modes .
it can be solved in the coordinate set ( @xmath112 ) , by substituting the differential operators defined in appendix [ appendix : differential - operators ] . here
, we want to consider the shear alfvn wave propagation near the resonant flux surfaces where the energy is absorbed by continuum damping .
therefore , we focus on the dynamics of continuum modes , namely modes that are characterized by radial singular structures .
this allows us to drop the second and third terms in eq .
[ eq : general - modes ] . the linear equation for radially localized shear alfvn modes in a compressible nonuniform tokamak plasma can be written in the form @xcite : @xmath130 where @xmath131 .
we adopt the value of the frequency of the low - frequency saw continuum accumulation point ( cap ) , delimiting the frequency gap of the beta induced alfvn eigenmode ( bae ) @xcite , given in ref .
@xcite : @xmath132 where @xmath133 and @xmath134 are the ion and electron temperatures , and @xmath135 is the ion mass . here , we focus on frequencies higher than @xmath136 and consistently neglect kinetic effects associated with wave - particle resonances @xcite
. the operators @xmath137 and @xmath138 are the gradients calculated along and perpendicularly to the equilibrium magnetic field ( given by the sum of the axisymmetric field plus the island field )
. in section [ subsec : coordinates - inside ] , we have described the region inside a magnetic - island flux tube with a cylinderlike set of coordinates , labeling the magnetic flux surfaces with a radial - like coordinate @xmath139 .
this magnetic flux tube is assumed here to be the static equilibrium of the problem .
we want to put the equation for radially localized shear alfvn modes , eq .
[ eq : continuum - modes ] , which is written in a general vectorial form , in an explicit form , using a cylinderlike set of coordinates for a straight flux tube and the differential operators defined in appendix [ appendix : differential - operators ] .
we substitute the parallel and perpendicular differential operators : @xmath140 and we impose boundary conditions of periodicity in @xmath72 and @xmath6 on @xmath141 circles . moreover , using the symmetry of the equilibrium in the @xmath6 coordinate ,
we can write the general solution as @xmath142 , where @xmath143 is the mode number in the @xmath6-direction . continuous spectrum of saw , @xmath104 , obtained by solving numerically eq .
[ eq : eigenvalue ] , and plotted versus the radial position inside the island , @xmath59 .
the case of small eccentricity case , @xmath105 ( corresponding to @xmath106 ) , is considered .
typical values of the equilibrium parameters have been chosen and @xmath111 .
the o point is at @xmath144 and the separatrix at @xmath100 .
modes with @xmath107 are considered .
the miae gap is found at frequencies @xmath145.,scaledwidth=47.0% ] we also note that the problem we are considering is greatly simplified if we change coordinates from @xmath65 to a field aligned set of coordinates @xmath146 , where @xmath147 , and @xmath148 is the normalized incomplete elliptic integral of the first kind .
in fact , we have @xmath149 and the parallel derivative takes the simple form @xmath150 . finally , eq . [ eq : continuum - modes ] is written in the form of an eigenvalue problem : @xmath151 \tilde\phi '' = 0\ ] ] where @xmath152 is the eigenvalue , and @xmath153 is the eigenfunction .
the boundary condition @xmath154 now reads @xmath155 , with the safety factor @xmath0 defined in eq .
[ eq : safety - factor ] . before solving eq .
[ eq : eigenvalue ] in its exact form , we note here that we can have an approximate solution by assuming cylindrical symmetry , and using the safety factor as calculated in eq .
[ eq : safety - factor ] . in this case , the solution of eq .
[ eq : eigenvalue ] can be written in the form : @xmath156 where we have used @xmath157 , with @xmath110 being the poloidal mode number .
when cylindrical symmetry is assumed , the continuum branches of modes with different @xmath110 are not coupled , and therefore intersect without interacting ( see fig .
[ fig : cont - from - q ] ) . in the case of small eccentricity @xmath105 , and close to the o point @xmath158
, we have @xmath159 and the formula given in eq .
[ eq : cont - approx ] is a good approximation of the solution of eq .
[ eq : eigenvalue ] : @xmath160 . in next section
, we provide the exact solution of eq .
[ eq : eigenvalue ] , namely retaining all geometrical effects due to the noncircularity of the flux surface s cross section .
in previous section , the equation for saw continuous spectrum has been written in the form of an eigenvalue problem in the coordinates @xmath146 , eq .
[ eq : eigenvalue ] .
this is an ordinary differential equation in one variable , @xmath72 , where the radial position @xmath59 is treated as a parameter .
it can be solved numerically , with a shooting method code for each position @xmath161 , giving the continuous spectrum @xmath162 as result .
the case of small eccentricity , corresponding to @xmath163 ( and @xmath164 ) , is considered firstly .
typical values for the equilibrium parameters , @xmath165 , @xmath166 , @xmath167 , @xmath168 are chosen .
continuum modes with mode number @xmath107 are considered ( see fig .
[ fig : e - zero ] ) . in this case of small eccentricity
, we see that near the o point , @xmath158 , cylindrical symmetry is a good approximation and the result is well described by the formula given in eq .
[ eq : cont - approx ] , as shown in fig .
[ fig : cont - from - q ] .
on the other hand , for @xmath169 , the flux surfaces have a noncircular cross section and therefore mode coupling occurs between modes with different @xmath110 numbers .
this mode coupling is responsible for the presence of a gap in the spectrum analogous to the eae gap in tokamaks @xcite , dubbed here as magnetic - island induced ae ( miae ) gap .
the closer a flux surface is to the separatrix , the wider is the miae gap
. continuous spectrum of saw , @xmath104 , calculated as in fig .
[ fig : e - zero ] , for @xmath170 and a typical size magnetic island , @xmath171 ( corresponding to @xmath172 ) . in this case
, the structure of the continuous spectrum is the same as for @xmath163 , but the miae gap is much wider .
this implies that a wide range of frequencies exists , where modes contained within the magnetic island are not affected by continuum damping.,scaledwidth=47.0% ] the calculation is also repeated for a magnetic island with a typical size , with @xmath173 , using the same equilibrium parameters .
this corresponds to a magnetic island with @xmath174 , which is the order of magnitude of a saturated magnetic island in tokamaks . for this case , we considered modes with @xmath170 .
we find that the continuous spectrum structure , shown in fig .
[ fig : m - small ] , is the same as for @xmath163 , but in this case we obtain a much wider miae gap .
this is due to the fact that flux surfaces of a typical magnetic island have a high eccentricity , and therefore the mode coupling between modes with different @xmath110 numbers is stronger .
the miae - gap central frequency is proportional to the magnetic island half - width and near the o point we have : @xmath175 this is valid for any value of any value of magnetic island half width , corresponding to any value of eccentricity . for typical tokamak plasma parameters , @xmath176 , with @xmath4 denoting the ratio between plasma and magnetic pressures .
therefore , finite compression bae frequency gap and magnetic island induced miae gap are of comparable size when @xmath177 , i.e. for a typical saturated magnetic - island size in tokamaks .
we have considered a straight flux tube model , neglecting the effect of curvature coupling among different @xmath110 numbers .
therefore , a gap analogous to the toroidicity induced alfvn eigenmode gap @xcite is not found here .
nevertheless , such a toroidal miae gap is expected to be present inside magnetic islands , at frequencies lower than the frequencies of the ellipticity miae gap .
its gap width can be estimated by substituting @xmath127 with @xmath178 in eq .
[ eq : continuum - modes ] .
this gives a multiplication factor @xmath179 to the first and third terms of eq .
[ eq : continuum - modes ] , and consequently to the first term of eq . [ eq : eigenvalue ] .
therefore , we expect to have a toroidicity induced gap within a magnetic island , with a width of the order of @xmath180 , analogously to the theory of toroidicity induced ae .
the magnetic island also has a modulation in the tokamak toroidal direction , given by the number @xmath181 .
this helicity of the flux tube is neglected here and is expected to weakly couple modes of the saw continuous spectrum with different @xmath143 , analogously to the case of stellarators .
these effects are not expected to qualitatively modify the miae gap and will be investigated in a different work as further extension of the present theory . on the other hand ,
the bae gap is considered in our model in the definition of @xmath182 , resulting in an upward shift in the continuous spectrum of the frequencies @xmath183 .
one of the main damping mechanisms of saw instabilities in nonuniform plasmas is continuum damping , which occurs near the magnetic flux surfaces where the frequency of the instability matches the saw continuum frequency .
therefore , understanding the structure of the saw continuous spectrum is one crucial step in studying the stability properties of a tokamak plasma . in toroidal devices ,
a frequency gap in the continuous spectrum represents a range of frequency where discrete alfvn eigenmodes ( ae ) can grow unstable , practically unaffected by continuum damping .
the radial structure of the continuous spectrum of saw is calculated in this paper in an equilibrium inside a magnetic - island flux tube .
since the typical island frequency and growth rate are much lower than the saw oscillation frequency , we have assumed that the equilibrium magnetic field is given by the usual tokamak axisymmetric field plus a quasi - static helical distortion due to the magnetic island .
the flux tube is considered to be straight , in order to focus on the effects of the noncircularity of the cross section .
curvature effects are retained only in the formation of the bae gap in the low frequency part of the continuous spectrum .
a linear mhd model is adopted for finite - beta tokamak plasmas .
a generalized safety factor is defined inside the magnetic - island flux tube , giving information on the rational flux surfaces where different modes couple to create gaps in the saw continuous spectrum .
the linearized saw equation is solved with a shooting method code inside a finite - size magnetic island and the result compared with the saw continuous spectrum calculated in tokamak equilibria .
we have found that there exists a saw continuous spectrum within a magnetic island , similar to that calculated in tokamak equilibria @xcite . for the case of small eccentricity of the flux surfaces , modes with different poloidal numbers are not coupled and therefore , the branches of the continuous spectrum intersect and there is no formation of gaps .
this is shown to occur inside a magnetic island with very large ( not typical ) width , near the o point , where the flux surfaces are nearly circular .
on the other hand , a typical - size magnetic island is shown to have wide gaps in the continuous spectrum , due to the strong eccentricity of the flux surfaces , labeled here as miae gap .
this is analogous to the formation of eae gaps in tokamaks , but is found here inside the magnetic island equilibrium .
note that miae can exist as bound states within the island , essentially free of continuum damping , provided that plasma equilibrium effects and free energy sources can drive and bind them locally @xcite .
the equilibrium inside a magnetic island is shown here to be analogous to a tokamak equilibrium , and , consequently , we emphasize the analogies between ae within the magnetic islands ( miae ) and the well - known ae in tokamaks .
the miae peculiarity resides in the corresponding saw continuous spectrum dependence on the magnetic - island size .
moreover , we expect to find all kinds of alfvnic instabilities inside a magnetic island , and not only ellipticity - miae .
for instance , toroidicity - miae could grow inside a gap due to the flux tube toroidicity effects , which are not studied here and will be treated in a subsequent paper .
we point out the important implications of this result in understanding the dynamics and stability properties of a magnetic island .
in fact , the presence of miae inside a magnetic island could modify the equilibrium profiles and nonlinearly affect the magnetic island growth . on the other hand
, miae could nonlinearly affect energetic particles redistribution in the proximity of the magnetic island rational surface , caused in part by the radial magnetic field due to the magnetic island itself @xcite .
the frequency of the ellipticity miae gap has been shown to be proportional to the magnetic - island width [ see eq .
[ eq : miae - gap ] ] .
this suggests the possibility of using the miae frequency scalings as a novel magnetic - island diagnostics in a fashion similar to other commonly used alfvn spectroscopy techniques @xcite . unlike in the bae case ,
the radial miae localization at the center of the island makes them difficult to detect by external measurements , and makes the use of internal fluctuation diagnostics , such as electron cyclotron emission and soft x rays , necessary . by detecting miae inside the island and measuring their frequency ,
one has indirect information on the magnetic - island size .
this work was supported by the euratom communities under the contract of association between euratom / enea and in part by prin 2006 and create , as well as by doe grants no .
de - fg02 - 04er54736 and no .
de - fc02 - 04er54796 , and by the nsf grant no .
part of this work was done while one of the authors , a.b .
was at the university of pisa , in collaboration with university of california , irvine , and centro ricerche enea frascati , which are gratefully acknowledged for the hospitality .
one of the author , a.b . would also like to thank andreas bierwage for valuable suggestions in writing the shooting method code .
useful discussions with c. di troia , i. chavdarovski , x. wang and p. porazik are also acknowledged .
99 f. zonca _
et al . _ , plasma phys . controlled fusion * 48 * , b15 ( 2006 ) . l. chen and f. zonca , nucl .
fusion * 47 * , s727 ( 2007 ) .
h. grad , phys . today * 22 * , ( 12 ) 34 ( 1969 ) .
a. hasegawa and l. chen , phys .
* 32 * , 454 ( 1974 ) .
l. chen and a. hasegawa , phys .
fluids * 17 * , 1399 ( 1974 ) . c. e. kieras and j. a. tataronis , j. plasma physics * 28 * , 395 ( 1982 ) .
l. chen , phys .
plasmas * 1 * , 1519 ( 1994 ) . c. z. cheng , l. chen and m. s. chance , ann .
phys . * 161 * , 21 ( 1985 ) .
h. alfvn , `` cosmical electrodynamics '' , clarendon press , oxford ( 1950 ) . w. m. elsasser , rev . mod
* 28 * , 135 ( 1956 ) .
a. hasegawa and t. sato , `` space plasma physics ( stationary processes ) '' vol .
1 ( new york ; springer ) ( 1989 ) .
p. buratti _
et al . _ ,
fusion * 45 * , 1446 ( 2005 ) .
h. p. furth , j. killeen and m. n. rosenbluth , phys .
fluids * 6 * , 459 ( 1963 ) .
m. s. chu , j. m. greene , l. l. lao , a. d. turnbull and m. s. chance , phys .
fluids b * 4 * , 3713 ( 1992 ) .
a. turnbull _
et al . _ ,
fluids b * 5 * , 2546 ( 1993 ) .
f. zonca , l. chen and r. a. santoro , plasma phys .
controlled fusion * 38 * , 2011 ( 1996 ) .
a. biancalani , l. chen , f. pegoraro and f. zonca , phys .
lett . * 105 * , 095002 ( 2010 ) .
a. biancalani , l. chen , f. pegoraro and f. zonca , `` 2d continuous spectrum of shear alfvn waves in the presence of a magnetic island '' , submitted to plasma phys .
controlled fusion , ( 2010 ) .
k. swartz and r. d. hazeltine , phys .
fluids * 27 * , 2043 ( 1984 ) .
r. betti and j. p. friedberg , phys .
fluids b * 3 * , 1865 ( 1991 ) .
p. h. rutherford , phys .
fluids * 16 * , 1903 ( 1973 ) . c. z. cheng and m. s. chance , phys .
fluids * 29 * , 3695 ( 1986 ) .
m. brdgam `` nonlinear effects of energetic particle driven instabilities in tokamaks '' , phd thesis , tu mnchen , germany ( 2010 ) .
s. d. pinches _ et al .
_ , plasma phys . controlled fusion * 46 * , b187 ( 2004 )
. b. breizman , m. s. pekker and s. e. sharapov , phys .
plasmas * 12 * , 112506 ( 2005 ) .
et al . _ ,
fusion * 49 * , 085009 ( 2009 ) .
here , we calculate the differential operators in the set of coordinates inside the magnetic island @xmath65 . the domains of these coordinates are @xmath184 , @xmath185 and @xmath186 , where @xmath6 is the coordinate of translational symmetry , and periodicity in @xmath72 and @xmath6 is assumed for the perturbations .
the gradients of these coordinates are : @xmath187 where we use the notation @xmath188 for a unitary length vector .
moreover we have the relation @xmath189 . here
@xmath190 , where @xmath191 , and : @xmath192 a covariant basis ( @xmath193 ) can also be defined as orthogonal to the contravariant basis ( @xmath194 ) , namely satisfying : @xmath195 . with this notation
, any vector can be written as @xmath196 , where we call @xmath197 the _ physical _ components of the vector , @xmath198 .
finally , the differential operators for the cylinderlike set of coordinates @xmath199 are : -1em @xmath200 it is straightforward to see that , in the case @xmath82 and in proximity of the o point ( @xmath201 , @xmath202 ) , these differential operators reduce to the known differential operators in cylinder geometry . | the radial structure of the continuous spectrum of shear alfvn waves is calculated in this paper within the separatrix of a magnetic island .
geometrical effects due to the noncircularity of the flux surface s cross section are retained to all orders . on the other hand , we keep only curvature effects responsible for the beta - induced gap in the low - frequency part of the continuous spectrum . modes with different helicity from that of the magnetic island are considered .
the main result is that , inside a magnetic island , there is a continuous spectrum very similar to that of tokamak plasmas , where a generalized safety factor @xmath0 can be defined and where a wide frequency gap is formed , analogous to the ellipticity induced alfvn eigenmode gap in tokamaks .
the presence of this gap is due to the strong eccentricity of the island cross section .
the importance of the existence of such a gap is recognized in potentially hosting magnetic - island induced alfvn eigenmodes ( miae ) . due to the frequency dependence of the shear alfvn wave continuum on the magnetic - island size , the possibility of utilizing miae frequency scalings as a novel magnetic - island diagnostic is also discussed .
+ pacs : 52.55.tn , 52.35.bj |
the prediction of structural , dynamic and thermodynamic properties and the phase behavior of fluids based solely on the shape of the intermolecular interaction potential is an important goal in the study of liquids and in the understanding of the peculiarities of the liquid state .
the simplicity of this surmise has not only theoretical interest but also practical consequences@xcite . in spite of these motivations and precepts of developing a much better molecular - level picture of the liquid matter , this goal is far from being achieved .
although for some simple potential the use of integral equations in the form of hypernetted - chain approximation can give an estimate of the phase stability @xcite , the mathematical complexity of the interaction potentials used even for simple liquid does not allow an analytical approach .
despite of this , some progress can be made choosing only liquids that share the same type of structure , for instance , a tetrahedrical local structure .
the most familiar tetrahedral liquid is water not only due to its abundance but also because of the occurrence of some remarkable features such as the presence of structural , dynamic or thermodynamic anomalies .
the question is to determine which characteristics of the intermolecular interaction potential would be needed to induce anomalies in the liquid phase behavior of fluids governed by this potential .
one of the most familiar mysteries in water s behavior is the density anomaly : at sufficient low temperatures , warming the liquid causes it to shrink .
there is an increasing awareness that including this one property in the potential will reveal the origins of the others .
recently , following an earlier hypothesis of kamb @xcite , and supported by experimental results @xcite-@xcite it has been shown that the increase in density on melting is related to a distortion in the non - nearest - neighbor coordination , i.e. bending of hydrogen bonds .
the key point is therefore modeling the behavior of the system based not on a potential mimicking the first neighbors interactions , as in mostly current models , but on a potential capable of describe the peculiar interactions between second neighbors @xcite@xcite .
these interactions are double - ranged : the competition between two scales is responsible for many of the anomalies of water . in this spirit
one can understand the success of a simple model as the core - softened potential introduced by stell et al .
it was shown that an one - dimensional hard - core system showing a first order phase transition could be capable of exhibiting two ( or even three ) phase transitions if the hard - core is softened in a reasonable way .
it is , however , not clear if the same essential features of this model would produce anomalies in three - dimensional systems .
another example is the double - well takahashi model@xcite , with an inner shallow well and a outer deeper well .
this model exhibits a density maximum in one - dimensional systems , with the correct pressure dependence : this density maximum becomes broader and shifts to lower temperatures by increasing the pressure .
this approach , however , seems not to work in three - dimensional systems@xcite and the question was posed if the three - dimensional potential , besides being double - ranged , also must have an angular dependence in order to show anomalies .
cho and co - workers @xcite also proposed an analytic double - well potential without angular dependence , with a slight modification on a lennard - jones potential , adding a gaussian - shaped perturbation .
this potential seems to have anomaly in the density .
however , a more consistent exploration of the parameter space is necessary in order to determine if this is the case for all range of parameters .
later , cho et al .
@xcite argued that only non local potentials could reproduce all water s anomalies . within their analysis
the potential should have a strong angular dependence and favor the build - up of a second - neighbor structure @xcite .
following a similar approach as the original takahashi model , jagla showed@xcite that several anomalies could be described by a simple interaction potential with two competing equilibrium distances .
the model including a global term for attractions displays a liquid phase with a first order line of liquid - gas transition ending in a critical point .
this potential , however , is difficult to implement in a molecular dynamics simulation , because the attraction term is not explicitly taken into account and the system needs long equilibration times . besides the three models mentioned above , a number of lattice @xcite and continuous @xcite core - softened potentials has been proposed . however , besides the thermodynamic , water also has dynamic anomalies .
while for most materials diffusivity decreases with pressure , liquid water has an opposite behavior in a large region of the phase diagram @xcite-@xcite . a good model for water and tetrahedral liquids
should not only exhibit the thermodynamic but also the anomaly in the mobility .
these dynamical anomalies are related to the competition between the local tetrahedral structure of first neighbors - which tends to decrease the mobility and the distortion of the structure of first and second neighbors , with weakened hydrogen bonds and interstitial water molecules - which tends to increase the mobility .
this anomaly in the diffusivity is enhanced at low temperatures where the two scales phenomena becomes more proeminent . in the present paper ,
we check if a simple model in which the particles are model as modified soft spheres interacting with two competing ranges exhibits the anomalies present in water .
particularly we study the diffusivity of two model cases : one in which the outer minimum is deeper and another in which the inner minimum is deeper .
this paper is organized as follows . in next section the model and method
are described and in sec .
3 the results and the conclusions are shown .
= - 1.2 , @xmath0 = 1.7 , @xmath1 = 0.15 , model potential ii ( dashed line ) with a inner deeper minimum , and parameters @xmath2 = - 0.6 , @xmath0 = 1.7 , @xmath1 = 0.15 , compared with a normal lennard - jones potential ( dotted line),width=377,height=226 ]
our model system consists of softened spheres with an intermolecular potential displaying a repulsive short - ranged component showing @xmath3 dependence and a double - ranged attractive component , by adding a gaussian to a normal lennard - jones potential @xcite , as follows : @xmath4 + \alpha \exp \left ( - \frac{(r-\beta)^2}{\gamma^2 } \right)\ ] ] this potential can represent a whole family of intermolecular interactions , from weak repulsive shoulder to deep double well potentials , depending on the choice of the values of @xmath2 , @xmath0 and @xmath1 . figures 1 - 2 show two choices of parameters that illustrate two significative cases : one in which the outer minimum is deeper and another in which the inner minimum is minimum .
depending on the choice of parameters , we can obtain either a potential similar to the one employed by cho@xcite or the one introduced by jagla@xcite .
we carried out molecular dynamics simulations in the microcanonical ensemble of 256 particules interacting with an intermolecular potential given by the expression above .
a broad range of thermodynamic conditions , expressed as several densities and temperatures was chosen to explore the phase diagram of the fluid described by this potential .
using reduced units for density @xmath5 , for temperature @xmath6 , the values for @xmath7 and @xmath8 were 0.60 to 1.00 and 0.30 to 1.00 respectivelly .
thermodynamic parameters ( internal energy and pressure ) were calculated over 3000000 steps long simulations .
analysis of the dependence of the pressure against density along isotherms , aided by visual inspection of the final structure allows to check if the system is unstable or not .
we measure the mean - square displacement averaged over different initial times @xmath9 ^ 2\rangle$ ] and then the diffusion constant is calculated using the relation @xmath10 .
the diffusion constant is calculated for two types of potentials : ( i)@xmath11 , @xmath12 and @xmath13 , ( ii ) @xmath14 , @xmath12 and @xmath13 .
potential i exhibit a competition between a first attractive shell and a second deeper minimum .
this competing interactions can generate two liquid phases with different densities @xcite@xcite@xcite and a resulting density anomaly .
the mobility of potential i , illustrated in figure 2 , is similar to the behavior we found @xcite in spce @xcite supercooled water .
the diffusivity increases as the density is lowered , reaches a maximum and decreases .
it is not clear if for this potential the diffusivity increases at very low densities @xcite because the system cavitates before this happens .
potential ii has a competition between a deeper inner well and an shallow outer shell .
this competition can also generate two liquid phases with different densities @xcite and a resulting density anomaly .
the diffusion constant related to potential ii , illustrated in figure 3 , is also anomalous , however its anomalous behavior occurs inside the unstable region where the system already is phase separated . according cho et .
@xcite@xcite a double well takahashi model would exhibit a density maximum only if the outer well is deeper than the inner well .
although our potential is not of a takahashi type , its shape roughly resembles that potential by an adequate choice of parameters .
our main conclusion is that even if the outer well is shallower than the inner well the anomalies are still present , however they are covered by the phase separation .
a systematic study in progress of the dependence of the fluid behavior with the choice of parameters will enable us to better understand the hierarchy of the several anomalies in such a two - ranged analytic model . | using molecular dynamics simulations we investigate the ability of an analytical three - dimensional double well in reproducing static and dynamic anomalies found experimentally in liquid water .
we find anomalous behavior in the stable region of the phase diagram if the outer minimum is deeper than the inner minimum . in the case of a deeper inner minimum , anomalous behavior is also present but inside the unstable region .
core - softened potential , diffusion |
light stops are a quintessential feature of a natural supersymmetric theory @xcite , being responsible for the cancellation of the dominant higgs mass quadratic divergence coming from the top quark . however , after run 1 of the lhc the possible existence of light stops has been strongly constrained by a suite of dedicated searches by atlas and cms @xcite .
for instance , in a simplified scenario containing a neutralino as the lightest supersymmetric particle ( lsp ) and the stop as the next - to - lightest supersymmetric particle ( nlsp ) , with a few gaps and caveats , stop masses as high as @xmath15 gev and lsp masses as high as @xmath16 gev have been excluded @xcite . still , as is always the case with direct searches , the limits depend strongly on the spectrum and decays of the stops , and there remain open windows in which a light stop can hide from lhc searches .
examples include compressed @xcite or stealth @xcite stops and r - parity violating stop decays @xcite .
it is therefore critical to continue to develop new strategies to directly search for light stops @xcite .
another opportunity to probe light stops is presented when the stop is long lived , which naturally occurs in a number of motivated scenarios .
indeed , provided the stop does not have an _ unsuppressed _
2-body decay , stop pairs produced through gluon fusion can form a stoponium bound state due to the coulombic attraction mediated by the strong force @xcite .
once the bound state is formed , it can decay via annihilation to @xmath17 and @xmath4 , leaving a distinctive resonance signature .
as we will emphasize below , the branching ratios to these final states depend essentially on one parameter , the trilinear higgs - stop - stop ( @xmath18 ) coupling .
therefore , provided there is enough time for the stoponium to form and annihilate decay , the signature is independent of the details of the light susy spectrum and thus offers a highly complementary probe to direct stop searches . in this paper
we will use the null results from @xmath19 and @xmath4 resonance searches to constrain the production of stoponium during the first run of the lhc and in turn derive new limits on light stops . assuming the stop is sufficiently long - lived , constraints on stoponium production in the @xmath0 , @xmath20 , and @xmath2 channels _ alone _ limit stops lighter than @xmath21 gev for _ any _ value of the stop mixing angle and heavy stop mass @xmath22 .
furthermore , for mostly left - handed light stops , the @xmath10 searches constrain light stops up to about 170 gev .
we will also present projections for 14 tev lhc with low ( 30 fb@xmath23 ) and high ( 3 ab@xmath23 ) luminosities .
one new aspect of our study which has not been emphasized in the literature is the interplay between stoponium searches and other indirect constraints on light stops . in particular , since the stoponium branching ratios depend primarily on the @xmath18 coupling , our new constraints are highly complementary to those suggested by higgs coupling measurements .
light stops with a large coupling to the higgs boson significantly alter higgs production in the gluon - gluon fusion channel and are already strongly constrained except for a `` blind spot '' in the parameter region where the @xmath18 coupling is small @xcite . on the other hand , the @xmath0 resonance signature of stoponium
is enhanced in the region where the @xmath18 coupling is small . combining both constraints ,
we can rule out the existence of light stops up to about 130 gev independently of the other parameters in the stop sector . besides the interplay with higgs couplings , we will also explore the impact of other indirect constraints on light stops , including vacuum stability and electroweak precision data .
we also discuss the consequences of our results for two motivated scenarios in which stoponium annihilation decays are relevant .
the first is the canonical r - parity conserving case with a bino lsp and a stop nlsp . for @xmath24
the stoponium annihilation decays are visible , providing a clean and unambiguous probe of light stops which is independent of possible kinematic degeneracies or model dependencies of stop branching ratios .
second , we discuss the case of a stop lsp decaying via the r - parity violating @xmath12 operator to two jets , which is challenging to probe directly due to the large qcd backgrounds .
again , stoponium can provide a clear - cut test of light stops in this scenario .
it is worth emphasizing that there is a potentially large source of theoretical uncertainty present in the stoponium production cross section coming from the assumed potential model and contribution of the excited states to the signal .
the limits we present here are based on the empirical quarkonia potential model of hagiwara et al .
@xcite . furthermore , as discussed in ref .
@xcite , it is likely that higher @xmath6-wave stoponium states contribute to the resonance signals , and we therefore include the first two states in our estimate . as we will argue below , we believe our assumptions in this regard are conservative , but we will also discuss in detail how different choices affect the stoponium production cross section and the resulting limits .
we begin in section [ sec : stoponium ] with a brief review on the production and decay of stoponium at hadron colliders .
our new limits and future projections from lhc resonance searches on several benchmark scenarios for each diboson channel are presented in sec .
[ sec : limits ] .
the current limits from stoponium and other indirect constraints on the full stop sector parameter space are combined in sec .
[ sec : interplay ] . in sec .
[ sec : scenarios ] we discuss the implications of our results on two motivated scenarios containing lights stops .
section [ sec : outlook ] presents our conclusions and outlook .
finally we include several appendices [ app : param]-[app : searches ] which collect and summarize our conventions for the stop sector , the indirect constraints on light stops , the stop decay widths , our procedure for limit extrapolation , and the lhc diboson resonance searches used in our analysis .
in this work we will investigate the annihilation decay signatures of the @xmath6-wave @xmath25 stoponium bound state , @xmath26 , at the lhc .
these signatures result from the following processes : @xmath27 searches for stoponium resonances provide a highly complementary probe of light stops , testing different assumptions about susy spectra and parameters than both direct searches for stops and indirect probes such as higgs coupling measurements , precision electroweak data , and vacuum stability constraints . in particular , direct searches for stops ( and other new particles )
are inherently model dependent , requiring assumptions about the superpartner spectrum and stop decay chains .
in contrast , the limits and projections we will derive for stoponium will depend on two _ different _ assumptions : 1 ) the stop does not have an unsuppressed 2-body decay , and 2 ) the value of the trilinear higgs - stop - stop coupling . as we will discuss
shortly , the first assumption of a narrow stop width is actually satisfied in a number of interesting scenarios .
the second assumption on the value of the @xmath28 coupling , which is determined by the stop sector masses and mixing angle , will govern the stoponium branching ratios for the various final states in eq .
( [ eq : stoponium ] ) . in this section
we will collect the relevant ingredients entering into our study of stoponium .
our analysis relies on a number of previous works @xcite , notably those of martin and collaborators @xcite .
next we describe in more detail the basic requirements for observable stoponium annihilation decays . following this
we will review the production of stoponium at the lhc and the stoponium decay branching ratios . before beginning ,
let us note here that we will work exclusively in this paper with the physical stop parameters @xmath29 , in which @xmath30 @xmath31 denotes the lightest ( heaviest ) physical stop state , and a mixing angle @xmath32 corresponds to a purely left - handed ( right - handed ) stop .
we will also restrict to the decoupling limit .
we refer the reader to appendix [ app : param ] for our stop sector conventions .
what conditions are required for the distinctive resonance signatures of stoponium to be observable ?
provided they are kinematically accessible , stop pairs will be copiously produced in @xmath5 collisions . however ,
whether or not the bound state forms and subsequently decays via the annihilation processes in eq .
( [ eq : stoponium ] ) depends on the natural width of the stop .
the bound state formation time is on the order of the inverse binding energy which , assuming a coulombic potential , is given by @xmath33 , where @xmath34^{-1}$ ] is the `` bohr radius '' of the system . therefore , stoponium will form provided the natural width of the stop is smaller than the binding energy .
subsequently , the decay of the bound state may proceed either through the prompt decay of the constituent stop or through the annihilation decays .
for instance , if the @xmath18 coupling is small the dominant annihilation decay is to a pair of gluons , with partial width ( see also sec . [
sec : branching ] below ) @xmath35 where @xmath36 is the radial wavefunction at the origin , for which we employ the parameterization in ref . @xcite . for @xmath37 gev , the factor @xmath38 gev ,
leading to a partial width @xmath39 mev in the light stop mass range of interest . if the annihilation decay width dominates over the natural stop width we can potentially observe the stoponium through its resonance signature . in summary we require , @xmath40 in practice , since the annihilation decay width is typically smaller than the binding energy , it sets the upper bound in eq .
( [ eq : cond ] ) .
in fact , these conditions are satisfied in a number of interesting of scenarios and over a wide range of model parameters , particularly when the stops are light .
we will examine two such scenarios in section [ sec : scenarios ] . at leading order ( lo ) the production of stoponium in @xmath5 collisions at the lhc proceeds via gluon - gluon fusion , with a cross section given by @xcite @xmath41 where @xmath42 is the gluon parton luminosity as a function of @xmath43 , with @xmath44 the squared center - of - mass energy , defined in eq .
( [ eq : pab ] ) in appendix [ app : extrap ] . the partial decay width for @xmath45
is given in eq .
( [ eq : gg ] ) .
next - to - leading order ( nlo ) perturbative qcd corrections to stoponium production have been computed in ref .
@xcite and with the threshold resummation in ref .
@xcite . to account for these corrections
we will use the results of ref .
in particular , we extract the cross section prediction from their fig.9 and furthermore estimate a conservative @xmath46 scale uncertainty on this prediction from their fig.7 @xcite .
we can also reasonably estimate a @xmath47 pdf uncertainty in the mass range of interest , similarly to that for higgs production via gluon - gluon fusion @xcite .
there are likely larger sources of theoretical uncertainties than the qcd scale uncertainty , which are , however , difficult to quantify .
notice in particular that the cross section in eq .
( [ eq : lo ] ) depends on the value of the stoponium wavefunction at the origin squared , @xmath48 , through eq .
( [ eq : gg ] ) , and our imprecise understanding of this matrix element potentially provides a significant source of uncertainty .
@xcite employs the @xmath49 mev parameterization of the wavefunction at the origin from the study of hagiwara et al .
this parameterization is based on an empirical quarkonia potential model , which deviates significantly from the pure coulombic form at large @xmath50 due to the effect of higher order qcd corrections .
for instance , the prediction for @xmath48 for the ground state is smaller by roughly a factor of 2 in comparison to that obtained from a coulomb potential , suggesting that our estimates of the production rate are conservative .
@xcite also present @xmath51 mev and @xmath52 mev parameterizations , which lead to a @xmath53 reduction or enhancement in the stoponium production rate , respectively , giving some sense of the uncertainties involved .
we note that the study of ref .
@xcite is 25 years old , and given the significant experimental and theoretical progress in the studies of quarkonia and qcd in the interim , it would be extremely useful to perform a modern analysis in the spirit of @xcite , with the particular aim of assessing the theoretical uncertainties .
another source of uncertainty in the prediction is the contribution of higher @xmath6-wave stoponium states to the annihilation decay signals .
@xcite includes the first two states in their calculation , but reasonably argues that higher states will likely add to the signal , and in this way our estimates of the production rate are likely conservative . in figure
[ fig : xs ] we display the stoponium production cross section at the lhc under several different assumptions regarding the potential and contribution of excited states .
needless to say , it would be extremely interesting to study the stoponium system in more detail to both quantify the uncertainty due to the modelling of the bound state potential as well as understand the phenomenological consequences of the higher radial states .
tev ( solid blue ) and @xmath54 tev ( solid green ) from ref .
@xcite , which assume the @xmath49 mev parameterization of the stoponium wavefunction at the origin from ref .
@xcite and include the contribution of the first two @xmath6-wave states .
these cross sections are used to derive our limits and projections . for @xmath55
tev we also display for comparison the cross section derived with the quarkonium potential @xcite including instead only the ground state ( blue dashed ) and the first 10 @xmath6-wave states ( blue dotted ) , as well as the cross section assuming a coulomb potential ( red ) ( including only the ground state ) with @xmath56 evaluated at @xmath57 .
, scaledwidth=49.0% ] for the stoponium decay branching ratios , we use the lo results presented in ref .
@xcite for all decay channels except for the @xmath58 channel . for the diphoton channel
we obtain the partial width @xmath59 by multiplying the lo partial width @xmath60 @xcite by the nlo ratio @xmath61 computed in ref
. @xcite .
this approximation reproduces the correct nlo br(@xmath8 ) when br(@xmath62 ) dominates , and it is this case in which the diphoton channel is most important channel is also important in this parameter region , the nlo corrections for this channel are not available . ] .
when other decay modes such as @xmath4 or @xmath3 dominate instead , the diphoton branching ratio is anyway too small to induce an observable signal .
we use the two - loop running @xmath63 and the one - loop running @xmath64 renormalized from @xmath65 and @xmath66 .
( upper ) and stoponium branching ratios ( lower ) for @xmath67 120 gev ( left ) and 160 gev ( right ) as functions of the stop mixing angle . a strong correlation between @xmath68 and the stoponium branching ratios
is observed .
the shaded regions are constrained indirectly by higgs signal strength data , precision electroweak measurements , and vacuum stability as discussed in appendix [ sec : indirect ] .
here we have fixed @xmath69 tev .
note that in the left panel we have terminated the plot at small @xmath70 where the sbottom becomes tachyonic.,title="fig:",scaledwidth=49.0% ] ( upper ) and stoponium branching ratios ( lower ) for @xmath67 120 gev ( left ) and 160 gev ( right ) as functions of the stop mixing angle .
a strong correlation between @xmath68 and the stoponium branching ratios is observed .
the shaded regions are constrained indirectly by higgs signal strength data , precision electroweak measurements , and vacuum stability as discussed in appendix [ sec : indirect ] .
here we have fixed @xmath69 tev .
note that in the left panel we have terminated the plot at small @xmath70 where the sbottom becomes tachyonic.,title="fig:",scaledwidth=49.0% ] + ( upper ) and stoponium branching ratios ( lower ) for @xmath67 120 gev ( left ) and 160 gev ( right ) as functions of the stop mixing angle . a strong correlation between @xmath68 and the stoponium branching ratios
is observed .
the shaded regions are constrained indirectly by higgs signal strength data , precision electroweak measurements , and vacuum stability as discussed in appendix [ sec : indirect ] .
here we have fixed @xmath69 tev .
note that in the left panel we have terminated the plot at small @xmath70 where the sbottom becomes tachyonic.,title="fig:",scaledwidth=49.0% ] ( upper ) and stoponium branching ratios ( lower ) for @xmath67 120 gev ( left ) and 160 gev ( right ) as functions of the stop mixing angle . a strong correlation between @xmath68 and the stoponium branching ratios
is observed .
the shaded regions are constrained indirectly by higgs signal strength data , precision electroweak measurements , and vacuum stability as discussed in appendix [ sec : indirect ] .
here we have fixed @xmath69 tev .
note that in the left panel we have terminated the plot at small @xmath70 where the sbottom becomes tachyonic.,title="fig:",scaledwidth=49.0% ] we will work in the decoupling limit .
the minimal low energy spectrum will contain the stops and possibly the left - handed sbottom .
it is also possible that the spectrum contains a bino - like neutralino lsp , since the condition ( [ eq : cond ] ) is readily satisfied when @xmath71 ( see sec . [ sec : scenarios ] for a detailed discussion of this scenario ) , although the stoponium annihilation decays to neutralinos in this case are always negligible .
we will not consider higgsinos or winos that are lighter than @xmath30 , as in this case we expect the stop to have an unsuppressed two body decay to a lighter chargino and a bottom quark , generically violating the condition ( [ eq : cond ] ) plus missing energy @xcite . ]
all other heavy higgs scalars and superpartners are taken heavy and decoupled and will not influence the stoponium decays .
the mass of the left - handed sbottom is fixed by assuming vanishing sbottom mixing ( see eq .
( [ eq : sbottom ] ) ) .
we emphasize that left - handed sbottoms can have non - negligible effects on the stoponium branching ratios when the lightest stop is primarily left - handed .
as typically happens , unitarity and gauge invariance induce a ( partial ) cancellation in the @xmath3 partial width between @xmath72-channel stop and sbottom exchanges if the stop is mostly left - handed @xcite .
finally , the stoponium decays to @xmath73 and @xmath74 are also subdominant and will not be considered further .
the main decay modes , @xmath75 and @xmath4 , are described well by the three physical stop parameters , @xmath76 , and @xmath77 .
these parameters are defined in appendix [ app : param ] .
the dependence on @xmath78 is weak in most cases , so we fix @xmath79 throughout in this paper .
the trilinear coupling of the 125 gev higgs to the light stops , @xmath68 , plays a crucial role in determining the stoponium decay pattern .
this coupling is written in terms of the physical stop parameters as & = & v \ { s_t^2 c_t^2 } .
[ eq : lamhtt ] due to the last term in eq .
( [ eq : lamhtt ] ) , this coupling can become much larger than the electroweak scale , @xmath80 gev , when the stop mixing is large , inducing sizable partial decay widths for stoponium to @xmath81 via the @xmath44-channel exchange of @xmath82 .
this is illustrated in fig .
[ fig : br ] , where we display the mixing angle dependence of the coupling @xmath68 and the stoponium branching ratios .
a strong correlation is observed between the value of @xmath68 and the pattern of branching ratios .
for large values of mixing , the coupling @xmath68 grows and the stoponium dominantly decays to @xmath3 , @xmath2 , and , if kinematically allowed , @xmath4 .
instead , when @xmath68 is small , the @xmath62 decay dominates and , importantly , the @xmath0 and @xmath10 branching ratios are maximized ( see sec .
[ sec : diphoton ] and sec .
[ sec : zgamma ] for further discussion ) .
large values of @xmath68 are constrained , particularly for light stops , by indirect tests including higgs signal strength measurements , electroweak precision data , and vacuum stability .
these indirect constraints are reviewed in appendix [ sec : indirect ] and shown as the shaded regions in fig .
[ fig : br ] . in particular , we observe that only a small region around @xmath83 is consistent with these indirect constraints for very light stops .
it is therefore very interesting to consider the @xmath84 ( and @xmath10 ) channel as it can provide an independent probe of this unconstrained region of parameter space .
we now derive the current limits on light stops from stoponium annihilation decays and estimate the future reach of each diboson channel in eq .
( [ eq : stoponium ] ) : @xmath85 and @xmath4 . in this section
, we simply assume that the stoponium forms and decays via annihilation , but we will discuss particular models and relevant parameter space in sec .
[ sec : scenarios ] . for each diboson channel , we define and consider the _
ideal _ or _ benchmark _ branching ratio that best represents the optimal or characteristic reach , respectively , over a wide range of the stop masses . based on these benchmarks , we present the limits as a function of the stoponium mass . next , in sec .
[ sec : interplay ] we exhibit the exclusion limits in the general stop parameter space , , as a function of the stop masses and mixing angles , accounting for the realistic branching ratios and comparing with the indirect constraints described in appendix [ sec : indirect ] . our limits and projections are obtained by recasting the results of the latest heavy resonance searches at lhc 7 + 8 tev , which are collected in appendix [ app : searches ] . we will present the strongest limit in each diboson channel . for the future sensitivities , we extrapolate the current expected limits to lhc 14 tev with 30 fb@xmath23 and 3 ab@xmath23 .
our extrapolation method and assumptions on statistical and systematic uncertainties are described in appendix [ app : extrap ] .
scenario defined in sec .
[ sec : diphoton ] .
atlas(right ) and cms(left ) @xmath8 resonance search results are used ; the cms result is extended to a higher mass .
four theoretical predictions are shown as in fig .
[ fig : xs ] : considering the quarkonium - inspired potential of ref .
@xcite and including first two @xmath6-wave states ( thick solid ) , 1s only ( bottom dashed ) , first 10 @xmath6-wave states ( mid - dashed ) as well as considering a 1s with coulomb potential ( top dot - dashed ) . a @xmath86 scale+pdf uncertainty is also shown as a band around the thick line .
vertical shaded regions are excluded if 1s+2s states are considered .
a constant fiducial cut efficiency 0.66 is unfolded for the atlas results.,title="fig:",scaledwidth=49.0% ] scenario defined in sec .
[ sec : diphoton ] .
atlas(right ) and cms(left ) @xmath8 resonance search results are used ; the cms result is extended to a higher mass .
four theoretical predictions are shown as in fig .
[ fig : xs ] : considering the quarkonium - inspired potential of ref .
@xcite and including first two @xmath6-wave states ( thick solid ) , 1s only ( bottom dashed ) , first 10 @xmath6-wave states ( mid - dashed ) as well as considering a 1s with coulomb potential ( top dot - dashed ) . a @xmath86 scale+pdf uncertainty is also shown as a band around the thick line .
vertical shaded regions are excluded if 1s+2s states are considered .
a constant fiducial cut efficiency 0.66 is unfolded for the atlas results.,title="fig:",scaledwidth=49.0% ] scenario defined in sec .
[ sec : diphoton ] .
atlas(right ) and cms(left ) @xmath8 resonance search results are used ; the cms result is extended to a higher mass .
four theoretical predictions are shown as in fig .
[ fig : xs ] : considering the quarkonium - inspired potential of ref .
@xcite and including first two @xmath6-wave states ( thick solid ) , 1s only ( bottom dashed ) , first 10 @xmath6-wave states ( mid - dashed ) as well as considering a 1s with coulomb potential ( top dot - dashed ) . a @xmath86 scale+pdf uncertainty is also shown as a band around the thick line .
vertical shaded regions are excluded if 1s+2s states are considered .
a constant fiducial cut efficiency 0.66 is unfolded for the atlas results.,title="fig:",scaledwidth=49.0% ] scenario defined in sec .
[ sec : diphoton ] .
atlas(right ) and cms(left ) @xmath8 resonance search results are used ; the cms result is extended to a higher mass .
four theoretical predictions are shown as in fig .
[ fig : xs ] : considering the quarkonium - inspired potential of ref .
@xcite and including first two @xmath6-wave states ( thick solid ) , 1s only ( bottom dashed ) , first 10 @xmath6-wave states ( mid - dashed ) as well as considering a 1s with coulomb potential ( top dot - dashed ) . a @xmath86 scale+pdf uncertainty is also shown as a band around the thick line .
vertical shaded regions are excluded if 1s+2s states are considered .
a constant fiducial cut efficiency 0.66 is unfolded for the atlas results.,title="fig:",scaledwidth=49.0% ] the diphoton channel has long been recognized as a promising place to search for stoponium @xcite . at lo ,
the @xmath87 and @xmath88 partial decay widths do not depend on stop sector parameters other than the lightest stop mass , @xmath89 .
they are induced solely by the strong and electromagnetic interactions , such that the ratio , , [ eq : r ] is fixed by gauge couplings and independent of @xmath89 . as discussed in sec .
[ sec : branching ] , in our numerics we include the nlo corrections to the ratio in eq .
( [ eq : r ] ) from ref .
these corrections , along with the running gauge couplings , induce a mild dependence on the lightest stop mass , with the ratio ( [ eq : r ] ) varying from @xmath90 as the lightest stop mass @xmath50 varies from 100 - 400 gev .
the nlo corrections actually reduce the ratio ( [ eq : r ] ) by 35%25% from the lo result @xcite .
since the @xmath62 and @xmath0 decay modes are always present and the size of the corresponding partial decay widths are rather model - independent , the diphoton branching ratio is maximized simply when the other possible decay channels ( @xmath3 , @xmath2 , @xmath4 ) are minimized .
we will therefore define the _
ideal @xmath8 scenario _ by the idealized limit in which the branching ratio @xmath91 is given by the ratio in eq .
( [ eq : r ] ) ( including nlo corrections as discussed above ) .
we note that the idealized branching ratios are @xmath92 @xmath93 times bigger than that of the sm higgs boson with the same mass in the range of @xmath94 gev . in the full
stop parameter space , the ideal diphoton scenario is approximately realized in the region where the higgs - stop - stop coupling @xmath95 is small , corresponding to mixing angles @xmath96 in this region of parameter space the @xmath97 diagrams involving @xmath44-channel higgs exchange are of a similar size or smaller than those mediated by the weak gauge couplings , thus explaining the dominance of the @xmath98 mode .
this can also be seen from the dependence of @xmath95 and the stoponium branching ratios on the stop mixing angle in fig .
[ fig : br ] , where we observe that the @xmath62 branching ratio is nearly 100% when the @xmath68 is small . for the stop mass @xmath99 gev ( @xmath100 tev ) , the ideal @xmath8 scenario
is approximately realized for @xmath101 or @xmath102 .
the actual maximum diphoton branching ratio in these parameter regions is smaller than the ideal one of eq .
( [ eq : r ] ) only by about @xmath103 fraction . in the next section
, we will translate our limits to the full stop parameter space , accounting for the realistic branching ratios .
we note that the ideal @xmath8 scenario lies within the blind spot region consistent with current indirect constraints , even for very light stop masses of order 100 gev .
in particular , the constraints from the higgs signal strength data are relaxed in this region because the coupling @xmath95 is small and the modification of the gluon - gluon fusion cross section from stop loops is minimized .
this makes the @xmath104 channel particularly interesting to study as it can provide a complementary probe of this parameter region . in fig .
[ fig : diphoton ] , we show the current limits from atlas and cms diphoton resonance searches and the future sensitivities for the ideal diphoton scenario .
stoponium masses @xmath105 in the the vertical blue bands are excluded by the observed limits .
we note that the atlas and cms observed limits display complementary patterns of statistical fluctuations , excluding somewhat different stop mass ranges .
the union of the two searches excludes stop masses up to about @xmath106 gev .
we emphasize that these exclusions are based on the assumption of including the first two @xmath6-wave stoponium states in the production rate , as described in sec .
[ sec : prod ] .
this is a conservative assumption as it is likely that higher modes also contribute to the diphoton signal .
following ref .
@xcite , for comparison we have also displayed in fig . [
fig : diphoton ] the cross section including only first @xmath6-wave mode or first 10 @xmath6-wave modes , allowing the reader to see the effect of this uncertainty .
the reader can also see the difference between the two potential models : the thick blue line indicates the empirical quarkonia potential in ref .
@xcite and the dot - dashed red line represents the coulomb potential .
we also show in fig .
[ fig : diphoton ] the 14 tev projections , based on the extrapolation of the current expected limits ; see appendx [ app : extrap ] for a description of the methodology behind our extrapolation . at the early stage of running with 30/fb ,
stops as heavy as about 180 gev can be excluded in the ideal diphoton scenario , while a high luminosity run with 3/ab can exclude the stop up to about 370 gev .
another important channel is @xmath10 since the branching ratio is also maximized when the coupling @xmath68 is small and the indirect constraints are absent . as for the diphoton channel ,
we can write the ratio of @xmath1 to @xmath62 partial widths at lo : ( c_t^2 - s_w^2 ) ^2 ( 1- , ) . [ eq : zgammaratio ] as in the diphoton case , the br(@xmath10 ) is maximized when the br(@xmath62 ) dominates and the other decay modes ( @xmath107 ) are suppressed , corresponding to small @xmath68 .
however , unlike the diphoton case the mixing angle dependence in eq .
( [ eq : zgammaratio ] ) implies that the br@xmath108 is largest when the lightest stop @xmath30 is mostly left - handed .
this asymmetry is also clearly visible in fig .
[ fig : br ] .
we therefore define the _ ideal @xmath10 scenario _ by evaluating br(@xmath10 ) with stop parameters corresponding to @xmath109 and a mostly left - handed @xmath30 . in practice , we choose different mixing angles satisfying this condition for different stop masses . the mixing angle and the br(@xmath10 ) for the ideal @xmath10 scenario gradually increase from @xmath110 to @xmath111 and from 0.84% to 1.37% , respectively , as the stop mass varies in the range @xmath112 gev .
we note that these branching ratios are 100 3000 times larger than those of the sm higgs boson with the same mass in the range of @xmath94 gev .
scenario defined in sec .
[ sec : zgamma ] from atlas(right ) and cms(left ) results . theoretical predictions and excluded regions are as in fig .
[ fig : diphoton ] . a constant fiducial cut efficiency 0.6 is unfolded for atlas results.,title="fig:",scaledwidth=49.0% ] scenario defined in sec .
[ sec : zgamma ] from atlas(right ) and cms(left ) results .
theoretical predictions and excluded regions are as in fig .
[ fig : diphoton ] .
a constant fiducial cut efficiency 0.6 is unfolded for atlas results.,title="fig:",scaledwidth=49.0% ] scenario defined in sec . [
sec : zgamma ] from atlas(right ) and cms(left ) results . theoretical predictions and excluded regions are as in fig .
[ fig : diphoton ] .
a constant fiducial cut efficiency 0.6 is unfolded for atlas results.,title="fig:",scaledwidth=49.0% ] scenario defined in sec .
[ sec : zgamma ] from atlas(right ) and cms(left ) results . theoretical predictions and excluded regions are as in fig .
[ fig : diphoton ] .
a constant fiducial cut efficiency 0.6 is unfolded for atlas results.,title="fig:",scaledwidth=49.0% ] we show the current limits on the ideal @xmath10 scenario from the @xmath113 resonance searches in the upper panels of fig . [
fig : zgamma ] .
notably , the cms and atlas results differ somewhat in the excluded stop mass range ; the cms result is stronger and excludes stops up to about 170 gev in the ideal @xmath10 scenario , while instead the atlas result is sensitive only up to 120 gev stops . in terms of the excluded cross - section the cms result is about 2 - 3 times stronger for the stop mass range shown . since the @xmath10 result can potentially give the strongest constraint in the currently unconstrained parameter space with a mostly left - handed stop and small @xmath68 , we encourage further dedicated analysis between the two experiments .
we also show the 14 tev prospects in the lower panels of fig .
[ fig : zgamma ] . the cms - extrapolated result can potentially probe the stop up to 225 gev with 30/fb and the atlas - extrapolated result can be sensitive to 300 gev with 3/ab . as we will see in fig . [
fig : summary ] in the next section , the diphoton channel is most complementary to the indirect constraints , as it is sensitive to the entire small @xmath68 region including primarily right - handed stops . on the other hand ,
the @xmath10 channel provides stronger limits than the diphoton channel for mostly left - handed stops .
it is important to mention that our @xmath10 results are based on the lo calculation while the @xmath8 result in the previous subsection is from the nlo calculation in ref .
@xcite . for the latter case ,
the nlo effects reduce the ideal @xmath0 branching ratio by 35%25% for the stop mass in the range 100 - 400 gev .
if the same nlo reduction is applied to the ratio of the @xmath10 to @xmath62 partial widths ( hence , to the ideal br(@xmath10 ) ) , the expected exclusion on the stop mass is approximately reduced by about 20 gev .
in addition to the nlo perturbative qcd effects , uncertainties from the number of @xmath6-wave stoponium states contributing to the signal and from potential models are again significant as shown by multiple theoretical predictions in fig .
[ fig : zgamma ] .
the branching ratios br(@xmath2 ) and br(@xmath3 ) are intimately related with the coupling @xmath68 as discussed in sec .
[ sec : branching ] .
the indirect probes of light stops discussed in sec .
[ sec : interplay ] place strong constraints on large values of this coupling and in turn determine the maximum phenomenologically allowed branching ratios br(@xmath2 ) and br(@xmath3 ) , as is depicted in fig .
[ fig : br ] .
scenario with br(@xmath2)=0.06 .
cms @xmath114 search result is used .
, title="fig:",scaledwidth=49.0% ] scenario with br(@xmath2)=0.06 .
cms @xmath114 search result is used .
, title="fig:",scaledwidth=49.0% ] scenario with br(@xmath3)=0.12 .
cms @xmath115 search result is used .
, title="fig:",scaledwidth=49.0% ] scenario with br(@xmath3)=0.12 .
cms @xmath115 search result is used .
, title="fig:",scaledwidth=49.0% ] unlike the diphoton channel discussed above , there are several reasonable ways to define the benchmark @xmath2 , @xmath3 scenarios .
for instance , one could simply take the largest theoretically allowed branching ratio , or alternatively the largest branching ratio consistent with indirect constraints ; in both cases one is led to vary the mixing angle as a function of the stop masses .
one could also choose a fixed mixing angle for a range of the stop mass .
for simplicity , we will instead choose a prescription with a fixed branching ratio that is broadly consistent with the current indirect constraints .
referring to fig .
[ fig : br ] of @xmath116 gev case , we define the _ benchmark @xmath2 , @xmath3 scenarios _ by choosing br(@xmath2)=6% and br(@xmath3)=12% .
in fact , for the range of @xmath70 allowed for the 160 gev stop we obtain maximum br(@xmath117 4 - 8% and br(@xmath118 11 - 16% for the stop mass @xmath119 gev ( with @xmath120 tev ) .
thus , the maximum branching ratios in this range of @xmath70 do not vary much with the stop mass .
the current indirect constraints are somewhat weaker for heavier stops and a wider range of @xmath70 is viable , but we envisage that indirect constraints will also improve with more data , probing a broader range of mixing angles and heavier stop mass .
we refer to the @xmath121 gev case as a representative benchmark scenario in view of the interplay between stoponium and indirect constraints .
again , we provide limits on the full stop parameter space accounting for true branching ratios in the next section .
using a fixed branching ratio as a benchmark scenario for a wide range of @xmath122 gev stop masses is also consistent with the softened goldstone enhancement of the stop decay into longitudinal @xmath123 and @xmath124 bosons , which could have sensitively increased the branching ratio with the stop mass .
the decay into gauge bosons are maximized when the amplitude is dominated by the @xmath44-channel exchange of higgs boson , corresponding to values of the @xmath68 coupling strong enough to dominate the decay but weak enough to be consistent with indirect constraints . but
this contribution softens the goldstone enhancement factor @xmath125 of the amplitude by a suppression @xmath126 from the propagator , leaving the dominant amplitude relatively insensitive to the stop mass for the benchmark parameter space . in fig .
[ fig : zz ] and fig .
[ fig : ww ] we show the current limits and prospects of the benchmark @xmath2 and @xmath3 scenarios from the @xmath127 and the @xmath115 channels , respectively .
the constraint from the @xmath2 channel is stronger , just as it is for the sm higgs .
the @xmath128 channel gives similar bounds as the @xmath129 channel but is not shown for simplicity .
currently , the @xmath2 channel constrains stops up to about 110 gev and a small range of 121 - 126 gev for the benchmark scenario . in the future
, the @xmath2 channel is expected to probe 140 ( 250 ) gev stops at @xmath54 tev with 30/fb ( 3/ab ) for the @xmath2 benchmark scenario .
while there are no limits yet , the lhc at 14 tev with 3/ab is expected to probe 180 gev stops in the benchmark @xmath3 scenario .
we note that in the full stop parameter space , with realistic branching ratios , there are in fact new limits from the @xmath3 channel for very light stops , where the @xmath4 mode is absent . similarly to the @xmath2 and @xmath3 branching ratios , the maximum branching ratio br@xmath130 is limited by indirect constraints on the coupling @xmath68 .
we define the _ benchmark @xmath4 scenario _ by a fixed br(@xmath4)=0.25 .
the actual maximal br(@xmath4 ) for the range of @xmath70 currently allowed for the 160 gev stop gradually increases from 16% to 41% for the 130 gev to 300 gev stop . in fig .
[ fig : hh ] , we display the current sensitivity and 14 tev projections on the benchmark @xmath4 scenario from the @xmath131 resonance search .
no new limits can be obtained with this channel from the run 1 data . however , in the long term this channel is expected to probe stops as heavy as 215 gev in the benchmark @xmath4 scenario at @xmath132 tev with 3/ab .
scenario with br(@xmath4)=0.25 .
atlas @xmath131 search result is used.,title="fig:",scaledwidth=49.0% ] scenario with br(@xmath4)=0.25 .
atlas @xmath131 search result is used.,title="fig:",scaledwidth=49.0% ] -@xmath22 plane for fixed @xmath133 gev , including @xmath8 ( blue hatched ) , @xmath113 ( green ) , @xmath127 ( red ) , @xmath115 ( orange ) .
we also display the union of indirect constraints ( gray ) shown in fig .
[ fig : indirect ] . the strongest limit from atlas or cms in each channel are shown .
finally , we indicate the stop parameters for which @xmath109 ( red dashed contour).,scaledwidth=92.0% ]
our main results are presented in fig . [
fig : summary ] , which displays the constraints from diboson resonance searches on stoponium in the physical stop parameter space .
we show six slices of the parameter space , fixing the lightest stop mass @xmath133 gev and showing the limits in the @xmath70-@xmath22 plane .
the individual constraints from the @xmath0 , @xmath10 , @xmath2 , and @xmath3 channel are represented ( no bounds from @xmath4 channel can be obtained from the current data ) . for comparison
we also display the region excluded by indirect constraints coming from higgs signal strength measurements , electroweak precision data , and vacuum stability ( see appendix [ sec : indirect ] ) .
in contrast to the previous section in which we defined ideal or benchmark branching ratios , here we account for the realistic branching ratios , computed at each point in the physical stop parameter space as described in sec . [ sec : branching ] . assuming the stoponium forms and annihilate decays , we observe that light stops with masses @xmath134 gev are excluded by the combination of diboson resonance searches and the indirect constraints .
stoponium searches alone probe stops lighter than about 125 gev .
remarkably , a strong complementarity is seen between the @xmath0 resonance searches and the indirect constraints ( particularly the higgs coupling constraints ) , with the @xmath0 channel probing the remaining open stop blind spot region corresponding to small or vanishing coupling @xmath68 .
moreover , we also see that primarily left - handed stops with masses @xmath135 gev are constrained by @xmath10 resonance searches .
again , we emphasize that the limits we present here rely on the @xmath136 mev parameterization of the stoponium wavefunction at the origin from ref .
@xcite and the inclusion of the first two @xmath6-wave states @xcite . while it is likely that this estimate is conservative , particularly in light of the possible additional signal coming from the higher excited states , it is worthwhile to examine how the limits change if different assumptions are made .
for example , if we instead include only the contribution of the 1s state , the expected limit on the stoponium mass will be weaker by about 20 gev , as can be seen by examining figs .
[ fig : diphoton ] and [ fig : zgamma ] . alternatively , if one assumes the pure coulomb potential , the expected limit on the stoponium mass will be stronger by about 70 gev . in any case
, it is clear that stoponium is now probing the hypothesis of light stops beyond the lep limits and will explore uncharted territory with the next run of the lhc .
we now discuss two motivated examples scenarios in which stoponium provides a complementary probe to direct stop searches and indirect tests .
the condition ( [ eq : cond ] ) will generically hold when the stop does not possess an unsupressed 2-body decay .
this can happen due to kinematics or small couplings , and to demonstrate this we will focus two examples - 1 ) a canonical " case of r - parity conservation with a bino lsp and a stop nlsp , and 2 ) a r - partiy violating case with a stop lsp which decays to a pair of jets through the @xmath12 operator . a canonical benchmark scenario with light stops consists of a stop nlsp and a neutralino lsp , under the assumption that r - parity is conserved . in this case
the neutralino is stable , leading to the signature of missing energy at colliders . as is well known , in this scenario the stop width is very sensitive to the mass spectrum , and several kinematic regions can be distinguished : a. @xmath137 : the stop decays via the 2-body process @xmath138
. b. @xmath139 : the stop decays via the 3-body process @xmath140 . c. @xmath141 : the stop decays via the 4-body process @xmath142 , or possibly through the flavor - violating 2-body decay @xmath143 . to illustrate these features , in the left panel of fig .
[ fig : stform ] we have plotted the decay width of the stop as a function of its mass for the case of a 10 gev bino lsp ( see the appendix [ app : decay ] for the 3-body partial decay width of the stop ) . for comparison , in the same plot
we show the stoponium binding energy and annihilation decay width . as the stop
mass approaches the boundary between 2- and 3-body decays , we observe that @xmath144 becomes larger than the natural stop width . therefore , in region i above , the stop generally decays too quickly for stoponium resonance signatures to be visible , although it may be possible near the edge of region i where phase space suppression is relevant .
however , in regions ii and iii the dominant stop decay channels are 3- and 4-body ( or suppressed 2-body ) , respectively , allowing both sufficient time for the formation of the stoponium and the dominance of the bound state annihilation decays .
gev , @xmath100 tev , and @xmath145 .
the vertical line separates the two- and three - body decays of the stop . *
( right ) : * in the red shaded region above each red line , the stoponium annihilation is quicker than the individual stop decay , and searches for stoponium resonances are possible .
three benchmark scenarios are presented as red lines : the top two have almost ideal @xmath8 branching ratios ( defined in sec .
[ sec : diphoton ] ) with @xmath146(top ) , 0.18(middle ) , and the bottom one with @xmath147(bottom ) . for comparison
, we also display a union of atlas limits on the stops ( orange dotted ) , which assume 100% branching ratios to the final states under consideration @xcite as well as a branching ratio - independent exclusion limit taken from ref .
@xcite ( orange shaded ) .
the blue - hatched region is excluded from stoponium fig .
[ fig : summary].,title="fig:",scaledwidth=47.0% ] gev , @xmath100 tev , and @xmath145 .
the vertical line separates the two- and three - body decays of the stop . *
( right ) : * in the red shaded region above each red line , the stoponium annihilation is quicker than the individual stop decay , and searches for stoponium resonances are possible .
three benchmark scenarios are presented as red lines : the top two have almost ideal @xmath8 branching ratios ( defined in sec .
[ sec : diphoton ] ) with @xmath146(top ) , 0.18(middle ) , and the bottom one with @xmath147(bottom ) . for comparison
, we also display a union of atlas limits on the stops ( orange dotted ) , which assume 100% branching ratios to the final states under consideration @xcite as well as a branching ratio - independent exclusion limit taken from ref .
@xcite ( orange shaded ) . the blue - hatched region is excluded from stoponium fig .
[ fig : summary].,title="fig:",scaledwidth=50.0% ] therefore , regions ii and iii are prime targets for resonance searches for stoponium , which is clearly illustrated by the `` phase '' diagram in the stop
neutralino mass plane shown in the right panel of fig .
[ fig : stform ] .
we represent the current exclusion limit from stoponium resonances by the blue hatched region ; this region is excluded for the given lightest stop mass for all other stop parameters ( see fig .
[ fig : summary ] ) . for comparison
, we have overlaid the current atlas limits from direct stop searches @xcite , represented by the orange dotted line .
while these direct limits naively appear to cover most of light stop region , it is important to emphasize that they rely on the assumption of a 100@xmath148 branching ratio to the final state under consideration .
if this assumption is relaxed , the limits are weakened considerably .
for example , in region iii if the stop has comparable branching ratios in the 4-body @xmath149 and 2-body @xmath150 channels , a significant portion of parameter space opens up @xcite , and only the orange - shaded region can be excluded regardless of the assumption on the branching ratio .
in contrast , constraints from stoponium resonances are already starting to unambiguously probe this region .
it is also worthwhile to note that the `` stealth '' stop region the phase space - suppressed part of region i near @xmath151 , @xmath152 is also potentially amenable to stoponium searches ( within the red - shaded region in the bottom part of fig .
[ fig : stform ] ) this region is challenging to probe with direct stop searches , and several dedicated observables such as spin correlations @xcite , top pair production rates @xcite and special kinematic variables @xcite have been studied .
however , they primarily rely on precision measurements and calculations , which are often subject to subtle uncertainties .
on the other hand , the stoponium resonance searches are cleaner and less ambiguous in most cases .
thus , the stoponium can provide an important complementary probe of the stealth region . in this region
, stops with a mostly left - handed mixture ( small @xmath70 ) can be probed via stoponium over a wider range of stop masses ( see fig .
[ fig : stform ] right ) because the individual stop decay width is smaller with a smaller @xmath70 . in any case , stoponium does not yet have sensitivity to stealth stops ( the blue - hatched region extends to about 130 gev ) , but can potentially be sensitive to a significant portion of this challenging region in the early part of run 2 with @xmath153/fb of data ( see fig .
[ fig : diphoton ] ) .
let us finally comment on the case in which the lsp is a very weakly interacting particle such as the gravitino . in this scenario ,
the stop nlsp is typically long lived due to its weak coupling to the gravitino , and therefore stoponium annihilation decays can be relevant .
for instance , taking a low supersymmetry breaking scale @xmath154 tev ( corresponding to an essentially massless gravitino ) , a 200 gev stop has a decay width @xmath155 gev @xcite , which is much smaller than the typical stoponium annihilation width . in practice , however , this scenario is already strongly constrained by direct searches covering both prompt and displaced decays @xcite , and only a small window around the stealth stop parameter region for prompt decays is still open .
the stoponium searches will be able to fill this open region .
gev ( red ) and @xmath156 gev ( blue ) for typical values of stoponium binding energies and annihilation decay widths .
we also display a contour of the stop lifetime @xmath157 mm .
stops with lifetimes in this range and longer are subject to stringent lhc searches for long - lived charged particles @xcite .
stops with prompt rpv decays in the red - shaded region are also constrained from rpv searches @xcite .
stoponium can therefore be relevant for a broad range of rpv couplings , @xmath158 , and the current exclusion from stoponium resonances is shown as blue - hatched .
, scaledwidth=49.0% ] another benchmark scenario of interest is a stop lsp that decays to a pair of jets due to a small @xmath12 r - parity violating coupling , which we customarily denote as @xmath159 .
such a rpv stop is challenging to probe with direct searches at lhc . despite its large production cross section ,
the signature of four jets ( paired dijet resonances ) is difficult to disentangle from the qcd background , and stops as light as @xmath160 gev are still allowed @xcite although some range of the stop masses @xmath161 gev has recently been excluded @xcite .
stoponium resonance searches provide another means of probing this interesting scenario .
since the decay of the stop is through a 2-body process , the rpv coupling must be somewhat suppressed in order for the condition ( [ eq : cond ] ) to apply .
in fact , this may be motivated in explicit models of r - parity violation which aim to protect the lifetime of the proton through other symmetries .
for example , in the scenario of mfv susy @xcite , the dominant rpv coupling is @xmath162 , which is small enough for ( [ eq : cond ] ) to hold .
note also that if the rpv coupling is extremely small , @xmath163 then the stop will have a macroscopic lifetime . in this case , there are additional strong constraints from searches for displaced dijets @xcite or long - lived colored / charged particles @xcite . in fig .
[ fig : decayrpv ] we display the rpv stop lsp phase diagram .
we conclude that stoponium resonance searches can provide new complementary probe of this scenario for a broad range of couplings , @xmath164 .
a natural supersymmetric solution to the hierarchy problem predicts light stops , which can lead to a variety of novel phenomena . besides the suite of direct searches at the lhc ( which depend on the susy spectrum ) and indirect tests such as higgs couplings , precision electroweak data and vacuum stability requirements , it is possible that light stops can manifest through a stoponium annihilation decay leading to a resonance signature in the @xmath165 and @xmath4 channels
this will happen if the stop does not have an unsuppressed two body decay and naturally occurs in several motivated scenarios .
in this work we have derived new limits on light stops from atlas and cms diboson resonance searches using run 1 data .
our limits are derived using the empirical quarkonia potential model of @xcite and assuming the first two @xmath6-wave states contributing to the resonance signal , which we have argued is conservative . assuming the stoponium can form and annihilate decay
, light stops below about 130 gev are excluded by the combination of constraints on stoponium resonances and other indirect constraints , such as higgs couplings measurements .
notably , we have demonstrated that the @xmath0 channel is especially complementary to these indirect probes , as the resonance signature in this channel is maximized in the blind spot region where the higgs - stop - stop trilinear coupling is small and these constraints are weakest .
furthermore , @xmath10 resonance searches provide even stronger limits for primarily left handed stops , limiting stop masses below about 170 gev . in the long term , the lhc experiments will be able to probe stops in the 300 - 400 gev range with a high - luminosity run with 3/ab at 14 tev .
we have also discussed the implications of these searches for two specific susy scenarios with lights stops . in the case of a bino lsp and stop nlsp , stoponium annihilation decays can be relevant since the stop width is naturally suppressed over a wide range of parameters .
stoponium resonance searches provide a relatively clean and unambiguous probe , and can give a handle on the challenging compressed and stealth stop regions of parameter space .
instead for the case of a r - parity violating stop lsp decaying to two jets , stoponium already provides unique sensitivity to stops lighter than about 130 gev and rpv couplings in the range @xmath166 , which is difficult to probe directly due to the large qcd backgrounds . looking forward
, it is clearly of interest on the theoretical side to bring under better control the uncertainties coming from our imprecise understanding of the stoponium bound state system . a modern study of empirical quarkonia potential models along the lines of ref .
@xcite is warranted , and perhaps lattice studies could provide further insight into the non - perturbative matrix elements entering into stoponium production and decays . a more detailed investigation into the decay patterns of the excited states would also be helpful in order to understand their contribution to the resonance signals and perhaps uncover additional signatures of stoponium . on the experimental side , stoponium clearly provides a well motivated target for resonance searches , which should be a high priority during the next run of the lhc .
we thank roberto franceschini , seyong kim , adam leibovich , florian staub , brock tweedie and yeo woong yoon for helpful discussions .
we are also grateful to guillelmo gomez ceballos , louis helary , elisabeth petit , marco pieri , pierre savard , and livia soffi for helpful discussions and correspondence regarding the atlas and cms analyses .
bb is supported in part by a cern cofund fellowship .
sj is supported in part by the national research foundation of korea under grant 2013r1a1a2058449 .
sj thanks the mainz institute for theoretical physics and cern theory group where some of this work was done .
in terms of the gauge eigenstates @xmath167 , the stop mass matrix is given at tree level by @xmath168 where @xmath169 , @xmath170 are the left- and right - handed squark soft mass parameters , @xmath171 , with @xmath172 the soft trilinear coupling , @xmath173 the supersymmetric higgs mass parameter , and @xmath174 the ratio of up and down type higgs vacuum expectation values , and @xmath175 , @xmath176 . for simplicity , we will assume all parameters are real .
the physical stop mass eigenstates are related to the gauge eigenstates through the orthogonal transformation : @xmath177 where the mixing angle @xmath70 satisfies @xmath178 which we take in the range @xmath179 $ ] .
the sbottom sector is described in a similar fashion .
the sbottom mass matrix can be obtained from eq .
[ eq : stop - mass ] with the replacements @xmath180 , @xmath181 , @xmath182 , @xmath183 , and @xmath184 .
in this appendix we summarize the existing indirect constraints on light stops coming from higgs signal strength measurements , precision electroweak data , and vacuum stability . in the mssm , there will be additional constraints on the stop parameters coming from the requirement of obtaining a 125 gev higgs mass , but we will remain open to new contributions to the higgs quartic coming from physics beyond the mssm . furthermore ,
if higgsinos or gluinos are light there can be further constraints from flavor physics , which we will not include here .
several summary plots of these constraints are shown in fig .
[ fig : indirect ] .
light stops modify the @xmath185 and @xmath186 couplings at one loop and are therefore subject to constraints by atlas and cms measurements of the higgs signal strength parameters @xcite . to derive the exclusions , we use the method of ref .
@xcite , with the recent updates in ref .
@xcite , which include the latest run 1 results as of summer 2014 . in this approach
, the raw signal strength data are combined to derive a @xmath187 function for the @xmath0 , @xmath188 , and @xmath189 channels as a function of two combined production mode signal strengths : 1 ) gluon - gluon fusion and @xmath190 ( ggf+tth ) , and 2 ) vector boson fusion and associated production ( vbf+vh ) .
this yields eight composite signal strengths , @xmath191 . in terms of the ratios @xmath192 ,
the signal strength predictions are given by @xmath193 where we have defined @xmath194 with br@xmath195 .
with these ingredients we construct the @xmath187 function . as @xmath208 and @xmath209
are correlated , we define a one parameter best - fit region in the physical stop parameter space at 2@xmath210 c.l . by demanding @xmath211 .
where @xmath214 and we have defined the function @xmath215 we have used the left - handed sbottom mass , @xmath216 , in eq .
( [ eq : sbottom ] ) with the zero sbottom mixing .
we apply the constraint from ref .
@xcite : @xmath217 in the field space of @xmath218 , we study the ( meta)stability conditions for our electroweak vacuum characterized by @xmath219 gev and vanishing stop vevs @xcite .
the tree - level potential that we use is ( in terms of real scalar fields with @xmath220 normalizations factored out ) v^0 & = & m_h^2 h^2 m_q^2 t_l^2 m_u^2 t_r^2 ( ( t_l t_r)^2 ( h t_l)^2 ( h t_r)^2 ) + & + & ( h^2 + ^2 - ^2 ) ^2 ( h^2 ^2 ) ^2 ( t_l^2 ^2 ) ^2 + & + & h t_lt_r h.c . h^4 .
we have assumed that the effects of @xmath221 are negligible , appropriate in the large @xmath78 regime , and have substituted @xmath222 , @xmath223 , and @xmath224 . to have a proper electroweak vacuum and the measured higgs mass
, we set @xmath225 , @xmath226 gev , and @xmath227 .
in particular , we assume the presence of some physics , either radiative stop corrections in the mssm or physics beyond the mssm , that generates an appropriate correction to the higgs quartic coupling , @xmath228 . to arrive at the above form of the @xmath229 terms , we assume that the stops are aligned or anti - aligned in the @xmath230 color space
. the negative @xmath231-linear term can induce new minima when all three scalar fields obtain vevs .
we numerically scan the field space to determine all local minima , and we use the ` cosmotransitions ` program @xcite to calculate the tunnelling rate between our electroweak minimum and any deeper charge - color breaking minima . the classical action for this tunneling , @xmath232 is required for the metastability of our vacuum .
we find that the parameter space with large values of the higgs - stop - stop coupling , @xmath68 , is constrained from the requirement of vacuum ( meta)stability .
our vacuum stability bounds rely on the simplifying assumptions discussed above , and it is possible that stronger constraints can be obtained if some of these assumptions are relaxed ( e.g. @xmath221 vevs or misaligned stop vevs ) , or when one accounts for one - loop corrections to the potential and thermal effects ( for a useful comparative study , see ref .
@xcite ) . nevertheless we believe our bounds give a qualitatively correct representation of the current limits , and we observe that for light stops the vacuum stability limits are generally weaker than those coming from higgs couplings .
when the left - handed stop is light , we also check whether the mass of the accompanying left - handed sbottom is positive ( and evading lep bounds ) . for
the given stop parameters and heavy enough right - handed sbottoms , the left - handed sbottom is the heaviest when the sbottom mixing is zero ( the equality below holds in this case ) m_b_l^2 ( m_t_1 ^ 2 c_t^2 + m_t_2 ^ 2 s_t^2 ) - m_t^2 - c_2 m_w^2 .
[ eq : sbottom ] throughout in this work , we use this left - handed sbottom mass with the equality sign by assuming the zero sbottom mixing .
= ( 100,130,160,200,300 ) gev the constraints in the @xmath70 - @xmath22 plane coming from higgs signal strength measurements ( gray ) , precision electroweak data ( orange ) , vacuum stability ( yellow ) , tachyonic sbottoms ( blue ) and sbottoms below the lep kinematic reach ( light blue ) .
the white regions are currently unconstrained .
the red dashed line indicates where the higgs - stop - stop coupling @xmath68 vanishes .
here we have fixed @xmath233 and the sbottom mixing to zero ( @xmath234).,title="fig:",scaledwidth=40.0% ] = ( 100,130,160,200,300 ) gev the constraints in the @xmath70 - @xmath22 plane coming from higgs signal strength measurements ( gray ) , precision electroweak data ( orange ) , vacuum stability ( yellow ) , tachyonic sbottoms ( blue ) and sbottoms below the lep kinematic reach ( light blue ) .
the white regions are currently unconstrained .
the red dashed line indicates where the higgs - stop - stop coupling @xmath68 vanishes .
here we have fixed @xmath233 and the sbottom mixing to zero ( @xmath234).,title="fig:",scaledwidth=40.0% ] + = ( 100,130,160,200,300 ) gev the constraints in the @xmath70 - @xmath22 plane coming from higgs signal strength measurements ( gray ) , precision electroweak data ( orange ) , vacuum stability ( yellow ) , tachyonic sbottoms ( blue ) and sbottoms below the lep kinematic reach ( light blue ) .
the white regions are currently unconstrained .
the red dashed line indicates where the higgs - stop - stop coupling @xmath68 vanishes .
here we have fixed @xmath233 and the sbottom mixing to zero ( @xmath234).,title="fig:",scaledwidth=40.0% ] = ( 100,130,160,200,300 ) gev the constraints in the @xmath70 - @xmath22 plane coming from higgs signal strength measurements ( gray ) , precision electroweak data ( orange ) , vacuum stability ( yellow ) , tachyonic sbottoms ( blue ) and sbottoms below the lep kinematic reach ( light blue ) .
the white regions are currently unconstrained .
the red dashed line indicates where the higgs - stop - stop coupling @xmath68 vanishes .
here we have fixed @xmath233 and the sbottom mixing to zero ( @xmath234).,title="fig:",scaledwidth=40.0% ] + = ( 100,130,160,200,300 ) gev the constraints in the @xmath70 - @xmath22 plane coming from higgs signal strength measurements ( gray ) , precision electroweak data ( orange ) , vacuum stability ( yellow ) , tachyonic sbottoms ( blue ) and sbottoms below the lep kinematic reach ( light blue ) .
the white regions are currently unconstrained .
the red dashed line indicates where the higgs - stop - stop coupling @xmath68 vanishes .
here we have fixed @xmath233 and the sbottom mixing to zero ( @xmath234).,title="fig:",scaledwidth=40.0% ]
the partial decay width for @xmath236 can be written as @xmath237 where we have defined the off - shell " partial widths : @xmath238 \nonumber \\ \gamma_{t^*\rightarrow w b}(s ) & = & \ ! \frac{n_c g^2 |v_{tb}|^2 } { 192 \pi}\frac{s^{3/2}}{m_w^2 } \lambda^{1/2}\left ( 1 , \frac{m_w^2}{s } , \frac{m_b^2}{s}\right ) \left [ \lambda\left(\ ! 1,\frac{m_w^2}{s},\frac{m_b^2}{s}\right)\ ! + \ !
3\frac{m_w^2}{s } \left(\!1\!- \ ! \frac{m_w^2}{s}\ ! + \ !
\frac{m_b^2}{s}\right ) \right ] , \nonumber \\ & & \label{eq:3body } \end{aligned}\ ] ] here , we have defined the couplings , @xmath239 with @xmath240 the neutralino mixing matrix elements , and the kinematic function @xmath241 .
notice that the three - body partial decay width reduces to the two - body one in the appropriate kinematical regime through the use of the narrow - width approximation .
to estimate the future prospects for probing stoponium at the lhc we extrapolate the current _ expected _ limits of the diboson resonance searches to @xmath242 tev with integrated luminosities @xmath243 of 30/fb and 3/ab . to accomplish this ,
we employ several simplifying assumptions . for the uncertainties we consider two extreme cases : ( 1 ) statistical uncertainties dominate , and ( 2 ) systematic uncertainties dominate and improve in proportion to @xmath244 .
the results in the main text of section [ sec : limits ] are obtained with the former assumption , but we will offer a comparison between the two cases below . at a specified collision energy and integrated luminosity denoted collectively by a superscript @xmath245 , we have = , [ eq : stat ] where @xmath246 is the upper bound on the signal cross - section , @xmath247 is the background cross - section , and @xmath248 is the number of background events after all cuts .
the latter two quantities are related as @xmath249 , where @xmath250 is the partonic background cross - section ( which does not depend on @xmath245 ) , @xmath251 is the integrated luminosity , @xmath252 is the selection efficiency , and @xmath253 is the parton luminosity for initial partons @xmath254 , defined as ^i_ab _ ^i^1
dx f_a(x , q^2 ) f_b(^i / x , q^2 ) , ^i .
[ eq : pab ] in eq .
( [ eq : stat ] ) we have taken @xmath255 since the dominant backgrounds in the most sensitive channels , @xmath0 and @xmath2 , arise from @xmath256 initial states . in our numerics we take the up quark parton luminosity for simplicity , as it will dominate in @xmath5 collisions .
we furthermore assume a constant efficiency , @xmath257 , which is reasonable since the signal - to - background ratio near a resonance is largely determined by the resonance mass .
this is similar to gluino pair searches @xcite and perhaps more generally useful @xcite .
taking ratios , we can extrapolate the current bounds to those for other collision energies and luminosities in a simple way in terms of parton luminosity ratio and luminosity ratio = = = .
[ eq : statextra ] we assume that systematic uncertainties improve with 1/@xmath244 , again taking ratios , we are able to extrapolate the current limits : = = .
[ eq : sysextra ] comparing eq .
( [ eq : statextra ] ) and eq .
( [ eq : sysextra ] ) , we see that if errors are dominantly from systematics , one obtains a weaker cross - section bound by a factor @xmath258 .
numerically , this equals to 1.45 and 1.73 for @xmath259 160 , 400 gev at 14 tev collision for @xmath260 . these results
roughly show how sensitively the future bounds may depend on the assumptions on errors .
in table [ tab : dataused ] , we collect the latest experimental results on diboson resonance searches .
the searches shown in bold give the strongest constraints in each diboson channel and are used to derive limits on stoponium .
the cms and atlas @xmath8 limits are similar but constrain complementary mass ranges due to different statistical fluctuations , and we therefore show both results .
the cms @xmath1 is found to be somewhat stronger than that of atlas , but due to the importance of this channel , we show both results and encourage more careful analysis .
all searches use 8 tev datasets except for the ones with @xmath26124/fb and @xmath26110/fb , in which case the combined 7 + 8 tev dataset is used .
the cms @xmath114 result is stronger than the atlas one partly because cms presents combined results of all production channels and uses the 7 + 8 tev dataset .
the @xmath128 result has similar sensitivity to that of the @xmath262 , but we show only the latter result for simplicity .
the searches listed in the last four lines give weaker constraints .
lastly , atlas @xmath8 and @xmath10 present limits on the fiducial cross - section , so we assume and unfold a constant fiducial efficiency for each channel as mentioned in the relevant part of text . s. dimopoulos and g. f. giudice , phys . lett .
b * 357 * , 573 ( 1995 ) [ hep - ph/9507282 ] .
a. g. cohen , d. b. kaplan and a. e. nelson , phys .
b * 388 * , 588 ( 1996 ) [ hep - ph/9607394 ] . c. brust , a. katz , s. lawrence and r. sundrum , jhep * 1203 * , 103 ( 2012 ) [ arxiv:1110.6670 [ hep - ph ] ] .
m. papucci , j. t. ruderman and a. weiler , jhep * 1209 * , 035 ( 2012 ) [ arxiv:1110.6926 [ hep - ph ] ] .
g. aad _ et al . _ [ atlas collaboration ] , jhep * 1406 * , 124 ( 2014 ) [ arxiv:1403.4853 [ hep - ex ] ] .
g. aad _ et al . _ [ atlas collaboration ] , jhep * 1409 * , 015 ( 2014 ) [ arxiv:1406.1122 [ hep - ex ] ] .
g. aad _ et al . _ [ atlas collaboration ] , arxiv:1406.5375 [ hep - ex ] .
g. aad _ et al . _ [ atlas collaboration ] , arxiv:1407.0583 [ hep - ex ] .
g. aad _ et al . _ [ atlas collaboration ] , phys .
d * 90 * , 052008 ( 2014 ) [ arxiv:1407.0608 [ hep - ex ] ] .
the atlas collaboration , atlas - conf-2014 - 056 , atlas - com - conf-2014 - 076 .
cms collaboration [ cms collaboration ] , cms - pas - sus-13 - 009 .
s. chatrchyan _ et al . _
[ cms collaboration ] , eur .
j. c * 73 * , 2677 ( 2013 ) [ arxiv:1308.1586 [ hep - ex ] ] .
cms collaboration [ cms collaboration ] , cms - pas - sus-13 - 015 .
cms collaboration [ cms collaboration ] , cms - pas - sus-14 - 011 .
https://twiki.cern.ch/twiki/pub/cmspublic/susysmssummaryplots8tev/t2tt_ichep2014_all.pdf s. p. martin , phys .
d * 78 * , 055019 ( 2008 ) [ arxiv:0807.2820 [ hep - ph ] ] .
j. fan , m. reece and j. t. ruderman , jhep * 1111 * , 012 ( 2011 ) [ arxiv:1105.5135 [ hep - ph ] ] .
j. a. evans and y. kats , jhep * 1304 * , 028 ( 2013 ) [ arxiv:1209.0764 [ hep - ph ] ] .
y. bai , a. katz and b. tweedie , jhep * 1401 * , 040 ( 2014 ) [ arxiv:1309.6631 [ hep - ph ] , arxiv:1309.6631 ] . w. porod and t. wohrmann , phys . rev .
d * 55 * , 2907 ( 1997 ) [ erratum - ibid .
d * 67 * , 059902 ( 2003 ) ] [ hep - ph/9608472 ] .
m. carena , a. freitas and c. e. m. wagner , jhep * 0810 * , 109 ( 2008 ) [ arxiv:0808.2298 [ hep - ph ] ] .
m. a. ajaib , t. li and q. shafi , phys .
d * 85 * , 055021 ( 2012 ) [ arxiv:1111.4467 [ hep - ph ] ] .
m. drees , m. hanussek and j. s. kim , phys .
d * 86 * , 035024 ( 2012 ) [ arxiv:1201.5714 [ hep - ph ] ] .
t. plehn , m. spannowsky and m. takeuchi , jhep * 1208 * , 091 ( 2012 ) [ arxiv:1205.2696 [ hep - ph ] ] .
d. s. m. alves , m. r. buckley , p. j. fox , j. d. lykken and c. t. yu , phys .
d * 87 * , no .
3 , 035016 ( 2013 ) [ arxiv:1205.5805 [ hep - ph ] ] . c. brust , a. katz and r. sundrum , jhep * 1208 * , 059 ( 2012 ) [ arxiv:1206.2353 [ hep - ph ] ]
a. choudhury and a. datta , mod .
a * 27 * , 1250188 ( 2012 ) [ arxiv:1207.1846 [ hep - ph ] ] . z. h. yu , x. j. bi , q. s. yan and p. f. yin , phys .
d * 87 * , no .
5 , 055007 ( 2013 ) [ arxiv:1211.2997 [ hep - ph ] ] . m. l. graesser and j. shelton , phys .
lett . *
111 * , no .
12 , 121802 ( 2013 ) [ arxiv:1212.4495 [ hep - ph ] ] .
k. krizka , a. kumar and d. e. morrissey , phys .
d * 87 * , no .
9 , 095016 ( 2013 ) [ arxiv:1212.4856 [ hep - ph ] ] .
a. delgado , g. f. giudice , g. isidori , m. pierini and a. strumia , eur .
j. c * 73 * , 2370 ( 2013 ) [ arxiv:1212.6847 [ hep - ph ] ] .
p. agrawal and c. frugiuele , jhep * 1401 * , 115 ( 2014 ) [ arxiv:1304.3068 [ hep - ph ] , arxiv:1304.3068 ] . y. bai , h. c. cheng , j. gallicchio and j. gu , jhep * 1308 * , 085 ( 2013 ) [ arxiv:1304.3148 [ hep - ph ] ] . g. belanger , d. ghosh , r. godbole , m. guchait and d. sengupta , phys .
d * 89 * , 015003 ( 2014 ) [ arxiv:1308.6484 [ hep - ph ] ] .
r. boughezal and m. schulze , phys .
d * 88 * , no .
11 , 114002 ( 2013 ) [ arxiv:1309.2316 [ hep - ph ] ] .
j. a. evans , y. kats , d. shih and m. j. strassler , jhep * 1407 * , 101 ( 2014 ) [ arxiv:1310.5758 [ hep - ph ] ] .
t. eifert and b. nachman , arxiv:1410.7025 [ hep - ph ] . j. eckel , s. su and h. zhang , arxiv:1411.1061 [ hep - ph ] .
g. ferretti , r. franceschini , c. petersson and r. torre , arxiv:1502.01721 [ hep - ph ] .
m. j. herrero , a. mendez and t. g. rizzo , phys .
b * 200 * , 205 ( 1988 ) .
v. d. barger and w. y. keung , phys .
b * 211 * , 355 ( 1988 ) .
h. inazawa and t. morii , phys .
lett . * 70 * , 2992 ( 1993 ) . c. kim and t. mehen , phys .
d * 79 * , 035011 ( 2009 ) [ arxiv:0812.0307 [ hep - ph ] ] .
v. barger , m. ishida and w .- y .
keung , phys .
rev . lett .
* 108 * , 081804 ( 2012 ) [ arxiv:1110.2147 [ hep - ph ] ] .
y. kats and m. j. strassler , jhep * 1211 * , 097 ( 2012 ) [ arxiv:1204.1119 [ hep - ph ] ] . m. drees and m. m. nojiri , phys . rev .
lett . * 72 * , 2324 ( 1994 ) [ hep - ph/9310209 ] .
m. drees and m. m. nojiri , phys .
d * 49 * , 4595 ( 1994 ) [ hep - ph/9312213 ] .
y. kats and m. d. schwartz , jhep * 1004 * , 016 ( 2010 ) [ arxiv:0912.0526 [ hep - ph ] ] . c. kim , a. idilbi , t. mehen and y. w. yoon , phys .
d * 89 * , 075010 ( 2014 ) [ arxiv:1401.1284 [ hep - ph ] ] . s. p.
martin , phys .
d * 77 * , 075002 ( 2008 ) [ arxiv:0801.0237 [ hep - ph ] ] .
s. p. martin and j. e. younkin , phys .
d * 80 * , 035026 ( 2009 ) [ arxiv:0901.4318 [ hep - ph ] ] .
j. e. younkin and s. p. martin , phys .
d * 81 * , 055006 ( 2010 ) [ arxiv:0912.4813 [ hep - ph ] ] .
n. kumar and s. p. martin , phys .
d * 90 * , 055007 ( 2014 ) [ arxiv:1404.0996 [ hep - ph ] ] .
n. craig , m. farina , m. mccullough and m. perelstein , jhep * 1503 * , 146 ( 2015 ) [ arxiv:1411.0676 [ hep - ph ] ] .
j. fan , m. reece and l. t. wang , arxiv:1412.3107 [ hep - ph ] .
k. hagiwara , k. kato , a. d. martin and c. k. ng , nucl .
b * 344 * , 1 ( 1990 ) .
z. han , a. katz , d. krohn and m. reece , jhep * 1208 * , 083 ( 2012 ) [ arxiv:1205.5808 [ hep - ph ] ] .
g. aad _ et al . _ [ atlas collaboration ] , arxiv:1412.4742 [ hep - ex ] .
m. czakon , a. mitov , m. papucci , j. t. ruderman and a. weiler , phys .
* 113 * , no .
20 , 201803 ( 2014 ) [ arxiv:1407.1043 [ hep - ph ] ] .
g. aad _ et al . _ [ atlas collaboration ] , eur .
j. c * 74 * , no . 10 , 3109 ( 2014 ) [ arxiv:1406.5375 [ hep - ex ] ] . c. kilic and b. tweedie , jhep * 1304 * , 110 ( 2013 ) [ arxiv:1211.6106 [ hep - ph ] ] .
b. dutta , w. flanagan , a. gurrola , w. johns , t. kamon , p. sheldon , k. sinha and k. wang _ et al .
_ , phys .
d * 90 * , no . 9 , 095022 ( 2014 ) [ arxiv:1312.1348 [ hep - ph ] ] .
m. r. buckley , t. plehn and m. j. ramsey - musolf , phys .
d * 90 * , no . 1 , 014046 ( 2014 ) [ arxiv:1403.2726 [ hep - ph ] ]
. w. s. cho , j. s. gainer , d. kim , k. t. matchev , f. moortgat , l. pape and m. park , arxiv:1411.0664 [ hep - ph ] .
u. sarid and s. d. thomas , phys .
* 85 * , 1178 ( 2000 ) [ hep - ph/9909349 ] . c. l. chou and m. e. peskin , phys .
d * 61 * , 055004 ( 2000 ) [ hep - ph/9909536 ] .
y. kats and d. shih , jhep * 1108 * , 049 ( 2011 ) [ arxiv:1106.0030 [ hep - ph ] ] .
z. liu and b. tweedie , arxiv:1503.05923 [ hep - ph ] .
g. aad _ et al . _ [ atlas collaboration ] , eur .
j. c * 73 * , no . 1 , 2263 ( 2013 ) [ arxiv:1210.4826 [ hep - ex ] ] . v. khachatryan
_ et al . _
[ cms collaboration ] , arxiv:1412.7706 [ hep - ex ] . c. csaki , y. grossman and b. heidenreich , phys .
d * 85 * , 095009 ( 2012 ) [ arxiv:1111.1239 [ hep - ph ] ] .
e. nikolidakis and c. smith , phys .
d * 77 * , 015021 ( 2008 ) [ arxiv:0710.3129 [ hep - ph ] ] .
v. khachatryan _ et al .
_ [ cms collaboration ] , phys .
d * 91 * , no . 1 , 012007 ( 2015 ) [ arxiv:1411.6530 [ hep - ex ] ] . m. carena , s. gori , n. r. shah and c. e. m. wagner , jhep * 1203 * , 014 ( 2012 ) [ arxiv:1112.3336 [ hep - ph ] ] .
a. arvanitaki and g. villadoro , jhep * 1202 * , 144 ( 2012 ) [ arxiv:1112.4835 [ hep - ph ] ] .
d. carmi , a. falkowski , e. kuflik and t. volansky , jhep * 1207 * , 136 ( 2012 ) [ arxiv:1202.3144 [ hep - ph ] ] .
t. cohen , d. e. morrissey and a. pierce , phys .
d * 86 * , 013009 ( 2012 ) [ arxiv:1203.2924 [ hep - ph ] ] .
d. curtin , p. jaiswal and p. meade , jhep * 1208 * , 005 ( 2012 ) [ arxiv:1203.2932 [ hep - ph ] ] .
p. giardino , k. kannike , m. raidal and a. strumia , jhep * 1206 * , 117 ( 2012 ) [ arxiv:1203.4254 [ hep - ph ] ] . k. blum , r. t. dagnolo and j. fan , jhep * 1301 * , 057 ( 2013 ) [ arxiv:1206.5303 [ hep - ph ] ] .
d. carmi , a. falkowski , e. kuflik , t. volansky and j. zupan , jhep * 1210 * , 196 ( 2012 ) [ arxiv:1207.1718 [ hep - ph ] ] . j. r. espinosa , c. grojean , v. sanz and m. trott , jhep * 1212 * , 077 ( 2012 ) [ arxiv:1207.7355 [ hep - ph ] ] .
r. t. dagnolo , e. kuflik and m. zanetti , jhep * 1303 * , 043 ( 2013 ) [ arxiv:1212.1165 ] .
r. s. gupta , m. montull and f. riva , jhep * 1304 * , 132 ( 2013 ) [ arxiv:1212.5240 [ hep - ph ] ] .
b. batell , c. e. m. wagner and l. t. wang , jhep * 1405 * , 002 ( 2014 ) [ arxiv:1312.2590 [ hep - ph ] ] .
j. fan and m. reece , jhep * 1406 * , 031 ( 2014 ) [ arxiv:1401.7671 [ hep - ph ] ] .
t. gherghetta , b. von harling , a. d. medina and m. a. schmidt , jhep * 1404 * , 180 ( 2014 ) [ arxiv:1401.8291 [ hep - ph ] ] .
g. belanger , b. dumont , u. ellwanger , j. f. gunion and s. kraml , phys .
d * 88 * , 075008 ( 2013 ) [ arxiv:1306.2941 [ hep - ph ] ] .
j. bernon , b. dumont and s. kraml , phys .
d * 90 * , no . 7 , 071301 ( 2014 ) [ arxiv:1409.1588 [ hep - ph ] ] .
s. p. martin , phys .
d * 66 * , 096001 ( 2002 ) [ hep - ph/0206136 ] .
a. djouadi , phys .
rept . * 459 * , 1 ( 2008 ) [ hep - ph/0503173 ] .
r. barbieri and l. maiani , nucl .
b * 224 * , 32 ( 1983 ) . c. s. lim , t. inami and n. sakai , phys .
d * 29 * , 1488 ( 1984 ) .
m. drees and k. hagiwara , phys .
d * 42 * , 1709 ( 1990 ) .
s. heinemeyer , w. hollik and g. weiglein , phys .
rept . * 425 * , 265 ( 2006 ) [ hep - ph/0412214 ] .
v. barger , p. huang , m. ishida and w. y. keung , phys .
b * 718 * , 1024 ( 2013 ) [ arxiv:1206.1777 [ hep - ph ] ] . j. m. frere , d. r. t. jones and s. raby , nucl .
b * 222 * , 11 ( 1983 ) .
m. claudson , l. j. hall and i. hinchliffe , nucl .
b * 228 * , 501 ( 1983 ) .
g. gamberini , g. ridolfi and f. zwirner , nucl .
b * 331 * , 331 ( 1990 ) .
j. a. casas , a. lleyda and c. munoz , nucl .
b * 471 * , 3 ( 1996 ) [ hep - ph/9507294 ] .
t. falk , k. a. olive , l. roszkowski and m. srednicki , phys .
b * 367 * , 183 ( 1996 ) [ hep - ph/9510308 ] .
a. riotto and e. roulet , phys .
b * 377 * , 60 ( 1996 ) [ hep - ph/9512401 ] .
a. kusenko , p. langacker and g. segre , phys .
d * 54 * , 5824 ( 1996 ) [ hep - ph/9602414 ] .
m. carena , m. quiros and c. e. m. wagner , phys .
b * 380 * , 81 ( 1996 ) [ hep - ph/9603420 ] .
j. m. cline , g. d. moore and g. servant , phys .
d * 60 * , 105035 ( 1999 ) [ hep - ph/9902220 ] . c. le mouel , nucl .
b * 607 * , 38 ( 2001 ) [ hep - ph/0101351 ] . c. le mouel , phys .
d * 64 * , 075009 ( 2001 ) [ hep - ph/0103341 ] .
h. h. patel , m. j. ramsey - musolf and m. b. wise , phys .
d * 88 * , no . 1 , 015003 ( 2013 ) [ arxiv:1303.1140 [ hep - ph ] ] .
m. reece , new j. phys . *
15 * , 043003 ( 2013 ) [ arxiv:1208.1765 [ hep - ph ] ] .
j. e. camargo - molina , b. oleary , w. porod and f. staub , jhep * 1312 * , 103 ( 2013 ) [ arxiv:1309.7212 [ hep - ph ] ] .
d. chowdhury , r. m. godbole , k. a. mohan and s. k. vempati , jhep * 1402 * , 110 ( 2014 ) [ arxiv:1310.1932 [ hep - ph ] ] . n. blinov and d. e. morrissey , jhep * 1403 * , 106 ( 2014 ) [ arxiv:1310.4174 [ hep - ph ] ] .
n. chamoun , h. k. dreiner , f. staub and t. stefaniak , jhep * 1408 * , 142 ( 2014 ) [ arxiv:1407.2248 [ hep - ph ] ] .
u. chattopadhyay and a. dey , jhep * 1411 * , 161 ( 2014 )
[ arxiv:1409.0611 [ hep - ph ] ] . j. e. camargo - molina , b. garbrecht , b. oleary , w. porod and f. staub , phys .
b * 737 * , 156 ( 2014 ) [ arxiv:1405.7376 [ hep - ph ] ] . c. l. wainwright , comput .
commun .
* 183 * , 2006 ( 2012 ) [ arxiv:1109.4189 [ hep - ph ] ] .
s. jung and j. d. wells , phys .
d * 89 * , no . 7 , 075004 ( 2014 ) [ arxiv:1312.1802 [ hep - ph ] ] .
cms collaboration [ cms collaboration ] , cms - pas - hig-14 - 006 .
g. aad _ et al . _ [ atlas collaboration ] , phys .
lett . * 113 * , no .
17 , 171801 ( 2014 ) [ arxiv:1407.6583 [ hep - ex ] ] .
cms collaboration [ cms collaboration ] , cms - pas - hig-14 - 031 .
g. aad _ et al . _ [ atlas collaboration ] , phys .
b * 738 * , 428 ( 2014 ) [ arxiv:1407.8150 [ hep - ex ] ] .
s. chatrchyan _ et al .
_ [ cms collaboration ] , phys .
d * 89 * , no . 9 , 092007 ( 2014 ) [ arxiv:1312.5353 [ hep - ex ] ] . [ atlas collaboration ] , atlas - conf-2013 - 013 , atlas - com - conf-2013 - 018 .
cms collaboration [ cms collaboration ] , cms - pas - hig-13 - 027 .
s. chatrchyan _ et al . _
[ cms collaboration ] , jhep * 1401 * , 096 ( 2014 ) [ arxiv:1312.1129 [ hep - ex ] ] .
the atlas collaboration , atlas - conf-2013 - 067 , atlas - com - conf-2013 - 082 .
cms collaboration [ cms collaboration ] , cms - pas - hig-13 - 032 .
g. aad _ et al .
_ [ atlas collaboration ] , phys .
lett . * 114 * , no . 8 , 081802 ( 2015 ) [ arxiv:1406.5053 [ hep - ex ] ] . s. chatrchyan _ et al . _
[ cms collaboration ] , eur .
j. c * 73 * , 2469 ( 2013 ) [ arxiv:1304.0213 [ hep - ex ] ] .
the atlas collaboration , atlas - conf-2014 - 039 , atlas - com - conf-2014 - 057 .
v. khachatryan _ et al . _
[ cms collaboration ] , arxiv:1503.04114 [ hep - ex ] .
the atlas collaboration , atlas - conf-2014 - 005 , atlas - com - conf-2014 - 007 .
g. aad _ et al . _ [ atlas collaboration ] , phys .
d * 89 * , no . 3 , 032002 ( 2014 ) [ arxiv:1312.1956 [ hep - ex ] ] . | we derive new limits on light stops from diboson resonance searches in the @xmath0 , @xmath1 , @xmath2 , @xmath3 and @xmath4 channels from the first run of the lhc .
if the two - body decays of the light stop are mildly suppressed or kinematically forbidden , stoponium bound states will form in @xmath5 collisions and subsequently decay via the pair annihilation of the constituent stops to diboson final states , yielding striking resonance signatures .
remarkably , we find that stoponium searches are highly complementary to direct collider searches and indirect probes of light stops such as higgs coupling measurements . using an empirical quarkonia potential model and including the first two @xmath6-wave stoponium states , we find that in the decoupling limit @xmath7 gev is excluded for _ any _ value of the stop mixing angle and heavy stop mass by the combination of the latest resonance searches and the indirect constraints .
the @xmath8 searches are the most complementary to the indirect constraints , probing the stop `` blind spot '' parameter region in which the @xmath9 trilinear coupling is small .
interestingly , we also find that the @xmath10 searches give a stronger constraint , @xmath11 gev , if the stop is primarily left - handed . for a scenario with a bino lsp and stop nlsp , several gaps in the direct collider searches for stops can unambiguously be filled with the next run of the lhc . for a stop lsp decaying through an r - parity violating @xmath12 coupling
, the stoponium searches can fill the gap 100 gev @xmath13 gev in the direct searches for couplings @xmath14 . |
the process of analysing a time series of astronomical images to measure the variations in brightness of an object over time and to calibrate those variations against reference sources is a very time - consuming task .
an accurate understanding of the variation in brightness of an object over time , i.e. its photometric light curve , allows modelling of key characteristics of the object ( e.g. the shape of an asteroid , the type of a supernova or gamma - ray burst optical afterglow ) .
the measurement process often varies only little depending on the type of object , whether it is an asteroid tumbling through space , a supernova or the rapidly fading optical afterglow from a gamma - ray burst .
the process usually involves examining the images in the time series for at least one suitable reference source in order to calibrate the brightness .
some processes require the reference sources to have particular characteristics , e.g. a specific spectral class or magnitude .
other processes may need sources that have been identified as a photometric standard .
if more than one calibration source is available , the precision can be improved enormously ; however , this is also much more time - expensive .
the process must be repeated for each image and corrections made from one image to another .
these corrections can incorporate predictable adjustments such as extinction due to changing air mass .
in addition , often the corrections that need to be made in order to obtain a consistent light curve are variable , non - uniform and unpredictable , and are based on experience and intuition . the ability to quickly analyse a set of images and to produce a high quality light curve is particularly valuable on automated telescopes , e.g. the fully robotic zadko telescope in gingin , western australia @xcite .
participation in programmes like the prompt follow - up of gamma - ray burst alerts for the detection of transient optical emissions as part of the tlescope action rapide pour les objets transitoires ( tarot ) network @xcite and the gaia follow - up network for solar system objects ( gaia - fun - sso , http://www.imcce.fr/gaia-fun-sso ) @xcite results in the creation of data sets containing transient sources that need rapid analysis .
we have developed a high efficiency image detection & identification ( heidi ) pipeline software for processing images and analysing the variations in brightness of optical transients .
we take the approach that all sources in an image are potential reference sources . the variations in the brightness of the sources in each image of a time series are assessed automatically .
corrections are made to compensate for these variations while also taking into account that some sources may intrinsically vary on short time scales ( e.g. short period variable stars ) . to facilitate production of photometric light curves solar system objects are identified automatically during the image processing . as a result the automated analysis of an image set and the production of a calibrated light curve is achieved in a matter of minutes .
heidi thus fulfils an emerging need for the rapid analysis of time series image data .
heidi is implemented in the programming language python ( http://www.python.org ) for linux platforms .
it processes astronomical images in fits format and requires only a minimal amount of user - supplied detail .
it can also be installed on linux virtual machines .
this makes the software extremely versatile .
preprocessing tasks such as astrometric calibration and image alignment are accomplished using existing applications as described in [ section : fitsalign ] .
the main functions of identification and selection of calibration sources , identification of targets and production of light curves are accomplished using newly developed python code as described in
[ section : correlstar ] and [ section : photometry ] .
heidi has been benchmarked on a pentium class system and a virtual machine , i.e. low - performance , low - cost systems .
the first step in the image analysis process is to perform the astrometric calibration .
the amount of information contained in fits image headers depends heavily on the system in place at the observing site . while an integrated system would typically produce fits images which include calibration information in the image headers
, some sites may not even include the pointing information .
we have therefore adopted a standard approach of uploading one image from the set to the astrometry.net server in order to obtain a new astrometric calibration @xcite . by using the blind astrometric calibration system provided by astrometry.net
, heidi thus removes any reliance on the images having already had astrometric calibrations or having to verify the pointing information , orientation or scale . after obtaining the astrometric calibration
, heidi assumes that all the images in the time series have the same pointing and updates the fits image headers .
the world coordinate system ( wcs ) information from astrometry.net is included in the headers of the updated fits images .
the images are then processed through sextractor in order to build a catalogue of all objects found in the image set .
the sextractor catalogue is subsequently processed through scamp @xcite which produces image headers that are ready to be used by swarp @xcite .
swarp aligns the images and includes the appropriate header information from any translation or rotation applied by swarp during the alignment process .
most of the processing time is taken up with the sextractor - scamp - swarp workflow . on a single - cpu pentium - class system ,
the astrometric calibration and image alignment stage for a set of 50 images takes about 10 min ( vs. 1.25 min for the automated production of the photometric light curve by heidi , [ section : correlstar ] , [ section : photometry ] ) . in the correlation stage
the images from the astrometric calibration and alignment stage are reprocessed through sextractor to produce catalogues of all the sources that have been detected across the entire image set .
these catalogues are correlated by heidi to identify the sources which are common to every image in the time series .
these common sources are the candidate reference sources for the photometric calibration of the image set .
heidi calculates the natural logarithm of the flux value for each candidate source as a representation of relative ( instrumental ) magnitude .
we refer to the natural logarithm of the flux hereafter simply as log flux .
examination of the log flux value for each candidate source may show some variation across the image set .
heidi tests the log flux variation of each candidate source and rejects sources as candidates if the variation is greater than a predetermined threshold .
an examination of the log flux values for the candidate sources in a time series of 50 images for asteroid 939 isberga shows a variation common to all candidate sources ( figure [ fig : figure1 ] ) . as this variation appears to affect every source in a consistent manner ,
we conclude that this variation can be analysed and corrected . the selected stars from observations of asteroid 939 isberga show a variation in the log flux values across the time series .
as this variation appears to affect every source in a consistent manner it can be analysed and a correction applied . ] for a given time series of @xmath0 images , heidi finds all sources that are common to all @xmath0 images in that time series .
these @xmath1 sources are selected as candidates for photometric calibration and are tested for variability over the time series .
step 1 : for the @xmath0 images and @xmath1 sources , heidi reads the flux values @xmath2 , @xmath3 , @xmath4 from the sextractor catalogues and calculates the natural logarithm of the flux .
we will denote the natural logarithm of the flux as @xmath5 , @xmath3 , @xmath4 and refer to it hereafter as log flux .
the initial log flux set is hence @xmath6 , @xmath7 and contains @xmath8 values .
step 2 : for each image @xmath3 in the time series , heidi calculates @xmath1 log flux means : for each source @xmath4 , the source @xmath9 is excluded ( denoted by @xmath10 ) and the log flux mean of the remaining sources @xmath11 is determined : @xmath12 step 3 : heidi determines the global maximum of the log flux : @xmath13 for all @xmath3 , @xmath4 the difference between the global maximum log flux @xmath14 and the log flux means @xmath15 ( determined in step 2 ) are calculated : @xmath16 this will be our correction factor .
step 4 : to smooth out the variations across the image set , the log flux values @xmath5 , @xmath3 , @xmath4 are adjusted , yielding an adjusted log flux set : @xmath17 step 5 : for each source @xmath4 heidi determines the adjusted log flux mean , the maximum adjusted log flux and the minimum adjusted log flux of this source across all @xmath0 images : @xmath18 @xmath19 @xmath20 step 6 : for each source @xmath4 heidi tests whether the deviation of the adjusted log flux values @xmath21 from the adjusted log flux mean @xmath22 is within a predetermined tolerance @xmath23 .
this test rejects sources that are varying in brightness on short time scales ( e.g. short period variable stars ) or are otherwise not varying in accordance with the other candidates in the set ( e.g. edge effects or vignetting ) . only sources that pass this test will remain in the candidate list : @xmath24 sources that do not pass this test will be excluded from the candidate list . if @xmath25 is the number of excluded sources then the reduced candidate list consists of @xmath26 candidates .
this concludes the first pass and reduces the original candidate list @xmath6 to @xmath27 : @xmath28 step 7 : repeat step 2 to step 6 for the reduced candidate list @xmath29 .
this process of filtering variable sources is applied iteratively .
the tolerance @xmath23 for the deviation test of the log flux ( step 6 ) is 0.5 , 0.2 and 0.1 for the first , second and third pass , respectively .
we found that additional iterations with finer tolerances do not significantly improve the result as one reaches the limit of measurement precision .
the candidate stars that survive this filtering process are used in the photometric calibration of the image set for producing the light curve for the target as described in [ section : photometry ] .
in addition , the candidates that fail the test are recorded for further study .
figure [ fig : figure2 ] shows a sample image from a time series of @xmath30 images for asteroid 939 isberga .
this image contains exactly @xmath31 sources that are common to all images in the time series , marked with a cross @xmath32 .
these sources are selected as candidate sources and are tested for their suitability for use as calibration sources .
circles @xmath33 indicate the reduced set of 37 sources that were found to be suitable for use as calibration sources by heidi
. the position of the target , asteroid 939 isberga is indicated by a square ( @xmath34 ) .
sample image from the time series for asteroid 939 isberga .
this image is centred on right ascension @xmath35 , declination @xmath36 ( j2000 ) and has a field of view of @xmath37 arcmin .
sources that are common to all images in the time series are indicated by a cross @xmath32 .
these sources are tested for their suitability to be used as calibration sources .
circles @xmath33 indicate those sources that have been selected as calibration sources .
the position of the target , asteroid 939 isberga is indicated by a square ( @xmath34 ) . ] on a single - cpu pentium - class system the correlation stage for a time series of 50 images takes only about 1 min .
heidi has been designed to facilitate the production of photometric light curves for solar system objects ( target sources ) . following the selection of calibration sources as described in
[ section : correlstar ] , the target sources are identified and selected .
heidi uses the institut de mcanique cleste et de calcul des phmrides ( imcce ) sky body tracker ( skybot ) cone - search and resolver services @xcite to quickly identify the known solar system objects in the time series , our target sources .
heidi processes the time series very quickly : for a set of 50 images it takes heidi only approximately 15 s to identify one target source , to test for its presence in the image set , to compile the list of the corresponding flux values and to produce the photometric light curve for this object , i.e. @xmath38 s per object per image . as a consequence
we have implemented heidi to produce light curves for all of the objects found in the time series rather than just the primary target .
the imcce skybot cone - search service identifies which solar system objects lie within a specified field of view at a given epoch : heidi calls the cone - search service and obtains a list of the objects that may be found in the time series by providing the centre coordinates in right ascension and declination and the epoch of observation . for each object returned by the cone - search service
, heidi calls the imcce skybot resolver service to obtain accurate coordinates : heidi provides the name of the object and the epoch for the first image in the time series to the resolver service .
if known , heidi also provides the international astronomical union ( iau ) observatory code for the site at which the images were recorded .
the resolver service then returns accurate coordinates for the specified object .
if the iau observatory code is provided then the resolver service returns coordinates in the topocentric frame , otherwise the coordinates are given for the geocentric frame . to determine the object coordinates in the remaining images of the time series ,
heidi finds the record for each object in the iau minor planet center orbit ( mpcorb ) database ( http://www.minorplanetcenter.net ) .
the coordinates for each object are computed from the orbital elements in the mpcorb database on a per - image basis across the time series .
these computations are much faster than obtaining coordinates from the resolver service .
it takes only a fraction of a second to calculate the coordinates for an object for all images in the time series , compared with several seconds per image when retrieved from the resolver service .
however , due to rounding errors in the computation , the coordinates provided by the resolver service may be more accurate than the ones computed by heidi .
heidi will thus determine the difference between the coordinates provided by the resolver service and the ones computed locally and apply this offset as a correction to all computed coordinates across all images .
the respective coordinates in each image of the time series are then tested to determine the presence of the target sources .
to minimise false detections without introducing false non - detections , this is done in a two - step process : in a first step , an area of 10 x 10 pixels around the coordinates of the target is tested for the presence of a source ; if no source is found , the area is expanded to 15 x 15 pixels and the test repeated .
it is not possible to determine whether a failure to detect a target source is due to the faintness of the object or an error in the position calculation .
we thus assume that the primary target for the measurement and production of a light curve has a well - known orbit and also a good signal - to - noise ratio in the time series .
if the object is detected in the images of the time series , the log flux of the object is recorded for each image .
plotting the log flux of the object and the log flux of the calibration sources without correcting for the variations between images in the time series does not give a clear indication of the light curve of an object .
figure [ fig : figure3 ] shows the uncorrected log flux measurements from calibration sources and the target in a time series for asteroid 939 isberga .
uncorrected log flux measurements from calibration sources @xmath39 and target @xmath40 in the time series for asteroid 939 isberga . the calibration sources can clearly be seen to vary in accordance with each other .
the target , asteroid 939 isberga , follows a similar trend .
the light curve is not readily apparent . ]
the log flux mean of the calibration sources identified in
[ section : correlstar ] is calculated for each image ( similar to equation [ eqn : mean_log_flux ] but this time over the entire candidate source set ) : @xmath41 the global log flux mean is also determined : @xmath42 this is used for reference and correction of variations across the time series . for each image @xmath3 in the time series , the log flux values of the calibration sources are adjusted by adding the difference between the global log flux mean and the log flux mean for that image . for all @xmath3 , @xmath4 the difference between the global log flux mean @xmath43 and
the log flux mean @xmath44 is calculated : @xmath45 for each image @xmath3 in the time series , the log flux value of the target object @xmath46 , @xmath3 ( where @xmath47 denotes the target source ) is adjusted by the same correction factor @xmath48 . to smooth out the background variations across the image set ,
the log flux values @xmath46 , @xmath3 for the target object @xmath47 are adjusted : @xmath49 applying this correction to the time series for asteroid 939 isberga shows a light curve that is readily apparent ( figure [ fig : figure4 ] ) . this light curve for asteroid 939 isberga
produced by heidi clearly shows a primary period consistent with the expected rotation period of 2.9173 h ( 0.12155 d ) @xcite .
corrected log flux measurements for calibration sources @xmath39 and target @xmath40 in the time series for asteroid 939 isberga .
the variation in the calibration sources is minimal @xmath50 .
the light curve of the target is clearly apparent . ]
the consistency of the results achieved by heidi is further demonstrated by extending the analysis to multiple time series for asteroid 939 isberga that were obtained on three separate dates .
the light curves determined by heidi for these three time series are shown superimposed in figure [ fig : figure5 ] .
it can be easily seen that the three curves follow a very similar pattern .
the light curves produced by heidi for asteroid 939 isberga using time series that were obtained on three separate dates show the consistency of the results .
it can be easily seen that the three curves follow a very similar pattern . a primary periodicity consistent with the expected rotation period of 2.9173 h is evident . ]
these curves can also be overlaid on one another to show that both the period and the amplitude are consistent ( figure [ fig : figure6 ] ) .
light curves produced by heidi for asteroid 939 isberga from time series obtained on three separate dates ( 22 , 23 and 25
november 2011 ) overlaid on one another to show that both the period and the amplitude of each curve is consistent with the period and amplitude of the other curves . ]
the current version ( as of november 2013 ) has good astrometric accuracy but performs only relative photometry , i.e. does not produce calibrated magnitudes . we plan to add the capability for the production of light curves calibrated to a standard photometric system .
the preprocessing tasks of astrometric calibration and image alignment described in [ section : fitsalign ] are independent of the type of object being studied .
the candidate sources that are rejected by the filtering process described in
[ section : correlstar ] warrant further study to determine whether they were excluded due to some intrinsic variability .
these excluded sources can include non - moving sources that have characteristic light curves , e.g. variable stars , supernovae and gamma - ray burst optical afterglows . in the next version of heidi
, we will include the ability to also produce photometric light curves for all the excluded objects as a standard function .
uncertainties on the produced light curves can be inferred from the dispersion of the magnitudes of the calibration sources and will be used to include error bars in the output . if the target source is near , i.e. within 2 to 3 arcsec , a star of similar or greater brightness , it is possible that the star is selected instead of the target source .
the next version of heidi will include a step to automatically exclude detections in which the position information falls outside a predetermined tolerance .
in addition , an algorithm to detect moving objects will be implemented for targets with less accurate position information or unknown targets .
the high efficiency image detection & identification ( heidi ) pipeline software rapidly processes sets of astronomical images .
it is very fast and reliable .
heidi takes a set of images from a night of observing and produces a light curve in a matter of minutes .
heidi is implemented in the programming language python .
it has been tested on various linux systems and linux virtual machines .
it has minimal hardware requirements and the installation process is relatively straightforward .
heidi runs in virtual machines without impacting performance and can thus be installed on non - linux systems ( if a linux virtual machine is available ) .
the high efficiency image detection & identification ( heidi ) pipeline software we developed has a small footprint yet is very powerful and extremely versatile .
given programmes like the prompt follow - up of gamma - ray burst alerts for the detection of transient optical emissions and the gaia follow - up network for solar system objects , which require prompt follow - up and rapid analysis of time series data , heidi will be very useful and provide a service and a speed that have not been available .
testing heidi on multiple fairly large time series for asteroid 939 isberga yielded extremely satisfying results .
the light curves produced using time series obtained on three separate dates showed consistency with the three curves following a very similar pattern .
heidi will be installed at the zadko telescope in gingin , western australia .
this will allow us to analyse very rapidly all the images in the data archive of the zadko telescope and search for interesting variable sources .
it will also allow prompt analysis and rapid issuing of results for time critical events such as gamma - ray burst optical afterglow emissions .
the heidi package is available from the first author on request .
mt thanks frdric vachier for assistance with photometry .
mt thanks andrew williams for assistance with processing fits image files .
the work reported on in this publication has been supported by the european science foundation ( esf ) , in the framework of the great research networking programme .
dmc is supported by an australian research council future fellowship .
berthier , j. , vachier , f. , thuillot , w. , fernique , p. , ochsenbein , f. , genova , f. , lainey , v. , arlot , j .- e . , 2006 , astronomical data analysis software and systems xv , 351 , 367 bertin , e. , arnouts , s. , 1996 , a&as , 117
, 393 bertin , e. , mellier , y. , radovich , m. , missonnier , g. , didelon , p. , morin , b. , 2002 , astronomical data analysis software and systems xi , 281 , 228 bertin , e. , 2006 , astronomical data analysis software and systems xv , 351 , 112 coward , d. m. , et al . , 2010 , pasa , 27 , 331 klotz , a. , bor , m. , eysseric , j. , damerdji , y. , laas - bourez , m. , pollas , c. , vachier , f. , 2008 , pasp , 120 , 1298 lang , d. , hogg , d. w. , mierle , k. , blanton , m. , roweis , s. , 2010 , aj , 139 , 1782 molnar , l. a. , et al . , 2008 , minor planet bulletin , 35 , 9 todd , m. , coward , d. m. , tanga , p. , thuillot , w. , 2013 , pasa , 30 , 14 | the production of photometric light curves from astronomical images is a very time - consuming task , taking several hours or even days .
larger data sets improve the resolution of the light curve , however , the time requirement scales with data volume .
the data analysis is often made more difficult by factors such as a lack of suitable calibration sources and the need to correct for variations in observing conditions from one image to another .
often these variations are unpredictable and corrections are based on experience and intuition .
the high efficiency image detection & identification ( heidi ) pipeline software rapidly processes sets of astronomical images , taking only a few minutes .
heidi automatically selects multiple sources for calibrating the images using a selection algorithm that provides a reliable means of correcting for variations between images in a time series .
the algorithm takes into account that some sources may intrinsically vary on short time scales and excludes these from being used as calibration sources .
heidi processes a set of images from an entire night of observation , analyses the variations in brightness of the target objects and produces a light curve all in a matter of minutes .
heidi has been tested on three different time series of 50 images each of asteroid 939 isberga and has produced consistent high quality photometric light curves in a fraction of the usual processing time .
the software can also be used for other transient sources , e.g. gamma - ray burst optical afterglows , gaia transient candidates .
heidi is implemented in the programming language python and processes time series astronomical images in fits format with minimal user interaction .
heidi processes up to 1000 images per run in the standard configuration .
this limit can be easily increased , with the only real limit being system capacity , e.g. disk space , memory .
heidi is not telescope - dependent and will process images even in the case that no telescope specifications are provided .
heidi has been tested on various linux systems and linux virtual machines .
heidi is very portable and extremely versatile with minimal hardware requirements .
[ firstpage ] methods : analytical methods : data analysis methods : numerical techniques : image processing techniques : photometric astrometry |
[ [ section ] ] ir blocking filters are critical for optimizing bolometer - based receivers in the millimeter and submillimeter spectral region . in these bands ,
the ir power emitted from the telescope and surroundings ( typically 250 - 300 k for ground and balloon - based instruments ) is much brighter than the sky background .
it is therefore crucial to control this out - of - band power .
filters serve to substantially reduce what would be a dominant source of noise and minimize the radiative loading on the cryogenic system .
in addition , the filters must not significantly radiate , reflect , or scatter in the band of interest .
this requires high thermal conductivity and good heat sinking for any filter containing absorptive components .
[ [ section-1 ] ] meeting all these requirements in one system is difficult . a variety of approaches have been used previously .
reflective frequency selective surfaces ( e.g. , patterned onto thin plastic substrates and stacked , as pioneered by ulrich @xcite , and summarized by ade @xcite ) , bulk filters of absorptive materials @xcite , and scattering filters @xcite are commonly employed . these filtering approaches ,
however , are not without their complications .
the reflective filters patterned on plastic substrates are subject to the intrinsic limits of multi - layered reflectors@xcite as well as heating due to absorption of the plastic . in practice ,
additional reflective layers improve the filter rejection only incrementally and absorption leads to reradiated power that falls onto detectors@xcite .
absorptive filters , such as those made of bulk alumina , present difficulties with antireflection coatings , requiring the use of lossy materials that reduce the overall transmission in the desired pass - band by 5% or more @xcite .
additionally , alumina has reststrahlen bands that open up upon cooling , diminishing its overall filtering performance @xcite . in this work we present an example of a hybrid approach based on a combination of reflective frequency selective structures patterned on silicon substrates , scattering and absorptive layers based on composites of powdered crystals exhibiting the reststrahlen effect , and metamaterial antireflection coatings to control the in - band reflections .
we present a particular example of the construction and performance of a blocking filter designed to pass the 70 - 170 ghz band in section [ sec : filterconstruction ] , and discuss its performance in section [ sec : filterperformance ] .
we conclude with a discussion of the scalability and applicability of this design in section [ sec : applications ] .
figure [ filter - schematic ] shows the anatomy of the composite absorptive / reflective ir - blocking filter .
this filter consists of lithographically defined frequency selective surfaces patterned on two silicon wafers , a 25 @xmath0 layer of an absorptive mixture of epoxy and reststrahlen powders placed between the two patterned surfaces , and a metamaterial antireflection coating on both vacuum silicon interfaces . at ir wavelengths
, light is scattered and reflected off the front silicon wafer and frequency selective surface .
the front metamaterial surface scatters light both specularly and diffusely for frequencies above the single - moded limit of the structure . for the metamaterial surfaces
described here ( tuned to pass 1 - 2 mm wavelengths ) , this frequency falls well below the infrared emission from a 300 k blackbody , which peaks at 10 @xmath0 ( 30 thz ) .
infrared light that passes the frequency selective metal mesh is subject to both scattering and absorption by the reststrahlen - epoxy composite .
an additional metal mesh layer reflects most of the remaining infrared light back into the epoxy - powder layer , boosting absorption and ( to a lesser extent ) reflection .
this approach reduces the load on the cryogenic stage by reflecting a significant portion of the ir power , and uses an absorbing layer to further attenuate ir power passing the first reflective layer . at millimeter and submillimeter wavelengths the metal mesh frequency selective surfaces and metamaterial silicon have a high transmission , and the extinction length of the epoxy mixture is inconsequential , leading to low absorption .
thus in the bands of interest , this structure behaves nearly as if it were a slab of solid low loss silicon treated with a high quality antireflection coating @xcite . in the remainder of this section
we describe the design and performance of the frequency selective surfaces , powder - epoxy mixes , metamaterial antireflection coatings , and conclude with predictions for the integrated performance of our filters .
the first filtering component of these composite ir - blocking filters is a low - pass frequency selective surface formed by a mesh of resonant metal squares .
these squares act as a grid of capacitive elements and pass low frequencies while reflecting high frequencies . in the low frequency limit ,
the metallization layer is effectively nonexistent giving nearly unity transmission , and in the high frequency limit , these features reflect according to the fill factor of the metallization . in the resonant region between , there is some additional reflection , with the cutoff frequency set by the grid spacing .
we selected the grid parameters to place the cutoff frequency well above the upper end of our desired signal band ( 170 ghz ) , but below the peak emission frequency of a 300 k blackbody ( 18 thz ) .
we selected a grid period of 23.8 @xmath0 and street widths of 5 @xmath0 , for a cutoff frequency in silicon of @xmath13.6 thz ( corresponding to a freespace wavenumber of 120 @xmath2 @xcite , clearly visible as the first resonance in the measurement in figure [ mesh - picture ] ) .
these dimensions were additionally constrained to be within the capabilities of large - diameter liftoff lithographic techniques ( limiting us to minimum features of few micron scales ) .
a prototype of this design was thoroughly characterized ( see figure [ mesh - picture ] ) , and the performance is in good agreement with our theoretical expectations , with a total reflectivity of 83% for a 300 k blackbody .
these features are straightforward to tune to a desired cutoff frequency and can be fabricated on the silicon using standard lithographic techniques .
design of these features was carried out via the analytical techniques described by ulrich @xcite , with additional optimization and verification via modeling in ansys hfss @xcite .
[ [ section-2 ] ] the reststrahlen effect , from the german for residual rays , " is the absorption of light at characteristic frequencies @xcite of the bound ion pairs in crystalline materials that falls in the infrared @xcite .
these structural resonances prevent light in the reststrahlen band from propagating through the material , and effective low - pass filters have been previously demonstrated using powdered reststrahlen materials @xcite .
transmissive and reflective filters have also been realized from bulk reststrahlen crystals @xcite .
[ [ section-3 ] ] powder filters , formed by reststrahlen powders mixed into a polyethylene carrier , were demonstrated as ir blocking filters by yamada et al .
@xcite these filters , due to their plastic carrier , suffered from heating and reradiation .
we fabricated similar powder filters by mixing reststrahlen powders into a mixture of toluene thinned epotek 301 optical epoxy and applying thin layers with a commercial spraygun mounted on a robotic gantry .
this allowed for the creation of thin ( 25 @xmath0 ) uniform layers of the epoxy - powder composite .
the powders were chosen from the assortment of materials characterized by yamada @xcite to provide good blocking coverage for wavelengths from < 10 @xmath0 up to 150 @xmath0 ( corresponding to 67 - 1000 + @xmath2 , 2 - 30 + thz ) , overlapping transmissive regions of one powder with absorptive regions of another .
additionally , the powders were chosen for ease of use and acquisition , limiting us to non - hazardous materials ( excluding materials such as thallium and beryllium salts , which have been previously used @xcite ) that are common chemical reagents . as a result of these constraints ,
our filters consist of magnesium oxide ( @xmath3 ) and calcium carbonate ( @xmath4 ) , both with 5 - 20 @xmath0 typical particle sizes .
see table [ table:1 ] for the composition of the reststrahlen composite layer , obtained from the maxwell - garnett theory @xcite .
the measured optical transmissions for these powder filters are shown in figure [ reststrahlen ] .
these transmission spectra exhibit the characteristic absorption features expected for both materials @xcite .
we combined these powders in equal parts ( by mass ) to toluene thinned optical epoxy and applied it with the spraygun . for the full composite filter , this epoxy layer also adhered the two silicon wafers together .
the particle size is sufficiently small that the epoxy layer can be treated as a dielectric mixture that is well described by mean field effective medium approximations @xcite in the instrument band of 70 - 170 ghz ( see figure [ lf ] ) .
.the composition of the epoxy composite layer . [ cols="^,^,^,^",options="header " , ] [ [ section-4 ] ] the performance was characterized at cryogenic temperatures to ensure their proper cryogenic functioning .
it is known that some reststrahlen materials have absorption bands that open up when the material is cooled down .
in particular , alumina ( @xmath5 ) is known to have a section of its absorption band ( between 30 and 300 @xmath0 ) open up at temperatures of tens of kelvin @xcite . to explore whether this would be problematic with our materials ,
a powder filter consisting of our mixture of @xmath4 and @xmath3 was measured in a fourier transform spectrometer ( fts ) at a range of temperatures between 4 k and 300 k. plots of the extremal temperatures are shown in the right panel of figure [ reststrahlen ] . from this , it is apparent that changes in the absorption spectrum are minimal , and do not compromise the filtering performance of a 300 k blackbody significantly . high resistivity silicon is an excellent material for millimeter and submillimeter wave optics . the low loss ( tan @xmath6 < 7e-5 ) @xcite and high refractive index allows for refractive optics with negligible loss .
high purity , single - crystal silicon is available in large diameters and can be readily obtained .
additionally , silicon has a high thermal conductivity ( @xmath7)@xcite , which prevents filters with an absorptive component from heating up .
the high refractive index of silicon presents the problem of high reflectivity for optics , but this problem has been successfully managed with a machined sub - wavelength anti - reflective surface on the outer surfaces of the optics .
these features allow for the creation of antireflection coatings `` with simulated dielectric layers .
these metamaterial antirreflection surfaces allow for high transmission optics ( > 99% transmission across a 70 - 170 ghz band for a three - layer simulated dielectric coating @xcite ) fabricated entirely from silicon . a thorough discussion of this antireflection surface treatment approach is given in datta et al @xcite .
figure [ ar - picture ] shows a photograph of a three - layer ' ' coating " as well as a comparison of the simulated and measured reflection at 15@xmath8 incidence for two orthogonal polarizations .
the spectral performance of our composite filters and their constituent components was evaluated using fts measurements .
the thermal performance and integrated measurements were made in a cryostat open to a 300 k blackbody , using a carbon loaded disk bolometer .
the performance of the individual components was measured across both the low - frequency signal band ( down to 10 @xmath2 , 300 ghz ) limited by the low frequency capability of the fts ) and high - frequency blocking band ( up to 5000 @xmath2 , 150 thz ) . additionally , the overall performance of a full composite was measured across the blocking band . [
[ measurement - methods ] ] measurement methods : + + + + + + + + + + + + + + + + + + + + the reflectance and transmittance in the range 10 - 5,000 @xmath2 ( 0.3 - 150 thz ) were measured with the bruker ifs 125 fts using the following two combinations of source - beamsplitter - detector : hg arc lamp multilayer mylar liquid helium cooled bolometer ( 30 - 700 @xmath2 , 0.9 - 21 thz ) , and globar ge - coated kbr
dlatgs ( 500 - 5,000 @xmath2 , 150 - 1500 thz ) .
the spectral response in the overlap region agreed to within 0.5% .
the data sets were merged into one spectra by equating the areas underneath the curves in the overlap region using a weighted average .
the reflectance was measured in a collimated beam geometry with an @xmath9 angle of incidence .
the transmittance was measured in a focused beam ( f/6.5 ) geometry at normal incidence .
the total hemispherical transmittance in the range 500 - 5,000 @xmath2 was collected using a bruker - made integrating sphere ( 75 mm diameter ) accessory for the fts with its own internal dlatgs detector .
a diffusely reflecting gold surface , which matches the inner surface of the sphere , was placed over the sample port to collect the reference scan . for integrated testing
, a cryostat was used to measure the radiative properties and total infrared blocking .
[ [ ir - blocking - performance ] ] ir blocking performance : + + + + + + + + + + + + + + + + + + + + + + + + the infrared blocking performance of these filters was measured on the fts up to 5000 @xmath2 ( 2 @xmath0 wavelength , 150 thz ) , giving a full characterization of the transmission across the spectrum of a 300 k blackbody .
these measurements reproduce the characteristic reststrahlen powder filter shape in thicker ( 75 @xmath0 ) layers @xcite ( see figure [ reststrahlen ] ) , and demonstrate excellent ir blocking in layers as thin as 25 @xmath0 . in the fts measurements , the composite filter
specularly reflected > 40% of the light incident from a 300 k blackbody ( indicating reflection off the front silicon surface and metal mesh features ) , and diffusely reflected another 10% , indicative of backscattering off the powder layer .
the overall reflectance was lower than would be naively expected from the single reflector measurements in section [ sec : filterconstruction ] .
this is likely explained by coupling of fields from the reflectors into the lossy epoxy composite below .
should higher reflectivity be needed , the reflective layers could be separated from the lossy composite and multiple reflective layers could be employed .
[ [ low - frequency - performance ] ] low frequency performance : + + + + + + + + + + + + + + + + + + + + + + + + + + the low frequency performance ( below 1 thz ) of a 75 @xmath0 layer of the powder filter component was measured down to 10 @xmath2 ( 300 ghz ) using an fts .
these data were then fit with a simple transmission line model .
the index of the powder layer was estimated via a maxwell - garnett effective medium approximation @xcite , and treated as a layer of continuous material of an unknown thickness .
the thickness of this layer and the underlying silicon layer were then fit using a standard least - squares approach . for a 75 @xmath0 mix layer ( consisting of a binder of epotek 301 , @xmath10
as measured in the 1 - 10 thz range with a sample etalon formed by silicon wafers , loaded with powdered @xmath3 , @xmath11 , and @xmath4 , @xmath12 in 0.073 and 0.092 volumetric fractions respectively ) on 500 @xmath13 thick silicon ( @xmath14 ) , the model accurately reproduced the measured total thickness .
this model was then used to extrapolate to lower frequencies and simulate the effect of adding a three - layer antireflective coating to a filter made using this material .
this model shows that the reststrahlen powder filter introduces minimal loss ( dominated by the epoxy carrier ) in a signal band from 70 - 170 ghz , and that an instrument - band transmission of > 99% should be achievable for a filter using this technology ( with the total transmission limited by the antireflection coating performance ) . in the low frequency region ,
the filter performance is well represented with a simple transmission line model , taking into account the effective index of the composite determined from an effective medium approximation .
the measured and modeled low frequency performance is shown in figure [ lf ] .
[ [ thermal - performance - and - cryostat - testing ] ] thermal performance and cryostat testing : + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + an integrated test of the composite filter performance was carried out in a 3-stage cryostat , to measure the total blocking efficiency of a 15 cm diameter prototype ( figure [ lf ] ) . a 5.8 cm diameter disk bolometer ( made from carbon - blackened copper )
was held at a base temperature of 5 k. the filter was mounted on the 20 k stage directly above the bolometer , and blocked light being emitted from a blackbody at 300 k visible through a hole in the lid of the 100 k stage . the total power blocked by the filter
was then estimated by measuring the bolometer temperature with and without the filter .
additionally , the heating of the 20 k stage was measured to estimate the power absorbed by the filter , and the filter temperature was measured at its center and edge to characterize the thermal gradient across it .
figure [ lf ] shows a schematic of the cryostat .
there was negligible heating of the center of the filter ( with thermal gradients of less than 1 k from center to edge ) when the filter was mounted to the 20 k stage and used to block the power from a 7 cm diameter window open to 300 k. the filter as a whole heated to 2 k above the stage temperature , demonstrating effective heat - sinking and power removal from the filter as well . in this configuration ,
the power deposited on a carbon disk bolometer at 5 k was measured , and this measurement established a lower limit on the blocking of > 98% of a 300 k blackbody .
this lower limit is in agreement with the fts measurements of the full composite filter . [
[ comparison - to - competing - filter - technologies ] ] comparison to competing filter technologies : + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + compared to existing reflective filters , built of reflective frequency selective surfaces patterned onto plastic substrates and stacked , our filters offer several distinct advantages .
the high thermal conductivity of silicon and low absorption means that there is negligible heating of the filter center relative to the heatsunk edge . additionally , silicon is a stiffer , more mechanically robust substrate , which allows for better control of lithographic features and finished filters that are stronger and less susceptible to deformation and wrinkling due to thermal cycling .
the composite construction of our filters additionally integrates absorptive components to improve the overall out - of - band rejection beyond what can be attained with a reasonable number of stacked reflective components .
compared to bulk absorptive filters such as alumina , our filters offer the ability to reflectively reject a significant portion of the power , reducing the load on the cryogenic system .
the reststrahlen materials we used are also easy to obtain .
the use of multiple reststrahlen materials allows for selection of complementary sets of materials that improves the frequency coverage and prevents sections of the band from becoming transmissive ( such as happens with alumina when cooled ) .
silicon is a lower loss material than alumina and the antireflection coatings , being metamaterial silicon , are significantly lower loss than the epoxies used to antireflection coat bulk alumina filters , and the thermal conductivity of silicon is significantly higher than alumina or ptfe . finally , the composite filter offers a number of adjustable parameters not present in bulk filters that allow better tailoring of the filter characteristics to achieve the desired performance .
the components of this style of silicon substrate composite filters all have applicability across a broad range of frequencies .
the limiting frequencies for the different components vary , but filters can be constructed using these techniques and some subset of the components for frequencies ranging from tens of gigahertz to hundreds of terahertz depending on the components used .
the frequency scaling of each of the four important components is discussed in this section .
the combination of several of these components can form effective filters across a wide range of frequencies .
the frequency selective metal mesh features can be fabricated across the full range of sizes available with modern lithographic techniques .
this results in applicability across the full range over which silicon is transparent , up to the beginning of the absorption band at 1 @xmath0 wavelength ( 300 thz ) .
silicon offers better control and repeatability of lithographic features than plastic substrates , allowing smaller features ( and therefore higher frequencies ) to be attained .
additionally , the metal features can be formed with both high- and low - pass frequency responses , as well as with anisotropy , which enable a range of filtering characteristics , including band - defining filters , and filters with polarization dependence .
the high - resistivity silicon substrate offers excellent transmission performance from low frequencies up to the beginning of its absorption band in the near - infrared .
it begins to become absorptive at a wavelength of approximately 1 @xmath0 , which corresponds to the plasma frequency ( 300 thz ) in the media@xcite . at frequencies lower than this ( longer wavelengths ) ,
it remains highly transmissive and low loss , with typical loss tangent below 0.0002 across the thz region @xcite , and remaining low through the ir ( with the exception of two lattice absorption features at 600 and 650 @xmath2 , corresponding to 18 and 19 thz or wavelengths of 15 and 16 @xmath0)@xcite .
the metamaterial antireflective coatings have applicability to a similarly broad range of frequencies .
at lower frequencies , the coatings can be machined into the surfaces using a dicing saw or conventional grinding techniques .
this approach to the antireflection coatings has been successfully demonstrated for frequencies up to 1.5 thz with excellent performance .
with increasing frequency , higher precision fabrication approaches become necessary . using laser machining to cut smaller features
is one option , up to the limits of current laser machining capabilities .
additionally , lithographic techniques involving the patterning of the surface and etching of features , such as deep reactive ion etching ( drie ) offer a solution to still higher frequencies , with attainable feature sizes of tens of nanometers that would be suitable for antireflection coatings well above the 1 @xmath0 absorption cutoff of silicon . for higher frequency and non - cryogenic applications , conventional antireflection coating approaches
can also be used , substituting layers of bulk dielectric materials ( or applied thin films of dielectric materials ) for the simulated dielectric formed by the metamaterial silicon .
the scattering and absorptive powder layer is useful only for lower frequencies due to the fixed reststrahlen bands and limits on attainable powder size .
the particulate size of the powders can be selected to move the scattering peak to higher or lower frequencies as needed , but the absorption bands of the materials are fixed by the material choice .
for some applications , suitable materials are available that will enhance the overall light rejection of the filter . for filters where blocking in these bands
is not desired , the scattering and absorptive powder can be removed from the design . the epoxy binder can likewise be removed in favor of direct bonding the wafers , for applications where the epoxy would increase the loss and is not needed as a carrier for scattering and absorptive powders .
in addition to forming effective free - space ir blocking filters , this filtering approach offers several novel possibilities for silicon - substrate optical elements .
lower frequency - selective metal elements can be incorporated into these filters to aid in defining the instrument signal band .
anisotropic application of these filtering techniques can form birefringent materials .
these filters can be easily and inexpensively integrated into other optical components , such as silicon lenses .
filters with higher cutoff frequencies and better uniformity can also be constructed for effectively blocking higher frequencies ( into the mid - ir and higher ) due to the high quality of silicon substrates available ( superior material properties and surface finish allow for finer lithography , leading to better high frequency performance , a current limit of plastic - substrate filters @xcite ) .
the frequency range can be tuned by adjusting the reflective grid parameters , the scattering particle size , and the specific material used for scatterers , and multi - layer reflective structures can be formed to increase the overall reflectivity .
direct wafer bonding can reduce or eliminate the need for lossy epoxy components , and reflective layers can be precisely spaced using standard lithographic techniques to set the layer thicknesses .
collectively , these techniques will further improve control of the filter properties , enabling higher performance and better customization of the filter characteristics to the desired application .
finally , more complex metamaterial behaviors can be added through more complicated lithographic features and machined subwavelength features .
this enables a wide array of novel optical characteristics with potentially broad consequences for future imaging systems .
this work was supported by a nasa office of the chief technologist s space technology research fellowship # nnx12am32h .
lithography was performed using the lurie nanofabrication facility at the university of michigan .
jjm was supported by de - sc0015799 for this work .
y. inoue , t. matsumura , m. hazumi , a. t. lee , t. okamura , a. suzuki , t. tomaru , and h. yamaguchi , `` cryogenic infrared filter made of alumina for use at millimeter wavelength , '' appl .
opt . * 53 * , 17271733 ( 2014 ) .
r. datta , c. d. munson , m. d. niemack , j. j. mcmahon , j. britton , e. j. wollack , j. beall , m. j. devlin , j. fowler , p. gallardo , j. hubmayr , k. irwin , l. newburgh , j. p. nibarger , l. page , m. a. quijada , b. l. schmitt , s. t. staggs , r. thornton , and l. zhang , `` large - aperture wide - bandwidth antireflection - coated silicon lenses for millimeter wavelengths , '' appl . opt . *
52 * , 87478758 ( 2013 ) .
a. hadni , j. claudel , x. gerbaux , g. morlot , and j .-
munier , `` sur le comportement diffrent des cristaux et des verres dans labsorption de linfrarouge lointain ( 401500@xmath15 ) la temprature de lhlium liquide , '' appl
. opt . * 4 * , 487494 ( 1965 ) .
r. datta , j. austermann , j. beall , d. becker , k. coughlin , s. duff , p. gallardo , e. grace , m. hasselfield , s. henderson , g. hilton , s. ho , j. hubmayr , b. koopman , j. lanen , d. li , j. mcmahon , c. munson , f. nati , m. niemack , l. page , c. pappas , m. salatino , b. schmitt , a. schillaci , s. simon , s. staggs , j. stevens , e. vavagiakis , j. ward , and e. wollack , `` design and deployment of a multichroic polarimeter array on the atacama cosmology telescope , '' journal of low temperature physics ( 2015 ) .
p. h. bolivar _
_ , `` measurement of the dielectric constant and loss tangent of high dielectric - constant materials at terahertz frequencies , '' ieee transactions on microwave theory and techniques ( 2003 ) .
m. n. afsar and x. li , `` millimeter wave complex refractive index , complex dielectric permittivity and loss tangent of high purity and compensated silicon , '' international journal of infrared and millimeter waves ( 1994 ) . | infrared ( ir ) blocking filters are crucial for controlling the radiative loading on cryogenic systems and for optimizing the sensitivity of bolometric detectors in the far - ir .
we present a new ir filter approach based on a combination of patterned frequency selective structures on silicon and a thin ( 50 @xmath0 thick ) absorptive composite based on powdered reststrahlen absorbing materials . for a 300 k blackbody ,
this combination reflects @xmath150% of the incoming light and blocks > 99.8% of the total power with negligible thermal gradients and excellent low frequency transmission .
this allows for a reduction in the ir thermal loading to negligible levels in a single cold filter .
these composite filters are fabricated on silicon substrates which provide excellent thermal transport laterally through the filter and ensure that the entire area of the absorptive filter stays near the bath temperature .
a metamaterial antireflection coating cut into these substrates reduces in - band reflections to below 1% , and the in - band absorption of the powder mix is below 1% for signal bands below 750 ghz .
this type of filter can be directly incorporated into silicon refractive optical elements . |
in a vector space we enjoy the ability to combine vectors by addition and to scale by elements of a field .
but we can not , in general , multiply vectors .
if we upgrade " our space to permit multiplication ( subject to a few restrictions ) , we obtain an algebra . generalizing from vector spaces to free modules over a domain , we make the following definition .
let @xmath0 be a domain ( with @xmath2 ) .
@xmath0-_algebra _ is an associative ring @xmath3 ( with @xmath2 ) equipped with an embedding of @xmath0 into the center of @xmath3 that is free as an @xmath0-module .
rank _ of a free @xmath0-algebra is its rank as an @xmath0-module .
we will assume that all algebras have finite rank .
we identify @xmath0 with its image @xmath4 .
algebras allow us to multiply tuples .
for example , we can multiply pairs of real numbers in the quadratic ( rank 2 ) @xmath1-algebra @xmath5 , and we can multiply quadruples in the quaternion algebras of sir william rowan hamilton . in this thesis , we contribute to the classification of finite - rank algebras , up to isomorphism .
algebras are classified first by rank .
much work has been done for rings of low rank @xcite .
there are different results for different ranks and choice of @xmath0 . the results for free quadratic ( rank 2 ) algebras and finite - rank division algebras over @xmath1 ( see @xcite and sections 2 , 3 ) suggest that possession of a standard involution is a useful feature with which to compare algebras . [
def : involution]an _ involution _
@xmath6 is an @xmath0-linear map satisfying @xmath7 for all @xmath8 , @xmath9 , and @xmath10 for all @xmath11 .
an involution is _ standard _ if it additionally satisfies @xmath12 for all @xmath8 . in section 4
we include some notes on standard involutions and the related topic of exceptional rings . in section 5
our first result ( theorem [ thm : associative ] ) is a universal multiplication table for a cubic algebra over a domain @xmath0 .
our main result , theorem [ thm : main ] in section 5 , is that free cubic algebras ( algebras of rank 3 ) are either commutative or possess a standard involution .
the unique algebra in the intersection of these cases we call the _ nilproduct _ ring , where the product of non - scalar basis elements is zero . in section 6
, we generalize the classification theorem of frobenius from finite - rank division algebras over @xmath1 to finite - rank division algebras over a field @xmath13 of characteristic not equal to 2 which possess a standard involution .
we conclude with some discussion of future work .
in this section , we present some results in the classification of ( associative ) free quadratic algebras over a commutative ring @xmath0 .
first suppose @xmath0 is a domain with @xmath14 .
we show that the isomorphism classes of quadratic @xmath0-algebras are in bijection with the set of discriminant classes @xmath15 .
the results of this section are standard . for a reference , see dummit and foote @xcite or voight @xcite .
let @xmath0 be a commutative ring ( with 1 ) and with @xmath14 , and let @xmath3 be a free @xmath0-algebra @xmath3 of rank 2 : we say that @xmath3 is _ free quadratic r - algebra_. [ lem : basis1 ] there exists @xmath8 such that @xmath16 is a basis for @xmath3 .
let @xmath17 be a basis for @xmath3 .
then @xmath18 , for some @xmath19 .
let @xmath20 be the ideal generated by @xmath21 .
consider the ideal of @xmath3 @xmath22 observe that @xmath23 , so @xmath24 .
we claim that @xmath25 .
an element @xmath26 is a sum of elements @xmath27 .
thus @xmath28 .
next for @xmath29 consider @xmath30 . with respect to the basis
@xmath17 we may write @xmath31 .
so @xmath32 we have that @xmath33 .
projecting in the first coordinate , we see that @xmath34 .
so for @xmath35 there exists @xmath36 such that @xmath37
. then @xmath38 .
since @xmath39 and @xmath0 is a domain , we have @xmath40 , so @xmath41 .
hence @xmath42 . thus there
exist @xmath43 such that @xmath44 .
so the change of basis @xmath45 mapping @xmath46 is invertible since @xmath47 = at - sb=1 $ ] , and therefore @xmath48 is a basis for @xmath3 .
let @xmath16 be a basis for @xmath3 as in lemma [ lem : basis1 ] .
consider multiplication in @xmath3 : for @xmath49 , @xmath50 with @xmath51 , we have by distributivity and associativity in @xmath3 , @xmath52 ( notice that multiplication is commutative in @xmath3 , since @xmath3 is generated by @xmath53 as an @xmath0-algebra . )
so to define multiplication in @xmath3 , it is necessary and sufficient to specify @xmath54 such that @xmath55 .
consequently , we may write @xmath56}{(x^2 - tx + n)}$ ] . thus the set of quadratic algebras equipped with a basis is in bijection with the set of pairs @xmath57 .
now we consider the isomorphism classes .
given @xmath58}{(x^2 - tx + n)}$ ] , we define the _ discriminant _ of @xmath3 by disc@xmath59 [ lem : completesquare ] if @xmath14 , then @xmath60}{(x^2 - tx + n)}\cong \frac{r[y]}{(y^2 - d ) } = r[\sqrt{d}]$ ] , where @xmath61disc(@xmath62 .
consider the ring homomorphism @xmath63 , which is bijective since @xmath14 . to show that relations are preserved compute @xmath64 by lemma [ lem : completesquare ] , up to isomorphism , we may take @xmath65 .
we have completed the square . "
[ lem : equaldisc ] if @xmath66 as free quadratic @xmath0-algebras , then disc@xmath67 disc@xmath68 ( as classes in @xmath69 ) .
express @xmath70}{(x^2 - tx + n)}$ ] with basis @xmath16 and @xmath71}{(y^2 - ty + n)}$ ] with basis @xmath72 , so that disc@xmath73 and disc@xmath74 . by hypothesis
there exists an isomorphism @xmath75 and @xmath76 , @xmath77 such that @xmath78 .
compute @xmath79 since @xmath80 is free , collecting coefficients of @xmath72 and using the fact that @xmath76 , we get @xmath81 and @xmath82 .
therefore @xmath83 , which shows that disc@xmath84 disc@xmath68 .
[ thm : congquadratics ] let @xmath14 and @xmath58}{(x^2 - d)}$ ] and @xmath85}{(y^2 - d)}$ ] be quadratic algebras .
then @xmath66 as algebras if and only if @xmath86 .
suppose @xmath86 .
then @xmath87 for some @xmath76 .
consider the @xmath0-linear map @xmath88\to b$ ] given by @xmath89 and @xmath90 .
let @xmath91 and @xmath92 be elements of @xmath3 . then @xmath93 together with linearity this shows that @xmath94 is an algebra homomorphism .
we compute the kernel of @xmath94 : @xmath95 which shows @xmath96 .
next suppose @xmath97 for @xmath98 .
then @xmath99 so @xmath100 .
since @xmath101 is a basis element , @xmath102 if and only if @xmath103 , so @xmath104 . hence @xmath105 . since @xmath76
, @xmath94 is surjective .
therefore by the first isomorphism theorem , @xmath66 . together with lemma [ lem : equaldisc ] this completes the proof .
[ cor : quadraticsbij ] if @xmath14 , then there is a bijection @xmath106&\longmapsto \rm{disc}(a)\\ \left[\left.\raisebox{.2em}{$r[x]$}\middle/\raisebox{-.2em}{$(x^2-d)$}\right.\right ] & \longmapsfrom dr^{\times 2}\end{aligned}\ ] ] by theorem [ thm : congquadratics ] , the isomorphism classes of quadratic @xmath0-algebras correspond to discriminants in @xmath69 .
[ ex : disc0 ] consider the class of discriminant 0 .
this class is represented by the algebra with @xmath107 , namely @xmath108}{(x^2)}$ ] , and @xmath53 is nilpotent with @xmath109 .
[ ex : disc1 ] a representative of the class of discriminant 1 is the algebra with @xmath110 , so that @xmath111 , and @xmath108}{(x^2-x)}$ ] . in this case , @xmath53 is idempotent . since @xmath112 factors as @xmath53 and @xmath113 , by the chinese remainder theorem , we have @xmath114}{(x-1)}\times \frac{r[x]}{(x)}\cong r\times r$ ] . at ( -2.5 , 3.6 )
@xmath115}{(x^2 - x)}$ ] ; at ( -2.5 , 2.8 ) @xmath116 ; at ( -2.5 , 2.2 ) @xmath117 ; at ( 3 , 3.6 ) @xmath115}{(x-1)}\times \dfrac{r[x]}{(x)}$ ] ; at ( 4.5 , 2.8 ) @xmath118 ; at ( 0 , 0 ) @xmath119 ; at ( .2 , 3.8 ) @xmath120 ; at ( .2 , 3.4 ) ; ( .2 , .8 ) ( 1.8,2.4 ) ; ( -1 , 3.6 ) ( 1.4 , 3.6 ) ; ( -.2 , .8 )
( -1.8 , 1.6 ) ; the induced multiplication is componentwise .
multiplication in this class is given by @xmath121 .
this class of algebras also has representative @xmath115}{(x^2 - 1)}$ ] .
[ ex : risz ] if we remove the requirement that @xmath14 , we sometimes achieve the same classification .
the following was known to gauss @xcite .
for example , take @xmath122 .
let @xmath123}{(x^2 - tx + n)}$ ] , which has discriminant @xmath124 .
suppose @xmath125}{(y^2 - ty + n)}$ ] is isomorphic to @xmath3 via @xmath94 given by @xmath126 , @xmath127 for some @xmath128 .
then @xmath129 thus @xmath130 and @xmath131 .
computing the discriminant of @xmath80 we get @xmath132 .
letting @xmath133 leads to identical results .
since @xmath134 , this covers all possible invertible maps .
suppose conversely that @xmath3 and @xmath80 have equal discriminants @xmath135 ( since @xmath136 , these are single - element classes ) .
then @xmath137 , from which we see that @xmath138 and @xmath139 are even and that @xmath140 then consider the isomorphism @xmath141 given by @xmath126 , @xmath142 .
we check that relations are satisfied : @xmath143y + \frac{1}{4}(t^2 - 2tt+t^2 ) - \frac{1}{2}t^2 + \frac{1}{2}tt+n\\ & = y^2 - ty + \frac{t^2}{4 } - \frac{tt}{2}+ \frac{t^2}{4 } - \frac{t^2}{2}+\frac{tt}{2}+n\\ & = y^2 - ty + \frac{t^2 - t^2}{4 } + n\\ & = y^2 - ty + n,\end{aligned}\ ] ] which shows that @xmath94 induces a surjective homomorphism @xmath144 . to check that @xmath94 is injective let @xmath145 satisfy @xmath146 .
then @xmath147 which shows that @xmath148 .
thus @xmath94 is an isomorphism @xmath149 . if we relax the condition that @xmath14 , then the classification becomes much more complicated .
for example , when @xmath13 is a field of characteristic 2 , then there exist nonisomorphic quadratic @xmath13-algebras with equal discriminants .
consider @xmath150}{(x^2+x+1)}$ ] and @xmath151}{(y^2 + y)}$ ] .
note that disc@xmath67 disc@xmath152 .
since @xmath153 is irreducible , @xmath3 is a field .
algebra @xmath80 has zero divisors @xmath101 and @xmath154 , so @xmath80 is not a field . thus @xmath155 .
therefore we can not in general classify quadratic algebras by discriminant when @xmath156 .
however , when @xmath156 we can classify separable quadratic algebras @xmath3 using the artin - schrier group in @xmath0 .
[ def : separable]let @xmath157 $ ] be a polynomial of degree @xmath158 .
let @xmath159 , where @xmath160 are the roots of @xmath161 .
we say @xmath161 is _ separable _ if @xmath160 are distinct in @xmath162 .
define the @xmath13-algebra @xmath163}{(f(x))}$ ] to be _ separable _ when @xmath161 is separable .
[ lem : char2separable ] for a field @xmath13 of characteristic @xmath164 , polynomial @xmath165 $ ] is separable if and only if @xmath166 , and @xmath167}{(f(x))}\cong \dfrac{f[y]}{(y^2 + y + n)}$ ] for some @xmath168 .
let @xmath169 be the factorization of @xmath161 over @xmath170 .
by multiplying @xmath171 , we see that @xmath172 . thus @xmath166 , if and only if @xmath173 are distinct .
thus a separable quadratic polynomial over @xmath13 is of the form @xmath174 where @xmath175 . for separable @xmath163}{(x^2 - tx + n)}$
] , we have @xmath176 , so @xmath177
. then @xmath178}{(y^2 - y + t^{-2}n)}$ ] with isomorphism @xmath179 .
therefore we may take any separable quadratic @xmath13-algebra to be of the form @xmath167}{(x^2 + x + n)}$ ] .
let @xmath180 be the _ artin - schreier group _ of @xmath13 , with operation addition . to confirm that this is a group , check that for @xmath181 , we have @xmath182 : @xmath183 define the class of @xmath184 in @xmath13 to be @xmath185 .
[ prop : artinbij ] for any field @xmath13 with char@xmath186 , there exists a bijection of sets @xmath187}{(x^2 + x + a ) } \end{aligned}\ ] ] for the class @xmath188 , consider the map @xmath189}{(x^2+x+a)}$ ] .
first we show that the map is well - defined .
suppose @xmath190 .
then @xmath191 for some @xmath192 .
since @xmath167}{(x^2 + x + a)}\cong \dfrac{f[x]}{(x^2 + x + a + r+ r^2)}$ ] by the isomorphism @xmath193 , the map is well - defined . by lemma
[ lem : char2separable ] , the map is surjective .
next we show that the map is injective .
suppose @xmath194 , and that @xmath195}{(x^2 + x + a)}$ ] , and @xmath196}{(y^2+y+b)}$ ] with @xmath66 .
then there exists an isomorphism @xmath141 with @xmath126 and @xmath197 for some @xmath198 with @xmath199 .
then @xmath200 collecting like terms we see @xmath201 . since @xmath202 , @xmath203 .
then @xmath204 , which contradicts @xmath205 .
therefore the map is injective , and so the sets are in bijection . for @xmath58}{(x^2 - tx + n)}$ ] , consider the @xmath0-linear map @xmath206 induced by @xmath207 .
observe @xmath208 , that @xmath209 , and @xmath7 ( also that @xmath210 ) . thus this map is a standard involution ( definition [ def : involution ] ) .
therefore all quadratic algebras possess a standard involution .
it is known that when @xmath3 is quadratic this involution is unique @xcite .
a _ division ring _ is a ring in which each nonzero element is a unit . in 1877
, ferdinand frobenius @xcite categorized the finite - dimensional division algebras over the real numbers as consisting of @xmath211 , and the hamiltonians @xmath212 . in this section
, we give a modern elementary formulation of the proof .
in later sections , we generalize from the real field to a domain and from a division algebra to one possessing a standard involution ( so this argument serves as a partial model for our later investigations ) .
[ def : hamiltonians ] the _ hamiltonians _
@xmath212 are the @xmath1-algebra with basis @xmath213 where multiplication is defined @xmath214 and @xmath215 . [ rem : hamiltonianinvolution ] for @xmath216 , the map @xmath217 is a standard involution ( definition [ def : involution ] ) . define the _ trace _ of @xmath53 to be tr@xmath218
. we will show ( corollary [ cor : divisioninvolution ] ) that @xmath219 .
define the _ norm _ of @xmath53 to be nr@xmath220 .
observe that @xmath53 satisfies @xmath221@xmath222@xmath223 $ ] .
thus every @xmath224 satisfies a quadratic polynomial in @xmath225 $ ] .
[ lem : irreducibledegree ] a real polynomial @xmath226 $ ] that is irreducible over @xmath1 has degree @xmath2 or @xmath164 . by the fundamental theorem of algebra , the polynomial @xmath227 factors over @xmath5 as @xmath228 , for some @xmath229 , where @xmath230 are in @xmath5 . since @xmath227 is irreducible over @xmath1 , either @xmath227 is linear or all @xmath231 and thus occur in conjugate pairs @xmath232 .
so @xmath227 may be written @xmath233 for some @xmath234 .
since @xmath235 and @xmath236 are real , if the degree of @xmath227 is greater than 2 , we have a contradiction .
thus the degree of @xmath227 is less than or equal to 2 .
[ lem : divisionminimal ] if @xmath237 is a finite - dimensional division ring containing @xmath1 and @xmath238 , then the minimal polynomial of @xmath239 is irreducible over @xmath1 .
let @xmath226 $ ] be the minimal polynomial of @xmath239 over @xmath1 .
suppose @xmath240 with @xmath241 $ ] , both of nonzero degree ( less than @xmath158 ) ; then @xmath242
. then since @xmath237 is a division ring , @xmath243 or @xmath244 and so @xmath239 satisfies a real polynomial of degree less than @xmath158 , a contradiction .
therefore , @xmath227 is irreducible .
[ cor : fieldrorc ] a finite extension @xmath162 over @xmath1 is either @xmath1 or @xmath5 .
assume @xmath245 and let @xmath246 .
let @xmath227 be the minimal polynomial of @xmath239 over @xmath1 .
by lemma [ lem : divisionminimal ] we have that @xmath227 is irreducible , and since @xmath227 is of degree at least 2 , by lemma [ lem : irreducibledegree ] , @xmath227 is quadratic . since @xmath227 factors completely over @xmath5
, we have @xmath247 .
hence @xmath248 .
then since @xmath249\ge 2 $ ] , we have @xmath250 .
[ thm : frobenius ] the finite - dimensional division @xmath1-algebras are @xmath1 , @xmath5 , and @xmath212 .
let @xmath237 be a finite - dimensional @xmath1-algebra .
we have @xmath251 if and only if @xmath252
. we will show that if @xmath253 then @xmath254 , and then that if @xmath255 , then @xmath256 .
suppose @xmath253 .
then there exists @xmath257 .
since @xmath237 is finite dimensional , it can not be the case that @xmath258 are @xmath1-independent for all @xmath259 .
thus there exists @xmath260 such that @xmath261 are linearly dependent .
so @xmath262 satisfies a real polynomial .
let @xmath227 be the monic polynomial of smallest degree @xmath158 satisfied by @xmath262 .
then the @xmath1-subalgebra of @xmath237 generated by @xmath262 is @xmath263 = \ { a_0 + a_1 j + \cdots + a_{n-1 } j^{n-1 } : a_k\in \mathbb{r } \}\cong \frac{\mathbb{r}[x]}{(p(x))}.\ ] ] since @xmath1 is commutative and @xmath262 commutes with itself , @xmath264 $ ] is commutative .
[ lem : rjc ] @xmath264\cong \mathbb{c}$ ] . since @xmath237 is a division ring , @xmath227 is irreducible by lemma [ lem : divisionminimal ]
. then the ideal @xmath265 is maximal . therefore @xmath266
\cong \frac{\mathbb{r}[x]}{(p(x))}$ ] is a field .
since @xmath264 $ ] is a field containing @xmath1 and @xmath267 , by corollary [ cor : fieldrorc ] we have @xmath264\cong \mathbb{c}$ ] .
so if @xmath268 then @xmath254 , and @xmath269 if and only if @xmath270 .
next suppose that @xmath271 .
then there exists @xmath272 where @xmath273 is @xmath1-independent of @xmath2 and @xmath262 . by the argument in lemma [ lem : rjc ]
, we have @xmath274 \cong \mathbb{c}$ ] .
[ lem : isquared ] if dim@xmath275 then there exist @xmath276 such that @xmath277 are independent and @xmath278 . since dim@xmath275 there exist @xmath279 @xmath1-independent elements of @xmath237 .
since the minimal polynomial of @xmath280 is quadratic , @xmath281 for some @xmath282 .
replace @xmath280 with @xmath283 .
then @xmath284 are independent and @xmath285 .
note that @xmath286 because the minimal polynomial @xmath287 of @xmath288 is irreducible .
finally , replace @xmath288 with @xmath289 with @xmath290 chosen so that @xmath278 .
consider the map given by conjugation by @xmath273 in @xmath237 : @xmath291 a check of the axioms shows that @xmath94 is an @xmath1-algebra isomorphism .
observe that @xmath292 i d .
since @xmath293 in end@xmath294 , @xmath94 satisfies the polynomial @xmath295 , which has roots @xmath296 .
thus the minimal polynomial of @xmath94 idivides @xmath297 .
so the roots of the minimal polynomial of @xmath94 are distinct , and then @xmath94 is diagonalizable .
[ prop : deigenspaces ] let @xmath298 be the eigenspace of eigenvalue @xmath299 and @xmath300 be the eigenspace of eigenvalue @xmath2 with respect to the above map @xmath94 , as vector spaces over @xmath1
. then @xmath301 .
since @xmath94 is diagonalizable , @xmath237 has a basis of eigenvectors of @xmath94 .
consider the bijection @xmath302 the existence of @xmath303 shows that any basis of @xmath300 , union a basis of @xmath298 is a basis of eigenvectors of @xmath237 .
so @xmath301 .
note that since @xmath304 , we have @xmath274\cong \mathbb{c}\subset d^+$ ] .
[ prop : dplusc ] @xmath305 = \mathbb{c}$ ] .
let @xmath306 .
since @xmath306 , @xmath307 .
so @xmath308 .
thus the @xmath1-algebra generated by @xmath274 $ ] and @xmath53 , @xmath274[x]\cong \mathbb{c}[x]$ ] is commutative .
we check that @xmath300 is closed under multiplicative inverse : @xmath309 so @xmath310 , which shows that @xmath311 . since @xmath312 $ ] is a commutative division algebra , it is a field .
then by corollary [ cor : fieldrorc ] , @xmath313 , so @xmath314 .
we have shown that @xmath315 if and only if @xmath269 .
so now suppose that @xmath316 ; then there exists @xmath317 .
then @xmath318 and so @xmath319 .
note that @xmath320 . by the same reasoning as in proposition [ prop : dplusc ] , @xmath321 .
[ prop : dminusc ] @xmath322 $ ] .
observe that for @xmath323 , @xmath324 : @xmath325 let @xmath326 .
note that @xmath327 . since @xmath328 , we have @xmath329 .
conversely , if @xmath329 , then @xmath326 since @xmath330 and @xmath331 .
therefore , @xmath332 . therefore if dim@xmath333 , then @xmath334 . as in lemma
[ lem : isquared ] we may take @xmath335 .
so we have @xmath237 generated by @xmath336 where @xmath337 and @xmath319 .
this describes @xmath212 .
algebras of rank 1 , quadratic rings , and finite - dimensional division algebras containing @xmath1 each possess a standard involution .
consequently , each element in such an algebra satisfies a polynomial of degree @xmath338 with coefficients in that algebra s ring of scalars . in our classification , it is natural to consider which algebras possess such a map . in this section
we discuss standard involutions , give a few important examples , and review some of their properties . as in 1 ,
let @xmath0 be a domain and let @xmath80 be an @xmath0-algebra .
[ def : involution ] an _ involution _
@xmath339 is a self - inverse ( i.e. @xmath7 for all @xmath8 ) @xmath0-linear map satisfying @xmath10 for all @xmath340 .
an involution is _ standard _ if it additionally satisfies @xmath12 for all @xmath341 .
[ lem : xplusxbar ] if @xmath342 is a standard involution on @xmath80 , then @xmath343 for all @xmath341 . for @xmath341
, we have @xmath344 , so @xmath345 ; since @xmath346 , subtracting @xmath347 from each side we get @xmath343 .
let @xmath80 have standard involution @xmath342 .
for @xmath348 let @xmath349 be the element of @xmath0 such that @xmath350 .
then @xmath351 , which shows that @xmath342 is determined by the functional @xmath352 , @xmath353 .
observe that @xmath354 is the sum of @xmath342 and the identity , each of which is @xmath0-linear , and so @xmath354 is @xmath0-linear .
thus @xmath354 is an @xmath0-linear map .
[ rem : invpoly ] if @xmath3 has a standard involution then every @xmath8 is a root of a polynomial @xmath355 in @xmath356 $ ] .
[ ex : quaternioninvolution ] a quaternion algebra over a field possesses a standard involution .
let @xmath13 be a field and @xmath357 be a quaternion @xmath13-algebra with basis @xmath213 satisfying @xmath358 , @xmath359 , and @xmath215 .
then consider the @xmath13-linear map @xmath360 given by @xmath361 .
clearly , @xmath342 is self - inverse and fixes @xmath13 . to check that @xmath342 is an involution
we show that @xmath362 for all @xmath363 .
let @xmath364 and @xmath365 .
then @xmath366 next we compute @xmath367 : @xmath368 so @xmath362 , which shows that @xmath342 is an involution .
next we check that @xmath342 is standard : @xmath369 [ cor : divisioninvolution ] finite - dimentional division algebras containing @xmath1 possess a standard involution .
on @xmath1 , the identity is a standard involution .
the standard involution on @xmath5 is complex conjugation .
the hamiltonians @xmath212 are a quaternion algebra and have the standard involution described in example [ ex : quaternioninvolution ] .
[ ex : adjoint ] if @xmath13 is a field then the adjoint map
@xmath370\mapsto [ \begin{smallmatrix } d & -b\\-c & a \end{smallmatrix}]$ ] is a standard involution on @xmath371 .
consider the map that sends matrix @xmath372 to its adjoint : @xmath373 to check that @xmath342 is an involution , we first verify that @xmath342 is self - inverse : @xmath374 next we show that @xmath342 is @xmath13-linear .
clearly , @xmath342 fixes @xmath375 $ ] . to verify that @xmath342 is additive we compute : @xmath376 then we show that @xmath342 respects scaling by @xmath377 : @xmath378 for @xmath342 to be an involution it remains to verify that @xmath379 for all @xmath380 : @xmath381 therefore @xmath342 is an involution . to determine whether @xmath342 is standard we check that @xmath382 for all @xmath383 : @xmath384 [ ex : fcrossf ] on @xmath385 , the map @xmath386 is a standard involution .
this map is clearly @xmath13-linear and self - inverse . since ( component - wise )
multiplication is commutative , @xmath387 for all @xmath388 . to see that this involution is standard we compute @xmath389 an important class of rings possessing standard involutions are exceptional rings @xcite .
[ def : exceptional ] an _ exceptional ring _ is an @xmath0-algebra @xmath80 with the property that there is a left ideal @xmath390 with @xmath391 such that the left multiplication map @xmath392 end@xmath393 factors through a linear map @xmath394 , i.e. @xmath395 for @xmath396 .
[ prop : exceptionalinvolution ] the map @xmath397 for @xmath398 extends to a standard involution on @xmath80 , ( and so any exceptional ring possesses a standard involution ) .
let @xmath399 be the @xmath0-linear extension of the map @xmath400 , fixing 1 . to see that @xmath342 is self - inverse compute @xmath401 let @xmath19 and @xmath402 . to check that @xmath342 is an involution we compute @xmath403 next we find the product @xmath404 thus @xmath342 is an involution .
to check that @xmath342 is standard , compute @xmath405 in @xcite , voight shows that for algebras of rank 3 , possession of a standard involution is equivalent to being an exceptional ring .
an @xmath0-algebra @xmath80 of rank 3 has a standard involution if and only if it is an exceptional ring @xcite .
we give an elementary proof of this theorem in the following section .
let @xmath0 be a domain . in this section we classify @xmath0-algebras that are free of rank 3 as an @xmath0-module . first we produce a universal multiplication table for such algebras ( equation [ eqn : mtable ] ) .
then our main result is as follows .
[ thm : main ] a free @xmath0-algebra of rank @xmath406 is either commutative or possesses a standard involution .
we further show the only commutative algebra which also possesses a standard involution is the _ nilproduct algebra _ , defined to be the algebra in which the product of non - scalar basis elements is zero : @xmath407}{(x^2 , y^2 , xy , yx ) } = \dfrac{r[x , y]}{(x , y)^2}$ ] . in @xcite
, voight shows that @xmath0 is a direct summand of @xmath3 , i.e. , @xmath3 has a basis containing 1 .
let @xmath277 be a basis of @xmath3 .
then multiplication in @xmath3 is defined as follows , for some @xmath408 @xmath409 this multiplication table can be simplified by adjusting the basis . as in the paper by gross and lucianovic @xcite , we define a _ good basis _ to be a basis of the form @xmath277 where @xmath410 .
this adjustment is analogous to completing the square " in the quadratic case .
[ prop : goodbasis ] every @xmath0-algebra of rank @xmath406 has a good basis .
following @xcite , ( which assumes that the algebra is commutative , though it is not necessary to do so ) we apply the change of basis @xmath411 .
then @xmath412 which is in the span of 1 .
thus @xmath3 can be rewritten with the multiplication table : @xmath413with @xmath414 .
the requirement that @xmath3 be associative restricts the possible values of the coefficients in the above multiplication table . for example , by associativity , we have the identity @xmath415 . applying the above multiplication table to the identity
we obtain : @xmath416 subtracting @xmath417 from each side and applying the relation in table [ goodbasis ] we get @xmath418 and so @xmath419 .
this allows the simplifications @xmath420 and @xmath421 .
the relations revealed by associativity are as follows : 1 .
from @xmath415 we have @xmath422 ; 2 .
from @xmath423 we have @xmath424 ; 3 .
from @xmath415 we have @xmath425 ; 4 .
from @xmath426 we have @xmath427 ; 5 .
from @xmath428 we have @xmath429 ; 6 .
from @xmath430 we have @xmath431 ; 7 .
from @xmath432 we have @xmath433 .
the relations in @xmath434 allow the substitutions @xmath435 and @xmath436 .
from @xmath437 we may substitute @xmath438 . from @xmath439 ( and @xmath440 )
we see that we may replace @xmath441 with @xmath442 .
thus we have : @xmath443with @xmath444 satisfying the ( incomplete ) list of relations @xmath445 , @xmath446 , @xmath447 , and @xmath448 . applying these relations to the identity @xmath428
we see : @xmath449 and so @xmath450 must also satisfy @xmath451 .
we show two more relations that must be satisfied by the coefficients in the multiplication table .
[ lem : my ] we have @xmath452 .
suppose @xmath453 , so @xmath454 . since @xmath0 is a domain , the relations @xmath455 imply @xmath456 . then @xmath457 , so @xmath458 and @xmath459 .
so the system reduces to @xmath460 next we apply the reduced multiplication table [ reduced ] to the product @xmath461 : @xmath462 subtracting @xmath463 from each side we get @xmath464 since @xmath453 , this is a contradiction .
[ lem : cn ] we have @xmath465 .
the relations @xmath455 require @xmath465 except in the case that @xmath466 .
in that case , the relations @xmath467 and @xmath468 require that @xmath469 or @xmath470 . in the case
@xmath158 is not zero the system reduces to @xmath471 a computation on the identity @xmath472 similar to that in lemma 5.3 shows that @xmath473 , from which we conclude that @xmath474 .
therefore , @xmath3 has multiplication defined by @xmath475[eqn : relations ] where @xmath476 satisfy @xmath477 that is @xmath478 , @xmath479 , @xmath468 and @xmath480 .
[ thm : associative ] the multiplication table [ eqn : mtable ] with relations [ eqn : relations]@xmath481 determines an associative algebra .
the quotient of the free @xmath0-algebra on @xmath2 , @xmath273 , and @xmath262 and the ideal generated by the relations in [ eqn : mtable ] and [ eqn : relations]@xmath481 is an @xmath0-algebra of rank @xmath482 let @xmath372 be the rank 3 @xmath0-module with basis @xmath277 written @xmath483 .
define multiplication @xmath484 as in the table .
let @xmath485 , @xmath486 , and @xmath487 be elements of @xmath372 . by comparing @xmath488 and @xmath489
we verify ( in appendix a ) that multiplication is associative , thus proving the first assertion .
let @xmath490 , be the free associative @xmath0-algebra in @xmath491 .
let @xmath3 be the quotient of @xmath13 by the relations in the multiplication table [ eqn : mtable ] with coefficients @xmath476 that satisfy [ eqn : relations]@xmath481 : @xmath492 let @xmath493 stand for the ideal generated by the relations in the table , so that @xmath494 .
the relations in the table show that every element of @xmath3 may be expressed as an @xmath0-linear combination of @xmath495 , and @xmath262 , so @xmath3 is generated by @xmath277 as an @xmath0-module , and the rank of @xmath3 is less than or equal to 3 .
next we construct a cubic algebra @xmath80 .
consider the following three elements of @xmath496 : @xmath497 to obtain these matrices assume ( for a moment ) the independence of @xmath277 . in that case
we may identify @xmath498 with the column vector @xmath499 .
then the columns of @xmath500 and @xmath501 record the action of left - multiplication by @xmath273 and @xmath262 respectively on @xmath277 , as given in the multiplication table .
however , the existence of these matrices does not depend on properties of @xmath491 .
now recall that we have yet to show that that @xmath502 and @xmath262 are @xmath0-independent .
let @xmath80 be the sub @xmath0-algebra of @xmath496 generated by @xmath503 . by comparing the first columns of @xmath504 , and @xmath501 we see that @xmath504 , and @xmath501 are @xmath0-independent , so @xmath80 is free of rank 3 or greater as an @xmath0-algebra .
consider the surjective @xmath0-linear map from @xmath505 given by @xmath126 , @xmath506 , and @xmath507 . by the universal property of free algebras
, this map extends to a homomorphism @xmath508 .
we will show that @xmath94 factors as follows : at ( -2 , 2 ) @xmath13 ; at ( 2 , 2 ) @xmath80 ; at ( 0 , .25 ) @xmath3 ; at ( 0 , 2.5 ) @xmath94 ; at ( 1.25 , .75 ) @xmath509 ; at ( -1.25 , .75 ) @xmath510 ; ( -1.5 , 2 ) ( 1.5 , 2 ) ; ( -1.5 , 1.5 ) ( -.5 , .5 ) ; ( .5 , .5 ) ( 1.5 , 1.5 ) ; where @xmath511 is the quotient map sending @xmath53 to the coset @xmath512 , and @xmath513 is the map @xmath514 .
it is necessary and sufficient to show that the composition @xmath515 is independent of representative of the coset @xmath516 ; that is if @xmath517 , then @xmath518 , which is true if and only if @xmath519 . to check that @xmath519 , we verify the following identities by direct computation : @xmath520 first we compute the product @xmath521 : @xmath522 we compare @xmath521 to the corresponding expression in the table : @xmath523 applying the relations @xmath524 and simplifying we observe that indeed @xmath525 next we verify that @xmath526 .
we compute the product @xmath527 simplifying and applying the relations @xmath455 , we obtain @xmath528 thus @xmath526 .
similar computations demonstrate that @xmath94 respects the relations @xmath529 and @xmath530
. therefore , @xmath519 .
since @xmath94 is surjective and @xmath94 factors as @xmath515 , we have that @xmath531 is a surjective algebra homomorphism @xmath149 .
. then @xmath533 . by the independence of @xmath503
, we have @xmath534 .
thus @xmath535 , which shows that @xmath509 is injective , and hence bijective .
then @xmath509 is an algebra isomorphism @xmath149 , and @xmath3 is a cubic algebra .
[ lem : cases ] the relations ( [ eqn : relations].10 ) on the coefficients of the multiplication table of @xmath3 ( [ eqn : mtable ] ) require , in particular , one of the two following cases : ( c ) : : @xmath536 , in which @xmath3 is commutative , or ( e ) : : @xmath537 are not both @xmath538 , and @xmath539 , in which @xmath3 is noncommutative . to satisfy @xmath465 ,
first let @xmath540 .
then by @xmath541 , we have cases @xmath542 or @xmath539 .
next satisfy @xmath465 by taking @xmath469 .
so by @xmath543 we have either @xmath536 or @xmath544 . in case ( c )
we have the following multiplication table for @xmath3 @xmath545 where @xmath546 are arbitrary elements of @xmath0 .
observe that @xmath547 , and so @xmath3 is commutative .
these algebras are examined in detail in gross and lucianovic @xcite . for completeness
we include this result from their paper .
[ def : binaryform ] a _ binary cubic form _ is an expression @xmath548 $ ] such that @xmath549 is homogeneous of degree @xmath406 in @xmath164 variables : @xmath550 the _ discriminant _ @xmath551 of @xmath552 is given by @xmath553 the set @xmath372 of binary cubic forms over @xmath0 is a free @xmath0-module of rank 4 , denoted sym@xmath554 .
we define the ( twisted ) action of @xmath555 on @xmath372 by @xmath556 this action is faithful , and @xmath557 .
there is a discriminant - preserving bijection between the set of orbits of the twisted action of gl(@xmath0 ) on the @xmath0-module sym@xmath554 and the set of isomorphism classes of commutative cubic rings @xmath3 over @xmath0 .
[ rem : grossletters ] in their paper @xcite gross and lucianovic use different characters in their multiplication table for commutative cubic algebras .
the notation correspondence from this paper to theirs is @xmath558 , @xmath559 , @xmath560 , and @xmath561 .
next consider case ( e ) , in which @xmath539 and @xmath537 are not both zero . if @xmath562 , from @xmath563 we obtain @xmath564 , and from @xmath480 we see that @xmath565 .
taking @xmath566 produces identical results . thus in case ( e ) the multiplication table for @xmath3 is @xmath567 where @xmath537 are arbitrary elements of @xmath0 ( with @xmath537 not both 0 ) . comparing @xmath568 and @xmath569
we see that @xmath3 is not commutative in this case .
[ prop : exceptional ] in case ( e ) , algebra @xmath3 is an exceptional ring .
let @xmath570 .
note that @xmath571 and @xmath572 .
let @xmath372 be the left ideal @xmath573 .
we check that @xmath574 : @xmath575 then @xmath576 by independence of @xmath503 we have @xmath577 .
consider left - multiplication in @xmath372 : @xmath578 thus left - multiplication in @xmath372 factors through the @xmath0-linear map @xmath579 , @xmath580 , which shows that @xmath3 is an exceptional ring .
it follows from proposition [ prop : exceptionalinvolution ] that in case ( e ) , @xmath3 possesses a standard involution .
[ prop : exceptionalstandard ] in case ( e ) , the algebra @xmath3 possesses the following standard involution : @xmath581 extended linearly .
first we check that @xmath342 is an involution .
observe that @xmath342 fixes @xmath0 .
since @xmath582 ( similarly , @xmath583 ) , @xmath342 is self - inverse . since @xmath342 is @xmath0-linear , to check that @xmath379 for all @xmath11 , it suffices to check that @xmath584 , @xmath585 , @xmath586 and @xmath587 .
first @xmath588 by symmetry , @xmath585 .
next we check @xmath589 and also @xmath590 thus @xmath342 is an involution . to check that @xmath342 is standard , take @xmath591 .
then @xmath592 in proposition [ prop : exceptionalstandard ] we have shown that a noncommutative cubic algebra possesses a standard involution . [ lem : nilproduct ] suppose @xmath3 possesses a standard involution and is commutative .
then @xmath3 has multiplication table @xmath593 ( @xmath3 is the nilproduct ring ) . since @xmath3 is commutative , we have @xmath594
. then @xmath3 has multiplication table : @xmath595 since @xmath3 has a standard involution , @xmath273 and @xmath262 satisfy monic quadratic equations in @xmath356 $ ] ; i.e. , there exist @xmath596 such that @xmath597 and @xmath598 . equating coefficients with those in the above multiplication table we see that @xmath599 , @xmath600 , @xmath601 , and @xmath602 , so that @xmath603 and @xmath604
. since @xmath605 , there exist @xmath606 such that @xmath607 and @xmath608 .
then since @xmath609 we see that @xmath610 and @xmath611 .
next recall that @xmath586 , so that @xmath612 subtracting @xmath613 from each side we obtain @xmath614 .
then @xmath615 , which shows that @xmath3 is the nilproduct ring .
thus we have shown that a cubic algebra @xmath3 is either commutative or possesses a standard involution , and that the unique commutative algebra with a standard involution is the nilproduct ring .
thus we have the following picture of the parameter space of cubic algebras ; the commutative algebras ( c ) have four degrees of freedom ( multiplication table [ eqn : commutative ] ) , the algebras which are exceptional rings ( e ) have two degrees of freedom ( multiplication table [ eqn : exceptional ] ) , and the intersection of the two is the unique nilproduct ring @xmath616 .
lr at ( -1 , 2.5 ) ( c ) ; at ( 1 , 2.5 ) ( e ) ; at ( 0 , 1 ) @xmath617 ; at ( .3 , 1.5 ) @xmath616 ; at ( 3 , 1.5 ) @xmath618 ; at ( 3.4 , 0 ) @xmath158 ; at ( -4.2 , 1.3 ) @xmath619 ; at ( -3.5 , 2 ) @xmath620 ; at ( -2.8 , .5 ) @xmath101 ; at ( -3.2 , -.1 ) @xmath621 ; ( 0 , 1 ) ( 3 , 0 ) ; ( 0 , 1 ) ( 2.5 , 1.5 ) ; ( 0 , 1 ) ( -3 , -.2 ) ; ( 0 , 1 )
( -2.5 , .5 ) ; ( 0 , 1 ) ( -3.2 , 2 ) ; ( 0 , 1 )
( -4 , 1.2 ) ; ( 0,-.5 ) ( 0,3 ) ; & at ( 0 , 1 ) @xmath617 ; at ( -1.5 , 1.5 ) ( c ) ; at ( 2.25 , .3 ) ( e ) ; at ( .4 , 1.3 ) @xmath616 ; ( 0 , 0 )
( 0 , 2 ) ; ( -2,2 ) ( 0,2 ) ; ( 0 , 2 )
( -1 , 1 ) ; ( -3 , 1 )
( -2 , 2 ) ; ( -1 , 1 ) ( -3 , 1 ) ; ( 0,0 ) ( -1,-1 ) ; ( -1,-1 ) ( -3,-1 ) ; ( -1,1 ) ( -1,-1 ) ; ( -3 , -1 ) ( -3 , 1 ) ; ( 0 , 1 ) ( 2.5 , -.5 ) ; ( 0 , 1 )
( 3.5 , .8 ) ; ( 2.5 , -.5 ) ( 3.5 , .8 ) ; [ thm : cubicexceptional ] a cubic algebra is an exceptional ring if and only if it possesses a standard involution . in proposition
[ prop : exceptionalinvolution ] we showed that exceptional rings possess a standard involution .
next suppose that @xmath3 possesses a standard involution . if @xmath3 is commutative ( by lemma [ lem : nilproduct ] ) , @xmath3 is the nilproduct ring , which is exceptional .
if @xmath3 is noncommutative , then @xmath3 is in case ( e ) of lemma [ lem : cases ] .
algebras in case ( e ) were shown to be exceptional rings in proposition [ prop : exceptional ] . [ ex : charpoly ] specializing to case ( e ) , we compute the characteristic polynomial of an arbitrary element @xmath498 . since @xmath622 satisfies a quadratic equation , we expect to see a repeated factor . from multiplication table
[ eqn : exceptional ] we have the following left - regular representation of @xmath623 @xmath624 and so @xmath622 corresponds to the lower - triagular matrix @xmath625 then the characteristic polynomial @xmath626 of @xmath622 is @xmath627 .
[ cubicfield ] if @xmath3 is a cubic algebra over a field @xmath13 and @xmath3 possesses a standard involution , then @xmath3 is the nilproduct ring or isomorphic to the algebra of upper - triangular @xmath628 matrices over @xmath13 . in the special case
that @xmath0 is a field @xmath13 ( so @xmath3 is a vector space ) , we can compute the isomorphism classes of exceptional @xmath13-algebras . for @xmath629 ,
let @xmath3 be the exceptional ring with multiplication given by @xmath630 . for @xmath631 , the equation
@xmath632 has a solution @xmath633\in gl_2(f)$ ] if @xmath634 and @xmath635 , or is true for any @xmath636 if @xmath637 .
therefore the two isomorphism classes are the singleton class of the nilproduct ring and all other exceptional rings over @xmath13 .
consider @xmath493 , the ring of upper triangular @xmath628 matrices over @xmath13 .
observe that @xmath493 is an @xmath13-algebra of rank @xmath406 with basis @xmath638,$ ] @xmath639,$ ] @xmath640.$ ] computing the multiplication table : @xmath641 we see that @xmath493 is not the nilproduct ring .
as in example [ ex : adjoint ] , the adjoint map is a standard involution on @xmath493 . therefore cubic @xmath13-algebras which possess a standard involution are either the nilproduct ring or isomorphic to the algebra of upper triangular @xmath628 matrices .
in this section we use wedderburn s theorem and an extended version of frobenius s theorem to classify semisimple algebras over a field @xmath13 ( with char@xmath642 ) that have a standard involution .
we show that such an algebra must be @xmath13 , a quadratic extension of @xmath13 , a quaternion algebra over @xmath13 , or @xmath385 .
these results are known @xcite .
let @xmath3 be an @xmath13-algebra .
[ def : simple ] an @xmath13-algebra is _ simple _ if it has no nontrivial ( two - sided ) ideals .
we say @xmath3 is _ semisimple _ if @xmath3 is a finite direct product of simple algebras @xcite .
[ def : degree ] the _ degree _ of @xmath3 is the smallest integer @xmath643 such that every element @xmath8 satisfies a monic polynomial @xmath644 $ ] of degree @xmath618 ; if no such integer exists , we say @xmath3 has degree @xmath645 . define the _ degree _ of an element @xmath8 to be the degree of the minimal polynomial of @xmath53 , or @xmath645 if no such polynomial exists .
[ ex : lowdegree ] if @xmath3 has degree 1 , then @xmath646 .
if @xmath3 possesses a standard involution , then since each @xmath8 satisfies a monic quadratic in @xmath647 $ ] , the degree of @xmath3 is at most 2 .
the following proof is similar to the proof of frobenius s theorem ( theorem [ thm : frobenius ] ) and appears in @xcite .
we include it for completeness .
[ thm : extendedfrobo ] let @xmath3 be a division @xmath13-@xmath648 of degree @xmath164 and suppose that char@xmath649 .
then either @xmath3 is a quadratic field extension of @xmath13 or @xmath3 is a quaternion @xmath13-algebra . since the degree of @xmath3 is not 1 , @xmath650 by example [ ex : lowdegree ]
let @xmath651 . since the degree of @xmath3 is 2 there exist @xmath652 such that @xmath653 @xmath654 .
thus the extension @xmath655 over @xmath13 is quadratic .
since @xmath656 is a field we may consider @xmath3 as a ( left ) @xmath162-vector space .
then @xmath657 if and only if dim@xmath658 , if and only if dim@xmath659 .
since @xmath660 , as in lemma [ lem : completesquare ] we may take @xmath661 for some @xmath662 by completing the square .
" let @xmath663 be conjugation by @xmath273 ; @xmath664 .
note that since @xmath661 , we have @xmath665 and @xmath666 . thus @xmath667 .
a check of the axioms shows that @xmath94 is a @xmath162-linear endomorphism of @xmath3 .
observe that @xmath668 is the identity on @xmath3 . since @xmath669 in end@xmath670 , the roots of the minimal polynomial of @xmath94 are distinct , so @xmath94 is diagonalizable with eigenvalues @xmath296 .
let @xmath671 be the eigenspace of @xmath299 and @xmath672 be the eigenspace of @xmath2 , so that @xmath673 .
since @xmath674 we have @xmath675 .
next we show that dim@xmath676 .
let @xmath677 .
consider the field @xmath678 .
the degree of @xmath679 is at most 2 because @xmath53 satisfies a quadratic in @xmath680 .
then the degree of @xmath681 is finite ( at most 4 ) .
let @xmath682 and let @xmath683 $ ] be the minimum polynomial of @xmath101 .
suppose @xmath684 factors in @xmath685 as @xmath686 .
then @xmath687 , a contradiction .
therefore @xmath685 is separable . since @xmath685 is finite and separable over @xmath13 , by the primitive element theorem there exists @xmath688 such that @xmath689 .
since @xmath621 satisfies a quadratic polynomial in @xmath690 $ ] , @xmath685 is at most a quadratic extension of @xmath13 . together with the fact that @xmath685 is an extension of @xmath162 we see that dim@xmath691 , and so @xmath692 . since @xmath675 and @xmath693 for
all @xmath677 we obtain dim@xmath676 , so @xmath694 .
now we show that dim@xmath695 .
let @xmath696 .
then @xmath697 , so @xmath698 , and hence all elements of @xmath671 conjugate @xmath273 to @xmath699 . also note that @xmath700 .
let @xmath701 .
then @xmath702 so the product of two elements of @xmath671 is in @xmath672 and also is in the centralizer of @xmath273 . let @xmath703 .
since @xmath704 we see that any element of @xmath671 differs multiplicatively from @xmath262 by an element of @xmath694 .
thus @xmath705 and so dim@xmath695 .
finally we show that @xmath3 is a quaternion algebra . by completing the square
, we may assume that @xmath359 .
since @xmath706 we have @xmath707 so @xmath708 .
therefore @xmath709 .
[ cor : divisionstandard ] if @xmath237 is a degree @xmath164 division @xmath13-algebra then @xmath237 possesses a standard involution . by theorem [ thm : extendedfrobo ] , @xmath237 is either @xmath13 , a quadratic field extension of @xmath13 , or a quaternion algebra over @xmath13 .
if @xmath710 then the identity is a standard involution . if @xmath237 is a quadratic field extension over @xmath13 then @xmath237 is a quadratic algebra over @xmath13 .
we showed that quadratic algebras possess a ( unique ) standard involution in section 2.3 .
if @xmath237 is a quaternion algebra , then @xmath237 has a standard involution by example [ ex : quaternioninvolution ] .
[ lem : involutionmnd ] let @xmath237 be a division algebra containing field @xmath13 and consider the matrix ring @xmath711 as an @xmath13-algebra . if @xmath711 possesses a standard involution then @xmath712 , and if @xmath713 then @xmath710 .
let @xmath713 .
by example [ ex : adjoint ] , @xmath371 has a standard involution ( the adjoint map ) .
let @xmath714 and consider the matrix @xmath715 $ ] . since @xmath716 possesses a standard involution , @xmath717 $ ] has degree at most 2
so there exist @xmath652 such that @xmath717 ^ 2 - t[i ] +
n = 0 $ ] . then in particular , @xmath718 .
hence every element @xmath714 satisfies a monic quadratic polynomial in @xmath719 $ ] . since @xmath660 we can complete the square and assume @xmath358
now consider @xmath720\in m_2(d)$ ] .
there exist @xmath721 such that @xmath720 $ ] satisfies @xmath722 $ ] .
then in particular we have @xmath723 from which we see that @xmath724 and @xmath725 , and so @xmath726 . from @xmath727 , it follows that @xmath728 .
since @xmath237 is a division algebra @xmath273 must equal @xmath296 , a contradiction
. therefore if @xmath713 then @xmath710 .
suppose @xmath729 .
let @xmath730 be the diagonal matrix with diagonal entries @xmath731 .
the characteristic polynomial @xmath732 of @xmath372 is @xmath733 . since the characteristic polynomial has three distinct roots ( @xmath734 ) , the minimal polynomial @xmath735 of @xmath372 has deg@xmath736 , a contradiction + ( * ? ? ?
* section 12.2 ) .
wedderburn s theorem greatly simplifies our classification .
[ thm : wedderburn ] let @xmath80 be a simple @xmath13-algebra
. then @xmath80 is isomorphic to @xmath711 for some unique ( up to isomorphism ) division @xmath13-algebra @xmath237 and unique @xmath737 .
[ thm : simple ] let @xmath3 be a simple @xmath13-algebra of degree less than or equal to @xmath164 , with char(@xmath738
. then @xmath3 is isomorphic to @xmath13 , a quadratic field extension of @xmath13 , or a quaternion algebra over @xmath13 . if deg@xmath739 then @xmath646 .
suppose deg@xmath740 . by wedderburn
s theorem , @xmath741 for a unique ( up to isomorphism ) division @xmath13-algebra @xmath237 and a unique @xmath737 .
if @xmath742 then @xmath3 is a division @xmath13-algebra , of degree 2 .
then by theorem [ thm : extendedfrobo ] , @xmath3 is is isomorphic to @xmath13 , a quadratic field extension of @xmath13 , or a quaternion algebra over @xmath13 .
for @xmath743 we apply lemma [ lem : involutionmnd ] to complete the proof .
note that @xmath371 is a quaternion algebra @xcite .
we include two remarks about degree .
[ prop : relativedegree ] let @xmath3 and @xmath80 be @xmath13-algebras of finite degree . then deg@xmath744 if and only if there exist @xmath8 and @xmath745 such that deg@xmath746 , deg@xmath747 , and the respective minimal polynomials of @xmath53 and @xmath101 are relatively prime .
let @xmath748 denote the minimal polynomial of @xmath53 in @xmath647 $ ] .
let @xmath8 and @xmath749 satisfy deg@xmath746 , deg@xmath747 , and that @xmath748 and @xmath750 are relatively prime .
note that @xmath751 has degree deg@xmath752 .
if @xmath753 satisfies a polynomial @xmath754 $ ] , then @xmath755 . then since @xmath756 are relatively prime , deg@xmath757 , so deg@xmath758 .
since @xmath753 satisfies @xmath751 in particular , we have deg@xmath759 .
thus deg@xmath760 .
let @xmath761 . since @xmath762 satisfies @xmath763 and deg@xmath764 , we have deg@xmath765 . note that for all @xmath766 and @xmath749 we have deg@xmath767 and deg@xmath768 , so that deg@xmath764 , with equality if and only if deg@xmath769 and deg@xmath770 .
suppose that deg@xmath771 take @xmath761 such that the minimal polynomial @xmath772 of @xmath762 has degree @xmath773 . since @xmath762 satisfies @xmath763 we have deg@xmath774 .
therefore deg@xmath769 and deg@xmath770 , and since deg@xmath775 we see that @xmath776 and @xmath777 are relatively prime . for two @xmath13-algebras of finite degree @xmath3 and @xmath80
it is not always true that deg@xmath744 .
for example , for @xmath778 , consider the boolean ring @xmath779 .
for all @xmath780 we have @xmath781 , so deg@xmath782 , not @xmath783 . in proposition
[ prop : relativedegree ] the hypothesis that the respective minimal polynomials of @xmath53 and @xmath101 are relatively prime avoids such exceptions .
[ prop : semisimple ] let @xmath3 be a nonsimple semisimple @xmath13-algebra of degree @xmath164 , with char(@xmath738
. then @xmath3 is isomorphic to @xmath385 .
an example of such an algebra is @xmath385 which has standard involution @xmath784 ( example [ ex : fcrossf ] ) .
so @xmath385 has degree 2 . by wedderburn s theorem , @xmath785 where @xmath786 are division @xmath13-algebras uniquely determined ( up to isomorphism ) by @xmath13 , as are the integers @xmath787 .
let @xmath788 .
since deg@xmath789 , there exist @xmath652 such that @xmath176 .
in particular , @xmath790 for @xmath791 .
since @xmath53 is arbitrary it follows that deg@xmath792 for @xmath791 .
then by theorem [ thm : simple ] , @xmath793 is isomorphic to @xmath13 , or a quadratic field extension of @xmath13 , or a quaternion algebra over @xmath13 .
next we show that @xmath794 for @xmath795 .
let @xmath796 . since @xmath660 we can ( up to isomorphism ) take @xmath358 by completing the square .
consider the matrix @xmath797 .
let @xmath798 .
let @xmath652 satisfy @xmath799 .
then ( by considering the matrix entries ) since @xmath800 we have @xmath801 and @xmath802 in @xmath803 , and so @xmath804 and @xmath805 .
therefore @xmath806 , from which we see that @xmath728 . since @xmath803 is a division algebra , @xmath807 , a contradiction to @xmath796 .
therefore @xmath808 .
finally we show that @xmath712 .
let @xmath809 .
let @xmath652 satisfy @xmath799 .
from the first coordinate we obtain @xmath469 . from the second we get @xmath810 , and from the third coordinate
we see that @xmath811 , a contradiction
. therefore @xmath712 .
we would like to extend our investigation to algebras of high ( greater or equal to 4 ) rank .
much work has been done in in rank 4 because of the importance of quadratic algebras .
however , little is known about algebras of rank 5 or greater . in rank 5
we may be able to write down a multiplication table , ( as we did in rank 3 ) , which we could then use to find isomorphism classes .
it may be that the results for ranks 5 and lower suggest a general result for finite - rank algebras .
999 david s. dummit and richard m. foote , _ abstract algebra , @xmath406rd ed . , _ wiley , hoboken , nj , 2004 .
benedict h. gross and mark w. lucianovic , _ on cubic rings and quaternion rings , _
j. number theory * 129 * ( 2009 ) , no .
@xmath812 , 14681478 .
mark w. lucianovic , _ quaternion rings , quadratic forms , and fourier coefficients of modular forms on @xmath813 _ , harvard university ph.d .
thesis , may 2003 .
r. s. palais , _ the classification of real division algebras , _ amer .
math . monthly .
* 75 * ( 1968 ) no .
4 , 366368 .
john voight , _ the arithmetic of quaternion algebras _ , in preparation .
john voight , _ rings of low rank with a standard involution , _
illinois j. math , to appear .
here we include the computations showing associativity for theorem [ thm : associative ] . by lemma @xmath814
, we may consider case ( c ) , in which @xmath815 , and case ( e ) , in which @xmath816 . in case ( c )
we have the following multiplication table for @xmath3 @xmath545 where @xmath546 are arbitrary elements of @xmath0 .
let @xmath817 , @xmath818 , and @xmath819 be three elements of @xmath372 .
first we compute @xmath488 : @xmath820(p_3 + q_3i + r_3j ) \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \end{aligned}\ ] ] @xmath821 @xmath822 @xmath823 @xmath824 observe that in case ( c ) , @xmath831 .
next consider case ( e ) , in which the multiplication table for @xmath3 is @xmath567 where @xmath537 are arbitrary elements of @xmath0 ( with @xmath537 not both 0 ) .
+ we compute @xmath488 : @xmath820(p_3 + q_3i + r_3j ) \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \end{aligned}\ ] ] @xmath832 @xmath833 @xmath834 next we compute @xmath825 : @xmath826 \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \end{aligned}\ ] ] @xmath835 @xmath836 @xmath837 observe that in case ( e ) , @xmath838 . since associativity holds in both cases , the multiplication on @xmath372 defined by table [ eqn : mtable ] and relations [ eqn :
relations]@xmath481 is associative . | we classify ( possibly noncommutative ) algebras of low rank over a domain @xmath0 .
we first review results for algebras of rank 2 and for finite - dimensional division algebras over @xmath1 .
these results motivate us to consider which algebras possess a standard involution .
our main result is that algebras of rank 3 are either commutative or possess a standard involution .
thanks to my supportive parents barry and mara . comments and encouragement from my fellow master s degree students have been invaluable , particularly eric roma , mike klug , jared krogsrud , sam schiavone , brian lang , andy reagan , and jonathan godbout . i could not have completed this work without the background and skills i developed in my coursework , especially in the algebra and real analysis sequences taught by professors john voight and jonathan sands , respectively .
thank you to the members of the committee and to my advisor professor john voight , without whose guidance , tireless editing , and infectious energy this thesis would not exist .
thanks above all to my beloved wife , andrea . |
narrow , low - energy ( @xmath2 kev ) spectral lines in grbs have been reported using the data of several instruments ( e.g. , see review @xcite ) . based on these reports , the batse team expected lines to be easy to find and looked for them manually @xcite .
the reality was different : lines were not obvious in burst spectra . in order to be sure that the manual search had not missed any lines
, we implemented a comprehensive , automatic computer search @xcite .
bursts with at least one spectrum with a normed signal - to - noise ratio ( snr ) @xcite near 40 kev above 5.0 are searched . for each burst , we form spectra from each individual spectral record , every pair , triple , etc . the spectra so formed overlap in many cases .
once a burst is selected for the search , spectra are searched regardless of the presence of burst flux low snr spectra serve as controls .
each spectrum is fit with a continuum model and then a series of fits are made adding narrow lines at a closely spaced grid of line centroids extending to 100 kev .
line candidates are identified by a @xmath3 change @xmath4 of more than 20 , corresponding to a chance probability in a single spectrum of @xmath5 .
the probability is calculated for two line parameters , intensity and centroid , since the intrinsic width is assumed narrow compared to the detector resolution .
so far , 133,000 spectra formed from 12,000 spectral records from 117 bursts have been searched .
most of these spectra have snr too low to support the detection of a line .
only 16,000 have normed snr @xmath6 5 , which our simulations show is needed to have a reasonable sensitivity to lines similar to those found in the _ ginga _ data @xcite .
the search identified several cases with @xmath4 values exceeding our significance threshold , but which we rejected either because they occur in background intervals or because they become insignificant when a better fit requiring a non - negative flux is made .
these cases were located in spectra of low snr , and because of the large number of low - snr spectra , they are consistent with statistical fluctuations .
several possible candidates with angles between the burst direction and detector normal exceeding @xmath7 are inconsistent with the data of other detectors .
we believe these cases to be due to inadequacies in the detector model and we have set them aside .
after these exclusions , our candidate list has 12 members with @xmath4 values ranging from the threshold of 20 to 31.8 ( @xmath8 ) .
the normed snrs of the spectra in which these candidates are most significant range from 2.1 to 18.4 , with 10 normed snr values exceeding 5 .
these snr values are reasonable for real features .
the number of independent trials is a nebulous concept . while there are roughly 16,000 spectra sufficiently bright for there to be a high probability of detecting a real line , many of these spectra overlap , e.g. , starting a few records earlier or later , or differing in length by a few records .
consequently the number of bright , independent spectra is one or more orders of magnitude below 16,000 . the number of independent energy resolution elements averages about 5 per spectrum .
we estimate that the ensemble chance probability of the most - significant candidate is below @xmath9 , and probably much lower than this value .
we therefore believe that few , if any , of these candidates are statistical fluctuations . of the 12 candidates , in all but one
the best line fit is an emission feature with a centroid near 40 kev .
the twelfth candidate is an absorption feature at 60 kev .
typically the lowest energy of the data is about 20 kev , so we are unable to find lines much below 40 kev .
we have previously presented results @xcite for a candidate in grb 940703 ( trigger 3057 ) which was usefully observed by only one batse spectroscopy detector ( sd ) .
observations are useful only from sds in high - gain mode which point to within @xmath7 of the burst . since that time we have searched more bursts and begun more detailed analyses of the nine candidates that were usefully observed by more than one sd .
these multiple - detector cases allow stringent tests of the reality of the candidates .
ideally , the data from another detector will confirm the feature found by the automatic search .
while some features may not be confirmed by another detector due to differing snrs between the detectors or from plausible statistical fluctuations , the data of all detectors must be consistent .
grb 941017 ( trigger 3245 ) was usefully observed by sds 0 and 5 .
the automatic search identified a candidate in the data of sd 0 during the rising portion of the burst ( fig .
a spectrum is not available for sd 5 for the time interval with the best @xmath4 value for sd 0 .
we therefore continued our analysis using spectra from a similar time interval ( 9.728 to 23.936 s ) which is available for both detectors .
the results of fits to this interval are listed in table 1 .
because of the non - optimum time interval , the significance of the feature in the data of sd 0 is reduced .
the line was discovered in the data of sd 0 so we may regard the centroid as a priori determined when we analyze the data of sd 5 .
a one parameter line fit yields a significance for the feature in sd 5 of @xmath10 .
the joint fit ( fig .
2 ) yields a significance of @xmath11 , better than the value obtained with sd 0 alone despite the forced change to the common time interval .
sd 5 provides evidence for the feature independent of the detector in which the feature was found by the automatic search .
.line fits to the common time interval , grb 941017 .
[ cols="^,^,^,^,^,^ " , ] to demonstrate the existence of a line , we must show that the data require a line independent of which reasonable continuum model is assumed @xcite .
the above fits were made using the standard ` grb ' continuum function of band . however , if we use an alternative continuum model with a low - energy break _ in addition _ to the standard high - energy break , the significance of the line in sd 0 is reduced to @xmath12 ( @xmath13= 4% ) . while this model still involves a low - energy spectral feature ( fig .
3 ) , it means that for this burst we can not prove that the feature is a line .
since our last report @xcite , we have completed the automatic search for bright bursts through may 1996 .
we have identified 12 line candidates and have begun detailed analysis . because the 12 candidates appear as expected in the small fraction of spectra with high snr , their probability of appearing in the ensemble by chance is low . for all except one of the 12 candidates ,
the best line fit is an emission line with a centroid near 40 kev .
our simulations show that we have useful sensitivity to _ ginga_-like absorption lines at 40 kev @xcite , so the lack of detections of absorption lines gives a constraint on their frequency .
however , with the small number of bright bursts seen by batse and _ ginga _ , there is no significant discrepancy between the two instruments @xcite .
observations made with multiple detectors have the strong advantage of redundantly verifying the existence of the features .
we have now carefully investigated several candidate features , including some observed with multiple detectors .
so far the evidence appears good for the existence of spectral features , however we do not want to make a blanket statement that batse has observed spectral features until we have carefully examined all 12 candidates .
note that we are stating that the evidence appears good for the existence of _ spectral features .
_ in the cases
examined so far , we have not been able to demonstrate that the spectral features must be narrow lines
an alternative explanation is possible : a low - energy break in addition to the normal break in the few 100 kev to 1 mev region . | we have developed an automatic search procedure to identify low - energy spectral features in grbs . we have searched 133,000 spectra from 117 bright bursts and have identified 12 candidate features with significances ranging from our threshold of @xmath0 5e@xmath15 to @xmath0 1e@xmath17 .
several of the candidates have been examined in detail , including some with data from more than one batse spectroscopy detector .
the evidence for spectral features appears good ; however , the features have not conclusively been shown to be narrow lines .
= = = 1=1=0pt = 2=2=0pt |
the recent revival of interest in the properties of two - band ( or , more generally , multiband ) superconductors has been largely stimulated by the discovery of superconductivity in mgb@xmath0 ( ref . ) .
other candidates for multiband superconductivity include nickel borocarbides ( ref . ) , nbse@xmath0 ( ref . ) , cecoin@xmath1 ( ref . ) , and also the iron - based high - temperature superconductors , see ref . for a review .
these discoveries have shown that multiband superconductivity , which is characterized by a significant difference in the order parameter magnitudes and/or phases in different bands , might be a much more common phenomenon than it was previously thought .
one important class of multiband superconductors is noncentrosymmetric compounds with a strong spin - orbit ( so ) coupling of conduction electrons with the lattice . since the discovery of superconductivity in cept@xmath2si ( ref . ) , the list of noncentrosymmetric superconductors has grown to include dozens of materials , such as uir ( ref . ) , cerhsi@xmath2 ( ref . ) , ceirsi@xmath2 ( ref . ) , y@xmath0c@xmath2 ( ref . ) , li@xmath0(pd@xmath3,pt@xmath4)@xmath2b ( ref . ) , and many others . in noncentrosymmetric crystals ,
the bloch bands are split by the so coupling and the cooper pairing of electrons from different bands is suppressed , resulting in an effectively two - band picture of superconductivity .
both the spin structure of the bands and the momentum - space topology of the bloch wavefunctions are nontrivial , which brings about a number of novel properties , such as the magnetoelectric effect in the superconducting state,@xcite topologically protected gapless boundary modes and quantum spin hall effect,@xcite and the anomalous de haas - van alphen and hall effects in the normal state.@xcite a comprehensive review of the recent developments in the field can be found in ref . .
one of the most striking features of noncentrosymmetric superconductors is the existence of various unusual nonuniform superconducting states in the presence of a magnetic field @xmath5 ( refs . ) , or even without any field ( ref . ) .
the best studied example is the helical states , with the order parameter proportional to @xmath6 , where @xmath7 is linear in @xmath5 , which originate from the deformation of the bloch bands by the field , see ref . for a review .
the helical states , as well as their nonlinear modifications , such as `` multiple-@xmath7 '' states , have been previously studied only in some limiting cases , either using a bardeen - cooper - schrieffer ( bcs ) model with a spin - singlet attraction@xcite or assuming that superconductivity appears only in one of the bands.@xcite in both cases , the order parameter has one component .
the purpose of this paper is to develop a phenomenological theory of nonuniform states in noncentrosymmetric superconductors in the general case , fully taking into account the two - component nature of superconductivity in these materials .
generalization of the bcs theory to the case of two spin - degenerate bands was originally introduced in ref . .
subsequent work has shown that many properties of multiband superconductors differ qualitatively from the single - band case , with the most spectacular features associated with the additional degrees of freedom the relative phases of the pair condensates in different bands .
for example , if the condensate phases have different windings around a vortex core , than the vortex will carry a fractional magnetic flux.@xcite in addition to exotic vortices , there is another type of topological defects specific to two - band superconductivity , namely phase solitons , in which the relative phase exhibits a kink - like variation by @xmath8 between its asymptotic mean - field values.@xcite increasing the number of bands opens up even more intriguing possibilities .
for instance , superconducting states that break time - reversal symmetry can exist in `` frustrated '' three - band systems,@xcite with domain walls separating degenerate ground states.@xcite it has been proposed that the phase solitons in two - band superconductors can be dynamically created in nonequilibrium current - carrying states,@xcite or by the proximity effect with a conventional @xmath9-wave superconductor,@xcite but experimental observation of these effects has remained a challenge . in this paper , we show how the elusive phase solitons can be spontaneously formed in noncentrosymmetric superconductors in a sufficiently strong magnetic field .
we focus on two - dimensional ( 2d ) noncentrosymmetric superconductors in a parallel field .
these systems can be realized experimentally at an interface between two different non - superconducting materials,@xcite near a doped surface of an insulating crystal,@xcite or near the surface of a topological insulator.@xcite in all these systems the mirror symmetry between two half - spaces separated by an interface is explicitly broken .
we use the two - band ginzburg - landau ( gl ) formalism@xcite modified to include first - order gradient terms ( the lifshitz invariants ) to catch the effects specific to noncentrosymmetric superconductors .
the paper is organized as follows : in sec .
[ sec : two - band picture ] we develop a two - band description of noncentrosymmetric superconductors . in sec .
[ sec : helical state ] , the helical instability in a weak magnetic field is considered and the quasiparticle spectrum in the helical state is calculated . in sec . [ sec : phase solitons ] we discuss the phase solitons and soliton lattices in a strong field .
throughout the paper we use the units in which @xmath10 .
in order to see how nondegenerate bands are formed in a noncentrosymmetric crystal with the so coupling , we start with the following hamiltonian of noninteracting electrons : @xmath11a^\dagger_{\bk s}a_{\bk s'}.\ ] ] here @xmath12 are spin indices , @xmath13 are the pauli matrices , @xmath14 is the band dispersion without the so coupling , and the sum over @xmath15 is restricted to the first brillouin zone .
the so coupling of electrons with the crystal lattice is described by @xmath16 , which satisfies @xmath17 .
the momentum dependence of the so coupling is dictated by the requirement that it must be invariant under the crystal symmetry operations , in the following sense : if @xmath18 is any operation from the point group @xmath19 of the crystal , then @xmath20 .
the complete list of representative expressions for @xmath16 for all noncentrosymmetric point groups can be found in ref . .
for instance , in the least symmetric case of a triclinic crystal , i.e. if @xmath21 , the direction of @xmath22 is not related to the crystallographic axes and we have @xmath23 , where the coefficients @xmath24 form a real @xmath25 matrix . in a cubic crystal with @xmath26 , which describes the point symmetry of li@xmath0(pd@xmath3,pt@xmath4)@xmath2b , the simplest form compatible with all symmetry requirements is @xmath27 , where @xmath28 is a constant .
expressions become more complicated if the point group contains improper elements . for example , for the tetragonal group @xmath29 , which is relevant for cept@xmath2si , the so
coupling is given by @xmath30 .
since we focus on the purely 2d case , we can set @xmath31 .
then the so coupling takes the rashba form : @xmath32 which was originally introduced to describe the effects of the absence of mirror symmetry in semiconductor quantum wells.@xcite diagonalizing eq .
( [ h - free ] ) we arrive at the following band dispersion functions : @xmath33 where @xmath34 is called helicity .
therefore , the bands are nondegenerate ( except the lines or points where @xmath35 ) , invariant with respect to all operations from @xmath19 , and also even in @xmath15 . the last property is a consequence of time reversal symmetry .
indeed , the bloch states @xmath36 and @xmath37 belong to @xmath15 and @xmath38 , respectively , and have the same energy . here
@xmath39 is the time reversal operation for spin-@xmath40 particles and @xmath41 is the complex conjugation .
since the bands are nondegenerate , one can write @xmath42 , where @xmath43 is a nontrivial phase factor.@xcite the latter is given by @xmath44 for the model described by eq .
( [ h - free ] ) .
let us now introduce an external magnetic field @xmath5 .
we consider only a 2d superconductor , with the field parallel to its plane , therefore the vector potential and the orbital effects of the field can be neglected . adding the zeeman interaction @xmath45 to eq .
( [ h - free ] ) , where @xmath46 is the bohr magneton , and diagonalizing the resulting hamiltonian we obtain the following energy eigenvalues : @xmath47 .
assuming that the zeeman energy is small compared with the so coupling and expanding these eigenvalues to the first order in @xmath5 , we have@xcite @xmath48c^\dagger_{\bk\lambda}c_{\bk\lambda}.\ ] ] therefore , the bands are asymmetrically deformed by the field and no longer even in @xmath15 .
this allows for the cooper pairing with a nonzero centre - of - mass momentum to occur , leading to a variety of field - induced nonuniform superconducting states , which are discussed in the subsequent sections .
we assume the following hierarchy of the energy scales : @xmath49 , where @xmath50 is the superconducting critical temperature , @xmath51 is the energy cutoff of the pairing interaction , and @xmath52 is the fermi energy , which is a good assumption in realistic noncentrosymmetric superconductors.@xcite then the so - split bands are sufficiently separated for the pairs to form independently in each band , and it is natural to use the basis of the helicity band eigenstates to introduce an attractive interaction between electrons in the cooper channel . following the standard bcs ideology ,
we assume that the pairing interaction is only effective when the quasiparticle momenta are close to the fermi surfaces and , therefore , the interband pairing , i.e. the formation of the cooper pairs of electrons with opposite helicities , is suppressed .
this leads us to the pairing hamiltonian @xmath53 where @xmath54 is the system volume and @xmath55 is the pairing interaction function .
the latter has the form @xmath56 where @xmath57 can be represented as a bilinear combination of the basis functions of irreducible representations of @xmath19 and , therefore , is amenable to the usual symmetry analysis.@xcite note that the dependence of the pairing interaction on the center - of - mass momentum of the pairs is neglected , which is legitimate since @xmath58 is small compared to the fermi momenta in the two bands .
the simplest model corresponds to isotropic basis functions of the unit representation , with @xmath59 the strength of intraband pairing is described by the coupling constants @xmath60 and @xmath61 , while that of interband pairing by @xmath62 . in a uniform superconducting state , using the mean - field approximation to decouple the interaction in eq .
( [ h int reduced ] ) we arrive at the following expression : @xmath63,\ ] ] where @xmath64 is the gap function in the @xmath65th band ( we have omitted a @xmath66-number term on the right - hand side of the last equation ) . in the isotropic pairing model , see eq .
( [ isotropic - pairing ] ) , the gap functions have the form @xmath67 where @xmath68 play the role of the superconducting order parameters . the model defined by eqs . ( [ h_0 ] ) and ( [ h int reduced ] ) is formally similar to the two - band bcs theory,@xcite whose applications to mgb@xmath0 , iron - based superconductors , and other systems have been extensively studied recently .
the order parameter is represented by two complex functions @xmath69 and @xmath70 . an obvious difference is that in our case the two bands are nondegenerate and the gap functions contain the phase factors @xmath71 .
however , these phase factors do not affect the bulk observable quantities which are determined by the quasiparticle excitations , such as spin susceptibility and electronic specific heat , nor do they enter the gl free energy , which is expressed in terms of @xmath69 and @xmath70 .
the order parameters in the helicity band representation are related to the singlet ( @xmath72 ) and `` protected '' triplet ( @xmath73 ) components in the spin representation as follows : @xmath74 and @xmath75 ( ref . ) . in particular , in the case of a bcs - like point attraction all the coupling constants in eq .
( [ isotropic - pairing ] ) are the same : @xmath76 , and we have @xmath77 , i.e. the pairing is isotropic singlet .
in the opposite case of a `` single - band '' model , either @xmath69 or @xmath70 is zero . in general , however , both components are nonzero and different , and can also depend on coordinates . in the spin representation
this translates into an order parameter having both singlet and triplet components .
nonuniform superconducting states are most efficiently treated using the gl formalism .
the phenomenological gl functional of a noncentrosymmetric superconductor can be obtained in the standard fashion by expanding the free energy ( or , more precisely , the difference between the free energies in the superconducting and normal states ) in powers of @xmath69 and @xmath70 and their gradients and keeping all terms allowed by symmetry . in the case of a 2d noncentrosymmetric superconductor in a parallel magnetic field
@xmath5 we have the following expression for the gl free energy density : @xmath78 the intraband contributions are given by @xmath79+l_\lambda h^2|\eta_\lambda|^2,\ ] ] while @xmath80 is the band - mixing term describing the `` josephson coupling '' of the two order parameters , which originates in the interband pairing terms in eq .
( [ h int reduced ] ) .
the free energy ( [ f_gl ] ) yields the following expression for the supercurrent : @xmath81 where @xmath82 is the absolute value of the electron charge .
the first three terms in eq .
( [ f - pm ] ) are the usual gl uniform and gradient terms .
the temperature dependence enters through the coefficients @xmath83 , which have the form @xmath84 , where @xmath85 and @xmath86 is the transition temperature the @xmath65th band would have at zero field in the absence of any interband coupling . the fourth term , sometimes called the lifshitz invariant , is linear both in gradients and in the magnetic field , and is specific to noncentrosymmetric superconductors .
its origin can be traced to the deformation of the helicity bands by the field , see eq .
( [ h_0 ] ) .
microscopic derivation of the lifshitz invariant can be found in ref . .
it is the lifshitz invariants that are responsible for unusual nonuniform states created by the field , see secs .
[ sec : helical state ] and [ sec : phase solitons ] below .
the last term in eq .
( [ f - pm ] ) describes the paramagnetic suppression of the critical temperature , which is due to the change in the paramagnetic susceptibility in the superconducting state compared with the normal state .
microscopic theory yields expressions for the coefficients in the gl functional in terms of the fermi - surface averages of the order parameter , the fermi velocity , and the so coupling direction @xmath87 , see ref . .
assuming isotropic bands , one has the following order - of - magnitude estimates : @xmath88 where @xmath89 is the fermi - level dos and @xmath90 is the fermi velocity in the @xmath65th band , while @xmath91 is the superconducting critical temperature at zero field .
one can recover the single - band and the singlet bcs limits from the general gl functional ( [ f_gl ] ) as follows .
the single - band case corresponds to the absence of the interband coupling , i.e. @xmath92 . assuming @xmath93 , superconductivity appears only in the `` @xmath94 '' band , i.e. @xmath95 , at least near the critical temperature . in the bcs case , @xmath96 , and one obtains a one - component gl free energy in terms of @xmath97 , with the lifshits invariant
the critical temperature , @xmath98 , of the second - order superconducting phase transition is obtained by solving the linearized gl equations . assuming @xmath99 and anticipating a possible helical instability modulated perpendicularly to the field
, we seek the order parameters in the form @xmath100 .
then from the free energy ( [ f_gl ] ) it follows that @xmath101 where @xmath102 . setting the determinant of the matrix on the left - hand side to zero ,
one arrives at the following expression for the critical temperature as a function of the magnetic field and the helical modulation wavevector : @xmath103 the actual transition temperature @xmath98 and the modulation wavevector are obtained by maximizing the above expression with respect to @xmath104 . at zero field
the maximum critical temperature is achieved in the uniform state , i.e. at @xmath105 , and is given by @xmath106 at a small but nonzero field , we seek the wavevector in the form @xmath107 . expanding eq .
( [ t_c - hq ] ) in powers of @xmath108 and maximizing with respect to @xmath104 , we obtain : @xmath109 which corresponds to @xmath110 here @xmath111 and @xmath112 thus we see that , similarly to the single - component helical state,@xcite the superconducting order parameter is nonuniform , with the modulation wavevector linearly proportional to the field .
the suppression of the critical temperature is quadratic in @xmath108 , which is typical of a paramagnetic pair breaking .
according to eq .
( [ t_c - h ] ) , the pair breaking is weakened in the presence of the helical instability .
we would like to stress that , in contrast to the larkin - ovchinnikov - fulde - ferrell nonuniform state,@xcite which only exists in paramagnetically - limited superconductors at sufficiently strong magnetic fields , the helical state appears at an arbitrarily weak field .
the last term on the right - hand side of eq .
( [ f - pm ] ) , which describes the paramagnetic suppression of superconductivity , is crucial for maintaining the stability of the system at @xmath113 .
if this term were not included , the linear in gradient terms would result in the critical temperature unphysically increasing as a function of the field . to avoid this
, one has to assume that @xmath114 .
despite the fact that the order parameter in the helical state is proportional to @xmath115 , the supercurrent is equal to zero .
indeed , eq . ( [ supercurrent ] )
yields the following expression : @xmath116 where @xmath117 , and @xmath118 . substituting eq .
( [ q - h ] ) , we obtain : @xmath119 here we used the ratio of the order parameters at zero field , @xmath120 , which follows from eq .
( [ linear - gl ] ) . that the current must vanish
can also be understood using the following simple argument .
uniform supercurrent is obtained by differentiating the total free energy with respect to a uniform vector potential : @xmath121 , where @xmath66 is the speed of light .
due to the gauge invariance , we have @xmath122 and , therefore , @xmath123 , because the free energy has a minimum at the equilibrium value of the helical modulation . although the above calculations apply in the case of a 2d isotropic noncentrosymmetric superconductor , they can be straightforwardly extended to an arbitrary in - plane symmetry . in the general case , the lifshitz invariant is bilinear in both the order parameter gradients and the magnetic field and eq .
( [ f - pm ] ) is replaced by @xmath124 where @xmath125 , and the einstein summation convention is assumed .
the matrices @xmath126 and @xmath127 are symmetric , while @xmath128 is neither symmetric nor antisymmetric , in general .
the isotropic case is recovered when @xmath129 , @xmath130 , and @xmath131 . in the least symmetric case of @xmath21
all elements of the matrices @xmath126 , @xmath128 , and @xmath127 are nonzero .
we seek the order parameter in the form @xmath132 and obtain for the critical temperature the same expression as eq .
( [ t_c - hq ] ) .
the only difference is that @xmath133 are now given by @xmath134 maximizing the critical temperature with respect to @xmath7 , we obtain : @xmath135 and @xmath136h_ih_j,\ ] ] where @xmath137 therefore , the critical temperature is suppressed by the field , quadratically in @xmath5 , but the helical modulation wavevector is no longer perpendicular to the field , in general .
the supercurrent is equal to zero , for the same reason as explained above .
the helical states can be observed , for instance , in tunneling experiments , which measure the density of states ( dos ) of quasiparticle excitations . in this subsection
we calculate the quasiparticle spectrum in the state with @xmath138 , by solving the bogoliubov - de gennes ( bdg ) equations independently in each band .
derivation of the bdg equations for a noncentrosymmetric superconductor is presented in appendix [ sec : bdg - helical ] .
the helical state order parameter in the momentum representation is given by @xmath139 , and the bdg hamiltonian , see eq .
( [ h - bdg - transformed ] ) , takes the following form : @xmath140 where @xmath141 . to make the above hamiltonian diagonal in @xmath15-space
, we perform a unitary transformation @xmath142 , where @xmath143 and @xmath144 is the position operator in @xmath15-space .
one can see that the exponentials in @xmath145 act as finite displacement operators in momentum space , because @xmath146 .
after the unitary transformation , the bdg hamiltonian becomes @xmath147 in a weak field , the helical modulation wavevector is small and one can expand the last expression in powers of @xmath7 and @xmath5 , with the following result : @xmath148 where @xmath149 describes the field - induced deformation of the fermi surface and @xmath150 is the band velocity of quasiparticles .
now we are in a position to calculate the quasiparticle dos in the the @xmath65th band : @xmath151,\ ] ] where @xmath152 are the eigenvalues of the hamiltonian ( [ h - bdg - weak - field ] ) . near the fermi surface
, one can integrate eq .
( [ dos ] ) over @xmath153 and obtain : @xmath154 ^ 2-|\eta_\lambda|^2}}\right\rangle_{fs},\ ] ] where the angular brackets denote the average over the @xmath65th fermi surface , restricted by the condition @xmath155 .
since the dos satisfies the bdg electron - hole symmetry , @xmath156 , we focus only on the upper half of the spectrum , i.e. on @xmath157 . to make analytical progress we assume a cylindrical fermi surface with the rashba so coupling , see eq .
( [ rashba - soc ] ) , and @xmath99 , with @xmath158 . then , @xmath159 it follows from eqs .
( [ as ] ) and ( [ coeff - estimates ] ) that both the zeeman and `` helical '' contributions to @xmath160 , described by the first and second terms in @xmath161 , respectively , are of the same order . from eq .
( [ dos - final ] ) we obtain the following expression for the dos : @xmath162 where @xmath163 the angular integration in here is restricted by the condition @xmath164 .
we calculated the last integral numerically , with the result presented in fig .
[ fig : dos ] .
this plot reveals two prominent features .
first , the gap in the quasiparticle spectrum is given by @xmath165 , i.e. it is smaller in the helical state than in the uniform state .
second , the inverse - square - root dos singularity , which is a hallmark of the bcs theory , is replaced by a much weaker singularity at @xmath166 . a straightforward analytical evaluation of eq .
( [ ixy ] ) gives @xmath167 immediately above the gap edge , and @xmath168 near the peak .
tunneling experiments probe the total dos , which is given by @xmath169 and sketched in fig .
[ fig : dos - total ] . th band , for @xmath170 .
the gap edge is shifted to @xmath171 , while the logarithmic singularity occurs at @xmath172.,width=226 ]
in this section we show that a soliton - like texture is spontaneously formed in the superconducting state above a certain critical magnetic field .
phase solitons are the simplest topological defects that can exist in a two - band superconductor . in a phase
soliton the relative phase of the two order parameters exhibits a kink - like variation of @xmath8 , approaching its mean - field value at infinity.@xcite according to sec .
[ sec : two - band picture ] , noncentrosymmetric superconductors can be viewed as two - band systems and , therefore , are expected to support phase solitons .
we assume the in - plane isotropy and set @xmath99 . to make analytical progress , we work at zero temperature and employ the london approximation , in which the gap magnitudes are constant . considering a planar texture perpendicular to the @xmath173 axis ,
the order parameter components are given by @xmath174 .
the supercurrent in this state is given by eq .
( [ current - x ] ) .
the current conservation implies that @xmath175 , where the value of the constant is set by external sources and can be assumed to be zero .
the condition @xmath123 allows one to express the gradients of @xmath176 and @xmath177 in terms of the gradient of the relative phase @xmath178 : @xmath179 where @xmath180 it follows from eq .
( [ phi - grads ] ) that one can have two very different physical situations , depending on whether the phases @xmath176 and @xmath177 are locked together , with @xmath181 taking a constant value throughout the system , or they are allowed to vary independently , leading to a spatially - nonuniform @xmath181 . while the former case corresponds to the helical states considered in sec . [
sec : helical state ] , in the present section we focus on the latter possibility . substituting eq .
( [ phi - grads ] ) into eq .
( [ f_gl ] ) , we obtain : @xmath182 . here
the ellipsis denote the terms which depend only on @xmath183 and @xmath184 , while the contributions containing the relative phase have the form @xmath185 where @xmath186 and @xmath187 according to eq .
( [ coeff - estimates ] ) , we have the following estimate : @xmath188 . without loss of generality
we assume that @xmath189 .
, in the two - band helical state.,width=226 ] for concreteness we consider only the case @xmath190 ( interband attraction ) , which means that if the relative phase is constant then it is equal to zero .
the results for @xmath191 are obtained by shifting the whole phase texture by @xmath192 . for the free energy difference , @xmath193 , between the state with a nonuniform @xmath194 and the state with @xmath195
we obtain : @xmath196 where @xmath197 .
variational minimization of the last expression yields the following nonlinear differential equation for the relative phase : @xmath198 in addition to the uniform solution @xmath195 , this equation also has various nonuniform ones corresponding to phase solitons or soliton lattices .
for example , a single soliton connects @xmath195 at @xmath199 and @xmath200 at @xmath201 and has the following explicit form : @xmath202 $ ] , where @xmath203 is the soliton width . while the last term in eq .
( [ delta f ] ) is a full derivative and does not contribute to the equation of motion , it does affect the free energy .
in fact it is easy to see that this term is responsible for a phase transition in the system , which is controlled by the external field .
qualitatively , the parameter @xmath204 , which is proportional to the magnetic field , provides a bias favoring a nonzero average gradient of the relative phase . when this bias becomes large enough to overcome the energy cost of creating the solitons , the latter will be spontaneously formed in the system . denoting the single - soliton energy by @xmath205 and assuming that the density of solitons @xmath206 is low , i.e. the spacing between them is much greater than @xmath207 , one can expand the difference between the total free energies of the system with and without solitons in powers of @xmath206 : @xmath208 where @xmath209 is the length of the system in the @xmath173 direction .
the neglected terms , denoted by the ellipsis , take into account the interaction between the solitons .
one can see that at @xmath210 , where @xmath211 the leading term in eq .
( [ delta - f - n_s ] ) becomes negative , resulting in the proliferation of phase solitons .
the transition between the state with a uniform relative phase and the soliton lattice state is mathematically similar to the commensurate - incommensurate transition of noble gas atoms adsorbed on a periodic substrate.@xcite banishing the technical details of the solution to appendix [ sec : sl - transition ] , here we present only the results .
the critical magnetic field above which the phase soliton lattice is formed is given by @xmath212 see eqs .
( [ h_s - v_0 ] ) and ( [ h - def ] ) .
note that the critical field of the superconducting transition in 2d noncentrosymmetric systems diverges at @xmath213 , see refs . and ,
therefore , the soliton instability of the superconducting state is always realized at strong enough fields .
according to eq .
( [ l near h_s ] ) , the spacing between solitons grows logarithmically at @xmath214 : @xmath215 the relative phase of the order parameter components in the soliton lattice at @xmath216 is shown in fig .
[ fig : theta ] .
it follows from eq .
( [ phi - grads ] ) that at low fields , @xmath217 , when the relative phase is uniform , the order parameters in the two bands are still nonuniform and given by @xmath218 .
thus the two - band helical state studied in sec .
[ sec : helical state ] is recovered .
, between the two bands , at high magnetic fields , @xmath216 ( @xmath219 is the soliton lattice period).,width=226 ]
we have developed a general theory of the helical instability in two - dimensional noncentrosymmetric superconductors , taking into account the two - component nature of the order parameter in these systems .
we have found that the paramagnetic pair breaking is weakened in the presence of the helical modulation .
the quasiparticle dos in the helical state is significantly different from that in a uniform state , showing field - dependent gap edges and logarithmic singularities , which could be probed in tunneling expreriments .
we have also found a novel type of field - induced nonuniform superconducting state , namely the lattice of phase solitons . unlike the previously studied nonuniform states in noncentrosymmetric superconductors , the phase solitons appear only in the two - component model .
in contrast to the phase solitons in centrosymmetric two - band superconductors , which are difficult to create , the soliton instability predicted in this paper is always present at sufficiently strong magnetic fields . the transition into the soliton state takes place when the bias provided by the lifshitz invariants in the gl free energy , which are unique to noncentrosymmetric systems , overcomes the energy cost of creating a soliton .
this phase transition should show up as a feature in the high - field low - temperature portion of the phase diagram .
one can also expect that the spatial inhomogeneity of the order parameters will result in a qualitative modification of the quasiparticle spectrum , similar to the soliton bound states in centrosymmetric two - band superconductors,@xcite which could be studied by tunneling .
this and other issues , such as the fate of the soliton state at finite temperatures or in the presence of the orbital effects of magnetic field , will be studied elsewhere .
this work was supported by a discovery grant from the natural sciences and engineering research council of canada .
the bdg equations , which determine the quasiparticle spectrum in an arbitrary nonuniform superconducting state , can be derived using the standard mean - field approach of the bcs theory , with some modifications pertinent to noncentrosymmetric superconductors .
the starting point is the hamiltonian @xmath220 , where @xmath221 is the noninteracting part , with @xmath141 being the band dispersion function deformed by the magnetic field , see eq .
( [ h_0 ] ) , and @xmath222 is the pairing interaction given by eq .
( [ h int reduced ] ) .
decoupling the latter in the mean - field approximation , we obtain : @xmath223\nonumber\\ -\frac{1}{2{\cal v}}\sum_{\bk\bk'\bq}\sum_{\lambda\lambda'}\delta^*_\lambda(\bk,\bq)v^{-1}_{\lambda\lambda'}(\bk,\bk')\delta_{\lambda'}(\bk',\bq).\end{aligned}\ ] ] here @xmath64 is the gap function in the @xmath65th band , which satisfies the self - consistency equation @xmath224 and @xmath225 in the second term in eq .
( [ h_int - mf ] ) should be understood as the inverse matrix both in @xmath15- and @xmath65-spaces . separating the phase factors @xmath71 introduced in sec .
[ sec : two - band picture ] , we have @xmath226 and , therefore , @xmath227 , where the @xmath15-dependence of @xmath228 is determined by the basis functions of an irreducible representation of the point group . assuming an isotropic pairing , which corresponds to the unit representation , we have @xmath229 and @xmath230 two complex functions @xmath69 and
@xmath70 comprise the order parameter of our noncentrosymmetric superconductor . introducing the nambu operators in each band , @xmath231 ,
one can write the fermionic part of the mean - field hamiltonian in the following form : @xmath232 where @xmath233 are the matrix elements of the bdg hamiltonian in momentum representation . thus the spectrum of the bogoliubov quasiparticles can be found independently in each band .
the phase factors in the off - diagonal elements of the hamiltonian ( [ h - bdg ] ) can be removed by a unitary transformation : @xmath234 where @xmath235 the matrix elements of the transformed hamiltonian in momentum space have the following form : @xmath236 we see that the phase factors @xmath71 do not affect the quasiparticle spectrum in a nonuniform superconducting state ( as long as there is no impurities or external fields in the diagonal elements of the bdg hamiltonian ) . in a uniform state , the order parameter has the form @xmath237 , and the diagonalization of eq .
( [ h - bdg - transformed ] ) yields @xmath238 for the energy of an excitation with wavevector @xmath15 .
it is straightforward to check that eq .
( [ theta - eq ] ) has the following integral of motion : @xmath239 this admits a simple mechanical analogy : interpreting @xmath181 as a coordinate and @xmath173 as a time variable , @xmath240 has the meaning of the total energy of a pendulum of mass @xmath241 in the potential @xmath242 .
if @xmath243 , then the pendulum oscillates near one of the minima of the potential , which corresponds to a periodic modulation of the relative phase in real space .
the case @xmath244 corresponds to the pendulum completing just one full rotation from @xmath245 to @xmath8 , or to a single phase soliton connecting @xmath195 at @xmath199 and @xmath200 at @xmath201 . if @xmath246 then the pendulum has enough energy to complete an infinite number of full rotations , which corresponds to a soliton lattice .
it follows from eq .
( [ int - of - motion ] ) that @xmath247 which implicitly determines @xmath181 as a function of @xmath173 . from the last expression
we can immediately obtain the soliton lattice period as a function of @xmath240 : @xmath248 focusing on the vicinity of the transition at @xmath244 we have @xmath249 with logarithmic accuracy at @xmath250 .
the next step is to relate @xmath251 and @xmath219 to the magnetic field @xmath204 . using eq .
( [ delta f ] ) and the fact that the phase winding per soliton is equal to @xmath8 , we obtain the following expression for the free energy density : @xmath252 where @xmath253\ ] ] is the free energy per one cell of the soliton lattice .
since the lattice period diverges at the transition , we obtain from eq .
( [ delta f over l ] ) the following expression for the critical field : @xmath254 . introducing the single - soliton energy @xmath255
, we recover eq .
( [ h_s ] ) . in order to calculate @xmath256 and @xmath205
, we use the integral of motion , eq .
( [ int - of - motion ] ) , and obtain : @xmath257\nonumber\\ & = & \int_0^{2\pi}d\theta\sqrt{2{\cal e}+2v_0(1-\cos\theta)}-{\cal e}\ell.\end{aligned}\ ] ] the second line follows after changing the variable , @xmath258 , in the first one and using the expression for @xmath259 from eq .
( [ x vs theta ] ) .
putting @xmath244 , we have @xmath260 .
the lattice period is determined by minimizing the free energy , eq .
( [ delta f over l ] ) , with respect to @xmath219 .
in this way , we obtain : @xmath261 it follows from eqs .
( [ epsilon_1 ] ) and ( [ sl - period - gen ] ) that @xmath262 . substituting this and eq .
( [ epsilon_1 ] ) into eq .
( [ epsilon - derivative ] ) , we have @xmath263 this last equation implicitly determines @xmath251 and , therefore , the lattice period , as functions of the field .
the critical field of the soliton transition is given by @xmath264 m. a. tanatar , j. paglione , s. nakatsuji , d. g. hawthorn , e. boaknin , r. w. hill , f. ronning , m. sutherland , l. taillefer , c. petrovic , p. c. canfield , and z. fisk , phys .
lett . * 95 * , 067002 ( 2005 ) .
d. r. tilley , proc .
. soc . * 84 * , 573 ( 1964 ) ; m. e. zhitomirsky and v .- h .
dao , phys .
b * 69 * , 054508 ( 2004 ) ; m. silaev and e. babaev , phys . rev .
b * 85 * , 134514 ( 2012 ) ; e. babaev and m. silaev , phys . rev .
b * 86 * , 016501 ( 2012 ) .
the zeeman interaction is not diagonal in the helicity basis , leading to the interband transitions caused by the magnetic field .
the effect of these transitions on the helicity band quasiparticles is quadratic in @xmath108 and is neglected if the zeeman energy is much smaller than the so band splitting . | we show how the two - band nature of superconductivity in noncentrosymmetric compounds leads to a variety of novel nonuniform superconducting states induced by a magnetic field . at low fields ,
a two - band helical state is realized , with a distinctly non - bcs quasiparticle spectrum . at high fields ,
the superconducting state becomes unstable towards the formation of a lattice of topological phase solitons . |
minority games ( challet and zhang , 1998 ) , and more recently evolutionary minority games ( emg ) ( johnson et .
al . , 1999a ; johnson et .
al . , 2000 ; de cara et .
al . , 2000 ; johnson et .
al . , 2003 ; hod and nakar , 2002 ; hod , 2003 ; sysi - aho et .
al . , 2003 ;
johnson et .
al . , 1999b ; lo et .
al . , 2000 ) , have received widespread attention in recent years as a useful model to describe competition for highly limited resources in complex systems , especially in economics .
these games are essentially based on a minority rule ( challet and zhang , 1998 ) according to which @xmath0 agents compete repeatedly for some resources by choosing between two options a or b. each agent makes its choice , and those agents belonging to the less ( most ) frequently chosen option are considered the winners ( losers ) , so they are rewarded ( fined ) .
so , the idea behind this game is that the agents must always try to be in the minority : few individuals choosing the same option as yourself means less competitors , and so it should be easier to obtain the resource .
the decisions taken by the agents are chosen according to a pool of strategies available , and these strategies are based on the @xmath1 previous outcomes in the game , as that information is assumed to be accessible to all of the agents . to give a simple example , a specific strategy in a minority game with @xmath2 has the form @xmath3
this means that if the two previous winning options in the game were ( a , a ) , an agent following strategy @xmath4 will choose option a the next time ; if the last winning options were ( a , b ) , that agent will choose b , and so on . at the beginning of the game several strategies are assigned to each agent , and the agent tends to choose from among them the strategy that gave better results in the past ; however , many different versions of the minority game exist , where the rules that determine the strategies chosen by the agents are different . here , we will skip the minor details on the mechanisms of the minority game , since that is outside the scope of the current work ; an exhaustive compilation of works on minority games can be found in http://www.unifr.ch/econophysics . in the evolutionary version ( emg ) of the game ( johnson et .
al . , 1999a )
, all the agents are assigned the same strategies but they can _
i ) _ follow that given strategy with probability @xmath5 or _
ii ) _ do exactly the opposite with probability @xmath6 , where @xmath5 is different for each agent ( the subindex @xmath7 denotes the @xmath7th agent ) .
those agents performing the worst ( losing many times ) are forced to change their value of @xmath5 ; so , in the emg there is an implicit learning process based on _ trial and error .
_ as a consequence , the system tends towards an optimal distribution of @xmath5 values for which the number of winners is as close to @xmath8 as possible ( note that , by definition , in a minority game the number of winners can not be higher than @xmath8 ) . as reported in ( johnson et .
al . , 1999a )
, the most striking result arising from the emg is the natural emergence of segregated behaviour : those agents that behave in an extreme way ( @xmath9 and @xmath10 ) perform better than those with intermediate behaviour , so that the individuals tend to segregate into two groups : those who always follow the given strategy and those who never follow the strategy . from the point of view of complex systems , it has been claimed that this result may help to understand some levels of organization such as crowding ( johnson et .
al . , 2000 ; cont and bouchaud , 2000 ) and cooperation ( de cara , 2000 ) , which are common in many social and biological systems .
specifically , within the context of the emg some authors have coined the term _ unintentional _ or _ indirect cooperation _ to illustrate the behaviour observed ( quan et .
al . , 2003 ;
hod and nakar , 2004)_. _ this concept refers to the fact that in the emg many agents tend to behave similarly ( either @xmath9 or @xmath10 ) , but not consciously , but rather because the global winning probability is higher that way .
this is different from other games ( for example , the well - known prisoner s dilemma ) where cooperation is a conscious option given to the agents ( nowak et .
al . , 2004 ; nowak 2006 ) .
in general , minority games are helpful to describe multi - agent systems where each agent ( individual ) is able to analyze the history of the system ( i.e. , the success of the different strategies used before ) in order to make its next decision . for this reason
they have been especially designed and used to explain the complex dynamics of some financial markets ( johnson et .
al . , 2003 ) , albeit some authors have stated that similar ideas could also hold within an ecological context ; probably the best example being foraging behaviour ( hod and nakar , 2002 ; hod , 2003 ) .
however , as far as we know very few real efforts have been made to extend minority games to ecological situations . in ( tella et .
al . , 2001 ) the authors presented a model , inspired by the rules of the minority game , to explore the _ colonial versus solitary _ behaviour in birds as a function of predation pressure , and some discussions on the connection between minority games and ecological evolution were provided in ( de cara et .
al . , 2000
al . , 2003 ) .
the apparent lack of interest by ecologists in these games is probably due to the fact that the most interesting and dramatic situations concerning decision - making in animals are not well described by such concepts as _ trial and error _ and _ pool of strategies _ involved in minority games .
instead , in ecology most of the interest lies in understanding those situations where individuals perform just one or a few critical decisions throughout their whole life ( concerning , for example , timing in reproduction or choice of habitat ) ; these decisions have been called fitness - critical actions in a very recent work by heesch and little ( 2006 ) .
intuitively , decision - making in these fitness - critical actions follows quite simple mechanisms ( compared to the complex rules of minority games ) : the individuals need to use their skills or their experience to predict the _ a priori _ best option . by _ a priori _ best option we mean that option which would be the winning one in the case where half of the agents choose a and the other half choose b. in the basic minority game described above we have considered that the agents choose between two identical options a or b , so there is no _ a priori _ best option .
however , it is easy ( and more realistic ) to consider a game where a and b are intrinsically different .
for example , in the case of habitat selection , individuals usually need to choose between different options with different habitat qualities .
some individuals may be able , from past experience , to know in advance which the best choice is e.g. that where the availability of food is higher .
but if all the individuals are able to do this , then all of them will choose the same option and the availability of food will decrease there ; in that case the _ a priori _ best option is not necessarily the winning option .
those individuals that are not able to determine what the _ a priori _ best option is will probably behave _ randomly _ or _ persistently _ ( always choosing the same option ) .
the role of evolution and natural selection is thus expected to be crucial in these processes , as stated in ( heesch and little , 2006 ) .
we note that these decision - making mechanisms are also common in human behaviour . for instance
, drivers who have to choose between two alternative routes in order to avoid traffic jams do not analyze every past experience and make a decision according to a pool of strategies ( contrary to what is suggested by some authors ( hod , 2003 ) ) , but mainly use simpler strategies like _ persistent _ behaviour ( they always choose the same route because they do not like to take risks ) or they may simply listen to the traffic news to find out what the _ a priori _ best route is . according to these arguments ,
some essential elements which are absent in the emg must be considered in order to get a realistic implementation of minority games in ecology .
so , the aim of this work is to present a new game where competition for resources is also introduced by means of a minority rule , but the dynamics and strategies followed by the agents aim to capture the dynamics of some ecological systems . in
what follows , we will refer to this new model as the evolutionary learning game ( elg ) .
we now introduce a specific problem that has attracted the interest of ecologists in recent years ( van noordwijk et .
al . , 1995 ;
visser et .
al . , 1998 ;
grieco et .
al . , 2002 ;
visser et .
al . , 2004 ;
gienapp and visser , 2006 ) and has strongly motivated our approach . in many species of bird , individuals must face the problem of choosing the correct time for egg laying .
this choice becomes dramatic if the availability of food is restricted to a very short period of time .
so , for survival in breeding , the correct timing of egg laying is necessary , so that the feeding period matches the food peak .
this process has been studied in recent decades for some species , such as great tits ( _ parus major _ ) and blue tits ( _ parus caeruleus _ ) , whose main prey ( caterpillar ) is only available for two or three weeks in the late spring ( visser et .
al . , 2004 ) .
at the moment of egg laying ( approximately one month before ) ,
the birds do not know when the food peak will happen .
the problem is partially overcome by the way many of these birds develop with age the ability to follow some cues ( based on climate and other environmental parameters ) to predict the right time for laying ( van noordwijk et .
al . , 1995 ;
grieco et .
al . , 2002 ; gienapp and visser , 2006 ) . in general , this capacity of an individual to adapt its behaviour to the environmental conditions is known as phenotypic plasticity , and is usually a heritable trait .
specifically , it has been demonstrated ( nussey et .
al . , 2005 ) that plasticity in egg laying for birds is heritable .
the effects of global climate change , however , have put many biological species to the test ( parmesan , 2006 ) . as a consequence of warmer springs ,
caterpillars have advanced their hatching date in many habitats ( visser et .
al . , 1998 ;
visser et .
al . , 2004 ) , so those birds with a higher plasticity in laying are expected to adapt better to the new situation .
according to the observational data , some bird populations have become adapted , but in some other cases a very weak response to the new situation has been observed ( visser et .
al . , 2004 ;
gienapp and visser , 2006 ) . in the latter case
, the mismatching between the feeding period and the food peak will probably lead to a decline in the number of individuals ( both et .
al . , 2006 ) or the habitat fitness ( visser , 2007 ) .
although different explanations have been provided , there is no clear understanding of why different populations show different responses to the changing conditions ( gienapp and visser , 2006 ) .
as we discuss below , our model provides some arguments that support the idea that resource constraints from each specific habitat may be responsible for these differences .
we need to introduce two basic ideas that are missing from the original formulation of the emg , in order to reach a more realistic description of ecological systems : _ i ) _ first , reproduction and death processes must play a fundamental role in the dynamics of the system . in the emg
the agents continue to play indefinitely , but in ecology the consequences of choosing a wrong option can obviously be dramatic . if we want to explore the dynamics of systems over representative time scales , it is necessary to assume that the individuals may disappear ( die ) and/or be replaced by new individuals ( newborns ) with some probability .
moreover , the outcome obtained from any decision taken by the agents must affect in some way their reproduction / survival probabilities . _
ii ) _ secondly , in the emg all of the agents have access to the same information , and so all of them may use the same strategy . in ecology , however , as long as an individual grows up it gains experience and , in consequence , it is expected to choose better options .
so , an individual learning process must be considered somehow .
in fact , the situation where the strategies chosen by the agents are based on their individual histories has already been studied for the emg ( see ( de cara , 2000 ) and the references therein ) , but here we will explore the concept of learning from a different perspective . in our model , each of the @xmath0 agents competing in the game must repeatedly choose between options a or b. whether or not the decision taken by the agent is the good one will be determined by a minority rule with an arbitrary cutoff , as defined in ( johnson et .
al . , 1999b ) .
this means that we assign a resource capacity @xmath11 to one of the two options ( we consider @xmath12 without loss of generality ) and a resource capacity @xmath13 to the other option .
if the real number of agents choosing the first option is below @xmath11 , then the resources _ per capita _ in that option are higher than in the other one , so those agents are the winners and the agents choosing the other option are the losers .
if the number of agents choosing the first option is above @xmath11 , then the contrary arguments hold .
we will consider that the option with capacity @xmath11 is not always the same , but is chosen randomly every time step in order to incorporate the effects of a fluctuating environment , so sometimes option a will be the _ a priori _ best option ( that with a higher capacity ) and sometimes not . at each time step , the winners are rewarded with the possibility of reproductive success .
every winner is given the possibility of producing a _ newborn _ agent with a probability @xmath14 .
the newborn will replace one of the agents in the game ( to keep @xmath0 constant ) chosen randomly , so we assume that all of the agents are equally likely to die .
the agents choose option a or b according to the following rules .
younger agents act _ persistently _ : they make their first choice randomly and , after that , they continue to choose the same option .
however , after each time step all the _ persistent _ agents are given the possibility of _ learning _ with probability @xmath15 ( here @xmath16 stands for phenotypic plasticity and the subindex @xmath7 denotes the @xmath7th agent ) .
if they _ learn _ , it means that they give up _ persistent _ behaviour ; from then on , they always choose the option with a higher capacity , so we will say that they become _ wise _ agents .
phenotypic plasticity in our model is thus considered equivalent to a learning capacity .
this capacity can be inherited as follows : when a _ newborn _ appears , its characteristic probability @xmath15 is chosen randomly from an interval of width @xmath17 centred on the value of @xmath18 from its _ father _ , with reflecting boundary conditions at @xmath19 and @xmath20 .
so , we introduce an evolutionary dynamics for the probabilities @xmath15 into the model in a similar way as in the original formulation of the emg ( johnson et .
al . , 1999a ) .
but note that in this case the meaning of the width @xmath17 is extremely important for the dynamics of the system , as it measures the heritability of @xmath15 , so the best strategies @xmath21 will be transmitted to the _ newborns _ only if @xmath17 is not too high .
these are all the rules for our elg .
all of the agents will become _ wise _ sooner or later unless they die first , but if there are too many _ wise _ agents then the _ a priori _ best option will be crowded and will probably be the wrong one .
a complex dynamic thus emerges where learning as fast as possible is not necessarily the best strategy , which may seem counterintuitive at first .
as some of the rules presented could be considered too simple or unrealistic from a biological point of view , we tried to implement many different models with increasingly complex rules in order to compare their performances : _ i ) _ we tried to introduce explicit reproduction and death algorithms in many different ways ( for example , by using exponential or logistic growth ) , so that the number of agents @xmath0 was allowed to change with time . _
ii ) _ we tried to replace the minority rule with some other competition rules , even rules that allowed all the agents to be winners ( or losers ) at the same time .
for instance , we considered two independent capacities @xmath22 and @xmath23 for the two possible options , so if the number of individuals choosing option a is above ( below ) @xmath22 those agents are considered losers ( winners ) .
_ iii ) _ we tried to reward and/or fine agents on their reproductive success and/or their probability of survival .
of course , it is not necessary for the reproductive success to be completely suppressed for the losers as in the simplified version we have described ; we could consider two reproductive rates @xmath24 and @xmath25 for winners and losers respectively , with @xmath26 .
_ iv ) _ we tried to consider that the switching from _ persistent _ to _ wise _ behaviour is not so radical , but the agents learn progressively according to a rate given by @xmath15 . after these and many other trials
, we have found that the qualitative behaviour exhibited by the elg ( which is shown in the following section ) is highly robust .
according to our results , it seems that there are only two elements which are strictly necessary in order to obtain that behaviour : _
i ) _ a learning process regulated by the probabilities @xmath15 and _ ii ) _ that the number of agents rewarded ( fined ) is proportional to the number of winners ( losers ) .
the version of the elg we have presented here is one of the simplest possible , and so it offers the advantage that some analytical treatment is possible , as we will show below .
the greatest interest of our model lies in the form of the distribution of phenotypic plasticities @xmath27 that is reached in the steady state . in figure 1 we summarize the behaviour of @xmath27 as a function of the three parameters of the model : @xmath11 , @xmath14 and @xmath17 .
all the results shown here were obtained by computing the form of @xmath27 for @xmath28 after 10000 time steps ( which is far enough to reach the steady state ) , and carried out an average of 25 different realizations . initially all the agents were considered _
newborns _ and the values of @xmath15 were assigned randomly ; anyway , we have checked that our results are independent of the initial conditions chosen .
the series of plots from _
1.a _ to _ 1.d _ shows how @xmath27 changes when the value of @xmath11 is modified . figures _ 1.a _ and _ 1.d
_ correspond to extreme situations that are clearly predictable . in the first case ,
when @xmath29 both options a and b have similar resource capacities
. therefore , performing as a _ wise _ agent does not represent an advantage , because the resource of choice ( i.e. the larger one ) reaches its smallest possible size leading to overcrowding among wise agents . as a consequence ,
learning is avoided and there is a tendency @xmath30 . in the regime
@xmath31 one of the two options is much better than the other one . in this situation , performing as a _ wise _ agent is a strong advantage , and so the tendency @xmath32 should be expected .
but , surprisingly , there is a wide range of intermediate values of @xmath11 where segregated ( obviously asymmetric ) behaviour is found .
this means that in intermediate situations the dynamics of the system tends to favour individuals which either learn as fast as possible or avoid learning as much as possible .
although segregated behaviour was also found for the emg , the situation reported here is clearly different . in the case of the emg ( johnson et .
al . , 1999a )
the segregated behaviour in the steady state was independent of the initial conditions and the values of the parameters introduced .
however , hod and nakar ( 2002 ) proved later that the model is extremely sensitive to the prize - to - fine ratio , so for some parameters one observes a sharp transition where self - segregation is destroyed . on the contrary
, we have not noticed such effects in our model , but the form of @xmath33 always changes smoothly for any region of parameters considered .
the other main difference between the emg and the elg is that the results obtained here show an asymmetric distribution of @xmath27 .
the reason for this is that in the elg the _ persistent _ and _ wise _ agents do not necessarily choose different options ( while in the emg @xmath9 and @xmath34 represent opposite behaviours ) .
this , together with the different backgrounds considered and some ideas discussed below , shows that the general dynamics of the emg and the elg are different , although there are some major similarities between both
. one can observe in the series _
1.e _ to _ 1.h _ the role of the reproduction probability @xmath14 on @xmath27 while keeping the other parameters constant . a high value of @xmath14 involves the appearance of many _ newborns _ and , according to the discussion above , a _ wise _ strategy will then perform better that in a situation with few _
newborns_. for this reason , the model shows a tendency @xmath35 for low @xmath14 and a tendency @xmath36 for high @xmath14 . in intermediate situations ,
a segregated distribution is found again .
finally , the role of @xmath17 is shown in plots from _
1.i _ to _ 1.l_. as discussed above , the value of @xmath17 determines the heritability of the phenotypic plasticity .
low values of @xmath17 represent a high level of heritability and so the best strategies persist , while a high value of @xmath17 means that best strategies are not well transmitted to breeding ; so , for the latter @xmath27 is expected to tend to be uniform .
the value of @xmath17 also indirectly affects the number of winners and losers ; if @xmath17 is too high the best strategies do not persist and then the average number of winners decreases .
for this reason it is difficult to predict the exact form of @xmath37 as a function of @xmath17 .
actually , the specific role of @xmath17 in the game is fairly complicated , so this point will be addressed in detail in a further study .
note that this is another important difference from the case of the emg , where the final distribution @xmath38 is almost independent of @xmath17 ( johnson et .
al . , 1999a ) .
we can give some analytical support to the results shown in figure 1 by means of a mean - field - like approach as proposed before for the emg ( lo _ et .
_ , 2000 ) . here
we keep as much as possible to the notation used there in order to facilitate understanding .
first of all , note that we know the option that the _ wise _ agents will choose .
we also know that every _ persistent _ agent made its first choice randomly , so we should expect on average for half of them to choose a and the other half to choose b. therefore , the whole problem is reduced to finding out how many _ persistents _ are in the game .
we denote @xmath39 as the probability of @xmath40 of the @xmath0 agents in the game being _
persistent_. similarly , we define @xmath41 as the probability of @xmath40 of the agents being persistent , given that the @xmath7th agent is the only one that has not made its choice yet .
then , the following relation holds : @xmath42 where @xmath43 is the probability of the @xmath7th agent being _
persistent _ , given that its phenotypic plasticity is @xmath15 . in the following , we will use the simplified notation @xmath44 .
on the other hand , the winning probability @xmath45 of an agent that has a plasticity @xmath15 can be written as @xmath46 where @xmath47 denotes the integer part of @xmath48 following now the same treatment as in ( lo et .
al . , 2000 )
, we get from ( [ 1 ] ) and ( [ 2 ] ) the expression @xmath49 as stated in ( lo et .
al . , 2000 )
, the two first terms correspond to the winning probability of a @xmath7th agent whose action does not modify the result of the game , once the other @xmath50 agents have made their choice .
hence , the third term is the essential one , as it measures the influence of the @xmath7th agent on the final result . for example , imagine that after the first @xmath50 agents have made their choice , one half of them decide to follow option a ( so one half choose b ) ; then , the last agent?s decision will determine which the winning option is ; the influence of this final decision on the probability @xmath45 is what the last term in ( [ 3 ] ) measures . in order to analyze this term , we need to find the explicit expression of @xmath51 as a function of @xmath15 .
this is easy to do , because after each time step the _
persistents _ have a probability @xmath15 of becoming _ wise _ agents and an average survival probability @xmath52 ( according to the rules of our elg , only the range @xmath53 holds here ) .
then , the probability of the @xmath7th agent being _
persistent _ is @xmath54 } { \sum_{i=0}^{\infty } s^{i}}=\frac{1-s}{1-s(1-pp_{k})}.\ ] ] now , we can give an explicit expression for the term @xmath55 from equation ( [ 3 ] ) .
it can be seen that @xmath55 has the same appearance throughout the whole range @xmath56 and for the proper range of survival probabilities @xmath53 .
it is always a negative convex function with a relative minimum at @xmath57 .
this means that the third term in ( [ 3 ] ) always tends to favour extreme values of @xmath15 , which facilitates the emergence of segregated behaviour ; so , this gives some justification to our numerical results .
we stress , however , that stationary analytical approaches such as the one used here have some limitations , as the emg and similar models never reach a true stationary distribution ( hod , 2002 ) . in fact
, we have found for the elg that the number of _ persistents _ oscillates periodically over time , in accordance with similar results found for the emg ( hod , 2002 ) . in that work ,
it was also argued that when the amplitude of these oscillations increases , we observe in the emg a transition from self - segregated behaviour to clustering ( where clustering is characterized by a single - peaked distribution of @xmath5 values around @xmath58 ) .
it is interesting to note that , on the contrary , in the elg the amplitude of the oscillations increases in the region where self - segregated behaviour is found ; so , the oscillations in our model seem to enforce self - segregation rather than destroying it .
this situation is shown in figure 2 , where the number of _ persistents _ in the steady regime is plotted as a function of time for different situations ; dotted , dash - dotted and solid lines correspond to the cases _ 1.a _ , _ 1.c _ and _ 1.d _ reported in figure 1 , respectively .
from figure 2 , it is also clear that the mean number of _ persistents _ in the game increases as @xmath11 decreases , in accordance with our discussion above .
we have presented a model that sets the ideas of minority games , which are considerably popular tools for describing competition for resources in economics , into an ecological context .
the model presented here shows how social segregation emerges from an evolutionary learning process ( determined by the distribution @xmath27 ) in a group of individuals competing for strongly limited resources .
note that if the learning process were not introduced to the model , then the dynamics of the system would be trivial .
neither persistent nor wise behaviour on its own is an efficient strategy ; evolutionary learning is the key ingredient here for finding efficient cooperation between both .
this idea , together with the robustness shown by our model ( many other implementations with more realistic rules led to similar qualitative results ) seems to suggest that our model could be of interest for understanding social organization in complex evolutionary systems .
the results obtained here show that in the situation described by our model intermediate learning strategies can not persist for long times ; the tendency is always towards improvement ( @xmath36 ) , suppression ( @xmath30 ) or the coexistence of both ( segregated distribution )
. it also contradicts the intuitive idea that those individuals with a higher learning capacity must always be favoured by selection .
this is a consequence of the strong competition process which is assumed in our elg and in minority games in general : sometimes the _ a priori _ worst option can be the best because many agents tend to choose the _ a priori _ best option and then competition for the latter is higher . finally
, we come back to the problem of egg laying in birds described before , which can now be addressed using the ideas about learning and phenotypic plasticity discussed here . in order to provide an analogy with our model
, we could imagine that option a means laying early and option b means laying later .
those individuals with a higher plasticity very quickly become able to follow some environmental cues in order to predict the right option a or b. however , if the penalty for choosing a wrong option is not high ( because there are some other food resources available , or there are some other environmental constraints on laying ... ) then selection will not favour individuals with higher plasticity . in that case , when the individuals must face sudden environmental changes their capacity to respond will be weak . in those habitats where
phenotypic plasticity is strongly rewarded ( i.e. , for @xmath11 small in our model ) , individuals are expected to follow efficient learning strategies , so they will be able to respond better to environmental changes .
therefore , the results of our model provide an evolutionary basis to the idea that environmental constraints from each specific habitat could explain why different bird populations are responding differently to climate - driven changes in the behaviour of their prey , as recent experiments have suggested ( gienapp and visser , 2006 ) . however , empirical evidence supporting our ideas about learning and phenotypic plasticity is still lacking .
we believe that it would be of major interest if experimentalists were to attempt to check the predictions made by our model in real ecological systems .
this research has been partially supported by the generalitat de catalunya through grant 2006-bp - a-10060 ( dc ) , by the project cgl 2007 - 60797 ( jell ) and by grants fis 2006 - 12296-c02 - 01 , sgr 2005 - 00087 ( vm ) .
both , c. , bouwhuis , s. , lessells , c.m . , visser , m.e .
climate change and population declines in a long - distance migratory bird .
nature 441 , 81 - 83 .
challet , d. , zhang , y.c . , 1998 .
emergence of cooperation and organization in an evolutionary game .
physica a 246 , 407 - 418 ; on the minority game : analytical and numerical studies . physica a 256 , 514 - 532 .
tella , j.a .
, de cara , m.a.r . , pla , o. , guinea f. , 2001 . a model for predation pressure in colonial birds . in sixth granada lectures in computational systems .
garrido , p.l .
, marro , j. aip conf .
proc . 574 .
aip , new york .
plot of the distribution @xmath27 in the steady state ( after 10000 time steps ) for @xmath28 averaged over 25 realizations .
the series _ a _ to _ d _ shows the effect of varying @xmath11 while keeping @xmath59 and @xmath60 . the series _
e _ to _ h _ shows the effect of varying @xmath14 while keeping @xmath61 and @xmath59 . the series _
i _ to _ l _ shows the effect of varying @xmath17 while keeping @xmath61 and @xmath62 .
number of _ persistents _ as a function of time once in the steady state ( the results plotted correspond to times @xmath63 and @xmath28 ) .
the dotted line corresponds to the parameter values in figure _ 1.a . _ ; the dashed - dotted line corresponds to parameter values in figure _ 1.c . _ ; the solid line corresponds to the parameter values in figure _ 1.d . _ | we present an agent - based model inspired by the evolutionary minority game ( emg ) , albeit strongly adapted to the case of competition for limited resources in ecology .
the agents in this game become able , after some time , to predict the _ a priori _ best option as a result of an evolution - driven learning process .
we show that a self - segregated social structure can emerge from this process , i.e. , extreme learning strategies are always favoured while intermediate learning strategies tend to die out . this result may contribute to understanding some levels of organization and cooperative behaviour in ecological and social systems .
we use the ideas and results reported here to discuss an issue of current interest in ecology : the mistimings in egg laying observed for some species of bird as a consequence of their slower rate of adaptation to climate change in comparison with that shown by their prey .
our model supports the hypothesis that habitat - specific constraints could explain why different populations are adapting differently to this situation , in agreement with recent experiments . , and evolutionary learning , climate change , limited resources , predator - prey |
the ability to carry out high angular resolution observations of several types of masers has increased tremendously due to the availability of interferometers at several different frequencies .
for example , 36 ghz methanol maser observations are now routine with the karl g. jansky very large array ( vla ) , e.g. , @xcite and @xcite .
since masers allow for the observation of magnetic fields at high angular resolution via the zeeman effect , there has often been speculation on whether they trace the large - scale magnetic field or only the field within the small - scale environments in which such masers form ( @xcite ; @xcite ; @xcite ) .
since the observation of magnetic fields is challenging in general ( e.g. , @xcite ) , the ability to measure the large - scale magnetic field with masers would prove to be invaluable , especially because magnetic fields play such an important role in the process of star formation ( e.g. , @xcite ; @xcite ) . at a distance of 450
pc ( @xcite ) , the orion molecular cloud 2 ( omc-2 ) is considered one of the nearest intermediate - mass star forming regions ( e.g. , @xcite ) . this region is located at an angular distance of 12@xmath0 northeast of the trapezium ob cluster , and along with omc-3 , it is part of a single long filament in orion ( @xcite ) . in 2011 , we reported the discovery of the zeeman effect in the 44 ghz class i ch@xmath1oh maser line toward omc-2 ( @xcite ) . in this paper , we report follow - up observations on the zeeman effect of this methanol maser line .
these observations provide the opportunity to study whether the intensity or the magnetic field has changed since our first observations ( @xcite ) , and the implications this would have for measuring magnetic fields with 44 ghz ch@xmath1oh masers .
the observations and data reduction are presented in [ sodr ] . in
[ sanal ] , we present the analysis to derive the line of sight magnetic field value through the zeeman effect . finally , we present the results and the discussion in [ sr ] .
we observed the @xmath2 ch@xmath1oh ( methanol ) maser emission line at 44 ghz with the vla on 2011 september 25 .
the observations were carried out in dual polarization with a 1 mhz bandwidth and 256 spectral channels . at the time of the observations , the array was in the later stages of the reconfiguration from the most extended , a - configuration ( maximum baseline = 36 km ) , to the most compact , d - configuration ( maximum baseline = 1 km ) . on the day of the observations , all antennas except for one
were already placed in d - configuration .
the sole unmoved antenna that resulted in baselines @xmath3 km was edited out during the data reduction process .
the total observing time was 4 hours .
the calibrator source j0542@xmath44951 ( 3c147 ) was used to derive the antenna - based amplitude gain factors to calibrate the flux density scale . the uncertainty in the flux density calibration at the observed frequency , accounting for various observational parameters ( e.g. , weather , reference pointing , elevation effects ) , is expected to be up to 10% .
table [ top ] summarizes the parameters of these observations .
data reduction , including calibration , deconvolution and imaging , was carried out using the astronomical image processing system ( @xcite ) .
the flux density scale of the target source was set by applying the amplitude gain calibration solutions of 3c147 .
after doppler correcting the data of omc-2 , the spectral channel with the brightest maser emission signal was split off , and self - calibrated in both phase and amplitude and imaged in a succession of iterative cycles .
the final self - calibration solutions were applied to the full spectral - line data set of the target source .
hanning - smoothed stokes @xmath5=(rcp@xmath4lcp)/2 and @xmath6=(rcp@xmath7lcp)/2 image cubes were then constructed with a synthesized beamwidth of @xmath8 , and an effective velocity resolution of 0.053 km @xmath9 .
the resulting image cubes were further processed using the miriad data reduction package to derive the magnetic field value .
we note that these observations are one of the first successful zeeman effect observations and detection with the wide - band digital architecture ( widar ) correlator . in 2010
, the vla / widar commissioning team discovered a small but significant unanticipated spectral response , or `` spectral splatter '' effect while observing strong spectral lines at high frequency resolutions .
zeeman effect studies , which in this case rely on the difference of two strong signals , were particularly vulnerable .
the `` spectral splatter '' problem resulted in corrupt stokes @xmath6 spectra , and sometimes in false zeeman effect detections , e.g. , in the strongest 36 ghz methanol maser line in the star forming region dr21w @xcite .
this rendered the s - curve profiles in the stokes@xmath6 data very unreliable @xcite .
while the various zeeman effect commissioning tests were concluded in april 2012 , these tests showed that spectral line data acquired since august 2011 with widar would be suitable for zeeman effect and other spectral - line polarization studies .
the omc-2 observations reported in this paper were carried out in light of the various changes that have been implemented in the vla / widar system to address the spectral splatter problem .
the data have been checked carefully to verify that the detected signal in stokes v is astronomical in nature and not instrumental .
following the arguments presented in 3 of @xcite for maser lines with widths @xmath10 much greater than the zeeman splitting @xmath11 , we measured the zeeman effect by fitting the stokes @xmath6 spectra in the least - squares sense to the equation @xmath12 ( @xcite ; @xcite ) , where @xmath13 and @xmath14 are the fit parameters . here
, @xmath13 is usually the result of small calibration errors in rcp versus lcp , and it is on the order of @xmath15 or less in our observations . the magnetic field value is obtained through the fit parameter @xmath14 , where @xmath16cos@xmath17 . here , @xmath18 is the zeeman splitting factor ( hz @xmath19 ) , @xmath20 is the magnetic field , and @xmath17 is the angle of the magnetic field to the line of sight ( @xcite ) . the value of the zeeman splitting factor @xmath18 for ch@xmath1oh masers is not known but is likely to be very small because ch@xmath1oh , like h@xmath21o , is a non - paramagnetic molecule .
therefore , we will report only values of @xmath22 .
we note that while eq .
( [ evlco ] ) is true only for thermal lines , numerical solutions of radiative transfer equations ( e.g. , @xcite ) have shown that it also results in reasonable values for the magnetic fields in masers .
figure [ fivd ] ( left panel ) shows the stokes @xmath5 and @xmath6 spectra of the 44 ghz ch@xmath1oh maser toward omc-2 from our current observations . as described in [ sanal ] , we determined magnetic fields by fitting the stokes @xmath6 spectra in the least - squares sense using equation ( [ evlco ] ) . the fit parameter @xmath23@xmath24 ( see eq .
[ evlco ] ) obtained by this procedure is @xmath25 hz . by convention , a positive value for @xmath24 indicates a field pointing away from the observer .
also shown in figure [ fivd ] ( right panel ) is the stokes @xmath5 and @xmath6 spectra of the same 44 ghz ch@xmath1oh maser from @xcite .
we conclude that this is the same maser as in the follow - up observations because the masers in both epochs have the same velocity ( @xmath26 ) , and are at the same position relative to other , weaker masers in the field .
@xcite found for this maser that @xmath27 hz . compared to @xcite , the intensity of the maser is a factor of 1.5 higher in the current observations , while the values of @xmath18@xmath24 are the same , within the errors . in order to obtain the value of @xmath24 from @xmath14 , we need to know the zeeman splitting factor @xmath18 .
an empirical value for the land @xmath28-factor can be obtained by extrapolation from lab measurements of several methanol maser transitions near 25 ghz made by @xcite .
using this @xmath28-factor would result in a value of @xmath18 for the 44 ghz methanol maser line of 0.1 hz mg@xmath29 , implying a line - of - sight magnetic field value of order 0.2 g. such a value is an order of magnitude higher than those measured by water masers in star forming regions ( e.g. , @xcite ) .
moreover , the empirical scaling relation between the hydrogen molecule number density and magnetic field , @xmath30 @xcite , would result in a number density of @xmath31 @xmath32 , a value that is several orders of magnitude higher than that thought to be suitable for the 44 ghz methanol maser line ( e.g. , @xcite ) .
@xcite presented a detailed discussion on the uncertainty of the @xmath28-factor derived by @xcite while reporting the magnetic field values of various 6.7 ghz methanol masers .
these authors concluded that the @xmath28-factor may be uncertain by an order of magnitude . due to this , we have chosen to leave our results in terms of @xmath14 rather than @xmath24 .
our observations and the above noted considerations motivate the immediate need for an experimental measurement of the land @xmath28-factor in the 44 ghz methanol maser line . with the detection of consistent stokes v signals in the 44 ghz methanol maser line in omc-2 through two observing epochs ,
it is timely to assess whether the detected signal can be attributed to a magnetic field , and therefore being due to a real zeeman effect . as described by @xcite , a potential effect that could mimic a regular zeeman effect
is caused by a rotation of the axis of symmetry for the molecular quantum states , and can occur when the stimulated emission rate of the maser @xmath33 becomes greater than the zeeman splitting @xmath34 as the brightness of the maser increases while it becomes more saturated .
the detected splitting in our two epoch observations ( this work and @xcite ) is @xmath35 s@xmath29 .
the value of @xmath33 can be estimated using @xmath36 ( e.g. , @xcite ) , where @xmath37 is the einstein coefficient for the 44 ghz methanol maser transition , which is @xmath38 s@xmath29 @xcite , @xmath39 and @xmath40 are the boltzmann and planck constants , respectively , @xmath41 is the frequency of the maser transition , @xmath42 is the maser brightness temperature , and @xmath43 is the maser beaming solid angle .
higher angular resolution observations of the brightest 44 ghz methanol maser in omc-2 resulted in @xmath44 k and @xmath45 @xcite . from these parameters
, we derive @xmath46 s@xmath29 .
this is several orders of magnitude lower than @xmath34 , and suggests that the detected stokes v signal is unlikely to be due to this non - zeeman effect .
furthermore , this particular effect would introduce maser - intensity - dependent circular polarization because the molecules interact more strongly with the radiation field than with the magnetic field @xcite . since we observe no change in splitting over two epochs ( @xcite and this work ) in spite of the maser intensity having increased by a factor of 1.5 ,
this is additional proof that the detected signal is due to the zeeman effect in a magnetic field . as noted above
, the intensity of the maser is a factor of 1.5 higher in the current observations compared to @xcite .
very few observations exist of the variability of 44 ghz class i methanol masers , or indeed of any kind of class i methanol maser .
such masers are believed to be excited in outflows in star forming regions .
@xcite observed some variability in a number of sources , but due to differences in spatial and spectral resolution and coverage , were able to confirm variability in only two 44 ghz methanol maser sources .
more recently , @xcite have reported short term variability on the scales of days or possibly even hours , in 44 ghz class i masers toward the dr21 region . even though the 44 ghz methanol maser intensity in omc-2 has increased
, the measured magnetic field ( as given by @xmath22 ) has stayed the same , within the errors .
this leads us to speculate that the mechanism responsible for the change in maser intensity may be active on smaller scales than the shock compression that sets up the magnetic field . to elaborate , the post - shock magnetic field that we are detecting via the zeeman effect was likely amplified above its pre - shock value in proportion to the density of the post- and pre - shock regions ( e.g. , equation ( 2 ) of @xcite ) .
this does not appear to have changed between the two epochs , as @xmath47 has remained the same .
meanwhile , variability in collisionally pumped masers such as the 44 ghz class i methanol maser likely results from bulk gas flows and turbulence moving parts of the maser column in and out of velocity coherence along the line of sight ( @xcite ) . evidently , a larger column of gas has been moved into the line of sight to increase the maser intensity , but the compression factor between post- and pre - shock regions has remained the same , thereby keeping the magnetic field the same .
it would be very interesting , therefore , to observe other regions and compare maser intensities and magnetic fields between epochs in order to determine how frequently this happens .
if the effect were widespread , it might mean that we are detecting the large - scale magnetic field in the shocked region that is not affected by any changes on smaller scales that cause the intensity of the maser to change .
we have presented a second epoch of observations of the 44 ghz class i methanol maser line toward the star forming region omc-2 with the vla .
we detected a magnetic field value of @xmath18@xmath24 @xmath48 hz . the observations constitute one of the first successful zeeman effect detections with the new widar correlator of the vla . comparing to the result of our earlier epoch of data for this region @xcite , we find that the intensity of the maser has changed by a factor of 1.5 , but the magnetic field value has stayed the same , within the errors . we speculate that we may be detecting the large - scale magnetic field in the post - shock region that is unaffected by bulk motions or turbulence on smaller scales that may be causing the change in the maser intensity .
lccccccrrrrrr observation date & 2011 sep 25 + configuration & d + r.a . of field center ( j2000 ) & @xmath49 + decl .
of field center ( j2000 ) & @xmath705@xmath5009@[email protected] + total bandwidth & 1 mhz + no . of channels & 256 + total observing time & 4 hr + rest frequency & 44.069488 ghz + target source velocity & 11.6 km @xmath9 + hanning smoothing & yes + effective velocity resolution & 0.053 km @xmath9 + fwhm of synthesized beam & @xmath8 + & p.a .
@xmath52 + line rms noise & 10 mjy beam@xmath29 | we present a second epoch of observations of the 44 ghz class i methanol maser line toward the star forming region omc-2 .
the observations were carried out with the very large array , and constitute one of the first successful zeeman effect detections with the new widar correlator . comparing to the result of our earlier epoch of data for this region , we find that the intensity of the maser increased by 50% , but the magnetic field value has stayed the same , within the errors .
this suggests that the methanol maser may be tracing the large - scale magnetic field that is not affected by the bulk gas motions or turbulence on smaller scales that is causing the change in maser intensity . |
despite recent improvements in stellar modeling , several notable discrepancies remain between theoretical predictions of low mass stellar properties and high precision measurements from techniques such as eclipsing binary ( eb ) analysis and interferometry .
one such discrepancy is the inflated radius problem , where young , low - mass stellar radii are observed to be larger by @xmath0% than theoretical predictions .
a second discrepancy is between the observed lithium patterns of young clusters such as the pleiades , which host abundance dispersions in excess of an order - of - magnitude at fixed , and standard model theoretical predictions , which anticipate no dispersion at fixed .
notably , the most li rich stars in the pleiades are also the most rapidly rotating , the opposite sense of the correlation expected from rotational mixing .
in two recent papers , we suggested that if some mechanism induced inflated radii in rapidly rotating , low - mass stars during the pre - ms , then the li destruction rate in the inflated stars would be suppressed , and one could explain both the inflated radius problem and the li - rotation correlation in the pleiades , with a single mechanism ( ( * ? ? ?
* somers & pinsonneault 2014 ) ; ( * ? ? ?
* somers & pinsonneault 2015a ) ) . in this proceedings
, we summarize results from our recent work which explores this possibility using a specific inflation mechanism , starspots .
starspots are the visual manifestation of concentrated magnetic fields near the surfaces of stars .
spots inhibit energy transport through convection , and alter the pressure conditions at the photosphere , inducing a global structural response that can be modeled ( ( * ? ? ?
* gough & taylor 1966 ) ) .
spots have long been argued to impact the luminosities , temperatures , and radii of stars ( e.g. ( * ? ? ?
* spruit & weiss 1986 ) ) , but in only a few cases has their effect on the pre - main sequence ( pre - ms ) been examined ( e.g. ( * ? ? ?
* jackson & jeffries 2014 ) ) , and never with a modern stellar evolution code .
young , active stars are typically the most heavily spotted , so a systematic study of the structural impact of spots during the pre - ms and early ms may prove crucial to our understanding of early stellar evolution . to this end
, we have incorporated in our stellar evolution code a treatment of spots which accounts for two relevant physical processes : 1 ) the redistribution of flux in the interior , modeled by explicitly altering the radiative gradient ; 2 ) the altered boundary conditions at the surface , modeled by self - consistently solving for the pressure in the un - spotted regions ( the full details are presented in @xmath1 of ( * ? ? ?
* somers & pinsonneault 2015b ) ) . using this method , we calculated a grid of low mass ( 0.1 - 0.8 ) and solar mass ( 0.8 - 1.2 ) stars , with various spot properties .
our spotted models each have a spot@xmath2photosphere temperature contrast of 0.8 , and host a variety of spot filling factors up to 50% surface coverage , the upper limit of claimed properties in the literature .
we first explored the impact of spots on stellar radii . we find that heavy spot coverage leads to a substantial enhancement at all ages , developing on the early pre - ms and persisting onto the ms .
for all masses in our range , radius anomalies peak at 8 - 12% during the pre - ms , and converge to a mass dependent pattern at the zams , where fully convective stars are inflated by @xmath34% , and solar mass stars are inflated by up to 10% .
the larger anomalies on the pre - ms result primarily from the reduced hayashi contraction rate of spotted stars , a consequence of their suppressed luminosity ( @xmath4[sec4 ] ) .
by contrast , ms anomalies result predominately from the impact of spots on the pressure boundary conditions at the surface , and at the transport of flux in the upper envelope . in the left panel of fig .
[ fig1 ] , we compare the radius predictions of our spotted models to eb masses and radii .
each data point shows the fractional inflation percentage relative to our zero - age ms predictions .
the red region shows the impact of 10 gyrs of stellar evolution on the radius , indicating that a subset of the data could be explained solely by age effects , though more than half can not .
the blue region shows the area permitted when including both spot and age effects , which could both be at work in short - period , ( presumably ) tidally - locked ebs . with only one exception ,
our spot models can account for the anomalous radii of the comparison sample .
this demonstrates that our models predict reasonable inflation factors for observationally motivated spot properties .
next , we explore the destruction of lithium in our models .
li burns at @xmath5 k , and is depleted from stellar surfaces during the pre - ms until a substantial radiative core has formed .
standard models ( red line in the right panel of fig .
[ fig1 ] ) predict that the mass and composition of a star uniquely determine its li destruction during the ms , and consequently predict no spread in li at fixed within clusters .
by contrast , a large dispersion in li abundance at fixed ( grey points ) is observed in the 120 myr pleiades cluster , suggesting additional mechanisms are at play .
spotted models are physically larger during the pre - ms , leading to a reduction in the temperature at the base of the surface convection zone , and thus a suppressed rate of li destruction . the blue line in the right panel of fig .
[ fig1 ] shows the li abundances resulting from a constant filling factor of 50% during the pre - ms , and the blue shaded regions show abundances produced by intermediate spot properties .
the majority of the data fall within this range , suggesting that stars of equal mass must have dissimilar radii during the pre - ms to explain the data .
furthermore , the most li - rich stars are the fastest spinning , suggesting that this radius dispersion arises from a physical process linked to rotation @xmath2 spots are a prime candidate . collectively , li and radius data suggest that a range of radii at fixed mass , driven by magnetic activity , emerge in open clusters during the pre - ms and persist onto the ms .
we now discuss an important implication of this conclusion .
the left panel of fig . [ fig2 ] shows standard ( red ) and 50% spotted ( blue ) isochrones at 3 myr in the hr diagram .
the blue isochrone is displaced downward and to the right , indicating that spots reduce the luminosity and of their host stars .
now consider the location of the black cross .
if a pre - ms stars is observed at this hr - diagram location , one might infer a mass of 0.5 and an age of 3 myr through comparison with the standard isochrones .
however , if this star is actually spotted , then the blue isochrones should be used to derive its properties , in which case one finds @xmath6 , and @xmath7 myr .
this demonstrates that spotted stars are _ systematically older and more massive than often measured . _
to estimate the magnitude of this effect , we return to the pleiades li abundances in fig .
if the cause of the li dispersion is indeed pre - ms spots , then the current li abundance of each individual star tells us what its spot properties must have been on the pre - ms .
we can therefore derive the distribution of pre - ms spot properties that the pleiades _ must have had _
, in order to produce the current vs. li distribution .
we can then ask what masses and ages _ would we have measured _ from this distribution , using standard isochrones .
the results are in the right panel of fig .
the blue points show the masses derived from the pleiades , backwards modeled to a coeval age of 10 myr .
the red points show the masses and ages that isochrone fitting gives for this population .
as can be seen , there is a small systematic trend towards lower inferred masses , and a very strong trend towards lower ages , with errors reaching a factor of @xmath8 for the most vulnerable stars .
this is not a subtle effect !
we further note that the lowest mass stars appear to be the youngest , a feature which has been observed in some young clusters ( e.g. ( * ? ? ?
* herczeg & hillenbrand 2015 ) ) .
this suggests that many young associations may in fact be older than often quoted , that age spreads claimed in some clusters may be spurious , and that age errors are largest for the lowest mass stars .
we have found that the impact of spots on stellar models is to increase their radii , and to lower their luminosities and , in a mass dependent fashion .
the increased radii due to spots can explain the observed radius inflation of some low mass stars , and can account for the suppressed li depletion in rapid rotators during the pre - ms .
these data collectively suggest that stars of equal mass have different physical sizes at fixed age around 3 - 10 myr . as a consequence , masses and ages derived with isochrones will be erroneously low and young , respectively , and may create the appearance of age spreads in co - eval populations . | we investigate the impact of starspots on the evolution of late - type stars during the pre - main sequence ( pre - ms ) .
we find that heavy spot coverage increases the radii of stars by 4 - 10% , consistent with inflation factors in eclipsing binary systems , and suppresses the rate of pre - ms lithium depletion , leading to a dispersion in zero - age ms li abundance ( comparable to observed spreads ) if a range of spot properties exist within clusters from 3 - 10 myr .
this concordance with data implies that spots induce a range of radii at fixed mass during the pre - ms .
these spots decrease the luminosity and of stars , leading to a displacement on the hr diagram .
this displacement causes isochrone derived masses and ages to be systematically under - estimated , and can lead to the spurious appearance of an age spread in a co - eval population . |
_ conjunctive queries _ are basic and heavily studied database queries , and can be viewed logically as formulas consisting of a sequence of existentially quantified variables , followed by a conjunction of atomic formulas . since the 1977 article of chandra and merlin @xcite ,
complexity - theoretic aspects of conjunctive queries have been a research subject of persistent and enduring interest which continues to the present day ( as discussed and evidenced , for example , by the works @xcite ) . in this article , we study _ conjunctive query evaluation _ , which is the problem of evaluating a conjunctive query on a relational structure .
conjunctive query evaluation is indeed equivalent to a number of well - known problems , including the homomorphism problem on relational structures , the constraint satisfaction problem , and conjunctive query containment @xcite .
that this problem appears in many equivalent guises attests to its fundamental , primal nature , and this problem has correspondingly been approached and studied from a wide variety of perspectives and motivations . as has been eloquently articulated in the literature
, the employment of classical complexity notions such as polynomial - time tractability to grade the complexity of conjunctive query evaluation is not totally satisfactory : a typical scenario for example , in the database context is the evaluation of a relatively short query on a relatively large structure .
this asymmetry between the two parts of the input suggests that one might employ a notion of complexity wherein one relaxes the dependence on the query .
for example , in measuring time complexity , one might allow a non - polynomial dependence on the query while enforcing a polynomial dependence on the structure .
_ parameterized complexity theory _
@xcite is a comprehensive framework for studying problems where each instance has an associated parameter , and arbitrary dependence in the parameter is permitted ; in the query evaluation setting , the query ( or the query length ) is naturally and typically taken to be the parameter .
we use and focus on this viewpoint in the present article , and hence in this discussion .
conjunctive query evaluation is known to be computationally intractable in general , and consequently a recurring theme in the study of this problem is the identification of structural properties of conjunctive queries that provide tractability or other computationally desirable behaviors . a convenient framework in which to seek such properties is the family of parameterized homomorphism problems @xmath1 , for classes of relational structures @xmath0
: the problem @xmath1 is to decide , given a relational structure @xmath2 from @xmath0 and another relational structure @xmath3 , whether there is a homomorphism from @xmath2 to @xmath3 ; the parameter here is the first structure @xmath2 . studying this problem family amounts to studying conjunctive query evaluation on various sets of conjunctive queries , as it is a classical fact that each boolean conjunctive query @xmath4 can be bijectively represented as a structure @xmath2 in such a way that , for any structure @xmath3 , it holds that @xmath4 is true on @xmath3 if and only if @xmath2 admits a homomorphism to @xmath3 @xcite .
our primary interest is in understanding the interplay between structural properties on the classes @xmath0 and the complexity behavior of @xmath1 .
( we will focus on the case where @xmath0 has bounded arity , and assume this property throughout this discussion . )
* known classifications . * an exact description of the tractable problems of the form @xmath1 was presented by grohe @xcite .
in particular , a known sufficient condition for fixed - parameter tractability of @xmath1 was that the _ cores _ of @xmath0 have _ bounded treewidth _ @xcite ; grohe completed the picture by showing that for any class @xmath0 not satisfying this condition , the problem @xmath1 is w[1]-complete .
fixed - parameter tractability is a parameterized relaxation of polynomial - time tractability , and w[1]-hardness plays , in parameterized complexity , a role similar to that played by np - hardness in classical complexity .
intuitively , the _ core _ of a structure is an equivalent structure of minimal size ; _ bounded treewidth _ is a graph - theoretical property that one might conceive of as tree - likeness .
later , a classification of the tractable @xmath1 problems was presented @xcite .
this classification is exhaustive up to parameterized logarithmic space reductions ; parameterized logarithmic space ( _ para - l _ ) relaxes logarithmic space in a way analogous to that in which fixed - parameter tractability relaxes polynomial time .
let @xmath5 denote the class of all trees , let @xmath6 denote the class of all paths , and for a class of structures @xmath0 , let @xmath7 be the class of structures obtainable by taking a structure @xmath2 in @xmath0 and adding each element of @xmath2 as a relation .
the classification states that each tractable @xmath1 problem is para - l equivalent to @xmath8 , para - l equivalent to @xmath9 , or decidable in para - l .
the properties determining which of the three behaviors occurs are _ bounded pathwidth _ and _ bounded tree depth _ , established graph - theoretical properties . in that work @xcite , it is also shown how the problems @xmath8 and @xmath9 can be characterized as complete problems for machine - defined complexity classes . in this article , we present a significantly refined complexity classification of the homomorphism problems @xmath1 which is exhaustive up to an extremely simple and computationally weak notion of reduction based on quantifier - free interpretations from first - order logic .
our classification generalizes the just - described known classifications , and indeed provides a uniform , modular , and self - contained treatment thereof .
we now give an overview of the contributions and obtained results ; we refer the reader to the technical sections for precise statements and more information . * a graph - theoretic hierarchy . * to derive our classification , we first focus on graphs . previous work @xcite related the complexity of conjunctive queries to the named graph - theoretic properties by showing that certain relationships on graph classes implied reductions for the corresponding homomorphism problems , for example : * if @xmath10 is a graph class and @xmath11 is the class of minors of graphs in @xmath10 , then @xmath12 reduces to @xmath13 .
* if the members of a graph class @xmath10 have bounded width tree decompositions whose trees lie in a graph class @xmath14 , then @xmath13 reduces to @xmath15 . we give a unified explanation for these reductions by defining a relation @xmath16 on graph classes .
this relation has the key property that @xmath17 implies that @xmath13 reduces to @xmath15 , and this key property is shown to imply the two just - named results .
the definition of the relation @xmath16 is simple and is based on a notion which we call _ graph deconstruction _ and which is strongly related to and inspired by the notion of tree decomposition .
when @xmath18 and @xmath19 are graphs , we define an @xmath19-deconstruction of @xmath18 to be a family @xmath20 of subsets of the vertex set of @xmath18 which is indexed by the vertex set @xmath21 of @xmath19 and which satisfies properties similar to those in the definition of tree decomposition ( definition [ def : deconstruction ] ) ; note that here , the graph @xmath19 is not restricted to be a tree , as it is in the definition of tree decomposition .
each such deconstruction @xmath20 has associated with it a measure called its _ width _ which is based on the sizes of the subsets @xmath22 . for graph classes @xmath10 and
@xmath14 , we define @xmath17 if and only if there is a constant @xmath23 where for each graph @xmath18 in @xmath10 , there is a graph @xmath19 in @xmath14 such that @xmath18 has a @xmath19-deconstruction of width at most @xmath23 . for a computer scientist
, this relation @xmath16 can be intuitively thought of as a notion of reduction between graph classes , and indeed some of its properties are akin to those of complexity - theoretic reductions ; for example , any graph set is equivalent to a finite variation thereof . is a _ finite variation _ of a set @xmath10 if the symmetric difference of @xmath10 and @xmath14 is finite .
] however , since the definition of the @xmath16 relation is purely graph - theoretic , we are able to unconditionally derive a complete hierarchy of graph classes with respect to this relation .
define @xmath24 to be the class of trees of height at most @xmath25 ; @xmath26 to be the class of forests of height at most @xmath25 ; and @xmath27 to be the class of all graphs .
we present the following hierarchy of graph classes ( theorem [ thm : graph - hierarchy ] ) : @xmath28 we prove that this hierarchy is strict and _ comprehensive _ in that every single graph class is equivalent to one of the sets in the hierarchy . to understand the upper levels of the hierarchy ,
we make use of known _ excluded minor theorems_. note indeed that these upper levels are highly related to established , already - mentioned conditions : as examples , a graph class @xmath10 has bounded treewidth if and only if @xmath29 , and bounded pathwidth if and only if @xmath30 ( proposition [ prop : bounded - characterizations ] ) . to determine the lower part of the hierarchy ( below @xmath6 ) , we introduce a new complexity measure on graph classes which we call _ stack depth_. * a complexity - theoretic hierarchy . *
having understood the introduced relation @xmath16 on graph classes , we then turn to study homomorphism problems . as mentioned , we prove that @xmath17 implies @xmath13 reduces to @xmath15 ( theorem [ theo : dechom ] ) ; this property was indeed a primary motivation for the definition of @xmath16 .
we prove this with respect to a computationally weak notion of reduction that we call _ quantifier - free after a precomputation ( qfap)_. this notion of reduction naturally unifies and incorporates two modes of computation that have long been studied . for each parameter
, this reduction provides a quantifier - free , first - order _ interpretation _ that defines the output instance in the input instance ; interpretations as reductions have a tradition in descriptive complexity @xcite , and dawar and he studied a type of quantifier - free interpretation in the parameterized setting @xcite .
in addition , our reduction allows for _ precomputation _ on the parameter of an input instance .
this follows an established schema in the definition of parameterized modes of computation : for example , fixed - parameter tractability can be defined as polynomial time after a precomputation , and para - l can be defined as logarithmic space after a precomputation @xcite .
note that it is readily verified that our reduction is finer than para - l reduction ( lemma [ lem : plred ] ) . since by the comprehensiveness of our graph hierarchy every graph class @xmath10 is equivalent ( under @xmath16 ) to a class @xmath14 from the hierarchy , we obtain the following : for every graph class @xmath10 , the problem @xmath13 is equivalent under qfap - reduction to a problem @xmath15 , + where @xmath14 is a graph class in the above graph hierarchy .
let @xmath0 denote an arbitrary class of bounded - arity structures .
we proceed to show that _ every _ problem of the form @xmath1 is equivalent under qfap - reduction to a problem @xmath13 , where @xmath10 is a set of graphs ( precisely , the graphs of the cores of the structures in @xmath0 ) .
this implies the following strengthening of the just - given statement on graph classes ( theorem [ theo : coregraphhom ] ) : each problem of the form @xmath1 is equivalent under qfap - reduction to a problem @xmath15 , + where @xmath14 is a graph class from the graph hierarchy .
this interestingly implies that , with respect to qfap - reduction , the complexity degrees attained by problems @xmath1 are linearly ordered according to the classes in our graph hierarchy , in particular , linearly ordered in a sequence of order type @xmath31 . in brief
, our approach for understanding the problems @xmath1 is to present a graph hierarchy and then show that this hierarchy induces a complexity hierarchy for these problems .
we wish to emphasize the unifying nature , the modularity , and the conceptual claenliness of this approach .
our definition and presentation of the graph hierarchy cleanly and neatly encapsulates the graph - theoretic content needed to present the complexity hierarchy .
we obtain a uniform and self - contained derivation of the mentioned known classifications @xcite , which derivation we find to be clearer and simpler than those of the original works .
our treatment also strengthens the known classifications , since the problems classified as being computationally equivalent are here shown to be so under qfap - reductions ; in particular , it follows from our treatment that all of the w[1]-complete problems from grohe s theorem @xcite are pairwise equivalent under qfap - reductions and hence in a very strong sense .
* consistency , pebble games , and logarithmic space . * the study of so - called consistency algorithms has a long tradition in the study of constraint satisfaction and homomorphism problems .
such algorithms are typically efficient and simple heuristics that can detect inconsistency ( that is , that an instance is a _ no _ instance ) and are based on local reasoning .
identifying cases of these problems where such algorithms provide a sound and complete decision procedure has been a central theme in the tractability theory of these problems ( see for example @xcite )
. seminal work of koalitits and vardi @xcite showed that certain natural consistency algorithms could be viewed as determining the winner in certain ehrenfeucht - frass type pebble games @xcite .
since this work , there has been sustained effort devoted to presenting pebble games that solve cases @xmath1 of the homomorphism problem . in general , this is determined by the class @xmath10 of graphs of cores of structures from @xmath0 .
for example , there is study of pebble games that solve @xmath1 when @xmath10 has bounded treewidth ( that is , when @xmath29 ) @xcite , and also when @xmath10 has bounded pathwidth ( that is , when @xmath32 ) @xcite .
we complete the picture by presenting natural pebble games that are shown to solve @xmath1 when @xmath10 has bounded tree depth ( section [ sect : pebble - games ] ) .
our pebble games are finite - round games that can be thought of as homomorphism variants of the classical ehrenfeucht - frass game .
we develop the theory of our games , showing for example that it is decidable , given a structure @xmath2 , whether or not a particular game solves the homomorphism problem on @xmath2 ( theorem [ thm : char - v - game - solves ] ) .
we also show strong correspondences between the number of pebbles needed to solve @xmath1 and the tree depth of @xmath10 ; and , between the number of rounds needed to solve @xmath1 and the stack depth of @xmath10 ( section [ subsect : game - resources ] ) . as a fruit of our development of pebble games ,
we obtain a characterization of the classical homomorphism problems @xmath33 decidable in classical logarithmic space : these are precisely those where the cores of structures from @xmath0 have bounded tree depth . this result is conditional on a certain hypothesis from parameterized complexity theory
. * outlook . * motivated by the desire to understand the complexity of homomorphism problems , we defined a relation on graph classes and then analyzed the resulting hierarchy .
we then showed that the resulting graph hierarchy can be used to derive a complexity hierarchy on the problems @xmath1 which generalizes existing classification results .
this uniform derivation is a testament to the relevance of the graph hierarchy , and we view the simple definition of the relation @xmath16 , as well as the development of its basic theory , as conceptual contributions . the graph hierarchy presented here is strongly reminiscent of a hierarchy given by blumensath and courcelle @xcite , and this situation naturally suggests the possibility of establishing connections between this work and theirs .
our results show that the given hierarchies along with the notion of qfap - reduction provide a clean and comprehensive understanding of the complexity degrees of parameterized homomorphism problems .
we expect that the given hierarchies can be meaningfully used to obtain a fine - grained understanding of the complexity of other problems of independent interest . as an indication favoring this perspective
, we show , in the final section of the article , that the hierarchy can be used to classify the complexity of model - checking existential sentences having bounded quantifier rank .
for @xmath34 we let @xmath35 $ ] denote @xmath36 and understand @xmath37=\emptyset$ ] . a _ vocabulary _ is a finite set of relation symbols , where each symbol @xmath38 has an associated arity @xmath39 .
structure with vocabulary @xmath40 _ , for short , a _ @xmath40-structure _
@xmath3 is given by a non - empty set @xmath41 called its _ universe _ together with an _ interpretation _
@xmath42 of @xmath38 for every @xmath43 .
we only consider finite structures , i.e. structures with finite universe .
when @xmath3 is a @xmath40-structure and @xmath44 a non - empty subset of @xmath41 , we let @xmath45 denote the @xmath40-structure _ induced _ in @xmath2 on @xmath44 : it has universe @xmath44 and interprets every @xmath43 by @xmath46 . the class of all @xmath40-structures is denoted by @xmath47 $ ] , and the class of all structures by @xmath48 . for a vocabulary @xmath40 ,
a ( first - order ) @xmath40-formula @xmath49 is built from _
atoms _ @xmath50 and @xmath51 where @xmath52 and the @xmath53 are variables and @xmath54 , by means of boolean combinations @xmath55 and existential and universal quantification @xmath56 .
we write @xmath57 for @xmath49 to indicate that the free variables of @xmath49 are among the components of @xmath58 ; for a @xmath40-structure , and @xmath59 we write @xmath60 to indicate that @xmath61 satisfies @xmath57 in @xmath2 , further we write @xmath62 .
formulas without free variables are _ sentences_. the _ quantifier rank _ @xmath63 of a formula @xmath49 is defined as follows : @xmath64 for atoms @xmath49 ; @xmath65 ; @xmath66 ; @xmath67 . two structures @xmath2 and @xmath3 interpreting the same vocabulary are called _
similar_. in this case , a _ homomorphism _ from @xmath2 to @xmath3 is a function @xmath68 such that for every @xmath43 and every tuple @xmath69 , it holds that @xmath70 .
we write @xmath71 to indicate that such a homomorphism exists .
a _ partial homomorphism _
@xmath72 from @xmath2 to @xmath3 is either @xmath73 or a homomorphism from @xmath74 to @xmath3 ; by @xmath75 we denote the domain of @xmath72 and by @xmath76 its image .
a structure @xmath2 is a _ core _ if all homomorphisms from @xmath2 to itself are injective .
the _ core of a structure _ @xmath2 is a structure @xmath3 such that there is a homomorphism from @xmath2 to @xmath3 , @xmath3 is a core , @xmath77 , and @xmath78 for each symbol @xmath38 .
it is well - known that each finite structure has at least one core , and that all cores of a finite structure are isomorphic ; for this reason , one often speaks of _ the _ core of a finite structure @xmath2 , which we denote here by @xmath79 .
for example , all structures of the form @xmath80 are cores . here , @xmath80 is the expansion obtained from @xmath2 by interpreting for each @xmath81 a new unary relation symbol @xmath82 by @xmath83 . for a class of structures
@xmath0 we let @xmath84 in this article , a _ graph _ is a @xmath85-structure @xmath18 for a binary relation symbol @xmath86 such that @xmath87 is irreflexive and symmetric .
a graph @xmath18 is a _
subgraph _ of another graph @xmath19 if @xmath88 and @xmath89 .
given a graph @xmath18 we write @xmath90 when @xmath2 is a @xmath40-structure , we let @xmath91 be the graph with universe @xmath92 and an edge between @xmath93 if @xmath94 and there are @xmath95 and a tuple @xmath96 such that @xmath97 are components of @xmath61 .
we call @xmath2 _ connected _ if so is the graph @xmath91
. a _ ( connected ) component _ of @xmath2 is a structure induced in @xmath2 on a ( connected ) component of @xmath91 .
a _ path _ is a graph isomorphic to the graph with universe @xmath35 $ ] and edges @xmath98 for some @xmath99 .
a _ grid _ is a graph that is isomorphic to an _
@xmath25-by-@xmath25 grid _ for some @xmath34 ; this is the graph with universe @xmath35 \times [ n]$ ] and edges @xmath100 \times [ n])^2 \mid acyclic . by a _ rooted tree _
, we mean a tree that interprets a unary relation symbol @xmath101 by a set containing a single element , called the _ root _ of the tree . by a _ rooted forest _ ,
we mean a graph @xmath18 where , for each component @xmath102 , the graph @xmath103 is a rooted tree .
when @xmath104 and @xmath105 are elements of a rooted tree , we say that @xmath104 is an _ ancestor _ of @xmath105 and that @xmath105 is a _ descendent _ of @xmath104 if @xmath104 lies on the unique path from the root to @xmath105 ; if in addition @xmath106 , we say that @xmath104 is a _ proper ancestor _ of @xmath105 and that @xmath105 is a _ proper descendent _ of @xmath104 .
height _ of a rooted tree is the maximum length of a path from the root to some vertex .
the height of a tree @xmath107 is the minimum height of all rootings of @xmath107 .
the height of a forest @xmath108 is the maximum height of a ( connected ) component of @xmath108 .
we use the following notation .
let @xmath109 ; * @xmath27 denotes the class of all graphs ; * @xmath110 denotes the class of all grids ; * @xmath5 denotes the class of all trees ; * @xmath6 denotes the class of all paths ; * @xmath111 denotes the class of all trees of height at most @xmath105 ; * @xmath112 denotes the class of all forests of height at most @xmath105 .
the _ tree depth _ of a connected graph @xmath18 is the minimum @xmath113 such that there exists a rooted tree @xmath107 with universe @xmath114 of height @xmath115 such that @xmath116 is contained in the closure of @xmath107 .
the _ closure _ of @xmath107 is the set of edges @xmath117 such that either @xmath118 is an ancestor of @xmath119 in @xmath107 or vice versa . for an arbitrary graph ,
its tree depth is defined to be the maximum tree depth taken over all components of @xmath18 .
a _ tree decomposition _ of a graph @xmath18 is a tree @xmath19 along with an @xmath21-indexed family @xmath20 of subsets of @xmath120 satisfying the following conditions : * for every pair @xmath121 , there exists @xmath122 such that @xmath123 . *
( connectivity ) for each element @xmath124 , the set @xmath125 is connected in @xmath19 .
the _ width _ of a tree decomposition is @xmath126 .
a _ path decomposition _ of a graph is a tree decomposition where the tree @xmath19 is a path .
the _ treewidth _ of a graph @xmath18 is the minimum width over all tree decompositions of @xmath18 ; likewise , the _ pathwidth _ of @xmath18 is the minimum width over all path decompositions of @xmath18 .
a class @xmath10 of graphs has _
bounded treewidth _ if there exists a constant @xmath23 such that each graph @xmath127 has treewidth @xmath128 ; the properties of _ bounded pathwidth _ and _ bounded tree depth _ are defined similarly .
a graph @xmath129 is a _ minor _ of a graph @xmath18 if there exists a _ minor map _ from @xmath129 to @xmath18 , which is a map @xmath130 defined on @xmath131 where * for each @xmath132 , it holds that @xmath133 is a non - empty , connected subset of @xmath120 ; * the sets @xmath133 are pairwise disjoint ; and , * for each @xmath134 there exist @xmath135 and @xmath136 such that @xmath137 .
when @xmath18 is a graph , we use @xmath138 to denote the class of all minors of @xmath18 , and we extend this notation to a class @xmath10 setting @xmath139 .
the following theorem is known ; the first two parts are due to robertson and seymour s graph minor series ( see @xcite ) and the third is due to blumensath and courcelle @xcite . [
thm : excluded ] let @xmath10 be a set of graphs . 1 . ( excluded grid theorem ) @xmath10 has bounded treewidth if and only if there exists a grid that is not contained in @xmath140 .
( excluded tree theorem ) @xmath10 has bounded pathwidth if and only if there exists a tree that is not contained in @xmath140 .
( excluded path theorem ) @xmath10 has bounded tree depth if and only if there exists a path that is not contained in @xmath140 .
we will make use of the following known , readily verified fact .
[ prop : depth - gives - decomposition ] let @xmath18 be a connected graph , and suppose that @xmath107 is a rooted tree with height @xmath72 witnessing that @xmath18 has tree depth @xmath115 , i.e. @xmath107 it has height @xmath115 and its closure contains @xmath87 . then the tree @xmath141 together with @xmath142 is a tree decomposition of @xmath18 of width @xmath72 where @xmath143 . to prove proposition [ prop : depth - gives - decomposition ] , the key observation is the following .
for any two vertices @xmath144 connected by an edge in @xmath18 , one is an ancestor of the other in @xmath107 , and if , say , @xmath119 is an ancestor of @xmath118 , then @xmath145 .
[ def : deconstruction ] when @xmath18 and @xmath19 are graphs , an _ @xmath19-deconstruction of @xmath18 _ is an @xmath21-indexed family @xmath20 of subsets of @xmath120 that satisfies the following two conditions : * ( coverage ) for each pair @xmath121 , there exists a pair @xmath146 such that @xmath147 . * ( connectivity ) for each element @xmath124 , the set @xmath125 is connected in @xmath19 .
the _ width _ of a @xmath19-deconstruction @xmath20 is defined as @xmath148 .
we will refer to the subsets @xmath22 as _
bags_. note that the definition of an @xmath19-deconstruction is similar to that of a tree decomposition , but one important difference is that , in the definition of @xmath19-deconstruction , it is not required that an edge @xmath137 lie inside a single bag @xmath22 , but rather , may lie inside the union @xmath149 of two bags where @xmath150 .
[ ex : grid ] let @xmath151 .
when @xmath19 is the @xmath25-by-@xmath25 grid ( defined in section [ subsect : graphs ] ) , any graph @xmath18 on @xmath25 vertices has an @xmath19-deconstruction of width @xmath152 .
assume without loss of generality that @xmath153 $ ] .
the desired deconstruction is @xmath154 defined by @xmath155 .
coverage holds , since each pair @xmath156 ^ 2 $ ] has @xmath157 .
connectivity holds , since for each @xmath158 $ ] , the set @xmath159 forms a cross in the grid .
[ ex : self - deconstruction ] for any graph @xmath18 , the family @xmath160 defined by @xmath161 is a @xmath18-deconstruction of @xmath18 having width at most @xmath162 .
[ prop : inheriting - connectivity ] let @xmath163 be graphs , @xmath164 connected in @xmath18 and @xmath165 an @xmath19-deconstruction of @xmath18 . then @xmath166 is connected in @xmath19 .
suppose that @xmath137 , and define @xmath167 and @xmath168 .
it suffices to show that @xmath169 is connected in @xmath19 . by coverage , there exists @xmath146 such that @xmath147 .
if there exists such a pair @xmath170 with @xmath171 , then the sets @xmath172 and @xmath173 share the vertex @xmath72 in common ; otherwise , there is a pair @xmath150 with @xmath174 and @xmath175 , so @xmath176 and @xmath177 , implying connectedness of the union @xmath169 in @xmath19 .
let @xmath10 and @xmath14 be sets of graphs .
* we write that @xmath10 has @xmath14-deconstructions of width @xmath178 if for each graph @xmath127 , there exists a graph @xmath179 such that @xmath18 has an @xmath19-deconstruction of width @xmath178 .
* we write that @xmath10 has @xmath14-deconstructions of bounded width if there exists @xmath180 such that @xmath10 has @xmath14-deconstructions of width @xmath178 .
we will employ analogous terminology to discuss , for example , _ nice deconstructions _ which will be defined later .
we define the binary relation @xmath16 on sets of graphs as follows : @xmath17 if and only if @xmath10 has @xmath14-deconstructions of bounded width .
we write @xmath181 if @xmath17 and @xmath182 , and we write @xmath183 if @xmath17 and @xmath184 .
we now present some basic properties of the relation @xmath16 .
[ prop : preorder ] the relation @xmath16 is reflexive and transitive .
reflexivity of @xmath16 follows directly from example [ ex : self - deconstruction ] .
note that throughout , we will tacitly use the fact that , if @xmath185 , then @xmath17 .
transitivity of @xmath16 follows immediately from the following lemma .
[ lemma : compose - deconstructions ] let @xmath18 , @xmath19 , and @xmath186 be graphs ; suppose that @xmath18 has an @xmath19-deconstruction of width @xmath187 , and that @xmath19 has an @xmath186-deconstruction of width @xmath128 .
then , @xmath18 has an @xmath186-deconstruction of width @xmath188 .
suppose that @xmath18 has the @xmath19-deconstruction @xmath20 of width @xmath187 , and that @xmath19 has the @xmath186-deconstruction @xmath189 of width @xmath128 .
define the family @xmath190 by @xmath191 .
we claim that @xmath190 is an @xmath186-deconstruction of @xmath18 , which suffices . to verify coverage , let @xmath121 . by coverage of @xmath192
, there exists @xmath146 such that @xmath147 . by coverage of @xmath193
, there exists @xmath194 such that @xmath195 . by definition of @xmath196
, we have that @xmath197 , from which it follows that @xmath198 . to verify connectivity ,
let @xmath124 be an arbitrary element .
by connectivity of @xmath192 , it holds that @xmath199 is connected in @xmath19 .
it follows from proposition [ prop : inheriting - connectivity ] that @xmath200 is connected in @xmath186 .
we claim that @xmath201 .
this is because @xmath202 it is an immediate consequence of proposition [ prop : preorder ] that @xmath203 is an equivalence relation .
[ prop : equivalence - with - minors ] for any set of graphs @xmath10 , it holds that @xmath204 .
it is clear that @xmath205 , since @xmath206 . to show that @xmath207 , we prove that when a graph @xmath129 is a minor of a graph @xmath18 , it holds that @xmath129 has a @xmath18-deconstruction of width @xmath208 .
let @xmath130 be a minor map from @xmath129 to @xmath18 , and define , for all @xmath124 , the set @xmath209 to be @xmath210 .
clearly , for each @xmath124 , it holds that @xmath211 .
we claim that @xmath160 is a @xmath18-deconstruction of @xmath129 . for each @xmath132 , since @xmath133 is non - empty , there exists @xmath124 such that @xmath212 . for each @xmath134 , by definition of minor , there exists @xmath137 with @xmath135 and @xmath213 ; it thus holds that @xmath214 . for each @xmath132 , the set
@xmath215 is equal to @xmath133 , and is hence connected since @xmath130 is a minor map .
we now generalize the notion of tree decomposition to arbitrary graphs , and then compare the resulting notion with the presented notion of @xmath19-deconstruction .
when @xmath18 and @xmath19 are graphs , define an _
@xmath19-decomposition _ of @xmath18 to be an @xmath21-indexed family @xmath20 of subsets of @xmath120 satisfying the following conditions : * for each pair @xmath121 , there exists @xmath216 such that @xmath123 .
* connectivity ( as defined in definition [ def : deconstruction ] ) this is a natural generalization of the definition of _ tree decomposition _ :
a tree decomposition is precisely a @xmath19-decomposition where @xmath19 is required to be a tree .
[ prop : decomposition - vs - deconstruction ] let @xmath18 and @xmath19 be graphs , and let @xmath217 . 1 .
when @xmath20 is an @xmath19-decomposition of @xmath18 , it holds that @xmath20 is also an @xmath19-deconstruction of @xmath18 .
if @xmath19 is a tree such that @xmath18 has an @xmath19-deconstruction of width @xmath128 , then @xmath18 has an @xmath19-decomposition of width @xmath218 .
consequently , when @xmath14 is a set of trees , it holds that @xmath17 if and only if @xmath10 has @xmath14-decompositions of bounded width .
\(1 ) is straightforwardly verified . for ( 2 ) ,
suppose that @xmath20 is an @xmath19-deconstruction of @xmath18 where each bag has size @xmath128 .
root the tree @xmath19 , and define @xmath219 as follows : if @xmath72 is the root of @xmath19 , define @xmath220 , and otherwise define @xmath221 to be the parent of @xmath72 . define @xmath222 by @xmath223 ; it is straightforward to verify that @xmath222 is an @xmath19-decomposition of @xmath18 , and each of its bags has size @xmath224 .
we find that working with the notion of @xmath19-deconstruction is cleaner than working with the notion of @xmath19-decomposition ; this is a primary reason for our focus on the notion of @xmath19-deconstruction . indeed ,
note that while reflexivity of the @xmath16 relation is straightforward to prove , we do not know of a simple proof of reflexivity of the analogous relation defined via @xmath19-decomposition . note that while there exists a constant @xmath23 such that each graph @xmath18 has a @xmath18-deconstruction of width @xmath128 ( in particular , one can take @xmath225 ) , there does not exist a constant @xmath23 such that each graph @xmath18 has a @xmath18-decomposition of width @xmath128 : the bags of a @xmath18-decomposition of width @xmath128 can cover at most a number of edges that is linear in the number of vertices ( @xmath226 many ) , but graphs in general may have quadratically many edges .
recall that @xmath111 denotes the class of all trees of height @xmath227 . for @xmath228 , @xmath180 , define @xmath229 to be the tree with universe @xmath230^{\leq h}$ ] and with @xmath231^{<h } , i \in [ k ] \}$ ] .
here , @xmath230^{\leq h}$ ] and @xmath230^{<h}$ ] denote the sets of strings over alphabet @xmath230 $ ] of length @xmath115 and of length @xmath232 , respectively .
define the _ stack depth _ of a set @xmath10 of graphs to be @xmath233 ; let it be understood that this maximum should be considered @xmath234 if the set is infinite .
( note , though , that in this article we generally only measure the stack depth of sets of graphs for which it is finite . )
[ thm : stack - depth - decomps ] suppose that @xmath235 are constants and that @xmath10 is a set of trees having stack depth @xmath105 and where each tree has height @xmath236 .
then @xmath237 .
let @xmath129 , @xmath18 be rooted trees .
let us say that an @xmath129-deconstruction of @xmath18 is _ nice _ if it is of the form @xmath238 such that the following hold : * @xmath130 is a minor map from @xmath131 to @xmath120 .
* @xmath239 , where @xmath240 and @xmath241 denote the roots of @xmath18 and @xmath129 , respectively . *
if @xmath242 is a child of @xmath243 , @xmath135 , @xmath136 , and @xmath137 , then @xmath119 is a child of @xmath118 .
[ lemma : transitivity - niceness ] suppose that @xmath244 , @xmath129 , and @xmath18 are rooted trees , that @xmath245 is a nice @xmath244-deconstruction of @xmath129 , and that @xmath238 is a nice @xmath129-deconstruction of @xmath18 .
then @xmath246 is a nice @xmath244-deconstruction of @xmath18 , where here @xmath247 denotes @xmath248 .
it is straightforward to verify that @xmath249 is a minor map from @xmath244 to @xmath18 .
it follows from the proof of lemma [ lemma : compose - deconstructions ] that @xmath246 is an @xmath244-deconstruction of @xmath18 .
we verify that this deconstruction is nice , as follows . by the niceness of @xmath250 and @xmath251 , we have that @xmath239 and @xmath252 , so @xmath253 ( here , we use @xmath240 , @xmath241 , and @xmath254 to denote the roots of @xmath18 , @xmath129 , and @xmath244 , respectively ) .
next , suppose that @xmath255 is a child of @xmath25 in @xmath244 , that @xmath256 , that @xmath257 , and that @xmath137 .
there exist @xmath258 such that @xmath135 , @xmath259 , @xmath136 , and @xmath260 . since @xmath261 , we have @xmath262 ; as @xmath137 , we then have @xmath134 , and from the niceness of @xmath251 , we have that @xmath242 is a child of @xmath243 in @xmath129 .
it then follows from the niceness of @xmath250 that @xmath119 is a child of @xmath118 . for @xmath263 and @xmath180 , let us say that a node @xmath264 of a rooted tree has property @xmath265 if @xmath266 or if @xmath267 and @xmath264 has @xmath268 pairwise incomparable descendents each having property @xmath269 .
we consider two nodes @xmath270 to be incomparable if neither is an ancestor of the other . )
let us say that a rooted tree has property @xmath271 if its root has property @xmath265 .
observe that if a rooted tree has property @xmath271 , then the tree contains @xmath272 as a minor .
[ lemma : property - transfer ] suppose that @xmath129 and @xmath18 are rooted trees and that @xmath238 is a nice @xmath129-deconstruction of @xmath18 .
for @xmath263 and @xmath180 , if @xmath129 has property @xmath265 , then @xmath18 has property @xmath265 . for each @xmath132
, since @xmath133 is connected , it has a unique highest element , by which we mean the element with shortest distance to the root ; denote this element by @xmath273 .
suppose that @xmath242 is a child of @xmath243 in @xmath129 ; the parent of @xmath274 must , by the definition of deconstruction , lie in @xmath275 where @xmath276 is adjacent to @xmath242 ; but it must be that @xmath277 by the niceness of @xmath250 .
we thus have that , when @xmath242 is a child of @xmath243 in @xmath129 , it holds that @xmath274 is a descendent of @xmath273 ; it can then be verified ( by induction on @xmath105 ) that if a node @xmath132 has property @xmath265 in @xmath129 , then @xmath273 has property @xmath265 in @xmath18 .
( theorem [ thm : stack - depth - decomps ] ) let @xmath278 be a constant .
suppose that @xmath10 is a set of rooted trees of height @xmath236 which do not have property @xmath279 ; we prove that @xmath10 has nice @xmath111-deconstructions of bounded width .
( this suffices , since the assumption that @xmath10 has stack depth @xmath105 implies that there is a constant @xmath278 such that @xmath280 , which in turn implies that the trees in @xmath10 do not have property @xmath279 . ) we proceed by induction on @xmath105 . [ [ case - d-0 ] ] case @xmath266 + + + + + + + + + + + + + + + + + + + + + + + + we have that the trees in @xmath10 do not have property @xmath281 .
consider a rooted tree from @xmath10 .
the number of leaves is bounded above by @xmath282 ; since each node is the ancestor of a leaf , the total number of nodes is bounded above by @xmath283 .
thus @xmath10 has nice @xmath284-deconstructions of width @xmath285 .
[ [ case - d-0 - 1 ] ] case @xmath267 + + + + + + + + + + + + + + + + + + + + + + + + we argue by induction on @xmath286 .
if @xmath287 , then we have that @xmath288 and we are done ( for each @xmath127 , use the @xmath18-deconstruction of @xmath18 discussed in conjunction with reflexivity in proposition [ prop : preorder ] ) .
so suppose that @xmath289 .
define a set of trees @xmath290 as follows : for each tree @xmath18 in @xmath10 , and for each child @xmath291 of the root of @xmath18 , if @xmath291 does not have property @xmath292 , then place the subtree of @xmath18 rooted at @xmath291 in @xmath290 .
the trees in @xmath290 have bounded height and do not have property @xmath292 ; so , by induction , we have that there is a constant @xmath23 such that @xmath290 has nice @xmath293-deconstructions of width @xmath128 .
let @xmath18 be a tree in @xmath10 .
let @xmath294 denote the children of the root that have property @xmath292 , and let @xmath295 denote the remaining children of the root . since the root of @xmath18 does not have property @xmath296 , we have that @xmath297 .
let @xmath298 denote the subtrees of @xmath18 rooted at @xmath295 , respectively .
each tree @xmath299 is contained in @xmath290 , and so for each @xmath300 $ ] , there is a tree @xmath301 such that @xmath299 has a nice @xmath302-deconstruction @xmath303 of width @xmath128 .
now define the tree @xmath19 to be the minor of @xmath18 obtained from @xmath18 by contracting together the vertices @xmath304 to obtain @xmath305 , and by replacing each @xmath299 with @xmath302 .
observe that the height of @xmath19 is @xmath306 .
the following map @xmath130 is a minor map from @xmath19 to @xmath18 : @xmath307 , @xmath308 is equal to @xmath309 if @xmath310 , and @xmath311 for all other vertices @xmath312 .
it is straightforward to verify that @xmath313 gives a nice @xmath19-deconstruction of @xmath18 having width @xmath314 .
let @xmath14 denote the set of all trees @xmath19 obtained from @xmath315 in this way .
we just saw that @xmath10 has nice @xmath14-deconstructions of bounded width .
since @xmath14 has height @xmath306 and does not have property @xmath279 by lemma [ lemma : property - transfer ] , by induction , @xmath14 has nice @xmath316-deconstructions of bounded width . as a consequence of lemma [ lemma : transitivity - niceness ]
, we obtain that @xmath10 has nice @xmath111-deconstructions of bounded width .
recall that @xmath317 denote the classes of graphs , trees and paths respectively , and @xmath318 denote the classes trees respectively forest of height at most @xmath105 ( cf .
section [ subsect : graphs ] ) .
( hierarchy theorem ) [ thm : graph - hierarchy ] the hierarchy @xmath319 presents correct relationships , and is comprehensive in that each set of graphs is equivalent ( under @xmath203 ) to one of the sets therein .
we give in this section a sequence of lemmas , and then show how the theorem follows . at the end of the section .
we begin by observing that the established conditions of bounded treewidth , bounded pathwidth , and bounded tree depth can be formulated using the @xmath16 relation and the defined graph sets .
[ prop : bounded - characterizations ] let @xmath10 be a set of graphs . 1 . @xmath10 has bounded treewidth if and only if @xmath29 .
@xmath10 has bounded pathwidth if and only if @xmath30 .
3 . @xmath10 has bounded tree depth if and only if there exists @xmath263 such that @xmath320 . the first two claims follow immediately from proposition [ prop : decomposition - vs - deconstruction ] . for the third claim , we reason as follows . for the forward direction ,
let @xmath105 be an upper bound on the tree depth of @xmath10 , and let @xmath18 be a graph in @xmath10 ; proposition [ prop : depth - gives - decomposition ] ( along with proposition [ prop : decomposition - vs - deconstruction ] ) implies that each component of @xmath18 has a @xmath111-deconstruction of width @xmath227 . to establish the backward direction , proposition [ prop : decomposition - vs - deconstruction ] yields that there exists a constant @xmath23 such that @xmath10 has @xmath112-decompositions of width @xmath321 , which implies that @xmath10 has tree depth @xmath322 ( see ( * ? ? ?
* remark 4.3(a ) ) ) . in the next two lemmas , we present the negative results needed to give the hierarchy , showing that various pairs of graph classes are not related by @xmath16 .
[ lemma : top - hierarchy ] the following relationships hold .
1 . @xmath323 2 .
@xmath324 3 .
@xmath325 , for all @xmath263 immediate from theorem [ thm : excluded ] and proposition [ prop : bounded - characterizations ] .
[ lemma : bounded - td - hierarchy ] for each @xmath263 , the following hold : 1 .
@xmath326 2 .
@xmath327 we prove this by induction on @xmath105 . when @xmath266 , the claim that @xmath326 is clear , since @xmath284 contains only one - vertex trees , but @xmath328 contains graphs with arbitrarily many vertices .
the claim that @xmath327 always follows from @xmath326 , as follows : the assumption @xmath329 immediately implies that @xmath330 by proposition [ prop : inheriting - connectivity ] ; this in turn implies that @xmath331 .
let @xmath267 ; we need to prove that @xmath326 .
suppose for a contradiction that @xmath23 is a constant such that @xmath112 has @xmath111-deconstructions of width @xmath128 .
we show that @xmath332 , which contradicts the induction hypothesis .
given an arbitrary tree @xmath107 in @xmath111 , create @xmath333 copies of it ; the resulting graph has a @xmath19-deconstruction @xmath20 of width @xmath128 , with @xmath334 .
root @xmath19 with a vertex @xmath305 to witness that its height is @xmath335 .
since the bag @xmath336 has size @xmath128 , there must be a copy of @xmath107 that is disjoint from @xmath336 ; call this the _ key copy _ and denote its universe by @xmath282 . by proposition [ prop : inheriting - connectivity ]
, it is possible to take a subtree @xmath337 of @xmath19 that excludes @xmath305 such that @xmath338 gives an @xmath337-deconstruction of the key copy .
the graph @xmath337 has height @xmath339 .
we thus showed that @xmath111 has @xmath293-deconstructions of width @xmath128 , implying that @xmath332 , as desired .
[ lemma : placing - g ] let @xmath10 be a set of graphs . 1 . if @xmath340 , then @xmath29 .
2 . if @xmath341 , then @xmath30 .
3 . if @xmath342 , then @xmath10 has bounded tree depth .
we begin with the third claim .
suppose that @xmath342 .
it holds by proposition [ prop : equivalence - with - minors ] that @xmath343 , from which it follows that @xmath344 and hence , by theorem [ thm : excluded ] , @xmath10 has bounded tree depth . to establish the second claim , we reason in an analogous way .
if @xmath341 , it follows from proposition [ prop : equivalence - with - minors ] and theorem [ thm : excluded ] that @xmath10 has bounded pathwidth , and it follows from proposition [ prop : bounded - characterizations ] that @xmath30 .
we can also prove the first claim in an analogous way , but in order to use theorem [ thm : excluded ] , we need to prove that @xmath345 , where @xmath110 denotes the set of all grids ; it follows from this that @xmath340 implies @xmath346 .
that @xmath345 follows from example [ ex : grid ] : each graph has @xmath110-deconstructions of width @xmath208 .
[ lemma : graphs - to - forests ] let @xmath10 be a set of graphs .
if @xmath10 has bounded tree depth , then there exists a set @xmath14 of forests of bounded height such that @xmath181 and such that @xmath10 and @xmath14 have the same stack depth .
( namely , one can take @xmath14 to be the set of all forests that are minors of @xmath10 . )
let @xmath14 be as described .
the set @xmath14 has bounded height , as @xmath10 has bounded tree depth .
the sets @xmath10 and @xmath14 have the same tree minors , so they have the same stack depth
. we have @xmath182 by proposition [ prop : equivalence - with - minors ] .
we prove @xmath17 as follows .
the proof is along the lines of that of ( * ? ? ?
* lemma 4.8 ) ; we provide a proof here for completeness . for each graph @xmath127 ,
let @xmath19 be a forest that contains a depth - first search tree of each component of @xmath18 ( for precise details , we refer to diestel @xcite , who speaks of _ normal spanning trees _ ; see proposition 1.5.6 ) .
since @xmath19 is a subgraph of @xmath18 , we have @xmath179 .
each component of @xmath18 is contained in the closure of its corresponding component in @xmath19 ( in the sense of the definition of tree depth ) , so proposition [ prop : depth - gives - decomposition ] and the fact that @xmath14 has bounded height imply that @xmath17 .
[ lemma : stack - depth - in - hierarchy ] let @xmath10 be a set of forests having bounded height , and let @xmath105 denote the stack depth of @xmath10 .
it holds either that @xmath347 or that @xmath348 .
let @xmath349 be the set of connected graphs that appear as components of graphs in @xmath10 ; note that @xmath105 is the stack depth of @xmath349 . for each @xmath180 and each graph @xmath127
, define @xmath350 to be the number of components of @xmath18 having the tree @xmath272 as a minor .
we consider two cases .
* case 1 : * suppose that , for all @xmath180 , the set @xmath351 has infinite size .
we claim that @xmath348 .
it follows directly from theorem [ thm : stack - depth - decomps ] that @xmath320 . for each @xmath180 , let us use @xmath352 to denote the graph consisting of @xmath268 disjoint copies of @xmath272 . by assumption , we have @xmath353 . we also have that ( up to isomorphism ) each graph in @xmath112 is a subgraph of a graph of the form @xmath354 ; hence , by proposition [ prop : equivalence - with - minors ] , it holds that @xmath355 . *
case 2 : * when the assumption of the first case does not hold , one can choose a sufficiently large @xmath278 such that , for all @xmath127 , it holds that @xmath356 .
we claim that @xmath347 .
that @xmath357 follows from the hypothesis that @xmath105 is the stack depth of @xmath349 .
we show that @xmath237 .
let @xmath14 be the subset of @xmath349 that contains a graph @xmath358 if and only if @xmath359 is not a minor of @xmath19 . by theorem [ thm : stack - depth - decomps ]
, it holds that @xmath360 ; let @xmath217 be such that @xmath14 has @xmath293-deconstructions of width @xmath128 .
let @xmath127 ; let @xmath361 be the components of @xmath18 having @xmath359 as a minor , and let @xmath362 be the other components of @xmath18 . by the choice of @xmath282 , it holds that @xmath363 . since @xmath364
, there exists @xmath365 such that @xmath349 has @xmath111-deconstructions of width @xmath187 .
let @xmath366 be a sufficiently large tree so that each @xmath367 has a @xmath107-deconstruction @xmath368 of width @xmath187 ; then , the disjoint union of the @xmath367 has a @xmath107-deconstruction of width @xmath369 , namely , @xmath370 .
each @xmath371 has a @xmath372-deconstruction of width @xmath128 , where @xmath373 .
let @xmath374 be equal to @xmath107 but augmented so that the root of each @xmath372 is a child of the root of @xmath107 ; we have that the depth of @xmath374 is @xmath105 .
the graph @xmath18 has a @xmath374-deconstruction where each bag is defined as it was in the respective @xmath107-deconstruction or @xmath372-deconstruction ; this @xmath374-deconstruction has width @xmath375 .
the following is a consequence of the previous two lemmas .
[ lemma : stack - depth - placement ] let @xmath10 be a set of graphs having bounded tree depth , and let @xmath376 .
the set @xmath10 has stack depth @xmath105 if and only if @xmath347 or @xmath348 .
we first prove the forward direction .
suppose that @xmath10 has stack depth @xmath105 .
by lemma [ lemma : graphs - to - forests ] , there exists a set of forests @xmath14 having bounded height and stack depth @xmath105 such that @xmath181 . by lemma [ lemma : stack - depth - in - hierarchy ]
, we have that @xmath377 or @xmath378 , implying that @xmath347 or @xmath348 .
for the backward direction , suppose that @xmath347 or @xmath348 .
by lemma [ lemma : bounded - td - hierarchy ] , it holds that @xmath379 .
if @xmath10 does not have stack depth @xmath105 , then we have a contradiction to the forward direction .
( proof of theorem [ thm : graph - hierarchy ] ) it is clear that @xmath380 from which it follows that @xmath381 lemmas [ lemma : top - hierarchy ] and [ lemma : bounded - td - hierarchy ] imply that @xmath382 to prove that the hierarchy is comprehensive , let @xmath10 be an arbitrary set of graphs . consider the first of the following conditions that holds . * if @xmath383 , then clearly @xmath384 . + otherwise , by lemma [ lemma : placing - g ] , it holds that @xmath29 . * if @xmath385 , then @xmath386 .
+ otherwise , by lemma [ lemma : placing - g ] , it holds that @xmath30 . * if @xmath387 , then @xmath388 . + otherwise , by lemma [ lemma : placing - g ] , @xmath10 has bounded tree depth . *
if @xmath10 has bounded tree depth , it follows from lemma [ lemma : stack - depth - placement ] that either @xmath347 or @xmath348 , where @xmath105 is the stack depth of @xmath10 .
a _ parameterized problem @xmath389 _ is a subset of @xmath390 . given an instance @xmath391 of @xmath389 we refer to @xmath268 as its _
parameter_. the _
@xmath268th slice _ of @xmath389 is the classical problem @xmath392 ; here we are refering to subsets of @xmath393 as _ classical problems_. following flum and grohe @xcite , we say @xmath389 is in l _ after a pre - computation _ if there exists a computable function @xmath394 and a classical problem @xmath395 in l such that for all @xmath396 @xmath397 where @xmath398 is some standard pairing function for binary strings .
equivalently , this means that @xmath399 is decidable in space @xmath400 for some computable @xmath401 .
the class of such problems is denoted _ para - l_. this mode of speech makes sense not only for l but for any classical complexity class , and we refer to @xcite for the corresponding theory .
e.g. fpt is the class of parameterized problems which are in p after a pre - computation .
a _ parameterized reduction _ from a parameterized problem @xmath389 to another @xmath402 is a function @xmath403 such that there is a computable @xmath404 such that for all @xmath396 we have for @xmath405 that @xmath406 and : @xmath407 .
if there is a comptable @xmath408 such that @xmath409 is computable in space @xmath410 ( on a turing machine with write - only output tape ) , then we speak of a @xmath411_-reduction_. in descriptive complexity one considers classical problems as isomorphism closed classes of ( finite ) structures of some fixed vocabulary . in the parameterized
setting we are led to consider the slices of parameterized problems as such classes of structures .
[ def : paramprobl ] a _ parameterized problem _ is a subset @xmath412 such that for every @xmath413 there is a vocabulary @xmath414 such that the _ @xmath268-the slice of @xmath389 _ , i.e. @xmath415 , is an isomorphism closed class of @xmath414-structures . if there is @xmath416 such that @xmath417 for all @xmath418 , we say that @xmath389 has bounded arity.s are assumed to be pairwise equal and only _ ordered _ structures are considered . ]
this definition is not in conflict with the mode of speech above if one views binary strings as structures in the usual way .
flum and grohe @xcite transferred capturing results ( cf .
* chapter 7 ) ) of classical descriptive complexity to the parameterized setting via the concept of _ slicewise definability_. for example , a parameterized problem @xmath389 is _ slicewise fo - definable _ if there exists a computable function @xmath105 mapping every @xmath413 to a first - order sentence @xmath419 defining the @xmath268-th slice of @xmath389 ( cf .
if one additionally assumes that there is a constant @xmath420 bounding the quantifier rank of all @xmath421 then it follows by standard means that slicewise fo - definable probelms are in para - l .
in fact , in @xcite flum and grohe gave a characterization of para - l along these lines via slicewise definability in deterministic transitive closure logic .
many parameterized classes could be characterized in this way @xcite . in this section
we study the complexity of the parameterized homomorphism problems associated to classes of structures @xmath0 : @xmath422 here , @xmath423 is a natural number coding the structure @xmath2 in some natural way .
the goal of this section is to show that the complexities of homomorphism problems are captured in a strong sense by the hierarchy from section [ sec : hierarchy ] .
we start by carefully defining a very weak notion of reduction .
central to descriptive complexity are first - order reductions which take a structure @xmath2 to the structure @xmath424 where @xmath425 is a first - order interpretation ( see e.g. ( * ? ? ?
* chapter 12.3 ) ) .
we recall the definition ( see e.g. ( * ? ? ?
* chapter 11.2 ) ) .
[ df : interpetation ] let @xmath426 be ( finite , relational ) vocabularies and @xmath427 be a unary relation symbol outside @xmath428 .
an _ interpretation ( of @xmath428 in @xmath40 ) _ is a sequence @xmath429 of @xmath40-formulas such that there exists @xmath420 such that for all @xmath430 we have @xmath431 where every @xmath432 is a tuple of @xmath23 variables .
the number @xmath23 is the _ dimension _ of @xmath425 .
the vocabularies @xmath40 and @xmath428 are the _ input _ and _ output vocabulary of _ @xmath425 , respectively .
an interpretation is _ quantifier - free _ if all its formulas are .
the @xmath428-structure @xmath424 is defined only if @xmath2 is a @xmath40-structure such that @xmath433 . in this case
it has universe @xmath434 and interprets @xmath435 by @xmath436 ; here we view @xmath437 as a subset of @xmath438 .
the following is well - known and easy to see ( see e.g. ( * ? ? ?
* exercises 12.3.2 , 11.2.4 ) ) .
[ lem : int ] there is a computable function mapping a pair of interpretations @xmath439 to an interpretation @xmath440 such that for every structure @xmath2 with @xmath424 and @xmath441 both defined , also @xmath442 is defined and @xmath443 .
if both @xmath444 are quantifier - free , then so is @xmath440 . if @xmath425 has dimension @xmath23 and @xmath445 has dimension @xmath446 , then @xmath440 has dimension @xmath447 .
we need to agree upon a way how to consider pairs of structures as a single structure .
here is a straightforward way to do this : given a pair @xmath448 of a @xmath40-structure @xmath2 and a @xmath428-structure @xmath3 , define the structure @xmath449 by taking the disjoint union of @xmath2 and @xmath3 and interpreting two new unary relation symbols @xmath450 and @xmath451 by the ( copies of the ) universes of @xmath2 and @xmath3 respectively .
naturally here , the disjoint union of @xmath2 and @xmath3 has universe @xmath452 and interprets @xmath453 by @xmath454 where @xmath455 if @xmath456 and else @xmath457 ; @xmath458 is defined analogously . for @xmath459 many structures
@xmath460 we inductively set @xmath461 .
it is well - known that np contains problems that are complete under quantifier - free reductions , i.e. reductions computed by a quantifier - free interpretation @xmath425 as above .
dawar and he @xcite transferred the notions to the parameterized setting and asked whether central completeness results for the classes of the w - hierarchy exhibit a similar robustness .
more precisely , dawar and he defined a parameterized reduction @xmath462 from @xmath389 to @xmath402 to be _ slicewise quantifier - free definable _ if there exists @xmath463 and a computable function @xmath105 that maps every @xmath413 to some quantifier - free interpretation @xmath419 each of whose formulas has at most @xmath23 variables such that @xmath464 ; here , one views the pair @xmath465 in some suitable way as a single structure .
[ df : qfred ] let @xmath466 be parameterized problems ( definition [ def : paramprobl ] )
. for @xmath413 let @xmath414 denote the vocabulary of the @xmath268-th slice of @xmath389 .
a parameterized reduction @xmath462 from @xmath389 to @xmath402 is _ quantifier - free after a pre - computation _ if there are @xmath463 and computable functions * @xmath467 * @xmath468 * @xmath105 mapping @xmath413 to a quantifier - free interpretation @xmath419 of dimension @xmath23 , such that for all @xmath469\times{\mathbb{n}}$ ] the structure @xmath470 is defined and @xmath471 .
we write @xmath472 to indicate that such a reduction exists , and we write @xmath473 to indicate that both @xmath472 and @xmath474 .
note the new _
p_arameter @xmath475 is computed by @xmath476 from @xmath268 alone , @xmath104 is the pre - computation providing an _
a_uxiliary structure , and @xmath105 provides the _ d_efinition of the new structure @xmath477 . to compare definition [ df : qfred ] with dawar and he s notion , observe that if @xmath402 has arity bounded by @xmath462 , then the formulas in @xmath419 have at most @xmath478 many variables .
we shall only be interested in problems satisfying this assumption .
has size @xmath479 for @xmath462 the arity of the @xmath475-th slice of @xmath402 . ]
[ lem : plred ] let @xmath466 be parameterized problems and assume @xmath402 has bounded arity .
if @xmath472 , then @xmath480 .
let @xmath462 be quantifier - free after a pre - computation and choose @xmath481 witnessing this .
given @xmath465 we need to explain how to compute @xmath482 in parameterized logarithmic space .
note the computation of @xmath483 and @xmath419 requires an amount of space that depends on the parameter @xmath268 only .
it thus suffices to show that @xmath477 can be computed from @xmath484 and @xmath419 in parameterized logarithmic space . to compute @xmath477
one needs to check for every @xmath485 from @xmath419 and for each length @xmath486 tuple @xmath61 from the universe of @xmath484 whether @xmath61 satisfies @xmath57 in @xmath484 .
this can be done in space @xmath487 if @xmath243 is the size of the universe of @xmath484 : note that if @xmath23 is the dimension of @xmath419 and @xmath462 bounds the arity of @xmath402 , then @xmath49 has at most @xmath488 many variables @xmath489 , and this is an absolute constant .
[ lem : trans ] let @xmath490 be parameterized problems . if @xmath472 and @xmath491 , then @xmath492 .
let @xmath481 and @xmath493 witness @xmath472 and @xmath491 respectively .
we define @xmath494 witnessing @xmath492 as follows .
we set @xmath495 and @xmath496 .
to define @xmath497 write @xmath498 , and @xmath499 .
we have to explain how to compute from @xmath444 a ( quantifier - free ) interpretation @xmath440 such that @xmath500 .
let @xmath23 be the dimension of @xmath501 .
we construct an interpretation @xmath502 of dimension @xmath23 such that @xmath503 .
let @xmath504 be a quantifier - free formula defining the universe of ( the copy of ) @xmath505 in @xmath506 .
write @xmath507 and @xmath508 for @xmath509 and set @xmath510}\neg \psi(x_i))\vee(\psi(x_1)\wedge \bigwedge_{i\in[w]}x_i = x_1),\\ \psi_{p_1}(\bar x)&:=&\psi(x_1),\\ \psi_{p_2}(\bar x)&:=&\neg\psi(x_1),\\ \psi_r(\bar x_1,\ldots,\bar x_{{\mathrm{ar}}(r)})&:=&\textstyle(\varphi_r(\bar x_1,\ldots,\bar x_{{\mathrm{ar}}(r)})\wedge\bigwedge_{j\in[{\mathrm{ar}}(r ) ] } \neg\psi(x_{j1}))\vee ( rx_{11}x_{21}\cdot x_{{\mathrm{ar}}(r)1}\wedge\bigwedge_{j\in[{\mathrm{ar}}(r)]}\psi(x_{j1})),\end{aligned}\ ] ] for @xmath38 appearing in both , the output vocabulary of @xmath425 and the vocabulary of @xmath505 ; if @xmath38 does not appear in the former ( latter ) , we omit the former ( latter ) disjunct in @xmath511 .
then @xmath512 is as required . from @xmath513
compute @xmath440 according to lemma [ lem : int ] . then @xmath514 .
[ [ convention ] ] convention + + + + + + + + + + for technical reasons we need to consider homomorphism problems @xmath1 also for classes @xmath0 which are not necessarily decidable .
in such a case we slightly abuse notation and write @xmath515 for a parameterized problem @xmath389 to mean that there are _ partially _ computable functions @xmath516 whose domain contains @xmath517 such that for all @xmath518 and similar @xmath3 we have that @xmath519 is defined and : @xmath520 let @xmath521 be computably enumerable classes of graphs . in this subsection we show that the associated homomorphism problems @xmath13 and @xmath15 are @xmath522-equivalent if the graph classes @xmath10 and @xmath14 are @xmath203-equivalent : [ theo : dechom ] if @xmath17 , then @xmath523 choose @xmath23 such that graphs in @xmath10 have @xmath14-deconstructions of width at most @xmath23 . using that @xmath14 is computably enumerable , it is not hard to see that there is an algorithm that computes given @xmath524 a graph @xmath525 and an @xmath19-deconstruction @xmath165 of @xmath18 of width at most @xmath23 .
given an instance @xmath526 of @xmath13 with @xmath524 and @xmath3 similar to @xmath527 the reduction outputs @xmath528 where @xmath525 is as above and @xmath529 is defined as follows .
assume first that all bags @xmath530 are nonempty .
for @xmath531 let @xmath532 denote the set of pairs @xmath533 such that @xmath534 is a partial homomorphism from @xmath527 to @xmath3 . for each @xmath535
choose a tuple @xmath536 that lists the elements of @xmath22 . define @xmath537 we claim that @xmath538 if @xmath408 is a homomorphism from @xmath527 to @xmath3 , then @xmath539 is a homomorphism from @xmath540 to @xmath529 .
conversely , suppose that @xmath541 is a homomorphism from @xmath540 to @xmath529 .
if @xmath541 maps @xmath535 to @xmath542 , let @xmath543 be the map @xmath544 .
any two such maps @xmath543 and @xmath545 are compatible in the sense that they agree on arguments on which they are both defined : indeed , if @xmath546 then @xmath547 and @xmath548 , so there is a path in @xmath19 from @xmath72 to @xmath549 ; if @xmath550 then there exists neighbors @xmath551 on this path such that @xmath552 ; then @xmath553 is not a function , and in particular @xmath554 ; as @xmath555 this contradicts @xmath541 being a homomorphism . therefore and since every @xmath556 appears in some @xmath22 , @xmath557 is a function from @xmath120 to @xmath41 . to verify
it is a homomorphism we show that it preserves @xmath86 ; that is preserves the colours @xmath558 can be seen similary .
so , given an edge @xmath559 we have to show @xmath560 .
choose @xmath561 such that @xmath562 .
then @xmath563 is a partial homomorphism from @xmath527 to @xmath3 : this is clear if @xmath564 ; otherwise @xmath565 , so @xmath566 and is follows by definition of @xmath567 that @xmath568 is a partial homomorphism . but @xmath568 is defined on @xmath569 , so @xmath568 and hence @xmath408 maps @xmath117 to an edge in @xmath570 .
we are left to show that there is a quantifier - free interpretation producing @xmath529 from @xmath571 for some structure @xmath572 computable from @xmath527 .
for @xmath572 we take the expansion of @xmath527 that interprets for every @xmath535 a @xmath23-ary relation symbol @xmath22 by @xmath573 . for @xmath23-tuples of variables @xmath574 and @xmath575 consider the formula @xmath576 } p_1x_i\wedge \bigwedge_{i\in[w ] } p_2y_i\wedge\bigwedge_{(i , i')\in[w]^2 } ( x_i = x_{i'}\to y_i = y_{i ' } ) \wedge\bigwedge_{(i , i')\in[w]^2 } ( ex_ix_{i'}\to ey_iy_{i'})\ ] ] and let @xmath577 be similarly defined for @xmath578-tuples
. set @xmath579 this is a quantifier - free interpretation @xmath425 of dimension @xmath578 such that @xmath580 .
this finishes the proof for the case that the bags @xmath530 are nonempty . in the general case
we can assume that any @xmath22 is nonempty whenever there exists @xmath581 in the connected component of @xmath72 in @xmath19 such that @xmath582 .
thus we can assume @xmath19 is the disjoint union of @xmath583 and @xmath584 such that @xmath585 for all @xmath586 and @xmath587 for all @xmath588 .
as seen above we get a @xmath529 such that @xmath589 define @xmath590 as follows using a new vertex @xmath591 : @xmath592 it is straightforward to check that @xmath529 can be defined by a quantifier - free interpretation as above : e.g. as formula @xmath593 one may now take @xmath594}(c_{g_0}x_i\wedge c_{g_0}y_i),\ ] ] for some fixed vertex @xmath595 . furthermore , it is easy to see that @xmath596 from and we see that @xmath597 is a reduction from @xmath13 to @xmath15 as desired .
let @xmath0 be a computably enumerable class of structures .
we show : [ theo : coregraphhom ] suppose @xmath0 has bounded arity .
let @xmath10 denote the class of graphs from the hierarchy ( [ eq : hierarchy ] ) such that @xmath598
. then @xmath599
thus , the complexity of @xmath1 is determined in a strong sense by the level @xmath600 takes in our hierarchy .
for example , because the reductions are weaker than fpt - reductions it is the level @xmath600 takes in our hierarchy what determines whether @xmath1 is w[1]-complete or fixed - parameter tractable ( cf .
@xcite ) , and because it is weaker than pl - reductions ( cf .
@xcite ) , this level determines e.g. whether @xmath1 is in para - l or path ( cf .
@xcite ) .
we devide the proof into a sequence of lemmas .
the first two of them build on two lemmas from @xcite .
recall that the _ direct product _
@xmath601 of two @xmath40-structures @xmath2 and @xmath3 has universe @xmath602 and interprets @xmath95 by @xmath603 suppose @xmath0 is a class of cores .
then @xmath604 let @xmath605 with @xmath606 be an instance of @xmath607 .
let @xmath608 be the restriction of @xmath3 to the vocabulary @xmath40 of @xmath2 .
set @xmath609 using that @xmath0 is a class of cores , it is proved in ( * ? ? ? * lemma 3.9 ) that @xmath610 to see that the reduction @xmath611 is quantifier - free after a pre - computation , we define @xmath481 witnessing this .
set @xmath612 and @xmath613 .
recall the interpretation @xmath614 is meant to produce @xmath529 from @xmath615 .
it has dimension @xmath162 and is given by @xmath616 [ cor : core ] @xmath617 by lemma [ lem : trans ] it suffices to establish the following sequence of reductions : @xmath618 to see @xmath619 , map an instance @xmath620 with @xmath606 to @xmath621 where @xmath529 expands @xmath3 interpeting @xmath622 by @xmath623 .
it is easy to see that this is quantifier - free after a pre - computation .
that @xmath624 follows from the previous lemma . finally , to see @xmath625 we use ( here and only here ) that @xmath0 is computably enumerable
: this ensures that there exists a partially computable function @xmath476 such that for all @xmath626 we have that @xmath627 is the code of a structure in @xmath0 whose core is @xmath2 ; the reduction then maps @xmath620 to @xmath628 .
again it is clear that this is quantifier - free after a pre - computation .
@xmath629 because @xmath0 is computably enumerable there is a partially computable function @xmath476 that maps any @xmath630 with @xmath631 to ( the code of ) a structure @xmath80 with @xmath606 and @xmath632 ; in particular , @xmath633 and we write @xmath634 . from a further given structure @xmath3
similar to @xmath527 define the following structure @xmath529 : @xmath635:\textup { if } a_i\neq a_j , \textup { then } ( b_i , b_j)\in e^{{\mathbf{b } } } \big\},\end{aligned}\ ] ] for @xmath38 in the vocabulary @xmath40 of @xmath2 .
this structure is produced from @xmath615 by the following interpretation @xmath425 of dimension 2 : @xmath636 ^ 2}(x_i= x_j\vee ey_iy_j),\end{aligned}\ ] ] for @xmath95 .
clearly , @xmath425 is computable from @xmath630 . setting @xmath637 and @xmath638 we get a reduction as desired provided @xmath639
this has been proved in ( * ? ? ?
* lemma 3.8 ) .
next , we show a partial converse to the previous lemma .
[ lem : classtograph ] suppose @xmath0 has bounded arity .
then @xmath640 we consider two cases depending on whether or not @xmath0 has bounded treewidth .
[ [ case-1 ] ] case 1 + + + + + + suppose @xmath0 has unbounded treewidth
. then @xmath641 by the hierarchy theorem . by theorem [ theo : dechom ]
we get @xmath642 . since trivially @xmath643 for every class of graphs @xmath14 it suffices to show @xmath644 for some class of graphs @xmath14 .
in fact , we show that for every @xmath0 with bounded arity ( not necessarily of the form @xmath7 ) there exists a class of graphs @xmath14 such that @xmath645 .
let @xmath40 be a vocabulary and @xmath2 be a @xmath40-structure .
we define the following graph @xmath646 , reminiscent of the incidence graph .
its universe has `` left '' vertices @xmath647\big\}\ ] ] together with `` right '' vertices @xmath648 ; for notational simplicity we assume that @xmath649 .
its edges @xmath650 are also devided into two kinds , namely we have edges `` on the left '' between @xmath651 and @xmath652 for @xmath653 together with `` left to right '' edges between @xmath654 and @xmath655 . given @xmath606 and a structure @xmath3 ,
say both of vocabulary @xmath40 , define the following structure @xmath529 .
it expands @xmath656 to a structure interpreting the language of @xmath657 .
namely , @xmath658\}\subseteq l({\mathbf{b } } ) , \qquad\textup { for } ( r,\bar a , i)\in l({\mathbf{a}}),\\ c_{a}^{{\mathbf{b}}'}&:= & r({\mathbf{b}})=b , \qquad\textup { for } a\in r({\mathbf{a}}).\end{aligned}\ ] ] we claim that @xmath659 to see the forward direction , assume @xmath72 is a homomorphism from @xmath2 to @xmath3
. define @xmath660 setting @xmath661 for @xmath662 and @xmath663 for @xmath651 .
note that indeed @xmath664 because @xmath665 as @xmath72 is a homomorphism .
it also follows that @xmath549 preserves the colours @xmath666 ; that it preserves the colours @xmath82 is clear .
it is also clear that it preserves edges `` on the left '' . finally consider a `` left to right '' edge , say between @xmath654 and @xmath655 . then @xmath667 and this is in @xmath668 .
conversely , let @xmath549 be a homomorphism from @xmath657 to @xmath529 and let @xmath72 be its restriction to @xmath669 .
since @xmath549 preserves the colours @xmath622 the function @xmath72 takes values in @xmath670 .
we claim @xmath72 is a homomorphism from @xmath2 to @xmath3 .
let @xmath671 for @xmath95 .
for each @xmath672 $ ] there exists a tuple @xmath673 such that @xmath674 because @xmath549 preserves the colour @xmath666 .
but , in fact , @xmath675 does not depend on @xmath676 : if there would be @xmath677 $ ] such that @xmath678 , then @xmath549 would not preserve the edge `` on the left '' between @xmath679 and @xmath680 .
it now suffices to show that @xmath681 , that is , @xmath682 for all @xmath672 $ ] ( where we write @xmath683 ) . if this would fail for @xmath672 $ ] , then @xmath684 while @xmath685 , contradicting that @xmath549 is a homomorphism .
thus , @xmath686 defines a parameterized reduction from @xmath1 to @xmath687 where @xmath688 is a class of graphs .
we are left to show that this reduction is quantifier - free after a pre - computation .
it is here where we are going to use the assumption that @xmath0 has bounded arity .
it suffices to show that some isomorphic copy of @xmath529 can be produced from @xmath689 by a suitable interpretation @xmath425 where @xmath477 is some auxiliary structure .
it will be clear that @xmath425 and @xmath477 are computable from @xmath423 .
the dimension of @xmath425 is @xmath690 where @xmath691 bounds the arity of @xmath0 .
if @xmath2 has vocabulary @xmath40 , the auxiliary structure @xmath477 is @xmath692))^*$ ] , i.e. take the @xmath73-structure with universe @xmath693)$ ] and add colours for the elements .
the interpretation produces the isomorphic copy of @xmath529 where @xmath694 is replaced by @xmath695 where @xmath696 is a @xmath697-tuple of @xmath698s .
similarly , @xmath699 is replaced by @xmath700 with an @xmath462-tuple of @xmath698s .
recall , @xmath529 has vocabulary @xmath701 . letting @xmath58 and @xmath702 be @xmath462-tuples of variables ,
the interpretation @xmath425 reads as follows .
abbreviate for @xmath703 $ ] @xmath704\setminus[{\mathrm{ar}}(r)]}(p_1x_j\wedge c_0x_j),\\ \textit{right}(z\bar x)&:=&\textstyle p_2z\wedge \bigwedge_{j\in[r]}(p_1x_j\wedge c_0x_j),\end{aligned}\ ] ] and define @xmath705 } \textit{left}_{r , i}(z\bar x)\vee \textit{right}(z\bar x),\\ \varphi_{c_{(r,\bar a , i)}}(z\bar x)&:=&\textit{left}_{r , i}(z\bar x),\\ \varphi_{c_a}(z\bar x)&:=&\textit{right}(z\bar x),\\ \varphi_e(z\bar x , z'\bar x')&:= & \varphi'_{e}(z\bar x , z'\bar x')\vee \varphi'_{e}(z'\bar x',z\bar x),\end{aligned}\ ] ] where @xmath706}\big(\big(\textit{left}_{r , i}(z\bar x)\wedge \bigvee_{i'\in[r]\setminus\{i\}}\textit{left}_{r , i'}(z'\bar x')\wedge\bigwedge_{j\in[r]}x_j = x'_j\big)\\ & & \qquad\qquad\qquad\qquad\textstyle \ \vee\ \big(\textit{left}_{r , i}(z\bar x)\wedge\textit{right}(z'\bar x ' ) \wedge z'=x_i\big)\big).\end{aligned}\ ] ] [ [ case-2 ] ] case 2 + + + + + + suppose @xmath0 has bounded treewidth . then @xmath707 and by the hierarchy theorem there exists a class of forests @xmath14 such that @xmath708 .
moreover , we can assume that @xmath14 is computably enumerable and closed under deleting proper ( connected ) components since so are the classes of graphs listed in this theorem .
theorem [ theo : dechom ] implies that @xmath709 , so it suffices to show @xmath644 .
the proof of this is similar to the proof of theorem [ theo : dechom ] .
there is a @xmath23 such that for every @xmath606 there exists @xmath525 such that @xmath2 has an @xmath19-decomposition .
given an instance @xmath605 of @xmath607 with @xmath710 compute @xmath525 and a @xmath19-decomposition @xmath165 of @xmath2 of width @xmath23 .
as in the proof of theorem [ theo : dechom ] we can assume that each component of @xmath19 either has all its bags empty or all its bags nonempty .
we can even assume that all bags are nonempty - otherwise use the above closure assumption on @xmath14 and throw away all components with empty bags .
define @xmath529 as in the proof of theorem [ theo : dechom ] , now with @xmath80 replacing @xmath527 there .
we show @xmath711 the direction from left to right is seen as in the proof of theorem [ theo : dechom ] .
conversely , let @xmath541 be a homomorphism from @xmath540 to @xmath529 and let @xmath712 be a component of @xmath2 . choose a component , i.e. a tree , @xmath583 of @xmath19 such that @xmath713 is a tree decomposition of @xmath712
. for @xmath586 define @xmath714 as in the proof of theorem [ theo : dechom ] .
again we show that @xmath715 is a homomorphism from @xmath712 to @xmath3 . to see this ,
let @xmath38 be a relation symbol in the vocabulary of @xmath2 and @xmath716 .
then the components of @xmath61 form a clique in @xmath717 .
it is well - known for tree decompositions , that cliques are contained in some bag ( see e.g. ( * ? ? ?
* lemma 4 ) ) , that is , there is @xmath586 such that @xmath22 contains all components of @xmath61 . as @xmath714 is a partial homomorphism from @xmath712 to @xmath3 and contains @xmath61 in its domain
, we get @xmath718 as desired . that @xmath719 is quantifier - free after a pre - computation is seen as in the proof of theorem [ theo : dechom ] .
the last two lemmas imply : [ cor : graph ] suppose @xmath0 has bounded arity .
then @xmath720 ( theorem [ theo : coregraphhom ] ) note that with @xmath0 also @xmath721 is computably enumerable .
we thus get @xmath722 by corollaries [ cor : core ] and [ cor : graph ] , respectively .
this shows the first equivalence in theorem [ theo : coregraphhom ] .
the second follows from theorem [ theo : dechom ] .
a _ game vector _ is a tuple @xmath723 where @xmath724 .
we say @xmath725 has @xmath462 _ rounds _ and @xmath726}p_i$ ] _
pebbles_. let @xmath2 , @xmath3 be similar structures .
duplicator winning strategy for the @xmath725-game on @xmath727 _ is a sequence @xmath728 of partial homomorphisms from @xmath2 to @xmath3 satisfying the following conditions : * for each @xmath729 with @xmath730 , the set @xmath731 contains a mapping with domain @xmath44 .
* for each @xmath676 with @xmath732 , if @xmath733 , @xmath44 is a superset of @xmath734 with @xmath735 , then @xmath736 contains an extension @xmath119 of @xmath118 with domain @xmath44 .
we write @xmath737 to indicate that such a winning strategy exists .
[ prop : transitivity - of - strategy ] the @xmath738 relation is transitive , that is , when @xmath2 , @xmath3 , and @xmath739 are similar structures , if @xmath737 and @xmath740 , then @xmath741 .
let @xmath728 be a duplicator winning strategy for the @xmath725-game on @xmath727 , and let @xmath742 be a duplicator winning strategy for the @xmath725-game on @xmath743 .
for each @xmath744 $ ] , define @xmath745 to be the set @xmath746 .
it is straightforward to verify that @xmath747 is a duplicator winning strategy for the @xmath725-game on @xmath748 .
let @xmath749 be a game vector .
a _ @xmath725-decomposition _ of a structure @xmath2 is an @xmath19-decomposition @xmath20 such that : * @xmath19 is a rooted forest of height @xmath750 , * for each node @xmath72 of @xmath19 at level @xmath751 , it holds that @xmath752 , and * when @xmath150 and @xmath753 are at levels @xmath676 and @xmath754 , respectively , it holds that @xmath755 and @xmath756 . by the _ level _ of a node in a rooted forest
, we refer to the number of vertices in the unique path from the root ( of the node s component ) to the node .
so , roots are considered to be at level @xmath751 . [ prop : hom - gives - strategy ] if @xmath2 and @xmath3 are structures such that there exists a homomorphism @xmath72 from @xmath2 to @xmath3 , then for any game vector @xmath725 , there is a duplicator winning strategy for the @xmath725-game on @xmath727 . such a strategy is given by taking @xmath757 to be the set containing each restriction of @xmath72 to a subset @xmath729 with @xmath758 .
[ thm : strategy - to - hom ] let @xmath749 be a game vector , and let @xmath759 be similar structures and assume that @xmath107 has a @xmath725-decomposition .
if there exists a duplicator winning strategy for the @xmath725-game on @xmath760 , then there is a homomorphism from @xmath107 to @xmath3 .
let @xmath19 be a forest such that @xmath20 gives a @xmath725-decomposition of @xmath107 .
for each @xmath312 , we define a map @xmath761 from @xmath22 to @xmath3 that is a partial homomorphism from @xmath107 to @xmath3 , in the following inductive manner .
let @xmath728 be a duplicator winning strategy for the @xmath725-game . for each root @xmath305 of @xmath19
, define @xmath762 to be a map in @xmath731 that is defined on @xmath336 .
when node @xmath549 is the child ( in @xmath19 ) of a node @xmath72 at level @xmath676 having @xmath761 defined , we define @xmath763 to be a map in @xmath736 that is defined on @xmath764 and that extends @xmath761 ; such a map exists by the definition of duplicator winning strategy and by the definition of @xmath725-decomposition . for every element @xmath765 , by the definition of @xmath19-decomposition
, there exists an @xmath312 such that @xmath766 .
if for an element @xmath765 it holds that @xmath767 where @xmath72 is the parent of @xmath549 in @xmath19 , by the way in which we defined the mappings @xmath761 , @xmath763 , it holds that @xmath768 .
it follows by the connectivity condition of a decomposition that for any @xmath769 such that @xmath770 , one has @xmath768 .
hence the @xmath761 together naturally define a mapping @xmath408 from @xmath771 to @xmath41 , namely , the mapping given by @xmath772 .
this mapping @xmath408 is a homomorphism from @xmath107 to @xmath3 , since for any tuple of a relation of @xmath107 , its entries are contained in a bag @xmath22 , and @xmath408 extends @xmath761 , which is a partial homomorphism from @xmath107 to @xmath3 defined on @xmath22 . for each structure
@xmath2 and each game vector @xmath749 , we define the structure @xmath773 as follows .
let us say that a tuple @xmath774 is a _ set vector _
( of @xmath725 ) if @xmath775 , @xmath776 , @xmath777 , and for each @xmath778 $ ] , it holds that @xmath779 . let @xmath44 be the set of all set vectors of @xmath725 .
for a set vector @xmath780 of @xmath725 , when @xmath781 , define @xmath782 to be the smallest prefix of @xmath783 with @xmath784 ( equivalently , with @xmath785 ) , and define @xmath786 to be the set @xmath787 .
the universe @xmath788 of @xmath773 is @xmath789 , and every symbol @xmath38 from the vocabulary of @xmath2 is interpreted by @xmath790 we now present two basic facts on the structure @xmath773 .
[ prop : tva - has - v - decomp ] for every game vector @xmath725 and every structure @xmath2 , the structure @xmath773 has a @xmath725-decomposition .
we use the notation in the definition of @xmath773 above .
one can naturally define a rooted forest with vertices @xmath44 ( the set of all set vectors of @xmath725 ) by making two vectors @xmath791 adjacent if and only if @xmath792 extends @xmath783 by one entry ( or vice - versa ) ; the roots are the length @xmath751 elements of @xmath44 . with respect to this rooted forest
, @xmath793 gives the desired @xmath725-decomposition .
[ prop : hom - from - tva - to - a ] for each game vector @xmath725 and each structure @xmath2 , there is a homomorphism from @xmath773 to @xmath2 . the homomorphism is the projection onto the first coordinate , that is , the mapping that sends an element @xmath794 of @xmath788 to @xmath104 .
[ thm : char - strategy ] let @xmath795 be similar structures with vocabulary @xmath40 , and let @xmath749 be a game vector .
the following are equivalent . 1 .
there exists a duplicator winning strategy for the @xmath725-game on @xmath727 .
2 . for each structure
@xmath107 on @xmath40 that has a @xmath725-decomposition , if there is a homomorphism from @xmath107 to @xmath2 , then there is a homomorphism from @xmath107 to @xmath3 .
3 . there exists a homomorphism from @xmath773 to @xmath3 .
( 1 @xmath796 2 ) : we have that @xmath797 and that @xmath737 . by proposition [ prop :
hom - gives - strategy ] , we have @xmath798 . by proposition
[ prop : transitivity - of - strategy ] , we have @xmath799 . we obtain @xmath800 by theorem [ thm : strategy - to - hom ] .
( 2 @xmath796 3 ) : this is immediate from propositions [ prop : tva - has - v - decomp ] and [ prop : hom - from - tva - to - a ] .
( 3 @xmath796 1 ) : let @xmath72 be a homomorphism from @xmath773 to @xmath3 . for each @xmath744 $ ] , define @xmath801 where @xmath802 is the map defined on @xmath803 that takes @xmath804 to @xmath805 . by the definition of @xmath773 , each @xmath757 contains only partial homomorphisms .
we verify that the @xmath757 give a duplicator winning strategy , as follows .
first , when @xmath806 and @xmath807 , we have that @xmath731 contains a map defined on @xmath102 , via the set vector @xmath808 .
next , suppose that @xmath809 $ ] , that @xmath733 , and that @xmath102 is a superset of @xmath734 with @xmath810 .
by the definition of @xmath757 , there exists a set vector @xmath783 of length @xmath676 such that @xmath811 .
let @xmath792 be equal to the length @xmath812 set vector that extends @xmath783 with @xmath102 .
the mapping @xmath813 has domain @xmath102 , and extends @xmath814 because for each @xmath815 , it holds that @xmath816 .
let @xmath2 be a structure , and let @xmath725 be a game vector .
let @xmath817 denote the classical problem @xmath818 .
we say that _ the @xmath725-game solves _
@xmath817 if for every structure @xmath3 similar to @xmath2 , the existence of a duplicator winning strategy for the @xmath725-game on @xmath727 implies that there is a homomorphism from @xmath2 to @xmath3 . for a class of structures @xmath0 , we say that _ the @xmath725-game solves _
@xmath1 if , for each @xmath518 , the @xmath725-game solves @xmath817 .
[ thm : char - v - game - solves ] let @xmath2 be a structure , and let @xmath749 be a game vector .
the following are equivalent . 1 .
the @xmath725-game solves @xmath817 .
there exists a homomorphism from @xmath2 to @xmath773 .
3 . the structure @xmath79 has a @xmath725-decomposition .
( 1 @xmath796 2 ) : by the equivalence of 1 and 3 in theorem [ thm : char - strategy ] , and the fact that @xmath819 we have @xmath820 . then @xmath821 by the assumption that the @xmath725-game solves @xmath822 . ( 2 @xmath796 3 ) : if @xmath821 then there is a homomorphism @xmath72 from @xmath79 to @xmath773 . by proposition [ prop :
hom - from - tva - to - a ] @xmath823 , and clearly @xmath824 , so there is a a homomorphism @xmath549 from @xmath773 to @xmath79 .
it follows that @xmath825 is a homomorphism from @xmath79 to itself and thus has to be injective .
hence @xmath72 is injective . by proposition [ prop :
tva - has - v - decomp ] , we know that @xmath773 has a @xmath725-decomposition @xmath826 .
it is straightforward to verify that @xmath827 defined by @xmath828 is a @xmath725-decomposition of @xmath79 ; here , @xmath829 denotes the universe of @xmath79 .
( 3 @xmath796 1 ) : suppose that @xmath737 ; we want to show that @xmath71 . since @xmath830
, we have @xmath831 ( by proposition [ prop : hom - gives - strategy ] ) . by the transitivity of @xmath738 ( proposition [ prop : transitivity - of - strategy ] )
, we have @xmath832 , and by theorem [ thm : strategy - to - hom ] , we obtain @xmath833 . as @xmath824 , we may conclude that @xmath71 . [ thm : pebbles - td ] ( correspondence between number of pebbles and tree depth ) let @xmath0 be a set of structures , let @xmath10 denote @xmath600 , and let @xmath151 .
the following are equivalent . 1 .
there exists a game vector @xmath725 with @xmath25 pebbles such that the @xmath725-game solves @xmath1 .
2 . the @xmath834-game solves @xmath1 .
3 . the set @xmath10 has tree depth @xmath835 .
( 1 @xmath836 2 ) : the implication from 2 to 1 is immediate .
to prove that 1 implies 2 , let @xmath749 be a game vector with @xmath726 } p_i = n$ ] ; it suffices to show that if there is a duplicator winning strategy @xmath837 for the @xmath838-game , then there is a duplicator winning strategy for the @xmath725-game .
the sequence @xmath839 is straightforwardly verified to be a duplicator winning strategy for the @xmath725-game .
( 2 @xmath836 3 ) : by appeal to theorem [ thm : char - v - game - solves ] , it suffices to show that each graph in @xmath10 has a @xmath838-decomposition if and only if condition 3 holds . for the forward direction ,
let @xmath20 be a @xmath838-decomposition of a graph @xmath127 with respect to the rooted forest @xmath19 . by removing any vertex @xmath312 with @xmath840 and by contracting together adjacent vertices @xmath769 with @xmath841
, we obtain that each root of @xmath19 has @xmath842 and that each node @xmath549 having a parent @xmath72 satisfies @xmath843 and @xmath844 .
rename each root @xmath845 by the unique element in @xmath764 and rename each @xmath845 with a parent @xmath72 by the unique element that is in @xmath846 ; then , it holds that the new @xmath19 has height @xmath835 , and witnesses that the tree depth of @xmath18 is @xmath835 . for the backward direction ,
let @xmath107 be a rooted tree of height @xmath835 witnessing that a component @xmath102 of a graph @xmath847 has tree depth @xmath835 .
the component @xmath102 has a @xmath838-decomposition given by @xmath142 defined by @xmath848 ; this is straightforwardly verified . combining the @xmath838-decompositions of the components of @xmath18
, we obtain the desired decomposition of @xmath18 . note that in particular , theorem [ thm : pebbles - td ] implies that for a single structure @xmath2 , the number of pebbles needed to solve @xmath849 is equal to the tree depth plus one .
we now study the number of rounds needed to solve a problem @xmath1 .
note that this question has an asymptotic nature : in the case that @xmath0 consists of a single structure , one can solve @xmath1 in one round , that is , by the @xmath725-game where @xmath725 is the one - round vector @xmath850 .
( correspondence between number of rounds and stack depth ) let @xmath0 be a set of structures , let @xmath10 denote @xmath600 , and let @xmath851 .
the following are equivalent . 1 .
there exists a game vector @xmath725 with @xmath690 rounds such that the @xmath725-game solves @xmath1 .
2 . @xmath852 .
3 . the set @xmath10 has bounded tree depth and stack depth @xmath853 .
( 1 @xmath836 2 ) : by appeal to theorem [ thm : char - v - game - solves ] , it suffices to show that each graph in @xmath10 has a @xmath725-decomposition if and only if @xmath852 . for the forward direction , denote @xmath725 by @xmath854 .
we have by definition that a @xmath725-decomposition is a @xmath19-decomposition for a rooted forest @xmath19 of height @xmath853 .
also , it follows from definition that each bag @xmath22 has size @xmath855 .
thus @xmath10 has @xmath856-decompositions of bounded width , and by proposition [ prop : decomposition - vs - deconstruction ] , we obtain @xmath852 . for the backward direction , by proposition [ prop : decomposition - vs - deconstruction ] , we have that there exists @xmath217 such that @xmath10 has @xmath856-decompositions of width @xmath321
. when @xmath857 and @xmath20 is a @xmath19-decomposition of width @xmath321 for a graph @xmath18 , define @xmath858 by @xmath859 where the union is over all ancestors @xmath104 of @xmath72 .
it is readily seen that @xmath858 is a @xmath860-decomposition of @xmath18 .
( 2 @xmath836 3 ) : by proposition [ prop : bounded - characterizations ] , whenever condition 2 holds , @xmath10 has bounded tree depth .
hence , the desired equivalence follows from lemma [ lemma : stack - depth - placement ] .
for a class of structures @xmath0 consider the classical problem @xmath861 in this section we prove the following .
[ thm : logchar ] assume @xmath862 and let @xmath863 have bounded arity .
then the following are equivalent . 1 .
@xmath600 has bounded tree depth .
2 . @xmath864 .
this result characterizes the classical homomorphism problems in logarithmic space .
the characterization is conditional on the hypothesis @xmath862 , an hypothesis from parameterized complexity theory .
the class path has been discovered by elberfeld et al .
@xcite and is defined as follows . a parameterized problem @xmath865 is in if and only if there are a computable @xmath401 and a nondeterministic turing machine @xmath866 that accepts @xmath389 and for all inputs @xmath391 and all runs on it uses space @xmath867 and performs at most @xmath868 many nondeterministic steps .
one can argue that path is a natural and important parameterized class , e.g. there are some fundamental problems which turn out to be complete for path under @xmath869 .
the above characterization further underlines its importance .
we refer to @xcite for more information . here , we just mention the following result . [ prop : pathcomplete ] @xmath9 is complete for under @xmath869 .
[ lem : vgame - in - l ] for each game vector @xmath725 , the problem there exists a logarithmic space algorithm that , given as input a pair @xmath727 of similar structures , decides if there exists a duplicator winning strategy for the @xmath725-game .
write @xmath870 and @xmath871 } p_i$ ] .
for an instance @xmath448 of @xmath33 consider the following boolean formula in the variables @xmath872 for @xmath873 : @xmath874 further , consider the assignment evaluating @xmath875 by 1 or 0 depending on whether @xmath876\}$ ] is a partial homomorphism from @xmath2 to @xmath3 or not .
this assignment satisfies the formula if and only if @xmath737 .
both the formula and the assignment are computable from @xmath448 in logarithmic space , and so is the truth value .
( theorem [ thm : logchar ] ) ( 1 @xmath796 2 ) let @xmath877 bound the tree depth of @xmath600 . by theorem [ thm
: pebbles - td ] , the @xmath725-game solves @xmath33 where @xmath878 .
it follows that @xmath879 .
this problem is in l by lemma [ lem : vgame - in - l ] and the assumption @xmath863 .
( 1 @xmath796 2 ) first observe that @xmath880 : this follows from our assumption @xmath862 and proposition [ prop : pathcomplete ] because path is @xmath869-closed .
we assume that @xmath600 has unbounded tree depth and claim @xmath881 .
lemma [ lemma : placing - g ] implies that @xmath882 . by theorem [ theo : dechom ]
( and lemma [ lem : plred ] ) we get @xmath883 . but
the latter problem is @xmath884 by theorem [ theo : coregraphhom ] . since @xmath880 we conclude @xmath885 .
this clearly implies that the classical problem @xmath33 is not in l.
in this section we study the complexity of the parameterized model - checking problems associated with sets of ( first - order ) sentences @xmath886 , namely @xmath887 here , @xmath888 is a natural number coding in some straightforward sense the sentence @xmath49 .
existential _ sentence is one in which the quantifier @xmath889 does not occur and negation symbols @xmath890 appear only in front of atoms .
a _ primitive positive _ sentence is one built from atoms by means of @xmath891 and @xmath892 . for @xmath893 let @xmath894 $ ] and @xmath895}$ ] denote the sets of existential and , respectively , primitive positive sentences of quantifier rank at most @xmath896 where all appearing relation symbols have arity at most @xmath462 .
the proof proceeds by standard means defining a `` canonical query '' @xcite .
details follow . given a tree @xmath901 and @xmath902
we define a primitive positive formula @xmath903 such that for every structure @xmath2 similar to @xmath904 and every @xmath81 @xmath905 moreover , @xmath906 for @xmath72 the height of @xmath107 when rooted at @xmath462 .
we give the definition by induction on @xmath72 .
for @xmath907 the tree @xmath107 consists of one node @xmath462 , and we set @xmath908 . otherwise , let @xmath909 list the neighbors of @xmath462 in @xmath107 . for @xmath910 $ ] let @xmath911 denote the connected component of @xmath912 containing @xmath913
. then @xmath911 rooted at @xmath913 has height at most @xmath914 , so @xmath915 is defined and we can set @xmath916 } \exists y(exy\wedge \varphi_{{\mathbf{t}}_i , t_i}(y)).\ ] ] given an instance @xmath917 of @xmath918 , the forest @xmath108 is the disjoint union of , say , @xmath291 many trees @xmath919 . for @xmath920
$ ] choose a root @xmath921 witnessing that @xmath922 has height at most @xmath923 and set @xmath924}\exists x\varphi_{{\mathbf{t}}_i , r_i}(x)$ ] .
then @xmath925 is a reduction as desired .
given an instance @xmath927 of @xmath928)}$ ] , let @xmath428 denote the vocabulary of @xmath49 .
for each @xmath435 choose a new relation symbol @xmath929 of the same arity .
the reduction maps @xmath927 to @xmath930 where @xmath477 expands @xmath2 interpreting every new symbol @xmath929 by @xmath931 .
the sentence @xmath932 is obtained from @xmath49 by replacing every subformula of the form @xmath933 by @xmath934 , and then moving disjunctions out using the equivalence @xmath935 and de morgan rules .
this preserves the quantifier rank .
it is easy to see that @xmath936 is a reduction as desired .
given an instance @xmath939 of @xmath940})}$ ] we can assume that the existential quantifiers in @xmath49 quantify pairwise distinct variables .
following a known construction of @xcite define a structure @xmath941 interpreting the vocabulary @xmath428 of @xmath49 as follows .
its universe is the set of variables of @xmath49 and a relation @xmath435 is interpreted by those tuples @xmath489 such that the atom @xmath942 appears in @xmath49 .
then @xmath943 this shows @xmath940})}{\le_{\mathit{qfap}}}{p\textsc{-hom}({\mathcal{a}})}$ ] where @xmath944 } \big\}$ ] . note that @xmath0 is decidable and of bounded arity .
we claim that @xmath945 .
we argue similarly as in ( * ? ? ?
* theorem 3.12 ) . define a graph on the universe of @xmath941 , i.e. the variables of @xmath49 , by putting an edge between @xmath950 and @xmath951 if an only if @xmath952 and @xmath953 are consecutive quantifiers in the formula tree of @xmath49 .
this graph is a forest @xmath954 of height at most @xmath923 when we root each component by the first - quantified variable in it , i.e. the variable which is closest to the root in the formula tree of @xmath49 .
the closure of @xmath954 contains @xmath955 : if @xmath956 then there exists @xmath435 and a tuple @xmath489 of variables containing @xmath950 and @xmath951 such that @xmath942 appears in @xmath49 ; since @xmath49 is a sentence every variable in @xmath489 and especially @xmath950 and @xmath951 are quantified in @xmath49 , so @xmath952 and @xmath953 appear on some branch in the formula tree of @xmath49 ; this means there is a path from @xmath950 to @xmath951 in @xmath954 , so @xmath957 is in the closure of @xmath954 .
let an instance @xmath960 of @xmath961})}$ ] be given , where @xmath268 is a positive natural and @xmath962}$ ] for all @xmath963 $ ] .
for @xmath963 $ ] let @xmath964 be the output of the reduction of the previous lemma . then @xmath965 , so @xmath966 is a disjoint union of , say , @xmath967 many trees @xmath968 .
then @xmath969:{\mathbf{b}}\models\varphi_i\\\nonumber & \longleftrightarrow&\exists i\in[k ] : { \mathbf{f}}^*_i{\stackrel{h}{\to}}{\mathbf{b}}_i\\\nonumber & \longleftrightarrow&\exists i\in[k]\ \forall j\in[c_i ] : { \mathbf{t}}^*_{ij}{\stackrel{h}{\to}}{\mathbf{b}}_i\\\label{eq : skolemtuple } & \longleftrightarrow&\textstyle \forall\bar\jmath\in\prod_{i\in[k]}[c_i]\ \exists i\in[k]:{\mathbf{t}}^*_{ij_i}{\stackrel{h}{\to}}{\mathbf{b}}_i,\end{aligned}\ ] ] where we write @xmath970 .
we can assume that the @xmath971 s are pairwise disjoint and for each of them a root is chosen witnessing that it has height at most @xmath923 .
fix some @xmath972}[c_i]$ ] .
define the tree @xmath973 as follows : take the ( disjoint ) union of the trees @xmath974 and then merge their roots to a new node @xmath462 .
then @xmath462 witnesses that @xmath975 and all @xmath976,$ ] are subtrees of @xmath973 pairwise intersecting precisely in @xmath462 .
we define a structure @xmath977 such that @xmath978:{\mathbf{t}}^*_{ij_i}{\stackrel{h}{\to}}{\mathbf{b}}_i&\longleftrightarrow&{\mathbf{t}}^*_{\bar\jmath}{\stackrel{h}{\to}}{\mathbf{b}}^{\bar\jmath}.\end{aligned}\ ] ] the structure @xmath977 is the disjoint union of the structures @xmath979,$ ] defined next . to ensure ( [ eq : bigt ] ) it suffices that @xmath980 satisfies @xmath981 indeed , assuming ( [ eq : bibarj ] ) for every @xmath963 $ ] we derive ( [ eq : bigt ] ) as follows .
if @xmath982 for some @xmath676 , then @xmath983 for such an @xmath676 by ( [ eq : bibarj ] ) and , clearly , this implies @xmath984 .
conversely , assume @xmath984 .
since @xmath985 is connected , any homomorphism from @xmath985 to @xmath977 has image in some @xmath980 , i.e. @xmath986 for some @xmath676 ; for such an @xmath676 then ( [ eq : bibarj ] ) implies @xmath982 .
given @xmath987 it is not hard to define @xmath980 such that ( [ eq : bibarj ] ) is satisfied .
first forget all interpretations of symbols outside the vocabulary of @xmath985 : these are all colours @xmath988 for @xmath989 as well as @xmath990 where @xmath991 is the root chosen for @xmath992 . to get a structure in the vocabulary of @xmath985
, we add interpretations of @xmath988 for @xmath993 : all these @xmath988 are interpreted by @xmath994 . finally , we add loops to the elements in @xmath994 , i.e. set @xmath995 then ( [ eq : bibarj ] ) is straightforwardly verified , and thus we know ( [ eq : bigt ] ) .
finally , let @xmath108 be the disjoint union of the trees @xmath996}[c_i]$ ] , that is , make disjoint copies of the @xmath973 s and then take the union .
for every @xmath997 every node @xmath998 has a copy @xmath999 in @xmath1000 .
we adapt @xmath977 by renaming @xmath988 by @xmath1001 , more precisely , let @xmath1002 be the structure with universe @xmath1003 interpreting @xmath86 by @xmath1004 and @xmath1001 by @xmath1005 for @xmath998 .
this structure is the disjoint union of @xmath1006,$ ] where @xmath1007 is obtained by analoguous renaming applied to @xmath1008 .
then the disjoint union @xmath529 of the @xmath1009}[c_i],$ ] is similar to @xmath1010 , and @xmath1011}[c_i ] : { \mathbf{t}}^*_{\bar\jmath}{\stackrel{h}{\to}}{\mathbf{b}}^{\bar\jmath}\\ & \longleftrightarrow & { \mathbf{f}}^*{\stackrel{h}{\to}}{\mathbf{b } } ' . \end{aligned}\ ] ] the first equivalence follows from ( [ eq : bigt ] ) and ( [ eq : skolemtuple ] ) , and the second is trivial .
it is clear that @xmath1012 can be computed from @xmath1013 .
we show how to get a structure isomorphic to @xmath529 by a suitable interpretation from @xmath1014 where @xmath2 is a suitable auxiliary structure .
choose @xmath481 for the reduction used above , i.e. the one producing @xmath964 from @xmath1015 .
choose @xmath420 such that @xmath105 maps numbers to interpretations of dimension @xmath23 .
for @xmath963 $ ] write @xmath1016 and note @xmath1017 .
we define our auxiliary structure to be @xmath1018)^*\ , \
( \prod_{i\in[k]}[c_i])^*\ \big\rangle.\ ] ] note @xmath1019)$ ] is the @xmath73-structure with universe @xmath230 $ ] .
there is a quantifier - free formula @xmath1020 that defines in @xmath449 the union of ( the copies of ) the universes of @xmath1021 and the universe of @xmath3 let @xmath1022 be obtained from @xmath1023 by replacing each formula @xmath1024 in @xmath1023 by @xmath1025 . then @xmath1026 . in the following
we understand that @xmath676 ranges over @xmath230 $ ] and @xmath997 over @xmath1027}[c_i]$ ] .
it is easy to construct an interpretation of dimension 1 producing @xmath980 from @xmath987 .
combining this interpretation with @xmath1022 according to lemma [ lem : int ] gives a quantifier - free interpretation @xmath1028 of dimension @xmath23 such that @xmath1029 .
as @xmath980 is obtained by renaming colours we get @xmath1030 with @xmath1031 .
the structure we want to produce is the disjoint union of @xmath1032 .
let @xmath1033 be @xmath23-tuples of variables and let @xmath1034 and @xmath1035 be the formulas in @xmath1030 .
define an interpretation @xmath502 of dimension @xmath1036 as follows .
@xmath1037 where @xmath676 is the unique index such that @xmath1038
. then @xmath1039 is isomorphic to the disjoint union of the family @xmath1040 , i.e.@xmath1041 as desired .
( theorem [ theo : modcheck ] ) applying lemmas [ lem : oed ] , [ lem : mainmc ] and [ lem : canquery ] in row , we get @xmath898)}{\le_{\mathit{qfap}}}{p\textsc{-mc}({\textsc{dpp}^{q}[r ] } ) } { \le_{\mathit{qfap}}}{p\textsc{-hom}({\mathcal{f}}^*_{q-1})}{\le_{\mathit{qfap}}}{p\textsc{-mc}({\textsc{pp}^{q}[2]})}.\ ] ] noting @xmath1042})}\subseteq { p\textsc{-mc}(\sigma^q_1[r])}$ ] , theorem [ theo : modcheck ] follows .
a. atserias and m. weyer .
decidable relationships between consistency notions for constraint satisfaction problems . in _ proc .
18th annual conference of the european association for computer science logic ( csl ) _
, pages 102116 , 2009 .
hubie chen and vctor dalmau .
beyond hypertree width : decomposition methods without decompositions . in _ eleventh international conference on principles and practice of constraint programming _ ,
pages 167181 , 2005 .
nicole schweikardt , thomas schwentick , and luc segoufin .
database theory : query languages . in mikhail
j. atallah and marina blanton , editors , _ algorithms and theory of computation handbook _ , volume 2 : special topics and techniques , chapter 19 .
crc press , second edition , nov 2009 . | we study the problem of conjunctive query evaluation over sets of queries ; this problem is formulated here as the relational homomorphism problem over a set of structures @xmath0 , wherein each instance must be a pair of structures such that the first structure is an element of @xmath0 .
we present a comprehensive complexity classification of these problems , which strongly links graph - theoretic properties of @xmath0 to the complexity of the corresponding homomorphism problem .
in particular , we define a binary relation on graph classes and completely describe the resulting hierarchy given by this relation .
this binary relation is defined in terms of a notion which we call graph deconstruction and which is a variant of the well - known notion of tree decomposition .
we then use this graph hierarchy to infer a complexity hierarchy of homomorphism problems which is comprehensive up to a computationally very weak notion of reduction , namely , a parameterized form of quantifier - free reductions .
we obtain a significantly refined complexity classification of homomorphism problems , as well as a unifying , modular , and conceptually clean treatment of existing complexity classifications .
we also present and develop the theory of pebble games which solve the homomorphism problems where the cores of the structures in @xmath0 have bounded tree depth ; as a fruit of this development , we characterize the homomorphism problems solvable in logarithmic space .
we also use our framework to classify the complexity of model checking existential sentences having bounded quantifier rank .
[ section ] [ theorem]conjecture [ theorem]corollary [ theorem]example [ theorem]examples [ theorem]proposition [ theorem]proposition [ theorem]lemma [ theorem]definition [ theorem]remark |
the quantum dynamics of the radical - pair mechanism @xcite , underlying the avian magnetic compass @xcite and spin transport in photosynthetic reaction centers @xcite , have recently attracted the attention of the quantum physics community @xcite , since it was shown @xcite that radical - pairs offer an ideal system to study quantum coherence effects and explore quantum information processing in a complex biochemical setting .
radical - pair reactions consist of a coherent spin motion in a multi - spin system embedded in a biomolecule , interrupted by an electron transfer that results in the spin - dependent charge recombination of the radical - ion - pair and the termination of the reaction .
it is known that the coherent spin motion as well as the measurable reaction yields in radical - pair reactions are also influenced by the external magnetic field through the unpaired electrons zeeman interaction .
hence radical - pair reactions are no different than other quantum systems used to measure a classical parameter , as for example are the well - developed atomic magnetometers @xcite using e.g. alkali vapors @xcite or nitrogen vacancy centers @xcite .
central in these studies have been the fundamental measurement precision limits set by the quantum dynamics of the system under consideration .
we here establish a venue for studying quantum metrology in a biological context @xcite .
we introduce the full machinery of quantum parameter estimation @xcite in order ( i ) to establish the _ exact _ value of @xmath0 , the fundamental magnetic sensitivity of the reaction , and ( ii ) design an optimal molecular system approaching this fundamental limit . to this end
we consider the quantum fisher information obtained from the radical - pair reaction and the resulting cramr - rao bound .
we then treat the intra - molecule hyperfine couplings as free design parameters , and obtain their optimum value by maximizing the quantum fisher information .
this leads to the fundamental limit @xmath0 , which we explicitly derive for any radical - pair . knowing the absolute quantum limit on @xmath0 , we address the well - known measurement of reaction yields and show it is sub - optimal .
we then modify a recently proposed method of reaction control @xcite , introducing a quantum circuit analysis of the controlled reaction , and reducing @xmath0 by a factor of 3 compared to @xcite .
the outline of the paper is the following . in sec .
ii we briefly introduce the dynamics of radical - pair reactions , and in sec .
iii the basic tools of quantum metrology , in particular the analytic form of the parameter - generator , a useful tool recently introduced @xcite . in sec .
iv , the eigenvalues of this operator are then used to find the maximum quantum fisher information and the resulting bounds on @xmath0 for radical - pair reactions .
in sec . v we discuss a common observable in radical - pair reactions , the reaction yield , in the context of magnetic sensitivity .
we demonstrate that the resulting maximum possible sensitivity is an order of magnitude smaller than the absolute quantum limit . in sec .
vi we present the optimum measurement scheme that can realize the optimum quantum limit on @xmath0 .
since this scheme does not appear to be chemically realistic , a natural question is whether some sort of quantum reaction control can improve the magnetic sensitivity of reaction yields .
this is indeed the case as shown in sec .
vii , where we take advantage of the spin - exchange interaction naturally occurring in radical - pairs , and known from quantum metrology work to simulate a controlled - not gate .
the spin - exchange interaction effects a state - preparation and readout before and after the actual magnetometric state evolution , respectively , reminding of ramsey interferometry .
together with the reaction control method of @xcite , which is a factor of 6 away from the absolute quantum limit , our measurement scheme is shown to approach this limit within a factor of 2 .
radical - pairs ( rps ) are the cornerstone system of spin chemistry , the field of physical chemistry and photochemistry dealing with the effect of electron and nuclear spins on chemical reactions .
the radical - pair mechanism was introduced by closs and closs @xcite and by kaptein and oosterhoff @xcite as a reaction intermediate explaining anomalously large epr and nmr signals observed in organic molecule reactions in the 1960 s .
the quantum degrees of freedom of radical - ion pairs are formed by a multi - spin system embedded in a biomolecule .
the multi - spin system is comprised of the two unpaired electrons of the two radical - ions and a usually large number of nuclei .
their coherent spin motion is driven by intramolecule magnetic interactions , as for example hyperfine couplings between each radical s magnetic nuclei and the respective unpaired electron .
the magnetic field effects resulting from such interactions in this spin - dependent biochemical reaction have been extensively explored theoretically and experimentally @xcite .
in particular , a charge transfer following the photoexcitation of a donor - acceptor dyad da leads to the radical - pair ( also called charge - separated state ) @xmath1 , where the two dots represent the two unpaired electron spins of the two radicals .
the initial spin state of the two unpaired electrons of the radical - pair is usually a singlet , denoted by @xmath2 .
now , both d and a contain a number of magnetic nuclei which hyperfine - couple to the respective electron .
neither singlet - state nor triplet - state rps are eigenstates of the magnetic hamiltonian , @xmath3 , hence the initial formation of @xmath2 is followed by singlet - triplet ( s - t ) mixing , i.e. a coherent oscillation of the spin state of the electrons , designated by @xmath4 .
concomitantly , nuclear spins also precess , and hence the total electron / nuclear spin system undergoes a coherent spin motion driven by @xmath3 . as will be detailed later , the subscript @xmath5 in @xmath3 is a reminder that the hamiltonian depends parametrically on the magnetic field @xmath5 to be estimated .
this coherent spin motion has a finite lifetime .
charge recombination , i.e. charge transfer from a back to d , terminates the reaction and leads to the formation of the neutral reaction products , conserving during the process the electronic angular momentum .
that is , there are two kinds of neutral products , singlet ( the original da molecules ) and triplet , @xmath6da .
the percentage of the initial radical - pair population ending up in the singlet ( triplet ) neutral product defines the singlet ( triplet ) reaction yield .
singlet and triplet recombination takes place at the rate @xmath7 and @xmath8 , respectively .
both rates are in principle known parameters of the specific molecular system under consideration , as of course are the hyperfine couplings entering @xmath3 .
the whole process is schematically depicted in fig.[schematic]a . , which is coherently converted to the triplet radical - pair , @xmath9 , due to intramolecule magnetic interactions embodied in the spin hamiltonian @xmath3 .
simultaneously , spin - selective charge recombination leads to singlet ( da ) and triplet neutral products ( @xmath6da ) .
( b ) quantum metrology aspect of the radical - pair reaction , where an initial spin state is transformed into a final spin state , which depends on the magnetic field @xmath5 through the spin hamiltonian .
the measurement of the final state conveys information about @xmath5 . ] the schematic of fig.[schematic]a should not be taken too literally , as it suggests that only pure singlet ( triplet ) radical - pairs can recombine to the singlet ( triplet ) neutral reaction products .
this is not the case , like it is not the case for a two - level atom that only an excited - state atom can decay to the ground state . as in atoms having ground - excited state coherence , radical - pairs
can be in coherent superpositions of singlet and triplet states , continuously evolving by @xmath3 .
a major aspect of our previous work has been to understand the physics of this coherence , its fundamental dissipation properties , and the role thereof in establishing the fundamental master equation , @xmath10 , accounting for the radical - pair reaction s quantum dynamics .
this issue is still hotly debated @xcite .
this work , however , is decoupled from this debate , as we consider only the hamiltonian contribution to the quantum metrology aspect of the reaction .
the non - trivial quantum dynamics previously alluded to mainly appear in the case of unequal recombination rates , @xmath11 .
here we consider the simple exponential model , where @xmath12 and we further neglect the presence of s - t decoherence @xcite , which is unavoidable even in the case @xmath13 .
the rationale of this approach is that it allows a step - wise understanding of the quantum metrology aspect of radical - pair reactions , starting from the most evident features stemming from the coherent spin motion , and progressing to more complex properties of the system , which is an open and leaky quantum system .
earlier works @xcite attempted to explore some metrological aspects of these reactions , however using the traditional ( called haberkorn s ) master equation and considering phenomenological sources of decoherence .
our understanding is ( i ) that haberkorn s master equation scrambles the quantum dynamics of the system and is a phenomenological description valid only in the regime of strong spin relaxation , and ( ii ) consideration of decoherence and its role in this new kind of biochemical quantum metrology is a non - trivial task , intertwined with the understanding of the fundamental master equation .
therefore we opt to first establish the fundamental limits to the magnetic sensitivity of radical - pair reactions in the simple case of equal recombination rates and a purely coherent spin motion .
thus , the only effect of the finite radical - pair s lifetime @xmath14 relevant to this work is that the radical - pair population decays exponentially as @xmath15 .
hence the sensitivity limit @xmath0 @xcite will be shown to be directly dependent on @xmath16 .
this is not unexpected , since measurement time is a central resource in quantum metrology .
in other words , we here explore the equivalence of fig.[schematic]a with fig.[schematic]b , which describes the usual scheme of quantum estimating a classical parameter like a magnetic field .
an initial state @xmath17 evolves under the unitary action of the hamiltonian @xmath3 into @xmath18 . measuring the final state conveys information about @xmath5 . in radical - pair reactions , an initial spin state ( comprising
the spin of the two electrons and the present nuclei ) evolves under @xmath3 , and the recombination process effects the measurement in the singlet / triplet basis , i.e. the measurement stage is naturally inbuilt into the radical - pair mechanism .
we here treat radical - pair reactions as scalar magnetometers , and conclusively address the questions : ( a ) what is the fundamental quantum limit , @xmath0 , to the precision of estimating @xmath5 ?
( b ) is this limit realized when the physical observable carrying the information on @xmath5 is the reaction yield ?
( c ) if not , can we control the reaction in order to better approach the fundamental limit @xmath0 ? if we consider a radical - pair with @xmath19 nuclear spins in the donor and @xmath20 nuclear spins in the acceptor , the hyperfine hamiltonian in the presence of an external magnetic field @xmath21 , where @xmath5 is to be estimated , is _
b =- b(s_dz+s_az)+_j=1^n_d_d_j_j+_k=1^n_a_a_k_k.[ham ] here we denote by @xmath22 ( @xmath23 ) the electron spin of the donor ( acceptor ) radical , @xmath24 ( @xmath25 ) is the @xmath26-th ( @xmath27-th ) nuclear spin of the donor ( acceptor ) radical , and @xmath28 ( @xmath29 ) the hyperfine tensor coupling the @xmath26-th ( @xmath27-th ) nuclear spin of the donor ( acceptor ) radical to the donor s ( acceptor s ) electron . the gyromagnetic ratio of the electrons setting the frequency scale , @xmath30 , has been set to @xmath31 in eq . and
we will keep this convention from now on .
we also set @xmath32 , thus the hyperfine couplings and the magnetic field @xmath5 have units of frequency , while the spin operators are dimensionless . in the following , in order to get actual magnetic field values , one should divide the derived expressions for @xmath5 with @xmath33 . before embarking on our analysis
, we will first lay out the tools of quantum parameter estimation and apply them to two pedagogical cases , a single electron zeeman interaction and a two - electron zeeman interaction .
these considerations will form a baseline for comparing fundamental sensitivity limits in radical - pairs .
the formalism developed by brun and coworkers @xcite is ideally suited to treat the estimation of @xmath5 , which is not a multiplicative parameter of the radical - pair hamiltonian .
specifically , the authors in @xcite consider the time - evolution operator , @xmath34 , which obviously depends on the parameter @xmath5 we wish to estimate .
if the initial state of the system is @xmath35 , then the time - evolved state will be @xmath36 , where the subscript in @xmath37 reminds us that the time evolved state also depends on @xmath5 .
the generator of @xmath5 translations is then shown to be @xmath38 .
the utility of this generator is that it directly leads to the _ maximum _ quantum fisher information f_b^max=[_max(h_b)-_min(h_b)]^2 , where @xmath39 ( @xmath40 ) is the maximum ( minimum ) eigenvalue of @xmath41 .
the authors in @xcite then derive two general results .
first , knowing the @xmath42 different eigenvalues , @xmath43 , and corresponding eigenvectors , @xmath44 of @xmath3 , where @xmath45 , with @xmath46 being the degeneracy of eigenvalue @xmath43 , one can obtain @xmath47 from @xmath48 by @xmath49 we denote the projector to the @xmath27-th eigenspace of @xmath3 . secondly , the maximum fisher information is obtained for the initial state = 1(+e^i),[opt_init ] where @xmath50 and @xmath51 are the eigenkets of @xmath41 corresponding to its maximum and minimum eigenvalues , respectively . finally , the uncertainty @xmath0 in estimating @xmath5 is limited by the cramr - rao bound @xcite b,[dbfmax ] where @xmath52 is the number of independent repetitions ( number of radical - pairs in our case ) of the measurement . before proceeding with radical - pairs ,
we analyze two intuitive and simple examples .
consider first a single electron in a magnetic field @xmath53 , the hamiltonian being @xmath54 .
the eigenvectors and eigenvalues of @xmath55 are @xmath56 and @xmath57 , respectively . since the eigenvectors are @xmath5-independent , the second term in eq .
is zero , while the first term leads to @xmath58 .
the maximum and minimum eigenvalues of @xmath41 are @xmath59 and @xmath60 , respectively , hence @xmath61 .
thus , we recover the well - known time - scaling limit , namely the magnetic sensitivity resulting from measuring the electron s larmor frequency during a time interval @xmath62 is limited by @xmath63 .
repeating this measurement @xmath52 times , the so - called shot - noise limited sensitivity will be @xmath64 .
we will now demonstrate the particle - number scaling limit by considering a 4-dimensional hamiltonian consisting of the zeeman interaction of two electrons , @xmath65 .
now the eigenvalues are @xmath66 with corresponding eigenstate @xmath67 , @xmath68 with corresponding eigenstate @xmath69 , and @xmath70 , which is doubly degenerate , with corresponding eigenstates @xmath71 and @xmath72 . again
, the eigenstates are @xmath5-independent , hence @xmath73 .
the eigenvalues of @xmath41 are @xmath74 , hence now it is @xmath75 .
the magnetic sensitivity is now @xmath76 .
this is @xmath77 times better than repeating the one - electron measurement two times .
generalizing to an @xmath78-electron system , where @xmath79 , we find @xmath80 .
this is @xmath81 times better than repeating the one - electron measurement @xmath78 times .
this enhancement is due to a possible multi - partite entanglement in the @xmath78-electron state .
that is , a notable feature of @xmath82 is that it automatically takes into account of such a possibility in the system s state preparation .
it should be noted that the scaling with the particle number should not be confused with the scaling with the number @xmath52 of the experiment s repetition . since the experimental realizations are independent , the scaling with @xmath52 is the ordinary statistical scaling @xmath83 .
that is , repeating the 2-electron measurement @xmath52 times , the so - called heisenberg limited magnetic sensitivity will be @xmath84 .
similarly , in the @xmath78-electron system , the heisenberg - limited magnetic sensitivity obtained by averaging @xmath52 independent measurements will be @xmath85 .
as shown previously with the simple scenario of free electrons in a magnetic field , the sensitivity @xmath0 depends on the measurement time @xmath62 . in radical - pair reactions there is a natural time scale limiting the magnetic sensitivity , the radical - pair s lifetime .
this is determined by the recombination rates @xmath7 and @xmath8 .
for the reasons outlined in sec .
ii.a , we here consider the so - called exponential model , where @xmath12 .
when @xmath86 , the quantum dynamics of the radical - pair reaction simplify considerably .
this is because , neglecting singlet - triplet decoherence which we understand is inherent in the system ( even when @xmath13 ) , in the exponential model radical - pairs can be considered to evolve unitarily by the magnetic hamiltonian @xmath3 , while their population decays exponentially at the rate @xmath27 .
equivalently , at the single - molecule level , each radical - pair evolves unitarily until the random instant in time when it recombines .
this time follows the exponential distribution @xmath87 , where @xmath14 . since the quantum fisher information is time - dependent , we have to take into account the fact that from each radical - pair in the ensemble we can extract a different fisher information , depending on the time it recombined .
if @xmath88 is the radical - pair population at time @xmath62 and @xmath89 is the initial population , then in each time interval @xmath90 there will be @xmath91 radical - pairs contributing to @xmath82 . if the variance of the magnetic field estimate resulting from one molecule is @xmath92 , then the @xmath91 molecules contribute independently to the measurement during @xmath90 .
the inverse uncertainties of @xmath5 add in quadrature , hence the inverse variance stemming from those @xmath91 molecules will be @xmath93 . since @xmath94 , the magnetic sensitivity for the whole reaction is b=1.[delta_b_f ] clearly , we are not concerned with the absolute value of @xmath0 as determined by how many molecules participate in the experiment .
we rather focus on optimizing @xmath82 , which depends on the state preparation and measurement scheme .
thus , in the following we will take @xmath95 . for those cases where @xmath96 , where @xmath97 is some constant
, it follows that @xmath98
. we will first derive exact analytic results for @xmath82 and @xmath0 for a radical - pair with one nuclear spin-1/2 contained in e.g. the donor . the hyperfine hamiltonian is _ b =- b(s_dz+s_az)+a_xs_dxi_x+a_ys_dyi_y+a_zs_dzi_z.as it formally turns out , the maximum quantum fisher information does not depend on @xmath99 . intuitively , this is because the term @xmath100 just produces a shift in the magnetic field `` seen '' by the electron spin @xmath101 along the z - axis and hence does not `` produce '' any information on @xmath5 .
we therefore have to distinguish two cases : @xmath102 and @xmath103 . after dealing with the single - nuclear - spin radical - pair
, we generalize to multiple nuclear spins . for the spheroidal hyperfine coupling it
b =- b(s_dz+s_az)+as_dxi_x+as_dyi_y+as_dzi_z . a special case , occurring when @xmath104 , is the commonly encountered isotropic hyperfine coupling .
the eight eigenvalues of @xmath3 are given in appendix a , along with the eigenvalues of @xmath41 calculated from eq . .
it is found that the maximum and minimum eigenvalues of @xmath41 are @xmath62 and @xmath105 , respectively .
hence in this case , the maximum quantum fisher information is @xmath75 , leading to the quantum limit b_f=1.[deltab_f ] : the minimum uncertainty , @xmath0 , for determining a magnetic field @xmath5 by using a radical - ion - pair reaction , the single nuclear spin of the radical - pair having a spheroidal hyperfine coupling , is given by eq . .
it is worthwhile noting that the maximum fisher information , @xmath106 , is the same with the case of two free electrons studied in section iii.b .
one would perhaps expect that having three particles in the system ( two electrons and one nucleus ) , the optimal sensitivity should gain ( according to the heisenberg scaling ) a factor of 3 compared to the single - electron case , or a factor of 3/2 compared to the two - electron case .
the reason behind the absence of such enhancement is that the nuclear spin does not strongly couple to the magnetic field .
hence it does not provide any independent information on the magnetic field , but only serves to drive the time evolution of the radical - pair s electronic spin state .
the lack of enhancement by the nuclear spin is not because we omitted the nuclear zeeman interaction in the hamiltonian . indeed ,
if we include the nuclear zeeman term in @xmath3 , we find that @xmath107 , where @xmath108 is the nuclear gyromagnetic ratio ( scaled to @xmath33 ) .
thus the correction to @xmath109 is on the order of @xmath110 and hence negligible .
however , if it were @xmath111 , then we would get the expected factor of 3 in sensitivity gain compared to the single - electron case of sec .
iii.a . in other words ,
the information about the magnetic field essentially stems from the strength of the field s coupling to the spins .
we will now consider the general hyperfine coupling , where @xmath103 , _ b =- b(s_dz+s_az)+a_xs_dxi_x+a_ys_dyi_y+as_dzi_z .
again , we can find analytic expressions for the eigenvalues of @xmath41 , which are given in appendix b. there it is shown that @xmath112 and @xmath113 , hence the resulting maximum quantum fisher information is bound by @xmath106 , which we found previously for the spheroidal hyperfine coupling .
* we thus arrive at our second general result * : for a radical - pair with a single nuclear spin-1/2 , the spheroidal hyperfine coupling ( the isotropic being a special case ) leads to the smallest uncertainty , @xmath0 , for determining a magnetic field @xmath5 along the spheroid s symmetry axis .
this uncertainty depends only on the radical - pair s lifetime @xmath16 , and is given by @xmath114 . as a numerical estimate , for @xmath115 and @xmath116 radical - pairs
we obtain @xmath117 .
radical - pairs contain many ( sometimes tens ) of nuclear spins .
based on the above , we can readily generalize and state * the third general result of this work * : for any radical - pair with a spin - independent lifetime ( i.e. @xmath118 ) , the maximum magnetic sensitivity ( minimum @xmath0 ) that can be obtained with _ any measurement method and any initial state _ is @xmath119 .
this follows from the same physical argument used in sec .
iv.a , namely that the uncertainty @xmath0 is determined just by the two electron spins .
the nuclear spins do not couple to the external magnetic field , i.e. they are spectators just driving the spin state evolution .
a formal proof of this general result follows . for any operator
@xmath120 depending parametrically on @xmath5 , it is @xcite @xmath121 .
we take @xmath122 , calculate the above derivative with @xmath5 and multiply with @xmath123 in order to find @xmath124 .
taking the operator norm we get @xmath125 , hence indeed the maximum ( minimum ) eigenvalue of @xmath41 is smaller ( larger ) or equal than @xmath62 ( @xmath105 ) .
typically , when studying the magnetic sensitivity of radical - pair reactions , one considers the singlet reaction yield , which quantifies the percentage of the reactants ( number of radical - pairs starting out in the electronic singlet state at @xmath126 ) ending up in the singlet neutral product state . to define the singlet reaction yield ,
@xmath127 , we first need to introduce two basic operators , the singlet and triplet projectors , @xmath128 and @xmath129 , respectively . for a radical - pair with a single nuclear spin they are written as @xmath130 and @xmath131 .
they leave the nuclear spin state untouched and project out of a general state @xmath132 the electronic singlet or triplet component .
the expectation value of @xmath128 in the state @xmath132 is thus @xmath133 , hence the singlet reaction yield is written as @xmath134 , where @xmath135 .
it is obviously irrelevant whether one chooses to measure the singlet or the triplet reaction yield , since it is always @xmath136 , where @xmath137 .
now , since the magnetic field enters @xmath138 through the hamiltonian , the reaction yield is a function of @xmath5 . in particular , in order to find the magnetic sensitivity @xmath0 , we first need to distinguish two cases .
( a ) if one can measure the instantaneous singlet yield given by @xmath139 , one can estimate @xmath5 just from those radical - pairs that recombined into the singlet channel during @xmath90 through the relation @xmath140 , where @xmath141 is the variance of @xmath128 at time @xmath62 , and @xmath142 is the magnetic sensitivity of the instantaneous yield .
all such estimates can then be statistically combined ( inverse uncertainties add in quadrature ) to yield the total uncertainty b=^-1/2,[dbin ] where in the expression for the variance of @xmath128 appearing in the denominator of the integrand in eq .
we took into account that @xmath143 , since @xmath128 is a projector .
for this measurement scheme to be realistic , the time resolution of the measurement of the instantaneous yield must be much better than @xmath144 .
if this is not the case , we are led to case ( b ) integration over the whole reaction , i.e. measurement of the total yield @xmath127 .
then the magnetic sensitivity @xmath0 is given by @xmath145 , where @xmath146 is the precision with which @xmath127 is measured .
this is calculated as follows . in each time
step @xmath90 , the instantaneous yield , proportional to @xmath147 , is a random variable following a binomial distribution with probability @xmath147 .
thus , the total yield follows the sum of binomials having different probabilities , which is the poisson binomial distribution .
its variance is @xmath148 , hence b=^1/2.[dbav ] it is expected that the magnetic sensitivity of case ( b ) is smaller than case ( a ) , or equivalently @xmath149 , since in case ( a ) we have access to much more information along the reaction than the integrated yield relevant to case ( b ) .
nevertheless , we here opt to provide exact expressions for @xmath0 in the integrated case , as we think that this is most relevant for physiological conditions . for completeness
, we then report the corresponding sensitivities for case ( a ) .
we first consider an isotropic hyperfine hamiltonian , @xmath150 .
we calculate @xmath0 for initial state ( i ) @xmath151 , ( ii ) @xmath152 , and ( iii ) an equal mixture of ( i ) and ( ii ) , which is usually taken to describe the initial state of radical - pair reactions , as it accounts for thermal equilibrium ( practically zero ) nuclear spin polarization .
we denote the respective uncertainties by @xmath153 , @xmath154 and @xmath155 .
the analytic expressions for these uncertainties follow from the analytic expressions for the reaction yields @xmath156 , @xmath157 and @xmath158 and their derivatives with respect to @xmath5 entering the denominator of eq . , as well as from the analytic expressions for the respective nominators .
the resulting formulas are too cumbersome to list here . in appendix c
we provide for reference the exact expressions for the reaction yields .
we here use the obtained analytic expressions for the uncertainties @xmath0 to display their inverses as a function of the hamiltonian parameters @xmath5 and @xmath159 in the contour plots of fig.[db](a)-(c ) for the cases ( i)-(iii ) , respectively .
we first note that the minimum of @xmath154 ( see fig.[db]b ) is smaller by about 30% than the minimum of @xmath153 ( see fig.[db]a ) , and both minima appear at a finite ( and different in each case ) value of the hyperfine coupling @xmath159 and at a different field @xmath5 .
this is due to the different singlet - triplet mixing frequencies caused by the nuclear spin in the @xmath160 or in the @xmath161 state . in the @xmath161 state
the nuclear magnetic field opposes @xmath5 and hence reduces the mixing frequency , thus its @xmath5-dependence becomes relatively more significant . in case ( iii ) , shown in fig.[db]c , the minimum of @xmath155 is achieved for @xmath162 .
although the sensitivity @xmath163 is linear in the density matrix , the magnetic sensitivity @xmath0 depends on the absolute value of @xmath163 , hence @xmath155 is not trivially related to @xmath153 and @xmath154 .
for example , at low @xmath5 and @xmath159 where @xmath153 and @xmath154 are close to their minimum , the respective derivatives @xmath163 are opposite in sign , and this is why @xmath155 is large in this region . in any case , taking the limit of large @xmath159 we find the exact expression b_s^iso(b)^1/2.the minimum occurs at @xmath164 and takes the value b_s^iso=5.14.[dbs ] : for a radical - pair with one isotropically coupled nuclear spin , the maximum possible magnetic sensitivity obtained by measuring the time - integrated reaction yield is 15 times lower , @xmath165 , than the highest possible sensitivity allowed by quantum physics and given by eq . .
this means that there is ample room for improvement . by changing to an anisotropic hyperfine interaction
we can already get about a factor of 2 improvement in @xmath0 .
that is , we repeat the calculation for @xmath0 taking @xmath166 .
we find that @xmath0 is minimized either for @xmath167 and @xmath168 or for @xmath169 and @xmath170 .
for both cases the minimum is the same for both initial states ( i ) @xmath151 and ( ii ) @xmath152 , and hence the same for ( iii ) the mixed singlet initial state .
this is expected , since both pure initial states are symmetric with respect to the hamiltonian anisotropy .
we thus denote the uncertainty common to all three initial states ( i)-(iii ) by @xmath171 . as in the isotropic case ,
the resulting expressions are long . in appendix c
we provide for reference the reaction yield .
as shown in fig.[db]d , @xmath172 increases with increasing @xmath159 .
like before , we take the limit @xmath162 and find b_s^aniso(b)^1/2.the minimum occurs at @xmath173 , and takes the value b_s^aniso=2.27 , which is still a factor of 6.4 away from @xmath174 . * to summarize our fifth main result * : the measurement of the integrated reaction yield can at best provide 6.4 times worse magnetic sensitivity than the absolute quantum limit , and this is achieved for the maximally anisotropic hyperfine interaction .
the reason the anisotropic coupling outperforms the isotropic in the reaction yield magnetic sensitivity will be given in sec .
vi.a after we introduce the optimal measurement strategy .
furthermore , we stress that for a given magnetic field @xmath5 to be estimated , the optimum reaction - yield sensitivity @xmath0 is obtained for a particular lifetime of the radical - pair on the order of @xmath175 .
the reason will be given in sec .
vii.c . for completeness
, we produce in figs.[db]e , f the results of eq . , i.e. the case when our measurement time resolution is enough to monitor the instantaneous yield along the reaction .
in both cases studied , isotropic and anisotropic , this kind of measurement yields about a factor of 2 improvement in magnetic sensitivity .
specifically , we find @xmath176 and @xmath177 , obtained for @xmath162 at @xmath178 .
moreover , both minimums become broader , i.e. there is a larger range of @xmath5 values close to the optimal @xmath0 .
the usual measurement scheme of radical - pair reactions , namely the singlet initial state and the measurement of the singlet reaction yield , is enforced by the very nature of these reactions .
as shown in the previous section , this measurement is sub - optimal . towards a possible improvement in magnetic sensitivity
, we first need to point to the optimal initial state and the optimal measurement operator . according to the general result of eq . , the optimal initial state for a single - nuclear spin radical - pair is the greenberger - horne - zeilinger state @xmath179 .
clearly , @xmath17 belongs to the triplet manifold , and exhibits maximum tri - partite entanglement .
it is expected that by measuring the electronic spin precession of this state in the magnetic field one would obtain the optimum sensitivity . indeed , for the isotropic hyperfine hamiltonian , which we know already is optimal ( see section iv.a ) , the time - evolved state ( taking @xmath180 ) is @xmath181 .
we choose @xcite as measurement operator @xmath182 .
we will now analyze the two scenarios mentioned in section v , that of a time - resolved measurement and that of an integrated measurement . in the former case we get @xmath183 , hence @xmath184 .
now during @xmath90 there will be @xmath185 molecules contributing to this measurement of @xmath186 .
the resulting inverse variance in @xmath5 is @xmath187 , where @xmath188 is the variance of @xmath186 .
thus we find @xmath189 , which is exactly equal to the maximum quantum fisher information , leading to @xmath190 .
it thus follows that with this measurement strategy one achieves the limit @xmath174 at any @xmath5 .
in contrast , an integrated measurement is not as capable .
now the integrated @xmath186-``yield '' is @xmath191 , and its magnetic sensitivity is @xmath192 .
the square error in @xmath193 will be the integrated variance of @xmath186 , weighted by the exponential population decay , i.e. @xmath194^{1/2}=[8b^2/(16b^2+k^2)]^{1/2}$ ] .
finally , the magnetic sensitivity will be @xmath195^{1/2}$ ] .
it is seen that @xmath196 , with the equality sign valid only for @xmath197 .
that is , in the integrated measurement with the optimal initial state and optimal measurement operator we achieve the optimal sensitivity only at @xmath197 .
the optimum magnetic sensitivity follows from the optimal measurement strategy outlined before , choosing as initial state a maximally entangled state of the triplet electronic manifold , and measuring its spin coherence while it is evolving , always within the triplet manifold .
clearly , this is far from how radical - pairs evolve in reality .
this leads to a natural question that we will affirmatively address in the following section , i.e. can we control the reaction in a chemically and physically realistic way in order to approach the optimum magnetic sensitivity ? before addressing the previous question
, we will explain the fact that the maximally anisotropic hyperfine interaction gives a factor of 2 improvement in @xmath0 , as was demonstrated in sec .
this can be seen to result from the overlap of the state evolved by the magnetic hamiltonian @xmath3 , which is @xmath198 , with the optimal state @xmath199 previously defined . for the isotropic
hamiltonian the overlap is zero , while for the anisotropic hamiltonian it is @xmath200 .
finally , it might sound as contradicting that on the one hand we obtain the maximum fisher information for the isotropic case , while the maximum _ reaction - yield _ magnetic sensitivity for the anisotropic case .
the latter finding does not contradict the former , as the _ reaction - yield _ sensitivity limit is well below the quantum limit defined by the fisher information .
in section v we have rigorously proved that the singlet reaction yield with a maximally anisotropic hyperfine interaction can at most provide a magnetic sensitivity 6.4 times worse than the absolute quantum limit .
a natural question is , how can one do better ?
in particular , given the discussion of the previous section , how can one do better in a chemically realistic way ?
towards addressing this question we will ( i ) take advantage of a very promising approach of optically switching the conformation of the radical - pair , recently proposed in @xcite , and ( ii ) include a realistic exchange interaction in the hamiltonian , which changes ( a ) the initial spin state before the radical - pair commences its magnetometric state evolution , and ( b ) the effective measurement basis before it recombines . in summary ,
given the maximally anisotropic coupling that resulted from the optimization of section v.c , the reaction control proposed in @xcite , and the modified initial state and measurement basis we introduce in the following , we will show that the obtained sensitivity @xmath0 is just a factor 2 away from the quantum limit @xmath174 of eq . .
moreover , compared to the approach of @xcite , we reduce @xmath0 by a factor of 3 . ,
where @xmath201 , for a mixed singlet initial state @xmath202 , and ( a ) an isotropic and ( c ) a maximally anisotropic hyperfine interaction . in ( b ) and ( d ) we depict the corresponding probability per unit time , @xmath203 , for the molecular switches to close by the reaction control laser pulses , which are tuned to coincide with the positive swings of @xmath204 . in the middle of trace ( b )
there are a number of pulses missing , since there the corresponding @xmath204 is on average zero and hence will not contribute to the singlet yield magnetic sensitivity .
for all plots it was @xmath205 and @xmath206.,width=6 ] we briefly reiterate the method of @xcite , since the added advantage we introduce by the exchange hamiltonian is based on the same method of optically pulsing the conformation of the radical - pair .
in particular , the authors in @xcite suggest binding the donor and acceptor parts of the radical - pair to the two ends of a molecular switch , the conformation of which can be laser controlled .
schematically , this is shown in fig.[rc ] . the rationale behind this idea is the following .
as shown previously , the magnetic sensitivity depends on @xmath207 , where @xmath208 is the so - called singlet fidelity of the radical - pair state at time @xmath62 , and @xmath159 , @xmath5 the hyperfine coupling and the magnetic field . for reference ,
the functions @xmath204 are given in appendix d for the hamiltonians considered in this work . in fig.[pm]a
we plot an example of @xmath209 , which is seen to be symmetric about zero .
thus , when integrated with the exponential population decay , @xmath210 , and the lifetime @xmath144 is long enough to contain many positive and negative swings of @xmath211 , magnetic sensitivity is suppressed . the idea of @xcite is to pulse the conformation of the molecular switch by an external laser .
when the switch is open , the radical - pair evolves unitarily by the magnetic hamiltonian , which for later use we call @xmath212 .
for example , this would be either @xmath213 for the isotropic or @xmath214 for the anisotropic case . when the switch is closed , the authors in @xcite argue , the short distance between d and a will turn on the exchange interaction @xcite , @xmath215 . for large exchange coupling @xmath216 ,
pertinent to the small d - a separation at the closed switch position , the singlet and triplet energy levels separate by @xmath216 and singlet - triplet mixing is suppressed , so only recombination can take place . if the reaction control laser in turned on at those instances ( fig.[pm]b ) where @xmath211 is positive ( and does not have fast oscillations , as in the middle part of fig.[pm]a ) , then the reaction yield magnetic sensitivity will be enhanced , as demonstrated in @xcite .
taking advantage of the @xmath217 modulation that can be externally controlled by the reaction control laser pulses , we now analyze our measurement scheme approaching the absolute quantum limit @xmath174 . as shown in fig.[scheme ] , we first prepare the radical - pair state in the electron singlet state .
the nuclear spin is usually in an equal mixture of the states @xmath160 and @xmath161 . towards better exhibiting the connection of this biochemical reaction with quantum metrology ,
we take the quantum circuit perspective and depict the electron singlet state as produced from @xmath218 by a hadamard gate followed by a controlled - not gate . in radical - pairs , this state preparation is naturally realized by the electron transfer producing the charge separated state , since the precursor neutral molecule is already in the singlet state . at @xmath126
all molecular switches are in the `` closed '' conformation and the radical - pairs in the state @xmath219 , which describes a singlet state for the electrons and a mixed state for the nuclear spin .
now , while the authors in @xcite open the switch at this time , using a laser pulse strong enough to open all molecular switches , we wait for a time @xmath220 and act on the initial state with the hamiltonian @xmath221 . while the authors in @xcite consider an exchange coupling @xmath216 too large to allow any s - t mixing , we take @xmath216 to be a finite optimization parameter . we thus take @xmath222 .
the duration @xmath220 of the action of @xmath221 is a free parameter , however constrained by @xmath223 , so that the radical - pairs do nt have enough time to recombine through the singlet channel .
essentially , the action of @xmath221 for a time @xmath220 prepares the initial state of the radical - pair in a state other than @xmath35 . at time
@xmath224 a strong reaction control laser pulse opens all molecular switches , and the two radicals are now far apart , so that @xmath225 is a good approximation , given the exponential dependence of @xmath216 on inter - radical distance @xcite . from @xmath224 until @xmath226 the hamiltonian @xmath212 effects the singlet - triplet conversion forming the main magnetometric state evolution . at time @xmath226
a weak reaction control laser pulse closes some of the switches .
the pulse energy is chosen so that the rate of closing is equal to the radical - recombination rate @xmath27 .
this pulse is the first pulse shown in the pulse sequence of fig.[pm]d . now in our model , the hamiltonian @xmath221 will act again until the radical - pairs of those switches that closed recombine . in the model of @xcite ,
the radical - pairs just recombine at some time after the switches close without any state evolution taking place before recombination .
the radical - pairs of those switches that did not close in step 3 continue to evolve under @xmath212 .
step 3 is then repeated with the next weak reaction control pulse , and so on .
thus , the pulse repetition period is @xmath227 , which is the envelope period of the function @xmath209 shown in fig.[pm]c , while the pulse width is @xmath228 , so that only the positive swings of @xmath209 contribute to the yield s magnetic sensitivity .
hence for any given radical - pair , the time @xmath229 during which @xmath212 is acting is some odd multiple of @xmath228 , plus the time within the pulse , at which this radical - pair recombines .
the magnetic sensitivity resulting from the quantum circuit of fig.[scheme ] is shown in fig.[result ] in two equivalent ways . in fig.[result]a
we plot the yield sensitivity @xmath230 , in order to directly compare with the result of @xcite . in fig.[result]b
we plot the absolute value of @xmath0 , normalized to the optimum quantum limit @xmath114 .
it is evident that ( a ) the choice of the maximally anisotropic hamiltonian and ( b ) the inclusion of the action of the exchange interaction in @xmath221 leads to an enhancement by a factor of 3 compared to @xcite , and puts the scheme of fig.[scheme ] a factor of 2 away from the absolute quantum limit .
the factor of 3 is equally attributed to ( a ) and ( b ) .
the physical interpretation of the enhancement of the magnetic sensitivity by the exchange interaction is the initial phase difference between the singlet and triplet states resulting from the initial action of @xmath221 . due to this phase difference
, the action of @xmath212 fully transforms the @xmath231 into a @xmath232 state , thus sensitively affecting the results of the recombination measurement . without this phase , i.e. setting @xmath233 , the singlet and triplet states both have significant populations at the end of the circuit and dilute the magnetic sensitivity of the recombination products .
our quantum circuit scheme of fig.[scheme ] reminds of ramsey spectroscopy , where an initial @xmath234 pulse produces an atomic hyperfine coherence , which evolves under the clock transition hyperfine hamiltonian , and is refocused by the final @xmath234 pulse . to further clarify the workings of this quantum reaction control the following remarks are in order .
\(1 ) the pulse sequence of the reaction control laser shown in fig.[pm]d is synchronized with the positive swings of @xmath204 shown in fig.[pm]c .
this necessitates some prior ( and approximate ) knowledge of the magnetic field , a feature common with the reaction control scheme of @xcite .
\(2 ) the time interval @xmath220 during which @xmath221 acts before the molecular switch opens is taken @xmath235 , so that radical - pair recombination is negligible ( it actually increases the obtained @xmath0 by 5% ) .
after the switch closes , @xmath221 acts for time @xmath236 before the radical - pairs recombine .
this time is taken to follow the exponential distribution with parameter @xmath27 , i.e. @xmath221 acts for a time as long as the radical - pair takes to recombine on average , i.e. @xmath237 .
\(3 ) the inclusion of @xmath238 , which includes the exchange interaction was motivated by ( i ) other works @xcite , where a controlled - not gate is shown to be a crucial element in metrology , and ( ii ) the fact that the controlled - not gate is naturally realized by the exchange interaction , as analyzed in @xcite .
we let the exchange coupling @xmath216 be a free optimization parameter .
the minimum @xmath0 was found for @xmath239 . for a typical hyperfine coupling @xmath159 of several gauss ,
the resulting value of @xmath216 is also on the order of several gauss .
now , @xmath240 , where @xmath241 is the donor - acceptor distance , and typical values @xcite of @xmath242 and @xmath243 are @xmath244 and @xmath245 , respectively .
for @xmath216 to be on the order of several gauss , the distance @xmath241 in the closed position of the switch must be around 1.8 nm .
this is quite larger than the d - a distance of 0.5 nm in the closed position of azobenzene @xcite , proposed in @xcite as a molecular switch .
so for the reaction control studied here azobenzene is not an ideal candidate .
furthermore , in fig.[result]c we plot the minimum value of the obtained sensitivity , @xmath246 ( i.e. the minimum of the red solid trace of fig.[result]b ) as a function of the exchange coupling @xmath216 .
however , as the exchange coupling depends on inter - radical separation , which is modulated by molecular vibrations , in reality we have to average the trace of fig.[result]c .
indeed , evaluating @xmath240 around @xmath247 , and taking a variation of @xmath241 by 0.05 nm , which is typical for studies on the relaxation effect of @xmath216-modulation due to molecular vibrations @xcite , leads to a factor of 2 change in @xmath216 , similar to the @xmath216-range of fig.[result]c .
we thus obtain a final @xmath248 , i.e. 10% higher than the value for a constant ( and optimum ) @xmath216 .
\(4 ) further , there are two points that might cause a misunderstanding .
we first note that although the reaction control pulse sequence introduces a timing in the measurement of reaction yields , the measurement is not of the instantaneous type described in sec .
v.a , since we still measure an integrated yield , as does the scheme in @xcite .
secondly , the reader might argue that we use an exchange interaction , which was absent in the optimization presented in secs .
iv and v. however , the exchange interaction is used in @xmath221 , which is just a state - preparation process , changing the initial singlet state and the final measurement basis .
we thus engineer an initial state which is more optimal than @xmath35 , and the actual magnetometry takes place during the action of @xmath212 , which does not include any spin exchange . , and ( b ) error @xmath0 in the estimation of the magnetic field , normalized by the absolute quantum limit @xmath114 .
the black dashed line reproduces the result of the reaction control scheme of @xcite , while the red solid line is the result of this work .
our reaction control scheme approaches @xmath114 within a factor of 2 .
( c ) the minimum of the red solid line in ( b ) is plotted as a function of the exchange coupling @xmath216 . for @xmath249
, we obtain @xmath250 . but
nearby values of @xmath216 are induced by molecular vibrations , hence averaging trace ( c ) leads to the realistic uncertainty 2.2 times away from @xmath174.,width=8 ] finally , we elaborate on a subtle point regarding practical implementation . in section
v , evaluating the optimum sensitivity of the reaction yield in case of the maximally anisotropic coupling , we found @xmath171 to be 6.4 times away from the absolute quantum limit @xmath114 .
this optimum , however , is realized for a specific value of @xmath5 , e.g. @xmath251 for the anisotropic case , and a hyperfine coupling @xmath162 .
in other words , if one wants to realize the limit @xmath171 at e.g. earth s field , one needs to find a radical - pair having a lifetime @xmath252 .
we can now explain this earlier finding : because at that lifetime the reaction is almost complete during one ( positive or negative ) swing of the sensitivity function @xmath204 , and further swings do not suppress sensitivity .
now , it is evident by looking at fig.[result]b that the optimum sensitivity @xmath171 we obtain just by using the optimal rp lifetime ( i.e. without any reaction control ) is the same as the one achieved by the authors of @xcite using the reaction control , but taking @xmath253 , which is far from the magnetic field value at which @xmath171 is optimized .
this leads to the following statement summarizing our findings .
one can realize the optimum uncertainty @xmath0 at a desired magnetic field @xmath5 if it is possible to engineer a radical - pair with the specified lifetime and an anisotropic hyperfine coupling approaching the maximal anisotropy .
for example , the lifetime engineering could result from molecular bridges @xcite interleaving the donor and acceptor . on the other hand , if such experimental control of the radical - pair s lifetime is not possible , then the reaction control scheme of @xcite and its modification presented here offer a generally useful alternative .
in this work we introduced the tools of quantum metrology to put formal and fundamental limits to the magnetic sensitivity of radical - pair reactions , a class of spin - dependent biochemical reactions central in the field of spin chemistry and relevant to the avian compass mechanism . knowing what is the fundamental limit is crucial for understanding how successful a particular measurement scheme is , and for motivating the search for new measurement schemes if there is room for improvement . this has been shown to be the case with the reaction yield measurement , which we have shown to be sub - optimal by almost an order of magnitude .
we then took advantage of a recently proposed reaction control scheme , modified the scheme by inclusion of the exchange interaction along the lines of a quantum circuit and ramsey interferometry , and demonstrated a close approach to the absolute quantum limit .
regarding future work , we point to two venues of research naturally following from here .
a recurring discussion @xcite in the quantum dynamics of radical - pair reactions , in particular in relevance to the avian compass , is whether electron spin entanglement is a resource .
in other words , whether the initial singlet electron state , which is maximally entangled , and its subsequent evolution , more or less maintaining the initial entanglement , enhances whatever biological performance radical - pair reactions have . regarding the radical - pair magnetometer we have considered in this work , the answer is clear : considering a radical - pair with a spin - independent lifetime ( @xmath118 ) , and neglecting the intrinsic singlet - triplet decoherence mechanism we introduced @xcite , electron spin entanglement obviously helps _ in principle_. indeed , based on the discussion of sec .
iii.a and sec .
iii.b , for a system consisting of just two _ uncorrelated _ electron spins the optimum magnetic sensitivity is @xmath254 .
allowing quantum correlations one can in principle obtain a @xmath77 improvement , i.e. @xmath255 .
our reaction control scheme of fig.[scheme ] leads to @xmath256 , but this does not imply that entanglement `` does not help '' . in other words , it is not straightforward to arrive at a definitive statement with such comparisons . on the one hand
it is inconceivable how to experiment with two free electron spins in a chemical environment .
radical - pairs offer such a possibility .
similarly , there is no immediate way to controllably `` switch - off '' entanglement within the radical - pair reactions . put differently ,
even though the achieved sensitivity @xmath256 happens to be worse than the two - uncorrelated - spins case , further analysis is required to demonstrate whether or not ( or what part of ) @xmath256 is attributed to entanglement . moreover , according to our understanding @xcite , singlet - triplet decoherence is an unavoidable feature of the radical - pair mechanism itself , and in the case of equal recombination rates ( @xmath86 ) leads to a master equation for @xmath257 that reads @xmath258-k({{\rm q}_{\rm s}}\rho+\rho{{\rm q}_{\rm s}}-2{{\rm q}_{\rm s}}\rho{{\rm q}_{\rm s}})-k\rho$ ] . in other words , in this work we omitted the second term of the previous equation , firstly because its validity is not generally accepted and we wish to decouple this work from the relevant debate , secondly because omitting it considerably simplifies the calculations , and thirdly we obtain the sought after fundamental limits in the idealized and intuitive physical context of unitary evolution .
nevertheless , the role of decoherence in the magnetic sensitivity @xmath0 ought to be addressed in detail , as it is known that the advantage due to entangled states might deteriorate @xcite .
hence it remains an unsettled issue if entanglement is a resource for this kind of biochemical magnetometers .
a natural extension of this work is to study the fundamental limit @xmath259 in estimating the angle of the magnetic field with respect to a molecular frame of reference .
this is directly relevant to the avian compass function of radical - pair reactions , and the relevant study will be undertaken elsewhere . for the spheroidal hyperfine interaction @xmath260 considered in sec .
iv.a , the eigenvalues of @xmath3 are @xmath261 ( doubly degenerate ) , @xmath262 , @xmath263 and @xmath264 . taking care of the degeneracy in the calculation of @xmath41 ,
the eigenvalues of @xmath41 are found to be 0 ( doubly degenerate ) , @xmath265 , @xmath266^{1/2}/2(a^2+b^2)$ ] , and @xmath267^{1/2}/2(a^2+b^2)$ ] . by inspection
it is seen that @xmath268 , but due to the cosine term it is not immediately obvious how @xmath269 compares to @xmath270 .
we can prove that for all times @xmath271 .
indeed , take @xmath272 and @xmath273 and subtract from both the common term @xmath59 .
we need to show that @xmath274^{1/2}/2(a^2+b^2)$ ] is less than @xmath59 , or their ratio smaller than 1 .
the maximum value of the term involving the cosine occurs at @xmath275 , where @xmath276 .
then the maximum value of the ratio is @xmath277 for all @xmath42 .
thus the maximum and minimum eigenvalues of @xmath41 are @xmath62 and @xmath105 , respectively .
for the ellipsoidal hyperfine coupling discussed in sec .
iv.b , the eigenvalues of @xmath41 are found to be @xmath278 , @xmath279^{1/2}/[(a_x+a_y)^2 + 4b^2]$ ] , @xmath280^{1/2}/[(a_x+a_y)^2 + 4b^2]$ ] , and @xmath281^{1/2}/[(a_x - a_y)^2 + 4b^2]$ ] .
now it is less straightforward to find the maximum ( and similarly the minimum ) eigenvalue , as for some times @xmath272 is the maximum , while at other times it is @xmath273 .
however , we can prove as in appendix a that _ at any time _ the maximum eigenvalue is smaller or equal than @xmath62 , and similarly the minimum eigenvalue is larger or equal than @xmath105 .
hence the ellipsoidal case can not exceed the spheroidal @xmath82 .
for the hamiltonian @xmath282 , we calculate the singlet reaction yields @xmath283 and @xmath284 corresponding to the initial states @xmath151 and @xmath152 , respectively . for the hamiltonian @xmath285 , and for all three initial states considered before we find a common singlet reaction yield @xmath286 .
the results are the + ( - ) sign in the third term of corresponds to @xmath151 ( @xmath152 ) . taking the average @xmath287
, we reproduce the result of @xcite .
the sensitivities @xmath288 can be readily evaluated , but are too long expressions to list here . the magnetic field sensitivity of the singlet fidelity @xmath289 , where @xmath201 , is given ( after setting @xmath290 ) by the expressions , and for the isotropic hamiltonian @xmath150 , and initial states ( a ) @xmath151 , ( b ) @xmath152 and ( c ) an equal mixture of ( a ) and ( b ) , respectively .
for the maximally anisotropic hamiltonian @xmath285 , all three initial states produce the same expression for @xmath291 , given ( after setting @xmath292 ) by .
@xmath293\big[\alpha\sin({{(a+b)t}\over 2})+b\sin({{\alpha t}\over 2})\big]\label{e1}\\ g_{t}(a , b)&={a^2\over { 4\alpha^4}}\big[\alpha t\cos({{\alpha t}\over 2})-2\sin({{\alpha t}\over 2})\big]\big[\alpha\sin({{(a - b)t}\over 2})-b\sin({{\alpha t}\over 2})\big]\label{e2}\\ g_{t}(a , b)&=-{a^2\over { 4\alpha^4}}\big[\alpha t\cos({{\alpha t}\over 2})-2\sin({{\alpha t}\over 2})\big]\big[\alpha\cos({{at}\over 2})\sin({{bt}\over 2})+b\sin({{\alpha t}\over 2})\big]\label{e3}\\ g_{t}(a , b)&=-{a^2\over \beta^4}\sin({{bt}\over 2})\big[\beta t\cos({{\beta t}\over 4})-4\sin({{\beta t}\over
4})\big]\big[\beta\cos({{bt}\over 2})\cos({{\beta t}\over 4})+2b\sin({{bt}\over 2})\sin({{\beta t}\over 4})\big]\label{e4}\end{aligned}\ ] ] e. daviso et al . , _ the electronic structure of the primary electron donor of reaction centers of purple bacteria at atomic resolution as observed by photo - cidnp @xmath294c nmr _ , proc .
usa * 106 * , 22281 ( 2009 ) .
i. f. c@xmath295spedes - camacho and j. matysik , _ spin in photosynthetic electron transport _ , in _ the biophysics of photosynthesis _
goldbeck j , van der est a ( eds . ) , springer science + business media , new york ( 2014 ) .
xu , j. zou , j .- g . li and b. shao , _ estimating the hyperfine coupling parameters of the avian compass by comprehensively considering the available experimental results _ , phys .
e * 88 * , 032703 ( 2013 ) .
m. kritsotakis and i. k. kominis , _ retrodictive derivation of the radical - ion - pair master equation and monte carlo simulation with single - molecule quantum trajectories _ , phys .
e * 90 * , 042719 ( 2014 ) .
i. k. kominis , _ reply to the comment on `` quantum trajectory tests of radical - pair quantum dynamics in cidnp measurements of photosynthetic reaction centers '' by g. jeschke _ ,
lett . * 648 * , 204 ( 2016 ) . | radical - ion pairs and their reactions have triggered the study of quantum effects in biological systems .
this is because they exhibit a number of effects best understood within quantum information science , and at the same time are central in understanding the avian magnetic compass and the spin transport dynamics in photosynthetic reaction centers .
here we address radical - pair reactions from the perspective of quantum metrology .
since the coherent spin motion of radical - pairs is effected by an external magnetic field , these spin - dependent reactions essentially realize a biochemical magnetometer . using the quantum fisher information , we find the fundamental quantum limits to the magnetic sensitivity of radical - pair magnetometers . we then explore how well the usual measurement scheme considered in radical - pair reactions , the measurement of reaction yields , approaches the fundamental limits . in doing so , we find the optimal hyperfine interaction hamiltonian that leads to the best magnetic sensitivity as obtained from reaction yields .
this is still an order of magnitude smaller than the absolute quantum limit .
finally , we demonstrate that with a realistic quantum reaction control reminding of ramsey interferometry , here presented as a quantum circuit involving the spin - exchange interaction and a recently proposed molecular switch , we can approach the fundamental quantum limit within a factor of 2 .
this work opens the application of well - advanced quantum metrology methods to biological systems . |
the vortex state of cuprate superconductors is of central importance in understanding high-@xmath1 superconductivity ( htsc ) .
one the most well studied quantities is the internal magnetic field distribution @xmath7 associated with the vortex lattice ( vl).@xcite as discussed below , several methods can be used to measure @xmath7 , which depends on the london penetration depth @xmath8 , the coherence length @xmath9,@xcite and , to a lesser extent , the internal structure of the vortices,@xcite and non - linear and non - local effects.@xcite the form of @xmath7 has a distinctive asymmetric shape due to the spatial magnetic inhomogeneity characteristic of an ordered two - dimensional ( 2d ) lattice of vortices .
one basic feature in @xmath7 is a prominent high field tail associated with the vortex cores , which depends on the magnitude of @xmath9 .
there is also a saddle point in the local field profile located between two vortices .
this gives rise to a van hove singularity or sharp peak in @xmath7 below the average field .
the overall width or second moment of @xmath7 depends primarily on the london penetration depth @xmath8 , the lengthscale over which the magnetic field is screened .
anisotropy of the fermi surface or the superconducting order parameter can result in a different vl but the main features are similar for any ordered lattice.@xcite another general feature associated with any real vl is disorder arising from vortex pinning at structural defects and impurities.@xcite structural defects are present in all superconductors to some degree , but may be more prevalent in structurally complex compounds such as ybco .
for example , ybco s slightly orthorhombic structure facilitates crystal twinning , i.e. in a single crystal , there are generally domains with the nearly equal @xmath10 and @xmath11 directions interchanged . separating such twin domains
are well defined 45@xmath12 grain boundaries or twin boundaries which have been shown to be effective extended vortex pinning sites.@xcite scanning tunneling microscopy ( stm ) imaging of a twinned ybco crystal show that the areal vortex density is strongly modified by the twin boundaries.@xcite small - angle neutron scattering ( sans ) studies of ybco confirm that the twin boundaries strongly deform the vl.@xcite understanding the influence of such structural defects on the vl has been the subject of intense theoretical work,@xcite and is important for two main reasons .
firstly , it affects @xmath7 and thus adds uncertainty to measurements of fundamental quantities like @xmath8 and @xmath9 , since it can be difficult to isolate such extrinsic effects from changes in fundamental quantities of interest .
secondly , the degree of pinning of vortices determines the critical current density which is important for many applications.@xcite measurements of the vortex state field distribution @xmath7 are most often done using sans,@xcite nuclear magnetic resonance ( nmr),@xcite and conventional muon - spin rotation ( @xmath13sr).@xcite all these methods probe the vl in the bulk and can be applied over a wide range of magnetic fields .
it is also possible to probe the magnetic field distribution @xmath7 near the surface of the sample using low energy-@xmath13sr(le-@xmath13sr ) in low magnetic fields.@xcite recently we have demonstrated that similar information on @xmath7 near a surface can be obtained using @xmath0-nmr.@xcite this has the advantage that it can be applied over a wide range of magnetic fields . in this paper , we report measurements of the vl above the surface of the cuprate superconductor yba@xmath2cu@xmath3o@xmath4 using @xmath0-nmr.@xcite the @xmath14li@xmath6 beam was implanted into a thin silver overlayer evaporated onto several ybco samples . measuring in the ag allows one to isolate the contribution to @xmath7 from _ long wavelength disorder _ , i.e. disorder that occurs on length scales much longer than the vortex spacing and @xmath8 due to structural defects such as twin and grain boundaries .
this is possible because the field distribution broadening just outside the superconductor due to the vl inside has a very distinctive field dependence . in particular
, it vanishes in high magnetic fields where the vl spacing becomes less than the characteristic distance of the probe from the superconductor . on the other hand ,
long wavelength disorder has a much weaker dependence on magnetic field and dominates the observed @xmath7 in the high field limit .
our results show evidence for significant broadening of @xmath7 from such long wavelength disorder on the scale of @xmath15 , which is attributed to pinning at twin or other grain boundaries .
the magnitude of the broadening is similar to that observed in bulk @xmath13sr measurements , suggesting that the same broadening contributes to @xmath7 in bulk @xmath13sr measurements .
there is a crossover such that near @xmath1 , where @xmath16 , the broadening scales with the superfluid density , whereas at lower temperatures , where @xmath17 , the broadening does not track the superfluid density .
we discuss the consequences of this for the inference of @xmath18 from measurements of @xmath7 in polycrystalline superconductors .
the paper is organized as follows : section ii reviews the theory for the field distribution , and its second moment near the surface of a superconductor .
section iii contains all the experimental details . in section iv , we present the results . finally in section v we discuss the results and draw conclusions .
in a type ii superconductor , above the lower critical field @xmath19 , the magnetic field penetrates the sample _ inhomogeneously _ forming a lattice of magnetic vortices , each carrying a flux quantum , @xmath20 . in a perfect crystal ,
intervortex interactions lead to a long - range ordered 2d lattice of vortices , usually of triangular ( hexadic ) symmetry.@xcite at the core of each vortex ( a cylinder of radius approximately the superconductor s coherence length @xmath9 ) , the local magnetic field is maximal . outside the core
, concentric circulating supercurrents partially screen the field which thus falls exponentially with a lengthscale @xmath8 .
the average magnetic field in the vl is the applied field @xmath21 for flat samples , where demagnetization effects are negligible.@xcite at a given field , the average vortex spacing , i.e. the lattice constant of the vl , @xmath10 is fixed . for the triangular lattice
this is @xmath22 if one considers the profile of the magnetic field along a line in the lattice ( perpendicular to the direction of the applied field ) , it is thus _ corrugated _ with a period determined by @xmath10 .
this inhomogeneity in the magnetic field causes a characteristic broadening in local magnetic resonance probes such as the muons in @xmath13sr or the host nuclei in nmr . since the muon ( or host nuclear spin ) is at a well - defined lattice site(s )
, it samples the vl with a grid spacing given by the lattice constant of the crystal .
since this is much smaller than the vl constant , the resulting field distribution @xmath7 provides a random sampling of the spatially inhomogeneous field @xmath23 over the vl unit cell : @xmath24 where the integral is over a unit cell of the vl of area @xmath25 . in this paper
we are concerned with the @xmath26-component of the magnetic field @xmath27 ( parallel to the @xmath28-axis of ybco samples ) , and refer to it simply as @xmath29 .
for an ideal triangular vl , the spatial dependence of the @xmath26-component of the magnetic field in or outside a type ii superconductor follows the modified london equation , @xmath30 here @xmath31 when the applied field is along the @xmath28-axis and the screening supercurrents flow in the ab plane , @xmath32 is the 2d laplacian , @xmath33 for @xmath34 and zero otherwise , @xmath35 is a 2d vector in the xy plane , @xmath36 are the bravais lattice vectors for the vl .
we define the @xmath26 axis as the normal to the surface of a superconducting slab with negative @xmath26 outside the superconductor . the solution of eq .
( [ br - london - b1 ] ) is easily obtained using the fourier transform , @xmath37 , where the dimensionless fourier components , @xmath38 , are given by,@xcite @xmath39.}\ ] ] here @xmath40 , and @xmath41 $ ] are the reciprocal lattice vectors of the triangular vl , where @xmath42 .
a cutoff function @xmath43 , approximated by a simple gaussian @xmath44 , can be used to account for the finite size of the vortex core , where @xmath38 is replaced by @xmath45.@xcite however , the corrections due to @xmath43 are very small in our case , so it will be omitted . an approximate solution for the magnetic field along @xmath26 ( both inside and outside the superconductor )
is given by @xmath46 the second moment of @xmath7 at a depth @xmath26 , @xmath47 , where @xmath48 is the spatial average , is given by @xmath49 in an applied field of 52 mt at a distance 90 nm from the superconductor using eq .
( [ bzr_eq ] ) convoluted with a lorentzian of width @xmath50 .
( b ) the broadening of @xmath7 versus the applied field .
solid lines represent @xmath51 from eq .
( [ scm - vl ] ) in the bulk , and 40 and 90 nm away from the superconductor .
long dashed line shows @xmath51 at 90 nm from eq .
( [ scm - dvl ] ) for @xmath52 @xmath13 m and @xmath53 .
inset : sketch of a possible vortex arrangement including vortex trapping at twin boundaries spaced by @xmath54 and a regular triangular vortex lattice elsewhere .
( c ) @xmath55 ( normalized at @xmath56 ) , from eq .
( [ scm - dvl ] ) for @xmath57 mt , @xmath58 nm , and @xmath53 , is plotted against @xmath59 for different values of @xmath54 .
a @xmath60-wave temperature dependence of @xmath18 is used.@xcite in all figures @xmath61 nm.,title="fig : " ] in an applied field of 52 mt at a distance 90 nm from the superconductor using eq .
( [ bzr_eq ] ) convoluted with a lorentzian of width @xmath50 .
( b ) the broadening of @xmath7 versus the applied field .
solid lines represent @xmath51 from eq .
( [ scm - vl ] ) in the bulk , and 40 and 90 nm away from the superconductor .
long dashed line shows @xmath51 at 90 nm from eq .
( [ scm - dvl ] ) for @xmath52 @xmath13 m and @xmath53 .
inset : sketch of a possible vortex arrangement including vortex trapping at twin boundaries spaced by @xmath54 and a regular triangular vortex lattice elsewhere .
( c ) @xmath55 ( normalized at @xmath56 ) , from eq .
( [ scm - dvl ] ) for @xmath57 mt , @xmath58 nm , and @xmath53 , is plotted against @xmath59 for different values of @xmath54 .
a @xmath60-wave temperature dependence of @xmath18 is used.@xcite in all figures @xmath61 nm.,title="fig : " ] the field distribution for a perfectly ordered triangular vl calculated from eqs .
( [ pb - def ] ) and ( [ bzr_eq ] ) , at @xmath57 mt and outside the superconductor at @xmath58 nm and @xmath62 nm ( relevant to ybco at @xmath63 ) , is presented in fig .
[ simfig](a ) ( @xmath64 , defined below ) .
it shows the characteristic high field skewness with a cutoff corresponding to the field at the core of the vortices .
the sharp peak corresponds to the most probable field @xmath65 at saddle points in @xmath23 midway between adjacent vortices .
the low field cutoff occurs at the center of an elementary triangle of vortices . as we move farther from the superconductor , @xmath65 moves towards the applied field as the field approaches uniformity for @xmath66 , @xmath67 , @xmath68 .
this crossover occurs as @xmath69 , as the @xmath26 variation of the fourier components @xmath38 in eq .
( [ fourierb ] ) is controlled by @xmath70 which takes values equal to or larger than @xmath71 . however
, if instead we consider a simple non - superconducting overlayer instead of free space , then the limiting @xmath7 will be the intrinsic lineshape in the overlayer material .
as mentioned above @xmath7 is also affected by disorder in the vl due to pinning at structural defects in the crystal , where the superconducting order parameter is suppressed .
such disorder causes broadening of the magnetic resonance , obscuring the features expected from an ideal vl,@xcite adding uncertainty to parameters of interest such as @xmath8 and @xmath9 . relatively little
is known about the detailed characteristics of this disorder .
accounting for the disorder of the vl is most often done by smearing the ideal lineshape with a gaussian or lorentzian distribution of width @xmath50 , where the latter is a phenomenological measure of the degree of disorder.@xcite calculated distributions for an applied field @xmath72 mt are shown in fig . [ simfig](a ) for @xmath73 0.3 and 0.6 mt , together with the ideally ordered vl ( @xmath64 ) .
such disorder is more pronounced outside the superconductor and renders the lineshape symmetric when the depth dependent intrinsic vl broadening is smaller than @xmath50 .
one major difference between conventional @xmath13sr and @xmath0-nmr or le-@xmath13sr is the stopping range of the probe . in conventional @xmath13sr
the @xmath74 stopping range is @xmath75 mg/@xmath76 , yielding a fraction of a mm in ybco .
in contrast , in @xmath0-nmr or le-@xmath13sr , the mean depth of the probe can be controlled on a nm lengthscale from the surface . for implantation depths inside the superconductor , comparable or larger than @xmath8 ,
the @xmath13sr lineshape ( proportional to @xmath7 ) is nearly field independent for @xmath77 ( for @xmath78 ) , and the second moment of @xmath7 follows the formula,@xcite @xmath79 neglecting the cutoff field .
using the latter makes @xmath51 slightly field - dependent , but the corrections are small for fields @xmath80 . outside the superconductor ,
the magnetic field inhomogeneity of the vl vanishes over a lengthscale that depends on the spacing between vortices , @xmath10 .
in particular , the recovery to a uniform field occurs on a lengthscale of @xmath81.@xcite the field distribution is thus strongly field dependent when @xmath82 is of the order of @xmath83 .
this is shown in fig .
[ simfig](b ) , where @xmath51 due to the vl given in eq .
( [ scm - vl ] ) is plotted against the applied field at a distance of 90 nm and 40 nm above the surface . in low magnetic fields ,
the magnetic resonance lineshape outside the superconductor is sensitive to both the intrinsic inhomogeneity of the vl as well as any additional broadening from disorder .
however , in high magnetic fields the linewidth is dominated by vl disorder .
taking the view that the dominant source of disorder is due to twin or grain boundaries,@xcite one can model the effect of disorder on the regular vl in different ways .
the simplest is to assume that , in addition to the regular triangular lattice , a fraction of vortices is trapped along the structural defects such as twin or grain boundaries as shown in the inset of fig .
[ simfig](b ) .
the local field in real space will be the superposition of both contributions @xmath84 where @xmath85 is the field due to the regular vl , @xmath86 the field due to the vortices pinned by disorder and @xmath87 is some generally incommensurate wave vector related to the pinning , which for simplicity we take to be of the form @xmath88 , where @xmath89 is the spacing between vortices in a boundary , and @xmath54 is the separation between boundaries as drawn in the inset of fig . [
simfig](b ) .
the average field is then @xmath90 , where @xmath91 and @xmath92 is the fraction of pinned vortices ( @xmath93 ) .
the second moment of @xmath94 from eq .
( [ bz - dis ] ) can be then easily calculated : @xmath95.\ ] ] it is clear from eqs .
( [ fourierb ] ) and ( [ scm - dvl ] ) , that the broadening from the vl ( first term ) at a distance @xmath26 outside the superconductor becomes small at high magnetic fields where @xmath96 .
however the broadening outside the superconductor due to disorder ( second term ) remains large provided @xmath83 is not much larger than @xmath97 .
since @xmath54 and @xmath92 depend on the arrangement of twin boundaries we expect them to be sample dependent .
in addition , one may also anticipate that @xmath92 will decrease at high magnetic fields where the increased repulsive interaction between vortices overcomes vortex - pinning .
therefore , we assume a simplified phenomenological parameterization @xmath98 , where @xmath99 is temperature and sample dependent and @xmath100 .
the broadening of the field distribution due to a regular vl can be significantly larger when introducing the effect of disorder due to the twin and grain boundaries . when taking the disorder into account , @xmath51 of eq .
( [ scm - dvl ] ) is no longer zero at high magnetic fields as seen in fig .
[ simfig](b ) .
this is because the broadening has a disorder component which decays on a length scale of @xmath54 rather than @xmath10 , where @xmath101 ( we also assume @xmath102 and thus ignore the effect of the spacing within the twin boundaries ) .
consequently @xmath51 shows a strong deviation from the ideal vl result as seen in fig .
[ simfig]-c as @xmath54 increases . in this case
the second moment from eq .
( [ scm - dvl ] ) no longer scales with @xmath103 as predicted for an ideal vl ( see eq .
( [ scm - lambda ] ) ) .
in particular , at low @xmath104 , the broadening is almost @xmath104-independent irrespective of superconducting gap structure .
it is interesting to note that the first @xmath13sr studies on powder samples of cuprates showed a very flat variation in the linewidth.@xcite this was taken as evidence for @xmath105-wave superconductivity .
later measurements on high quality crystals of yba@xmath2cu@xmath3o@xmath4 showed a much different low temperature behaviour,@xcite and , in particular , a linear variation in @xmath106 consistent with @xmath60-wave pairing.@xcite although the lineshapes in powders are expected to be more symmetric than in crystals due to the additional disorder and random orientation , the different temperature dependence is surprising since it was thought that the line broadening from disorder should also scale with @xmath106.@xcite the current work provides a clear explanation for the discrepancy between powders and crystals . in powders ,
the line broadening is dominated by long wavelength pinning of vortices at grain boundaries .
consequently the resulting broadening at low temperature reflects variations in the vortex density and is thus only weakly dependent on temperature . in later work on crystals ,
the contribution from such long wavelength pinning is much less important .
this is evident from bulk @xmath13sr in crystals where one observes the expected characteristic lineshape associated with a vl.@xcite
the measurements were carried out on three different near - optimally doped ybco samples , two flux - grown single crystals and a thin film .
i ) the twinned single crystal in the form of a platelet @xmath107 mm thick with an area @xmath108 mm@xmath109 had @xmath110 k. it was mechanically polished with 0.05 @xmath13 m alumina , then chemically etched with a dilute ( 0.8% ) bromine solution followed by annealing at 200@xmath12 c in dry n@xmath111 to improve the surface quality .
it was then sputter coated with a 120 nm thick ag film ( 99.99 % purity ) at room temperature in an ar pressure of 30 mtorr .
the deposition rate was 0.5 / s , and to ensure ag uniformity , the crystal was rotated .
ii ) the optimally doped detwinned single crystal had @xmath110 k , @xmath107 mm thickness , and area @xmath112 mm@xmath109 .
the crystal was cleaned , annealed , and mechanically detwinned . a 120 nm thick ag , from the same source as above ,
was sputtered onto the prepared surface under similar conditions .
iii ) the film of @xmath113 87.5 k , critical current density @xmath114 a/@xmath115 and 600 nm thickness , supplied by theva ( ismaning , germany ) , was grown by thermal co - evaporation on a laalo@xmath116 substrate of area 9@xmath1178 mm@xmath109 .
the film was coated _ in situ _ with a 60 nm silver layer ( 99.99@xmath118 purity ) .
the experiments were performed using the @xmath0-nmr spectrometer at the isac facility in triumf , canada , where a highly nuclear - spin - polarized beam ( intensity @xmath119 ions / s ) of @xmath5li@xmath6 is produced using collinear optical pumping with circularly polarized laser light.@xcite the beam is directed onto the sample which is mounted on the cold finger of a he flow cryostat and positioned in the centre of a high homogeneity superconducting solenoid .
the beamline and entire spectrometer are maintained in ultrahigh vacuum ( @xmath120 torr ) . in @xmath0-nmr measurements ,
the @xmath14li@xmath6 nuclear spin polarization is monitored via its asymmetric radioactive beta decay ( lifetime @xmath121 s ) , where the high energy ( several mev ) beta electron is emitted preferentially opposite to the nuclear spin direction .
the experimental asymmetry , defined as the ratio @xmath122 of the count rates in two plastic scintillation detectors placed in front ( f ) and at the back ( b ) of the sample , is proportional to the probe s spin polarization.@xcite li@xmath6 at energies of ( a ) 5 kev into 60 nm of ag with the mean at 40 nm away from ybco film , and ( b ) 8 kev into 120 nm of ag with the mean at 90 nm away from ybco crystals , as calculated via trim.sp.@xcite solid lines are phenomenological fits . ]
the whole spectrometer can be biased at high voltage , allowing one to tune the implantation energy of @xmath14li@xmath6 ions and their implantation depth between 5 - 200 nm .
therefore , the implanted @xmath14li@xmath6 can monitor the depth dependence of the local magnetic field distribution in materials at nm scale by measuring the nmr lineshape in a manner analogous to conventional nmr.@xcite in this work , the @xmath5li@xmath6 ions are decelerated to stop in the ag overlayer deposited on each of the three ybco samples .
implantation profiles of @xmath5li@xmath6 were calculated using the trim.sp code,@xcite examples of which are shown in fig .
[ trimfig ] .
the implantation energies used in this study ( 8 kev in the crystal samples and 5 kev the film ) , were tuned to stop all the @xmath14li@xmath6 within the ag .
the mean distances are 90 and 40 nm from the ag / ybco interface in the crystals and film , respectively .
the @xmath0-nmr measurement is carried out by monitoring the time averaged nuclear polarization through the beta decay asymmetry , as a function of the radio frequency ( rf ) @xmath123 of a small transverse oscillating magnetic field @xmath124 , where @xmath125 mt . when @xmath123 matches the larmor frequency @xmath126 , where @xmath127 khz / mt is the gyromagnetic ratio and @xmath128 is the local field , the @xmath14li@xmath6 spins precess about @xmath128 , causing a loss of polarization . to establish the vortex state in the ybco samples ,
they are cooled in a static magnetic fields @xmath129 applied parallel to the @xmath28-axis of ybco ( normal to the film and platelet crystals ) .
@xmath130 is also parallel to both the initial nuclear spin polarization and the beam direction .
the local field sensed by the @xmath14li@xmath6 is determined by the applied field and the internal magnetic field generated by the screening currents associated with the vortex lattice .
thus , @xmath128 is distributed over a range of values , which can be calculated using @xmath131 where @xmath132 is the implantation profile calculated using trim.sp given in fig .
[ trimfig ] .
when @xmath5li@xmath6 is implanted in ag ( with no superconducting substrate ) at temperatures below 100 k , it exhibits a single narrow resonance at the larmor frequency.@xcite the resonance should yield an approximately gaussian distribution caused by nuclear dipolar moments.@xcite however , continuous wave rf leads to a power - broadened lorentzian lineshape , whose linewidth is small ( 1 khz @xmath133 mt ) and corresponds to the dipolar broadening due to the @xmath134ag nuclear moments and rf power broadening.@xcite in the presence of any additional magnetic inhomogeneity in the ag , due for example to a vl associated with a superconducting substrate , the observed resonance lineshape will be a convolution of the narrow rf power broadened lorentzian of ag with the ( depth dependent ) field distribution due to the vl in the substrate .
there are unique aspects of measuring the field distribution in the ag overlayer compared to the superconductor itself . as mentioned above , in high magnetic field
, it is possible to isolate and study the broadening due to vl disorder that occurs on a long length scale .
also in low magnetic fields , where the broadening is dominated by the vl , it should be possible to measure @xmath8 in magnetic superconductors since the field distribution above the sample is free of any internal hyperfine fields that make a bulk measurement impossible.@xcite nmr resonances in ag / ybco ( crystal i ) at temperatures 100 k , 80 k , 20 k , and 4.5 k measured in a magnetic field @xmath21 of @xmath135 mt applied along ybco @xmath28-axis .
the solid lines are best fits using a lorentzian . ] , @xmath136 of the twinned ( full symbols ) and detwinned ( opaque squares ) ybco crystals in an applied field @xmath130 .
@xmath137 is the linewidth at temperature @xmath104 of the lorentzian fits and @xmath138 is the constant linewidth in the normal state .
solid lines represent a fit using @xmath139 where @xmath51 is given in eq .
( [ scm - dvl ] ) and @xmath54 and @xmath92 are varied to fit the data .
a @xmath60-wave temperature dependence of @xmath18 in ybco is used,@xcite where @xmath61 nm . ] , the applied field which is 52.3 ( 51.7 ) mt for the crystals ( film ) .
solid lines are lorentzian fits and dashed lines are the simulation described in the text . ]
the @xmath0-nmr resonances were measured as a function of temperature under field - cooled conditions at fields ranging from @xmath140 mt to 3.3 t in each one of the three samples .
[ asymfig ] , shows typical resonance lineshapes at various temperatures in sample i with @xmath141 mt . above @xmath1 ,
the line broadening is small and temperature independent as expected from nuclear dipolar broadening .
below @xmath1 , the field distribution in the ag overlayer broadens dramatically from the vl in the underlying superconductor .
such broadening was observed in all samples and at all magnetic fields , although there are significant variations as a function of both magnetic field and sample as discussed below .
the first thing to note is that the lineshape is very symmetric and fits well to a simple lorentzian .
this is much different from the asymmetric lineshape observed with conventional @xmath13sr in samples similar to i and ii.@xcite the other significant difference between the current results and previous bulk @xmath13sr measurements@xcite on crystals is that the broadening at low temperatures is only weakly dependent on temperature , as may be seen by comparing the resonances at 20 k and 4.5 k. in contrast , the broadening from an ordered vl lattice scales with @xmath103 and consequently in ybco shows a strong linear @xmath104-dependence at low temperatures due to the @xmath60-wave superconducting order.@xcite of the @xmath0-nmr resonance spectra at temperatures @xmath142 4.5 - 10 k in the three samples .
the experimental broadening is compared with the linewidth of an ideal vl @xmath143 ( dashed lines ) , and @xmath144 of a disordered vl ( solid lines ) , which are both weighted by the @xmath14li@xmath6 profile given in fig .
[ trimfig ] . ]
the observed lineshape fits well to a convolution of two lorentzians , one from vortices in the superconducting state with a full width at half maximum ( fwhm ) @xmath145 , and one from other sources determined from the normal state of fwhm @xmath138 .
the width of a convolution of two lorentzians is the sum of the individual widths : @xmath146 .
therefore , the contribution from vortices in the superconducting state can be obtained by simply subtracting the temperature independent normal state width .
[ fwhmfig ] shows the resulting @xmath147 as one enters the superconducting state in samples i and ii . at low field ,
the measured width ( @xmath148 mt ) at low temperature is larger than expected from a regular vl and decreases significantly in the detwinned crystal to about @xmath149 mt . for comparison , simulations using eq .
( [ pb ] ) and the @xmath14li@xmath6 stopping profile in fig . [
trimfig](b ) ; indicate that the broadening due to a regular vl is only @xmath150 mt . at 3.33 t , the discrepancy between the observed width ( see fig . [ fwhmfig ] ) and that expected from a regular vl is even more dramatic . at this high field
the vortices are spaced so closely ( @xmath151 nm ) , that there should be no detectable broadening from a regular vl for our stopping depths .
this can be seen clearly from the simulation in fig .
[ simfig](b ) , where the vl broadening approaches zero at high fields .
in contrast , the data at 3.33 t shows significant broadening below @xmath1 which is therefore solely attributed to vortex disorder on a long length scale .
the temperature dependence of the broadening is also much weaker than expected from a regular vl in ybco , where @xmath103 has a strong linear term due to the @xmath60-wave order parameter.@xcite the observed temperature dependence fits well to our model of disorder , where @xmath147 is compared to an estimate of the fwhm given by @xmath152 , where @xmath51 is given in eq .
( [ scm - dvl ] ) .
this leads to an estimate of @xmath54 of the order of a micron , consistent with the the separation between twin boundaries or grain boundaries.@xcite in the detwinned crystal , @xmath54 is found to be larger but not infinite since the detwinning is not complete .
the fraction of vortices @xmath92 pinned by structural defects in the twinned crystal is about @xmath153 at low field ( 52 mt ) and decreases considerably at high field ( 3.33 t ) .
thus , the amplitude of the enhanced vortex density at the defects , @xmath154 , varies between 2 - 4 mt at all fields . in the detwinned crystal , @xmath155 ,
is an order of magnitude smaller than in the twinned crystal at the same field , with small variation in the vortex density ( 0.5 mt ) compared to the twinned crystal .
these results are consistent with expectations from pinning at twin boundaries . in particular
, one expects the fraction of vortices pinned will decrease in the partially detwinned crystals .
also , it is reasonable to expect that in a high magnetic fields the fraction of vortices pinned will decrease due to the smaller separation between vortices and resulting increase in the repulsive interaction . in fig .
[ ybcosfig ] , the spectra in all three samples above and below @xmath1 are compared with the corresponding simulated field distributions .
the observed lineshapes are all symmetric , and significantly broader than expected , showing little or no sign of the characteristic vl field distribution .
simulation of the vl lineshapes ( dashed lines ) was done using eqs .
( [ fourierb ] ) , ( [ bzr_eq ] ) , and ( [ pb ] ) , for @xmath62 nm , and was convoluted with a lorentzian representing the normal state spectra with @xmath156 mt . the theoretical lineshape for the film is broader and asymmetric because it is weighted by the @xmath14li@xmath6 stopping distribution which was on average closer to the superconductor .
the lineshapes for the crystals are almost symmetric as the ideal vl lineshape at the depths of an average 90 nm away from the superconductor are narrower than the lorentzian they are convoluted with .
the magnetic field dependence of the superconducting linewidth @xmath147 at low temperatures @xmath142 4.5 - 10 k is plotted in fig .
[ fwhmbfig ] . in all samples ,
we find that the broadening is largest at low field and decreases gradually with increasing field .
also , in all cases the broadening remains large and well above the prediction from a regular vl , approximated by @xmath157 where @xmath51 of an ideal vl is given in eq .
( [ scm - vl ] ) and weighted by the @xmath14li@xmath6 profile given in fig .
[ trimfig ] .
the broadening is substantially smaller in the detwinned crystal compared to the other samples .
one can account for all of the data using a linewidth due to a disordered vl , @xmath158 , where @xmath51 is now given in eq .
( [ scm - dvl ] ) ) and weighted by the @xmath14li@xmath6 profile plotted in fig .
[ trimfig ] .
the data is well fitted ( see fig . [ fwhmbfig ] ) by assuming that the twin / grain boundaries spacing @xmath54 is sample dependent of the order of a few microns , and by assuming a phenomenological form for the fraction of pinned vortices @xmath98 with @xmath159 and @xmath99 sample dependent .
it is clear that the @xmath0-nmr lineshapes in the ag overlayer differ substantially from that expected from a well - ordered vl field distribution .
this has little to do with the method of observation .
for example , in the conventional superconductor nbse@xmath111 , @xmath0-nmr shows the expected vl lineshape.@xcite the lineshapes reported here in the ybco film are also qualitatively different than that seen with le-@xmath13sr in a ybco film coated with a 60 nm thick ag layer.@xcite in that experiment the authors found a more asymmetric lineshape in the ag overlayer which was closer to that of a regular vl
. some of this difference may be due to the different pinning characteristics of the samples , although the ybco film used by niedermayer _ et al . _
was from the same source as sample iii .
also , the le-@xmath13sr experiment was probing the vl closer to the interface and in a lower applied field where the disorder is less important compared to the contribution from the ordered vl .
the symmetry and large broadening of the lineshape at low fields can not be accounted for by vl melting ( at a reentrant vortex liquid state near @xmath19 ) which would instead yield a motional narrowing of the field distribution.@xcite the observed resonances in the current experiment are dominated by long range variations of the vortex density across the face of the sample.@xcite such disorder in the vl can produce a symmetric lineshape,@xcite and can broaden the field distribution significantly compared to that of the corresponding ordered state.@xcite weak random pinning or point - like disorder due to oxygen deficiency may slightly distort the vl , and may also broaden the lineshape.@xcite however , the correlated disorder due to the twin and grain boundaries is dominant at long wavelengths , @xcite and therefore we are mostly sensitive to the twin / grain boundaries .
indeed , the position of the probe outside the superconductor enhances its sensitivity to long wavelength disorder , as the proximal fields fall off with distance as @xmath160 where @xmath54 is the wavelength of the inhomogeneity of the field.@xcite the broadening is reduced in a detwinned crystal where the twin boundaries are more sparse as shown in fig .
[ ybcosfig ] , thus the vortex density variation across the face of the sample is smaller than in the twinned crystals as @xmath92 is largely reduced .
the extrinsic broadening due to disorder at low temperature , @xmath161 , reported here is between 0.5 mt and 2.5 mt .
this is remarkably close to the additional gaussian broadening required to explain lineshapes in bulk @xmath13sr measurements on crystals.@xcite in the bulk , this extrinsic broadening is small compared to the intrinsic vl broadening , whereas outside the sample the reverse is true .
it is important to note the @xmath104-dependence of the extrinsic broadening in fig .
[ fwhmfig ] _ does not follow the superfluid density _ ( @xmath162 ) which varies linearly at low-@xmath104 because of the @xmath60-wave order parameter.@xcite instead , we observe a much weaker @xmath104-dependence .
this is expected from our model of disorder which occurs on a long length scale @xmath54 .
for example , at low-@xmath104 where @xmath8 is short compared to @xmath54 , the flux density outside the sample is determined solely by inhomogeneities in the vortex density , and is independent of the superfluid density as the vortices are static and well - pinned in the twin boundaries .
the current results may also explain early @xmath13sr work on htsc powders and sintered samples which mistakenly indicated an @xmath105-wave @xmath104-dependence of @xmath103.@xcite it is likely in these cases the linewidth was dominated by extrinsic vl disorder on a long length scale .
this tends to flatten the @xmath104-dependence of the linewidth and the effective @xmath8 obtained from the analysis.@xcite therefore we conclude that although the linewidth obtained from powders can be useful in making rough estimates of @xmath8 , one can not extract accurate measurements of @xmath8 or its @xmath104-dependence without additional information about the source of broadening and in particular vl disorder . in conclusion , we have measured the magnetic field distributions due to the vortex state of ybco using @xmath0-nmr .
we find a significant inhomogeneous broadening of the nmr attributed to the underlying vl in ybco .
however , the observed resonances have several unexpected properties .
in particular , they are broader and more symmetric than for an ideal vl .
the anomalous broadening is most evident in high field where there is no significant contribution from the regular vl .
these effects are attributed to long wavelength disorder from pinning at twin or grain boundaries .
the temperature dependence of the disorder - related broadening does not scale with @xmath103 , suggesting there is a contribution to the linewidth in the bulk of the vortex state that does not track the superfluid density .
this is likely to have only a minor effect on the interpretation of data on crystals where the observed lineshape is close to that expected from a well ordered vl .
however , it can be significant in powders or crystals where there is substantial disorder in the vl .
in particular , when the broadening is dominated by vl disorder on a long length scale ( i.e. much bigger than @xmath8 ) the temperature dependence of the linewidth does not scale with @xmath103 and therefore can not be used to determine the symmetry of the superconducting gap .
we would like to thank d. a. bonn , w. n. hardy , and r. liang for providing the ybco crystals .
we would also like to acknowledge r. abasalti , d. arseneau , and s. daviel for expert technical support , and nserc , cifar for financial support .
z. salman , d. wang , k. h. chow , m. d. hossain , s. kreitzman , t. a. keeler , c. d. p. levy , w. a. macfarlane , r. i. miller , g. d. morris , t. j. parolin , h. saadaoui , m. smadella , and r. f. kiefl , phys .
lett . * 98 * , 167001 ( 2007 ) .
r. f. kiefl , w. a. macfarlane , g. d. morris , p. amaudruz , d. arseneau , h. azumi , r. baartman , t. r. beals , j. behr , c. bommas , j. h. brewer , k. h. chow , e. dumont , s. r. dunsiger , s. daviel , l. greene , a. hatakeyama , r. h. heffner , y. hirayama , b. hitti , s. r. kreitzman , c. d. p. levy , r. i. miller , m. olivo , and r. poutissou , physica b * 326 * , 189 ( 2003 ) .
g. d. morris , w. a. macfarlane , k. h. chow , z. salman , d. j. arseneau , s. daviel , a. hatakeyama , s. r. kreitzman , c. d. p. levy , r. poutissou , r. h. heffner , j. e. elenewski , l. h. greene , and r. f. kiefl , phys .
lett . * 93 * , 157601 ( 2004 ) .
z. salman , k. h. chow , r. i. miller , a. morello , t. j. parolin , m. d. hossain , t. a. keeler , c. d. p. levy , w. a. macfarlane , g. d. morris , h. saadaoui , d. wang , r. sessoli , g. g. condorelli , and r. f. kiefl , nano lett .
* 7 * , 1551 ( 2007 ) .
m. xu , m. d. hossain , h. saadaoui , t. j. parolin , k. h. chow , t. a. keeler , r. f. kiefl , g. d. morris , z. salman , q. song , d. wang , and w. a. macfarlane , journal of magnetic resonance * 191 * , 47 ( 2008 ) .
h. saadaoui , w. a. macfarlane , g. d. morris , z. salman , k. h. chow , i. fan , m. d. hossain , r. liang , a. i. mansour , t. j. parolin , m. smadella , q. song , d. wang , and r. f. kiefl physica b * 404 * , 730 ( 2009 ) .
niedermayer , e. m. forgan , h. glckler , a. hofer , e. morenzoni , m. pleines , t. prokscha , t. m. riseman , m. birke , t. j. jackson , j. litterst , m. w. long , h. luetkens , a. schatz , and g. schatz , phys .
83 * , 3932 ( 1999 ) .
d. r. harshman , g. aeppli , e. j. ansaldo , b. batlogg , j. h. brewer , j. f. carolan , r. j. cava , m. celio , a. c. d. chaklader , w. n. hardy , s. r. kreitzman , g. m. luke , d. r. noakes , and m. senba , phys .
b * 36 * , 2386 ( 1987 ) .
r. f. kiefl , t. m. riseman , g. aeppli , e. j. ansaldo , j. f. carolan , r. j. cava , w. n. hardy , d. r. harshman , n. kaplan , j. r. kempton , s. r. kreitzman , g. m. luke , b. x. yang , and d. ll .
williams , physica c * 153 * , 757 ( 1988 ) .
j. e. sonier , r. f. kiefl , j. h. brewer , d. a. bonn , j. f. carolan , k. h. chow , p. dosanjh , w. n. hardy , ruixing liang , w. a. macfarlane , p. mendels , g. d. morris , t. m. riseman , and j. w. schneider , phys .
lett . * 72 * , 744 ( 1994 ) .
t. j. parolin , z. salman , k. h. chow , q. song , j. valiani , h. saadaoui , a. ohalloran , m. d. hossain , t. a. keeler , r. f. kiefl , s. r. kreitzman , c. d. p. levy , r. i. miller , g. d. morris , m. r. pearson , m. smadella , d. wang , m. xu , and w. a. macfarlane phys .
b * 77 * , 214107 ( 2008 ) .
s. l. lee , p. zimmermann , h. keller , m. warden , i. m. savi , r. schauwecker , d. zech , r. cubitt , e. m. forgan , p. h. kes , t. w. li , a. a. menovsky , and z. tarnawski , phys .
* 71 * , 3862 ( 1993 ) .
u. divakar , a. j. drew , s. l. lee , r. gilardi , j. mesot , f. y. ogrin , d. charalambous , e. m. forgan , g. i. menon , n. momono , m. oda , c. d. dewhurst , and c. baines , phys .
lett . * 92 * , 237004 ( 2004 ) .
t. m. riseman , j. h. brewer , k. h. chow , w. n. hardy , r. f. kiefl , s. r. kreitzman , r. liang , w. a. macfarlane , p. mendels , g. d. morris , j. rammer , j. w. schneider , c. niedermayer , and s. l. lee , phys . rev .
b * 52 * , 10569 ( 1995 )
. y. j. uemura , g. m. luke , b. j. sternlieb , j. h. brewer , j. f. carolan , w. n. hardy , r. kadono , j. r. kempton , r. f. kiefl , s. r. kreitzman , p. mulhern , t. m. riseman , d. ll .
williams , b. x. yang , s. uchida , h. takagi , j. gopalakrishnan , a. w. sleight , m. a. subramanian , c. l. chien , m. z. cieplak , g. xiao , v. y. lee , b. w. statt , c. e. stronach , w. j. kossler , and x. h. yu , phys .
lett . * 62 * , 2317 ( 1989 ) . | @xmath0-detected nmr ( @xmath0-nmr ) has been used to study vortex lattice disorder near the surface of the high-@xmath1 superconductor yba@xmath2cu@xmath3o@xmath4 ( ybco ) .
the magnetic field distribution from the vortex lattice was detected by implanting a low energy beam of highly polarized @xmath5li@xmath6 into a thin overlayer of silver on optimally doped , twinned and detwinned ybco samples .
the resonance in ag broadens significantly below the transition temperature @xmath1 as expected from the emerging field lines of the vortex lattice in ybco . however , the lineshape is more symmetric and the dependence on the applied magnetic field is much weaker than expected from an ideal vortex lattice , indicating that the vortex density varies across the face of the sample , likely due to pinning at twin boundaries . at low temperatures
the broadening from such disorder does not scale with the superfluid density . |
the many - body problem can not be solved exactly for realistic interactions and systems .
numerous approximations have been formulated and applied over the years .
one of the problems is that many particles require a large hilbert space to allow for the many possible correlations .
in fact , the space typically grows exponentially with the particle number .
the necessary reduction of the hilbert space to obtain a tractable problem has to be accompanied by a correponding transformation of the interaction which in turn has to be used in the smaller space .
this is the method of effective interactions in reduced spaces .
a different approach starts by an approximation applied to the interaction itself .
in many physical fields , the two - body interaction is a complicated entity , and therefore it is desirable to reduce this complexity .
one such direction is to perform a harmonic approximation of the interaction hamiltonian which leaves all terms ( one- and two - body ) as a polynomial of at most second order in the coordinates of all particles .
this yields an exactly solvable model for the many - body system @xcite .
however , to utilize such an approach , the input parameters must be adjusted to reproduce properties of realistic systems for low particle numbers .
this could be two - body binding and scattering information and structural expectation values .
this sort of approach has been pursued in nuclear and condensed - matter physics for many years @xcite as a source of exact insight into many - body problems @xcite when numerics are either intractable or hard to interpret .
more recently , cold atomic gases have emerged as an arena and testing ground for various models due to the large degree of controllability in experiments @xcite . in particular
, interaction can be controlled through interatomic feshbach resonances @xcite .
however , these systems are almost exclusively studied in the presence of an external confining potential which can often be modelled by a ( possibly deformed ) harmonic oscillator . for experimental setups with very deformed traps ,
the geometry can in fact be manipulated to study effective one- or two - dimensional dynamics .
very recently , it has even become possible to explore the few - body limit with trapping of just a few atoms near the degenerate regime @xcite . at the moment , the physics of ultracold atomic gases is , however , dominated by short - range interacting neutral atoms @xcite .
in addition , since the samples are often extremely dilute , the details of the potential at short - range does not matter , and only low - energy scattering information is important @xcite .
the replacement of the potential by a regularized zero - range potential is therefore often justified . in the presence of a harmonic confinement ,
the solution of the two - body problem with a zero - range potential was presented by busch _
this solution was subsequently confirmed in experiments @xcite .
more recently , this model has been used as a starting point for addressing the few- and many - body problem in a controlled manner in both nuclear and cold atom physics @xcite .
the zero - range approximation scheme is one approach to effective interactions .
here we pursue a different one via harmonic expansion of the interaction hamiltonian .
system that can be described in some regimes of parameter space by harmonic hamiltonians are extremely useful due to their solvability . of course ,
as mentioned above , the parameters have to be meningfully related to realistic physical systems .
this can be done through fits to two - body properties as described previously in ref .
@xcite . here
we take a parametric approach and study the details of the harmonic approximation for identical quantum particles as function of the two relevant frequencies for such a system ; the ( internal ) interaction frequency and the ( external ) trapping frequencies , both of which are described by ( isotropic ) harmonic terms .
deformation can be easily implemented and will not be considered here .
the objective of the present paper is to describe a method for calculating thermodynamical quantities for systems of identical bosons and fermions , a problem of relevance for any subfield of physics concerned with quantum mechanical behavior of multiparticle system in equilibrium .
this requires full access to the partition function and the thermodynamic quantities that can be derived from it .
the quantum statistics of ensemble of particles is needed here , and we present a general method for obtaining such information for harmonic hamiltonians that works for any number of particles in principle . there is great interest in the universal thermodynamics of strongly interacting fermi gases @xcite , and measurements have determined the equation of state of the system @xcite .
those studies are mostly concerned with trap - averaged quantities that are less sensitive to sudden changes in thermodynamic behavior as expected near phase transitions .
however , recently it has become possible to determine local properties of the gas by density fluctuation measurements @xcite .
quantities that involve derivatives of the energy or pressure like heat capacity and compressibility can thus be used to study for instance superfluid properties of both fermionic @xcite and bosonic systems @xcite .
in addition , the same experiments are also able to explore effects of changing the dimensionality of the system through optical lattice potentials . in the current presentation
we focus on extending th harmonic approximation into the statictical regime , and more particularly on the technicalities of the method used to achieve this . to demonstrate our method we consider the case of a two - dimensional system of identical bosons or fermions for different ratios of interaction to external trapping frequencies .
the paper is organized as follows . in sec .
[ secmethod ] we briefly recap the harmonic method and then discuss the partition function in the harmonic approximation . the most important part of the method is the counting of the degeneracies in the partition function which we discuss in extended detail .
the thermodynamic quantities are briefly outlined and a discussion of how to related the parameters in the harmonic hamiltonian to the model by busch _
@xcite is given .
practical numerical results are obtained and discussed in sec .
[ secthermo ] for a system of cold atoms confined to two spatial dimensions . both bosons and fermions
are explicitely investigated .
finally , give a summary and an outlook in sec .
[ seccon ] .
we consider a system of @xmath0 identical quantum particles , fermions or bosons , that have two - body interactions , all of which are confined by an external field .
all interactions are substituted by carefully chosen harmonic oscillators .
the solutions of the hamiltonian can then be found explicitly and treatment of thermodynamic properties is possible .
we first present the necessary ingredients of the model , and continue to develope the method for calculating the partition function .
this includes the key discussion of how to obtain the degeneracies of the many - body states .
the @xmath0-body hamiltonian contains kinetic energy , one- and two - body interaction terms , all of which we assume can be written in terms of second order polynomials in the particle coordinates when proper choice of effective interaction parameters have been made .
the resulting harmonic oscillator structure can be separated into terms related to each of the cartesian coordinates and the hamiltonian in one , two , and three dimensions is a sum of terms from each coordinate . for example , in the @xmath1-direction , the effective hamiltonian is @xmath2 where @xmath3 is the mass of the identical particles and @xmath4 is the reduced mass .
the one - body ( external ) harmonic trap frequency is @xmath5 , while the two - body interaction frequency between all pairs is denoted @xmath6 ( internal ) .
@xmath7 is a constant energy shift which is included for generality .
double counting of interactions account for the factor @xmath8 . for two and three dimensions ,
the total hamiltonian is @xmath9 and @xmath10 .
we will assume that the external harmonic trap is isotropic , extension to the deformed case are straightforward and will be addressed in future work .
all particles are identical and interact in the same way which means that @xmath6 is independent of @xmath11 and @xmath12 . also the one - body frequency ,
@xmath5 , is independent of @xmath12 . the schrdinger equation for the hamiltonian in eq .
can be solved analytically for any particle number ( details can be found in @xcite ) the pertinent properties in the present context are that the energies , @xmath13 , of the excited states have oscillator character . here
@xmath13 denotes the @xmath14th _ many - body _ excited state , i.e. the energy of all @xmath0 particles when they are distributed in a certain way in the oscillator levels that come out of the diagonalization of the full hamiltonian . in this paper
we will concentrate mostly on the case of a two - dimensional system , and we therefore specialize to this case now .
the energy of a many - body state , @xmath13 , can be written @xmath15 where @xmath16 and @xmath17 are quanta in the @xmath1-direction related to the relative and center - of - mass excitations with the former carrying the index @xmath12 running over all @xmath0 particles ( and similar terms for the @xmath18-direction ) .
the total number of relative and center - of - mass quanta in the many - body state @xmath14 are denoted @xmath19 and @xmath20 respectively .
note that @xmath16 , @xmath17 , @xmath21 , and @xmath22 must all be specified to completely characterize the many - body state of energy @xmath13 .
however , more than one set of individual quantum numbers can lead to the same energy @xmath13 and this determines the degeneracy of the state , which we denote @xmath23 and discuss at length below . @xmath13 and @xmath23 are all that is needed to compute the partition function and thus the thermodynamic properties of the system . the constants , @xmath24 and @xmath25 in the formula for @xmath13 , are from the two - dimensional zero - point energy of intrinsic and center of mass parts , respectively .
the one - body frequency , @xmath5 , correspondings to center of mass motion .
the other frequency , @xmath26 , is @xmath27 times degenerate for identical particles .
it is given by the relation @xcite @xmath28 if the two - body interaction goes to zero ( @xmath29 ) , @xmath30 , and the resulting eigenvalues and eigenfunctions converge to those of the external one - body trapping potential .
the solutions are @xmath0 non - interacting identical particles occupying @xmath0 states in a harmonic oscillator . for interacting particles
, we obtain instead @xmath27 identical frequencies with value given by eq . .
the extension of the above discuss to three dimensions involves also the @xmath31-direction and is straightforward , although the extra quantum numbers makes the degeneracy larger and the counting of many - body states more involved . with the harmonic hamiltonian ,
the exact spectrum of excited many - body states , @xmath13 , are thus available for a given number of particles . for the thermodynamic calculations ,
the natural choice is to use the canonical ensemble applicable for a definite number of particles .
the canonical partition function is @xmath32 where @xmath23 is the degeneracy of the @xmath14th many - body state and @xmath33 is the temperature measured in energy units ( the boltzmann constant , @xmath34 ) . once @xmath35 is obtained , all thermodynamic quantities can be derived from it .
the most involved question is how to obtain @xmath23 for a given energy of the many - body system .
this will now be discussed .
the only quantities in the partition function are the energies , @xmath13 , and their degeneracies , @xmath23 .
the energies are appropriate multiples of the harmonic oscillator frequencies introduced above .
the degeneracies depend strongly on which symmetry restriction is imposed .
problems arise when all , or a group of , particles can not be distinguished because they are identical and allow occupation of the same states .
we then have to count the number of ( anti-)symmetric states for the indistinguishable particles of a given excitation energy , @xmath13 , and the corresponding degeneracies , @xmath23 , of each _ many - body _ state of energy @xmath13 .
we now describe our procedure to obtain @xmath23 .
it is based on knowledge of the non - interacting system , a recursive procedure , and the fact that when the two - body interaction vanishes , the degeneracies remain the same as we will now explain .
we start with @xmath0 identical particles without any two - body interaction ( @xmath29 ) , and all in the same external trap with frequency @xmath5 .
this hamiltonian will have @xmath0 degenerate solutions of frequency @xmath5 per spatial dimension ( so @xmath36 for the two dimensional case ) .
the many - body solutions are the products of @xmath0 combinations of the single - particle harmonic oscillator wave functions
. for identical fermions , the pauli principle , or equivalently the antisymmetrization , requires that each of the single - particle states at most is occupied by one particle .
in addition , if this condition of occupancy by at most one particle for each state is fulfilled , then one and only one antisymmetric state is uniquely constructed as the slater determinant of the @xmath0 singly occupied single - particle states .
importantly , the energy can be characterized by the sum , @xmath37 , of single - particle oscillator quantum numbers for all particles in the state @xmath13 , as seen in eqs . and .
assume that we have computed the number , @xmath38 , of properly symmetrized states of non - interacting identical particles in the external potential .
the quantity @xmath38 can be computed in a simple manner by trail - and - error , i.e. taking all possible permutations and testing which are completely symmetric and which are completely antisymmetric .
this is , however , notoriously difficult since the computational effort is exponentially increasing with particle number .
also , it does not distinguish between center of mass motion and relative motion and this needs to be disentangled .
this can be done by recalling the symmetry of the center of mass as we now discuss .
the center of mass motion is a symmetric mode in all permutations of the particle coordinates , simply because the center of mass coordinate is the sum of the individual particle coordinates .
this motion is always symmetric , independent of the number of quanta in this mode , @xmath20 .
therefore , the relative motion determines the symmetry of the total wave function .
the excited states consist of a combination of two distinctly different types of excitations .
they are characterized by the number of quanta in the center of mass motion , @xmath20 , and the number of quanta in the relative motion , @xmath19 .
we denote the number of properly ( anti)symmetric @xmath0-body states by @xmath39 , which is the number of states that have the proper symmetry in terms of the _ relative _ coordinates .
it can be computed from knowledge of the total number of non - interacting many - body states , @xmath40 , for all values of @xmath37 .
to do so , one must subtract the number of states of different center of mass quanta from the total number of non - interacting many - body states in a recursive manner , i.e. @xmath41 where the degeneracy , @xmath42 , of the center of mass state depends on the spatial dimension of the oscillator ; @xmath43 for one , @xmath44 for two , and @xmath45 for three dimensions , respectively .
the formula comes from the fact that we know only the total number of quanta , @xmath46 , but _ not _ @xmath19 and @xmath20 individually .
we need to recursively subtract the states that correspond to @xmath47 , 2 , @xmath48 , @xmath46 as the formula prescribes .
the factor @xmath42 is due to the degeneracy of the center of mass motion itself for given number of quanta , @xmath20 .
note that @xmath49 is well - defined and corresponds to a state with no quanta of excitation , i.e. the ground state .
it can be zero or non - zero depending on @xmath0 and whether the particles are bosons or fermions .
. determines the degeneracy of an @xmath0-body state in a harmonic system and is one of the main results of this work .
the degeneracy can now be calculated iteratively from eq . for any number of quanta , @xmath46 from a starting point based on a non - interacting system of identical particles in a trap .
these states must then be either completely symmetric or antisymmetric with respect to interchange of all coordinates of any pair of particles to obey either bosonic or fermionic statistics .
now we add the two - body interaction on top of the external one - body potential .
the external frequency , @xmath5 , remains in the spectrum corresponding to the center of mass motion , and an @xmath27 times degenerate frequency , @xmath26 , appears corresponding to relative motion .
the absolutely crucial point is that the degeneracies do _ not _ depend on the two - body interaction strength which will only influence the value of the degenerate frequency .
the counting remains completely unchanged , because the number of states of a given symmetry is a discrete number which does not change with continuous variation of the interaction .
mathematically , this corresponds to a continuous map ( the scheme for counting degeneracies ) from a continuous interval ( the interaction frequency , @xmath26 ) to a discrete set ( number of states , @xmath23 , with given energy ) .
such a map must necessarily be a constant map .
the underlying point here is that our splitting into @xmath20 and @xmath19 is done in a way that moves the states that are related to the relative motion when @xmath26 is increased from zero to its full value , while the states that are related to center of mass motion remain constant in energy since @xmath5 remains a solutions .
we confirmed this simple counting method by an elaborate and computationally very slow brute force procedure which is much worse than computing @xmath38 above .
the ( anti-)symmetrization of each of the computed wave functions is performed by permutations of all the coordinates followed by extraction of a basis with a size equal to the smallest number of linearly independent states . for sufficiently small
@xmath0 this can be computed in reasonable time and comparison can be made . in all cases considered below we found perfect agreement between the numerics and the procedure outlined above .
the procedure above is , however , far superior since it needs only counting of states with given energy , _ not _ the full wave functions .
we can now use eq . to get the number of center of mass states and the number of relative states as functions of the sum of corresponding quanta .
along with the frequencies , this is all we need to calculate the partition function in eq . .
this simplification occurs since the structure of the wave functions do not enter in the partition function .
the center of mass mode of frequency @xmath5 separates in eq . because both the exponential function and the degeneracy factorizes , that is @xmath50 , where the center of mass contribution is analytical and given below eq . .
we have @xmath51 where @xmath52 is the dimension , @xmath53 , and @xmath54 . in practice
, the infinite sum over relative quanta has to be limited by a cut - off value which is chosen sufficiently high to achieve the required accuracy .
certain quantities are more sensitive to the cut - off than others , such as those with more derivatives of the partition function ( heat capacity , compressibility , etc . ) .
for all the results presented below we have checked convergence by increasing the cut - off and identifying the value of @xmath33 below which the relevant quantities remain unchanged . from the canonical partition function , eqs . or , we get immediately the basic quantities of energy , @xmath55 and free energy , @xmath56 , that is @xmath57 where @xmath58 is the entropy , and @xmath59 is the heat capacity @xcite . in thermodynamic formulations of macroscopic systems
, both temperature and volume is usually external parameters .
both the energy , @xmath55 , and heat capacity , @xmath59 , defined above , are in principle the derivatives for a fixed volume , @xmath60 ( and also fixed @xmath33 and @xmath0 ) . in our case with an externally confined @xmath0-body system
the volume is not the fixed quantity , but rather it is the external trapping frequency , @xmath5 , that is fixed .
the derivatives above are thus taken for fixed @xmath5 .
usually , variation of the volume allows information about pressure and compressibility .
definitions of these quantities and connection to their thermodynamic counterparts require a precise translation between @xmath5 and volume . here
we employ the simplest choice and define the volume via the length parameter related to the external trap , @xmath61 , which is defined by @xmath62 .
we now have @xmath63 , where @xmath64 and @xmath65 are the two and three dimensional angular surface areas .
derivatives of any quantity , @xmath66 , with respect to volume can now be written @xmath67 for the pressure , @xmath68 , we find @xmath69 which , from the definition in eq . , gives @xmath70 the derivatives are easily worked out since they only contain @xmath71 .
continuation to the second derivative with respect to the volume produce the isothermal bulk modulus , @xmath72 , defined as @xmath73 which is the reciprocal of the isothermal compressibility , @xmath74 . again
two terms arise related to center of mass and relative degrees of freedom .
several terms appear by carrying out the two derivatives .
this kind of compressibility , that of a response of a system to an applied pressure , is different to the compressibility of a self - bound system such as a nucleus @xcite . for self - bound systems , the compressibility refers to resistance to density fluctuations .
this can be written as the second derivate of the energy per particle with respect to the fermi momentum for fermionic systems . from eq . and the external trap frequency , @xmath5 , as function of the scattering length , @xmath75 , in two dimensions for the busch model @xcite with @xmath76-wave interactions for bosons ( upper ) and for the @xmath77-wave interaction identical ( i.e. spin - polarized ) fermions ( lower ) . in both plots ,
the different curves correspond to different particle numbers . from the bottom to top , the curves are for 3 , 4 , 5 , 6 , 10 , 20 , and 30 particles .
the length parameter is the trapping length , @xmath78 , introduces in the text.,scaledwidth=48.0% ] before we present results for degeneracies and various thermodynamic properties , we discuss how one can fix the two - body interaction term to capture the properties of the realistic system which have interactions that are not harmonic . in the case of identical bosons ,
the procedure was already presented in ref .
@xcite but we repeat the arguments here for completeness .
we assume that we are in the situation that the real two - body interaction is of much shorter range than the external trap length @xmath61 above . in this case
, we can use the zero - range model of busch _ et al .
_ @xcite to obtain the solutions in the trap .
the effective two - body interaction that we want to use is a harmonic oscillator , which in general contains two parts ; an oscillator frequency , @xmath6 , and a shift , @xmath79 . to fix these parameters
we employ the conditions that the two - body harmonic interaction should reproduce the correct two - body binding energy of the ground state in the busch model , and also the spatial extend of the wave function .
the latter condition is implemented by calculating the root - mean - square radius , @xmath80 , in the exact solution and fixing the oscillator parameter to reproduce value .
the energy shift is suqsequently tuned so as to also reproduce the two - body binding energy . clearly , since the busch model uses a zero - range interaction , the only scale left is the external trap frequency , @xmath5 .
therefore what we obtain from this is the quantity @xmath81 .
we ignore the shift from now on since it merely provides an overall shift in the @xmath0-body harmonic hamiltonian which is not of interest in the current paper . the next step is to insert this value into eq . to obtain @xmath82 , which will now naturally depend on @xmath0 .
in the upper panel in fig . [ busch ] we show the result of this procedure for the case of bosons in two dimensions ( more detail for bosons in both two and three dimensions can be found in ref .
the results are plotted as function of the two - dimensional scattering length @xmath75 .
note that when @xmath75 is small , the ground state becomes strongly bound , while for large @xmath75 it goes to the non - interacting limit .
this is clearly reflected in the behavior of @xmath82 .
the result of this procedure agrees rather well with predictions from the well - studied problem of bosons interacting through a zero - range interaction @xcite in the strong and weak coupling limits @xcite , also in the case where higher - order interaction terms are included @xcite .
recently , a very similar procedure has been used to study polar molecules in layered system @xcite where excellent agreement with exact methods has been found both for isotropic @xcite and anisotropic potentials @xcite which have reasonably large potential pockets . incidentally
, one - dimensional dipolar system also turn out to have such potential pockets @xcite and applying the harmonic approximation to these systems is an interesting direction for future research . for identical
( spin - polarized ) fermions a complication arises from the fact that the original busch model applies to particles interacting through an isotropic @xmath76-wave interaction only .
the pauli principle dictates that @xmath76-wave interactions are zero for identical fermions , and therefore the interactions must have @xmath77-wave ( or higher odd partial wave ) character . in two dimensions ,
the equivalent of the busch model with @xmath77-waves can be solved and the spectrum turns out to be very similar to the @xmath76-wave case except for ( unimportant ) shifts @xcite . however , a further complication arises since the corresponding ground state wave function is singular at the origin in a manner that does not allow it to be normalized @xcite .
this can be fixed by a properly defined scalar product as discussed for the three - dimensional case in ref .
we will not pursue this approach here but instead we note that any higher order structural average of the type @xmath83 with @xmath84 an integer is still perfectly convergent .
we thus consider a higher order average , @xmath85 , in order to fix the oscillator parameter for identical fermions .
we are thus tacitly using that the @xmath77-wave spectrum is very similar to the @xmath76-wave one and assume that the wave function should be so as well .
of course , it must be kept in mind that this is _ not _ at odds with the pauli principle since the two - body wave function in the @xmath77-wave channel is still zero at the origin . in any case , the pauli principle is exactly enforced on the @xmath0-body system as discussed above .
the results for @xmath82 are shown in the lower panel of fig .
[ busch ] .
the similarity to @xmath76-waves and bosons is quite clear , and the only difference seems to be slightly lower overall values for the fermionic case . below
we will illustrate our method by calculating thermodynamic quantities for both fermions and bosons . here
the only thing that matters is the ratio @xmath82 .
one can then reverse the process and use fig .
[ busch ] to obtain a corresponding value of the scattering length and thus compare to a realistic system with trapped bosons or single - component fermions . since this is mainly a discussion of method , we will not dwell on this any further .
an important thing to note , however , is that since the ratios @xmath82 are similar for bosons and fermions , the results below will truly isolate the statistical behavior coming from the exchange requirements in the different cases .
we now proceed to demonstrate our methods by computing the density of states and the thermodynamic quantities that have been introduced above .
they depend on particle mass , dimension , number of particles and their quantum statistics , temperature , one- and two - body frequencies . if we measure all energies in units of @xmath86 , all results for given particles in @xmath52 dimensions , depend only on the two ratios , @xmath87 and @xmath82 .
this implies that any kind of interaction model used to define the effective harmonic hamiltonian only has to provide these two quantities as pointed out above . , while it is @xmath88 for fermions.,title="fig:",scaledwidth=48.0% ] , while it is @xmath88 for fermions.,title="fig:",scaledwidth=48.0% ] first we discuss the degeneracies themselves which are found by the procedure discussed above , essentially through the recursive formula eq . . while the number of non - interacting ( anti-)symmetric states increases exponentially as function of excitation energy , the numbers can , however , be tremendously reduced by the symmetry requirements . a simple example indicating this tendency is that for bosons all particles are allowed to occupy the lowest level with zero quanta , @xmath89 .
in contrast , a given number of fermions has a minimum sum of quanta , @xmath90 , in the ground state , that is @xmath91 .
we show the numbers of given symmetry as functions of excitation energy in fig .
[ fermbos ] .
since we are dealing with harmonic oscillators , a given energy corresponds to a given total number of quanta .
the starting point for the ground state is naturally @xmath92 and @xmath93 , which is larger for fermions than for bosons due to the requirement that at most one fermion can occupy each state .
the corresponding higher degeneracy implies therefore that the number of states of given excitation energy is larger for fermions than for bosons because the latter do not need to obey the pauli principle . for bosons ,
the lowest excited state consisting of one quantum added to zero quanta in the ground state can only be a center of mass excitation , since one quantum in the relative motion would be antisymmetric in one pair of coordinates ( it would have to be a state of @xmath77-wave / parity - odd symmetry ) .
the lowest completely symmetric excited state of the relative motion appear for @xmath94 . in conclusion
, there is always a gap in the excitation spectrum for bosons corresponding to @xmath95 where fermions only have a gap of @xmath96 .
this will be reflected in the temperature dependence below . for bosons ( upper part of fig .
[ fermbos ] ) the same degeneracy is found for low energy ( small total number of quanta ) for all particle numbers . in the upper plot of fig . [ fermbos ]
the horizontal axis is @xmath97 since we have found that this is a good measure for the excitation energy in the system . for fermions
, one can give a qualitative argument for the exponential behavior which we discuss below .
the similarity of degeneracies at low energy for bosons can be understood by the direct counting procedure where a given number of particles , @xmath0 , has to be distributed to add up to the total number of quanta , @xmath19 .
first , @xmath27 particles are placed in the lowest oscillator level , and the one remaining particle then has to be placed in the level with total number of quanta equal to @xmath19 .
then we move the single particle one step down to @xmath98 , and simultaneously compensating by moving one particle one step up from the lowest level .
we continue with these combinations until we have @xmath19 particles in the second oscillator level and all others in the lowest level .
adding one particle and repeating the counting process we realize that their is a one - to - one correspondence between the configurations of @xmath0 and @xmath99 particles .
going from one to the other is simply by removing or adding one particle in the lowest oscillator level with zero contribution to the total number of quanta . for @xmath0 bosons ,
the deviation from this universal curve starts for @xmath100 .
the reason is again found by following the counting procedure .
the configurations with all particles in the lowest two levels are only possible when @xmath101 .
this implies that we find fewer states when @xmath102 for @xmath0 than for @xmath99 particles .
the curves break away from the universal curve for increasing @xmath0 when @xmath100 .
the degeneracies for fermions have very different behavior , as seen in the lower part of fig .
[ fermbos ] .
notice that the horizontal axis is different in the two plots .
we notice a regime of linear dependence which has the same origin as the exponential square root dependence of excitation energy of the free fermi gas level density @xcite . for two dimensions ,
the particle number dependence is roughly @xmath103 as reflected on the axis in the fermion plot .
this holds for intemediate excitation energies and for relatively large particle numbers .
a qualitative understanding of the behavior can be obtained in a manner following ref .
the density of _ single - partile _ states at the fermi level in a two - dimensional harmonic trap is roughly @xmath104 , while the typical excitation energy in the system is @xmath105 .
the _ many - body _ level density in this situation is then proportional to @xmath106 $ ] .
this explains the choice of horizontal axis for fermions in fig .
[ fermbos ] .
the suggested linear dependence is not very clear but the assumptions are not well fulfilled .
the requirements is that the excitation energy has to be sufficiently large to allow statistcial treatment and sufficiently small not to exhaust particles at the bottom of the potential .
this is not true for the relatively small particle numbers that we must necessarily work with to make the problem tractable by both brute force symmetrization and the counting scheme developed here . , for several
boson ( upper ) and fermion ( lower ) systems , identified by the number of particles and frequency ratio , @xmath107 .
the ground state energy is subtracted , and the energy unit is @xmath108 for both energies and temperatures .
the ground state energy , @xmath109 , has been substracted.,scaledwidth=50.0% ] we are now in a position to explore the thermodynamics and to compare bosonic and fermionic behavior in detail . the free energies per particle for a two - dimensional system are shown in fig .
[ bosf ] as function of temperature for different interactions and particle numbers , both for bosons ( upper panel ) and fermions ( lower panel ) .
note that the ground state energy has been subtracted as it is not of interest here .
the curves start out very flat at zero showing that the low - temperature dependence is at least quadratic .
the decrease with increasing temperature is actually rather similar to a quadratic behavior with @xmath33 .
the behavoir is modified when @xmath33 exceeds @xmath25 and the center of mass mode becomes active .
notice that the dependence on particle number is relatively weak .
this is in contrast to the interaction dependence where the free energy varies substantially more for the smallest two - body interactions .
in addition , we note that fermion free energies are always smaller than boson free energies .
this demonstrates that the free energy is entropy driven , i.e. the entropy term grows faster than the energy with temperature .
again the larger degeneracy of fermionic systems is seen by the lower free energies . , for several
boson ( upper ) and fermion ( lower ) systems , identified by the number of particles and frequency ratios , @xmath107 .
the ground state energy , @xmath109 , has been subtracted from each curve.,title="fig:",scaledwidth=48.0% ] , for several boson ( upper ) and fermion ( lower ) systems , identified by the number of particles and frequency ratios , @xmath107 .
the ground state energy , @xmath109 , has been subtracted from each curve.,title="fig:",scaledwidth=48.0% ] fig .
[ bosu ] displays the energy per particle for bosons and fermions .
we measure energy in units of @xmath110 to isolate the high temperature behavior which is @xmath111 for two dimensionsal harmonic system by the equipartition theorem .
again we subtract the uninteresting constant ground state energy .
the increase from zero at zero temperature is rather steep for both types of particles .
all curves continue to increase with temperature but much slower after a few units of @xmath25 .
the largest energy per @xmath112 is for the system with the least particles and the smallest frequency ratio .
this can be understood from the fact that for large @xmath82 , only the center of mass modes ( twice degenerate in two dimensions ) are active . however , since we divide by @xmath0 , the values become smaller for larger @xmath0 , but the behavior remains the same ( self - similar lines for @xmath113 in fig .
[ bosu ] ) .
this is similar for both bosons and fermions .
we also observe that the equipartition limit at high temperature is reached for substantially larger temperatures , a clear sign of the interaction effects .
the low - temperature behavior is , however , different for bosons and fermions .
the low - energy @xmath95 gap in the boson spectra arises due to a ground state with all particles with zero oscillator quanta of excitation .
then there are no states with one quanta of excitation . for fermions , there are only @xmath114 gaps .
therefore the energy exhibits flat regions at relatively low temperature , but these features are observed first for fermions and later for bosons due to the difference in energy gap which is related to the activation energy .
this demonstrates a clear signature of quantum statistics in the harmonic model , but which should be expected in generic interacting systems with identical particles . , for boson ( upper ) and fermion ( lower ) systems , identified by the number of particles and frequency ratios , @xmath107 .
, scaledwidth=48.0% ] next , we show the entropy , @xmath58 , in fig .
we note the increase of available states per particle as function of temperature goes from zero to values around a few times the temperature in units of @xmath25 . for lower ratios @xmath82 , the increase with temperature is faster since less energy is required to excite the internal modes of the system and more states are available , correspondingly increasing @xmath58 .
again , for larger values of @xmath82 , the center of mass is the only active mode at low temperature and the division by @xmath0 in the plots explains the lower value of @xmath58 for larger @xmath0 . only at substantially larger temperatures
will both internal and center of mass modes become active .
the effect of symmetry is not very pronounced , although it is noticeable that the entropy is larger for fermionic than for bosonic systems due to the larger degeneracy .
an interesting feature of the fermion plot is that for @xmath115 the limit of @xmath116 for @xmath117 is non - vanishing .
the number of available states is finite reflecting that the ground state itself is degenerate .
this is seen by simply counting of oscillator degeneracy for a two - dimensional harmonic oscillator .
the lowest three quantum levels can hold 6 identical fermions ( 1 , 2 , and 3 , respectively ) , and the fourth can hold additionally four particles .
this means that eight particles only occupy half of the last level , leaving the ground state as six times degenerate .
this beavior at zero temperature then nicely indicate the presence of shell structure . in two dimensions for boson ( upper ) and fermions ( lower ) systems , identified by the number of particles and frequency ratios , @xmath107 .
the black line shows the unscaled value for the mean field center of mass mode.,title="fig:",scaledwidth=48.0% ] in two dimensions for boson ( upper ) and fermions ( lower ) systems , identified by the number of particles and frequency ratios , @xmath107 .
the black line shows the unscaled value for the mean field center of mass mode.,title="fig:",scaledwidth=48.0% ] we now consider some derived thermodynamic quantities that are of experimental interest in many fields of physics . the heat capacity and the compressiblity are two such quantities that can be obtained from second derivatives of the partition function as discussed above . in fig .
[ boscv ] , we show the heat capacity per particle , @xmath118 , defined in eq . as function of temperature for both bosons and fermions .
they both start with an initial activation of the external trap mode , since the degenerate relative degrees of freedom all require higher temperature to be excited .
after a delay , these internal modes are activated , and the heat capacity increases with temperature . the delay and
the rate of increase depend strongly on @xmath82 with a slower increase of @xmath59 for larger interaction ratios .
the tendency to increase slower and in steps is related to gaps in the energy spectrum . at high excitation energy ,
the spectrum becomes denser , and gap sizes larger than the temperature can not appear .
however , the presence of a gap in the low - energy spectrum is important for the heat capacity in general .
in fact , this is clearly seen by the fact that bosons have a flat profile for a region of low temperature , while the fermions have two flat plateaus .
fermions rise faster initially , and always approach the equipartition heat capacity from below ( @xmath119 for two dimensions ) .
the larger gap causes a delay in the boson systems .
the heat capacity slightly overshoots the equipartition value , oscillate back below the equipartion value ( at a temperature outside the scale in fig .
[ boscv ] ) , and eventually approach the limit from below .
the curve marked com shows the heat capacity when assuming that only the center of mass mode is excited .
all the other curves will approach this at very large temperatures when the internal structure is washed out .
however , as the plots clearly show , the approach is very different for different particle numbers and interaction strengths . )
divided by @xmath0 for boson ( upper ) and fermions ( lower ) systems , identified by the number of particles and frequency ratios , @xmath107 .
, scaledwidth=48.0% ] in fig .
[ boskt ] , we show the isothermal compressibility per particle , @xmath120 , as defined in eq .. the high temperature behavior of a harmonic system is @xmath121 .
the compressiblity shows a small increase through a maximum at very small temperature followed by steady decrease towards zero at large temperature with a @xmath121 slope .
the compressiblity indicates how easy it is to squeeze the system and a large value indicates that the system is very susceptible to compression .
we see that the more strongly interacting systems ( larger @xmath82 ) have larger @xmath74 .
this comes from the fact that the attraction in these systems will make it energetically favourable to contract , due to the larger degeneracy of the interaction frequency , @xmath26 .
this also explains why @xmath74 increases with @xmath0 .
comparing bosons and fermions we find the same qualitative behavior .
the fermions are slightly less compressible , with a sharper dependence on particle number and interaction strength .
this is a common feature of fermi systems and usually attributed to the pauli principle . for comparison ,
[ 3dferkt ] shows the compressibility for fermions in three spatial dimensions , which shows the same qualitative behavior .
the results for bosons are similar and we do not show them here . notice that the peak features in the compressibility are sharper in the three - dimensional case .
our system is similar to an attractively interacting fermi gas where superfluidity is expected to show a signature in the compressibility @xcite .
our results are consistent with the fact that phase transitions are less pronounced in lower dimensions . ) divided by @xmath0 for fermionic systems in three dimensions , identified by the number of particles and frequency ratios , @xmath107 .
, scaledwidth=48.0% ]
the harmonic approximation is extremely useful because its simplicity allows transparent calculations of otherwise complicated properties .
the only approximation lies in the choice of harmonic potentials acting on each particle and between pairs of particles .
one can therefore think of the harmonic approximation as an effective interaction scheme and subsequently investiagate the behavior of its predictions under changes in the input parameters .
the latter should naturally be connected to whatever realistic physical system one is interested in studying . in this paper
, we explore the harmonic approximation scheme by considering the hamiltonian for a many - body system consisting of a given number of identical particles .
in particular , we explore the consequences of symmetry requirements on the properties of the system .
this is done with thermodynamic applications in mind since this is a venue where the quantum statistics plays a decisive role at low temperatures .
we develop a new method for counting the number of correctly symmetrized _ many - body _ states that reduces the complications that arise from this fundamental problem of statistical mechanics and thermodynamics .
the advantage of the harmonic approach is obviously that the energy spectrum is analytically known .
however , the degeneracy of each many - body state of given total energy still remains to be determined before the partition function is fully defined and possible to compute numerically .
we design a novel procedure to obtain the degeneracy of each state , subject to requirements of symmetry and antisymmetry appropriate for bosons and fermions , respectively .
we separate the completely permutation symmetric center of mass motion from the relative motion , which then has to carry the symmetry corresponding to bosons or fermions .
we count by subtracting the number of states of different quanta in the center of mass motion from the total number of non - interacting states of given symmetry .
to demonstrate the method , we consider the case of a two - dimensional system with identical bosons or fermions ( with no internal degrees of freedom ) . within the canonical ensemble ,
we compute the partition function , and from it the free energy , entropy , heat capacity , and compressibility of the system .
this is done for a relatively small number of particles ( up to 20 ) .
the method is a considerable improvement over the brute force method where one explicitly checks for symmetry properties by exchanging all pairs of particles one by one .
however , it is still computationally involved when going beyond the particle numbers considered here .
however , as is known from for instance the virial expansion @xcite , it is often enough to consider small particle numbers and then extrapolate to large system sizes from this information .
the effective harmonic interaction can be related to quantities in realistic systems and we discuss a case of great usefulness within the realm of ultracold atomic gases , that of two particles interaction in a harmonic trap through a two - body interaction of zero - range . while bosons can interact in the @xmath76-wave channel originally considered by busch _
@xcite , identical fermions must have an antisymmetric relative wave function and thus an @xmath76-wave interaction of short - range will vanish .
we therefore have to consider the @xmath77-wave channel , but we find that the effective harmonic interaction frequencies are very similar to the @xmath76-wave case in the two - dimensional setup that we consider here .
therefore we have chosen to parametrize the discussion of the thermodynamic quantities by the harmonic oscillator frequency of the two - body interaction itself .
one can then make the connection to a realistic system by working backwards through the model of busch _
et al._. our numerical results show that the low - temperature behavior reflects shell structure for the particle numbers we study .
the large - temperature equipartition limits are recovered for energy and heat capacity , and the transition from small to large temperature is fastest for smaller two - body interaction strength . the qualitative behavior is rather similar for bosons and fermions .
however , bosons always have a gap in the low - energy spectrum since one quantum of excitation of the relative motion must be antisymmetric and therefore forbidden .
this leads to a slower variation with temperature , since this gap has to be overcome before the number of available states goes up .
our results for the density of states turn out to scale with the number of relative excitation quanta in a manner that is very similar to the treatment of the many - body density of states for a uniform fermi gas .
surprisingly , the bosonic many - body level density scales similarly to the fermionic one , although the particle number enters differently in our interpolation formulas . in future studies it will be interesting to consider also multi - component systems , something which is simply done within the harmonic approach since the level counting can be factorized in the different components .
also , a study of the virial expansion based on the harmonic approach is currently on - going , both in two and three dimensions .
in addition , we note that one dimensional system have attracted a lot of attention recently due to their realization in ultracold atomic physics @xcite
. it would be interesting to test our prediction against some of the models that are being explored in the experiments and for which a number of exactly solvable many - body models are known @xcite .
furthermore , recent experiments studying ultracold few - body two - component fermi systems ( particle numbers of ten or less ) @xcite would be an interesting comparison for the harmonic approximation .
m. a. zaluska - kotur , m. gajda , a. orlowski , and j. mostowski , phys .
a * 61 * , 033613 ( 2000 ) ; j. yan , j. stat . phys . * 113 * , 623 ( 2003 ) ; m. gajda , phys . rev .
a * 73 * , 023603 ( 2006 ) ; j. r. armstrong , n. t. zinner , d. v. fedorov , and a. s. jensen , _ j. phys .
* 44 * ( 2011)055303 .
b. sutherland : _ beautiful models _ , ( world scientific publishing co. , singapore , 2004 ) ; r. j. baxter : _ exactly solved models in statistical mechanics _ , ( academic press , new york , 1982 ) ; v. e. korepin : _ exactly solvable models of strongly correlated electrons _ , ( world scientific publishing co. , singapore , 1994 ) .
t. stferle , h. moritz , k. gnter , m. khl , and t. esslinger , phys .
* 96 * , 030401 ( 2006 ) ; t. volz _
et al . _ , nature phys .
* 2 * , 692 ( 2006 ) ; g. thalhammer _
et al . _ ,
lett . * 96 * , 050402 ( 2006 ) ; c. ospelkaus _ et al .
lett . * 97 * , 120402 ( 2006 ) .
w. c. haxton and t. luu , phys .
lett . * 89 * , 182503 ( 2002 ) .
i. stetcu , b. r. barrett , and u. van kolck , phys .
b * 653 * , 358 ( 2007 ) ; i. stetcu , b. r. barrett , u. van kolck , and j. p. vary , phys . rev .
a * 76 * , 063613 ( 2007 ) ; y. alhassid , g. f. bertsch , and l. fang , phys . rev . lett . * 100 * , 230401 ( 2008 ) ; n. t. zinner , k. mlmer , c. zen , d. j. dean , and k. langanke , phys . rev .
a * 80 * , 013613 ( 2009 ) ; i. stetcu , j. rotureau , b. r. barrett , and u. van kolck , ann . phys . * 325 * , 1644 ( 2010 ) ; t. luu , m. j. savage , a. schwenk , and j. p. vary , phys . rev .
c * 82 * , 034003 ( 2010 ) ; j. rotureau , i. stetcu , b. r. barrett , m. c. birse , and u. van kolck , phys . rev .
a * 82 * , 032711 ( 2010 ) .
h. hu , p. d. drummond , and x .- j .
liu , nature phys .
* 3 * , 469 ( 2007 ) .
m. horikoshi , s. nakajima , m. ueda , and t. makaiyama , science * 327 * , 442 ( 2010 ) ; s. nascimbne _
et al . _ ,
nature * 463 * , 1057 ( 2010 ) ; n. navon , s. nascimbne , f. chevy , and c. salomon , science * 328 * , 729 ( 2010 ) ; c. cao _ et al .
_ , science * 331 * , 58 ( 2011 ) .
c. sanner _
et al . _ ,
105 * , 040402 ( 2010 ) ; t. mller _
lett . * 105 * , 040401 ( 2010 ) . m. j. h. ku , a. t. sommer , l. w. clark , and m. w. zwierlein , arxiv:1110.3309v1 . c. l. hung , x. zhang , n. gemelke , and c. chin , nature * 470 * , 236 ( 2011 ) .
d. a. mcquarrie : _ statistical mechanics _ , ( harpercollins , new york , 1976 ) .
k. huang : _ statistical mechanics _
( john wiley & sons , new york , 1987 ) .
s. s. m. wong : _ introductory nuclear physics _
, ( wiley interscience , new york , 1998 ) . r. v. e. lovelace and t. j. tommila , phys .
a * 35 * , 3597 ( 1987 ) ; g. baym and c. j. pethick , phys . rev
. lett . * 76 * , 6 ( 1996 ) .
h. fu , y. wang , and b. gao , phys .
a * 67 * , 053612 ( 2003 ) ; n. t. zinner and m. thgersen , phys .
a * 80 * , 023607 ( 2009 ) ; m. thgersen , n. t. zinner , and a. s. jensen , phys . rev .
a * 80 * , 043625 ( 2009 ) .
j. r. armstrong , n. t. zinner , d. v. fedorov , and a. s. jensen , europhys .
* 91 * , 16001 ( 2010 ) ; n. t. zinner , b. wunsch , d. pekker , and d .- w .
wang , arxiv:1009.2030v3 ; n. t. zinner _
_ , arxiv:1105.6264v1 .
a. g. volosniev _
et al . _ , j. phys .
b : at . mol .
44 * , 125301 ( 2011 ) ; a. g. volosniev , d. v. fedorov , a. s. jensen , and n. t. zinner , phys .
* 106 * , 250401 ( 2011 ) ; a. g. volosniev _
_ , arxiv:1112.2541v1 .
ho and e. j. mueller , phys . rev .
* 92 * , 160404 ( 2004 ) ; x .- j .
liu , h. hu , and p. d. drummond , phys . rev . lett . * 102 * , 160401 ( 2009 ) ; phys . rev .
a * 82 * , 023619 ( 2010 ) ; s .- g .
peng , s .- q .
li , p. d. drummond , and x .- j .
liu , phys .
a * 83 * , 063618 ( 2011 ) . | we describe a method to compute thermodynamic quantities in the harmonic approximation for identical bosons and fermions in an external confining field .
we use the canonical partition function where only energies and their degeneracies enter .
the number of states of given energy and symmetry is found by separating the center of mass motion , and counting the remaining states of given symmetry and excitation energy of the relative motion .
the oscillator frequencies that enter the harmonic hamiltonian can be derived from realistic model parameters and the method corresponds to an effective interaction approach based on harmonic interactions . to demonstrate the method
, we apply it to systems in two dimensions .
numerical calculations are compared to a brute force method that is considerably more computationally intensive . |
one of the goals for doing computer simulations is the production of information useful in the interpretation and design of experiments .
notwithstanding important issues regarding hamiltonian selection and parameterization , the interface of simulations with experiment is particularly challenging for quantum systems . the current monte carlo algorithms , whether they impose quantum particle statistics constraints or not , are performed either in real - time @xmath0 or in imaginary - time ( euclidean time ) @xmath1 . in real - time , the propagator @xmath2 for a system , described by a hamiltonian @xmath3 , oscillates wildly at long - times .
analytically , these rapid oscillations self - cancel , but a monte carlo process , as it is typically used , has difficulty achieving this cancellation . as a consequence , modifications of the basic algorithms have been proposed to extend the simulations as long as possible in the real - time domain @xcite . with these new algorithms ,
simulations typically produce dynamics extending to 2 to 3 times the natural periods of the systems . in imaginary - time , the propagator @xmath4 is diffusive and the rapid oscillations are avoided .
correlations functions @xmath5 , however , are now a function of imaginary - time , and such functions do not easily convey the actual dynamics of the system . in principle , real - time correlation ( green s ) functions @xmath6 can be obtained from the imaginary - time ones by the process of analytic continuation . in practice , this process is difficult because it is ill - posed and because the monte carlo data is incomplete and noisy .
recently , procedures were proposed to perform this analytic continuation @xcite .
these procedures draw heavily upon methods of bayesian statistical inference and the principle of maximum entropy to infer from imaginary - time correlation functions their associated spectral densities @xmath7 . through linear - response theory ,
the spectral densities represent the spectra associated with numerous real - time measurements of current - current , spin - spin , etc.correlations functions .
what apparently has not yet been tried is performing the hilbert transform of these spectral densities to obtain the frequency - dependent retarded correlation function and then fourier transforming this quantity to obtain the real - time correlation function . in this paper
, we will carry out these additional steps . by doing this ,
we hope to gain a greater understanding of the physical content present in the spectral density returned by the bayesian methods .
we expected that the resultant real - time information would be limited by the approximate and probabilistic nature of the analytic continuation methods .
we found , however , that the distance in real - time over which our results are valid is limited primarily by the ability of the simulation algorithm to produce good data . to interface profitably with the numerical analytic continuation ,
the simulation algorithm has to produce high quality data consistent with the assumptions of procedures .
the algorithm we used had problems doing this , and we will describe the measures taken to reduce this difficulty .
even so , in most cases we were able to extend in real - time up to factor of 10 natural periods of the physical systems .
longer extensions are possible and require longer monte carlo runs . for present purposes , we had no physical motivation to do so . in section ii
, we will discuss the various models studied .
we simulated a particle moving in single harmonic and double - well anharmonic potential and a collection of particles moving in chains of these potentials .
for these models we know the exact solutions . by calculating their properties numerically ,
we can benchmark our methods . certain properties of a single double - well potential , like the tunnel splitting , are easily obtained numerically . the phase diagram for a chain of such oscillators
is also known @xcite .
this type of chain can exist in a symmetric or displacive state . in section iii ,
we summarize the numerical analytic continuation procedure we used and discuss our simulation technique . modifying the simulation technique to be more naturally ergodic and to produce data with short statistical auto - correlation times was the most difficult and restrictive part of our study .
we present our results and conclusions in sections iv and v.
we simulated five hamiltonians .
one was that for a single harmonic oscillator @xmath8 which has the natural frequency @xmath9 .
another was that for a chain of @xmath10 such oscillators @xmath11 fourier transforming the displacements @xmath12 and momenta @xmath13 , we can of course rewrite this second hamiltonian as @xmath14 where @xmath15 and @xmath16 . in this form
, the hamiltonian is explicitly expressed as a collection on @xmath10 independent simple oscillators .
the natural frequency of an oscillator is @xmath17 .
the third hamiltonian was a variant of the harmonic chain @xmath18 which after fourier transforming the displacements becomes @xmath14 where @xmath19 . in this form ,
the hamiltonian is again explicitly expressed as a collection on @xmath10 independent simple oscillators but with a dispersion relation that has a non - zero frequency at @xmath20 .
the two other hamiltonians were a single , symmetric , double - well potential @xmath21 which has well bottoms at @xmath22 and a barrier height of unity at @xmath23 , and a chain of such potentials @xmath24 in the chain , @xmath25 sets the barrier height of every double - well . in both chains we assumed periodic boundary conditions and disallowed particle exchange . for these hamiltonians ,
our simulations obtained estimates of the imaginary - time displacement - displacement green s function @xmath26 here , the angular brackets denote thermal averaging .
it is more convenient and illuminating to work with the spatial fourier transform @xmath27 of @xmath28 and it is known @xcite that @xmath29 where @xmath30 is the spectral density .
this function has the properties that @xmath31 the odd symmetry of @xmath30 allows us to re - express ( [ eq : the_equation ] ) as @xmath32 } { 1-e^{-\beta\omega } } \frac{a_k(\omega)}{\omega } \label{eq : the_eq}\ ] ] and it is straightforward to show that @xmath30 obeys the sum rule @xmath33 the spectral density @xmath30 is also related to the frequency fourier transform @xmath34 of the real - time , retarded green s function @xcite @xmath35 \rangle\ ] ] via @xmath36 where @xmath37 , from which it follows that @xmath38 for an individual harmonic oscillator of frequency @xmath17 , the eigenstates and energies are exactly known , and all the quantities in the above paragraph are known analytically : @xmath39 \label{eq : gvstau}\\ a_k(\omega)&= & \pi [ \delta(\omega-\omega_k)-\delta(\omega+\omega_k)]/m\omega_k \label{eq : avsomega}\\ \hat g_k^r(\omega ) & = & \frac{1}{2m\omega_k } \left(\frac{1}{\omega-\omega_k+i\eta } - \frac{1}{\omega+\omega_k+i\eta}\right ) \label{eq : gvsomega}\\ \hat g_k^r(t ) & = & \theta(t)\frac{1}{m\omega } \sin(\omega_kt ) e^{-\eta t } \label{eq : gvst}\\ \langle x_k^2\rangle & = & \frac{1}{2m\omega_k } \coth\left(\beta\omega_k/2\right ) \label{eq : meansquare}\end{aligned}\ ] ] the eigenstates and energies of a single harmonic oscillator have definite , well - known characteristics : because the potential is symmetric about @xmath23 , the eigenfunctions have alternating parity .
the ground state has even parity and an energy @xmath40 .
the energies of the excited states are regularly spaced at intervals of @xmath17 .
the double - well potential is also symmetric about @xmath23 , and its eigenstates also alternate in parity with the ground state again having even parity .
the precise details about the energy spectrum , however , are only available numerically .
when these states lie below the barrier , and particularly for deep wells , they group into widely - separated , nearly degenerate pairs . the separation in energies within and between pairs
is called the _ tunnel spitting_. the spectral density is dominated by terms with matrix elements involving states 0 and 1 .
contributions from matrix elements of @xmath41 involving ( 0,3 ) , ( 3,4 ) , ( 0,5 ) , ( 3,5 ) , ( 5,6 ) , etc .
have smaller contribution to the spectral density because of smaller overlap between the eigenstates . additionally , most can be `` frozen - out '' by making the temperature at least comparable to @xmath42 .
this temperature range is the one in which we generally worked .
the nature of the energy levels and eigenstates is schematically represented in fig . 1
. the spatial fourier transformation of the hamiltonian of the double - well chain does not produce as system of independent oscillators .
this is the essence of its non - linearity .
the model , however , has an interesting zero - temperature phase diagram as a function of the parameters @xmath25 and @xmath43 @xcite .
roughly , @xmath25 is a measure of the barrier height relative to the frequency of inter - site oscillation and the frequency of oscillation associated with the well - bottom .
when the barrier height is large , the particles collectively are displaced to the left or the right of their classical equilibrium positions in a broken symmetry state characterized by a non - zero value of the mean - squared displacement .
as the barrier height is lowered , a critical value is reached where because of zero - point motion and tunneling , the particles collectively make a transition to a state where the mean - squared displacement of each is zero . accordingly , a simulation of the chain , done in one of these two thermodynamic phases , is expected to exhibit different quantum characteristics in the spectral density .
our monte carlo simulations will be based on the feynman path integral formulation of quantum mechanics . in imaginary - time , this formulation represents the partition function @xmath44 as @xmath45 } \label{eq : partition}\ ] ] where @xmath46 = \int_0^\beta d\tau\ , h[x(\tau ) ] \label{eq : action}\ ] ] is the action corresponding to the path @xmath47 and @xmath48 $ ] represents the path dependence of the hamiltonian .
the monte carlo method is used to perform the integration over the paths in ( [ eq : partition ] ) .
it is applied after the integral in ( [ eq : action ] ) is approximated by a sum over @xmath49 steps in imaginary - time each of length @xmath50 and the momenta are approximated by a forward - difference approximation between successive displacements in imaginary - time : @xmath51 for a one - dimensional system of @xmath10 particles , the action becomes @xmath52 this form is similar in appearance to a classical two - body potential energy defined on a @xmath53 lattice where at a given @xmath54 the particles interact through the potential energy function of the original problem ( scaled by @xmath50 ) , and at a given position they interact by a harmonic potential with a spring constant equal to @xmath55 . for a single particle ,
the summation over the spatial coordinate @xmath56 is dropped and the discretized action has the interpretation of a chain where the particles at imaginary - time move in a potential @xmath57 while interacting with particles at neighboring times by a harmonic potential with spring constant @xmath58 . for the models we are considering ,
the discretized actions are : 1 .
single harmonic oscillator : + @xmath59 ^ 2 + { \gamma\over 2}x_{j}^2,\ ] ] 2 .
harmonic chain : + @xmath60 ^ 2 + { \gamma\over 2}\left(x_{i+1,j}-x_{i , j}\right)^2,\ ] ] 3 . harmonic chain
plus on - site oscillator : + @xmath60 ^ 2 + { \gamma\over 2}\left(x_{i+1,j}-x_{i , j}\right)^2+\frac{1}{2}x_{i , j}^2,\ ] ] 4 . single double - well potential : + @xmath61 ^ 2 + { 1\over 4}\left(x_{j}^2 - 1\right)^2.\ ] ] 5 .
double - well chain : + @xmath62 ^ 2 + { \gamma\over 2}\left(x_{i , j}-x_{i , j+1}\right)^2 + { 1\over 4}\left(x_{i , j}^2 - 1\right)^2.\ ] ] the simplest way to apply the monte carlo method is to move repeatedly from point to point on the space - time lattice , at each point propose a change in the coordinate , @xmath63 , and accept the change via the metropolis algorithm with probability @xmath64 $ ] where @xmath65 is the change in the value of the action proposed by the proposed change @xcite .
this method is often called _ path - integral monte carlo _ ( pimc ) .
an alternative to the monte carlo evaluation of the path - integral is a molecular dynamics evaluation @xcite . here , a fictitious momentum @xmath66 is associated with each point to define a pseudo hamiltonian @xmath67\ ] ] and standard molecular dynamics techniques are used to sample phase space .
the approach takes advantage of the classical nature of the fields in the path - integral formulation and produces the correct statistical mechanics because in classical statistical mechanics the momentum degrees of freedom can be integrated out of the partition function .
the method is often called _ path - integral molecular dynamics_. at the core of the method we used is the hybrid monte carlo approach suggested by duane et al .
@xcite which combines the molecular dynamics approach with the monte carlo procedure to obtain the best features of both methods .
the general expectation is faster equilibration of the simulation and shorter auto - correlation times between measured quantities . with this method
, the following steps are cycled : for a given set of @xmath68 the corresponding pseudo momenta are assigned values randomly from a maxwell - boltzmann distribution for the velocities .
the energy is computed .
next , both the momenta and displacements are evolved by molecular dynamics for some pseudo time @xmath69 .
the energy is recomputed .
then the evolved displacements are accepted with probability @xmath70 $ ] , where @xmath71 is the difference in energy between the initial and final configuration .
normally , molecular dynamics is energy conserving so the evolved displacements would always be accepted . a guiding idea behind the hybrid method is to use for the molecular dynamics with emphasis on fast integration , as opposed to accurate integration , and to adjust the size of these steps @xmath72 and @xmath69 so the monte carlo decision accepts @xmath73 of the configuration .
the molecular dynamics method globally updates all the displacements and is a computationally efficient procedure .
the monte carlo procedure maintains detailed balance to ensure proper equilibrium averages and filters out the results of `` bad '' integrations .
we found , however , that this simple form of the hybrid method was inadequate for present purposes .
the output of our simulation is to be used as input to maximum entropy procedures to execute the analytic continuation .
as we will discuss below , the analytic continuation problem is an ill - posed problem and hence is very sensitive to the size of the errors associated with the input data . for a fixed amount of computer time ,
reducing the size of the error efficiently by a monte carlo method requires shortening the auto - correlation times between measurements . in our computed green s function , in spite of small estimates for the error bars , we would often see small ( within the error bars ) unphysical sawtooth - like structure in regions about @xmath74 .
following a simple procedure suggested by neal @xcite , we could generally remove this structure and also be more ergodic .
his suggestion was after each monte carlo decision to reverse the direction of the molecular time integration , i.e. , @xmath75 , with probability @xmath76 .
a smaller improvement is achieved by not using a fixed length for the time integration in the molecular dynamics simulation but rather to choose the length randomly from the interval @xmath77 where @xmath69 and @xmath78 are chosen so the monte carlo acceptance rate is in the 90 to 95 per cent range .
we will refer to the combined method as the _ hybrid path - integral monte carlo method _ ( hpimc ) .
adjusting the monte carlo acceptance rate is not the entire story .
first , it seems best to insure @xmath69 is several times larger than the natural period associated with the slowest significant modes in the systems and then choose @xmath78 to fix the acceptance ratio .
we remark that fahy and hamann @xcite observed for the standard hybrid method the existence of a critical time @xmath79 ( dependent of model parameters ) demarking non - ergodic and chaotic behavior in the results of the time integration . for a harmonic system
, @xmath79 is infinite which suggests the inapplicability of the method to a harmonic system .
we observed the behavior they found but whether @xmath69 was larger or smaller than @xmath79 had only small consequences on our measured results .
as we illustrate below , we achieved very accurate results for the harmonic models .
we also remark that the results of the simulations depend on the size of @xmath50 . by performing simulations for several different values of @xmath50 , we could in principle extrapolate the results to the @xmath80 limit .
we did not do this but instead observed that simulations performed with different values of @xmath50 gave very similar results .
the maximum entropy method @xcite is used to regularize the solution of ( [ eq : the_eq ] ) . dropping the subscript @xmath81 for convenience , we rewrite this equations as @xmath82 \label{eq : the_integral_eq}\ ] ] where the kernel @xmath83 } { 1-e^{-\beta\omega}}\ ] ]
because both the kernel and @xmath84 are regular at @xmath85 , we solve for @xmath84 and then trivially find @xmath7 . for discrete values of @xmath54 and @xmath86 we approximate ( [ eq : the_integral_eq ] ) as @xmath87 where @xmath88 , @xmath89 , and @xmath90 .
the monte carlo method will return estimates @xmath91 of @xmath92 and estimates of the sample variance @xmath93 on the @xmath92 . with this information
, a natural solution path to @xmath94 would be to find the values of @xmath94 that minimize @xmath95 this approach , however , almost always fails .
one reason is that it ignores the strong correlations that normally exist between the measured values of @xmath92 , i.e. , values of the green s function at different imaginary - times . at the very least
, we must modify ( [ eq : chi_1 ] ) to be @xmath96_{i , j}(\bar g_j - g_j ) \label{eq : least - square}\ ] ] where @xmath97 is the measured covariance among the values of @xmath98 .
the @xmath56-th diagonal element of @xmath99 is simply @xmath100 .
this modification of the definition of @xmath101 while necessary is insufficient .
the difficulty is the inverse problem , that is , solving ( [ eq : the_integral_eq ] ) for @xmath7 , is ill - posed .
this condition is caused by the exponential character of the kernel at large values of @xmath86 .
at large @xmath86 , large variations in @xmath7 make little change in @xmath5 .
the simulation , on other other hand , gives noisy and incomplete information about @xmath5 , and hence for a given set of @xmath91 , an infinite number of @xmath94 will satisfy the least - squared estimate ( [ eq : least - square ] ) .
the next level of solution seeks to regularize the minimization of @xmath101 by constraining it , i.e. , minimizing @xmath102 where the @xmath103 are lagrange multipliers and the @xmath104 are functions of @xmath94 representing possible constraints on the solution .
typical constraints include smoothness , non - negativity , sum rules , moments , etc .
the difficulty with this approach is choosing the lagrange multipliers . _
ad hoc _ choices are commonplace .
often small changes in the values of these parameters produce massive changes in the results .
the maximum entropy approach follows from the observation that the spectral density is interpretable as a probability density function .
the principle of maximum entropy states the probabilities should be assigned in such a way as to maximize @xmath105 here , the @xmath106 , called the _ default model _ , set the location of the maximum of @xmath107 and the value of @xmath107 at this point to be zero .
the default model is solution for @xmath94 in the absence of other constraints on @xmath94 .
the method of maximum entropy maximizes @xmath108 to fix @xmath109 , an ad hoc procedure called _ historic _ maximum entropy is often used @xcite .
an more modern alternative is the bayesian - based _ classic _ maximum entropy which uniquely determines @xmath109 provided certain conditions are meet @xcite . under these conditions solution for @xmath94
is the most probable one .
unfortunately , these conditions seem often violated in the analytic condition problem . accordingly , to estimate the @xmath94 , we adopted a procedure suggested by bryan . in _
bryan s method _
@xcite , for a given value of @xmath109 , we find the @xmath110 that maximizes @xmath111 .
for the solution to ( [ eq : the_matrix_eq ] ) , we take @xmath112 \label{eq : the_solution}\ ] ] where @xmath113 $ ] is the probability of @xmath109 given the data @xmath114 .
bayesian analysis shows that @xmath115 = \pr[\alpha]\int { \cal d}a\ , \frac{e^q}{z_lz_s(\alpha)}\ ] ] where @xmath116 is the normalization factor for @xmath117 , @xmath118 is the normalization factor for @xmath119 , and @xmath120 $ ] is jeffreys prior .
details on the computation of this joint probability are discussed elsewhere @xcite . with this function ,
the integral ( [ eq : the_solution ] ) is performed numerically . the most difficult part of the problem is not evaluating the maximum entropy equations but satisfying the statistical assumptions on which they are based .
the principal assumption is @xmath121= \frac{e^{-\frac{1}{2}\chi^2}}{z_l } \label{eq : likelihood}\ ] ] the meaning of this assumption is the measured values of @xmath92 are statistically independent and distributed according to a multi - variable gaussian distribution function defined by the covariance matrix @xmath99 .
proper estimation of @xmath99 is paramount . under normal circumstance
the data produced by the simulations do not satisfy these assumptions .
the procedures we use to have the data approximate these assumptions are discussed elsewhere @xcite .
when we have proper data , our solution ( [ eq : the_solution ] ) usually shows good insensitivity to the choice of the default model .
additionally , the historic and classic maximum entropy solutions usually agree well with it .
to determine spectral properties of models listed in section ii , we performed hpimc simulations with up to 800 bins of data ( @xmath122 ) , each with up to 4000 measurements ( @xmath123 )
. simulations with large bin - sizes were necessary to avoid nonergodic behavior of the hpimc method when used for chains with double well potentials close to the zero - temperature phase transition point .
furthermore , we set the value for the imaginary time step to @xmath124 .
this choice on the one hand was small enough to avoid errors associated by the discretization of otherwise continuous imaginary time scale @xmath54 , and on the other hand was large enough to avoid unphysical correlations between successive imaginary - time measurements of the green s function @xmath5 .
since our calculations were performed at the inverse temperatures @xmath125 between 1 and 10 , the corresponding number of imaginary time steps @xmath126 was between 40 and 60 . for a successful application of the hpimc
it is crucial to choose the proper value of the pseudo - time @xmath69 and the size of its step @xmath127 in the molecular dynamic part of the simulation .
following fahy and hamman , we determined the value of the critical value @xmath79 for each case under consideration and then took @xmath128 to avoid running the simulation in a non - ergodic regime .
typical values for @xmath69 were between 5 and 15 for double - well cases listed below .
we stress that there was not much difference in the quality of the hpimc data if we chose @xmath128 or chose @xmath129 .
in addition , we obtained good data for the harmonic wells by choosing @xmath130 even though for this particular case @xmath131 .
if we define @xmath132 as the smallest nonzero frequency of the system , optimal values of the step size @xmath72 are between @xmath133 and @xmath134 .
larger values of @xmath72 lead to larger errors in the pseudo - time propagation which then lead to small acceptance ratios .
smaller values of @xmath72 lead to longer compution times . unless specified otherwise , we always chose @xmath135 .
we emphasize that the hpimc method is insensitive of the choice of the mass @xmath136 associated with the fictitious momentum @xmath66 . in fig . 2
, we show the displacement - displacement green s function @xmath5 , for a single harmonic oscillator , obtained by evaluating ( [ eq : gvstau ] ) for various values of @xmath137 , as a function of the imaginary - time variable @xmath54 .
these curves look similar to the green s functions that we obtained numerically for this and the other models : for some parameters and temperatures , the @xmath5 varies little as a function of @xmath54 ; for others , it varies rapidly at the ends of the interval @xmath138 and is flat in the middle with values nearly equal to zero ; and for still other parameters , it has a featureless , parabolic looking shape .
the common features of these curves have several significant implications for the analytic continuation problem .
first , we remark that from the quantum monte carlo simulations we obtain estimates of @xmath5 only at a relatively small number of discrete values @xmath139 .
the smoothness of the curves implies that the @xmath140 at neighboring values of @xmath139 are correlated .
the computation of the covariance matrix in ( [ eq : least - square ] ) is thus an important part of the of the analysis of the data .
while the correlations among the different @xmath54 values of @xmath5 make the interpretation of the assignment of an `` error - bar '' to a given @xmath54 value delicate , such an assignment illustrates several difficulties inherent in the data that help to make the analytic continuation of the data often very difficult . in the case where @xmath5 is nearly flat ,
the errors bars mean that a number of values of @xmath5 are `` within the error - bars '' of each other .
this situation , along with the correlations implied by sizable off - diagonal elements of the covariance matrix , means that only a subset , and often a small subset , of the measured values of @xmath5 represent independent data useful for the analytic continuation procedure .
the analytic continuation near the classical limit can be very difficult .
the situation with the rapid end - point variation and the flat nearly zero values is another difficult case .
again the flat region generates a lose of useful values of @xmath5 and the smallness of @xmath5 in this region can engender situations where the error bars would imply that during the simulation estimates of @xmath5 , which must be non - negative , were derived from ensemble values that included negative ones .
the monte carlo algorithms in fact do not produce negative values but do produce highly skewed fluctuations about the mean .
the gaussian assumption for the likelihood function in ( [ eq : likelihood ] ) thus can only be approached in the limits of a large number of independent measurements when the central limit controls the data distribution .
the ratio of the mean value to the estimated variance ( signal to noise ratio ) also indicates that the most effective data comes from those in the rapidly decreasing region . at low temperatures ,
the analytic continuation problem can become very difficult .
the details of the simulation algorithm can also impact the quality of the results and data . in fig . 3
, we show @xmath141 as a function of @xmath54 obtained by two closely related simulation techniques for a single harmonic oscillator .
we remark that the scale of abscissa is @xmath142 of that of fig . 2 and the ordinate shows @xmath54 only in a narrow region at the symmetry point @xmath143 .
the dashed curve is the analytic result obtained from ( [ eq : gvstau ] ) .
the data is represented by square markers were obtained by the hpimc method without the time - reversed step ; the data represented by the circles were obtained with the hpimc method with the time - reversed step . in each case , the same number of monte carlo steps were made .
one sees that the fluctuations with the hpimc method without the time - reversed step are larger and that the error bars associated with the results suggest a dip into non - negative values of @xmath5 .
more significantly , the results deviate from the exact curve by more than one standard deviation in the immediate vicinity of @xmath144 .
the analytic continuation result for @xmath7 from the data partially shown in fig .
3 is shown in fig .
4 . from ( [ eq :
avsomega ] ) , the spectral density should be @xmath145
. the solid curve is obtained from the hpimc algorithm and shows a broadened delta - function at the right location with nearly the correct weight .
the fraction of a per - cent difference from the correct weight is most likely a consequence of the small error caused by discretizing the imaginary - time derivatives . on the other hand ,
the dashed curve , which is obtained from the data obtained from the hipmc algorithm without the time - reversed step , is broader , located incorrectly , and has a larger discrepancy in its weight .
the increased breadth is a consequence of the larger variance in the measured data .
the incorrect location and poorer weight is a consequence of the small deviation from the exact value near @xmath144 , which in turn is a consequence of the monte carlo algorithm performing badly . in fig . 5 , with the data obtained from the hpimc algorithm , we show the full analytic continuation form @xmath5 to @xmath146 . in fig . 5a , the data ( open circle )
is compared with the exact results for @xmath5 ( solid line ) obtained from ( [ eq : gvstau ] ) . in fig .
5b , the dashed curve is the @xmath7 obtained by the analytic continuation procedure , while the solid line is a lorentzian at the same location .
the real part of @xmath147 is shown in fig . 5c , where the dashed line is the quantum monte carlo results and the solid line is an analytic result obtained from the lorentzian from fig . 5c .
the width @xmath148 of the lorentzian shown in fig .
5b was adjusted so the @xmath85 values of the two curves agreed .
the single adjustment produced remarkably good agreement at high frequencies .
the principal differences between the two curves are at @xmath149 where divergences should exist as indicated by ( [ eq : gvsomega ] ) .
finally , @xmath146 is shown in fig .
the solid line was obtained analytically and used the lorentzian of fig . 5b , while the dashed line is the quantum monte carlo result .
the agreement between the results is satisfying and comparable in quality to that possible by real - time quantum monte carlo methods .
the width of the lorentzian was @xmath150 , i.e. , about 7 times the natural period of the harmonic oscillator .
this width is still controlled by the size of the variance in the measured values of @xmath5 . for the case of the harmonic chain , we computed the green s function @xmath27 for each independent wave number and from it found the corresponding spectral density @xmath30 .
as with the single oscillator , this function should be a delta - function located at @xmath17 and have a weight equal to @xmath151 .
as in the single oscillator case , instead of finding a delta - function , we found a lorentzian - like peak at the correct location with the correct weight .
the peaks for different values of @xmath81 , however , had different widths .
in general , the peak widths increased with increasing @xmath81 , correlating with the increased variance associated with the @xmath27 .
the spectral densities for the three lowest values of @xmath81 are shown in fig .
the peak positions as a function of @xmath86 give the phonon dispersion relation .
our determination of this relation is compared to the exact result in fig .
the agreement is excellent .
there is a difficulty that must be mentioned . at @xmath20 and @xmath152 , @xmath27 is flat and the weight of the peak approaches infinity . for @xmath20 ,
not surprisingly , we were unable to do the analytic continuation .
this situation is why we considered the model defined by ( [ eq : model3 ] ) . here , @xmath17 at @xmath20
is not zero and the determination of @xmath30 , and subsequently @xmath17 , is possible for all values of @xmath81 .
the dispersion relation found from the analytic continuation agrees very well with exact result .
for a single , double - well potential , the spectral density can give direct information about tunneling processes .
this situation is illustrated in fig .
8 where the spectral densities for the hamiltonian , described by ( [ eq : model4 ] ) , are shown for several different values of the parameter @xmath43 .
it is straightforward to discretize schroedinger s time - independent differential equation for this potential and find the eigenvalues of the resulting eigenvalue equation .
we adjusted the model parameters so only a very few ( usually one ) of the lowest eigenstate lied below the barrier height , similar to the situation depicted in fig .
( for deeper wells , our simulation methods would get stuck in one well or the other for large numbers of monte carlo steps , and hence the algorithm effectively lost ergodicity . ) in the cases reported , the temperatures of the simulations were also less than the separation between the two lowest lying pairs of eigenstates .
thus , our spectral densities only exhibited the transition between these two states , and the position of the peak gives a direct measure of the lowest frequency tunnel splitting .
this position agreed very well with the exact value calculated from schroedinger s equation .
additionally , the weight of the peak also agreed well with the sum rule ( [ eq : sumrule ] ) . for a chain of double - well potentials ( [ eq : model5 ] ) ,
still a different situation presents itself . at zero temperature ,
depending on the values of @xmath25 and @xmath43 , the system exists in either a broken symmetry phase , in which the mean - squared displacement is non - zero , or in a symmetric phase , in which the mean - squared displacement is zero .
the model has a quantum phase transition .
we found that simulation , if performed in the broken symmetry phase , stuck in one well or the other for large numbers of monte carlo steps , making it difficult to collect the amount of statistically independent information to do the analytic continuation . in fig .
9 , we report the @xmath30 for two simulations done in the symmetric phase .
the one in fig .
9b is close to the zero - temperature phase boundary .
as one moves closer to the @xmath153 phase boundary by increasing @xmath25 , the @xmath20 peak moves towards @xmath85 .
this movement is a consequence of the decreasing probability for tunneling between the two minima of the double - well potential as the barrier height is increased .
we used methods of bayesian statistical inference and the principle of maximum entropy to analytically continue imaginary - time green s function generated in quantum monte carlo simulations to obtain the real - time displacement - displacement green s functions . for test problems , we considered chains of harmonic and anharmonic oscillators whose properties we simulated by a hybrid path - integral quantum monte carlo method ( hpimc ) . from the imaginary - time , displacement - displacement green s function , we first obtained its spectral density . for harmonic oscillators
, we demonstrated that the peaks of this function were in the correct position and their area satisfied a sum rule .
additionally , as a function of wavenumber , the peak positions followed the correct dispersion relation . for a double - well oscillator , we demonstrated the peak location correctly predicted the tunnel splitting . transforming the spectral densities to real - time green s functions ,
we conclude that we can predict the real - time dynamics for length of times corresponding to 5 to 10 times the natural period of the model .
the length of time was limited by the simulation algorithm and not directly by the analytic continuation procedure .
the simulation algorithm influences the results in at least two ways .
one way is the ease with which statistically independent , gaussian - distributed data is obtained . in our experience , of the various quantum monte carlo methods we have used , path - integral monte carlo methods ( pimc )
tend to produce data with long - ranged correlations , thus making statistical independence sometimes difficult to achieve .
achieving statistical independence is important for proper error estimation because of the sensitivity of ill - posed problems to errors in the data . with out statistical indepence these error are usually underestimated . that the data is gaussian - distributed is necessary to satisfy the assumptions of the analytic continuation procedure .
currently , large amounts of data are used to force the simulation data to have proper statistical properties .
the method of binned averages @xcite , with many measurements in a bin , is used to achieve statistical independence .
the central limit theorem is used to obtain a gaussian distribution of the binned averages . because of the low computational intensity necessary for the test cases considers ,
producing the necessary large amounts of data was not problematic .
a second way the algorithm can influence the results is through broken ergodicity and unphysical results . here
, we are referring to the ragged structure we saw in @xmath5 with the hybrid method and to the small systematic difference between the exact and computed results .
we have not previously seen similar problems .
small modifications in the simulation algorithm removed the problems but it was at first difficult to determine the source of the difficulty . because the purpose of the research was not algorithmic development , we did not do any comparisons of hpimc and other possible methods to determine relative efficiency and other merits .
again , the computation times for these simulations is small ( a few hours on a modest workstation ) and so we were unmotivated to make such comparisons .
other recently suggested approaches , e.g. , @xcite , should be considered as part of further studies . in general , we believe we have demonstrated that analytically continuing imaginary - time correlation functions , obtained from a quantum monte carlo simulation , to obtain real - time correlation functions is a feasible alternative to obtaining such real - time functions directly from a quantum monte carlo simulation done in real - time .
one advantage in choosing this approach appears to be the longer length of time over which the real - time information is faithful to the correct result . of course
, this conclusion is based on one study of simple models ; however , this study does strongly indicate the direction for further and more extensive work .
a brief review is given by b.j .
berne and d.thirumalai , ann .
* 47 * , 401 ( 1986 ) .
more recent works include : j.d .
doll , r.d .
coalson , and d.l .
freeman , j. chem .
* 87 * , 1641 ( 1987 ) : j.d . doll , d.l .
freeman , and t.l .
beck , adv .
phys . * 78 * , 61 ( 1990 ) ; c.h .
mak and r. egger , phys .
a , to appear ; and references therein . | we used methods of bayesian statistical inference and the principle of maximum entropy to analytically continue imaginary - time green s function generated in quantum monte carlo simulations to obtain the real - time green s functions . for test problems , we considered chains of harmonic and anharmonic oscillators whose properties we simulated by a hybrid path - integral quantum monte carlo method . from the imaginary - time displacement - displacement green s function , we first obtained its spectral density . for harmonic oscillators
, we demonstrated the peaks of this function were in the correct position and their area satisfied a sum rule .
additionally , as a function of wavenumber , the peak positions followed the correct dispersion relation . for a double - well oscillator
, we demonstrated the peak location correctly predicted the tunnel splitting . transforming the spectral densities to real - time green s functions ,
we conclude that we can predict the real - time dynamics for length of times corresponding to 5 to 10 times the natural period of the model .
the length of time was limited by an overbroadening of the peaks in the spectral density caused by the simulation algorithm . |
an understanding of electron collision processes with oxygen molecules is important because of its role in chemistry of electrical discharge and the upper atmosphere . in recent attempts to an operate electrical discharge oxygen - iodine laser ,
a population inversion of iodine atoms was achieved by a near resonant energy transfer via the @xmath7 @xmath8 @xmath9 process .
in contrast to the traditional liquid chemistry singlet oxygen generator @xcite , recent studies @xcite utilize flowing electric discharges where electron collisions with o@xmath0 excited electronic states can be important @xcite . in such conditions ,
even highly excited metastable states of o@xmath0(@xmath10 , @xmath11 , @xmath12 ) may play roles @xcite , in addition to the lower o@xmath0 @xmath2 and @xmath3 excited states .
however , most previous work has concentrated on electron collisions with the o@xmath13 ground state , so our knowledge of electron impact transitions from the o@xmath14 and @xmath3 excited states is limited . the past experimental and theoretical works concerning electron o@xmath0(@xmath15 , @xmath2 ) collisions were summarized by brunger and buckman @xcite .
one work on the excited electronic states is measurement of differential and integral cross sections at 4.5 ev for excitation from the o@xmath0 @xmath2 state to the @xmath3 state by hall and trajmar @xcite .
their value is more than an order of magnitude larger than that for the @xmath15 @xmath16 @xmath3 cross section .
also , khakoo et al .
@xcite studied the energy - loss spectrum for electron impact excitation on discharged o@xmath0 and assigned the transitions o@xmath0 @xmath2 ( @xmath17=0 ) @xmath16 o@xmath0 @xmath18 ( @xmath19=0,1, .. 7 ) .
burrow @xcite and beli and hall @xcite studied dissociative electron attachment with the o@xmath0 @xmath2 state .
the later authors found that dissociation proceeds to 3 different limits , o@xmath20+o@xmath21 , o@xmath20+o@xmath22 and o@xmath20+o@xmath23 .
in contrast to the situation in electron collisions with the excited oxygen molecule , a lot of work has been performed on the ground state o@xmath0 , both experimental @xcite and theoretical @xcite .
notably , noble , burke and their co - workers extensively applied their r - matrix method to the electron o@xmath0 collision problems during 1992 - 1996 @xcite .
they studied electronic excitation processes from the o@xmath13 ground state to the @xmath2 , @xmath3 , @xmath10 , @xmath11 and @xmath12 states using the fix - bond r - matrix method @xcite and applied the non - adiabatic r - matrix method to vibrational excitations process of the @xmath24 @xmath16 @xmath25 transitions @xcite .
they also calculated differential cross sections for elastic electron collisions of the @xmath15 state @xcite as well as impact excitations from the @xmath15 state to the @xmath2 and @xmath3 states @xcite .
the effect of nuclear motion was included in the former elastic cross sections by vibrational averaging of the t - matrix @xcite .
other than these r - matrix calculations , teillet - billy et al .
@xcite applied effective range theory ( ert ) to excitations from the @xmath15 to the @xmath2 and @xmath3 states . because of the different treatment of the o@xmath26 resonances , the ert results deviate from the r - matrix cross sections at energies above 5 ev .
given the importance of electron collisions with excited o@xmath0 molecules , we perform r - matrix calculations for electron o@xmath0(@xmath2,@xmath3 ) collisions . we chose the r - matrix method because it has been successfully applied to many electron - molecule collisions including e - n@xmath0 , n@xmath0o and h@xmath0o @xcite .
the fixed - bond method was employed in this work , because it gave reasonably good results in previous studies @xcite for transitions from the o@xmath0 @xmath15 state to the @xmath2 , @xmath3 state , and the ` 6 ev states ' ( @xmath10+@xmath11+@xmath12 ) .
in addition to these 6 low lying o@xmath0 target states , previous calculations included three higher excited target states of o@xmath0 @xmath27,@xmath28 and @xmath29 , in order to improve quality of the r - matrix calculations @xcite . in this work
, we use a valence complete active space description of the o@xmath0 target states and add other valence target states , @xmath30,@xmath31,@xmath32,@xmath33 , in our calculations . since excitation energies of some of these states are lower than those of @xmath27,@xmath28 and @xmath29 states , some improvement can be expected by inclusion of these extra @xmath34 target states . in principle , a complete valence active space is not sufficient for the description of these targets , because some of them are mixed with n=3 rydberg states as described in buenker and peyerimhoff @xcite .
since expansion of this active space increases the calculation cost considerably , we limit ourselves here the inclusion of the valence states to test the effects of the higher excited states . in this paper , details of the calculations are presented in section 2 , and we discuss the results in section 3 comparing our results with previous theoretical and available experiments . then the summary is given in section 4 .
the r - matrix method has been described extensively in the previous literature @xcite , so here we only repeat the outline of the method . in this method ,
configuration space is divided by two regions according to the distance @xmath35 of the scattering electron and the center of mass of the target molecule having @xmath36 electrons . in the inner region
@xmath37 , the @xmath36 + 1 electrons problem is solved by usual quantum chemistry method with slight modifications to account for existence of boundary at @xmath38 . in the inner region ,
the total @xmath36 + 1 electrons wave functions are represented by @xmath36-electron ci target wave functions augmented by diffuse functions . here
the target wave functions are contained in the sphere @xmath37 , whereas the diffuse functions overlap the boundary at @xmath39 in order to describe the scattering electron . in the outer region @xmath40
, the problem is reduced to single electron scattering , ignoring exchange of the scattering electron with the target electrons .
interaction of the scattering electron and the target is considered through static multipolar interaction terms which introduce inter - channel couplings .
the wave functions obtained in the inner region are converted to the r - matrix at the boundary @xmath35=@xmath41 , then the coupled radial schrdinger equations are solved so as to extract scattering information at the asymptotic region . in the inner region , the @xmath36
+ 1 electronic wavefunctions are expanded as , @xmath42 where @xmath43 is an antisymmetrization operator , @xmath44 are the @xmath36 electron target ci wave functions , @xmath45 are the diffuse functions representing wave functions of a scattering electron , and @xmath46 are bound @xmath36 + 1 electron wave functions , while @xmath47 and @xmath48 are variational coefficients . in this expression ,
the first term represents the scattering of an electron from and to the asymptotic region .
the second summation involves purely @xmath49 integrable terms .
in addition to the target molecular orbitals included in the ci wavefunctions in the first summation , some extra target virtual orbitals are usually included in @xmath46 in order to account for short range polarization effects .
we used a modified version of the polyatomic programs in the uk molecular r - matrix codes @xcite .
these programs utilize gaussian type orbitals ( gto ) to represent target molecule as well as a scattering electron . although most of past r - matrix works on electron o@xmath0 collisions had employed the diatomic modules using slater type orbitals ( sto ) obtained by hartree fock(hf ) calculation , we select gto mainly because of simplicity of the input and availability of basis functions .
the state averaged complete active space scf ( sa - casscf ) orbitals are imported from the target calculations with molpro suites of programs @xcite .
this employment of sa - casscf orbitals improves the vertical excitation energies of the o@xmath0 target states compared to the energies obtained using hf orbitals .
these target orbitals are constructed from the [ 5s,3p ] contracted basis of dunning @xcite augmented by a d function with exponent 1.8846 , as in sarpal et al .
@xcite . in the r - matrix calculations
, we included 13 target states ; @xmath15,@xmath2 , @xmath50,@xmath10,@xmath11 , @xmath12,@xmath27,@xmath28 , @xmath29 , @xmath30,@xmath31,@xmath32 and @xmath33 , where the last 4 @xmath34 states were not included in previous calculations .
the potential energy curves of these target electronic states are shown in figure [ fig1 ] for reference .
further details of these target electronic states can be found in saxon and liu @xcite and minaev and minaeva @xcite for example . in our fixed - bond r - matrix calculations ,
these target states are evaluated at the equilibrium bond length @xmath51 = 2.3 a@xmath52 of the o@xmath0 @xmath15 ground electronic state .
note that all calculations were performed with d@xmath53 symmetry because of restriction of the polyatomic uk r - matrix codes , though natural symmetry of this system is d@xmath54 . the radius of the r - matrix sphere @xmath41 was chosen to be 10 a@xmath52 in our calculations . in order to represent the scattering electron
, we included diffuse gaussian functions up to @xmath55=5 with 9 functions for @xmath55=0 , 7 functions for @xmath55=1 - 3 and 6 functions for @xmath55=4 and 5 .
the exponents of these gaussians were fitted using the gtobas program @xcite in the uk r - matrix codes .
details of the fitting procedure are the same as in faure et al .
we constructed the @xmath56 electron configurations from the orbitals listed in table [ tab0 ] .
the ci target wave functions are composed from the valence orbitals in table [ tab0 ] with the 1@xmath57 and 1@xmath58 orbitals kept doubly occupied .
the first terms in equation ( 1 ) are constructed from configurations of the form , @xmath59 here we assume that the total symmetry of this 17 electrons system is @xmath60 .
the first 4 electrons are always kept in the 1@xmath57 and 1@xmath58 orbitals , then the next 12 electrons are distributed over the valence orbitals with restriction of target state symmetry , @xmath61 symmetry of the o@xmath0 ground state in this case .
the last electron , the scattering electron , occupies one of the diffuse orbitals , @xmath62 symmetry in this example . to complete the wave function with the total symmetry @xmath60
, we also have to include configurations with the other target states combined with diffuse orbitals having appropriate symmetry in the same way as in the example .
the second terms in equation ( 1 ) are constructed from configurations , @xmath63 where the scattering electron occupies a bound @xmath64 extra virtual orbital , instead of the diffuse continuum orbitals in the expression ( 2 ) . as in table [ tab0 ] , we included one extra virtual orbital for each symmetry .
the second terms in equation ( 1 ) also contain configurations of the form @xmath65 in this case , the last 13 electrons including the scattering electron are distributed over the valence orbitals with the restriction of @xmath60 symmetry . in this way
, the number of configurations generated for a specific total symmetry is typically about 17000 , though the final dimension of the inner region hamiltonian is reduced to be about 500 by using ci target contraction and prototype ci expansion method @xcite . in order to obtain the integral cross sections for the electron o@xmath0 collisions
, the r - matrix calculations were performed over all 8 irreducible representations of d@xmath66 symmetry , @xmath67,@xmath68,@xmath69,@xmath62,@xmath70,@xmath71,@xmath72 and @xmath73 with both doublet and quartet spin multiplicity .
figure [ fig1 ] shows the potential energy curves of the o@xmath0 target states .
these curves were calculated by the sa - casscf method which was used in the actual r - matrix calculations .
although not included in our r - matrix calculations , we also include the curves for the o@xmath74 states for reference .
table [ tab1 ] compares vertical excitation energies from the present calculations with previous hf / sto results .
compared to the experimentally estimated values , our results are of the same quality for the o@xmath0 @xmath2 state and are slightly worse , 0.05 ev , than the hf / sto result for the @xmath50 state .
however , the excitation energies are improved by about 0.1 ev for the ` 6 ev states ' ; @xmath10 , @xmath11 and @xmath12 . for the higher 3 electronic states ,
the improvement is about 1 ev . though discrepancies with the experimental values are still not small , 0.22 ev for the @xmath50 state and 0.45 ev for the @xmath11 state for examples , we believe that our choice of the gto basis set and the cas space is satisfactory for the present r - matrix calculations considering the differences with the previous hf / sto results . in figures
[ fig2]-[fig5 ] , the cross sections for the transitions from the o@xmath0(@xmath15 ) ground state are shown .
these cross sections were previously calculated using the r - matrix method , but with a different basis set and target descriptions @xcite . in figure [ fig2 ] , we compare our elastic scattering cross sections for the @xmath15 state with the previous theoretical and experimental results .
the theoretical elastic scattering cross sections are quite similar in shape and magnitude each other .
there is a sharp peak around 0.5 ev in each theoretical cross section , which comes from the o@xmath26 @xmath75 resonance . in our calculation with 13 target states including 4 extra @xmath34 targets ,
this @xmath75 resonance is located at 0.196 ev and the width is 0.00134 ev . when the number of targets are reduced to 9 by removing the @xmath34 target states ,
the location of the resonance is shifted to 0.548 ev with a width of 0.0161 ev .
the later resonance parameters with 9 target states are closer to the results of the previous calculations , reflecting inclusion of the same number of the target states .
when our results are compared to the experimentally measured elastic cross sections , agreement is good for energy above 10 ev but is poorer at lower scattering energy below 5 ev . at 1 ev , the theoretical cross section is about a factor of 2 larger than the experimental results .
this situation mirrors that in the previous r - matrix calculations . as discussed in noble and burke @xcite ,
this discrepancy is attributed to lack of long - range polarization effects in our and their model . to improve this low energy behaviour of the elastic cross sections
, we may need pseudostates method of gillan et al .
@xcite and gorfinkiel and tennyson @xcite for example .
figures [ fig3 ] and [ fig4 ] show excitation cross sections from the o@xmath0 @xmath15 state to the @xmath2 and @xmath50 states . in both figures
, there is a pronounced peak in the cross sections around 8 ev which comes from the o@xmath26 @xmath76 resonance located at 7.988 ev with its width being 0.906 ev . compared to the previous r - matrix calculations of noble and burke @xcite , our cross sections with 13 target states are slightly smaller at all scattering energies .
the peak height of our results around 8 ev is 30% smaller in excitation to the @xmath2 case and is 35% smaller in the @xmath50 case .
however , general feature of the cross section profiles are quite similar in our results and the previous r - matrix calculations .
we also compare our results with the effective range theory ( ert ) results of teillet - billy et al
their method relied on the existence of the o@xmath26 @xmath75 resonance around 0.2 ev , but did not include the effect of the o@xmath26 @xmath77 resonance located at 8 ev .
thus , their results and our cross sections agree well at low energy , below 6 ev , where the @xmath75 symmetry mainly contributes to the total cross sections
. however , the agreement is worse at energy range above 7 ev because of the @xmath77 resonance contributions . in both o@xmath0(@xmath2 ) and o@xmath0(@xmath50 ) cases , agreement with the experimental cross sections is modestly good in the energy regions away from the resonance peak . as shown in the figures [ fig3 ] and [ fig4 ] , the cross sections of middleton et al .
@xcite have peak in energy region around 10 ev , which is 2 ev larger than the theoretical position . to resolve this discrepancy , we need to include the effect of nuclear motion in calculation as discussed in higgins et al .
@xcite .
figure [ fig5 ] shows excitation cross sections from the o@xmath0 @xmath15 state to the ` 6 ev states ' .
these ` 6 ev states ' consist of the o@xmath0 @xmath10 , @xmath11 and @xmath12 states . in order to compare with previous experimental measurements
, we sum the cross sections for the transitions to these states . as in the case of transitions to the @xmath2 and @xmath50 states in figures [ fig3 ] and [ fig4 ] , a prominent peak exists around 8 ev in our results .
the cause of this peak is the o@xmath26 @xmath77 resonance located at 7.988 ev .
our results are quite similar in shape and magnitude to the previous r - matrix calculation of noble and burke @xcite , though the cross section peak at 8 ev is slightly lower in our case .
the ert results of gauyacq et al .
@xcite are also shown in the figure . as discussed above
, their cross sections do not have a peak around 8 ev because they did not include the o@xmath26 @xmath77 resonance effects .
recently , green et al.@xcite measured the integral cross sections from the @xmath15 state to the ` 6 ev states ' and discussed the discrepancy between the past theoretical results and their measurements .
though the theoretical cross sections have a peak around 8 ev , the experimental results do not show this peak nor the enhancement of the cross sections near 8 ev .
figure [ fig5 ] compares our results with the experimental cross sections , which shows that the discrepancy still exists below 10 ev .
this deviation may come from our use of fixed - bond approximation , because the equilibrium bond distances of ` 6 ev state ' are longer than those of @xmath15 , @xmath2 and @xmath50 states . in principle , we need to employ the non - adiabatic r - matrix method or vibrational averaging procedure to take into account those difference of the equilibrium distances . in this study
, we limit ourselves at the fixed - bond approximation and leave the treatment of nuclear motion for work in future . figure [ fig6 ] shows elastic cross sections for the @xmath2 state as functions of electron collision energy .
these cross sections have almost the same shape and magnitude as the @xmath15 state elastic cross sections .
we do not observe a sharp resonance peak in the @xmath2 elastic cross sections in contrast to the @xmath15 case , because the o@xmath26 @xmath75 resonance is located 0.7 ev below the o@xmath0 @xmath2 state , but 0.2 ev above the o@xmath0 @xmath15 state .
as shown in the figure [ fig6 ] , the @xmath78 symmetry is the main contributor to the cross sections at low energy .
this indicates that the @xmath55=0 component of the scattering electron is as important as for the @xmath15 elastic scattering .
the cross section for excitation to the @xmath50 state from the @xmath2 state is shown in figure [ fig7 ] .
the magnitude of the cross section is about 10 times larger than the corresponding cross section for excitation from the @xmath15 state to the @xmath50 state . at electron collision energy of 7.0 ev
, there is a large peak in the cross sections arising from the @xmath79 symmetry .
the origin of this peak is the o@xmath26 @xmath77 resonance as in the cross sections from the @xmath15 state shown in figures [ fig3]-[fig5 ] . because the cross sections are plotted as functions of electron collision energy , the positions of the peak in figure [ fig7 ] and figure [ fig3]-[fig5 ] are different by 0.93 ev which is the energy difference of the o@xmath0 @xmath15 state and the o@xmath0 @xmath2 state .
hall and trajmar experimentally determined differential and integral cross sections at 4.5 ev for this excitation @xcite . as in figure
[ fig7 ] , their value and our results agree well . however , we need more experimental data to make detailed comparisons .
excitation cross sections to the ` 6 ev states ' from the o@xmath0 @xmath2 state are plotted in figure [ fig8 ] .
there we show the summed cross sections as well as individual contributions of the o@xmath0 @xmath10,@xmath11 and @xmath12 states .
the summed total cross section has a similar shape to the excitation cross section from the @xmath15 state .
the origin of a peak at 7.0 ev is the o@xmath26 @xmath77 resonance , as in the case of the @xmath2 @xmath6 @xmath50 transition .
the magnitude of the cross sections at this resonance peak is about 3 times larger than the corresponding cross sections of the @xmath15 case in figure [ fig5 ] .
the difference is less pronounced in the energy region above 10 ev where the cross sections from the @xmath2 state are about 30% larger than those from the @xmath15 state .
figures [ fig9 ] and [ fig10 ] show the cross sections for electron collisions with the o@xmath0 @xmath50 excited state .
the overall features are quite similar to the corresponding cross sections from the @xmath2 state . in particular ,
the elastic cross sections of the @xmath50 state in figure [ fig9 ] are almost the same as those of the @xmath2 state shown in figure [ fig6 ] .
excitation cross sections to the ` 6 ev states ' in figure [ fig10 ] are slightly different from the excitation cross sections to the ` 6 ev states ' from the @xmath2 state .
the height of the cross section peak at 6.5 ev is about 35% larger than that of the @xmath2 case . as in the cases of the @xmath15 @xmath6
` 6 ev states ' and @xmath2 @xmath6 ` 6 ev states ' excitations , the o@xmath26 @xmath77 resonance causes this peak in the cross section .
the location of the peak in figure [ fig10 ] is shifted from the peak positions in figures [ fig5 ] and [ fig8 ] because of energy differences between the o@xmath0 @xmath15 , @xmath2 and @xmath50 states .
finally , we discuss the effect of the extra @xmath80 target states in our r - matrix calculations . in figures
[ fig2]-[fig5 ] , we compare cross sections of electron collisions with the o@xmath0 @xmath15 state from the 9 target states calculations and those from the 13 target states calculations including 4 extra @xmath80 target states .
inclusion of extra @xmath80 target states generally lowers the cross sections .
however , this lowering is less than 15% and is not significant .
our 9 target states cross sections have similar magnitude in general compared to the 9 target - states r - matrix calculations of noble and burke @xcite . however , they are slightly different for the excitations to the @xmath2 , @xmath50 states and ` 6 ev states ' in the energy region around 8.0 ev . in this o@xmath26 @xmath77 resonance region ,
our cross sections are about 20 - 30% smaller than their results .
these differences in the cross sections may be attributed to the different treatment of the basis set and the ci representations of the target states in ours and their calculations .
we have investigated electron collisions with the excited @xmath2 , @xmath3 states of the o@xmath0 molecule using the fixed - bond r - matrix method which includes 13 target electronic states , @xmath15,@xmath2 , @xmath50,@xmath10,@xmath11 , @xmath12,@xmath27,@xmath28 , @xmath29 , @xmath30,@xmath31,@xmath32 and @xmath33 .
these target states are described by ci wave functions in the valence cas space , using sa - casscf orbitals .
gaussian type orbitals are used in this work , in contrast to the stos in the previous works .
our vertical excitation energies are in good agreement the previous results and the experimental values .
we obtaine integral cross sections for @xmath2 @xmath16 @xmath2,@xmath3 and ` 6ev states'(@xmath10 , @xmath11 and @xmath12 ) , as well as @xmath3 @xmath16 @xmath3 and ` 6ev states ' .
the magnitude of the cross sections for the @xmath2 @xmath16 @xmath3 transition is consistent with the existing experimental value , which is 10 time larger than the one for @xmath15 @xmath16 @xmath3 .
the elastic cross sections for the @xmath2 state and the @xmath3 state have similar magnitude and shape when compared to the elastic cross sections of the @xmath15 state .
the transitions for the @xmath2,@xmath3 @xmath16 ` 6ev states ' have cross sections about 5 times larger than the corresponding transitions from the @xmath15 ground state .
our results will be important for modeling of plasma discharge chemistry which needs cross sections between the excited electronic states in some case .
gorfinkiel for her hospitality during his visit to ucl .
the present research is supported in part by the grant from the air force office of scientific research : the advanced high - energy closed - cycle chemical lasers project ( pi : wayne c. solomon , university of illinois , f49620 - 02 - 1 - 0357 ) .
computer resources were provided in part by the air force office of scientific research durip grant ( fa9550 - 04 - 1 - 0321 ) as well as by the cherry l. emerson center for scientific computation at emory university .
the work of m.t . was partially supported by the japan society for the promotion of science postdoctoral fellowships for research abroad .
the elastic cross sections of the o@xmath0 @xmath15 state .
thick full line represents cross sections obtained by 13 target states calculation including @xmath80 target states .
thick dotted line is the cross sections including 9 target states without @xmath80 target states .
the partial cross sections from the 13 target states calculation are represented by thin full lines marked with open symbols .
symmetries with minor contributions are not shown in the figure . for comparisons
, we also include the previous r - matrix results of noble and burke @xcite , the experimental cross sections of trajmar et al .
@xcite , kanik et al .
@xcite , shyn and sharp @xcite , sullivan et al .
@xcite and linert et al .
] the excitation cross section from the o@xmath0 @xmath15 state to the @xmath2 state .
our results are shown in thick full and dotted lines as in figure [ fig2 ] .
the partial cross sections of @xmath81 symmetries are also shown as thin full lines marked with open symbols . for comparison , we include the previous r - matrix results of noble and burke @xcite , the ert calculations of teillet - billy et al .
@xcite , the experimental cross sections of doering @xcite , shyn and sweeney @xcite and middleton et al .
] the excitation cross section from the o@xmath0 @xmath15 state to the @xmath50 state .
our results are shown in thick full and dotted lines as in figure [ fig2 ] .
we include the previous theoretical results of noble and burke @xcite , teillet - billy et al .
@xcite and the experimental cross sections of shyn and sweeney @xcite and middleton et al . @xcite .
] the excitation cross sections from the o@xmath0 @xmath15 state to the ` 6 ev states ' which consist of the o@xmath0 @xmath10,@xmath11 and @xmath12 states .
the total cross sections shown here are the sum of the individual cross sections of these 3 states .
our results for the total cross sections are shown in thick full and dotted lines as in figure [ fig2 ] .
the individual excitation cross sections from 13 states calculations are shown as thin full lines marked with open symbols .
we include the total cross sections from the previous r - matrix results of noble and burke @xcite , the ert calculations of gauyacq et al .
@xcite , the experimental results of teillet - billy et al . @xcite and green et al .
] the elastic cross sections for the o@xmath0 @xmath2 state .
the thick full line represents the cross sections obtained by 13 target states calculation including @xmath80 target states .
the contributions of the @xmath82 and @xmath83 total symmetries are also shown . ]
excitation cross section from the o@xmath0 @xmath2 state to the @xmath50 state .
our results are shown in thick full line with the contributions of @xmath84 and @xmath85 symmetries in thin lines as in figure [ fig6 ] . the experimental cross section of hall and trajmar @xcite
is also included for comparison . ]
excitation cross sections from the o@xmath0 @xmath2 state to the ` 6 ev states ' of the o@xmath0 @xmath10 , @xmath11 and @xmath12 states .
the cross sections for excitation to the individual state are shown in thin lines .
the thick full line represents the sum of these cross sections .
] elastic cross section for the o@xmath0 @xmath50 state .
the results are shown in thick full line with the contributions of @xmath84 , @xmath85 and @xmath86 symmetries in thin lines as in figure [ fig6 ] . ]
excitation cross sections from the o@xmath0 @xmath50 state to the ` 6 ev states ' of the o@xmath0 @xmath10 , @xmath11 and @xmath12 states .
the cross sections for excitation to the individual state are shown in thin lines .
the thick full line represents the sum of these cross sections . ] | low - energy electron collisions with o@xmath0 molecules are studied using the fixed - bond r - matrix method .
in addition to the o@xmath0 @xmath1 ground state , integrated cross sections are calculated for elecron collisions with the @xmath2 and @xmath3 excited states of o@xmath0 molecules .
13 target electronic states of o@xmath0 are included in the model within a valence configuration interaction representations of the target states .
elastic cross sections for the @xmath2 and @xmath3 excited states are similar to the cross sections for the @xmath1 ground state . as in case of excitation from the @xmath1 state , the o@xmath4 @xmath5 resonance makes the dominant contribution to excitation cross sections from the @xmath2 and @xmath3 states .
the magnitude of excitation cross sections from the @xmath2 state to the @xmath3 state is about 10 time larger than the corresponding cross sections from the @xmath1 to the @xmath3 state . for this @xmath2 @xmath6
@xmath3 transition , our cross section at 4.5 ev agrees well with the available experimental value .
these results should be important for models of plasma discharge chemistry which often requires cross sections between the excited electronic states of o@xmath0 . |
in the last decade the quality of chemical vapor deposited ( cvd ) diamond for particle detectors has greatly improved , and the use of this material has become a potentially attractive option for vertex detectors at high luminosity colliders ( such as the lhc after the proposed luminosity upgrade @xcite ) promising to provide the radiation resistance needed for the challenging particle fluxes expected .
diamond detectors are believed to be more radiation - hard than silicon @xcite . even after radiation exposures in excess of @xmath1
they can be operated at room temperature without significant leakage current .
a loss in the charge signal of about 40% was observed with pion / proton radiation of @xmath2 @xcite . with neutrons a loss of 30%
was observed at a fluence of @xmath3 @xcite .
the charge collection distance is the figure of merit for diamond particle detectors .
it is defined as @xmath4 where @xmath5 is the diamond film thickness , and @xmath6 is the collection efficiency of the ionization charge . the collection distance has been increased from a few micrometers in the beginning of diamond r&d to about 250 @xmath0 m , corresponding to a mean signal of 9000 electrons , today @xcite .
the charge collection distance depends on the presence of charge carrier traps @xcite .
once a deep level trap has captured a charge carrier the trap can be permanently passivated . as a consequence of this process
, cvd diamond shows a significant increase in the charge collection distance after being exposed to fluences of the order of @xmath7 minimum ionizing particles per @xmath8 , a process called pumping or priming .
cvd diamond remains for a long period ( months ) in the primed state if kept in the dark and at room temperature .
cvd diamond usually has a polycrystalline structure with an average crystallite size on the growth side of the order of 1/7 of the thickness @xcite . on the substrate side
the crystal size is only a few micrometers .
the charge collection distance can be different in different crystallites @xcite , leading to a non - uniform response and a broadening of the distribution of the charge signal from the whole substrate .
in this paper we report test - beam measurements and simulations of diamond pixel detectors equipped with a prototype electronics chip of the atlas pixel silicon detector @xcite . in section [ sec2 ] the analysis of test - beam data taken with two diamond pixel detectors is presented . the spatial resolution and detection efficiency of the detectors
are reported , and the observation of a non uniform response of the detectors across the sensor area is discussed . in section [ sec3 ] a numerical simulation of the charge collection processes inside the diamond sensor and of the detector response is presented .
it is shown that the non uniform response can be reproduced as a consequence of the polarization fields created by the charge trapped in the crystallite structure .
test beam experiments were performed at the cern sps accelerator during the years 2000 and 2001 with a pion beam of 180 gev/_c _ momentum .
a beam telescope consisting of 4 silicon microstrip modules was used to measure the transverse position of the incident beam particles . in the setup used in the year 2000 @xcite each module consisted of a pair of microstrip detectors , each providing one coordinate . in the setup used in 2001 each module consisted of a double - sided microstrip detector .
the new telescope used a readout architecture including zero suppression @xcite , which offered higher trigger rate capability .
for both set - ups , the position resolution of tracks projected onto the tested devices was about 6 @xmath0 m .
two diamond sensors , identified by uts-5 and cd91 , were tested in the beam @xcite .
the sensors were grown to a thickness of about @xmath9 m .
the grain size on the growth side is of the order of 100 @xmath0 m .
the sensors were then lapped on the substrate side by about @xmath10 m and on the growth side by about 50 @xmath0 m so that the final thickness is 432 @xmath0 m for uts-5 and 470 @xmath0 m for cd91 .
the lateral dimensions of the sensors were @xmath11 .
the electrode on the growth side was segmented in pixels of @xmath12 dimensions .
each pixel was electrically connected via electroplated pbsn solder bumps to a matching readout electronics cell in the front - end chip . in the following ,
@xmath13 is the coordinate along the short ( 50 @xmath0 m ) dimension of the pixels , @xmath14 is the coordinate along the long ( 400 @xmath0 m ) dimension of the pixels , and @xmath15 is the coordinate perpendicular to the pixel plane . before operation at the test beam , data were taken with a @xmath16am and a @xmath17sr source , with activities of 74 mbq and 62 kbq respectively .
the sources were kept for 12 h at about 5 mm from the diamond sensor .
the @xmath18 source delivered a fluence of the order of @xmath19 ( 0.3 gy dose ) to each sensor , which is enough to bring good quality diamond in the primed state @xcite .
this conclusions is supported by the test beam data , discussed below , since the average signal and the detection efficiency did not show any time dependence during the data taking .
the electronics chips were produced during the development phase of the atlas pixel front - end electronics program @xcite .
their design was similar to that of the final front - end electronics for atlas pixel @xcite .
in each front - end chip , 2880 channels are arranged into 18 columns by 160 rows .
the charge - sensitive preamplifiers feature a dc feedback scheme with a tunable current providing control over the shaping - time for a given input charge .
a discrimination stage sits behind the preamplifier in each channel which is sensitive to the leading edges and the trailing edges of pulses .
each channel is equipped with its own 3-bit dac for channel - to - channel threshold adjustments , thus a means of overall dispersion reduction is provided .
the chips have a 7-bit charge measurement capability using the time - over - threshold ( tot ) of the signal .
the tot is calibrated by injecting a known charge into every channel .
a 2880-bit pixel register plus one corresponding latch per channel enable individual pixels to be masked - off independently for calibration - strobing and readout .
the chip is operated at the 40 mhz lhc bunch crossing rate and the times are measured in multiples of the 25 ns clock period .
only pixel signals whose leading edge belongs to the period specified by an external trigger are read out . at the test - beam the trigger
was provided by two scintillator detectors and 16 consecutive 25 ns time windows were accepted . during the operation in the test beam the thresholds of the individual channels
were adjusted achieving a threshold dispersion of less than 100 e. a typical threshold setting was 1000 e while the average noise per pixel was about 200 e.
events were filtered with the requirement of one and only one track reconstructed by the silicon microstrip telescope in each event .
only events with a track reconstruction @xmath20 probability greater than 0.02 were kept .
tracks were required to extrapolate into a fiducial area . for cd91
this was the surface covered by the pixel array , excluding the region within 40 @xmath0 m of the border .
for uts-5 it was the smaller area with a good bump - bonding yield @xcite .
pixel clusters were built clustering together all adjacent pixels , independently of track extrapolations .
the cluster charge is defined as the sum of the charges measured by the pixels of the cluster .
this can be less than the total collected charge , since it does not include the pixel signals below the electronics threshold .
the cluster position is computed as the arithmetic mean of the coordinates of the pixels in the cluster .
the data discussed here have been taken at normal incidence .
the sensors were operated with a bias voltage of 450 v ( uts-5 ) or 470 v ( cd91 ) . at normal incidence ,
most clusters are composed of one or two pixels .
the average cluster size was measured to be @xmath21 pixels for cd91 and @xmath22 for uts-5 .
charge calibrations were available only for uts-5 , for which the cluster charge ( the sum of the charges measured by the pixels of the cluster ) was @xmath23 electrons .
the detection efficiency was @xmath24% for cd91 and @xmath25% for uts-5 .
[ residuals ] shows the distribution of the @xmath13 residuals between the track position as determined by the telescope and the pixel cluster position .
for comparison , the same distribution is reported for an atlas pixel silicon sensor , with a similar pixel geometry , bump - bonded to a similar electronics chip and tested with the same test - beam setup . the r.m.s .
of spatial residual distributions yields resolutions of @xmath26 m for the silicon detector , @xmath27 m for uts-5 and @xmath28 m for cd91 .
the usage of charge interpolation algorithms for the position of multi - pixel clusters improves these values only by a few tenths of a micrometer .
the values of mean cluster size , detection efficiency and spatial resolution are summarized in table [ datatab ] .
the reasons for the poor diamond resolution were investigated by looking at the mean @xmath13 residuals and @xmath14 coordinates of the track extrapolation . the pixel cluster position , in contrast , is determined with far better precision in the @xmath13 direction , because of the 8:1 aspect ratio of atlas pixels .
for this reason , the @xmath14 residuals have not been taken into consideration . ] as a function of the track position on the sensor .
the sensor was divided in bins of @xmath29 size , and for each bin @xmath30 the mean spatial @xmath13 residual @xmath31 has been computed .
the same analysis was also performed on a atlas pixel silicon detector with similar sensor geometry , electronics and test beam setup .
the results are reported in fig .
[ residualsmap ] .
the silicon sensor ( upper plot ) shows a more uniform distribution .
bins with large residuals are isolated and scattered across the sensor as one expects from statistical fluctuations . the diamond sensors ( middle and lower plot ) present regions which have systematically positive or negative residuals as large as 40 @xmath0 m .
no correlation is observed between these regions and the segmentation in pixels .
the scale of the residual clustering is of the order of 100 @xmath0 m which is also the typical size of the diamond crystallites .
this suggests that in some crystallites the drift of the charge carriers has a component parallel to the sensor surface , so that the average position of the pixel clusters is shifted away from the track position .
the statistical error @xmath32 on each mean residual @xmath31 is @xmath33 where @xmath34 is the spatial resolution and @xmath35 is the number of events inside the position bin .
the weighted r.m.s of mean residuals is @xmath36 .
this is @xmath37 m for uts-5 , @xmath38 m for cd91 and @xmath39 m for the silicon sensor .
the weighed distribution of the average residuals is shown in fig .
[ meanres ] . in absence of systematic effects on the spatial response of the detector one
expects the r.m.s .
of mean residuals r to be @xmath40 where @xmath41 is the average number of entries for each position bin . for silicon
it is @xmath42 m which is indeed close to the measured value . for diamond an important systematic contribution exists , since the measured values of @xmath43 are much larger than the values expected from statistic fluctuations , @xmath44 m for uts5 and @xmath45 m for cd91 .
to quantify the apparent clustering of residual shifts , the linear correlation coefficient between the residuals of all track pairs has been determined .
since each event which passes selection cuts has only one reconstructed track , the pairs are formed considering the tracks of two different events .
correlation coefficients are determined in bins of the distance between the two tracks of a pair .
an empirical fit function @xmath46 has been found to describe the data of the correlation coefficient of residuals as a function of track separation @xmath13 .
the function is the sum of two components .
a sinusoidal modulation with the periodicity of the pixel pitch describes the correlation due to the pattern structure of the sensor . a falling exponential accounts for the residual shifts introduced by the grain structure .
the denominator in the exponent @xmath47 we term the correlation length .
it is a measure of the scale of the cluster structure of similar residual shifts .
the residuals correlation is reported in fig .
[ autocorrelation ] . for silicon
the correlation is well described by the pitch modulation term only and the amplitude of the exponential term is consistent with zero . for diamond
the exponential peak is evident .
the fit gives a correlation length @xmath48 m for uts5 and @xmath49 m for cd91 .
the correlation amplitude @xmath50 is @xmath51 for uts5 and @xmath52 for cd91 .
the presence of the exponential peak in the diamond correlation distribution is evidence of non - homogeneous charge collection properties producing local systematic shifts in the position response .
the investigation of their origin is the subject of the studies described in the next chapter .
in order to understand the experimental results in terms of the microscopic properties of charge carriers drift in diamond , a numerical simulation of the response of diamond sensors to ionizing radiation has been developed .
the interactions of high - energy pions with the diamond sensor have been simulated using the geant4 package @xcite .
this step of the simulation produces a file with a list of the energy deposits of the particles inside diamond ( hits ) .
the energy of each hit is converted into a number of electron - hole pairs using the conversion factor of one pair for 13.1 ev of deposited energy .
the charge carriers are then drifted in the diamond in steps of @xmath53 m until they are trapped or they are collected at the electrodes .
* drift model : * the drift direction is determined by the local electric field and thermal diffusion .
the local electric field is the superposition of the external field of 1040 v / mm and the polarization field created by the charge carriers trapped in the sensor .
the diffusion function is a gaussian with @xmath54 , where @xmath55 is the diffusivity and @xmath56 the drift time .
the diffusivity @xmath55 is related to the low - field mobility @xmath57 and the temperature @xmath58 by the einstein relation @xmath59 , where @xmath60 is the boltzmann constant and @xmath61 the elementary charge .
the drift time corresponding to one drift step is @xmath62 where the drift velocity @xmath63 depends on the electric field .
it is proportional to the electric field for low field intensity , and reaches saturation in the high - field limit .
the following parametrization was used to connect the two asymptotic behaviors @xcite @xmath64}\ ] ] where @xmath57 is the low - field drift mobility and @xmath65 is the high - field saturation velocity .
the values of these parameters in diamond are controversial @xcite .
the following values have been used @xcite : @xmath66 , @xmath67 cm / s for electrons and @xmath68 , @xmath69 cm / s for holes .
* diamond grains generator : * the trapping model presented below requires the simulation of the grain structure of cvd diamond . the grains growth generator is a generalization to three dimensions of the generator described in @xcite .
the space is divided in small cubic cells ( @xmath70 ) and the grains are initially 2 x 2 cells wide . every time that a layer of cells is added to the diamond film , each cell near the grain border has a probability @xmath71 to be claimed by a neighboring grain .
the resulting conflicts between grains claiming the same space element are solved assigning to each grain a probability to defeat the other proportional to the grain width .
this results in large grains getting larger and small grains getting smaller and finally being overgrown .
the resulting simulated structure is shown in fig .
[ grainmap ] .
the only parameters of the model are the growth probability @xmath71 and the film thickness , which is the sensor thickness plus the material removed from the substrate side .
the film thickness ( 732 @xmath0 m ) is equal to the final sensor thickness plus the thickness of material removed from the substrate side .
the growth probability is chosen so that the resulting spatial residuals correlation length ( see sec .
[ results ] ) agrees with the data . *
trapping model : * the probability for a charge carrier to be trapped within the drift step @xmath72 is @xmath73 where @xmath63 is the velocity and @xmath74 the lifetime .
the lifetime is inversely proportional to the local density of active traps .
different initial trap density distributions have been implemented . in the simulation presented in this paper ,
the initial trap density depends on the shortest distance to the next grain boundary and is described by @xcite @xmath75\ ] ] with @xmath76 the distance to the next grain boundary , @xmath77 the minimum trap density ( deep in the grains bulk ) and @xmath78 the _ lifetime length_. the trap density becomes very high near the grain boundaries .
this model is supported by observations that the impurity concentration is strongly enhanced at grain boundaries @xcite thus resulting in a larger density of electrically active traps @xcite .
priming is simulated by implementing two classes of traps @xcite .
after they capture a charge , the traps of one class can act as recombination centers , while the traps of the other class are permanently passivated .
the electron and hole lifetimes are computed as @xmath79 @xmath80 @xmath81 is the density of unfilled traps of both classes , @xmath82 is the density of recombination centers filled with holes ( electrons ) .
@xmath18 can be interpreted as the product of the charge carriers thermal velocity and an effective trapping cross section .
it is the same for traps of the two classes but different values can be set for electrons and holes . when a charged carrier is trapped , the local density of filled and unfilled traps and the local trapped charge density are updated .
the polarization field created by the trapped charges is periodically updated and superimposed to the external field to get the electric field map .
as the diamond sensor is exposed to ionizing radiation , three dose - dependent effects occur : * the pure trapping centers are passivated , so that the effective trap density is reduced , simulating the priming effect . * near the electrodes the charge carrier which is drifting toward the electrode is dominant ( electrons near the pixels electrodes , holes on the other side ) .
thus a net space charge density develops .
the resulting polarization screens the external field . *
a net space charge develops also near the grain boundaries ( only in the scenario with non - uniform trap density ) . because of grain lateral growth the boundary surface makes in general a non - zero angle with the @xmath15 direction .
electrons and holes approach this surface from opposite directions , and are trapped before reaching the interface because of the very high trapping density .
hence a space charge distribution is created , with opposite signs on the two sides of the interface .
the electric field they produce is normal to the interface , and has a component which is transverse to the external field .
this is the source of the charge drift parallel to the sensor surface . the distribution of trapped charge for a two - dimensional simulation is shown in fig .
[ polarization ] .
a net negative trapped charge develops near the pixels at the bottom and a net positive charge near the other electrode .
the tree - like structures are the boundaries between grains ( growing from above ) .
there is a larger trapped charge there , negative above the boundary and positive below it .
the parameters of the trapping model discussed are @xmath83 and the fraction @xmath84 of pure trapping centers as explained above .
a constraint on the above parameters is provided by the measured value of the average collected charge ( 3000 @xmath85 ) .
this still leaves four degrees of freedom in the parameter choice .
since the other parameters are related to unmeasured properties ( such as the priming ratio , or the hole / electron lifetime ratio ) it is not possible to determine them from the data .
some educated guesses have been made instead , with the purpose of building a model which can qualitatively reproduce the features observed in the data , in particular the systematic position shifts .
the choice of final values for the parameters is reported below . *
the ratio between hole and electron trapping @xmath86 was chosen to be 1.5 since measurements with @xmath84 particles show that holes have a lower lifetime than electrons in diamond sensors @xcite . *
the fraction @xmath84 of traps which are permanently passivated upon capture of a charge carrier affects primarily the priming ratio ( the ratio between the signals observed in the primed and unprimed state ) .
the value @xmath87 was used .
the resulting priming ratio is 1.6 ( it is also affected by @xmath77 , discussed below ) .
the priming ratio of the sensor studied at the test beam is not known , but the simulated value is typical for sensors with good charge collection efficiency @xcite . the simulated priming curve is shown in fig . [ priming ] . *
the extent of the region near the crystallite boundaries where trapping is enhanced is determined by @xmath78 .
the value @xmath88 m was chosen .
this is smaller than the crystallite size so that in the crystallite bulk the charge lifetime is indeed larger than near the border .
a lower value does not change qualitatively the results presented in the next section , but would make them more sensitive to the cell length @xmath72 of the grid of points on which the trap densities are computed . * the total trap concentration @xmath77 at the center of crystallites . the value @xmath89 has been used . a smaller value results in a weaker polarization field and thus in a smaller lateral component of the charge drift , in contrast with the experimental evidence . a larger trap density results in too effective external field screening , making it difficult to reproduce a typical priming curve ( the signal reaches a maximum then decreases with dose ) .
the sensitivity of the results of the simulation on each parameter has been computed varying one parameter at a time and retuning @xmath18 ( keeping constant @xmath90 ) so that the mean signal was always equal the experimental value .
the resulting variations of efficiency , resolution and systematic shift r.m.s .
@xmath43 are reported in table [ sens ] .
* signal formation : * at the end of the charge drift algorithm , a list of the positions where the electrons and holes have been trapped or collected is produced .
the charge @xmath91 induced on a given pixel electrode is computed as @xmath92\ ] ] where @xmath93 are the drifting electric charges , @xmath94 is the position at which they were trapped or collected , @xmath95 their initial positions , and @xmath96 is the weighting potential @xcite for the pixel electrode under consideration . the experimental electronics threshold and noise ( 1000 @xmath85 and 200 @xmath85 r.m.s .
) are simulated .
the pixel hits are processed with the same algorithms used for the analysis of real data . to simulate the telescope extrapolation uncertainty of 6 @xmath0 m a gaussian smearing
is added to the true track position .
due to the severe cpu and memory requirements of a 3d simulation , only a @xmath97 volume was simulated .
trap density maps and charge drift were made with a discretization of space in @xmath98 m steps while the computation of the polarization field was made with a grid with @xmath99 m pitch .
the diamond samples were in a primed state when data were collected .
this was simulated with a sufficiently large number of traversing charged particles to reach a primed state before events for analysis were processed .
the simulated mean spatial residuals as a function of track position are shown in the lower plot of fig .
[ simresmaps ] . for comparison ,
the upper plot reports the results obtained with a uniform trap density ( i.e. , without simulating the crystallites structure ) .
the simulation with trapping enhanced near the crystallites boundaries shows systematic shifts as large as 30 @xmath0 m in the position of the collected charge , as observed with the data ( fig .
[ residualsmap ] ) .
the agreement with experimental results is much better than using the simulation with uniform trapping .
the same conclusion holds for the residual correlation plot ( fig .
[ simcorr ] ) . in the simulation with uniform trapping ( square points )
only the readout pitch correlation is observed , and a good fit is obtained omitting the exponential term of eq .
[ corrfit ] , as with the experimental silicon data .
for the simulation with non - uniform trapping , however , the correlation at small distance scales seen in the diamond data ( fig.[autocorrelation ] ) is reproduced .
the simulated residual correlation plot has been fitted with eq .
[ corrfit ] .
the correlation length is determined by the average crystallite size , determined by the growth parameter @xmath71 , and by the trapping parameters , especially the spatial scale for trapping non - uniformities @xmath78 .
the correlation length was tuned to the experimentally measured value by varying @xmath71 while keeping the trapping parameters constant , except for an overall lifetime scaling to get an average signal equal to the measured value .
as @xmath71 is increased the average crystal size a weight proportional to the probability of a track to cross it .
thus , @xmath100 .
] at the growth surface increases from 0.0119 @xmath8 to 0.0422 @xmath8 and the correlation length increases from @xmath101 m to @xmath102 m .
the final choice of @xmath103 , corresponding to an average grain size @xmath104 , gives a correlation length of @xmath105 m , in agreement with the observed values . in table
[ simtab ] the values of the spatial resolution , the r.m.s . of systematic shifts r
( computed as in section [ analysis ] ) , the detection efficiency and the residuals correlation length are reported for the two simulations and test - beam data .
the statistic error is quoted for simulated values .
the systematic shifts are absent in the simulation without the crystallite structure since they are produced by the lateral component of the polarization field created by charges trapped near the crystallite boundaries .
the simulation of the polycrystalline structure is also needed to reproduce the experimental observed detection efficiency of about 70% . when it is omitted
the efficiency ( respecting the experimental constraints on the average cluster charge and discriminator threshold ) is almost 100% .
the simulation of the crystallites results in a non - uniform charge collection and regions with reduced detection efficiency .
this reproduces the experimentally observed detection efficiency to reasonable accuracy . while the simulation with enhanced trapping near the grain borders is able to qualitatively reproduce the lateral displacement of charge collection and the residuals correlation at small distance scales observed in the data , these polarization effects are less pronounced in the simulation . as a result , the amplitude and the r.m.s . of systematic shifts @xmath43 are smaller and
the spatial resolution is better in the simulation than in the data .
it is possible to increase the strength of polarization effects with a different choice of the model parameters .
in particular , a larger value for @xmath77 increases the density of traps and thus the polarization charges .
however , this also makes the screening of the external electric field more effective , and it is no longer possible to obtain a priming curve with a flat plateau ( fig . [ priming ] ) , instead the signal reaches a maximum and then decreases , as the field screening prevails over the passivation effect .
we have restricted ourselves to choices of the model parameters which give a priming curve typical for diamond sensors exposed to minimum ionizing particles .
most likely , to achieve a quantitative agreement with the data requires a more accurate trapping model . in the model described in section [ model ] only two classes of traps have been implemented , pure trapping centres with cross sections @xmath106 and @xmath107 for electrons and holes respectively and trapping / recombination centres with the same cross sections @xmath106 and @xmath107 both for trapping and recombination processes . in general , several classes of trapping centres
should be anticipated , each one with four different cross sections for electron trapping , hole trapping , electron recombination and hole recombination , respectively . in order not to spoil the predictive power of the model by a too large number of free parameters we refrained from introducing these parameters . a detailed microscopic characterization of the number , spatial distribution and trapping cross sections of defect centres in the diamond sensor would be required to better assess the role of these trapping centres and include them in the simulation .
that is , however , beyond the scope of the study presented in this paper .
test - beam data were taken with diamond pixel detectors . a detection efficiency of about 70% and a spatial resolution of about 20 @xmath0 m were obtained .
the fine segmentation provided by the pixels and the tracking of the particles by a silicon microstrip telescope allowed a study of the spatial resolution of the detector as a function of the incident position of the particles .
the spatial resolution was found to be degraded by the presence of regions where the reconstructed impact position given by the pixel cluster and the true track position measured with a beam telescope deviate systematically .
these regions have the same dimension of the diamond crystallites ( about 100 @xmath0 m ) hence they appear to be related to the polycrystalline nature of the sensor .
a model was proposed which attributes the charge collection position shifts to the polarization field created by charges trapped near the crystallite borders , which can have a component parallel to the sensor surface .
a simulation was presented which implements a detailed description of the ionizing particle interactions , the drift and the trapping of charge carriers in diamond , polarization effects and signal induction on the electrodes . a model for cvd diamond growth
was deployed to simulate the crystallite shape in three dimensions , together with a model for the trap density distribution in crystallites .
the simulation is able to reproduce at least qualitatively the lateral displacement of charge collection and the residuals correlation at small distance scales observed in the data , albeit the effect is weaker in the simulation .
detection efficiency for the same amount of collected charge is also reproduced . with a uniform trap density
neither the charge collection displacements and residuals correlation nor the detection efficiency can be reproduced .
the diamond sensors of the pixel detectors used for the test beam studies were provided by the rd42 collaboration .
the front - end electronics chips were provided by the atlas pixel collaboration , which we also acknowledge for use of the test beam setup and beam time .
the decoding and reconstruction of test beam data was made with a program developed by a. andreazza , c. meroni , f. ragusa and c. troncon ( infn and university of milan ) .
f. gianotti et al . , hep - ph/0204087 ( 2002 ) .
+ d. meier et al .
( rd42 collaboration ) , nucl .
instr . and meth .
a426 ( 1999 ) 173 .
+ w. adam et al .
( rd42 collaboration ) , nucl .
instr . and meth .
a447 ( 2000 ) 244
. + g. lindstrm et al .
( rd48 collaboration ) , nucl .
instr . and meth .
a466 ( 2001 ) 308 .
+ w. adam et al . , ( rd42 collaboration ) , lhcc - status - report - rd-42 , cern - lhcc-2002 - 010 , cern ( 2002 ) .
+ s. zhao , characterization of the electrical properties of polycrystalline diamond films , ph.d .
thesis , ohio state university , 1994 .
+ a. oh , particle detection with cvd diamond , ph.d .
thesis , desy - thesis-1999 - 022 .
+ m. ackers et al .
, ieee trans . on nucl .
46 ( 1999 ) 2033 .
+ l. blanquart et al .
instr . and meth .
a439 ( 2000 ) 403 .
+ p. fischer , nucl .
instr . and meth .
a465 ( 2001 ) 153 .
+ t. lari , study of silicon pixel sensors for the atlas detector , cern - thesis-2001 - 028 ( 2001 ) .
+ j. treis et al . , nucl .
instr . and meth .
a490 ( 2002 ) 112 .
+ m. keil et al . , nucl .
instr . and meth .
a501 ( 2003 ) 153 .
+ l. blanquart et al .
, ieee trans . on nucl .
49 ( 2002 ) 1778 .
+ l. blanquart et al .
, fe - i2 : a front - end readout chip designed in a commercial 0.25 @xmath0 m process for the atlas pixel detector at the lhc , proceedings of ieee2003 , oct 2003 , portland , usa .
+ i. gorelov et al . , nucl .
instr . and meth .
a481 ( 2002 ) 204 .
+ the software provided by the geant4 collaboration is documented in http:///wwwinfo.cern.ch / asd / geant4/geant4.html + c. jacoboni et al . ,
solid state electronics 20 ( 1977 ) 77 .
+ l. s. pan et al .
, j. appl .
74 ( 1993 ) 1086 .
+ c. canali et al .
instr . and
meth . 160 ( 1979 ) 73 .
+ m. artuso et al .
, ieee trans . on nucl .
26 ( 1979 ) 308 .
+ p. polesello et al . , phys .
. sol . 154
( 1996 ) 327 .
+ t. behnke et al . , diam . and rel . mat . 7 ( 1998 ) 1553 .
+ p. polesello et al . , nucl .
instr . and meth .
a426 ( 1999 ) 156 .
+ t. behnke et al .
instr . and meth .
a414 ( 1998 ) 340 .
+ s. ramo , proc .
i.r.e . 27 ( 1939 ) 584 .
+ v. radeka et al . ,
38 ( 1988 ) 217 .
+ .[datatab ] summary of the measurements performed on the two diamond detectors .
the values for a silicon detector with the same geometry and electronics is presented for comparison .
[ cols="^,^,^,^",options="header " , ] mean spatial residual between the position of the center of the pixel clusters and the track position determined by the tracking telescope , as a function of position inside the diamond sensor .
the upper plot is for a silicon detector , the middle one for the diamond sensor uts-5 and the lower plot is for the diamond sensor cd91.,title="fig:",width=377,height=226 ] mean spatial residual between the position of the center of the pixel clusters and the track position determined by the tracking telescope , as a function of position inside the diamond sensor .
the upper plot is for a silicon detector , the middle one for the diamond sensor uts-5 and the lower plot is for the diamond sensor cd91.,title="fig:",width=377,height=226 ] mean spatial residual between the position of the center of the pixel clusters and the track position determined by the tracking telescope , as a function of position inside the diamond sensor .
the upper plot is for a silicon detector , the middle one for the diamond sensor uts-5 and the lower plot is for the diamond sensor cd91.,title="fig:",width=377,height=226 ] distribution of the mean spatial residuals shown in fig .
[ residualsmap ] .
each entry ( mean spatial residual at a given position ) is weighed with the number of tracks used in the computation of the mean residual .
the upper plot is for the silicon sensor , the middle one for the diamond sensor uts-5 and the lower plot is for the diamond sensor cd91.,title="fig:",width=377,height=226 ] distribution of the mean spatial residuals shown in fig .
[ residualsmap ] .
each entry ( mean spatial residual at a given position ) is weighed with the number of tracks used in the computation of the mean residual .
the upper plot is for the silicon sensor , the middle one for the diamond sensor uts-5 and the lower plot is for the diamond sensor cd91.,title="fig:",width=377,height=226 ] distribution of the mean spatial residuals shown in fig .
[ residualsmap ] .
each entry ( mean spatial residual at a given position ) is weighed with the number of tracks used in the computation of the mean residual .
the upper plot is for the silicon sensor , the middle one for the diamond sensor uts-5 and the lower plot is for the diamond sensor cd91.,title="fig:",width=377,height=226 ] three - dimensional simulation of the growth of diamond crystals .
the grey scale is proportional to the carrier lifetime , so that the crystallite boundaries are visible as dark region , and the crystallites bulk is white .
the upper plot shows the crystallite shapes in a plane perpendicular to the growth direction , on the growth side of the diamond film ( where the pixel implantations are made ) .
the lower plot shows the same in a lateral view ; the average dimension of crystallites increases from the substrate side of the film ( below ) to the growth side ( above ) .
a thickness of 0.3 mm on the substrate side has been removed after film growth and it is not used in the simulation either.,title="fig:",width=377,height=302 ] three - dimensional simulation of the growth of diamond crystals .
the grey scale is proportional to the carrier lifetime , so that the crystallite boundaries are visible as dark region , and the crystallites bulk is white .
the upper plot shows the crystallite shapes in a plane perpendicular to the growth direction , on the growth side of the diamond film ( where the pixel implantations are made ) .
the lower plot shows the same in a lateral view ; the average dimension of crystallites increases from the substrate side of the film ( below ) to the growth side ( above ) .
a thickness of 0.3 mm on the substrate side has been removed after film growth and it is not used in the simulation either.,title="fig:",width=377,height=302 ] trapped charge as a function of position inside the sensor , for a two - dimensional simulation with preferential trapping near the grain boundaries . the crystal growth direction and the external electric field
are along the @xmath15 direction ; the pixel implants are located at @xmath108 .
the electrons move downward .
the grain boundaries are visible as tree - like structures .
the pixel electrodes are at the bottom.,width=377 ] | a pixel detector with a cvd diamond sensor has been studied in a 180 gev / c pion beam .
the charge collection properties of the diamond sensor were studied as a function of the track position , which was measured with a silicon microstrip telescope .
non - uniformities were observed on a length scale comparable to the diamond crystallites size . in some regions of the sensor , the charge drift appears to have a component parallel to the sensor surface ( i.e. , normal to the applied electric field ) resulting in systematic residuals between the track position and the hits position as large as 40 @xmath0 m .
a numerical simulation of the charge drift in polycrystalline diamond was developed to compute the signal induced on the electrodes by the electrons and holes released by the passing particles .
the simulation takes into account the crystallite structure , non - uniform trapping across the sensor , diffusion and polarization effects .
it is in qualitative agreement with the data .
additional lateral electric field components result from the non - uniform trapping of charges in the bulk .
these provide a good explanation for the large residuals observed . , , , , , diamond pixel detector , non - uniform charge collection 29.40.wk , 29.40.gx |
high - harmonic generation ( hhg ) is a highly nonlinear optical phenomenon @xcite of increasing interest because it can provide coherent xuv and soft x - ray radiation with attosecond ( @xmath1 s ) durations .
this property offers the opportunity to investigate unexplored research areas in atoms and molecules with unprecedented time resolution @xcite .
the hhg optical spectrum has a distinctive shape : a rapid decrease of the intensity for the low - order harmonics consistent with perturbation theory , followed by a broad plateau region where the harmonic intensity remains almost constant , and then an abrupt cutoff , beyond which almost no harmonics are observed .
the hhg process can be understood by means of semi - classical pictures , such as the celebrated three - step model @xcite : ( i ) an electron escapes from the nuclei through tunnel ionization associated with the strong laser field , ( ii ) it is accelerated away by the laser field until the sign of the field changes , ( iii ) whereupon the electron is reaccelerated back to the nucleus , where it may emit a photon as it recombines to the ground state .
a key quantity emerging from the model is the maximum energy the field can provide to the electron , @xmath2 , where @xmath3 is the ionization potential and @xmath4 is the ponderomotive energy @xcite .
hhg has been studied for many years with theoretical methods solving the time - dependent schrdinger equation using a real - space representation of the wave function @xcite .
these grid - based methods are taken as the numerical reference for this kind of calculations .
indeed , these approaches have proven to be accurate enough to explain key features of atomic and molecular hhg spectra .
however , grid calculations imply memory and cpu requirements that rapidly become prohibitive with increasing numbers of electrons .
because of this limitation , multielectron systems are handled in practice via the use of effective potentials keeping a single - active electron . by contrast
, quantum - chemistry methods such as time - dependent configuration interaction ( tdci ) @xcite , multiconfiguration time - dependent hartree - fock @xcite , or time - dependent density - functional theory @xcite using local basis functions can more easily handle multielectron systems such as molecules , including the treatment of electron correlation .
the main problem of these methods lies in the difficulty to accurately represent the continuum part of the system eigenstate spectrum . addressing this issue can be done on one - electron systems , such as the h atom , since only one electron is promoted into the continuum during the hhg process . in this context , the tdci method with a gaussian - type orbital ( gto ) basis set and a heuristic lifetime model @xcite was recently applied to the calculation of the dipole form of the hhg spectrum for the h atom @xcite .
the role of the rydberg and the continuum states was discussed in detail , and reasonable hhg spectra ( plateau / cutoff ) have been obtained , when compared with the prediction from the three - step model @xcite and grid - based calculations @xcite .
however , the background region , beyond the harmonic cutoff , was higher than expected and spurious harmonics were present .
a possible reason of this behavior is that the basis sets adopted in ref .
describe rydberg states better than the continuum ones .
indeed , while gto basis sets have been successfully applied for calculations of bound - state electronic properties ( even for non - linear optical properties such as second - order hyperpolarizabilities , see e.g. ref . ) , the inherent local nature of gto functions makes it difficult to properly describe continuum states extending over large distances ( see , e.g. , ref . ) . in refs . and
, standard gto basis sets have been augmented with a large number of diffuse basis functions and/or basis functions centered away from the nucleus in order to cover the large spatial extension of the time - dependent wave function . however , this strategy has the serious drawback of only increasing the number of rydberg states while the number of continuum states is not substantially changed .
this results in an unbalanced description of the rydberg and continuum states .
few attempts have been reported in the literature to further improve gto basis sets for a better description of the continuum states .
et al . _
@xcite proposed to fit gto basis functions to slater - type orbital basis functions having a single fixed exponent @xmath5 , supposed to be adequate for scattering calculations .
nestmann and peyerimhoff @xcite proposed to fit a linear combination of gto basis functions to a set of spherical bessel functions , which are the spherically - adapted continuum eigenfunctions for zero potential .
et al . _
@xcite extended the work of nestmann and peyerimhoff to the possibility of fitting a linear combination of gto basis functions to a set of coulomb continuum functions ( i.e. , the continuum eigenfunctions obtained in presence of the coulomb potential @xmath6 , with @xmath7 the nuclear charge ) .
finally , some hybrid methods have also been proposed , combining gaussian functions with finite - element / discrete - variable representation techniques @xcite or with b - spline basis sets @xcite .
note that an alternative approach to gaussian basis sets is given by the use of sturmian functions @xcite .
+ in this article , we study the merits of the gaussian continuum basis functions proposed by kaufmann _ et al . _
@xcite for calculating the hhg spectra in atomic hydrogen within the tdci framework .
while the present results are focused on hhg , our work is relevant for the calculation of any property involving electronic transitions toward the continuum such as , e.g. , photoionization cross sections @xcite or above - threshold ionization rates @xcite .
the paper is organized as follows .
we first describe the theory and give computational details .
we then present and discuss our results .
in particular , we show velocity hhg spectra extracted from the dipole , velocity , and acceleration power spectra calculated for different laser intensities , and basis sets .
we study in detail the effect of increasing the basis set cardinal number , the number of diffuse basis functions , and the number of gaussian continuum basis functions .
we directly compare our results with data from grid calculations , for three values of the laser intensity and two values of the laser wavelength , and adjust the heuristic lifetime model .
finally , we conclude with final comments and perspectives . unless otherwise noted , hartree atomic units , i.e. @xmath8 , are used throughout the paper .
the time - dependent schrdinger equation for the h atom in an external time - dependent uniform electric field @xmath9 in the length gauge is @xmath10 where @xmath11 is the time - independent field - free hamiltonian and @xmath12 is the interaction potential between the atom and the field in the semiclassical dipole approximation .
we consider the case of an electric field @xmath9 linearly polarized along the @xmath13-axis , representing a laser pulse , @xmath14 where @xmath15 is the maximum field strength , @xmath16 is the unit vector along the @xmath13 axis , @xmath17 is the carrier frequency , @xmath18 is the carrier - envelope phase , and @xmath19 is the envelope function chosen as @xmath20 where @xmath21 is the full width at half maximum of the field envelope .
the target quantity to be computed is the power spectrum @xmath22 defined as @xmath23 where @xmath24 and @xmath25 are the initial and final propagation times . in eq .
( [ fft ] ) , the operator @xmath26 can be either equal to the position operator @xmath27 , or to the velocity operator @xmath28 $ ] , or to the acceleration operator @xmath29 $ ] ( where @xmath30 is the total time - dependent hamiltonian ) , defining three different forms of the power spectrum : the dipole @xmath31 , the velocity @xmath32 , and the acceleration @xmath33 forms . according to recent works @xcite , the velocity form @xmath32 best represents the hhg spectrum of a single atom or molecule .
the three forms are related to each other by ( see appendix [ app : powerspectrum ] ) : @xmath34 in this work , we always show the same quantity , i.e. the velocity hhg spectrum , either extracted directly from the velocity power spectrum , or indirectly from the dipole or the acceleration power spectrum with the appropriate frequency factors following eq .
( [ eq : equi ] ) .
the time - dependent schrdinger equation is solved using the tdci method ( see , e.g. , refs . )
applied to the special case of the h atom .
the wave function @xmath35 is expanded in the discrete basis of the eigenstates @xmath36 of the field - free hamiltonian @xmath37 ( projected in the same basis ) , composed of the ground state ( @xmath38 ) and all the excited states ( @xmath39 ) @xmath40 where @xmath41 are time - dependent coefficients .
inserting eq .
( [ tdcis_state ] ) into eq .
( [ tdse ] ) , and projecting on the eigenstates @xmath42 , gives the evolution equation @xmath43 where @xmath44 is the column matrix of the coefficients @xmath41 , @xmath45 is the diagonal matrix of elements @xmath46 ( where @xmath47 is the energy of the eigenstate @xmath48 ) , and @xmath49 is the non - diagonal matrix of elements @xmath50 . the initial wave function at @xmath51 is chosen to be the field - free ground state , i.e. @xmath52 . to solve eq .
( [ ciequation ] ) , time is discretized and the simple split - propagator approximation is used to separate the contributions of the field - free hamiltonian @xmath45 and the atom - field interaction @xmath49 @xmath53 where @xmath54 is a small time step . since the matrix @xmath45 is diagonal , @xmath55 is a diagonal matrix of elements @xmath56 .
the exponential of the non - diagonal matrix @xmath49 is calculated as @xmath57 where @xmath58 is the unitary matrix describing the change of basis between the original eigenstates of @xmath37 and a basis in which the atom - field interaction @xmath59 is diagonal , i.e. @xmath60 where @xmath61 is the diagonal atom - field interaction matrix and @xmath62 is the diagonal representation matrix of the position operator .
since the time dependence is simply factorized in a multiplicative function independent of @xmath63 , the unitary matrix @xmath58 is time - independent and can be calculated once and for all before the propagation .
once the time - dependent coefficients are known , it is possible to calculate the time - dependent dipole , velocity , or acceleration as @xmath64 which , after taking the square of its fourier transform , leads to the corresponding power spectrum of eq .
( [ fft ] ) .
the field - free states ( simply corresponding to the atomic orbitals for the h atom ) are expanded on a gaussian basis set , @xmath65 where @xmath66 are real - valued gto basis functions centered on the nucleus . in
spherical coordinates @xmath67 , @xmath68 where @xmath69 is a normalization constant , @xmath70 are exponents , @xmath71 are real spherical harmonics .
we built the gaussian basis set starting from the dunning basis sets @xcite , adding first diffuse gto functions to describe the rydberg states , and a special set of gto functions adjusted to represent low - lying continuum states .
for the latter , we follow kaufmann _ et al . _
@xcite who proposed to fit gto basis functions to slater - type orbital basis functions having a single fixed exponent @xmath5 . for each angular momentum @xmath72 , kaufmann _
et al . _ found a sequence of optimized gto exponents which are well represented by the simple formula @xcite @xmath73 where @xmath74 is not associated to the quantum principal number but is just an index identifying a given value in the list of all exponents for a fixed @xmath72 , and the parameters @xmath75 and @xmath76 are given in table 2 of ref . .
the gto basis functions obtained with these exponents ( collected in table [ tab : continuum ] ) will be in the following referred to as `` gaussian continuum functions '' or `` kaufmann ( k ) functions '' .
.exponents @xmath77 [ see eq .
( [ kauf_sca ] ) ] of the gaussian functions for describing the continuum proposed by kaufmann _
_ @xcite and used in the present work for @xmath78 and @xmath79 .
[ tab : continuum ] [ cols="^,^,^,^",options="header " , ] the gto basis set incompleteness is responsible for an incorrect description of the continuum eigenfunctions .
they decay too fast for large @xmath80 , which prevents the description of the above - threshold ionization and leads to unphysical reflections of the wave function in the laser - driven dynamics . to compensate for this
, we use the heuristic lifetime model of klinkusch _ et al . _
@xcite which consists in interpreting the approximate field - free eigenstates @xmath81 above the ionization threshold ( taken as the zero energy reference ) as non - stationary states and thus replacing , in the time propagation , the energies @xmath47 by complex energies @xmath82 , where @xmath83 is the inverse lifetime of state @xmath48 .
for the special case of the h atom the @xmath83 are chosen as @xcite @xmath84 where @xmath85 is an empirical parameter representing the characteristic escape length that the electron in the state @xmath48 is allowed to travel during the lifetime @xmath86 .
these complex energies are used in the propagation described by eq .
( [ coefficients ] ) , in the field - free hamiltonian matrix @xmath45 .
the heuristic lifetime model is a simple alternative to using complex scaling @xcite , a complex - absorbing potential @xcite , or a wave - function absorber @xcite . in this work ,
we also introduce and test a modified version of the original heuristic lifetime model . in this version , two different values of the escape length , @xmath87 and @xmath88 ,
are used to increase the flexibility in the definition of the finite lifetimes , adapted to the present context of hhg .
a large value of @xmath87 ( small value of @xmath83 ) is used for all the above - ionization - threshold states with positive energy below the energy cutoff of the three - step model @xmath89 , while a smaller @xmath88 ( larger @xmath83 ) is used for the continuum states with energies above @xmath89 , which are not expected to contribute to hhg .
this allows us to better retain the contribution of low - energy continuum states for the recombination step of the hhg process .
c c c c @xmath90 & 5@xmath9110@xmath92 w/@xmath93 & 10@xmath94 w/@xmath93 & 2@xmath9110@xmath94 w/@xmath93 + @xmath95 nm + @xmath96 & 1.51 & 1.06 & 0.76 + @xmath97 & 0.11 & 0.22 & 0.44 + @xmath89 & 0.85 & 1.20 & 1.89 + @xmath98 & 15 & 21 & 33 + @xmath99 & 23 & 33 & 46 + + @xmath100 nm + @xmath96 & 1.13 & 0.79 & 0.57 + @xmath97 & 0.19 & 0.40 & 0.78 + @xmath89 & 1.10 & 1.77 & 2.97 + @xmath98 & 26 & 41 & 69 + @xmath99 & 41 & 59 & 82 + the field - free calculations are performed using a development version of the molpro software package @xcite from which all the electronic energies , as well as the dipole , velocity , and acceleration matrix elements over the electronic states have been obtained .
the external code light @xcite is used to perform the time - propagation using a time step @xmath54 = 2.42 as ( 0.1 a.u . ) and the fourier transformations with a hann window function .
an escape length @xmath85=1.41 bohr is used for the original heuristic lifetime model , while @xmath101 bohr and @xmath102 bohr are chosen for the modified version of the heuristic lifetime model as explained in the section results and discussion .
correlation - consistent @xmath103-aug - cc - pv@xmath104z @xcite basis sets are used , where @xmath104 is the cardinal number ( @xmath105 t , q , 5 ) connected to the maximum angular momentum ( @xmath106 for the h atom ) , and @xmath103 is the number of shells of diffuse functions for each angular momentum .
we only employ @xmath107 or @xmath108 because @xmath107 can be considered as the minimum augmentation needed to reasonably describe hhg spectra for the h atom @xcite . in particular , the 6-aug - cc - pvtz basis set describes up to @xmath109-shell rydberg states , 6-aug - cc - pvqz up to @xmath110-shell rydberg states , and 6-aug - cc - pv5z up to @xmath111-shell rydberg states .
furthermore , we investigate the effect of adding to the 6-aug - cc - pvtz basis set 3 , 5 , and 8 gaussian continuum functions ( or k functions ) for each angular momentum .
the extra diffuse and continuum gaussian functions are uncontracted . for comparison
, we also perform accurate grid calculations in the length gauge .
the wave function is expanded on a set of spherical harmonics @xmath112 up to @xmath113 , and the resulting coupled equations are discretized on a radial grid with a step size of @xmath114 bohr ( see ref . ) .
a box size of 256 bohr is used with a mask function @xcite at 200 bohr to absorb the part of the wave function accounting for ionized electrons that will not rescatter towards the nucleus .
the mask function multiplying the wave function at each time step has been chosen to be cos(@xmath80)@xmath115 , which is effective in modeling the ionization @xcite .
the time step used is @xmath54=0.65 as ( 0.027 a.u . ) .
the grid - based calculations , being converged with respect to the parameters mentioned above , represent the numerical reference for the current gto results .
we note that performing the grid - based calculations takes hours on a standard workstation , while the field - free and time - propagation calculations in the gto basis sets take only a few minutes .
unless otherwise noted , the calculations are done with the carrier laser frequency @xmath116 = 1.550 ev ( @xmath117 nm ) , corresponding to a ti : sapphire laser . for the comparison with the grid calculations
, we also use the laser frequency @xmath116 = 1.165 ev ( @xmath118 nm ) for which higher - energy regions are probed .
the pulse duration is @xmath119 oc where 1 optical cycle ( oc ) is @xmath120 ( 110.23 a.u . ) .
we use three peak laser intensities @xmath121 : @xmath122 w/@xmath93 , @xmath123 w/@xmath93 , and @xmath124 w/@xmath93 .
we have thus chosen a range of intensities encompassing the over - barrier ionization threshold ( i.e. the critical intensity above which the electron can classically overstep the barrier ) of hydrogen , @xmath125 w/@xmath126 .
we can therefore study the performance of our method in realistic conditions for which hhg progressively becomes less pronounced with increasing laser intensity .
the physical parameters relevant to hhg are reported in table [ tab : lasers ] .
we start by studying the performance of several gaussian basis sets for the calculation of hhg spectra of the h atom , continuing the previous work of luppi and head - gordon @xcite .
the optimal basis set including gaussian continuum functions is then used for a direct comparison with reference hhg spectra from grid calculations .
( left ) , velocity @xmath127 ( middle ) , and acceleration @xmath128 ( right ) calculated with the 6-aug - cc - pvtz basis set for laser intensities @xmath12910@xmath92 w/@xmath126 ( top ) , @xmath123 w/@xmath126 ( middle ) , and @xmath13010@xmath94 w/@xmath126 ( bottom ) .
[ fig1],title="fig:",width=288 ] + ( left ) , velocity @xmath127 ( middle ) , and acceleration @xmath128 ( right ) calculated with the 6-aug - cc - pvtz basis set for laser intensities @xmath12910@xmath92 w/@xmath126 ( top ) , @xmath123 w/@xmath126 ( middle ) , and @xmath13010@xmath94 w/@xmath126 ( bottom ) .
[ fig1],title="fig:",width=288 ] + ( left ) , velocity @xmath127 ( middle ) , and acceleration @xmath128 ( right ) calculated with the 6-aug - cc - pvtz basis set for laser intensities @xmath12910@xmath92 w/@xmath126 ( top ) , @xmath123 w/@xmath126 ( middle ) , and @xmath13010@xmath94 w/@xmath126 ( bottom ) .
[ fig1],title="fig:",width=288 ] we have reported on figure [ fig1 ] the time evolution of the dipole @xmath131 , the velocity @xmath127 , and the acceleration @xmath128 with the 6-aug - cc - pvtz basis set for the three laser intensities . the evolution of @xmath131 , @xmath127 , and @xmath128 follows the shape of the laser field given in eq .
( [ enve ] ) , with the shape of their envelopes changing with the intensity of the pulse .
note that @xmath127 is one order of magnitude smaller than @xmath131 and its oscillations have a finer structure .
similarly , @xmath128 is one order of magnitude smaller than @xmath127 and has even more structured oscillations . even though some fast oscillations are still present after the laser is switched off due to the population of electronic excited states , the conditions @xmath132 and @xmath133 ( see appendix [ app : powerspectrum ] ) are approximately fulfilled , which will allow us to use eq .
( [ eq : equi ] ) .
our results are in reasonable agreement with the results of bandrauk _ et al . _
@xcite and those of han and madsen @xcite who used grid - based methods .
similar findings have been reported for the he atom in a low - field regime using time - dependent hartree - fock and time - dependent kohn - sham with gaussian basis sets @xcite .
( i.e. @xmath134 ) , the velocity power spectrum @xmath135 ( i.e. @xmath136 ) , and the acceleration power spectrum @xmath137 ( i.e. @xmath138 ) calculated with the 6-aug - cc - pvtz basis set and laser intensities @xmath12910@xmath92 w/@xmath126 , @xmath123 w/@xmath126 , and @xmath13910@xmath94 w/@xmath126 . the ionization threshold ( @xmath140 , vertical dashed line ) and the harmonic cutoff in the three - step model @xmath98 ( vertical dot - dashed line )
are also shown .
[ fig2],width=336 ] in figure [ fig2 ] the velocity hhg spectrum , extracted from the dipole , velocity , and acceleration power spectra according to eq .
( [ eq : equi ] ) , calculated with the 6-aug - cc - pvtz basis set and the three laser intensities are shown . the typical form of the hhg spectrum ( plateau / cutoff / background ) is obtained .
we note that the harmonic peaks that we obtained are sharper than those calculated by bandrauk _
et al . _
@xcite based on a direct propagation of the time - dependent schrdinger equation on a grid . the dipole , velocity , and acceleration formulations of the velocity hhg spectrum give similar spectra in the plateau region , but different backgrounds beyond the harmonic cutoff .
in particular , the hhg spectrum calculated from the dipole power spectrum presents a higher background than the hhg spectra calculated from the velocity and acceleration power spectra , in agreement with the calculations of bandrauk _ et al .
_ @xcite .
these differences reflect the sensitivity to the basis set .
indeed , the expectation value of the dipole operator probes the time - dependent wave function in spatial regions further away from the nucleus than the expectation values of the velocity and acceleration operators do . in the following ,
since the dipole is the most difficult to converge with our basis set we will focus on the basis set convergence of the ( velocity ) hhg spectrum computed from the dipole power spectrum .
we first analyze the effect of the basis - set cardinal number @xmath104 , before examining the effect of adding gaussian continuum basis functions in sec .
[ conti ] .
we use the following series of basis sets : 6-aug - cc - pvtz ( s , p , and d shells ) , 6-aug - cc - pvqz ( s , p , d , and f shells ) , and 6-aug - cc - pv5z ( s , p , d , f , and g shells ) .
the number of total , bound ( i.e. , energy below 0 ) , and continuum ( i.e. , energy above 0 ) states , and the maximum energy obtained with these basis sets are reported in the upper half of table [ angmom ] .
going from 6-aug - cc - pvtz to 6-aug - cc - pv5z the total number of states increases considerably , from 68 to 205 .
the percentage of continuum states also tends to increase with the cardinal number .
however , these added continuum states are not necessarily in the energy range relevant to the hhg spectrum .
indeed , the maximum energies obtained are 3.45 hartree for 6-aug - cc - pvtz , 7.74 hartree for 6-aug - cc - pvqz , and 15.94 hartree for 6-aug - cc - pv5z , while the maximal kinetic energy that can be transmitted to the electron ( @xmath141 ) in the three - step model are between 0.35 and 2.47 hartree for the parameters considered ( see table [ tab : lasers ] ) .
l > + m10 mm > + m05 mm > + m05 mm > + m05 mm > + m05 mm > + m10 mm & total & & & @xmath142 + 6-aug - cc - pvtz & 68 & 42 & ( 62% ) & 26 & ( 38% ) & 3.45 + 6-aug - cc - pvqz & 126 & 63 & ( 50% ) & 63 & ( 50% ) & 7.74 + 6-aug - cc - pv5z & 205 & 90 & ( 44% ) & 115 & ( 56% ) & 15.94 + 6-aug - cc - pvtz+3k & 95 & 42 & ( 44% ) & 53 & ( 56% ) & 6.31 + 6-aug - cc - pvtz+5k & 113 & 46 & ( 41% ) & 67 & ( 59% ) & 6.68 + 6-aug - cc - pvtz+8k & 140 & 51 & ( 36% ) & 89 & ( 64% ) & 6.93 + in figure [ fig3 ] , we compare the velocity hhg spectrum extracted from the dipole power spectrum for the 6-aug - cc - pvtz , 6-aug - cc - pvqz , and 6-aug - cc - pv5z basis sets for the laser intensity @xmath123 w/@xmath126 .
the three basis sets give very similar results , in the plateau as well as beyond the harmonic cutoff .
we thus conclude that the hhg spectrum is not strongly affected by the cardinal number @xmath104 of the basis set and therefore , in the following , we will use a triple - zeta ( @xmath105 t ) basis set . in figure
[ fig3 ] , we also compare the spectra calculated using the @xmath103-aug - cc - pv@xmath104z basis sets with @xmath107 and @xmath108 .
the results show that the convergence in terms of diffuse basis functions is achieved with 6 diffuse shells . calculated with the 6-aug - cc - pv@xmath104z and 9-aug - cc - pv@xmath104z basis sets with @xmath104= t ( left ) ,
q ( middle ) and 5 ( right ) .
the laser intensity is @xmath123 w/@xmath126 . the ionization threshold ( @xmath140 , vertical dashed line ) and the harmonic cutoff in the three - step model @xmath98 ( vertical dot - dashed line ) are also shown .
[ fig3],title="fig:",width=336 ] + the sensitivity of the hhg spectrum to the cardinal number and to the number of diffuse functions led us to select the 6-aug - cc - pvtz basis set as the reference basis set to include the gaussian continuum functions of kaufmann _ et al . _
we have added 3 , 5 , and 8 gaussian continuum functions ( denoted by k ) for each angular momentum in the 6-aug - cc - pvtz basis set . in the lower half of table [ angmom ] , the number of total , bound , and continuum states and the maximum energy obtained with these 6-aug - cc - pvtz+3k , 6-aug - cc - pvtz+5k , and 6-aug - cc - pvtz+8k basis sets is reported .
it is noteworthy that increasing the number of k functions hardly affects the number of bound states , in favor of positive energy states , thus focusing the improvement on the description of the continuum .
more precisely , as the maximum energy obtained with these three basis sets is nearly unchanged ( 6.313 , 6.681 , and 6.927 hartree , respectively ) , the k functions increase the density of states in the energetically important region of the continuum .
we show in figure [ fig4 ] the distribution of the state energies for the different basis sets . increasing
the number of k functions essentially does not change the energy spectrum below the ionization threshold , while an almost continuum distribution builds up in the low - energy region above the ionization threshold . when compared with the 6-aug - cc - pvtz basis set
, the distribution of the continuum states becomes more dense ( closer to a `` real '' continuum ) and the gaps between ( near-)degenerate sets of states become smaller . in particular , the density of states is improved in the region from the ionization threshold to around 1 hartree , which is also the most relevant energy region for hhg for the laser intensity range studied here , according to the three - step model .
by contrast , luppi and head - gordon @xcite showed that adding diffuse functions increases the density of rydberg states , leaving the density of continuum states mostly unchanged .
, width=288 ] the upper panel of figure [ fig6 ] compares the radial wave function @xmath143 of a s continuum state at the energy @xmath144 hartree obtained with the 6-aug - cc - pvtz+8k basis set with the analytical solution of the time - independent schrdinger equation @xcite . for completeness
, the radial wave function from the grid calculation is also shown and is perfectly superimposed with the analytical solution . the radial wave function obtained with the 6-aug - cc - pvtz+8k basis set is a reasonable approximation to the exact solution , the continuum gaussian functions correctly reproducing the oscillations of the function up to a radial distance as large as 30 bohr .
this radial distance is consistent with the maximum distance @xmath99 ( see table [ tab : lasers ] ) traveled by the electron predicted by the three - step model with the laser parameters used here . for comparison ,
the lower panel of figure [ fig6 ] shows the radial wave function @xmath143 obtained with the 6-aug - cc - pvtz basis set for a similar s continuum state at the closest energy obtained with this basis set , @xmath145 hartree .
clearly , the basis set without the continuum gaussian functions is only able to describe the short - range part of function @xmath143 but not the long - range oscillating part .
@xcite and the radial wave function obtained using the 6-aug - cc - pvtz+8k basis set for a s continuum state at the energy @xmath144 hartree ( upper panel ) . in the lower panel , the same comparison is done for a similar state of close energy @xmath145 hartree but with the 6-aug - cc - pvtz basis set , i.e. without the kaufmann basis functions .
the radial wave functions obtained in the grid calculations are also shown .
since continuum wave functions can not be normalized in the standard way , the curves have been scaled in order to approximately have the same value at the first minimum.,title="fig:",width=288 ] + @xcite and the radial wave function obtained using the 6-aug - cc - pvtz+8k basis set for a s continuum state at the energy @xmath144 hartree ( upper panel ) . in the lower panel ,
the same comparison is done for a similar state of close energy @xmath145 hartree but with the 6-aug - cc - pvtz basis set , i.e. without the kaufmann basis functions .
the radial wave functions obtained in the grid calculations are also shown . since
continuum wave functions can not be normalized in the standard way , the curves have been scaled in order to approximately have the same value at the first minimum.,title="fig:",width=288 ] calculated with the 6-aug - cc - pvtz basis set plus 3 ( left ) , 5 ( middle ) , and 8 ( right ) gaussian continuum ( k ) functions .
the laser intensity is @xmath12910@xmath92 w/@xmath126 ( top ) , @xmath123 w/@xmath126 ( middle ) , and @xmath13010@xmath94 w/@xmath126 ( bottom ) .
the ionization threshold ( @xmath140 , vertical dashed line ) and the harmonic cutoff in the three - step model @xmath98 ( vertical dot - dashed line ) are also shown .
[ fig5],title="fig:",width=288 ] + calculated with the 6-aug - cc - pvtz basis set plus 3 ( left ) , 5 ( middle ) , and 8 ( right ) gaussian continuum ( k ) functions .
the laser intensity is @xmath12910@xmath92 w/@xmath126 ( top ) , @xmath123 w/@xmath126 ( middle ) , and @xmath13010@xmath94 w/@xmath126 ( bottom ) .
the ionization threshold ( @xmath140 , vertical dashed line ) and the harmonic cutoff in the three - step model @xmath98 ( vertical dot - dashed line ) are also shown .
[ fig5],title="fig:",width=288 ] + calculated with the 6-aug - cc - pvtz basis set plus 3 ( left ) , 5 ( middle ) , and 8 ( right ) gaussian continuum ( k ) functions .
the laser intensity is @xmath12910@xmath92 w/@xmath126 ( top ) , @xmath123 w/@xmath126 ( middle ) , and @xmath13010@xmath94 w/@xmath126 ( bottom ) .
the ionization threshold ( @xmath140 , vertical dashed line ) and the harmonic cutoff in the three - step model @xmath98 ( vertical dot - dashed line ) are also shown .
[ fig5],title="fig:",width=288 ] in figure [ fig5 ] the velocity hhg spectrum extracted from the dipole power spectrum is shown for the 6-aug - cc - pvtz+3k , 6-aug - cc - pvtz+5k , and 6-aug - cc - pvtz+8k basis sets and for the three laser intensities .
we focus our attention to the post - cutoff background region of the spectrum since diminishing the background in this region is an important goal of the present work . considering the laser intensity @xmath146 w/@xmath93 and analyzing the spectra between the 20th and 40th harmonics , we observe that the hhg spectrum with the 6-aug - cc - pvtz+3k basis set resembles the one obtained with the original 6-aug - cc - pvtz basis set , with no obvious improvement . when adding 5 or 8 k functions the background is strongly diminished , while the harmonics before the cutoff are not substantially changed .
the same trend is also observed for laser intensities @xmath147 w/@xmath93 and @xmath148 w/@xmath93 , even if the lowering of the background is not as strong . as demonstrated in ref .
, the rydberg bound states strongly contribute to the background of the hhg spectrum .
the addition of gaussian continuum functions to the basis set allows one to appropriately describe the low - lying continuum states , leading to a more balanced basis set yielding a lower background and therefore a much clearer identification of the cutoff region .
of course , such an improvement depends on the intensity of the laser pulse , since larger intensities require to describe continuum states of higher energy and therefore require more gaussian continuum functions . calculated with the 6-aug - cc - pvtz+8k basis set with two lifetime models and with grid calculations , for the two laser wavelength @xmath95 nm ( upper panel ) and 1064 nm ( lower panel ) , the laser intensities
@xmath14910@xmath92 , 10@xmath94 , and 2@xmath9110@xmath94 w/@xmath93 . the ionization threshold ( @xmath140 , vertical dashed line ) and the harmonic cutoff in the three - step model @xmath98 ( vertical dot - dashed line )
are also shown .
[ fig8],title="fig:",width=288 ] calculated with the 6-aug - cc - pvtz+8k basis set with two lifetime models and with grid calculations , for the two laser wavelength @xmath95 nm ( upper panel ) and 1064 nm ( lower panel ) , the laser intensities @xmath14910@xmath92 , 10@xmath94 , and 2@xmath9110@xmath94 w/@xmath93 . the ionization threshold ( @xmath140 , vertical dashed line ) and the harmonic cutoff in the three - step model @xmath98 ( vertical dot - dashed line ) are also shown .
[ fig8],title="fig:",width=288 ] for the laser wavelength @xmath95 and the intensities @xmath12910@xmath92 w/@xmath126 ( left ) , @xmath123 w/@xmath126 ( middle ) , and @xmath13010@xmath94 w/@xmath126 ( right ) obtained with the 6-aug - cc - pvtz+8k basis set with the original ( one parameter ) and modified ( two parameters ) lifetime models , compared with the results from the grid calculations.,width=288 ] we now investigate in more detail the performance of the 6-aug - cc - pvtz+8k basis set by comparison with grid calculations . in figure [ fig8 ]
we compare the velocity hhg spectrum extracted from the acceleration power spectrum , obtained with the 6-aug - cc - pvtz+8k basis set with two lifetime models , and with grid calculations for the same laser intensities as before and for @xmath95 and 1064 nm .
the acceleration power spectrum , rather than the dipole power spectrum , was chosen here because the grid calculation is easier to converge for the acceleration power spectrum . for the intensity @xmath14910@xmath92 w/@xmath93 and for the two wavelengths ,
the spectra obtained with the gaussian basis set and the original lifetime model ( with @xmath150 bohr ) are in good agreement with the ones from the grid calculations . in particular
, the cutoff appears at almost the same energy .
however , for the larger intensities , the intensity of the higher harmonics in the plateau obtained with the gaussian basis set decrease too rapidly . this can be attributed to a limitation of the original lifetime model which assigns too large lifetimes to high - energy continuum states .
for this reason we introduce a modified version of the lifetime model , with two different values of the parameter @xmath85 : a large value , @xmath101 bohr , for continuum states with positive energies below the energy cutoff of the three - step model @xmath141 , and a small value , @xmath102 bohr , for continuum states with energies above @xmath141 .
the ionization rates @xmath83 are thus smaller than in the original model for low - lying continuum states , and larger for high - lying continuum states .
this choice allows us to get a more accurate description of the harmonics in the plateau and close to the cutoff .
the values of @xmath87 and @xmath88 have been chosen comparing with the corresponding grid hhg spectra . not surprisingly , the value of @xmath87 is of the order of magnitude of the electron excursion distance @xmath99 ( see table [ tab : lasers ] ) .
we test our modified lifetime model by calculating the ionization probability ( for both the grid and gaussian - basis - set calculations ) @xmath151 where the sum runs over all the bound states .
figure [ fig7 ] reports @xmath152 obtained with the original and the modified lifetime models and from the grid calculations for the three laser intensities .
the original lifetime model leads to largely overestimated ionization probabilities in comparison to the results obtained from the grid calculations .
our modified lifetime model reduces the ionization probability and is in better agreement with the grid calculations , especially for the intensities @xmath147 w/@xmath93 and @xmath148 w/@xmath93 . coming back to figure [ fig8 ]
, it is seen that the combined use of the 6-aug - cc - pvtz+8k basis set and of the modified lifetime model results in a hhg spectrum that is in good agreement with the one obtained with the grid calculation at wavelength @xmath95 nm and the intensity @xmath123 w/@xmath93 .
the general shape of the spectrum and the position of the harmonic cutoff are well reproduced with the gaussian basis set , the only remaining differences being larger peaks and a larger background after the cutoff in comparison to the grid results .
for the same wavelength and the largest intensity @xmath15310@xmath94 w/@xmath93 , the agreement is also fairly good even though the position of the harmonic cutoff predicted with the gaussian basis set is slightly too low . the longer laser wavelength @xmath100 nm represents a more stringent test for our method since higher - energy regions are probed ( see table [ tab : lasers ] ) .
the agreement between the hhg spectra obtained with the gaussian basis set and from the grid calculations is still pretty good for the intensity @xmath14910@xmath92 w/@xmath93 , while the position of the cutoff is slightly underestimated for the intensity @xmath123 w/@xmath93 and significantly underestimated for the largest intensity @xmath154 w/@xmath93 .
this likely comes from a too poor description of the continuum states above 1 hartree with the 6-aug - cc - pvtz+8k basis set , which can be populated for these wavelengths and intensities . a larger number of continuum gaussian functions is needed in order to improve the high - energy part of the hhg spectrum for the largest intensities .
we note , however , that increasing the number of continuum gaussian functions can lead to near - linear dependencies in the basis set ( seen with the presence of very small eigenvalues of the overlap matrix of the basis functions ) and thus numerical instability issues in self - consistent - field calculations .
in this work , we have explored the calculation of the velocity hhg spectrum of the h atom extracted from the dipole , velocity , and acceleration power spectra with gaussian basis sets for different laser intensities and wavelengths . while all the three power spectra give reasonable velocity hhg spectra with similar harmonic peaks before the cutoff
, they tend to differ in the background region beyond the cutoff .
the hhg spectrum extracted from the dipole power spectrum is the most sensitive to the basis set . with the 6-aug - cc - pvtz basis set
it leads to a high background which blurs out the location of the plateau cutoff .
increasing the cardinal number of the basis set ( from @xmath155 to @xmath156 ) or the number of diffuse basis functions ( from @xmath107 to @xmath108 ) does not improve the hhg spectrum .
by contrast , adding 5 or 8 gaussian continuum functions , as proposed by kaufmann _
et al . _
@xcite , leads to an improvement of the velocity hhg spectrum extracted from the dipole power spectrum at least for laser intensities up to @xmath157 w/@xmath93 by decreasing the background , which thus allows one to better identify the cutoff region .
the combined use of gaussian continuum functions and a heuristic lifetime model with two parameters for modeling ionization results is in a fairly good agreement with the reference hhg spectra from grid calculations , in terms of the general shape of the spectrum , the number and intensity of peaks , and the position of the cutoff .
the agreement is less satisfactory for the largest intensities because the high - energy continuum states are poorly reproduced by the gaussian basis set calculations .
improving the accuracy for the largest intensities would require a larger number of gaussian continuum functions .
gaussian continuum functions thus appear as a promising way of constructing gaussian basis sets for studying electron dynamics in strong laser fields , allowing one to define a balanced basis set to properly describe both bound and continuum eigenstates .
the present work therefore opens the way to the systematic application of well established quantum chemistry methods with gaussian basis sets to the study of highly nonlinear phenomena ( such as hhg , photoionization cross sections , above - threshold ionization rates , ... ) in atoms and molecules .
this work was supported by the labex michem and calsimlab part of french state funds managed by the anr within the investissements davenir programme under references anr-11-idex-0004 - 02 and anr-11-labx-0037 - 01 .
we thank a. savin for having pointed out to us ref . and v. vniard for useful comments .
in this appendix , we review the relationship between the dipole , velocity , and acceleration forms of the power spectrum @xcite . if we define @xmath158 , where @xmath159 stands for position @xmath13 , velocity @xmath160 , or acceleration @xmath161 , and its fourier transform @xmath162 the three forms of the power spectrum are commonly expressed as @xmath163 and the relationship between the three forms is the relationship between the three @xmath164 . applying eq .
( [ xiomega ] ) for @xmath165 , performing an integration by parts over @xmath166 , and using @xmath167 , gives @xmath168 which , if we have the condition @xmath169 , can be simplified as @xmath170 the relation between @xmath171 and @xmath172 is then @xmath173\bigl ) , \ ; \nonumber\\ \label{vel}\end{aligned}\ ] ] which , in the case where we can make the approximation @xmath174 , simplifies as @xmath175 similarly , applying now eq .
( [ xiomega ] ) for @xmath176 , using @xmath177 , and integrating by parts gives @xmath178 where we have used the condition @xmath179 .
this leads to the relation between @xmath172 and @xmath180 @xmath181\bigl ) , \nonumber\\\end{aligned}\ ] ] which , if we can make the approximation @xmath182 , gives the final approximate relations between the three forms of the spectrum @xmath183
g. sansone , f. kelkensberg , j. f. prez - torres , f. morales , m. f. kling , w. siu , o. ghafur , p. johnsson , m. swoboda , e. benedetti , f. ferrari , f. lpine , j. l. sanz - vicario , s. zherebtsov , i. znakovskaya , a. lhuillier , m. y. ivanov , m. nisoli , f. martin , and m. j. j. vrakking , nature * 465 * , 763 ( 2010 ) . | we explore the computation of high - harmonic generation spectra by means of gaussian basis sets in approaches propagating the time - dependent schrdinger equation .
we investigate the efficiency of gaussian functions specifically designed for the description of the continuum proposed by kaufmann _
et al . _ [ j. phys .
b * 22 * , 2223 ( 1989 ) ] .
we assess the range of applicability of this approach by studying the hydrogen atom , i.e. the simplest atom for which `` exact '' calculations on a grid can be performed .
we notably study the effect of increasing the basis set cardinal number , the number of diffuse basis functions , and the number of gaussian pseudo - continuum basis functions for various laser parameters .
our results show that the latter significantly improve the description of the low - lying continuum states , and provide a satisfactory agreement with grid calculations for laser wavelengths @xmath0 = 800 and 1064 nm .
the kaufmann continuum functions therefore appear as a promising way of constructing gaussian basis sets for studying molecular electron dynamics in strong laser fields using time - dependent quantum - chemistry approaches . |
is a pulsating white dwarf with a hydrogen atmosphere , a dav , also called a zz ceti star @xcite .
these stars show multi - periodic non - radial @xmath5-mode pulsations that can be used to measure their internal properties and rate of evolution .
@xcite found the star was variable , and @xcite studied its light curve , finding 6 simultaneous pulsations .
the dominant mode has a period of 215 s , a fractional amplitude of 22 mma ( milli - modulation amplitude = 1/1.086 milli magnitude ) , and is stable in amplitude and phase .
the other smaller pulsation modes vary in amplitude from night to night @xcite , suggesting the presence of unresolved components . because the davs appear to be normal stars except for their variability @xcite , i.e. , an evolutionary stage in the cooling of all white dwarfs , it is likely that the dav structural properties are representative of _ all _ da white dwarfs .
@xcite and @xcite discovered 46 new zz cetis and more than doubled the number of known variables , using @xmath6 and @xmath7 g values derived from the optical spectra obtained by the sloan digital sky survey @xcite .
while @xcite find the zz ceti instability strip to be pure , i.e. , contain no non - variable stars , @xcite find several stars inside the instability strip for which they could detect no variability above their detection threshold .
it is necessary to obtain @xmath8 spectra of the non - variables within the strip and re - analyze their @xmath6 and @xmath7 g values , and also to obtain additional time series photometry on these stars , to ensure that they are not low amplitude variables , and re - study the purity of the instability strip .
we report our continuing study of the star g 117b15a , also called ry lmi and wd 0921 + 354 , one of the hottest of the zz ceti stars .
we expect that the rate of change of a pulsation period with time for @xmath5-mode pulsations in white dwarf stars to be directly related to its evolutionary timescale @xcite , allowing us to infer the age of a cool white dwarf since its formation .
we have been observing the star since 1974 to measure the rate of period change with time ( @xmath9 ) for the largest amplitude periodicity .
using all the data obtained from 1974 through 1999 , @xcite arrived at a determination of @xmath10 .
g 117b15a was the first pulsating white dwarf to have its main pulsation mode index identified .
the 215 s mode is an @xmath11 , as determined by comparing the ultraviolet pulsation amplitude ( measured with the hubble space telescope ) to the optical amplitude @xcite . @xcite
confirm the @xmath12 measurement for the p=215 s pulsation , and show that the other large amplitude modes , at 271 s and 304 s , show chromatic amplitude changes that do not fit the theoretical models , using time - resolved spectra obtained at the keck telescope .
@xcite , and @xcite derive @xmath6 near 12,400 k , while @xcite using a less efficient model for convection , derive @xmath6=11,630 k. @xcite demonstrated that the observed variations in the light curve of g 117b15a are due to non - radial _ g_-mode pulsations .
@xcite show the models predict the effect of radius change due to still ongoing contraction to be an order of magnitude smaller than the cooling effect on the rate of period change .
g 117b15a is proving to be a useful laboratory for particle physics @xcite .
@xcite calculated the limit on the axion mass compatible with the then observed upper limit to the cooling , showing @xmath13 and @xcite demonstrates axion cooling would be dominant over neutrino cooling for the lukewarm white dwarf stars for axion masses of this order .
@xcite show that the @xmath14 upper limit published in @xcite limits the string mass scale @xmath15 for 6 dimensions , from the observed cooling rate and the emission of kaluza - klein gravitons , but the limit is negligible for higher dimensions .
@xcite show the observed rates of period change can also be used to constrain the dynamical rate of change of the constant of gravity @xmath16 .
@xcite used the mode identification and the observed periods of the 3 largest known pulsation modes to derive a hydrogen layer mass lower limit of @xmath17 , and a best estimate of @xmath18 , assuming @xmath19 for the 215 s mode , and 20:80 c / o core mass .
the core composition is constrained mainly by the presence of the 304 s pulsation .
@xcite show the seismological models with time - dependent element diffusion are only consistent with the spectroscopic data if the modes are the @xmath11 , k=2 , 3 , and 4 , and deduces @xmath20 , @xmath21 and @xmath22 k. their best model predicts : parallax @xmath23=15.89 mas , @xmath24 s / s , for the p=215 s , @xmath25 s / s , for the p=271 s , and @xmath26 s / s , for the p=304 s periodicities .
@xcite reported on the observations from 1974 to 2000 .
we report in this paper additional 19.3 h of time series photometry in 2001 , 30.6 h in 2002 , 24.4 h in 2003 , 4 h in 2004 , and 13.6 h in 2005 , most using the argos prime - focus ccd camera @xcite on the @xmath27 m otto struve telescope at mcdonald observatory .
we observed the light through a bg40 filter to maximize the signal - to - noise ratio ( s / n ) because the pulsation amplitudes are small ( 2% ) , the star is faint ( v=15.52 , @xcite ) , and also because the non - radial _
g_-mode light variations have the same phase in all colors @xcite but the amplitudes decrease with wavelength .
for example , a filter - less observation with argos gives an amplitude around 40% smaller for g 117b15a .
we reduce and analyze the data in the manner described by @xcite , and @xcite .
we bring all the data to the same fractional amplitude scale , and the times from utc to the uniform barycentric julian coordinated date ( tcb ) scale , using jpl de405 ephemeris @xcite to model earth s motion .
we compute fourier transforms for each individual run , and verify that the main pulsation at 215 s dominates each data set and has an amplitude stable up to 15% , our uncertainty in amplitude due to the lack of accurate time- and color - dependent extinction determination .
as the dominant pulsation mode at p=215 s has a stable frequency and amplitude since our first observations in 1974 , we can calculate the time of maximum for each new run and look for deviations .
we fit our observed time of maximum light to the equation : @xmath28 where @xmath29 , @xmath30 , @xmath31 is the epoch of the time of maximum , i.e , the integer number of cycles after our first observation , @xmath32 is the time of maximum assumed , @xmath33 is the time of maximum that best fit the parabola , @xmath34 is the period that best fit the parabola , and @xmath35 is the period assumed at @xmath32 $ ] by non - linear least squares gives unreliable uncertainty estimates and the alias space in p and @xmath0 is extremely dense due to 31yr data set span @xcite .
our non - linear least squares result is @xmath36s / s . ] .
the times of maxima are calculated by a linear least - squares fit of the light curve of each night to a sum of the six detected frequencies .
they are shown in table [ tmax ] . in figure
[ figure1 ] , we show the o
c timings after subtracting the correction to period and epoch , and our best fit curve through the data .
the size of each point is proportional to its weight , i.e. , inversely proportional to the square of uncertainty in phase .
the error bars plotted are @xmath37 . from our data through 2005 , we obtain a new value for the epoch of maximum , @xmath38 , a new value for the period , @xmath39 , and most importantly , a rate of period change of : @xmath40 we use linear least squares to make our fit , with each point weighted inversely proportional to the uncertainty in the time of maxima for each individual run squared .
we quadratically add an additional 1s of uncertainty to the time of maxima for each night to account for external uncertainty caused perhaps by the beating of possible small amplitude pulsations @xcite or the small modulation seen in figure [ figure1 ] .
the amplitude , 1s , is chosen from the fourier transform of the ( o - c ) shown in figure [ figure2 ] , and is in agreement with @xcite conclusion that the fourier analysis of the times of arrival ( toa ) gives unbiased information about the noise .
such external uncertainty is also consistent with @xcite who show that the true uncertainties of the times of arrival of the milli - second pulsars are generally larger than the formal uncertainties , and that a quadratic term is added to them to fit the observations .
while it is true that the period change timescale can be proportional to the cooling timescale , it is also possible that other phenomena with shorter timescales can affect @xmath9 .
the cooling timescale is the longest possible one . as a corollary ,
if the observed @xmath9 is low enough to be consistent with evolution , then other processes , such as perhaps a magnetic field or diffusion induced changes in the boundary layers , are not present at a level sufficient to affect @xmath9 .
for the first time we also report on the search for the rates of period changes for the other relatively large amplitude modes of g 117b15a , @xmath41 s / s for the 270s periodicity , and @xmath42 s / s for the 304s periodicity .
they are much larger than the one derived for the main pulsation , which has a region of period formation deeper in the core , but are in line with the measurements for the 274s modes in zz ceti @xcite .
the models to date , without rotation , differential rotation , and magnetic fields , do not explain such values , or the chromatic amplitude changes reported by @xcite .
@xcite show the 270s and 304s periodicities have significant amplitude changes , indicative of multiple components or resonances .
@xcite discusses the influence of the proper motion of the star on the measured @xmath9 : @xmath43 where @xmath44 is the radial velocity of the star . assuming @xmath45 he derived @xmath46 \left(\mu[\," / yr]\right)^2 d[{\mathrm{pc}}]\ ] ] where @xmath47 is the effect of the proper motion on the rate of period change , @xmath34 is the pulsation period
, @xmath48 is the proper motion and @xmath49 is the distance .
the proper motion , @xmath50 , and the parallax , @xmath51 , are given by @xcite .
but the parallax has a large uncertainty and does not agree with other estimates of the distance . @xcite
measured the same proper motion for g 117-b15a and g 117-b15b : @xmath52 mas / yr , and @xmath53 mas / yr , including 5 usno - b epochs and sdss positions .
considering @xmath54 @xcite , if we use @xcite estimate of @xmath55 derived from their optical spectra , we obtain a distance of @xmath56 pc , equivalent to a spectroscopic parallax of @xmath57 " .
but g 117-b15a has international ultraviolet explorer ( iue ) and hubble space telescope ( hst ) faint object spectrograph ( fos ) flux calibrated spectra @xcite that can also be used to estimate the distance , if we use the evolutionary models of @xcite or @xcite to estimate the radius from @xmath6 and @xmath58 and fit the observed fluxes to a model atmosphere calculated by detlev koester , similar to that described in @xcite , but with an ml2/@xmath59=0.6 convection description consistent with @xcite and @xcite .
the iue spectra re - calibrated according to the new spectroscopic image procession system ( newsips ) data reduction by nasa was published by @xcite . from the iue spectra
we get a distance of @xmath60 pc .
the hst spectra , shown in figure [ hst ] , fits a distance of @xmath61 pc .
table [ dist ] lists all the available distances to g 117-b15a . .g
117-b15a distance determinations [ cols= " < , > " , ] the mass of an m3.5v should be around @xmath62 @xcite . with a separation around 13.4 arcsec @xmath63 au , corresponding to an orbital period around ( @xmath64 ) years , we get @xmath65 .
the large uncertainty takes into account the possibility the orbit might be strongly elliptical .
@xcite , in his section 4.3.1 , discusses common proper motion pairs which possibly do not form a physical binary .
if g 117b15a and b form a real binary system , the contribution of an orbital reflex motion to the observed @xmath9 might account for one half of the observed @xmath0 . the whole observed rate of period change _
could _ also be caused by a planet of jupiter s mass orbiting the white dwarf at a distance of 31 au , which corresponds to an orbital period of 223 yr , or a smaller planet on a closer orbit ( see figure [ fergal ] ) .
a planet with jupiter s mass any closer to the white dwarf would lead to a larger @xmath0 .
@xcite show that such a planet would survive the post - main sequence mass loss .
note that reflex motion produces periodic variations on the @xmath66 , which are distinguishable from parabolic variations after a significant portion of the orbit has been covered .
as discussed by @xcite , any relative acceleration of the star with respect to the barycenter of the solar system will contribute to the observed @xmath0 .
their equations ( 2.2 ) for the differential galactic orbits , decomposed in a planar contribution ( 2.12 ) , where the second term is the proper motion correction , and a perpendicular contribution ( 2.28 ) , applied to g 117-b15a , show the galactic contribution to be exactly the one calculated above for proper motion , i.e. , the other terms are negligible ( 2 to 3 orders of magnitude smaller ) .
we have measured the rate of change of the main pulsation period for the @xmath67 k pulsating da white dwarf g 117b15a , the first zz ceti to have its evolutionary rate of change measured , confirming it is the most stable optical clock known , with a rate of change of 1 s in 8.9 million years and a precise laboratory for physics at high energy .
it is important to notice that mode trapping , a resonance between the local pulsation wavelength with the thickness of one of the compositional layers , can reduce the rate of period change by up to a factor of 2 @xcite , but the changes in the trapping layers are still caused by cooling .
it has taken a huge investment of telescope time to achieve such precision , but not only have we measured the cooling rate of this 400 million year old white dwarf @xcite , excluding the time the star took to reach the white dwarf phase , we have also demonstrated it does not harbor planetary bodies similar to jupiter in mass up to a distance around 30 au from the star or smaller planets with light travel time effects on the white dwarf larger than 1s .
we claim that the 215s periodicity in g 117-b15a is the most stable optical clock known . @xcite and @xcite
discuss projects to build optical atomic clocks based on single trapped ions , or several laser cooled neutron atoms , of strontrium or ytterbium , expected to reach an accuracy of @xmath68 s / s . considering their periods are @xmath69 s and @xmath70 s ,
even though they will be more accurate than g 117-b15a , they will be much less stable , as their timescales for period changes , @xmath71 , are 125 s and 3 h , compared to 2 gyr for g 117b15a .
even the hulse & taylor s millisecond pulsar @xcite , has a timescale for period change of only 0.35 gyr @xcite , but the radio millisecond pulsar psr j1713 + 0747 @xcite has @xmath72 s / s , and a timescale of 8 gyr , and psr b1885 + 09=psr j1857 + 0943 with @xmath73
s / s has a stability timescale of 9.5 gyr @xcite . together with these millisecond pulsars ,
g 117b15a may be used as the most stable known natural frequency and time - keeping standards [ e.g. @xcite ] .
the white dwarf has the advantage of not having glitches and less severe general relativistic corrections .
it is essential to note that the measured rate of period change is consistent with the cooling rate in our white dwarf models only for cores of c or c / o . with the refinement of the models ,
the observed rate of period change can be used to accurately measure the ratio of carbon to oxygen in the core of this normal white dwarf , and therefore help in constraining the @xmath74 cross section .
2442477.797089 & 32071 & 0.5 & 1.7 + 2442779.887934 & 153358 & 3.9 & 2.1 + 2442783.850624 & 154949 & 1.2 & 2.9 + 2442786.981458 & 156206 & 2.2 & 1.5 + 2443462.962774 & 427607 & 1.6 & 1.4 + 2443463.946592 & 428002 & 0.5 & 1.4 + 2443465.969049 & 428814 & 0.5 & 1.6 + 2443489.909755 & 438426 & 0.2 & 1.5 + 2443492.898616 & 439626 & 0.9 & 1.6 + 2443521.927837 & 451281 & 0.1 & 1.3 + 2443552.752879 & 463657 & 0.8 & 1.4 + 2443576.725940 & 473282 & -1.6 & 3.3 + 2443581.692438 & 475276 & 0.3 & 1.3 + 2443582.693698 & 475678 & -0.2 & 1.3 + 2443583.697469 & 476081 & 1.0 & 1.3 + 2443584.733602 & 476497 & 0.8 & 1.4 + 2443604.659292 & 484497 & 1.3 & 1.5 + 2443605.752703 & 484936 & 0.4 & 1.4 + 2443611.693050 & 487321 & 0.6 & 1.3 + 2443613.658222 & 488110 & 0.7 & 1.6 + 2443636.674971 & 497351 & 8.8 & 3.4 + 2443839.956765 & 578967 & 5.8 & 3.0 + 2443841.976708 & 579778 & 3.7 & 3.5 + 2443842.980413 & 580181 & -0.7 & 2.2 + 2443843.944332 & 580568 & 0.5 & 2.6 + 2443869.989703 & 591025 & 1.5 & 2.4 + 2443870.946182 & 591409 & 5.5 & 3.1 + 2443874.916339 & 593003 & 2.4 & 2.1 + 2443959.695117 & 627041 & 0.1 & 2.0 + 2443963.662836 & 628634 & 1.6 & 2.1 + 2443990.664641 & 639475 & 2.7 & 1.3 + 2444169.945954 & 711455 & 0.1 & 1.6 + 2444231.822666 & 736298 & -0.7 & 2.9 + 2444232.818992 & 736698 & 3.0 & 1.6 + 2444293.833896 & 761195 & 0.3 & 1.8 + 2444637.776174 & 899285 & 5.8 & 1.9 + 2444641.624287 & 900830 & 2.8 & 1.1 + 2444992.789531 & 1041820 & 0.1 & 1.6 + 2444994.689956 & 1042583 & 1.2 & 1.2 + 2444996.744801 & 1043408 & 2.0 & 1.3 + 2444997.723649 & 1043801 & 1.9 & 1.2 + 2445021.716661 & 1053434 & 1.7 & 1.4 + 2445703.860004 & 1327309 & 1.9 & 1.7 + 2445734.642701 & 1339668 & 2.4 & 1.2 + 2445735.643972 & 1340070 & 2.8 & 1.3 + 2446113.763716 & 1491882 & 2.9 & 1.2 + 2446443.775386 & 1624379 & 2.8 & 1.1 + 2446468.630178 & 1634358 & 2.1 & 1.3 + 2446473.718679 & 1636401 & 0.3 & 1.6 + 2446523.620086 & 1656436 & 2.2 & 1.6 + 2446524.613917 & 1656835 & 5.5 & 2.5 + 2446768.855451 & 1754896 & 2.9 & 1.4 + 2446794.935676 & 1765367 & 2.5 & 2.1 + 2446796.928219 & 1766167 & 0.3 & 1.6 + 2446797.924535 & 1766567 & 3.1 & 1.3 + 2446798.903378 & 1766960 & 2.6 & 1.8 + 2446823.663537 & 1776901 & 3.1 & 1.9 + 2446825.651132 & 1777699 & 3.7 & 1.5 + 2447231.328096 & 1940575 & 3.7 & 1.9 + 2447231.612054 & 1940689 & 5.1 & 3.5 + 2447232.396626 & 1941004 & 5.0 & 1.6 + 2447232.623291 & 1941095 & 5.9 & 2.9 + 2447233.343090 & 1941384 & 4.5 & 1.3 + 2447233.634506 & 1941501 & 4.7 & 2.3 + 2447234.319475 & 1941776 & 6.8 & 3.2 + 2447235.313250 & 1942175 & 5.2 & 1.4 + 2447235.607168 & 1942293 & 6.4 & 2.1 + 2447236.610922 & 1942696 & 6.2 & 1.6 + 2447589.375198 & 2084328 & 3.2 & 1.4 + 2447594.331735 & 2086318 & 5.2 & 1.6 + 2447595.323018 & 2086716 & 3.5 & 2.0 + 2447596.311907 & 2087113 & 10.1 & 2.3 + 2447597.315602 & 2087516 & 4.8 & 1.7 + 2447598.319339 & 2087919 & 3.1 & 3.1 + 2447499.072036 & 2048072 & 6.5 & 3.2 + 2447532.768799 & 2061601 & 1.3 & 1.4 + 2447853.846325 & 2190511 & 4.3 & 2.1 + 2447856.832697 & 2191710 & 5.2 & 1.9 + 2447918.644630 & 2216527 & 2.6 & 3.1 + 2447920.619811 & 2217320 & 6.7 & 3.3 + 2447952.622834 & 2230169 & -3.3 & 2.9 + 2447972.620899 & 2238198 & 9.6 & 6.1 + 2447973.709340 & 2238635 & 9.7 & 2.6 + 2447973.741682 & 2238648 & 6.5 & 1.4 + 2447978.770467 & 2240667 & 10.0 & 2.1 + 2447979.781717 & 2241073 & 11.8 & 3.1 + 2447980.319627 & 2241289 & 4.6 & 3.5 + 2447977.403038 & 2240118 & 7.5 & 2.3 + 2447978.327055 & 2240489 & 4.3 & 3.3 + 2447979.358189 & 2240903 & 2.6 & 3.4 + 2447979.358145 & 2240903 & -1.2 & 4.9 + 2447978.601069 & 2240599 & 7.4 & 2.5 + 2447980.621017 & 2241410 & 5.8 & 3.4 + 2447980.782929 & 2241475 & 7.2 & 2.3 + 2447981.325918 & 2241693 & 8.4 & 1.4 + 2447981.592393 & 2241800 & 5.7 & 1.4 + 2447981.779185 & 2241875 & 4.8 & 1.1 + 2447982.329663 & 2242096 & 7.4 & 1.8 + 2447982.743093 & 2242262 & 5.0 & 1.2 + 2447983.734400 & 2242660 & 5.4 & 1.2 + 2447979.281057 & 2240872 & 9.5 & 2.9 + 2447980.224899 & 2241251 & -2.4 & 2.9 + 2447984.735678 & 2243062 & 6.5 & 1.1 + 2448245.724666 & 2347847 & -3.3 & 5.1 + 2448267.799932 & 2356710 & 5.2 & 2.3 + 2448324.627972 & 2379526 & 4.3 & 1.2 + 2448325.708938 & 2379960 & 4.1 & 1.3 + 2448328.593208 & 2381118 & 6.4 & 1.6 + 2448331.661735 & 2382350 & 4.0 & 1.2 + 2448238.571479 & 2344975 & 8.3 & 2.2 + 2448622.833258 & 2499253 & 3.3 & 1.8 + 2448680.642683 & 2522463 & 6.3 & 1.2 + 2448687.614155 & 2525262 & 4.0 & 1.2 + 2448688.597979 & 2525657 & 3.4 & 1.2 + 2449062.660365 & 2675840 & 4.2 & 1.6 + 2449063.609354 & 2676221 & 6.7 & 1.9 + 2449066.615640 & 2677428 & 6.5 & 1.4 + 2449066.371558 & 2677330 & 7.2 & 2.0 + 2449066.326737 & 2677312 & 8.2 & 2.6 + 2449069.342967 & 2678523 & 6.4 & 1.7 + 2449298.239287 & 2770423 & 8.5 & 4.1 + 2449298.304041 & 2770449 & 8.2 & 4.1 + 2449294.214264 & 2768807 & 5.5 & 4.1 + 2449294.293897 & 2768839 & -0.5 & 4.1 + 2449295.439583 & 2769299 & -4.0 & 6.1 + 2449295.494387 & 2769321 & -3.3 & 7.1 + 2449036.809260 & 2665461 & 2.4 & 2.2 + 2449038.677300 & 2666211 & 3.1 & 2.2 + 2449040.687310 & 2667018 & 3.6 & 4.1 + 2449041.616360 & 2667391 & 4.9 & 4.1 + 2449799.723888 & 2971765 & 5.6 & 1.3 + 2450427.920960 & 3223981 & 8.2 & 3.8 + 2450429.973242 & 3224805 & 2.7 & 2.4 + 2450430.914779 & 3225183 & 6.9 & 2.5 + 2450431.843821 & 3225556 & 7.5 & 1.5 + 2450434.912392 & 3226788 & 8.8 & 2.0 + 2450436.929828 & 3227598 & 5.4 & 1.7 + 2450483.633189 & 3246349 & 9.6 & 1.8 + 2451249.5989069 & 3553878&10.1&1.3 + 2451249.7632895 & 3553944&9.7 & 1.7 + 2451250.6126098 & 3554285&8.7&2.0 + 2451526.8772586 & 3665203 & 10.4 & 1.2 + 2451528.8523866 & 3665996&10.0&1.5 + 2451528.9196061 & 3666023&7.4&1.4 + 2451528.9868422 & 3666050&6.3&1.9 + 2451529.8585943 & 3666400 & 6.6&2.0 + 2451530.9097492 & 3666822 & 13.4&2.41 + 2451960.8561629 & 3839442 & 10.1&1.62 + 2451962.7864775 & 3840217 & 11.3&1.48 + 2451967.6806926 & 3842182 & 8.1 & 1.93 + 2451988.7919772 & 3850658 & 10.5 & 2.00 + 2451990.7845255 & 3851458 & 8.8 & 1.59 + 2452037.6472583 & 3870273 & 10.1 & 3.39 + 2452045.6399770 & 3873482 & 12.5&1.85 + 2452225.9050927 & 3945857 & 7.6 & 1.34 + 2452225.9598927 & 3945879 & 8.0 & 0.65 + 2452263.8834810 & 3961105 & 10.6&0.58 + 2452316.6442205 & 3980721 & 13.1&1.0 + 2452317.8995164 & 3982288 & 12.2&0.67 + 2452319.7999417 & 3982691 & 12.0&0.79 + 2452317.6479750 & 3982792 & 10.3&0.95 + 2452321.8348344 & 3983555 & 11.4&1.23 + 2452322.7265266 & 3984372 & 9.9 & 3.03 + 2452312.7412881 & 3984730 & 11.4&3.5 + 2452373.6840808 & 4089983 & 9.9&1.1 + 2452373.6839702 & 4090425 & 12.6&1.2 + 2452373.7140655 & 4090791 & 10.6&0.68 + 2452375.6392709 & 4122465 &
12.8&1.0 + 2452374.7700070 & 4124076 & 12.5&1.23 + 2452581.9494464 & 4134940 & 12.0 & 1.83 + 2452583.9095168 & 4137288 & 11.5 & 1.74 + 2452584.8812628 & 4144503 & 13.8 & 0.92 + 2452585.9821875 & 4148953 & 12.0 & 1.16 + 2452586.8937641 & 4146104 & 12.3 & 0.7 + 2452665.7845581 & 4255806 & 13.7 & 0.72 + 2452669.7970851 & 4256260 & 10.6 & 1.52 + 2452696.8561592 & 4257768 & 12.8 & 0.96 + 2452724.6624548 & 4267394 & 10.5 & 0.81 + 2453381.7442572 & 4409917 & 11.5 & 1.41 + 2453439.7653860 & 4433212 & 13.9 & 1.0 + 2453446.6073830 & 4435959 & 15.2 & 0.71 + 2453473.6191370 & 4446804 & 15.1 & 0.75 + | we report our measurement of the rate of change of period with time ( @xmath0 ) for the 215s periodicity in the pulsating white dwarf g 117b15a , the most stable optical clock known . after 31 years of observations
, we have finally obtained a 4@xmath1 measurement @xmath2 .
taking into account the proper - motion effect of @xmath3 , we obtain a rate of change of period with time of @xmath4 .
this value is consistent with the cooling rate in our white dwarf models only for cores of c or c / o . with the refinement of the models ,
the observed rate of period change can be used to accurately measure the ratio of c / o in the core of the white dwarf . |
decoherence embodies the detrimental effects of noise on any quantum system whose coherence , in its widest sense , is smeared causing the loss of information of the initial state @xcite .
this represents a focal point in quantum information theory @xcite as it limits both the attainable fidelity and the variety of accessible protocols @xcite .
in particular , entanglement @xcite represents a fundamental resource in quantum computation @xcite and thus it should be protected against dechorence . in this regard ,
the most undesirablefamily of quantum processes is given by the so - called entanglement breaking ( eb ) maps @xcite , under whose action any entanglement initially installed between the system and an external ancilla is completely lost .
these are maps acting on one component of an entangled pair leaving unperturbed the other . _ amendable channels _ strongly related to eb maps @xcite .
they realize an eb map when applied twice consecutively on the same system , but they admit a _ filtering _ operation that , applied in between the first and the second action of the map , prevents the global transformation from being entanglement breaking . identifying the set of amendable channels and their associated filtering operations is an important quantum error correction task which have profound implications in many research areas . in particular this could be useful in developing efficient long - range communication schemes based on quantum repeaters architectures @xcite where the signaling process takes place through intermediaries ( the quantum repeaters ) who collect , process , and redistribute the messages sent by the communicating parties ( in this picture the action of an amendable channel simulates the transferring from two communicating parties and one repeater , while the filtering operation corresponds to the data processing performed by the latter ) .
the study of amendable channels is particularly relevant in the context of the so called bosonic gaussian channels ( bgcs ) @xcite .
these are completely positive trace preserving maps @xcite , which provide prototypical examples of decoherence processes that occurs in continuous variable ( cv ) systems @xcite , e.g. in the transmission of optical signals through lossy dispersive optical fibers and/or in free - space @xcite .
examples of bgcs which are amendable were first discussed in ref @xcite . moving from those observations , in this paper we propose and discuss in details a feasible quantum optics experiment for the realization and the experimental test of an amendable map using gaussian channels . in particular , having at disposal a two - mode squeezed vacuum state @xcite generated by a type - ii sub threshold opo @xcite , we show that by suitable passive linear optical manipulations it is possible to realize an eb gaussian channel .
then we prove that it is possible to amend the eb channel in a simple way thus preserving the initial entanglement of the probe state . the proposed experimental set
up is an effective realization of the conceptual scheme discussed in sec .
iii a of ref .
@xcite .
the paper is structured as follows : in sec .
[ sect : review ] we present a brief review of the theory of gaussian amendable channels introducing some useful notation ( see [ subsect : notation ] ) and the conceptual theoretical scheme ( see [ subsect : teoscheme ] ) . in sec .
[ sect : sketch ] we give a glance over the experimental proposal . in particular we prove that ( see [ subsect : expi ] ) a proper manipulation of the output of a single type
ii sub threshold opo is sufficient for generating both an entangled probe state and a local squeezed ancilla .
then , we show that an effective eb channel can be obtained by using only passive optical elements ( see [ subsect : expii ] ) . in sec . [ sect : results ] , we estimate the correlations of the output state looking for suitable experimental conditions that would make eb the resulting map .
then , we find the parameters setting that makes the map effectively amendable . finally , ( see [ subsect : losses ] ) we analyze the feasibility of the experiment in presence of losses , measurement uncertainty and detection inefficiencies .
in this section we review some basic theoretical notions and discuss a simple example of gaussian amendable channel .
a more detailed analysis can be found in ref .
@xcite .
consider @xmath0 optical radiation modes described by their position and momentum quadrature operators @xmath1 , @xmath2 , @xmath3 , @xmath4 , and @xmath5 , @xmath6 , @xmath3 , @xmath7 which we group in a vector of @xmath8 components : @xmath9 . such operators can be chosen to be dimensionless so that they obey the canonical commutation rules @xmath10=i\delta _ { i , j}$ ] , @xmath11=[p_{i},p_{j}]=0 $ ] . to any state @xmath12 of the system we can associate its first and second statistical moments defined respectively by the real vector @xmath13 and by the @xmath14 covariance matrix ( cm ) @xmath15 with entries @xmath16where the symbol @xmath17 indicates expectation values with respect to @xmath12 .
gaussian density matrices are fully characterized once @xmath13 and @xmath15 are assigned @xcite .
they correspond to states of the @xmath0mode system whose associated characteristic function is gaussian .
examples of gaussian states which will play an important role in the next section are the following pure states : single mode vacuum state , single mode squeezed vacuum state and two - mode squeezed vacuum state ( tmsv ) . according to this notation
the vacuum state is characterized by @xmath18 and @xmath19 the squeezed state by @xmath18 and @xmath20while the tmsv state has @xmath21 and @xmath22gaussian channels are quantum operations which map gaussian states into gaussian states @xcite .
therefore , they are completely defined by their action on the displacement vector @xmath23 and the matrix @xmath15 . moreover , since the level of entanglement of a state depends only on the correlations and it is insensitive to displacement operations , the action on @xmath13 can be completely neglected for the purpose of the present paper .
in particular in the following we will make extensive use of the transformation associated to a beam splitter of transmissivity @xmath24 . given two input modes with cm
@xmath15 it will produce at the output a two - mode state with cm @xmath25 , where @xmath26if we mix a single mode state with the vacuum on a beam splitter and we trace out one of the output modes we are left with a non - unitary attenuation ( or lossy ) channel @xmath27 @xcite acting on the cms as @xmath28where @xmath29 is the cm of the vacuum given in eq .
( [ vacuum ] ) .
another important single mode operation we will use in the following is the single mode squeezing acting as @xmath30 with @xmath31the previous states , operations and combinations thereof are the main ingredients of the scheme which will be presented in the following .
finally we stress that in a real experiment the cm of , at most , a two mode state can be fully reconstructed by a single homodyne detection scheme @xcite .
our goal is identifying an experimentally feasible scheme for realizing the paradigms of gaussian amendable channels .
as recalled in the introduction a channel @xmath32 is amendable if it is entanglement breaking of order @xmath33 , i.e. @xmath34 and there exists a unitary filter such that @xmath35 this problem is equivalent to the following one : find a channel @xmath36 and a unitary @xmath37 such that @xmath38 while @xmath39 indeed if eq.s ( [ amend3 ] ) and ( [ amend4 ] ) hold , it is straightforward to check that @xmath40 and @xmath41 satisfy eq.s ( [ amend1 ] ) and ( [ amend2 ] ) in view of the invariance of entanglement under local unitaries .
it turns out that , for gaussian channels , the second problem is simpler to address and therefore in this paper we focus on the latter pair of conditions , eq.s ( [ amend3],[amend4 ] ) . in ref .
@xcite , it was shown that an attenuation channel ( eq . ( [ attenuation ] ) ) and a local squeezing operation ( eq . ( [ squeezing ] ) ) are valid examples of @xmath42 and @xmath43 respectively .
indeed one has that , for some values of the channel transmissivity @xmath24 and squeezing parameter @xmath44 @xmath45 while @xmath46 a natural way to verify that @xmath47 is @xmath48 while @xmath49 is not would be to apply those maps to one part of a maximally entangled state and check whether the initial entanglement is preserved or not . in continuous variables systems , however maximally entangled states are not physically realizable but , as proven in ref .
@xcite , the test can be performed by using a pure two - mode squeezed state with finite entanglement and mean energy .
however , we note here that if the incoming state is mixed , e.g. due to the presence of losses in the state preparation stage , this equivalence property is not valid any more and the output state may result in being separable even if the map is not eb .
it goes without saying that even in this case it may happen that a unitary filter , acting at a proper stage , will restore the lost entanglement . the theoretical model of this test
is graphically depicted in fig .
fig : simplifiedsetups , where we sketch the scheme of @xmath50 and @xmath51 . in both cases the map
is applied to one side of a tmsv state with squeezing parameter @xmath52 ( see eq .
( [ tmsv ] ) ) . after the application of @xmath50
it has been proved ( see fig .
4a of ref .
@xcite ) that there exists a range of @xmath24 for which the output state is separable . on the other hand , when @xmath51 is applied the output state is always entangled , for all values of @xmath52 and @xmath24 ( see fig .
4b of ref .
the equivalence between eq.s ( [ amend3],[amend4 ] ) and eq.s ( [ amend1],[amend2 ] ) implies that the gaussian map @xmath53 is amendable via a squeezing unitary filter @xmath54
. this theoretical scheme will be our starting point for designing a more realistic experimental proposal .
in this section we propose an experimental set up in order to realize an effective amendable gaussian map , endowed with the appealing property of being quite simple to be realized in the laboratory .
furthermore , we will also take into account the effects of losses , detection inefficiency and , eventually , measurement indeterminacy .
our starting point is the theoretical model given in fig .
fig : simplifiedsetups .
notice that , even though appearing quite simple , the first set up associated with the channel @xmath47 in principle requires , in addition to the state preparation part , a non - trivial active operation on the system : the local squeezing between the two beam splitters . in quantum optics active operations can be realized by non - linear interactions . in most squeezing schemes , the initial state ( e.g. vacuum , coherent , squeezed and/or thermal ) interacts with a strong classical field in an optical nonlinear medium .
this can be achieved for example by a sub threshold optical parametric oscillator ( opo ) @xcite .
compared with passive transformations , active operations are relatively difficult to be engineered and experimentally costly . in order to get around this obstacle , our idea is to use a single initial active operation both for the generation of the entangled ( probe ) state and for the ( indirect ) realization of the local squeezing .
these resources will be obtained in the preparation stage described in fig .
[ figurepartscheme ] . ) .
the entangled modes @xmath55 and @xmath56 are orthogonally polarized . by applying a @xmath57 wave plate and a polarizing beam splitter ( pbs ) to @xmath55 and @xmath56 ,
we can obtain two _ independent _ single mode squeezed vacuum states @xmath58 and @xmath59 with orthogonal squeezing phases ( eq . ( [ resource1 ] ) ) . then , through a balanced beam splitter ( the red one ) , polarization insensitive , we mix mode @xmath59 with the vacuum state @xmath60 obtaining the pair @xmath61 , a pure tmsv ( eq . ( [ eq : vc1c2 ] ) ) .
this pair represents the entangled system on which we will test the entanglement breaking properties of the maps @xmath62 and @xmath49 defined in eq.s ( [ phi1],[phi2 ] ) .
the pure single mode beam @xmath58 will be used for implementing the squeezing transformation @xmath63 . ] at the output of a type ii opo one has at disposal two cross - polarized but frequency degenerate entangled modes @xcite , say @xmath55 and @xmath56 . as shown in fig .
[ figurepartscheme ] , by means of a @xmath64 wave plate and a polarizing beam splitter ( pbs ) it is possible to manipulate the entangled state in order to obtain two _ independent _ single - mode squeezed vacuum states @xcite @xmath58 and @xmath59 , with orthogonal squeezing phases . according to the notation introduced in section subsect : notation , the correlation matrix of the latter modes can be written as @xmath65then , combining the mode @xmath59 with the vacuum @xmath60 by means of a balanced polarization insensitive beam splitter ( the red plate in fig .
figurepartscheme with @xmath66 ) , we generate the pair of modes @xmath67 with correlation matrix @xmath68b\left ( \frac{1}{2}\right ) \\
& = & [ s(-\frac{r}{4})\oplus s((-\frac{r}{4})]v_{2}(-\frac{r}{2})[s((-\frac{r}{4})\oplus s((-\frac{r}{4})]\ , .
\notag\end{aligned}\ ] ] notice that , up to local ( single - mode ) operations , @xmath69 and @xmath70 are in a tmsv state with squeezing parameter @xmath71 , _ i.e. _ half of the original two mode squeezing characterizing the pair @xmath55 and @xmath56 at the opo output . at the same time , we will have at disposal an auxiliary single mode squeezed vacuum @xmath58 that , in a certain sense , carries the second half of the original squeezing . summarizing , at the output of the above described generation stage we have at disposal three optical modes : the entangled pair @xmath67 , which will play the role of the probe state for testing the entanglement breaking properties of the gaussian maps @xmath47 and @xmath72 , and the squeezed mode @xmath58 which will be used as a resource for mimicking a single - mode squeezer .
we expect that suitably setting the opo squeezing @xmath44 and the beam splitters transmissivity @xmath24 , we can find that the final state , of the pair @xmath73 , is separable under the action of @xmath74 and entangled for @xmath75 . in this section
, we will show how to implement the maps @xmath47 and @xmath49 defined in eq.s ( [ phi1],[phi2 ] ) , and graphically represented in fig .
[ fig : simplifiedsetups ] .
let us first consider @xmath47 .
as recalled in section subsect : notation , each attenuation map @xmath27 can be directly implemented by letting the incoming mode pass through a beam splitter of transmissivity @xmath76 .
less trivial is the passive implementation of the local squeezing @xmath63 without using another opo .
this can be indirectly achieved mixing the auxiliary squeezed mode @xmath58 with @xmath69 onto a beam splitter of transmissivity @xmath24 . as a matter of fact , by observing that @xmath77b(\eta ) = b(\eta ) [ s(r)\oplus s(r)]\ , , \label{commutation}\ ] ] it derives that combining an incoming mode with a single mode squeezed vacuum on a beam splitter is equivalent to indirectly attenuating and then squeezing the system , as graphically shown in fig .
[ fig : property ] . and a more complex operation consisting in the sequence @xmath78 . in both case
the _ idler _ mode is traced out .
this shows how a squeezed ancilla mode can be used to effectively realize a squeezing operation on a given input state . ] more precisely , we have that the effect of the first optical circuit of fig .
fig : property , tracing out the idler mode , is equivalent to the sequence @xmath79 ) ) followed by the implementation of @xmath47 and @xmath49 by means of two bss with transmissivity @xmath80 ( lower right corner of the picture ) .
three fictitious bss , see sec .
iv a , mimic the effects of losses ( bs with transmissivity @xmath81 ) and detection inefficiencies ( bss with transmissivity @xmath82 ) . ] by applying this equivalence property , we can easily implement the action of @xmath47 on the incoming mode @xmath69 , as pictorially represented in the full experimental set up given in fig .
[ figurecomplscheme ] . here
the green beam splitters act on the incoming mode @xmath69 as @xmath83 thus experimentally realizing the map of eq.([phi1 ] ) up to the unitary transformation @xmath84 .
we recall that the entanglement - breaking properties of a map , are invariant under unitary redefinition of the input and output spaces @xcite , that is @xmath85 iff @xmath86 . on the other hand , the implementation of the channel @xmath51 defined in eq .
( [ phi2 ] ) can be straightforwardly implemented by discarding the auxiliary mode @xmath87 and substituting it with the vacuum . in other words
, one should simply let the mode @xmath88 pass through the green beam splitters without feeding any light in the empty ports .
we can therefore conclude that the experimental setup represented in fig .
[ figurecomplscheme ] is , up to experimental losses , equivalent to the theoretical scheme in fig .
[ fig : simplifiedsetups ] . for
the sake of clearness , let us point out that while in the theoretical scheme of fig .
[ fig : simplifiedsetups ] the squeezing parameters @xmath89 and @xmath44 are totally independent , in the realistic setup of fig .
[ figurecomplscheme ] the structure of the scheme forces @xmath71 .
nonetheless , this lack of freedom does not affect the feasibility of the experiment .
as explained in section [ subsect : teoscheme ] ( see eq.s ( [ amend3]-phi2 ) ) , in order to experimentally prove the existence of gaussian amendable channels we need to show that @xmath47 is entanglement - breaking while @xmath49 is not .
this can be done by measuring the output state of our experimental circuit and checking its separability for different choices of the experimental parameters .
in particular we will apply the ppt criterion @xcite to the cm of the output state . here
we discuss the proposed experimental scheme and we theoretically estimate @xmath90 , the expected cm for the final state . by writing it in @xmath91 blocks @xmath92one can easily compute the minimum symplectic eigenvalue @xmath93 of the partially transposed state @xmath94}}{2 } } \label{eq : nu}\]]where @xmath95+\det [ \mathbf{b}]-2\det [ \mathbf{c}]$ ] . from the ppt criterion
it can be shown that the output state is entangled if and only if @xmath96(see @xcite and references therein ) .
the relation above , provides a necessary and sufficient criterion for testing the separability of the output state and thus for studying the entanglement - breaking properties of the applied map .
the behaviour of @xmath97 for initial squeezing @xmath98 is given in fig .
[ fig : nuquadro2 ] , that refers to the case of ideal ( lossless ) preparation stage and detectors with unit efficiency .
it results that , if on the one hand @xmath72 can never become @xmath99 for any value of the transmissivity @xmath24 ( _ i.e. _ @xmath100 is always lower than @xmath101 so that the state keeps its entanglement ) , on the other hand there exists a finite interval of @xmath24 such that @xmath102 and thus its output state is separable .
this separability interval has been computed in @xcite and corresponds to @xmath103 with @xmath104 in the next subsection , we will consider the effects of losses , detection inefficiencies and measurement uncertainties . in a realistic implementation
we can not neglect the statistical uncertainty affecting the measurement process and , at the same time , we also have to consider the effects of losses ( decoherence ) and detection efficiency . in order to take into account the experimental indeterminacy into eq .
( [ eq : nu ] ) we have considered typical experimental values for the uncertainties relative to the cm elements .
these values are used in propagating the measurement s errors into the formula that gives @xmath100 in terms of cm elements ( a detailed discussion on the errors affecting the different elements can be found in ref .
thus we obtain the statistical error @xmath105 for @xmath100 . from the experimental point of view claiming that @xmath106 requires that @xmath107 , _ _
i.e.__the distance from the separability threshold must overcome the measurement confidence interval ( _ i.e. _ twice the uncertainty ) . in fig .
[ fig : err05 ] we have fixed @xmath108 and plotted @xmath109 as a function of @xmath24 .
the dashed lines represent the boundaries of the confidence interval for @xmath100 . from this plot
we conclude that in this case the confidence interval @xmath110 would make ambiguous , from the experimental point of view , the statement that @xmath102 .
this ambiguity can be overcome by considering an increased level of the squeezing for the pure state generated by the type ii opo .
for example , it is sufficient to raise @xmath111 from @xmath112 to @xmath113 to obtained a clear experimental proof that @xmath102 for @xmath114 , as shown in fig .
[ fig : err065 ] . here
@xmath115 is plotted for @xmath116 , and @xmath117 is greater than the expected confidence interval in a range of values for @xmath24 contained in @xmath118 $ ] .
furthermore , real experiments face the effects of absorption losses and non ideal detection , which can be modeled by the three fictitious beam splitters we have introduced in fig .
[ figurecomplscheme ] .
the first one of transmissivity @xmath81 ( on the left ) simulates the effects of losses and in particular , the opo cavity escape efficiency @xcite that unavoidably makes any state at the output of an opo cavity a mixed one @xcite .
the last two beam splitters of transmissivity @xmath82 ( on the right ) model the inefficiency of the detectors .
it is interesting to see that the conclusions retrieved from the analysis performed in fig .
[ fig : err065 ] are still valid if losses and detection inefficiencies are taken into account . in fig .
[ fig : real065 ] we plot the behavior of @xmath119 in a realistic scenario , setting the losses at @xmath120 so that @xmath121 and detection efficiency at @xmath122 , and assuming the same statistical indeterminacy used in the case without losses .
the effect of @xmath123 and @xmath124 is , on one hand , to reduce the maximum value for @xmath100 inside the eb region ( the maximum also moves to a higher @xmath24 s value ) , on the other hand , to enlarge the @xmath76 interval for which @xmath125 .
we note that while reducing the weight of losses and detection inefficiencies is surely possible ( @xmath126 and @xmath127 have been recently reported schnabel2013 ) experimental indeterminacy can not be avoided and , as far as we know , the value used in ref .
@xcite is the lowest one for the experimental determination of the cm of a bipartite gaussian state .
in this work we have proposed a realistic quantum optics experiment based on continuous variable systems that would provide the existence of gaussian amendable maps and give more insight on entanglement breaking channels from a practical point of view .
the proposed scheme is translated into a rather simple experimental set up .
indeed , it is based on a single initial non - linear operation ( realized by a type ii sub threshold opo ) which has the role of preparing both the input entangled state and a squeezed ancilla .
the rest of the scheme is extremely simple since it requires only passive operations such as beam splitters and wave - plates .
the proposal has been realistically analyzed by taking into account the typical statistical uncertainty of gaussian state quantum homodyne tomography .
the effects of losses and detector inefficiencies have also been considered .
we have shown that , even in presence of such errors and losses , the experiment is still feasible .
indeed , a conclusive test can be achieved by appropriately tuning the experimental parameters .
the proposed scheme can be readily implemented in any laboratory having at disposal a running source of bipartite gaussian entangled states .
j. eisert and m. m. wolf , `` gaussian quantum channels '' _ in quantum information with continous variables of atoms and light _ , n.j .
cerf , g. leuchs , and e.s .
polzik eds .
( imperial college press , london , 2007 ) , pp .
2342 ; v. dauria , s. fornaro , a. porzio , s. solimeno , s. olivares , and m. g. a. paris , `` full characterization of gaussian bipartite entangled states by a single homodyne detector '' , phys .
* 102 * , 020502 ( 2009 ) . v. dauria , s. fornaro , a. porzio , e. a. sete , and s. solimeno , `` fine tuning of a triply resonant opo for generating frequency degenerate cv entangled beams at low pump powers '' , appl
b * 91 * , 309314 ( 2008 ) . v. dauria , a. porzio , s. solimeno , s. olivares , and m. g. a. paris , `` characterization of bipartite states using a single homodyne detector '' , j. opt .
b : quantum semiclass . opt .
* 7 * , s750s753 ( 2005 ) ; | we propose a quantum optics experiment where a single two - mode gaussian entangled state is used for realizing the paradigm of an amendable gaussian channel recently presented in phys .
rev . a , * 87 * , 062307 ( 2013 ) . depending on the choice of the experimental parameters the entanglement of the probe state
is preserved or not and the relative map belongs or not to the class of entanglement breaking channels .
the scheme has been optimized to be as simple as possible : it requires only a single active non - linear operation followed by four passive beam - splitters .
the effects of losses , detection inefficiencies and statistical errors are also taken into account , proving the feasibility of the experiment with current realistic resources . |
the low - mass x - ray binary ( lmxb ) exo 0748676 was discovered with _
exosat _ by parmar et al .
it showed periodic eclipses , irregular intensity dips , and type i x - ray bursts ( parmar et al . 1986 ) .
the eclipses , due to obscuration of the central x - ray source by the companion star , occurred at a period of 3.82 hr . from the eclipse duration
an inclination of 75@xmath0 to 82@xmath0 was derived .
dipping activity is seen at orbital phases @xmath1 and at @xmath2 ( @xmath3 corresponds to eclipse center ) .
the dips can be as deep as 100% ( church et al . 1998 ) .
it is believed ( white & swank 1982 ; walter et al . 1982 ) that such dips are caused by obscuration of the central x - ray source by a bulge in the outer accretion disk that is created by the impact of the gas stream from the companion star , or by the accretion stream itself if it penetrates the disk further in ( frank , king , & lasota 1987 ) . during the eclipses 4% of the 26 kev intensity remained , which parmar et al .
( 1986 ) attributed to the presence of an accretion disk corona ( adc ) . using
their ` progressive covering ' model church et al .
( 1998 ) found that @xmath470% of the out - of - dip flux is contributed by this adc , which they estimated to have a radius of 0.41.5@xmath510@xmath6 cm .
recently , a persistent 1 hz quasi - periodic oscillation ( qpo ) was found with rxte in the high inclination ( @xmath7 ) dipping lmxb 4u 132362 ( jonker , van der klis , & wijnands 1999 , hereafter jo99 ) .
jo99 concluded that its presence is most likely related to the high inclination of the system . in this letter , we report the discovery of a variable 0.582.44 hz qpo in exo 0748676 with properties similar to those found in 4u 132362 .
we made use of data obtained with the proportional counter array ( pca ; jahoda et al .
1996 ) on board the _ rossi x - ray timing explorer _ ( _ rxte _ ; bradt , rothschild , & swank 1993 ) . a log of all observations is given in table [ table1 ] .
observations 2 to 7 were originally performed to study the orbital evolution of the system .
each consisted of five @xmath42 ks data intervals that range in orbital phase from @xmath8 to @xmath9 .
all observations yielded data in the standard 1 and 2 modes , which have 1/8 s time resolution in 1 energy channel ( 260 kev ) and 16 s time resolution in 129 channels ( 260 kev ) , respectively . in addition , data were obtained in modes with at least 2@xmath10 s time resolution , in 67 ( obs .
1 ) , 32 ( obs .
2 ) , 256 ( obs .
27 ) , or 64 ( obs .
810 ) channels , covering the 260 kev range .
data during eclipses were removed from further analysis .
x - ray bursts were studied separately .
power spectra were created in the energy bands 260 , 25 , 58 , 813 , and 1342 kev .
the average 1/16128 hz power spectra of each observation were rms normalized , and fitted with a model consisting of a constant ( representing the poisson level ) , a power law , @xmath11 ( the noise component ) , and a lorentzian ( the qpo ) .
errors on the parameters were determined using @xmath12 .
the dependence on photon energy of the qpo strength was determined by fitting the power spectra in four energy bands , while keeping the qpo fwhm and frequency , and the power law index ( @xmath13 ) of the noise component fixed to the values obtained in the 260 kev band . unless stated otherwise the fit parameters are those in the 260 kev band .
the 95% confidence upper limits on the presence of the qpo and/or noise component were determined by fixing the fwhm and frequency , and/or by fixing @xmath13 . for observations 26 and 810 we compared power spectra obtained in- and outside the dips .
a count rate level was determined for each single orbit , below which data were assumed to be in a dip . for observations 810
this level was the persistent count rate during phases were normally no dipping is observed .
as mentioned before , observations 26 were taken during phases that normally show dipping , which is confirmed by increases in hardness .
for these observations we took the post - eclipse count rate as the discriminator , since dipping usually occurs less after the eclipse .
figure [ qpo ] shows two typical power spectra of exo 0748676 .
a qpo peak was found in the power spectra of all observations except the first one .
the qpo covered a frequency range of more than a factor 4 , between [email protected] hz ( obs . 10 ) and [email protected] hz ( obs .
2 ) , with rms amplitudes between 8.4%@xmath140.4% and 12.1%@xmath140.6% . within each observation ( @xmath41 day )
the qpo was only observed in a small frequency range , never changing more than 15% in frequency .
table [ table1 ] gives the fit parameters of all observations in the 260 kev energy band . in observation 1 , where the qpo was not detected ,
i.e. , at least a factor 4 weaker than otherwise , the background corrected count rate was a factor 2 to 3 higher than in the other observations . in the other observations ,
neither the qpo nor the noise component show any correlation between their parameters and the 260 kev count rate , which varies by @xmath420% .
however observations 8 , 9 , and 10 , for which ( due to the more extended phase coverage ) the persistent count rate could be estimated much better than for the other observations , show a clear anti - correlation between frequency and count rate .
the qpo peak becomes broader when its frequency increases ( see figure [ ratio ] ) .
the ratio @xmath15 , of frequency over fwhm of the qpo , is consistent with being constant ; a linear fit gives a @xmath15 of [email protected] , with @xmath16/d.o.f.=12.2/8 .
the noise component tends to become steeper as it gets stronger ( see figure [ noise ] ) .
the rms of the qpo decreases from @xmath412% to @xmath48.5% when the rms of the noise component increases from @xmath46% to @xmath410.5% .
although there is considerable scatter around a linear relation , there seems to be a connection between the two components ( correlation coefficient is .75 ) .
figure [ energy ] shows the photon energy spectra of the qpo in observations 2 to 10 ( dip and non - dip data combined ) .
it is obvious from this figure that they are quite flat .
the energy spectra of the noise component are similar .
all qpo parameters in- and outside the dips are consistent with each other , except for the qpo amplitude in observation 10 , which during the dips is 12.0%@xmath140.8% rms and outside the dips 7.5%@xmath140.4% rms .
the noise shows only a significant difference in observation 4 , where the dip rms is [email protected]% and the non - dip rms is [email protected]% .
four type i x - ray bursts were observed , all outside dips ; one in observations 5 and 8 each , and two in observation 9 ( hereafter burst 1 to 4 ) . in 16 s resolution data they had peak count rates ( 260 kev ,
3 detectors , background corrected ) of 2330@xmath1410 , 2240@xmath1410 , 2410@xmath1410 , and 2110@xmath1410 counts s@xmath17 , respectively , and they all lasted @xmath4400 s. in all four bursts the qpo was detected .
the qpo during burst 2 is the only that has width and frequency that are consistent with those of the qpo outside the burst .
the qpo frequency during burst 1 is 0.1 hz higher than that of the qpo outside the burst , and the fwhm is [email protected] hz , which is considerably smaller than 0.36 hz .
the qpos during bursts 3 and 4 are both poorly fitted with a lorentzian .
both peaks appear to have shifted in frequency during the bursts .
the peaked shape in the power spectrum of burst 4 clearly consists of two small peaks , which , when fitted with lorentzians , have frequencies of [email protected] hz and [email protected] hz and a fwhm of @xmath40.1 hz .
the qpo in burst 3 is very asymmetric .
the best fit with one lorentzian at [email protected] hz and a fwhm of @xmath40.1 hz , shows excess power at higher frequencies .
dynamical power spectra of burst 3 and 4 show that the qpo profile changed with time .
the qpos during the four bursts have rms amplitudes of 9.7%@xmath140.9% , 8.4%@xmath18% , at least 7% ( probably 8%10% ) , and @xmath410% ( the sum of the rms amplitudes of the two lorentzians ) , respectively , consistent with those found outside the bursts .
we have discovered a qpo in exo 0748676 with a frequency varying between 0.58 and 2.44 hz , and with rms amplitudes between 8.4% and 12.1% .
the qpo was detected throughout the persistent emission , the dips , and the type i x - ray bursts , with strengths that were consistent with each other , except in observation 10 , where the qpo was stronger during the dips than in the persistent emission . over the whole frequency range
@xmath15 was consistent with 3.5 , except in the x - ray bursts where it was higher .
the qpo was not detected in observation 1 , when the count rate was a factor 2 to 3 higher than otherwise .
the power spectral features show no changes that are correlated with variations of @xmath420% in the out - of - dip count rate , except for observations 8 to 10 , in which the frequency is anti - correlated with the out - of - dip count rate .
in addition to the qpo , a noise component at frequencies below 1 hz was present , with a typical strength of 6%10% rms . both the qpo and the noise component have a flat photon energy spectrum .
several of the properties of the qpo , notably its frequency , its relatively unchanged persistence during bursts and dips , and its flat energy spectrum , are remarkably similar to those of the @xmath41 hz qpo that was recently found by jo99 in 4u 132362 .
figure [ ratio ] shows that the ratio of frequency to fwhm of the qpo in 4u 132362 is also consistent with 3.5 , and figure [ noise ] shows that the noise component in 4u 132362 falls on the strength versus slope relation found for the noise component in exo 0748676 .
we conclude that both qpos are produced by the same mechanism .
since qpos like these have not been seen in other neutron star lmxbs the origin is most likely related to the high inclination of the sources ( jo99 ) .
it is unlikely that the qpos are produced by variations in mass accretion rate ( @xmath19 ) onto the neutron star : in that case they should also be visible in low - inclination sources .
moreover , other qpos that are thought to be produced by @xmath19 variations , show considerable energy dependence ( van der klis 1995 ) .
a medium modulating the radiation from a central source ( jo99 ) seems to be a more promising explanation ( see below ) .
we have found several new features of the qpo , that were not seen in 4u 132362 .
first , the large flux range in the observations of exo 0748676 allow us to say that the qpo is not present at high persistent count rates . in 4u 132362 ,
whose flux varied by only @xmath415% , the qpo was always present .
second , the properties of the qpo changed during three of the four bursts . in 4u 132362
the properties of the qpo remained the same during all bursts .
third , there is the difference in frequency range over which the qpos are observed : a factor @xmath44 in exo 0748676 and @xmath412% in 4u 132362 .
this could be due to the difference in the time span over which the sources were observed .
4u 132362 was observed for two consecutive days , whereas the exo 0748676 observations span 19 months .
the @xmath412% range in two days , is similar to the @xmath2015% range found for individual observations ( @xmath41 day ) of exo 0748676 .
finally , the noise has a flat photon spectrum , which suggests that the mechanism behind the noise is similar to that for the qpo .
much can be learned about the qpo from its behavior during the x - ray bursts .
the fact that the qpo is detected during all four bursts shows that the presence of the qpo does not directly depend on the instantaneous flux , and that the absence of the qpo in observation 1 is due to something else than just a higher count rate .
since the fractional rms amplitudes during the bursts are similar to those outside the bursts , variations in @xmath19 onto to neutron star can be ruled out as the origin of the qpo . a dramatic decrease in rms amplitude
is expected in that case .
the high count rates during the x - ray bursts enables us to detect the qpo in smaller time intervals , typically tens of seconds , than otherwise possible . in all bursts
the @xmath15 of the qpos is higher than 3.5 , and during three of the bursts we see the frequency shift .
this seems to indicate that the qpo is intrinsically narrower than measured by us .
the observed width of the qpo outside the bursts is then due to changes of the qpo frequency on time scales of a few hundred seconds , and/or due to the presence of several qpos with a higher @xmath15 that occur around a central frequency .
we now consider models that involve modulation of radiation from a central source . given the inclination of exo 0748 - 676 ( 75@xmath082@xmath0 ) , to be in the line of sight to the central source , a medium modulating the radiation from the central source must at least reach a height that is between 14% and 27% of its radial distance to the central source . de jong , van paradijs , & augusteijn ( 1996 ) found that accretion discs in lmxbs have opening semi - angles of @xmath412@xmath0 as seen from the central source .
hence only a small height above the accretion disc is required , typically a few percent of the radial distance to the neutron star , to be in the line of sight .
this suggests that the modulating medium is in or on the disc .
any model is constrained by the fact that the rms amplitude stays constant when the source goes into a dip ; the medium causing the dips has to remove more or less the same amount of modulated as unmodulated radiation from our line of sight .
models can be constructed , with varying geometries that depend on whether the central source is point - like or extended , and whether the media causing the dips and qpo are absorbing or scattering media . only a few combinations can be ruled out with certainty , e.g. most models involving a point - like source .
if the dips are being caused by a partial covering opaque medium , the central source has to be extended , and the medium causing the qpo should have an azimuthal extent comparable to that of the medium causing the dips , in order to explain the constancy of the rms amplitudes in- and outside the dips .
the mechanism also has to produce a central frequency that does not vary by more than 15% over the duration of a single observation ( typically one day ) . both partial covering by an opaque medium ( see jo99 ) and
thomson scattering can explain the flat photon spectrum of the qpo .
assuming the qpo frequency is the kepler frequency of a structure orbiting a 1.4 m@xmath21 neutron star , one derives radial distances of 1.02.4@xmath22 cm .
the viscous timescale ( frank , king , & raine 1992 ) at these radii , on which the orbiting structure is expected to alter , is comparable to the timescales that are seen for the frequency shifts during the burst .
the larger frequency changes between observations might be related to the timescale on which the accretion disc as a whole changes its structure ( typically days to weeks ) .
we remark that the inferred radii are similar to those in the @xmath13-disc model ( frank et al.1992 ) , at which the dominant opacity source changes from kramers opacity to electron scattering .
this radius increases with @xmath19 , and implies an anti - correlation between frequency and count rate , as seen in observation 810 .
the fact that the qpo is not found in observation 1 suggests that the higher count rate is accompanied by a change in the accretion disc structure
. both are probably due to an increase in @xmath19 . assuming that an accretion disc thickens with @xmath19 , the modulated flux might be obscured by a thicker disc , but the mechanism may also have disappeared altogether .
finally we remark that the 0.15 s jitter in the mid - eclipse timings reported by hertz et al .
( 1997 ) might be due the presence of the qpo reported here . a qpo with a frequency of 1 hz and an rms amplitude of 10% , arbitrarily superimposed on eclipse transitions lasting a few seconds
, can be expected to cause a jitter in the order of 0.1 s. bradt , h.v . ,
rothschild , r.e . ,
swank , j.h .
, 1993 , a&as , 97 , 355 church , m.j .
, bauciska - church , m. , dotani , t. , asai , k. , 1998 , , 504 , 516 de jong , j.a .
, van paradijs , j. , augusteijn , t. , 1996 , , 314 , 484 frank , j. , king , a.r . , lasota , j .-
, 1987 , , 178 , 137 frank , j. , king , a. , raine , d. , 1992 , accretion power in astrophysics ( cambridge ; cambridge univ .
press ) hertz , p. , wood , k.s .
, cominsky .
, l.r . , 1997 , , 486 , 1000 jahoda , k. , swank , j.h . , giles , a.b . ,
stark , m.j .
, strohmayer , t. , zhang , w. , morgan , e.h . , 1996 ,
spie , 2808 , 59 jonker , p.g . , van der klis , m. , & wijnands , r. , 1999 , , in press ( astro - ph/9811404 ) parmar , a.n . , white , n.e . ,
giommi , p. , & haberl , f , 1985 , , 4039 parmar , a.n . , white , n.e . ,
giommi , p. , & gottwald , m. , 1986 , , 308 , 199 van der klis , m. , 1995 , in x - ray binaries , ed . w.h.g .
lewin , j. van paradijs , e.p.j .
van den heuvel ( cambridge : cambridge univ .
press ) walter , f.m . ,
bowyer , s. , mason , k.o . ,
clarke , j.t .
, henry , j.p . , halpern , j. , grindlay , j.e .
, 1982 , , 253 , l67 white , n.e . ,
swank , j.h .
, 1982 , , 253 , l61 ccccccccccc 1 & 12 - 03 - 96 10:41 & 12 - 03 - 96 12:54 & 6.0 ( 0.9 ) & @xmath23 & & & & [email protected] & [email protected] & [email protected] + 2 & 03 - 05 - 96 08:37 & 04 - 05 - 96 00:30 & 12.0 ( 2.9 ) & 12.1@xmath24 & [email protected] & [email protected] & & [email protected] & [email protected] & [email protected] + 3 & 18 - 01 - 97 17:14 & 19 - 01 - 97 13:18 & 9.6 ( 3.7 ) & [email protected] & [email protected] & [email protected] & & [email protected] & [email protected] & [email protected] + 4 & 07 - 03 - 97 20:03 & 08 - 03 - 97 12:40 & 9.8 ( 5.6 ) & [email protected] & [email protected] & [email protected] & & [email protected] & [email protected] & [email protected] + 5 & 01 - 05 - 97 11:52 & 02 - 05 - 97 07:25 & 9.5 ( 3.3 ) & 9.3@xmath25 & [email protected] & [email protected] & & [email protected] & [email protected] & [email protected] + & burst 1 & & & [email protected] & [email protected] & [email protected] & & & & + 6 & 26 - 06 - 97 06:01 & 26 - 06 - 97 22:03 & 9.7 ( 2.7 ) & 9.8@xmath25 & [email protected] & [email protected] & & [email protected] & [email protected] & [email protected] + 7 & 07 - 10 - 97 00:51 & 07 - 10 - 97 20:32 & 8.6 & [email protected] & [email protected] & [email protected] & & 8.5@xmath26 & [email protected] & [email protected] + 8 & 13 - 08 - 97 19:38 & 14 - 08 - 97 03:02 & 17.3 ( 1.7 ) & [email protected] & [email protected] & [email protected] & & [email protected] & [email protected] & [email protected] + & burst 2 & & & [email protected] & [email protected] & [email protected] & & & & + 9 & 17 - 08 - 97 12:41 & 17 - 08 - 97 18:59 & 36.7 ( 6.9 ) & [email protected] & [email protected] & [email protected] & & [email protected] & [email protected] & [email protected] + & burst 3 & & & [email protected] & [email protected] & [email protected] & & & & + & burst 4 & & & [email protected] & [email protected] & [email protected] & & & & + & & & & 6.7@xmath27 & 0.13@xmath28 & 0.65@xmath29 & & & & + 10 & 22 - 08 - 97 00:13 & 22 - 08 - 97 18:56 & 24.5 ( 10.1 ) & [email protected] & [email protected] & [email protected] & & [email protected] & [email protected] & [email protected] + | we report the discovery of a quasi - periodic oscillation ( qpo ) in data obtained with the rossi x - ray timing explorer of the dipping and eclipsing low - mass x - ray binary exo 0748676 .
the qpo had a frequency between 0.58 and 2.44 hz changing on time scales of a few days , an rms amplitude between 8% and 12% , and was detected in the persistent emission , during dips and during type i x - ray bursts . during one observation , when the count rate was a factor 2 to 3 higher than otherwise , the qpo was not detected .
the strength of the qpo did not significantly depend on photon energy , and is consistent with being the same in the persistent emission , both during and outside the dips , and during type i x - ray bursts .
frequency shifts were observed during three of the four x - ray bursts .
we argue that the qpo is produced by the same mechanism as the qpo recently found by jonker et al .
( 1999 ) in 4u 132362 .
although the exact mechanism is not clear , it is most likely related to the high inclination of both systems .
an orbiting structure in the accretion disc that modulates the radiation from the central source seems the most promising mechanism . |
thz emission from intrinsic josephson junctions of the bscco type has received much attention recently .
several experiments have been reported @xcite , in which thz radiation emitted from bscco single crystals were observed .
however in most cases the detected power is rather small , or the frequency is rather low , or the emitted radiation is detected indirectly on an on - chip detector .
it has also been demonstrated that bscco can be considered a josephson junction with ac josephson effect even at frequencies as high as 2 thz @xcite . recently
a very convincing experiment was reported @xcite and it has attracted much focus and renewed experimental efforts .
parallel to the experimental work there has been theoretical / numerical work on fluxon dynamics in in layered superconductors of the bscco type @xcite .
the calculations demonstrate that the best way to obtain thz radiation is by having in - phase motion of the fluxons in the different layers .
this poses an interesting problem since the in - phase state of traveling fluxons is an energetically unfavorable state , and two fluxons of equal polarity will consequently repel each other .
however , due to the disparity of wave speeds for in - phase and out - of - phase solutions , it has been shown @xcite that the energetically unfavorable in - phase state of traveling fluxons are be stable above the asymptotic speed of the out - of - phase mode .
it has been assumed that the best way to obtain that is by having flux flow generated by a magnetic field applied parallel to the a - b plane , as studied theoretically in ref . .
however the successful experiment in ref . was done without a magnetic field , and it was suggested that an internal cavity ( based on the so - called fiske steps ) played a major role in the thz generation . in this paper
we study the fluxon modes in a stack of josephson junctions interacting with an external cavity .
we derive the conditions under which a large amount of current is induced in the cavity , and under which the external cavity may induce bunching of the josephson junction fluxons .
.,width=312 ] assuming that all the junctions in the stack are identical , the equations for a stack of long josephson junctions with @xmath0 superconducting layers and @xmath1 insulating layers can be written as @xcite @xmath2 where the @xmath3th element of @xmath4 , @xmath5 , is the gauge invariant phase difference across insulating layer @xmath3 .
the @xmath6 coupling matrix , @xmath7 , is given by ( only non - zero elements are shown ) @xmath8 with @xmath9 being the coupling parameter between the layers@xcite . the vector @xmath10 has the components @xmath11 where dampng parameter @xmath12 represents dissipation , and @xmath13 is the bias - current in the z - direction .
each component of @xmath14 is a current in the @xmath15-direction .
equations ( [ stackeq])-([current ] ) have been written in normalized units .
space @xmath16 is normalized to the josephson penetration depth , @xmath17 , and time @xmath18 is normalized to the inverse plasma frequency @xmath19 , where @xmath20 is the vacuum permeability , @xmath21 is the critical current of the individual josephson junctions , @xmath22 is the effective thickness of the insulating layer , and @xmath23 is the capacitance of the individual junctions , see refs . and for details . the model of the josephson stack coupled to a series cavity is shown in fig .
[ model ] , where @xmath24 is the cavity inductance , @xmath25 is the cavity resistance , and @xmath26 is the total cavity capacitance .
the boundary conditions for the phases can be written as@xcite @xmath27 where @xmath28 is the normalized capacitance and @xmath29 is the normalized charge in the cavity . defining @xmath30 to be the normalized cavity frequency and and @xmath31 as the quality factor
, the linear cavity equation becomes @xmath32 for more details on these equations see ref . .
two terms are present in eq .
( [ bc2l ] ) .
the first term couples the junctions to the cavity , by equally dividing the cavity current between the @xmath1 junctions .
the second term represents a direct coupling between the junction through the capacitors @xmath33 . in a real world situation
, the junctions would be embedded in a resonator and couple through electro - magnetic radiation from the edges .
the second term then models the radiation leaving one junction and ending up in another junction without being reflected by the cavity .
this is clearly not very efficient due to the geometry of the stack . with an efficient cavity , the second term is therefore expected to be much smaller than the first term and
may safely be neglected .
this may be justified by numerical calculations .
we thus choose to consider @xmath34 as the boundary conditions for eqs .
( [ stackeq])-([current ] ) and ( [ cavity ] ) .
we analyze the system in eqs .
( [ stackeq])-([current ] ) and ( [ cavity])-([bc ] ) in the case of weak inductive coupling where @xmath35 valid to first order in @xmath36 .
the solution to the linear cavity equation with initial conditions @xmath37 and @xmath38 is @xmath39 with @xmath40 .
the junction voltage at @xmath41 , @xmath42 , thus generates the cavity charge .
we look at the case where there is one fluxon in each junction and @xmath42 then becomes a voltage pulse . to simplify the integrations ,
these pulses are approximated by delta functions , i.e. , @xmath43 approximating voltage pulses at @xmath44 , @xmath45 , @xmath46 is the fluxon shuttling frequency in junction @xmath3 and @xmath47 is the phase shift of junction @xmath3 .
note , that since the present analysis is performed in the case of small @xmath36 , all the @xmath48-functions will have approximate the same amplitude , @xmath49 . with this ansatz , the cavity current becomes @xmath50e^{-\frac{\omega}{2q}\tilde{t}_n^i}\ , \label{gencur}\end{aligned}\ ] ] with @xmath51 and where @xmath52 is the heaviside step function .
limiting the analysis to the case of a high @xmath53 resonator , the steady state cavity current becomes @xmath54 for @xmath55 .
the phases , @xmath56 , are determined by [ eqlabel ] @xmath57 and @xmath58 only fluxons shuttling with the same frequency , @xmath59 for @xmath60 will be considered . in this case , eq .
( [ curgen ] ) can be reduced to @xmath61 with @xmath62 and @xmath63 for all @xmath3 .
thus , the cavity current is very simple when the cavity has reached a steady state .
note that the amplitude of the cavity current for an in - phase mode ( @xmath64 ) is @xmath65 . for an anti - phase mode ( @xmath66 )
the amplitude is @xmath67 for @xmath1 even and @xmath68 for @xmath1 odd . using eq .
( [ sinv ] ) , the hamiltonian of the stack of weakly coupled josephson junctions is @xmath69 with @xmath70 being the kronecker delta function . using eqs .
( [ stackeq])-([current ] ) , ( [ cavity ] ) and ( [ sinv ] ) the rate of change in energy is @xmath71_0^l\nonumber \ .\end{aligned}\ ] ] to determine the amplitude of the @xmath48-functions , we require that in the phase - locked state the energy - exchange of a `` collision '' with the boundary is the same for both a fluxon solution and the @xmath48-function approximation
. this energy - exchange is given by the time - integral of the last term in eq .
( [ dhdt ] ) , @xmath72_0^l dt\ , \nonumber \end{aligned}\ ] ] where @xmath73 and @xmath74 are taken such that they cover one collision with the boundary . to model a fluxon collision with the boundary , the following profiles are used@xcite @xmath75 with @xmath76 determining the fluxon polarity , @xmath77 being the lorentz factor , and the lowest characteristic velocity @xmath78 to first order in @xmath36 .
we take the same fluxon polarity in all junctions , thus @xmath79 for all @xmath3 .
following refs .
, using eqs .
( [ bc ] ) , ( [ iphaselock ] ) , and ( [ profile ] ) in eq .
( [ dhb ] ) yields @xmath80 \label{deltahb}\end{aligned}\ ] ] with @xmath81 and where the integration was carried out from @xmath82 to @xmath83 for mathematical convenience
. calculation of @xmath84 for the @xmath48-function approximation in eq .
( [ deltaapprox ] ) gives @xmath85\ .\end{aligned}\ ] ] requiring @xmath86 determines the amplitude of the @xmath48-functions to @xmath87 @xmath88 being given by eq .
( [ zeta ] ) and @xmath89 .
the asymptotic velocity , @xmath90 , present in eq . ( [ profile ] ) may be determined similarly to what is outlined in ref .
@xmath91 when the fluxons are shuttling with frequency @xmath92 .
the conditions for a steady state require that the energy averaged over one period is zero , thus @xmath93 using eq .
( [ dhdt ] ) , @xmath94 gives the condition @xmath95 where @xmath96 is determined from@xcite @xmath97 the current - voltage characteristics in eq .
( [ phaselock ] ) include the phase @xmath98 , such that at a given bias current the system can adjust this phase together with the collision times , @xmath47 , to satisfy condition ( [ phaselock ] ) ( if possible ) . the phase , @xmath98 , is related to the fluxon shuttling frequency , @xmath92 , through eqs .
( [ cphi ] ) and ( [ sphi ] ) and one may thus change the fluxon shuttling frequency by changing the bias current .
( [ phaselock ] ) @xmath13 is thus obtained as a function of @xmath92 . in the numerical simulations in section
iv we shall , inversely , obtain @xmath92 as function of @xmath13 . to calculate the conditions for the cavity to induce bunching ( in - phase motion )
, we consider a triangular fluxon configuration with one fluxon in each junction , modeled by @xmath99 .
the interaction energy between the fluxons is first calculated by considering an infinite line with a lattice spacing of @xmath100 , thus @xmath101 the interaction energy of this configuration is @xmath102 resulting in the well - known fluxon - fluxon force @xmath103 the force on the fluxons from the boundary can be calculated from @xmath104 using @xmath105 , valid for @xmath106 , resulting in @xmath107\ .
\nonumber\end{aligned}\ ] ] the separation between the fluxons , @xmath100 , may now be determined from the condition @xmath108 solving eq .
( [ balance ] ) for @xmath100 enables one to calculate the current voltage characteristics , eq .
( [ phaselock ] ) , and the cavity current , eq .
( [ iphaselock ] ) , in the steady state for a triangular fluxon lattice . + for the case of only two junctions , eq .
( [ fb ] ) reads @xmath109 in fig .
[ figfb ] we plot @xmath98 from eqs .
( [ cphi ] ) and ( [ sphi ] ) as well as @xmath110 from eq .
( [ fi ] ) and @xmath111 from eq .
( [ fb2 ] ) as a function of @xmath92 at constant @xmath100 .
it is seen that when the system is above the resonance frequency , the force from the boundary is negative while the @xmath110 is positive , thus they may balance each other .
below the resonance frequency the two forces has the same sign and the only steady state solution must be the one where the fluxons move in anti - phase .
the perturbation to the current - voltage characteristic by the cavity is contained in the @xmath84-term in eq .
( [ phaselock ] ) , given by eq .
( [ deltahb ] ) . in the case of anti - phase motion for two coupled junctions ,
this term will be zero . for three junctions ,
however , it will be non - zero .
( [ fi ] ) and ( [ fb ] ) will have the same direction for @xmath112 resulting in anti - phase motion below the resonance frequency .
this suggests that only for an odd number of junctions , we may observe an area of negative differential resistance in the current - voltage characteristics .
+ + + + + + + + numerical simulations of the full non - linear eqs .
( [ stackeq])-([current ] ) with boundary conditions ( [ cavity ] ) and ( [ bc ] ) has been done using second order finite differences for the spatial derivatives and a 5th order runge - kutta method with adaptive step size for the temporal integration@xcite .
the spatial resolution was kept at @xmath113 for all considered systems .
the initial fluxon configuration had one fluxon in each junction , each moving in anti - phase with the one in the neighboring junction(s ) .
the system was integrated until a stabilized cavity current was obtained or 20000 time units had passed .
the stable cavity current , @xmath114 , and the individual voltages at @xmath41 , @xmath115 , was analyzed using interpolation and fft@xcite to determine the most significant frequency in the power spectrum which is used as the cavity current frequency and the fluxon shuttling frequency , @xmath92 .
the difference in collision times , @xmath116 , can be calculated directly from the simulation using @xmath42 .
it may also be calculated analytically using eqs .
( [ uasymptotic ] ) and ( [ balance ] ) , where the latter eq . was solved numerically for @xmath100 . sometimes there were multiple solutions , @xmath117 , and we have chosen @xmath118 . when no solution was found in the interval
, we used @xmath119 corresponding to anti - phase motion .
the amplitude of the stabilized cavity current can be determined using a simple line search in the @xmath114 data and compared to the amplitude of eq .
( [ iphaselock ] ) using the value of @xmath100 obtained from eq .
( [ balance ] ) .
the frequency versus applied bias current can be compared to eq .
( [ phaselock ] ) , again using @xmath100 obtained from eq .
( [ balance ] ) .
when a steady state is found , we observe that @xmath120 with @xmath121 or @xmath122 , with @xmath123 giving a very low cavity current and therefore no significant difference from the unperturbed system . for high bias currents ,
the system was observed to be in the @xmath123 state and a switching to @xmath121 occurred when the system came close to the resonance in the current voltage characteristic . below resonance frequency , the system again switched to the @xmath123 state . to compare the analytical and numerical results
, we only show the numerical simulation points where a steady state was reached for the @xmath121 case . in figs .
[ l8 ] and [ l4 ] we have used eqs . ( [ iphaselock ] ) , ( [ phaselock ] ) , ( [ fi ] ) , and ( [ fb])/([fb2 ] ) for the analytical results ( shown with dashed lines ) . fig .
[ l8 ] shows the results on the system with @xmath124 and @xmath125 .
well above the resonance frequency we get very little current in the cavity .
as the shuttling frequency approaches the cavity frequency , the cavity current increases and reaches a maximum at near cavity frequency but suddently drops to near zero slightly above the cavity frequency . note that all numerical results are only shown for frequency values larger than the resonance frequency
this in general agreement with the findings in ref . where only oscillators with frequencies higher than the resonance frequency can be syncronised .
in addition , the forces shown in fig . [ figfb ]
are seen to be directed in opposite directions for frequencies larger than the resonance frequency and in the same direction for frequencies smaller than the resonance frequency .
the fluxon - fluxon distance will thus decrease only if the shuttling frequency is larger than the resonance frequency .
the corresponding current - voltage characteristic thus shows a deviation from the case without a cavity only above the resonance frequency .
the fluxon separation shows a similar behavior .
exactly at the resonance frequency , where the cavity current is at its maximum , the system exhibit anti - phase motion due to the boundary force being zero and we find the system to be in the @xmath123 state in the numerical simulations . slightly above resonance frequency
, the separation has a minimum and then it increases until it reaches at maximum at some point and then it decreases again .
the cavity current is thus largest slightly above resonance frequency , since the boundary force is zero exactly at resonance frequency , resulting in anti - phase fluxon motion .
[ l4 ] shows the corresponding results on the system with @xmath126 and @xmath125 .
the general behavior is the same as the one in fig .
[ l8 ] , except for the fluxon - fluxon separation .
below and at resonance frequency , we again see anti - phase behavior . slightly above the resonance frequency
, we see that the fluxon - fluxon separation has decreased to zero , i.e. the system has switched to a bunched state . at higher
fluxon shuttling frequency , the fluxon are separated at some distance @xmath100 and at some point this distance become so great that we again see anti - phase motion .
the minor discrepancies observed in figs .
[ l8 ] and [ l4 ] between theory and numerical experiment are primarily caused by the two core assumptions in the perturbation analysis ; namely the rigid collective coordinate approximation for the fluxon , and the idealized treatment of the fluxon reflection at the boundaries of the junction . among the approximations inherent to these assumptions are omission of phonons and the change in fluxon dynamics during reflections .
we notice , however , that the agreement between theory and simulations is very good , as can be seen in the figures .
the rather weak force induced by the cavity on the fluxons can only be used to obtain bunching in the weakly coupled case .
it is , however , essential that the fluxons do not move in perfect anti - phase in order to induce current into the cavity .
the top plot of fig .
[ l8highs ] the amplitude of the cavity current is shown for a simulation with similar parameters as fig .
[ l8 ] but with a much higher inductive coupling , @xmath127 , approaching the case of intrinsic junctions .
the simulation was started with a high bias current , resulting in a high fluxon - shuttling frequency and the bias current was gradually lowered resulting in lower fluxon shuttling frequencies .
( [ iphaselock ] ) gives zero cavity current in the case of a perfect anti - phase mode .
in the numerical simulations , however , we do not find zero cavity current , but rather that the system is in the @xmath123 state , i.e. the cavity is oscillating with twice the fluxon shuttling frequency . as the fluxon shuttling frequency
is lowered , the cavity current increases enough to slightly break the anti - phase motion and the cavity starts to oscillate at the fluxon shuttling frequency , seen in fig .
[ l8highs ] by following the line marked with a from the high frequency part to the low frequency part .
as the fluxon shuttling frequency is lowered still , the amplitude of the cavity current increases and the fluxon separation , @xmath128 , decreases . near the resonance the cavity current gets smaller and
the fluxon separation start to increase again . at some point ,
the fluxons start to behave erratic , meaning that we can not find a definitive value of @xmath100 in the simulations and thus we do not obtain steady state motion . the part of the curve near @xmath129 where the value of @xmath128 oscillates heavily shows this . at some point the system again switches to the anti - phase motion with the cavity oscillating at twice the fluxon frequency , seen by following the b-line from the low frequency part to the high frequency part of the figure .
we have not been able to determine if the erratic behaviour near @xmath129 is due to a too short simulation time before we give up finding a steady state of if this is the true behaviour of the system .
we have analytically calculated the cavity current and the current - voltage relation for @xmath1 weakly inductively coupled stacked josephson junctions coupled to a resonance cavity .
we have shown that the cavity introduces a force between the natively repulsive fluxons which may be used to obtain bunching in the weakly coupled case .
the effect is strongest for short junction , where the boundaries have larger influence .
our simple analysis show overall good agreement with numerical simulations . in the case of high inductive coupling
our perturbation results deviate from the simulations but the overall picture is still consistent with the theory .
sm and nfp would like to acknowledge the stvf framework program `` new superconductors : mechanisms , processes and products '' for financial support .
sm would also like to acknowledge financial support from the lundbeck foundation .
h. b. wang , p. h. wu , and t. yamashita , appl .
phys . lett . *
78 * , 4010 ( 2001 ) ; g. hetchfischer , w. walkenhorst , g. kunkel , k. schlenga , r. kleriner , p. mller , ieee trans .
. supercond .
* 7 * , 2723 ( 1997 ) ; k. kadowaki , i. kakeya , t. yamamoto , t. yamazaki , m. kohri , and y. kubo , physica c * 111 * , 437 ( 2006 ) . | we consider fluxon dynamics in a stack of inductively coupled long josephson junctions connected capacitively to a common resonant cavity at one of the boundaries .
we study , through theoretical and numerical analysis , the possibility for the cavity to induce a transition from the energetically favored state of spatially separated shuttling fluxons in the different junctions to a high velocity , high energy state of identical fluxon modes . |
the proposed experimental realization of the quantum simulator is based on the extension of an existing experiment . in the experiment @xcite ,
the potential of a tight optical dipole trap is populated with a degenerate fermi gas consisting of @xmath45li atoms in two hyperfine states . by applying a
magnetic - field gradient and varying the trap depth the atoms tunnel out of the trap in a fully controlled way , ending up with a defined number of particles in a determined quantum state .
the experimental dipole trap is described in good approximation by a gaussian - beam potential which results in a quasi 1d confinement with an aspect ratio of about 1:10 .
the beam profile in the longitudinal direction is approximately given by the lorentz potential @xmath46 \quad .\end{aligned}\ ] ] here , @xmath47 is a modulation factor which allows to vary the trap depth in the experiment , see @xcite , @xmath48 is the potential depth , @xmath49 the boltzmann constant , @xmath50 the rayleigh length with a laser wavelength of @xmath51 nm . in the experiment ,
the value of the waist @xmath52 has been @xmath53 .
achieving values lower than this is challenging but in principle a value of at least @xmath54 could be realized with a new experimental setup @xcite .
here we chose @xmath55 .
moreover , we chose a realistic value for the trapping depth @xmath56 . certainly with the chosen lorentz potential the long - range interaction of the nuclei and the electrons
can not be reproduced exactly . however , the choice of the parameters are such that energetic distribution of the low - lying bound - states resemble appropriately those of a 1d hydrogen atom with @xmath57 + on the other hand , systems with smaller binding potentials , like , e.g. , anions , could be simulated more accurately . for anions , the long - range interaction of an emitted electron with the remaining neutral atom scales as @xmath58 .
this is more comparable to the long range interatomic interaction which scales like @xmath59 for neutral atoms or with @xmath60 for dipolar atoms . for the simulator system natural units ( n.u .
) are introduced in which @xmath61 , the magnetic moment @xmath15 , the atomic mass @xmath14 and the trap length @xmath62 , which is equal to the extension of the ground - state wave function ( defined as the distance where the ground - state wavefunction has decreased to 1/e of its maximum value ) , are set to unity .
it should be emphasized that different to atomic units ( a.u . ) which are uniquely defined to reflect the electronic properties of the hydrogen atom , the introduced natural units change with the trapping potential . for a given simulator setup and a specific pulse ,
the corresponding values for the frequency @xmath33 and peak vector potential @xmath34 applied in the strong - field system are found by enforcing equal keldysh parameters @xmath63 and equal parameters @xmath64 here , @xmath65 and @xmath66 are the binding energies of the ground state of the field - free hamiltonians ( 1 ) and ( 3 ) in the main manuscript , respectively .
this mapping is not unique since @xmath66 , which is determined by the shape of the trapping potential , is a free parameter .
this freedom can be used to better adjust the trapping potential such that the energy - level distribution of the simulated system is resembled accurately . in the manuscript ,
the considered pulses are defined via the vector potential although the simulator mapping is performed in length gauge . since the simulator system is directly compared to the corresponding strong - field system ,
this is a convenient way to ensure that the pulse fulfills the zero net - force condition , i.e. the total integral over its electric ( or magnetic ) field is zero and the vector potential has the same value before and after the pulse .
the tdse @xmath67 for the hamiltonians @xmath68 and @xmath69 , ( 1 ) and ( 3 ) in the main manuscript , respectively , are solved by expanding the wavefunction @xmath70 in eigenstates @xmath71 of the field - free hamiltonians
@xmath72 and @xmath73 , ( 2 ) and ( 4 ) in the main manuscript , respectively . +
while the finally shown spectra have been calculated within velocity gauge , i.e. equations ( 7 ) and ( 8) in the main manuscript , to achieve faster convergence , it was verified for a selected number of laser parameters that length and velocity gauge results are in full agreement .
the simulator system is treated numerically in one dimension .
this approximation is well satisfied for the here considered experimental realization since for ultracold temperatures the transversal motion of a strongly anisotropic trapping potential as realized in @xcite is frozen out to the ground state .
for the solution of the 1d simulator [ soft - coulomb ] system , the eigenstates of the field - free hamiltonian are calculated via the matrix algorithm @xcite .
the results shown in the manuscript were obtained with a box size of @xmath74 n.u .
[ @xmath74 a.u . ] and a grid consisting of @xmath75 [ @xmath76 points .
for the time propagation , only @xmath77 [ @xmath78 states were used .
+ for the time propagation of the hydrogen atom , the field - free eigenstates are obtained in spherical coordinates @xmath79 by expanding the radial problem in b splines ( as in @xcite ) .
the results shown in the manuscript were obtained with a box size of @xmath80 a.u . including states with angular momenta up to @xmath81 . along the radial coordinate @xmath82 b splines of order @xmath83 and a linear knot sequence were used . for the time propagation , only states with energies up to @xmath84 a.u . above the ionisation threshold
were considered .
+ convergence of all spectra shown in the manuscript was ensured by varying the basis . in analogy to the tdse , the direct sfa ionization amplitude @xmath85 is easily translated from the corresponding formula for the strong - field amplitude @xcite applying equations ( 5 ) and ( 6 ) in the main manuscript . in length gauge
it reads @xmath86 where @xmath87 , @xmath88 is the classical action , @xmath89 is the vector potential according to the mapping ( 6 ) in the main manuscript , @xmath90 is the fourier transform of the initial state wavefunction , and @xmath91 is the volkov wavefunction . in velocity gauge ,
the amplitude is @xmath92 and thus the yield @xmath93 contains the momentum - space density @xmath90 as a factor .
this provides a way to extract the momentum - space density of the initial state from the measured energy - resolved spectra . in the limit of an infinite pulse ,
the integral even simplifies to an expression including generalized bessel - functions @xcite . | a quantum simulator based on ultracold optically trapped atoms for simulating the physics of atoms and molecules in ultrashort intense laser fields is introduced . the slowing down by about 13 orders of magnitude allows to watch in slow motion the tunneling and recollision processes that form the heart of attosecond science .
the extreme flexibility of the simulator promises a deeper understanding of strong - field physics , especially for many - body systems beyond the reach of classical computers .
the quantum simulator can experimentally straightforwardly be realized and is shown to recover the ionization characteristics of atoms in the different regimes of laser - matter interaction . in his renowned lecture , `` simulating physics with computers '' @xcite richard p. feynman suggested the use of quantum simulators , i.e. precisely controllable quantum systems , to simulate other quantum systems that can not be described theoretically due to their exponentially growing hilbert space .
for instance , the mott - insulator to superfluid phase transition in condensed - matter systems @xcite was predicted @xcite to be observable with ultracold atoms in an optical lattice and then successfully demonstrated @xcite .
also the higgs mechanism @xcite , high temperature superconductivity @xcite , or _
zitterbewegung _ @xcite ( to name just a few ) were successfully investigated by quantum simulation .
moreover , the quantum simulation of electrons in crystalline solids exposed to laser fields @xcite has been proposed .
strong - field physics has contributed considerably to the understanding of the light - matter interaction .
the progress leading to pulses on the attosecond timescale @xcite has even raised visions of real - time imaging of molecular processes @xcite and orbital tomography @xcite . yet
, attosecond many - body physics is challenging . an exact investigation on classical computers beyond
the single - active - electron approximation becomes prohibitively complex for many - electron systems .
in fact , the numerical treatment of two - electron systems like he or h@xmath0 is today still state of the art @xcite . thus , simplified models are widely used for interpreting modern experiments .
these models are controversial and their validation is difficult for several reasons .
first , the used light pulses are bound to the specifications of the laser . the wavelength range of lasers is limited , mostly ti : sapphire lasers are used .
the pulse shapes are restricted and can often only be reproduced and determined up to a considerable uncertainty . the intensity and timescale of laser pulses
are already pushed to a limit where further improvements require major technical or even principle developments with new limitations , like free - electron lasers .
second , atoms , ions , and molecules are complicated many - body systems .
their internal structure can not be simply manipulated . for example , a variation of the number of electrons or protons underlies constraints due to electroneutrality .
third , although the correlation of electronic and nuclear motion is known to influence the ionization behavior @xcite , in most theoretical models this effect is neglected by fixing the nuclei in space while investigating the electronic response to the laser field . in this work
, we introduce the concept of an ultracold - atom quantum simulator for attosecond science which offers great flexibility and control beyond the mentioned limitations .
this includes many - body quantum simulations that are impossible with any classical computer .
_ the attoscience simulator .
_ the simulator system consists of ultracold trapped atoms that replace the electrons in the atom , ion , or molecule , see fig .
[ fig : fig1 ] .
the core potential is replaced by an external , optical trapping potential .
the ability to implement single - well or multi - well trapping potentials allows for a simulation of atoms or molecules , respectively .
naturally , fermionic atoms may be chosen , but using bosons or distinguishable particles reveals effects of the exchange interaction .
the intense laser pulse is replaced by a periodically driven magnetic - field gradient which is generated by current - carrying coils .
restrictions for ultrashort laser fields like the zero - net - force condition @xcite do not apply here and thus fields of almost arbitrary shape can be created , even true half - cycle pulses and fields that formally correspond to sub - attosecond pulses .
certainly , the atom - atom interaction is shorter ranged than the coulomb interaction .
however , earlier quantum simulations like the famous superfluid to mott - insulator phase transition @xcite demonstrated that an equivalent physics is obtainable .
the use of ultracold atoms introduces the unique opportunity to arbitrarily vary the effective interaction strength _ via _ magnetic feshbach resonances .
this promises new insights on the influence of the interparticle interaction on the ionization behavior .
furthermore , theoretical studies which replace the core potential by , e.g. , a zero - range potential @xcite , can now be tested experimentally , and this even for many - particle systems .
since ultracold quantum systems are manipulated nowadays on the single - atom level @xcite , important tests of the widely used single - active - electron approximation and a detailed investigation of correlated many - body tunneling become accessible .
moreover , only the simulator allows for the experimental realization of fixed nuclei a task impossible with real molecules due to the heisenberg uncertainty principle .
the influence of a fixed nuclear geometry on the ionization behavior @xcite can thus be tested experimentally in a clean fashion . additionally , the differences between the quantum - mechanical nature of vibronic states and the simulation of a mechanical vibration of the nuclei can be investigated .
_ hamiltonian mapping . _
the formal equivalence of the quantum simulator hamiltonian to the electronic strong - field hamiltonian at a fixed nuclear configuration is demonstrated .
when treating the strong laser field classically , which is acceptable due to its high intensity , and applying dipole approximation and length gauge ( lg ) , respectively , the electronic strong - field hamiltonian reads @xmath1 where @xmath2 denotes the field - free hamiltonian for @xmath3 electrons .
@xmath4 is the electron mass , @xmath5 the electron charge , @xmath6 includes all electron - electron repulsion terms , and @xmath7 all the electron - nucleus interactions .
@xmath8 denotes the electric - field component of the pulse . in analogy ,
the hamiltonian of @xmath3 ultracold atoms confined in a trapping potential @xmath9 which are exposed to a time - dependent magnetic - field gradient @xmath10 reads @xmath11 where @xmath12 denotes the hamiltonian of the atoms in the trap without the gradient @xmath13 .
@xmath14 denotes the atomic mass , @xmath15 the magnetic moment of the atoms , @xmath16 includes all atom - atom and @xmath9 all atom - trap interactions , respectively .
the hamiltonians ( [ eq : sf_hamil ] ) and ( [ eq : uc_hamil ] ) are _ formally _ equivalent under the mapping @xmath17 it is important to note that the electrodynamical potentials , e.g. , the vector potential @xmath18 , map accordingly . in the ultracold simulator system
the `` vector potential '' is thus given by @xmath19 of course , the potential @xmath20 differs from the physical vector potential @xmath21 that generates the magnetic field @xmath22 and its gradient @xmath13 via @xmath23 . yet , equation ( [ eq : vector_pot ] ) is the formal consequence of the simulator mapping ( [ eq : mapping ] ) .
the simulator mapping ( [ eq : mapping ] ) is intrinsically defined in length gauge . however , it is particularly useful to consider the analog of the velocity - gauge ( vg ) formulation , too .
a gauge transformation of the strong - field hamiltonian ( [ eq : sf_hamil ] ) leads to the velocity - gauge form @xmath24 in analogy , a `` gauge '' transformation of the simulator hamiltonian ( [ eq : uc_hamil ] ) leads to the corresponding simulator hamiltonian in `` velocity gauge '' , @xmath25 again , the hamiltonians are _ formally _ equivalent .
the vector potential @xmath26 is the one resulting from the simulator mappings ( [ eq : mapping ] ) and ( [ eq : vector_pot ] ) .
+ _ experimental realization .
_ as a possible realization we consider the experiment of the group of selim jochim @xcite where a well defined number of fermionic atoms can be loaded into a tight optical dipole trap in a well defined quantum state .
this trap is in good approximation described by a one - dimensional lorentz potential @xcite .
a _ static _ magnetic - field gradient which tilts the continuum threshold of the trap , see fig .
[ fig : fig1 ] , is applied for the preparation and investigation of the system .
the here proposed attoscience quantum simulator is realized by replacing the static magnetic - field gradient by a periodically driven one .
similarly to strong - field physics , the pulse may be defined by its vector potential _ via _
@xmath27 here , @xmath28 is the number of cycles , @xmath29 is the carrier - envelope phase , @xmath30 the angular frequency , and @xmath31 from which @xmath32 is obtained _ via _ equation ( [ eq : vector_pot ] ) is the strength of the perturbation . for a given simulator setup and a specific pulse ,
the corresponding values for the frequency @xmath33 and peak vector potential @xmath34 applied in the strong - field system are found by enforcing equal keldysh parameters @xcite and an equal ratio of the binding energy to the frequency of the perturbing field @xcite .
is shown .
the sfa yields in ( c ) and ( d ) are rescaled in order to agree with the total ionization yield of the tdse calculation .
the factors are given in the figure legends .
, scaledwidth=48.0% ] _ validation of the quantum simulator .
_ in the experiment @xcite the simulator builds on , the atom loss is routinely measured .
this observable corresponds to a measurement of the total ion ( or electron ) yield in a strong - field experiment .
more detailed information on the underlying physics is obtained by a measurement of differential yields : energy - resolved electron or atom spectra for strong - field experiments or the simulator , respectively .
the measurement of energy - resolved atom spectra requires further experimental developments , similarly to strong - field physics where in the early days also only total yields were measured . to validate the simulator in more detail , energy - resolved electron spectra of a hydrogen atom
are compared to the corresponding energy - resolved atom spectra of the simulator setup , both initially in their ground state .
the spectra are calculated by solving the corresponding time - dependent schrdinger equations ( tdse ) @xcite , ensuring that the corresponding parameters for the quantum simulator are experimentally accessible @xcite .
the laser - matter interaction is typically divided into two characteristic regimes . in the low - frequency , high - intensity regime the system
is assumed to follow adiabatically the changes of the electric field of the laser . in this quasistatic regime
the electron is supposed to tunnel through or escape over the field - distorted potential barrier , see fig .
[ fig : fig1 ] . in the other limit of the high - frequency , low - intensity regime the multiphoton picture
is usually adopted in which the ionization is described within a simplified picture as an absorption of photons , despite the fact that in the theoretical treatment the electromagnetic field is treated classically . in the multiphoton regime ( figs .
[ fig : fig2]a and b ) , both spectra show the typical multi - peak structure ( above - threshold - ionization peaks ) where the peak distance reflects the frequency of the perturbing field . clearly , simulator and hydrogen atom show very good agreement . despite the different dimensionalities the tdse solutions
agree , in fact , almost quantitatively . in the quasistatic regime ( figs .
[ fig : fig2]c and d ) , a simple tunneling picture suggests an exponential decrease in the energy - resolved spectra , as is seen in the low - energy part ( up to @xmath35 where @xmath36 is the ponderomotive energy and @xmath37 the laser intensity ) .
however , in a periodically changing field the emitted electron or atom can reverse its direction of motion and recollide , see fig .
[ fig : fig1 ] . high - harmonic generation in strong - field physics is based on the recombination of the liberated electron with the parent ion at the recollision step .
using classical newtonian mechanics it had been found that high - harmonic spectra extend up to @xmath38 @xcite . for energy resolved electron spectra , the recollision process leads to a broad energy distribution of the rescattered electrons which manifests in a plateau as observed in @xcite and clearly seen in fig .
[ fig : fig2]d . in analogy to the high - harmonic cutoff law
classical newtonian mechanics predicts an extension of this plateau between @xmath39 and @xmath40 @xcite .
clearly , the simulator shows all expected features , both from tunneling and rescattering .
however , the more pronounced structures in the plateaus of the 1d systems ( simulator and 1d hydrogen atom ) reveal effects of the dimensionality .
such effects can be studied with the simulator even experimentally by varying the anisotropy of the trap a task impossible in strong - field experiments .
rescattering is the origin of nonsequential double ionization , high - energy above - threshold ionization , and high - order harmonic generation .
a controlled recollision , see fig . [ fig : fig1 ] , of an escaped atom on residual bound atoms prepared in a specific configuration with variable interaction strength can reveal insights into correlated recollision dynamics relevant , e.g. , for high harmonics @xcite and non - sequential double ionization @xcite .
on the other hand , inspired by the experiments on imaging molecular orbitals using laser - induced electron tunneling and diffraction @xcite controlled rescattering collisions can serve for the imaging of ultracold many - body wavefunctions .
_ strong - field approximation .
_ in the widely used strong - field approximation ( sfa ) @xcite bound states of the potential other than the initial state are neglected and the final continuum state is replaced by a volkov state , i.e. the solution of a free electron in a laser field . therefore , the interaction of the electron with the remaining ion is ignored in the final state .
thus , the sfa does not support rescattering as can also be seen in figs .
[ fig : fig2]c and d. on the other hand , the direct electrons and atoms ( up to @xmath39 ) in the quasistatic regime in figs .
[ fig : fig2]c and d are qualitatively well described by the sfa . similarly , the sfa reproduces the multi - peak structure in the multiphoton regime , see figs .
[ fig : fig2]a and b. note , the sfa is not gauge invariant and so far no arguments from first principles are known what gauge is to be preferred in which situation @xcite ( the gauge problem of sfa ) . in the simulator system ,
the number of trap states as well as the potential range can be varied which allows for an analysis of the assumptions and the gauge ambiguity of the sfa .
interestingly , the sfa in velocity gauge allows to obtain the momentum density of the initial state since the energy - resolved yield is a product of the momentum - space density @xmath41 and a prefactor @xmath42 @xcite . for a given momentum @xmath43 , the prefactor @xmath44 depends solely on the vector potential @xmath26 and the binding energy .
in contrast to the corresponding strong - field experiments , these parameters are known precisely for the quantum simulator because of the exactly known pulse shape .
note , this imaging technique relies on the agreement of the sfa in velocity gauge with the full tdse results , which is fulfilled as seen in figs .
[ fig : fig2]a and c despite the fact that the simulator mapping ( [ eq : mapping ] ) is bound to the length gauge .
thus , this imaging technique indeed allows to image the momentum density of an ultracold gas in a trap .
_ conclusion .
_ a proposal for a quantum simulator for attosecond physics is presented based on ultracold atoms in an optical trapping potential .
the simulator idea connects the very contrary physics of ultracold , trapped atomic gases and the one of atoms , ions , and molecules in ultra - intense , ultra - short laser fields .
the constraints one faces in strong - field experiments , such as the limitation to a specific molecular geometry , a fixed number of electrons per element or molecule , fixed interaction strengths , and limited pulse shapes are overcome in the simulator system .
moreover , the simulation can even reach parameter regions which are beyond those nowadays realizable in strong - field experiments , including , e.g. , exotic pulse shapes and effective pulse durations corresponding to the sub - attosecond regime .
in fact , the here proposed attosecond science in slow motion may shed light onto the ongoing debate on tunneling times @xcite .
the numerical analysis of the here proposed concrete experimental realization of the quantum simulator with realistic experimental parameters demonstrates that it reproduces in its simplest configuration the ionization characteristics of a hydrogen atom .
while this simple demonstrating example can be evaluated computationally , the simulator paves the way to systematically investigate many - body systems where the full numerical treatment is beyond the reach of any classical computer .
also the physics of ultracold atoms may profit from the quantum simulator by adopting concepts developed in attosecond science .
42ifxundefined [ 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 * * , ( ) link:\doibase 10.1126/science.1165449 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop @noop * * , ( ) @noop @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop @noop * * , ( ) @noop * * , ( ) |
stern s diatomic sequence @xmath0 is a particularly well studied sequence ( see , e.g. , @xcite , @xcite , @xcite and references therein , as well as @xcite ) .
the first section is devoted to showing that this sequence is _
interesting_. in particular , we shall look at the following properties .
* @xmath1 is a bijection between the positive natural numbers and the positive rational numbers , * @xmath2 extends to a continuous strictly increasing function on @xmath3 $ ] known as `` conway s box function '' ( it s inverse is @xmath4 , minkowski s question - mark function ) , * it shares a number of similarities to the fibonacci sequence ; in particular , it has a binet type formula .
.1 in the remaining three sections are devoted to three analogues of stern s sequence : * we replace addition by another binary operation ; in particular , we define @xmath5 where @xmath6 .
this sequence is related to stern s sequence and arises from certain sphere packings .
it has apparently not appeared before in the literature .
* we replace the complementary indexing sequences @xmath7 and @xmath8 by another pair of complementary sequences ; in particular , let @xmath9 where @xmath10 , @xmath11 form a specific pair of complementary beatty sequences .
this sequence has been extensively studied as @xmath12 is the number of ways @xmath13 can be represented as a sum of distinct fibonacci numbers . * the known binet type formula for stern s sequence @xcite is written in terms of the sequence @xmath14 ( : = the number of terms in the binary expansion of @xmath13 ) .
we replace @xmath14by @xmath15(:= the number of terms in the zeckendorf representation of @xmath13 ) . this new sequence ,
apparently not studied before , is an integer sequence with several interesting ( and several conjectural ) properties .
consider the following diatomic array " @xcite formed as a variant of pascal s triangle ; each entry is either the value directly above or else the sum of the two above it . @xmath16
the word diatomic " is used here since every entry of the diatomic array gets its value from either one or two entries above and gives that value to three entries below , hence has valence " 4 or 5 ( hence the diatomic array models a kind of crystalline alloy of two elements ) .
ignoring the right most column and reading the numbers as in a book , we get stern s diatomic sequence : @xmath17 the sequence is thus defined by the recurrence @xmath18 we define @xmath19 to be 0 ( the value consistent with @xmath20 ) .
perhaps the most celebrated property of this sequence is that every positive rational number is represented exactly once as @xmath21 .
see , for example , @xcite or @xcite .
we rephrase this fact as a theorem .
every ordered pair of relatively prime positive integers appears exactly once in the sequence @xmath22 . for an ordered pair ,
consider the process of subtracting the smallest from largest ( stop if equal ) .
for example , @xmath23 and @xmath24 . by the definition of stern s sequence ,
@xmath25 every relatively prime pair appears ( if not , then there is an ordered pair not on the list with lowest sum .
apply the process ; the result has lower sum and so is @xmath26 for some @xmath13 and so the original pair is either @xmath27 or @xmath28 ) .
every relatively prime pair appears _ exactly _ once since , if not , then there exist @xmath29 with @xmath30 and such that @xmath31 is as small as possible . applying
the process to both implies @xmath32 and thus @xmath33 which is impossible
. one can then rewrite any sum over relatively prime pairs in terms of stern s sequence . as an example
, we rephrase the riemann hypothesis .
first note that @xmath34 is an explicit bijection from @xmath35 to @xmath36 .
then the riemann hypothesis is equivalent to @xmath37 briefly why this is so : the mbius function can be written as @xmath38 and so the left hand side is really just merten s function @xmath39 .
the connection between merten s function and the riemann hypothesis is well - known ; see for example @xcite .
minkowski s question mark function was introduced in 1904 as an example of a singular function " ( it is strictly increasing yet its derivative exists and equals 0 almost everywhere ) .
it is defined in terms of continued fractions : @xmath40 where @xmath41 . by lagrange s theorem that states that the continued fraction representation of a quadratic surd must eventually repeat ,
it is clear that @xmath4 takes quadratic surds to rational numbers .
the function @xmath42 extends to a continuous strictly increasing function on @xmath3 $ ] .
this function is known as
conway s box function " and its inverse is minkowski s question mark function @xmath4 . and
its inverse @xmath43.,title="fig:",width=144 ] and its inverse @xmath43.,title="fig:",width=144 ] see @xcite for a proof .
the functions @xmath44 and @xmath4 extends to homeomorphisms ( or , equivalently , are restrictions of homeomorphisms ) between two fractals . stern s sequence is related to the fibonacci sequence in a number of ways . for example ,
the fibonacci sequence is a subsequence : @xmath45 it is easy to see that @xmath46 where @xmath47 is the jacobsthal sequence ( * ? ? ?
* a001405 ) . a new result by coons and tyler @xcite identifies and proves the asymptotic upper bound : @xmath48 the constants involved in this formula are perhaps not so surprising since , by formula ( 2 ) , it is clear that @xmath49 stern s sequence has a few remarkable similarities to the fibonacci sequence ( see @xcite and @xcite ) . for example , stern s sequence satisfies a modified fibonacci recurrence : @xmath50 next , certain diagonal sums across pascal s triangle yield the fibonacci sequence while the corresponding sums across pascal s triangle modulo 2 yield stern s sequence : @xmath51 @xmath52\ ] ] recall binet s formula @xmath53 stern s sequence satisfies a similar formula : @xmath54 where @xmath55 is a sixth root of unity and @xmath14 is the number of ones in the binary expansion of @xmath13 ( * ? ? ? *
a000120 ) : @xmath56 why this is true : if @xmath57 and @xmath58 is the generating function for @xmath59 , then @xmath60 and @xmath61 . for real @xmath62 , since @xmath63 , @xmath64 and the result follows by equating coefficients .
for non - negative real numbers @xmath65 , let @xmath66 if @xmath67 and @xmath68 then @xmath69 . if @xmath70 , then @xmath71 and thus @xmath72 by the fibonacci identity @xmath73 it follows that @xmath74 and so the sequence @xmath75 satisfies the modified fibonacci recurrence @xmath76 .2 in here we define the first new sequence .
let @xmath77 , and for @xmath78 , @xmath79 the sequence begins @xmath80 it is not immediately clear that this sequence must always be integral .
one way to show this is to express each @xmath81 as a product of elements of stern s sequence ( theorem 3.6 , below ) .
first we must prove a lemma . for @xmath82 , if @xmath83 then @xmath84 .
we prove this by induction on @xmath85 .
if @xmath86 , then @xmath87 and the result holds for @xmath88 . suppose now that the result holds for a fixed @xmath85 and that @xmath89 .
without loss of generality , @xmath90 and @xmath91 for some @xmath92 ( and so @xmath93 ) . then @xmath94 and the result follows . if @xmath95 , then @xmath96 if @xmath97 then , because @xmath98 , @xmath99 .
let @xmath100 where @xmath101 then @xmath102 and thus @xmath103 and , by lemma 3.5 and proposition 3.2 , @xmath104 hence @xmath105 for all @xmath106 , and the result follows . @xmath107 .
as seen in section 2 , stern s diatomic sequence leads to a construction of conway s box function @xmath44 , the inverse of minkowski s question - mark function .
the sequence @xmath108 gives rise to a similar function that turns out to be closely related to @xmath44 .
for @xmath109 , @xmath110 , let @xmath111 the function @xmath112 extends to a continuous function on @xmath3 $ ] that satisfies , for @xmath113 , @xmath114\ ] ] where @xmath44 is conway s box function .
let @xmath115 .
then @xmath116 for some @xmath117 . since @xmath118 it follows from theorem 3.6 that @xmath119 by ( * ? ? ?
* formulas ( 2 ) and ( 3 ) ) , @xmath120 and thus @xmath121=f(2^{j+1}x-1)[1-jf(2x)].\end{aligned}\ ] ] the extension of @xmath112 to a continuous function on @xmath3 $ ] follows from the facts that @xmath122 extends to a continuous function on @xmath3 $ ] and @xmath123 a scaled version of the graph of @xmath124 appears in the graph of @xmath125 . ,
width=144 ] the restriction of @xmath112 to @xmath126 $ ] is just a scaled version of @xmath44 : @xmath127 for @xmath128 $ ] .
recall that every positive rational number appears exactly once in the set @xmath129 .
we prove an analogue for the sequence @xmath108 .
we use the expression @xmath130 " to mean that @xmath131 for some integer @xmath13 .
every element of @xmath132 appears exactly once in the sequence @xmath133 .
consider the following analogue of the ( slow ) euclidean algorithm .
@xmath134 suppose @xmath135 , with @xmath136 . if @xmath137 then it is easy to see that @xmath138 and thus @xmath139 . in this case , since @xmath140 unless @xmath141 , the only possibility is @xmath142 .
hence , @xmath143 and , if this algorithm terminates at all , it must terminate at @xmath144 . with @xmath135 , with @xmath136 , let @xmath145 .
if @xmath146 , then @xmath147 and thus @xmath148 in general , the product of numbers in @xmath149 is strictly less than the product @xmath150 and thus the algorithm will eventually reach , without loss of generality , @xmath151 . if @xmath152 then the algorithm stops . on the other hand ,
if @xmath153 , it is easy to see that @xmath154 , and thus the algorithm will terminate at @xmath144 .
let @xmath155 . by the definition of the sequence @xmath108 ,
it s easy to see that for @xmath156 , @xmath157 and , moreover , if @xmath158 , then either @xmath159 or @xmath160 .
if @xmath135 , with @xmath136 is not of the form @xmath161 for some @xmath13 , then all of its successors under @xmath162 , including @xmath144 , are not either a contradiction .
hence every @xmath135 , with @xmath136 is of the form @xmath161 for some @xmath13 .
the pair @xmath144 appears only once and , in general , no pair appears more than once in @xmath163 for , otherwise , there exists a smallest @xmath156 such that @xmath164 for some @xmath165 . applying @xmath162 to both @xmath166 and @xmath161 forces @xmath167 and therefore @xmath168 .
thus @xmath169 , a contradiction .
a generalization of @xmath170 is as follows : for a given number @xmath171 , define @xmath172 @xmath173 solve @xmath174 defining @xmath175 in the obvious manner , @xmath176 every non - zero complex number @xmath177 can be represented uniquely as @xmath178 for some positive @xmath179 and some @xmath180 and so we define @xmath181 .
hence @xmath182 and @xmath175 are well defined for complex @xmath171 .
we may then generalize @xmath108 . given a ( complex ) number @xmath183 ,
let @xmath184 and , for @xmath78 , @xmath185 it turns out that such a sequence can be expressed as a linear combination of the sequences @xmath186 and @xmath108 .
we first need a lemma . for @xmath187 , @xmath188 let @xmath189 .
note that @xmath190 then @xmath191 and so @xmath192 similarly , @xmath193 and so @xmath194 note that @xmath195 and so the lemma holds for @xmath196 .
suppose the lemma holds for a particular @xmath106 .
we show it works for @xmath197 and @xmath198 and thus , by induction , the lemma will be shown .
@xmath199 and thus the lemma works for @xmath197 . @xmath200 and thus the lemma works for @xmath198 .
given @xmath201 , let @xmath202 then @xmath203 has @xmath184 and , for @xmath204 , @xmath185 @xmath205 and so @xmath206 and thus @xmath207 since @xmath208 the theorem is shown .
let @xmath209 , @xmath210 , we see that @xmath211 , @xmath212 , and thus @xmath213 .
if to every local cut point @xmath214 in the fractal cp appearing in figure 2 one attaches a sphere above but tangent to the plane at that point with curvature ( 1/radius ) equal to the sum of the curvatures of the two circles meeting there , then one gets a 3-dimensional generalization of ford circles .
the curvatures ( similarly , the product of local cut points and corresponding curvatures ) along any circular arc are from a sequence @xmath215 for appropriately chosen @xmath171 ( see @xcite and references therein for a discussion of various types of ford spheres " ) . consider the sequence @xmath108 written in tabular form : @xmath216 it is apparent that every column is an arithmetic sequence and , moreover , the defining differences are respectively @xmath217 the squares of stern s diatomic sequence @xmath186 .
this is , in fact , true .
we shall express this result as a formula .
for @xmath218 , @xmath219 assume @xmath218 .
since @xmath220 , theorem 3.6 implies @xmath221 by ( * ? ? ?
* formulas ( 2 ) and ( 3 ) ) , @xmath222 and thus @xmath223 the result follows by induction .
@xmath224 appears as ( * ? ? ?
* a119272 ) , the product of numerators and denominators in the stern - brocot tree . for a fixed @xmath225 , @xmath226 and @xmath227
are the two solutions of @xmath228 .4 in
a _ fibonacci representation _ of a number @xmath13 is a way of writing that number as a sum of distinct fibonacci numbers .
one such representation is , of course , the zeckendorf representation which is gotten by the greedy algorithm and which is characterized by having no two consecutive fibonacci numbers . in general
, a given @xmath13 has several fibonacci representations , the number of such we call @xmath12 .
the sequence @xmath229 is extremely well studied ; see papers by klarner @xcite , bicknell - johnson @xcite , and stockmeyer @xcite , for example .
a string of 0s and 1s is a finite word with alphabet @xmath230 ( equivalently , an element of @xmath231 ) .
often we denote such a word by @xmath232 .
we shall think of such strings as fibonacci representations : we shall assign a numerical value @xmath233 $ ] to a string @xmath232 by the formula @xmath234=
\sum i_j f_{k+2-j}.\ ] ] for example , @xmath235=[0011]=3 $ ] and @xmath236=21 + 8 + 2 + 1=32 $ ] .
the generating function for @xmath229 has an obvious product formulation .
the sequence @xmath237 satisfies @xmath238 where @xmath239 denotes the @xmath13th fibonacci number .
next , we define the _ fibonacci shift _ : @xmath240 that satisfies @xmath241)=[\omega 0]$ ] for every string @xmath232 . this shift has been studied before ; for example , it appears in ( * ? ? ?
* graffiti , p. 301 ) . for @xmath242 , @xmath243 , @xmath244 by binet s formula ( 3 ) , @xmath245 for any choice @xmath242 for @xmath243 , note that @xmath246 and therefore @xmath247 hence , @xmath248 in terms of @xmath249 , we may define @xmath229 recursively .
clearly , @xmath250 . a representation of @xmath13
either ends in @xmath251 in which case @xmath252 $ ] where @xmath241)=n$ ] or else it ends in @xmath253 in which case @xmath254 $ ] and so @xmath255=\rho([\omega])$ ] .
hence , for all @xmath78 , @xmath256 note that the function @xmath257 is an example of a beatty sequence ( i.e. , of the form @xmath258 ) and so has a complementary beatty sequence , namely @xmath259 .
for example , @xmath260 and @xmath261 the following characterization could be used as a new definition of @xmath229 . for @xmath78 , and @xmath259 , @xmath262 and @xmath263 since @xmath264 , @xmath265 since @xmath266 , @xmath267 and
so @xmath268 note that @xmath269 and therefore @xmath270 we show the first equation in the theorem by a counting argument . by the definition of @xmath249 , @xmath271 and so @xmath272 for a given @xmath13 , if @xmath273 $ ] then @xmath274 $ ] and @xmath275 equals either @xmath276 $ ] or @xmath277 $ ] .
suppose @xmath278 .
the map @xmath279 is a bijection from representations of @xmath13 to the representations of @xmath280 ending in 00 while the map @xmath281 is a bijection from representations of @xmath282 to the representations of @xmath280 not ending in 00 .
hence the first equation holds .
a similar argument holds when @xmath283 .
the sequence @xmath229 is thus analogous to the alternative form of stern s sequence : @xmath284 for every word @xmath285 , we let @xmath286 denote the length of @xmath232 and define a point in the complex plane @xmath287 we form a graph @xmath288 by putting an edge between @xmath289 and @xmath290 for @xmath291 , @xmath292 .
this graph is illustrated in figure 4 below .
.,width=4 ] note further that @xmath293 iff @xmath294 and @xmath233=[\omega']$ ] .
hence , we may consistently assign the integer @xmath233 $ ] to each vertex @xmath289 of the graph .
this shows that @xmath295}$ ] is the number of downward paths from @xmath296 to @xmath289 and the graph can be thought of as a kind of hyperbolic pascal s triangle .
in fact , the portion between 0,01,010,0101 , ... and 1,10,101,1010 ,
... is really just the `` fibonacci diatomic array '' appearing in @xcite .
for @xmath297 a vertex of the fibonacci representation graph , let @xmath298 $ ] be the number of downward paths from the top vertex to @xmath297 . along the @xmath13th row of the graph @xmath288 , the function @xmath298 $ ] forms an increasing sequence of consecutive integers @xmath299 .
iterates of @xmath300 starting at 0 yields the sequence @xmath301 ( provable by induction ) .
hence the last value of @xmath298 $ ] in the @xmath13th row is @xmath302 .
since @xmath303 , the lemma follows .
a consequence is the following surprising formula : @xmath304 this graph has numerous interesting properties : * every quadrilateral in the closure of the graph is either a square or a golden rectangle . *
all the squares ( actually hexagons ) are congruent in hyperbolic space with area @xmath305 ( and , as hexagons , each edge has length @xmath305 ) .
the figure is thus an aperiodic tiling of part of the upper half - plane @xmath306 ( and can be extended to all of @xmath306 ) where all the tiles are congruent ! * the points along any row , when embedded in @xmath307 form part of a one - dimensional quasicrystal .
the lengths of the segments , appropriately scaled , form a word : @xmath308 , the `` fibonacci word '' . *
the vertices form a quasicrystal in @xmath306 . *
the graph is the cayley graph of the fibonacci monoid " @xmath309 . *
the graph can be constructed by the following recursive procedure starting with a single vertex ; from each of the latest generation of vertices , draw two edges going southeast and southwest respectively , connect if a hexagon can be formed .
repeat .
something new with respect to the study of @xmath229 is the development of an analog of conway s box function . for @xmath310 , define @xmath311 for @xmath312 , @xmath313 and @xmath314 where @xmath315 denotes mediant addition " .
note that @xmath316 and so , if @xmath317 , @xmath318 then @xmath319 and the first equation follows .
similarly , @xmath320 and the second equation follows . as a consequence , if , as @xmath321 , @xmath322 converges to @xmath323 $ ] , then @xmath324 converges to some value , say @xmath325 .
the function @xmath326\rightarrow[0,1]$ ] is increasing and continuous .
note , however , it is not strictly increasing .
for @xmath327 , @xmath328 the inverse of @xmath329 satisfies , on its irrational points of continuity , @xmath330 where @xmath62 has continued fraction decomposition @xmath331 .
recall that @xmath332 and so @xmath333 letting @xmath334 where @xmath62 is a point of continuity of @xmath335 , we see that @xmath336 we may then rewrite : @xmath337 and the theorem follows .
the function @xmath325 extends past @xmath338 but is no longer monotonic .
patterns can be found by looking at the `` crushed array '' which is found by stacking rows of terms @xmath339 sliding terms to the left on rows : @xmath340 the @xmath106th column satisfies : @xmath341 with common difference @xmath342 ( @xmath343 ) .
alternatively , @xmath344 ( a dying rabbit " sequence ) .
@xmath345 characteristic polynomial factors @xmath346
so every example is of the form @xmath347 .
hence , @xmath348 and @xmath349 are arithmetic sequences .
@xmath350 e.g. , ( * ? ? ?
* a023434 ) @xmath351 so every example is of the form @xmath352 where @xmath353 is the plastic constant " , 1.32471795 ... , the smallest pisot number , and @xmath354 are its algebraic conjugates .
such examples are always a constant plus a padovan sequence @xmath355 .
e.g. , ( * ? ? ?
* a000931 ) @xmath356 is always a constant sequence .
let @xmath15 be the number of terms in the zeckendorf representation of @xmath13 ( e.g. , @xmath357 ) .
equivalently , @xmath15 is the least number of fibonacci numbers that sum to @xmath13 .
this sequence , for @xmath358 , is [ a007895 ] and starts @xmath359 using notation of the previous section , we see that @xmath15 satisfies the recursion : @xmath360)=s_f([\omega ] ) , s_f([\omega 01])=s_f([\omega])+1\ ] ] which translates to @xmath361 where @xmath249 is the fibonacci shift " defined in section 4 ( just after proposition 4.1 ) .
the crushed array for this sequence is @xmath362 note that columns are constant and that the limiting row is @xmath363 . replacing @xmath14 by @xmath15 in binet s formula for
stern s sequence yields our third variant of stern s sequence : @xmath364 the sequence starts , for @xmath365 , @xmath366 it is always integral since @xmath367 is an algebraic integer in @xmath368 $ ] invariant under complex conjugation .
a crushed array for this sequence is : @xmath369 the first column , @xmath370 apparently satisfies the padovan recurrence : @xmath371 .
moreover , every column is apparently a dying rabbit " sequence : @xmath372 or , more precisely , if @xmath373 , then @xmath371 .
this is indeed the case which we now prove .
given a string @xmath376 of integers , let @xmath377 denote the reverse of string @xmath376 , let @xmath378 denote the string @xmath376 with 1 added to every integer , and let @xmath379 denote the string @xmath376 with 1 subtracted from every integer ( e.g. , if @xmath380 , then @xmath381 , @xmath382 , and @xmath383 ) . if @xmath384 , let @xmath385 finally , given strings @xmath376 , @xmath386 , we let @xmath387 denote the concatenation of the two strings and @xmath388 denote the pointwise difference ( e.g. , if @xmath389 and @xmath390 then @xmath391 and @xmath392 ) .
let @xmath393 be the number of terms in the zeckendorf representation of @xmath13 .
for any interval @xmath394 , let @xmath395 denote the _ string _
@xmath396 where @xmath397 and @xmath398 .
then @xmath399 where the difference of two strings is the string of differences .
since we will use this formula , we let @xmath400 so that @xmath401 .
* five inequalities : @xmath418 * the minimum of each row in the crushed array is the leftmost element .
( if true , then the last inequality above , @xmath419 , is true ) . *
if @xmath419 for all @xmath13 , then what do these numbers count ?
* the following sequences have crushed arrays with columns satisfying @xmath420 for given @xmath85 : @xmath421 is there a general principle at work in this progression
? is there a similarly defined sequence with @xmath422 for example ? | we present three analogues of stern s diatomic sequence . |
an isothermal supersonic shock with a mach number @xmath0 creates density enhancements of @xmath3 , where @xmath4 and @xmath5 are the densities of the post- and pre - shocked gas ( e.g. , spitzer 1978 ) . since molecular clouds
are turbulent and supersonic , it can be expected that their internal density structure is , at first order ( i.e. , neglecting gravitational or thermal fragmentation ) , a direct consequence of the fragmentation by the chaotic , supersonic velocity field ( see , e.g. , the reviews by vzquez - semadeni et al . 2000 ; mac low & klessen 2004 ; scalo and elmegreen 2004 and references therein ) , a process which has been called _ turbulent fragmentation_. thus , it is reasonable to expect that supersonic turbulence plays a crucial role in determining the mass distribution of dense cores . in fact , the gravoturbulent scenario of star formation suggests that the cores are formed by compressible turbulent motions inside molecular clouds and that some of those cores may become gravitationally unstable and form stars , while others will redisperse in the ambient medium ( sasao 1973 ; hunter & fleck 1982 ; elmegreen 1993 ; ballesteros - paredes , vzquez - semadeni & scalo 1999 ; klessen et .
al . 2000 ; padoan et al . 2001 ; padoan & nordlund 2002 , hereafter pn02 ) .
on the other hand , the mass distribution of young stars follows a well - known distribution called the initial mass function ( imf ) . for stellar masses
@xmath6 it shows a power - law behavior @xmath7 , with slope @xmath8 ( salpeter 1955 ; scalo 1998 ; kroupa 2002 ; chabrier 2003 ) .
understanding the origin of the imf is one of the fundamental goals of a complete theory of star formation .
although important progresses have been achieved on the observational determination of the imf , there are still several proposed models to explain it ( see reviews by meyer et al . 2000 ; mac low & klessen 2004 , and references therein ) , and there is no agreement in the community for a standard one .
one of the more recent models suggests that the imf properties are a direct consequence of the core mass distribution ( cmd ) .
observational works ( e.g. , motte et al .
1998 ; testi & sargent 1998 ) have reported a slope of the high - mass wing of the dense core mass distribution ( or mass spectrum ) that is similar to the slope of salpeter s imf , suggesting that those cores are the direct progenitors of single stars .
since stars are born from dense cores this idea is , in principle , tempting .
however , there is a large number of physical processes that may play an important role during the core fragmentation and the protostellar collapse ( see . , e.g. , klessen & burkert 2000 ; goodwin , whitworth & ward - thompson 2004 ; bate & bonnell 2004 ) .
these make it unclear whether a single core will give birth to one or more stars , and what determines the masses of individual stars within a single core .
some of these processes are : ( a ) the mass distribution of cores changes with time as cores merge with each other ( e.g. , klessen 2001 ; schmeja & klessen 2004 ) .
( b ) cores generally produce not a single star but clusters of stars , and so the relation between the masses of cores and those of individual stars is unclear ( e.g. , larson 1985 , hartmann 2001 , goodwin et al .
in addition , there may be ( c ) competitive accretion influencing the mass - growth history of individual stars ( see , e.g. , bate & bonnell 2004 ) , ( d ) stellar feedback through winds and outflows , or ( e ) changes in the equation of state introducing preferred mass scales ( e.g. , scalo et al . 1998 ; li , klessen & mac low 2003 ; jappsen et al . 2005 ; larson 2005 ) . besides the uncertainties mentioned above
, there are other important caveats when looking for a direct relationship between the cmd and the imf .
for instance , even though some observational and theoretical works for dense , compact cores fit power - laws in the high - mass wing of the cmd , the actual shape of those cmds is not necessarily a single - slope power - law , but a function whose slope varies in a more continuous way , frequently similar to a log - normal distribution . from a theoretical point of view , pn02 have argued that the mass distribution of dense cores generated by turbulent fragmentation follows closely the salpeter distribution of intermediate- to high - mass newborn stars , with a slope of @xmath9 and that the slope of the cmd depends only on the slope of the turbulent energy spectrum .
these results have been taken as a proof that turbulent fragmentation is essential to the origin of the stellar imf .
recent numerical studies by tilley & pudritz ( 2004 ) and li et al .
( 2004 ) have reported that the prediction of the pn02 model for the cmd agrees with the results of their simulations .
we consider that some cautionary remarks concerning these theoretical results are necessary .
first , the dynamical range of the cmd where a power - law with slope @xmath10 is appropriate is usually smaller than one order of magnitude .
second , the reported cmds are calculated at one single epoch , instead of being averaged over several timesteps . due to the fact that stochastic fluctuations might be non - negligible in a single frame ( see [ convergence_considerations : sec ] ) , determining cmds at one single time is prone to significant statistical fluctuations .
this is particularly relevant when the core statistics is small , as in li et al .
third , the behavior of the cmds presented in those works seems to be a continuous change in the slope of the cmd , from zero at the maximum of the histogram , to large negative slopes in the high - mass range .
in fact , gammie et al .
( 2003 ) mentioned that the high - mass wing of the core mass spectrum in their simulations has a slope that appears to be consistent with the salpeter law , although other non - power - law forms for the mass spectrum may well be consistent .
finally , in these works there is no systematic study of the dependence of the cmds with the mach number .
for example , tilley & pudritz ( 2004 ) present simulations with mach numbers of 2 and 5 , and indeed a close inspection of their fig .
17 shows that the cmds are different in both models , even though they can be fitted with the power - law with the slope of @xmath10 in a small dynamical range . as has been discussed by klessen ( 2001 ) and by schmeja & klessen ( 2004 ) , at a given mach number the shape of the mass spectrum of dense cores changes with the wavenumber at which turbulence is driven . in a similar way , the slope of the mass spectrum may very well vary with the root - mean - square ( rms )
mach number @xmath0 of the flow , which is related to the kinetic energy density of the system .
in fact , using numerical simulations , kim & ryu ( 2005 , see also cho & lazarian 2004 ) had found that the density power spectra of isothermal , turbulent flows , vary with the mach number , becoming shallower when the mach number is increased .
this flattening is the consequence of the dominant density structures of filaments and sheets .
this occurs because , in a supersonic turbulent flow , density fluctuations are built up through a sequence of jumps , produced by a succession of new shocks within previously shocked regions ( vzquez - semadeni 1994 ; passot & vzquez - semadeni 1998 ) .
thus , when increasing the mach number , the time frequency and strength of shocks inducing density enhancements also increase .
since the density power spectra becomes shallower ( kim & ryu 2005 ) , this means that the flow contains smaller and denser structures .
thus , it can be expected that the cmd depends on the mach number . in the present work we test this idea by using two different numerical schemes that model the interior of molecular clouds . in our simulations
we vary the value of @xmath0 from 1 ( trans - sonic turbulence ) to 15 ( supersonic , highly compressible turbulence ) , and show that the form of the cmd does indeed depend on the mach number .
the plan of the paper is the following : in [ simulations : sec ] we briefly introduce the simulations used in the present work , and discuss the method for finding clumps , showing that turbulent systems reach statistical equilibrium after one turbulent crossing time .
we also discuss , the importance of averaging the cmd over several timesteps in order to erase statistical fluctuations . in [ results : sec ] we present our results .
[ discussion : sec ] discusses the application of the pn02 model to isothermal hydrodynamic turbulent fields and the relationship between the core mass distribution and the imf . finally , in [ conclusions : sec ] , we summarize our results .
in order to study how the mass spectra of the cores depend on the mach number , we analyzed two different sets of simulations made with two distinct numerical schemes . the first set uses a total variation diminishing ( tvd ) method , and the second one a smoothed particle hydrodynamics ( sph ) approach .
all the simulations presented here are three - dimensional and isothermal .
we concentrate on small subregions within a much larger cloud .
therefore , periodic boundary conditions are adopted in both numerical schemes .
since we are interested in the fragmentation produced by the sole presence of turbulence , the simulations do not include self - gravity and magnetic fields .
being isothermal , the simulations are scale - free , and the renormalization is somehow arbitrary , but in principle they reproduce the behavior of molecular clouds of sizes ranging from less than 0.1 pc up to 1020 pc .
table [ tabla : models ] presents the properties of each run . in the first four columns we list the name of the run , its rms mach number ( average over several timesteps ) , the numerical method used , and the resolution .
the remaining columns will be introduced in [ convergence_considerations : sec ] .
c c c c c c c c c c tvd1 & 0.97 & tvd & @xmath11 & 8 & 1.2 & 5 & 0.4 + tvd3 & 2.87 & tvd & @xmath11 & 12 & 1.2 & 4.5 & 0.3 + tvd6 & 6.2 & tvd & @xmath11 & 18 & 1.2 & 6 & 0.3 + + sph3 & 3.4 & sph & 205,000 & 96 & 1.02 & 4.25 & 0.034 + sph6 & 5.7 & sph & 205,000 & 64 & 1.008 & 4.536 & 0.056 + sph9 & 9.1 & sph & 205,000 & 77 & 3.375 & 6.795 & 0.0405 + sph15 & 14.98 & sph & 205,000 & 50 & 1.05 & 4.725 & 0.075 + +
tvd - hr & 3 & tvd & @xmath12 & 12 & 1.2 & 4.5 & 0.3 + sph - hre & 3.8 & sph & 9,938,375 & 6 & 1.368 & 1.558 & 0.038 + the tvd scheme is a second - order accurate upwind scheme whose implementation for isothermal flows is described in detail by kim et al .
here we just mention that the simulations performed used a turbulence random driver following the method presented in stone , ostriker & gammie ( 1998 ) .
each driving velocity component is generated in fourier space , and its power spectrum has a peak at a large scale ( specifically at wavenumber @xmath13 , where @xmath14 is the one - dimensional size of the computational box ) . the kinetic energy input rate is adjusted in order to maintain a roughly constant rms sonic mach number .
the three simulations used in the present work have a resolution of 2563 cells , and their three - dimensional rms mach numbers are 1 , 3 and 6 . an additional high - resolution simulation ( 5123 cells ) at mach 3 has been performed , in order to check that the log - normal type shape of the cmd is not an artificial result of the low - resolution simulation .
the sph method is a lagrangian scheme in which the fluid is represented by an ensemble of particles , and flow quantities are obtained by averaging over an appropriate subset of the sph particles ( benz 1990 ) .
the method is able to resolve large density contrasts as particles are free to move , and so naturally the particle concentration increases in high - density regions .
thus , this method has spatially varying spatial resolution as a function of density ( see ossenkopf , klessen & heitsch 2001 ) .
details on the method can be found in klessen ( 2001 ; see also klessen , heitsch , & mac low 2000 ; and klessen & burkert 2000 ) . in the four sph simulations analyzed in this paper the three - dimensional rms mach numbers are 3.4 , 5.7 , 9 and 15 .
again , we drive turbulence at large scales , but in this case in a narrow range of wavenumbers @xmath15 . as in the case of the tvd method
, we performed an additional high resolution sph simulation ( 9,938,375 particles ) at mach 3.8 , in order to check that the shape of the cmd is not an artificial result of the resolution . in this case
we used the parallel code gadget ( springel , yoshida & white 2001 ) , in which the same turbulent driving scheme as in klessen ( 2001 ) has been implemented ( see jappsen et al .
2005 for details ) .
we analyze all the simulations over many equally spaced timesteps covering several turbulent crossing times , @xmath16 ( where @xmath16 is defined as by the one - dimensional size of our computational box @xmath17 divided by the rms velocity of the field i.e . ,
the mach number ) . to compute the cmd at each timestep , we use an adapted version of the williams et al .
( 1994 ) clumpfinding algorithm to find cores in the three - dimensional density data - cubes calculated with the tvd scheme .
this modified version allows us to find clumps even in large ( @xmath12 ) density cubes . for the particle data
, we used its equivalent clumpfinding algorithm , developed by klessen & burkert ( 2000 , see their appendix a ) . in both schemes
, we used logarithmic contours separated by half order of magnitude .
the parameters of the clumpfinding algorithms in each data set are somehow different : for the tvd runs , the lower density threshold is 2 times the mean density , and for the sph runs the lowest density contour is 10 ( @xmath18 1/50 the mean density ) . in both cases , the separation of contours is logarithmic , but while in the tvd runs the increase of mass is in steps of half order of magnitude , in the sph we adopted the same procedure as klessen ( 2001 ) , having only 10 logarithmic isocontours starting from the lowest contour and increasing the exponent by 0.5 .
although the details of the cmd obtained with different parameters of the clumpfinding algorithms may change ( e.g. , exact mass of each core ) , the trend of the cmd variation with mach number does not depend on the numerical method and/or the details of the clumpfinding scheme ( see [ results : sec ] ) . as the main purpose of this work is to observe this trend , the conclusions we reach are also independent of those choices . since the simulations start from arbitrary random initial conditions with density and velocity physically decorrelated , it is convenient , first , to investigate the time at which the fluid reaches statistical equilibrium .
we have studied the temporal evolution of the energy spectrum ( not shown here for reasons of space ) , and found that after one turbulent crossing time , the energy spectrum of each run does not vary substantially . to verify that , in addition to the energy spectra , at this time the simulations have reached statistical equilibrium , we show in fig .
[ all_vs_time : fig ] the mach number as a function of time ( _ upper panel _ ) , as well as the number of cores in the sph ( _ middle panel _ ) and tvd ( _ lower panel _ ) simulations . from this figure , where solid lines represent the sph data , and dotted lines represent the tvd data ,
it becomes clear that , in fact , after 1 @xmath16 , the rms mach number ( upper panel ) in the flow approaches to a constant value still shows variations of order 5% around the mean value , specially in the sph data .
this is due to compressibility of the medium and the adopted fixed large - scale driving pattern , as discussed by klessen & lin ( 2003 ) .
the deviations from the mean value between individual time frames decrease with decreasing rms mach number . ] .
the quoted numbers give the rms mach number @xmath0 , averaged over all frames for which @xmath19 @xmath16 .
similarly , in the middle and lower panels , where the number of cores is plotted as a function of time , it is again seen that the transients are given only for the first turbulent crossing time . after this time
, each frame may give statistical fluctuations of the order of 20% in the number of cores .
an important observation from the middle and lower panels in fig .
[ all_vs_time : fig ] is that , given the fact that important fluctuations in the number of cores are present , making conclusions based on a cmd computed from individual frames is risky . to prevent statistical bias when reporting the shape of the cmd , it is therefore important to calculate it as an average over at least a few turbulent crossing times .
it is convenient to note that , if the average were calculated over a fraction of turbulent time , the statistical fluctuation of the small cores will be erased , but the statistical bias of the larger cores will not , since in a fraction of one turbulent time the small cores may change considerably , but the larger ones not .
getting back to our table [ tabla : models ] , in its last four columns we show , for each simulation , the number of frames used in the present analysis ( column 5 ) , the initial and final times ( in units of @xmath16 ) used in the analysis ( columns 6 and 7 ) , and the time separation between frames , in units of turbulent crossing times ( column 8) .
figure [ energy : fig ] shows the time - averaged energy spectrum of each simulation . for comparison , the expected slopes for an incompressible flow ( @xmath20 ) and for a shock - dominated one ( @xmath21 ) are indicated by dashed straight lines .
the kinetic energy density is proportional to the total area below the curve .
this area grows as the @xmath0 increases .
note that the energy spectra for the sph runs have somewhat shallower slopes than those for the tvd runs .
however , this can be understood as a consequence of the considerable differences between the two numerical schemes .
the former is a particle - based method while the latter uses finite differences on a equally - spaced mesh .
both methods are complementary ( see also the discussion in klessen et al .
2000 ) , and we base our discussion solely on trends that are common in both sets of simulations . the fact that both numerical schemes , tvd and sph , give rise to the same behavior of the cmd with rms mach number ( see [ results : sec ] ) , reassures us to consider the observed effect of varying @xmath0 as a real physical trend .
figures [ tvd : cubes : fig]a , b show snapshots of the density field in two tvd simulations with @xmath221 and 6 , respectively .
the snapshots have been acquired at @xmath23 ( @xmath24 ) and @xmath25 ( @xmath26 ) .
it is evident that the medium is more strongly fragmented in the high mach number case .
a similar conclusion holds for the sph data cubes .
the interactions of supersonic shocks and shocklets in high - rms mach flows produce strongly localized density enhancements , and consequently give rise to high degrees of fragmentation _ even _ on scales much smaller than the turbulent driving scale . to quantify that further , we present in figs .
[ tvd : histo : fig ] and [ sph : histo : fig ] the resulting core mass spectra resulting from the tvd and the sph models , respectively .
the total mass in the computational box for all the models has been renormalized to 64 jeans masses .
this differs from the spherical definition , @xmath27 , by a factor @xmath28 . ] .
the cmd clearly varies with @xmath0 .
in particular , the total number of cores as well as the peak value of the histogram increase with @xmath0 . to illustrate this result more clearly we show in fig . [
number_vs_mach : fig ] ( a ) the total number of cores and , ( b ) the maximum value reached by the histograms as function of @xmath0 . in this figure ,
filled and open squares correspond to the sph and the tvd models , respectively .
solid lines in each panel are least - square fits . as a consequence of mass conservation (
recall that each tvd and sph simulation has the same total mass ) the typical core mass shrinks as the number of cores increases .
this result is shown in fig .
[ mass_vs_mach : fig ] , where we plot ( a ) the mass of the most massive core and , ( b ) the mass at which the histogram peaks , both as function of @xmath0 .
a point of concern is whether our results depend on the resolution of the simulations . in fig .
[ cmd_hr : fig ] we show the cmd resulting from the two high - resolution ( hr ) simulations . recall that the hr - tvd runs has @xmath12 pixels , while the hr - sph runs has 9,938,375 particles .
this means that the ( spatial ) resolution is 8 times larger in the tvd - hr case , while the mass resolution in the sph - hr case is almost 50 times larger . from this figure
we stress that the shape of the cmd is , indeed , not a power - law function , but something more similar to a log - normal function , for which the slope varies from zero at the maximum of the distribution to large negative values .
recall at this point that , since the clumpfinding algorithm in the sph runs has been used with a smaller density threshold , the typical core mass in the sph models exceeds that in the tvd case
. however , the general trend of increasing number of cores and decreasing core mass with increasing rms mach number is independent of the used numerical method as well as the details of the clumpfinding scheme .
in this section we examine the implications of the results presented in [ results : sec ] , in the context of previous work on the cmd and its relationship with the stellar imf .
we find that the cmd is not a power - law and has a shape that depends on the rms mach number .
this fact contradicts previous results stating that , in the magnetic case , the cmd follows a power - law whose slope depends on the power index of the kinetic energy spectrum ( pn02 ) . as has been mentioned
, our simulations do not include magnetic fields and this is an important difference with the work of pn02 .
however , this fact is not the reason for the differences between the shape of cmd reported in previous section and the power - law they obtain .
indeed , it can be shown that a similar analysis to the one performed by pn02 but for the non magnetic case , also leads to a power - law in which the rms mach number of the flow does not enter in the functional form of @xmath29 .
it only determines the normalization factor , which , in turn , implies that larger @xmath30 leads to larger number of cores .
although this fact is consistent with our numerical results ( see [ results : sec ] ) , our results show that the actual shape of the cmd does vary with @xmath0 .
we identify the main reason why the model by pn02 does not include the dependency of the shape of the cmd on @xmath0 as hinted before : density fluctuations ( cores ) in a purely turbulent fluid are built - up by a statistical superposition of shocks .
this means that mass and density contrast in individual cores grow by a _ sequence _ of density jumps with pre - shock densities that can significantly deviate from the mean value @xmath5 ( see , e.g. , smith et al .
in addition , in a turbulent flow with characteristic mach number @xmath0 , the mach numbers of the individual shocks may deviate considerably from the mean value @xmath0 .
this translates into a large scatter of the larson ( 1981 ) @xmath31 relationship , which has been observed both in theoretical works ( e.g. , vzquez - semadeni et al .
1997 ; ostriker et al .
2001 ; ballesteros - paredes & mac low 2002 ) , as well as in works that include more recent sets of observations includding different tracers ( see garay & lizano 1999 ) . instead , the pn02 model and its hydrodynamic pendant introduced above , assumes that all cores form in one single event with fixed mach number and initial density . other caveat in the pn02 model is that stream collisions form sheets , not cores .
thus , the size @xmath32 of the shock - bounded region may not be a good representative of the size of cores . if cores are generated not by single shocks , but by statistical superposition of subsequent shocks of the turbulent flow , the dependency of the density structure on the rms mach number may have important implications for the fragmentation behavior of molecular clouds , clumps and their cores .
for instance , the internal density structure of a dense core ( even if it is subsonic ) may differ depending whether it is embedded in a highly supersonic , turbulent molecular cloud , or in a more quiescent one ( see , e.g. , ballesteros - paredes et al .
2003 , klessen et al .
2005 ) . for high mach numbers ,
the high- frequency part of the turbulent energy cascade lies in the supersonic regime and , hence , is able to produce strong density fluctuations on intermediate to small scales .
the amount of energy at the largest scales , where most of the power is , thus , should define the fragmentation on all scales , included those where , the non - thermal motions became subsonic .
thus , we speculate that cores of a given mass in more turbulent clouds ( e.g. , orion ) will have more substructure than cores of the same mass in less turbulent clouds ( e.g. , taurus ) . as mentioned in the introduction , there have been various attempts to make a direct connection between the cmd and the stellar imf . however , the result found in the present work that the cmd has an strong dependence on the mach number implies that the relationship between these two distributions is not immediate .
it is thus important to discus about the possible reasons of this fact in the context of turbulent fragmentation .
the supersonic turbulent motions ubiquitously observed in interstellar clouds , suggest that cores form in first place via shocks , with gravity taking over in the densest and most massive regions . once gas cores become gravitationally unstable , collapse sets in , but the details of that collapse are not well understood , yet .
it is tempting to suggest that turbulence plays a critical role in determining the stellar imf ( see , e.g. , the discussion in ballesteros - paredes 2005 ) . in this case
, the imf will directly follow from the cmd if individual molecular cloud cores map one - to - one onto stars .
however , there is a large number of physical processes that may play an important role during protostellar collapse .
thus , the main problem relating the cmd with the imf is that a core with a given mass will not necessarily form a star of the same mass , or of a mass that is proportional to the core mass .
however , even considering that these effects do not play an important role , and that every core of mass @xmath33 does indeed produce a star whose mass is directly proportional to @xmath33 , the numerical results presented in this paper suggest that turbulent fragmentation is not the sole agent defining the shape of the imf , since the shape of the cmd seems not to be a power - law and it varies ( as well as the cmd normalization ) with the total amount of kinetic energy .
thus , although turbulent fragmentation may play an important role in the determination of the cmd , it is difficult to argue that it is the main ingredient for determining the imf . in a supersonic turbulent fluid , where the gas is compressed into filaments and sheets ,
the initial cmd may depend on the parameters of the turbulent gas , however there is still room for other processes to take over and modify the initial cmd in its way to an intermediate mass distribution , and finally to the imf . finally , it has to be mentioned that , although it is generally accepted that the imf has a slope of @xmath10 at the high - mass range , the universal character of the imf is still a debated issue .
for instance , scalo ( 1998 ) notes that observations taken at face value reveal variations in the imf between different star clusters .
kroupa ( 2001 ; 2002 ) , on the other hand , argues that these variations can be explained by stellar dynamical processes during the early phases of star - cluster evolution , like mass segregation and the preferred evaporation of low - mass stars . on similar grounds ,
also elmegreen ( 2000 ) argues in favor of a more universal imf , although he recognizes statistical deviations .
using two different numerical schemes ( tvd and sph ) , we have shown that the core mass distribution ( cmd ) or core mass spectrum from purely turbulent ( i.e. without self - gravity and magnetic fields ) fragmentation follows a distribution with a continuously varying slope , not necessarily a single power - law .
the parameters of this distribution depend on the kinetic energy density in the system as characterized by the rms mach number of the flow .
interactions of shocks in high @xmath0 flows are able to produce strong density fluctuations on small and intermediate scales . we thus speculate that quiescent cores embedded in highly supersonic molecular clouds should exhibit more internal structure than quiescent cores embedded in more quiescent clouds . as consequence , for a molecular cloud of a given mas , the resulting resulting cmd is characterized by a large number of low - mass cores when the mach number is high .
low-@xmath0 flows , on the other hand , are characterized by small density contrasts and the cmd features fewer cores with on average larger masses .
we conclude that the rms mach number of the turbulent flow is a very important parameter that determines the shapes of cmds .
our calculations indicate that the exact shape of the cmd is not universal , but changes with the mach number .
this calls the existence of a simple one - to - one mapping between the purely turbulent cmd and the observed stellar initial mass function ( imf ) into question . finding the true relation between cmd and imf instead is a very complex and difficult task , and requires to take additional physical processes into account , such as protostellar feedback from new - born stars ( bipolar outflows , winds , and radiation ) , subfragmentation of individual collapsing protostellar cores into binary and low-@xmath34 multiple systems , competitive accretion to name just a few .
we thank e. vzquez - semadeni and s. kitsionas for fruitful discussions .
this work was supported in part by conacyt grant 27752-e .
the work by ag was supported by the dgapa grant papiit - in114802 .
j.s.k . was supported by the astrophysical research center for the structure and evolution of the cosmos ( arcsec ) of the korea science and engineering foundation through the science research center ( src ) program .
r.s.k . and a.k.j .
were supported by the emmy noether program of the dfg under grant kl1358/1 .
the tvd simulations were performed on the linux clusters at kasi and crya , with funding from kasi and arcsec , and conacyt grant 27752-e , respectively .
the sph calculations have been performed at the university of california at santa cruz as well as the pc cluster of the astrophysical institute potsdam .
this work has made extensive use of nasa s astrophysics abstract data service and lanl s astro - ph archives . | we investigate the core mass distribution ( cmd ) resulting from numerical models of turbulent fragmentation of molecular clouds .
in particular we study its dependence on the sonic root - mean - square mach number @xmath0 .
we analyze simulations with @xmath0 ranging from 1 to 15 to show that , as @xmath0 increases , the number of cores increases as well while their average mass decreases .
this stems from the fact that high - mach number flows produce many and strong shocks on intermediate to small spatial scales , leading to a highly - fragmented density structure .
we also show that the cmd from purely turbulent fragmentation does not follow a single power - law , but it may be described by a function that changes continuously its shape , probably more similar to a log - normal function .
the cmd in supersonic turbulent flows does not have a universal slope , and as consequence , cast some doubt on attempts to directly relate the cmd to a universal initial mass function . '
# 1#1i"13i # 1@xmath1 # 1@xmath2 # 110@xmath1 # 110@xmath2 |
@xmath3one of the most exciting recent discoveries in astronomy is the strong evolution and co - evolution of both the star - formation rate density , and black hole number density from @xmath1@xmath213 to today ( see e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
particularly interesting are composite or transition objects where quasars co - exist with powerful starbursts ( see e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
such objects are likely to be much more common at @xmath1@xmath22 , the peak epoch of both the quasar number density and the star - formation rate density .
in particular , recent _ spitzer _ space telescope @xcite results suggest that the ulirg number density was more than 200-fold greater at @xmath1@xmath22 than it is today ( for a review see * ? ? ? * ) .
this strong evolution of the luminosity function means that many more extremely luminous sources ( @xmath4@xmath2@xmath5@xmath6 ) have been discovered than are observed locally . by necessity therefore these extreme sources are typically compared with local ulirgs which are in fact much less luminous ( typically @xmath4@xmath2@xmath7@xmath8@xmath6 ) .
some of these @xmath1@xmath22 sources derive their luminosities from powerful obscured quasars , but others appear to be scaled - up versions of the local largely starburst - dominated ulirgs ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . due to the dearth of such sources nearby the physical conditions such as the dust composition at these highest luminosities is poorly understood .
one approach is to determine whether or not the detectability and relative strength of absorption and emission features arising from different dust components are consistent with those of local ulirgs in order to test how similar are their dust properties .
the mid - ir regime is well suited to such a study as it contains a wealth of molecular absorption features . by far
the strongest is the 9.7@xmath0 m silicate absorption feature ( si - o stretching mode ) .
this is thought to sample the large silicate grains , which are frequently coated in various ices including water .
these ices can be seen in the h - o - h bending mode ( 3.05@xmath0 m ) and the o - h stretching mode ( 6.15@xmath0 m ) . in more extreme environments such as observed toward the galactic center ,
the degree of hydrogenation is such that prominent hydrocarbon ( hac ) features are observed ( e.g. * ? ? ?
* ; * ? ? ?
the strongest of these is the 3.4@xmath0 m c - h stretching mode , but also includes the 6.85 and 7.25@xmath0 m c - h deformation modes .
studies of these features in external galaxies and in particular ulirgs are still quite challenging .
the first such study was that of @xcite who , out of a sample of 100 sources with _ iso_-sws spectra , detected the 6@xmath0 m water ice absorption in 18 galaxies ( including most of the ulirgs ) .
indeed ulirgs show the deepest silicate absorption features ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , and hence are the best candidates to look for the other mid - ir molecular absorption features .
the subaru study of the 34@xmath0 m spectra of 37 ulirgs @xcite supplemented by the @xmath2530@xmath0 m _ spitzer _ infrared spectrograph ( irs ; * ? ? ?
* ) study of local ulirgs @xcite , and the vlt study of the 34@xmath0 m spectra of 11 more ulirgs by @xcite all show molecular absorption features . over the 35@xmath0 m regime ,
the vlt isaac spectra of 5 ulirgs were presented in @xcite .
most recently _ akari _ spectra of 45 local ulirgs provide the first continuous 2.55@xmath0 m spectra @xcite of a large number of ulirgs and serve as an excellent local basis for comparison with our high-@xmath1 spectra .
lastly , co gas has a prominent 4.67@xmath0 m absorption feature , which has been detected in several deeply obscured local ulirgs @xcite .
an archival study of _ iso _ spectra of seyfert galaxies showed no sign of this feature , presumably due to different obscuration geometries , i.e. dust torus vs. buried nucleus @xcite .
all these absorption features arise in slightly different conditions and hence sample different obscuring layers @xcite .
however , how these features arise and their relative strengths are still not well understood even locally .
only a few dust models that address the relative strength of the different absorption features have been published ( e.g. * ? ? ?
* ; * ? ? ?
* ) , which along with laboratory ice spectra ( e.g. * ? ? ?
* ) help in interpreting the results .
most of what we know about interstellar ices comes from observations of embedded protostars ( for a review see * ? ? ? * ) which in may also be analogs to our most highly obscured @xmath1@xmath22 sources than local ulirgs . in this paper , we present unprecedented signal - to - noise rest - frame @xmath228@xmath0 m spectra of @xmath1@xmath22 ulirgs .
a spectrum of a @[email protected] galaxy with similar signal - to - noise is presented in @xcite , but starting at restframe @xmath25.5@xmath0 m .
spectra extending to the rest - frame @xmath22@xmath0 m have several advantages which will shed light on the physical conditions of these distant systems : 1 ) the continuum arises from hot dust near the power source , 2 ) this regime includes the 3.3@xmath0 m and 6.2@xmath0 m pah features both isolated and less severely affected by the mid - ir continuum model than the longer wavelength pah features , 3 ) dust absorption features including water ice at 3@xmath0 m and 6@xmath0 m , several hydrocarbon ( hac ) features , and 4 ) the warm co gas 4.67@xmath0 m absorption feature .
in this paper we concentrate on the various absorption and emission features in these spectra and compare our results with the available local ulirg data . throughout this paper
we assume a flat universe with @xmath9=0.27 , @xmath10=0.73 , and @xmath11=71kms@xmath12mpc@xmath12 @xcite .
@xmath3our group has been engaged in a multifaceted study of the properties of high-@xmath1 ulirgs , including low - resolution irs spectra , _ hst _ imaging , @xmath13-ray imaging , and keck spectroscopy .
we have @xmath2200 irs spectra of sources with @[email protected] , selected in the _ spitzer _ first look survey field , of which @xmath270 sources are at 1.5@xmath16@[email protected] , and likely @xmath4@xmath2@xmath5@xmath6 @xcite .
the analysis of the first 50 sources ( go1 program # 3768 ) of this sample has already been completed @xcite .
the remaining 150 sources ( go2 program # 20629 ) give a largely uniform , flux - limited sample ( dasyra et al . in prep . ) .
both the g01 and g02 samples have a limit of @[email protected] , but our g01 sources have two additional selection criteria : @xmath17(24@xmath0m)/@xmath17(0.64@xmath0m)@xmath151.0 and @xmath17(24)/@xmath17(8)@xmath150.5 .
the later selection cut means that the g01 sources tend to be fainter at 8@xmath0 m ( rest - frame 2@xmath0 m ) than the g02 sources .
we adopt three selection criteria for our follow - up program : 1 ) redshift , 2 ) brightness , and 3 ) silicate feature depth .
we adopt the redshift range 1.6@xmath16@[email protected] .
these boundaries are determined by practical reasons : beyond @[email protected] , irs spectra poorly sample the 9.7@xmath0 m silicate feature , while sources below @[email protected] require an observationally - expensive additional irs module ( sl2 ) . for the brightness , we simply require an 8@xmath0 m detection ( @xmath1860@xmath0jy ) or else the sl spectra would be too difficult to obtain .
the last criterion was most difficult to define .
as this is a largely exploratory program , we did not want to restrict ourselves to only the most extremely obscured sources ( see e.g. category 3 ; * ? ? ?
* ) , but also needed the sources to have relatively strong silicate absorption features since , as expected , only sources with silicate absorption have been observed to have ice absorption features @xcite .
we selected , @[email protected] ( based on the galactic center extinction curve used in sajina et al .
2007a ) as a rather generous lower limit . in terms of the observed 9.7@xmath0 m depth ( i.e. @xmath20
as is more commonly used , this translates to @xmath21@xmath22@[email protected] . for the sake of comparison with previous work , throughout this paper we use
observed optical depths ( see also [ sec_optdepth ] ) .
out of 52 sources in the go1 sample , 6 sources satisfy the above conditions . of the g02 sample ,
5 sources satisfy the conditions .
however , as we used very preliminary data reduction of this sample , when doing the initial sample selection , one of the sources that meets the criteria ( mips509 ) was not included , and another ( mips22633 ) was observed although its silicate feature depth is too low compared with the rest ( @[email protected] ) . in total ,
our sample consists of 11 sources .
only 3/11 of our sources have optical / near - ir redshifts ( see yan et al . 2007 and sajina et al . 2008 ) .
the rest are typically based on the 10@xmath0 m silicate absorption feature and hence are believed to have an uncertainty , @[email protected] ( see yan et al .
2007 ) . hence , for the sources without accurate redshifts ,
we look for example for absorption features within the range @xmath24@xmath25@xmath23 .
this constraint makes the reliability of single marginally - detected features doubtful . in cases of unambiguous and/or multiple absorption feature detections ,
these can be used to improve our redshift estimates . as the rest - frame peak of the molecular absorption features can also vary
for example the 3@xmath0 m ice feature can peak anywhere in the range @xmath23.05 3.1@xmath0 m @xcite , an uncertainty of up to @[email protected] remains .
in terms of both the molecular absorption features and hot dust continuum , we are interested in obtaining deeper spectra in the rest - frame @xmath228@xmath0 m spectral range . at the redshifts of our targets
this means using the _ spitzer _ irs s sl1 and ll2 modules .
one of our targets , mips22432 is very bright ( @xmath26=0.5mjy ) , and was already observed in sl1 as part of our go2 program .
its signal - to - noise is sufficient for our purposes ( see below ) and hence does not need to be re - observed in sl1 .
both mips22432 and mips15880 have sufficiently high signal - to - noise and do not need new ll2 observations .
lastly , mips8392 , was already observed in sl1 as part of the g01 program # 15 ( p.i .
houck ) and hence its sl1 observation was removed from our observations list .
unfortunately , the depth of the previous observation is 3@xmath27 lower than we needed here , but this is still considered a duplicate observation ( the limit is 4@xmath27 ) . the depth of the earlier observation is insufficient for our purposes hence it can serve only to constrain the continuum levels , but is not useful in looking for absorption features .
table [ table_obs ] gives a summary of the observations covering three _
spitzer _ observing cycles ( g01,g02 , and g04 ) .
the full go1 sample and observations are discussed in detail in @xcite .
the full go2 sample and observations are discussed in @xcite . the go4 observations ( pid#40793 )
are presented here for the first time . combining all these programs
, each source has typically @xmath22.5hrs on source of irs sl1 and ll2 observations , although brighter sources have less exposure time as can be seen in table [ table_obs ] .
the sl1 observations were mostly in mapping mode in order to improve the signal - to - noise ratio in these very faint sources . for our sources
, we adopted the irac and/or vla positions which are the most accurate available and sufficiently accurate for the sl1 module which has the narrowest slit width ( 3.6 compared with 10.5 for ll2 ) .
we use irslow , a data reduction code for irs low resolution observations custom - written by dario fadda ( see fadda et al .
2010 in prep . for further details ) .
the data reduction steps include bad pixel removal , background subtraction , co - adding different positions ( such as the 2 nod positions in stare mode , or the 4 positions we typically used in map mode ) , and finally extracting the spectra using optimal extraction .
bad pixels are determined as outliers across the available frames . for the sl data for example ,
fainter sources have 40 frames per source , but the bright sources have only 6 frames per source .
therefore this bad pixel flagging is less efficient for our bright sources with very few individual frames leading to somewhat noisier spectra than the nominal difference in exposure time .
the code is iterative in that after the first background subtraction , any sources present are masked , and a second background subtraction is done .
in addition , we test the reliability of any features we see , by displaying all points that go into the spectral extraction and manually clipping any obvious outliers ( on top of the bad pixel flagging of the code ) that might bias our results .
a few spurious spikes are removed in this way , but this is mostly a concern for narrow features of interest such as the 3.3@xmath0 m pah feature . in all cases , where we believe that this feature is or might be present we have used this option of the code to check for any possible outliers . for the broad water ice or hac absorption features , the quality of the background subtraction is more of a concern than any remaining outlier pixels .
below we describe our background subtraction quality test . as a double check on our per pixel errors ( and the background subtraction )
we also extract sky spectra from a position in the data image which does not include a source . the same psf is used for the off - source spectra extraction as the source spectra .
these sky spectra allow us to ascertain that the mean background level is zero and that no residual gradients are present .
the mean rms values from these sky spectra are found to be comparable with the typical per pixel errors of the extracted spectra ( but @xmath230% worse in the case of the two sl stare mode observations due to the less efficient bad pixel flagging ) . for the sl1 spectra the rms is typically @xmath22050@xmath0jy ( with the lower values typical of the fainter sources and higher values for the brighter sources ) with continuum snrs of @xmath248 . for the ll2 spectra the rms is typically @xmath260@xmath0jy with continuum snrs ranging from @xmath23 for mips22530 to @xmath225 for mips22303 , with typical values of @xmath210 .
we tested irslow s spectral extraction , by running spice after the background subtraction and obtained fully consistent results .
we find that our old reduction of the ll2 data for the go1 sources @xcite is frequently lower than the new data by up to 30% ( somewhat worse for mips464 ) .
however , when re - reduced the same way as the new data ( see above ) the two sets are consistent .
we attribute the difference to improved background subtraction and calibration . for consistency
we have re - reduced all available data ( including ll1 , ll2 and sl1 ) before co - adding into the final spectra .
figure[fig_specs ] shows the eleven spectra of our sources derived by co - adding all available data ( see table [ table_obs ] ) . for the ll2 region ( @xmath28@xmath21421@xmath0 m ) , the spectra represent the weighted mean of the old and new data .
the panels below show closer views of the spectra around the principal water ice absorption features at 3@xmath0 m and 6@xmath0 m .
we detect the continuum down to @xmath22.5@xmath0 m allowing us to look for the various molecular absorption features expected in this regime .
indeed , we see a number of interesting features in the spectra , the vast majority of which can be identified with the known molecular features including water ice , pah , hac , and co. figure[fig_specs ] also shows the sl1 and ll2 sky spectra for these sources ( i.e. spectra taken off - source , but with the same psf ) .
as expected , these sky spectra have a median of essentially zero , but do show some ` bumps ' and ` dips ' which in addition to the background noise , likely also include some low - level rogue pixels that are not sufficiently strong outliers to have been removed . such spurious features , however , could influence our source spectra as well .
essentially there are three considerations for believing the reality of a given feature : 1 ) whether or not it is stronger than the amplitude of the features in the sky spectra , 2 ) whether or not it has consistent redshift , and 3 ) whether or not it has roughly the expected shape ( i.e. width ) .
we address these issues in
[ sec_optdepth ] . before hand however , we examine briefly the hot dust continuum shape of the sources .
m regime .
here we show all sources showing evidence of features in this regime .
the crosses represent the actual spectra , while the solid red line is the spectra smoothed by a factor of 5 .
the dashed line is the best - fit continuum .
[ buried_conts],height=529 ] we can see in figure[fig_specs ] that there is considerable variety in the continuum shapes for our sources .
this likely is indicative of differences in obscuration geometry . to address this
fully requires comparison with radiative transfer power source+obscuration models which is beyond the scope of this paper .
here we merely wish to quantify the basic characteristics of these spectra such as spectral slope and monochromatic luminosity and compare with those of local ulirgs .
we adopt the standard measure of spectral slope which is a power law spectral index ( i.e. @xmath29@xmath30@xmath31 or @xmath32 ) .
we define continuum regions free of major features ( 2.6@xmath16@[email protected]@xmath0 m and 3.8@xmath16@[email protected]@xmath0 m ) .
we fit the above model to the spectra that fall within the continuum regions ( see figure[buried_conts ] ) . for mips464 , mips15880 , and mips22432
, their lower redshift means that the spectra do not reach rest - frame 2.8@xmath0 m . for these sources we adopt @[email protected]@xmath0 m as the continuum region
this assumes there is no water ice absorption , which is not observed for mips464 and mips16152 and is not possible to constrain for mips15880 or mips22432 .
we find fairly red spectra in this regime with slopes in the range @xmath34@xmath224 . to compare our values with those of local ulirgs without the bias of different continuum definitions , we take the ulirg akari spectra from @xcite and apply the same continuum fitting to them .
figure[fig_cont ] shows the distribution of the 34@xmath0 m slope for our sample compared with local ulirgs .
we find that the bulk of our sample has steeper slopes than the local ulirgs .
@xcite argue that a red 34@xmath0 m slope is an indicator of an agn dominant in the mid - ir ( although not necessarily bolometrically dominant ) .
as defined , @xmath34 is related to the slope @xmath35 as defined by @xcite via @xmath35=@xmath34 - 2 . using this we can translate the @xcite criterion to : @xmath34@xmath152 sources
are predominantly agn - dominated .
this is satisfied by most of our sources , which is not surprising as most are already believed to be agn - dominated @xcite .
mips22530 is our strongest - pah , starburst - dominated source and it has the flattest spectrum with @xmath34=1.9 .
m water ice and 3.4@xmath0 m hac opacity for the sources where these are observed .
we have applied a binning of 5 pixels to all spectra to increase the clarity and effective signal - to - noise ( @xmath36 ) .
the dashed and dotted lines are the average profiles of the 3.05@xmath0 m water ice and 3.4@xmath0 m hac features for galactic center stars from @xcite and the solid curve is the sum of the two .
note that these are composite features and hence a wide range of specific shapes is expected ( e.g. * ? ? ?
* ; * ? ? ?
mips16080 and mips22633 show evidence of both ice and hac features but the combined profile is inconsistent with expectations if these are the only contributors .
the observed spectra could be explained by the presence of the pah 3.3@xmath0 m emission feature as well ( e.g. * ? ? ?
[ hac_det],height=529 ] we follow the same procedure for the rest - frame @xmath257@xmath0 m spectra , where we define as continuum regions 5.0@xmath16@[email protected]@xmath0 m and 6.7@xmath16@[email protected]@xmath0 m .
the slopes there tend to be considerably flatter ( except where strong - pah contaminate this slope ) as can already be seen visually in figure[fig_specs ] .
lastly , we compute the monochromatic 3@xmath0 m and 5.5@xmath0 m luminosities based on these fits
. table[table_slopes ] gives the continuum parameters ( luminosites and slopes ) of our sources . in the previous section
, we discussed the determination of the continuum levels for our sources .
figure[buried_conts ] shows the best - fit continua for the sources with water ice and/or 3.4@xmath0 m hac absorption ( mips464 , mips16059 , mips16080 , mips22303 , and mips22633 ) .
this figure also shows that the absorption features are not driven by smoothing over a single outlier / spike .
none of the sky spectra of these sources show broad features of such strength which also excludes them being background subtraction artefacts .
after the continuum fitting , the optical depth of the various features is then determined based on : @xmath37=ln@xmath38 .
we should stress that here we use the convention of calling this observed feature strength an optical depth , even though the true optical depth is dependent on the exact radiative transfer solution .
in particular , in the case of the realistic clumpy dust distribution the two do not translate in any straightforward way @xcite .
as discussed previously , the reliability of weakly detected features is complicated by the frequently uncertain redshifts ( except for mips16059 , mips16080 , and mips22530 which sources have keck rest - frame optical spectra ) . for mips464 , mips22303 , and mips22633 ,
the observed features are of a redshift consistent with the silicate feature and sufficiently strong in themselves to make the feature identification convincing .
table[table_taus ] gives a summary of the detected features as well as the updated redshifts .
the uncertainties are estimated based on @xmath39@xmath251@xmath40 .
we ignore the uncertainty in the continuum level . in [ sec_optdepth ] we claim detections of water ice and/or hac absorption for 5/11 of the sources in our sample ( see table [ table_taus ] ) .
another test of the reliability of these features is to compare their observed profiles with the average profiles of these features as determined for galactic sources . in figure[hac_det ] we show the same five sources as in figure[buried_conts ] but with the continuum subtracted , and the average profiles of the 3.05@xmath0 m water ice and 3.4@xmath0 m hac features from @xcite overlaid .
the continuum is the same as determined in the previous section , although it should be kept in mind that there is some uncertainty in the continuum - fit parameters , as the fit is done on fairly small sections of essentially feature - free spectra .
we find that there is broad agreement although our spectra show noise about these profiles .
the @xmath23.0@xmath0 m absorption feature in mips16059 is at somewhat lower wavelengths than expected for water ice and therefore we treat this as an upper limit only and not as a detection .
overall , the most significant deviation is observed for mips16080 , mips22303 and mips22633 all of which show an excess at 3.3@xmath0 m in addition to the likely water ice and/or 3.4@xmath0 m hac absorption . to test the possibility of this being due to the 3.3@xmath0 m pah , in figure [ akari_comparison ]
, we compare the mips22633 and mips16080 spectra with reasonably closely matching local ulirgs where we degrade the resolution of the akari spectra to match that observed .
both local comparison sources show 3.3@xmath0 m pah @xcite , although in iras 19254 - 7245 it is much weaker than in iras14121 - 0126 .
we find that indeed the 3.3@xmath0 m excess can very well be due to pah emission especially for mips22633 .
for mips16080 the comparison is less convincing and therefore we treat its 3.3@xmath0 m emission as an upper limit only .
= 0.151 ) .
the later has been degraded to the spectral resolution of the mips22633 spectrum .
the combination 3.3@xmath0 m pah+3.05@xmath0 m water ice absorption for mips22633 is quite feasible .
[ akari_comparison],height=377 ] for mips22303 we find an excellent local analog in the _ iso_-sws spectrum of the galactic center source gc ring ne ( see figure[fig_mips22303 ] ) .
this comparison suggests , that mips22303 is well described by a combination water ice+3.3@xmath0 m pah without any clear need for hac absorption ( as shown in figure[hac_det ] ) . for this source
therefore , we only have an upper limit on the hac absorption . one of the goals of this program was to try to detect the warm co gas absorption feature at 4.67@xmath0 m . unfortunately , in several cases , rest - frame 4.67@xmath0 m falls right in the region of where we stitch together the sl1 and ll2 orders ( yellow band in figure [ fig_specs ] ) making any features observed there unreliable .
sources with @[email protected] ( i.e. where this region is entirely contained in the ll2 spectra ) or with @[email protected] ( i.e. where this region is entirely contained in the sl1 spectra ) are the only feasible ones to look at
. there are a total of 7 sources that satisfy these criteria . of these
, three sources show signs of absorption in this regime : mips464 , mips16059 , and mips22432 ( see figure[fig_specs ] ) .
these features are fairly convincing in themselves ( compared with the background ) , but do they agree with the expectations for the co feature ? in figure[fig_co ] we compare their smoothed normalized flux spectra ( i.e. data / cont ) , with the same for the hyperluminous galaxy f00183 - 7111 and the iras00397 - 1312 @xcite . while the nominal position of the line is at 4.67@xmath0 m , the peak absorption in our sources ( assuming it is co ) is typically blueward of that .
as the co gas absorption feature is composed of two broad branches ( r and p ) , if one of these is stronger than the other it results in the observed shift . a stronger r branch ( shorter-@xmath33 ) absorption
is also observed in iras08572 + 3915 @xcite and iras00397 - 1312 @xcite .
note that for all these galaxies , the redshift is well established on the basis of other features , without any additional shifts required to match the co feature .
for mips464 the redshift is based on the silicate absorption plus 3.4@xmath0 m hac feature . for mips16059 ,
the redshift is based on keck spectra , but is also consistent with the silicate and 3.4@xmath0 m features . for mips22432 , the only strong - pah source of the three
, the redshift is based on the 6.2@xmath0 m pah feature .
the observed strength of the co feature for these three sources ( the strongest in our sample ) is @[email protected] ( as discussed in 4.3 this is just the observed feature strength and does not translate to intrinsic optical depth ) .
this is somewhat weak compared with the co feature strengths observed by @xcite , which range from @xmath20.12.2 .
the relative weakness of our features can also be seen in figure[fig_co ] . an possible explanation for both the observed asymmetry and the overall weakness of the co feature might be the effects of a starburst - powered continuum originating from material outside the co absorption layer .
mips222432 ( a strong - pah source ) has a clear starburst - component and as we show in 4.8 , a starburst component is likely present in all sources even if not individually observed .
we look for the 3.3@xmath0 m and 6.2@xmath0 m pah features which are not as confused with the continuum model as for e.g. the strongest pah feature at 7.7@xmath0 m @xcite , and hence might reveal the suspected presence of star - formation even in sources dominated by nuclear activity . in @xcite
we showed that the stacked spectra of weak - pah ( based on the shallower spectra ) sources show pah emission .
there are several cases of weak - pah sources that have cold dust detections @xcite . in addition two our sources , mips22432 and mips22530 were identified as strong pah sources based on our previously defined criteria @xcite .
mips22530 was also shown to be a starburst - dominated source based on its rest - frame optical spectrum and cold dust detection @xcite .
mips22432 already had the necessary signal - to - noise ratio and hence no new data for it is presented here .
we merely re - reduced the spectra and again see a clear 6.2@xmath0 m pah feature .
unfortunately , the 3.3@xmath0 m feature falls right on the edge of its spectrum and hence we can not test for its detectability .
mips22530 , on the other hand , shows an unambiguous 3.3@xmath0 m pah feature ( see figure[fig_specs ] ) .
in addition , three more sources mips16080 , mips22303 and mips22633 show some evidence for the 3.3@xmath0 m pah feature but somewhat confused with water ice and hac absorption ( see figure,[hac_det ] ) . for mips22303 and mips22633
, a potential 6.2@xmath0 m feature falls in the unreliable wavelength range where we stich the different irs orders ( see figure[fig_specs ] ) .
mips16080 , whose 3.3@xmath0 m pah feature is too weak to be considered a detection , does show a tentative detection of the 6.2@xmath0 m feature .
the pah emission of this source is made more reliable due to its having an accurate optical spectroscopic redshift .
in addition , mips16080 has starburst - like optical line ratios , but agn - line ir spectrum suggesting it is a composite source @xcite .
we fit a simple local continuum+lorentz profile model to each of the detected or hinted at 3.3@xmath0 m pah features .
figure[pahfit_3um ] shows the results . wherever it was possible , we did the same for the 6.2@xmath0 m feature . for mips16080
we find @[email protected]@xmath27@xmath44erg / s/@xmath45 ( ew@xmath43120 nm ) . for mips22530
we find @[email protected]@xmath27@xmath44erg / s/@xmath45 ( ew@xmath43550 nm ) and @[email protected]@xmath27@xmath44erg / s/@xmath45 ( ew@xmath43370 nm ) .
these imply a 6.2/3.3 ratio of @xmath234 for these sources .
this is roughly consistent with what is observed for local ulirgs ( e.g. * ? ? ?
* ; * ? ? ?
* ) although the observed ratio has a large scatter .
note however , that for mips16080 , both the 3.3 and 6.2@xmath0 m pah features are somewhat unreliable ( see e.g. figure[akari_comparison ] ) and should really be considered upper limits . , height=529 ] figure[fig_comp ] shows our sources overlaid with local analogs . from top to bottom
we show sources in order of increasing silicate feature depth .
we picked iras19254 - 7245 , and iras17044 + 6720 due to their being among the few sources in the akri sample of @xcite which had comparably red 34@xmath0 m spectra .
here they are compared with mips16059 and mips464 respectively , to show that selecting on the basis of red 34@xmath0 m spectra alone , yields good agreement across the full observed range ( @xmath2211@xmath0 m ) .
mips22633 looks quite extreme and unusual , showing a very strong @xmath26@xmath0 m absorption complex , while being the weakest silicate absorption source in this sample .
however , we do find a local analog , iras23515 - 2917 , which has an even more extreme 6@xmath0 m absorption coupled with a relatively shallow 9.7@xmath0 m silicate absorption .
lastly our strongest silicate absorption source , mips8392 seems to have a mid - ir sed remarkably like that of the hyperluminous galaxy f00183 - 7111 @xcite .
this galaxy is one of the very few local examples that have comparable luminosities to those of our sample , and therefore finding such good agreement is not surprising . in the previous sections we have shown that a number of features are detected in the individual sources
spectra including the 3.4@xmath0 m hac feature as well as the 3.05@xmath0 m water ice absorption feature .
there are hints of some other features in the sources such as marked in figure[fig_specs ] , but they are typically of low significance . these include in particular the 6@xmath0 m absorption complex ( likely including the 6.15@xmath0 m water ice feature ) , the 6.85 and 7.25@xmath0 m hac features and the 4.67@xmath0 m co feature discussed in [ sec_co ] .
to address the question of whether or not these are present in our sample , we stacked the spectra as shown in figure[fig_stack ] .
because there are too few sources to test the robustness of the stack by taking randomly half of it , we used a different technique which involves removing each existing pair of sources , or 28 possibilities .
this implies that features that appear in each of the 28 stacks need to be present in the majority of sources in order not to disappear in at least some of the combinations .
figure[fig_stackzoom ] shows that in the stacked spectra , we see that the 3.4@xmath0 m hac feature is quite prominent , unsurprising given that we detect it in four of the individual spectra .
potentially there is also 3.05@xmath0 m water ice absorption and as well as potentially the 3.3@xmath0 m pah emission given some 3.3@xmath0 m excess .
the 3.0@xmath0 m feature is badly fit at the lower wavelength end for most , but not all permutations .
mips16059 is probably responsible for some fraction of these ( see 4.4 ) , the rest is probably due to edge effects in sources without any clear features .
figure[fig_stack ] , however , suggests that the 6@xmath0 m absorption complex is there in most sources .
the most prominent such feature is seen in mips22633 where the 6@xmath0 m absorption is clear in the individual spectrum as well .
one thing to keep in mind is that in a number of the sources , the @xmath26@xmath0 m features in particular are not individually detected due to their falling inside the unreliable irs modules overlap region .
this uncertainty , as well as the likely 3.3@xmath0 m pah make it difficult to estimate the average 3.0@xmath0 m water ice and 3.4@xmath0 m hac feature depths .
figure[fig_stackzoom ] suggests that these are @[email protected] and @[email protected] , but it should be born in mind that these are fairly uncertain .
the stacked spectra show signs of pah emission superimposed on the strong mid - ir continua ( recall that we have excluded the two individually strong - pah sources in the sample ) .
the @xmath28@xmath0 m profile suggests the presence of the 7.7@xmath0 m and 8.6@xmath0 m features .
in addition , the 3.3@xmath0 m feature ( seen tentatively in the individual spectra for mips16080 , mips22303 and mips22633 ) appears clearly present .
the stacks show a possible 6.2@xmath0 m feature as well ; however , caution must be taken as for several of the sources this feature falls in the unreliable edge region . to test the reliability of this ` detection ' we also stacked only the sources with reliable spectra in this region ( i.e. mips464 , mips15880 , mips16080 , mips16152 , mips22548 ) . the 6.0@xmath0 m absorption and the 6.85 and 7.25@xmath0 m hac
features appear in this as well , although the 6.2@xmath0 m ` bump ' is less pronounced than seen in figure[fig_stack ] .
the 4.67@xmath0 m co feature is hinted at but seems to be driven by the feature in mips464 as when this source is not included we have the four outlier stacks .
this re - inforces the conclusion that , apart from the three sources discussed before , the co feature is either not present or is extremely weak ( @[email protected] ) in the majority of our sample .
the silicate feature does not fall within our new deeper data and is discussed in more detail in @xcite for the go1 sources and @xcite for the go2 sources .
however here we address the measurement and reliability of the silicate feature depths for our 11 targets specifically .
this is necessary because in order to put our newly detected molecular absorption features in context with local sources we would like to look at the ice - to - silicate and hac - to - silicate feature ratios ( which we do in the discussion ) .
many of our sources have poorly defined continuum beyond rest - frame @xmath211@xmath0 m and also tend to have very low signal - to - noise inside the feature itself ( see figure[fig_specs ] ) .
these features were originally fit with a power law plus extinction law model ( see * ? ? ? * ) , which as discussed earlier relates to the observed silicate feature depth via @xmath50=1.4@xmath51 .
for some of the sources this is fairly well determined , but for others the fitted value has large uncertainty ( either due to poorly determined continuum , or uncertainty due to the pah features flanking the feature ) .
for example , for mips464 we derived model @[email protected] which translates to observed @[email protected] .
by contrast , mips15880 has considerably better defined feature , and we derive model @[email protected] which translates to observed @[email protected] . in the case of the strong - pah source , mips22530 ,
the silicate feature is saturated and we can only determine that it has observed silicate absorption of @[email protected] . in order to further test the uncertainty associated with our silicate feature optical depth estimates , we also fit all of them using the same procedure as used for the local ulirgs in @xcite .
we find that for the majority of the sources , the two approaches give consistent results , the values obtained differ most for the sources with poor continuum and noisy features , but are within the fitted uncertainty ( e.g. mips464 has @xmath21=2.5 consistent with our @[email protected] ) .
the most substantial difference comes for mips22530 where we obtain @xmath21=1.7 , considerably shallower than the value obtained from our model fitting .
this is a well known uncertainty arising from the different treatment of the pah features .
since mips22530 and mips22432 ( the other strong - pah source ) do not have either hac or ice detections , they do not feature in our optical depth ratio estimates ( see discussion ) and hence this substantial uncertainty does not affect our results .
figure [ pah_ratio ] shows the ew3.3 vs. ew6.2 for some representative local ulirgs .
we also show the derived values for mips22530 and mips16080 ( which is an upper limit only as neither feature is unambiguously detected by itself ) .
we show the starburst - dominated source limits as determined by @xcite based on optical spectral classification , and @xcite based on neon line ratios . on both counts , mips22530 satisfies the starburst - dominated ulirg criteria , consistent with our prior knowledge of this source .
mips16080 on the other hand is more likely a composite source , as expected .
four of the sources showed signs of the 3.3@xmath0 m pah feature , with mips22530 having the most unambiguous detection .
based on the fits presented in 4.6 , we can estimate 3.3@xmath0 m pah log luminosities of 9.6 , 9.6 , 9.8 and 10.1 for mips16080 , mips22303 , mips22530 , and mips22633 .
the total 31000@xmath0 m infrared luminosities for the first three sources are given in @xcite and for mips22633 we have @xmath52=12.9 ( sajina et al . in prep . ) . for our most reliable 3.3@xmath0 m pah source ( mips22530 ) , we have @xmath53=8.9@xmath27@xmath54 .
this is consistent with the @xmath55/@xmath4=8.5@xmath27@xmath54 reported by @xcite for a lensed @xmath1@xmath23 lyman break galaxy , the only other high-@xmath1 measurement of the 3.3@xmath0 m pah .
this ratio is somewhat larger than observed for local ulirgs ( @xmath2@xmath54 ; * ? ? ?
however , this is believed to be an aperture effect , as this ratio is observed to increase with distance with the more distant ulirgs having a ratio of @xmath2@xmath56 @xcite . for our other sources
we find @xmath53=6.6@xmath27@xmath54 ( mips16080 ) , 4.3@xmath27@xmath54 ( mips22303 ) and 1.6@xmath27@xmath56 ( mips22633 ) .
if we take a ratio of @xmath2@xmath56 to be indicative of starburst - dominated systems ( see * ? ?
* ) , then the above can give us an alternative look into what fraction of the @xmath4 is due to starburst activity and also what sfrs are feasible .
accordingly , mips16080 likely has equal agn and starburst contributions to the bolometric luminosity , whereas for mips22633 , the starburst may well be dominant . in mips22303 ,
the agn is likely the dominant component .
these are consistent with our earlier results @xcite .
note however , that using the above translation between @xmath55 and @xmath4 implies starburst - only luminosities of @xmath18@xmath7@xmath6 even for mips16080 and mips22303 , and consequently sfrs of @xmath2100@xmath57/yr .
however , as discussed in 4.6 , these features being of low signal - to - noise ratios and somewhat confused with the ice and hac absorption features , are not particularly reliable .
the stacked spectrum shows that pah emission is likely present in the majority of these sources , even excluding the two strong - pah sources .
the fact that star - formation is present is further supported by the fact that although most of these sources are not individually detected in the far - ir , their stacked 1.2 mm ( rest - frame @xmath2400@xmath0 m ) flux is strongly detected at @xmath20.5mjy which is comparable to fainter ( @xmath58@xmath223mjy ) sub - mm galaxies @xcite .
these numbers suggests that sfrs on the order of 100@xmath57/yr are feasible for the bulk of the weak - pah sample .
although in itself significant , star - formation at this level is insufficient to dominate the bolometric luminosities of these sources .
local ulirgs show a high level of detectability of water ice absorption features .
for example the 46 sources in the akari sample @xcite were selected out of the iras 1jy sample merely on the basis of visibility to akari and are therefore unbiased with respect to the local iras ulirg population as a whole .
the 3@xmath0 m feature was detected in 30/46 sources ( 65% ) , whereas the 3.4@xmath0 m hac feature was only observed in 4 of the sources ( 9% ) .
the presence of the hac feature has been suggests as an indicator of agn - dominance in the mid - ir @xcite and hence this low fraction is consistent with local ulirgs being predominantly starbursts .
to determine the appropriate detectability for our sample we need to consider first that mips8392 has too low a signal to noise in this regime , and three other sources are of @[email protected] which makes this feature too near the edge for reliable detections .
this means that 3/7 ( 43% ) of our sample show the 3@xmath0 m water ice absorption feature , this is roughly comparable to the local ulirgs given the small number of sources in our sample .
on the other hand ( 5/11 ) of the sources ( 45% ) show the 3.4@xmath0 m hac absorption feature , which fraction is significantly higher than observed for local ulirgs ( even accounting for the poisson error in such a small sample ) . however , the 3.4@xmath0 m hac feature can be practically difficult to detect in sources with strong 3.0@xmath0 m ice and/or 3.3@xmath0 m pah emission .
this is one of the explanation for the low hac detectability among the local ulirgs @xcite .
we have already shown that the best analogs to our sources are the sources with red 34@xmath0 m spectra rather than the more typical local ulirgs .
there are five sources in the @xcite akari sample that meet our criteria ( i.e. @xmath60@xmath152 ) . of these ,
3 show the hac feature , and a different set of 3 show the 3.0@xmath0 m ice feature .
this is roughly in better agreement with the observed detection rates for our sample ( but with very poor statistics ! ) .
figure[hac_si ] shows the hac / silicate ratio vs. the silicate feature strength for our sources compared with local ulirgs . to construct this figure
, we use the @xmath61 measurements for local ulirgs from several studies @xcite combined with the silicate feature depths from @xcite and @xcite .
we find that , while the hac detectability of our sample is higher than the local ulirg population , the hac / silicate ratio appears consistent with that observed among local ulirgs where the hac feature is detected . for comparison
we also show the data for embedded protostars from the review paper of @xcite .
note that the hac - to - silicate ratio observed in the mw is remarkably constant across a wide range of silicate feature depth .
this is consistent with the diffuse ism value of 0.050.06 @xcite .
moreover , a recent study of the spatial variation of the optical depth in the circinus galaxy ( the nearest agn to the mw ) found that although the absolute optical depth changed dramatically , this ratio remained nearly constant @xcite .
an enhanced hac - to - silicate ratio was first reported by @xcite who studied a sample of 4 local obscured agn nuclei including ngc1068 ( shown in figure[hac_si ] ) .
their favored interpretation of this enhanced ratio comes from the following explanation : each of the two optical depth samples the column density between us and the sources .
if the dust is well mixed and in a screen approximation is valid the ratio is expected to remain constant ( reflecting the relative abundance of the silicate and carbon grains and their optical properties ) .
however if a steep temperature gradient between the @xmath2300k dust at 9.7@xmath0 m and @xmath21000k dust at 3.4@xmath0 m is present , since the hotter dust is spatially closer to the power source , the hac feature arises from a larger column density than the silicate feature and hence the enhanced ratio .
this is what is commonly referred to as a ` buried nucleus ' ( see fig.2 * ? ? ?
notice that this model addresses only the obscuration geometry : it is neither a statement about the nature of the power source ( although commonly taken to be an agn in these cases ) , nor about the absolute level of obscuration ( the embedded protostars show comparable silicate feature depths ) .
an alternative explanation might be that deeply obscured ulirgs and our @xmath1@xmath22 sources simply have a higher proportion of hydrocarbons .
for example , bombardment with h atoms can hydrogenate amorphous carbon grains increasing the hac - to - silicate ratio ( see e.g. * ? ? ? * ; * ? ? ? * ) .
various other processes can be important , s.a .
uv radiation processing of ice mantles can lead to refractory materials on the grain surfaces that lead to the 3.4@xmath0 m absorption feature @xcite , consistent with the somewhat weaker ice features . ultimately , the degree of hac absorption depends on the balance of processes that support hydrocarbon generation to those that suppress it ( s.a . converting hydrocarbons to aromatics ) .
it is not unreasonable to suppose that such processes might be more efficient in our sources than the typical milky way sources .
the geometric explanation is simpler and therefore favored , however , a difference in composition can not be excluded .
figure[ice_si ] shows the ratio of the 3.0@xmath0 m ice feature optical depth to the 9.7@xmath0 m silicate dust optical depth . as before ,
we compare our sample with both local ulirgs and the data for mw stars .
we find that unlike the case for the hac - to - silicate ratio which is remarkably similar across for all available mw sources , the ice - to - silicate ratio shows a wide spread of more than an order of magnitude .
this ratio is strongly dependent on the ice mantle thickness ( e.g. * ? ? ?
* ; * ? ? ?
* ) , but radiative transfer details following the arguments presented in the previous section likely dominate .
when we consider only the sources with red 34@xmath0 m spectra , and where both ice and silicate optical depths are available , we find that iras08572 + 3915 has @xmath62=0.08 , and iras19254 - 7245 has @xmath62=0.15 .
although a much larger sampling is needed , this suggests ( especially in conjunction with our own sample ) that sources with redder 34@xmath0 m slopes have lower observed ice - to - silicate ratios .
the presence of co gas in at least some of these systems is known independently through mm - wave spectroscopic observations ( yan et al . , in prep . ) .
in particular , mips16080 and mips22530 both show significant masses of cold co gas .
unfortunately both of these are among the sources , where the 4.67@xmath0 m feature falls in the unreliable parts of the co - added spectra .
moreover , the presence of co does not automatically imply the presence of the 4.67@xmath0 m co feature which arises from the much warmer regions in the vicinity of the galactic nuclei .
we would expect to see this feature in our sources as it is seen in some deeply obscured local ulirgs @xcite .
we do show some evidence for the 4.67@xmath0 m co absorption feature in three of these sources .
although tentative , this is the first suggestion of this feature in such high redshift systems .
however this feature is very weak even in these sources and only seen in 3 of the 7 sources with usable spectra ( and is not present in the stacked spectra ) . we can therefore conclude that it is either not ubiquitous even among sources with silicate absorption , or more realistically that the feature manifests with a range of absorption depths , and we only see the most pronounced features . as discussed in 4.5
, a @xmath25@xmath0 m starburst continuum would effectively weaken the observed co feature depth ( as that likely originates close to the agn itself ) . for weak - pah sources ,
the @xmath25@xmath0 m continuum is most likely dominated by the agn @xcite , however a starburst contribution , sufficient to affect the co depth , can not be excluded .
the fact that one of our two strong - pah sources ( mips22432 ) is also one of three sources showing some hints of co absorption , suggests that this is much more complicated that a simple correction scaled to the starburst strength ( as revealed by the pah ) .
drawing a parallel with the dust - embedded galactic ysos , perhaps there is a range from purely solid form co to fully gaseous co as the temperatures rise and the co previously deposited on grain surfaces sublimates ( see * ? ? ?
* ; and references therein ) .
solid - phase co shows a much narrower feature than its gas - phase counterpart , and in our low - resolution spectra would be undetectable , unless exceptionally strong . in local ulirgs , this problem is present as well , and only the high resolution spectrum of iras08572 + 3915 has allowed to rule out co ice in this source @xcite . to date , only ngc4945 has a co ice detection @xcite .
following the argument in @xcite , we might also conclude that we are not seeing this feature due to dust tori obscuration rather than ` buried nuclei ' .
however , this runs counter to our conclusions from the strong 3.4@xmath0 m hac absorption features ( see 5.3 ) .
ultimately , there are too few 4.67@xmath0 m co detections even locally ( see introduction ) and therefore the statistics on what fraction of the sources we should expect to show this feature sufficiently strongly are lacking .
our sources are drawn from sample of 24@xmath0 m flux limited and color - selected sources @xcite .
the 0.9mjy 24@xmath0 m flux limit already restricts us at @xmath1@xmath22 to the highest luminosity sources , and indeed we find @xmath63@xmath2@xmath64@xmath6 for all our @xmath1@xmath22 sources
@xcite shows that locally , at comparable luminosities , @xmath270% of the ulirgs host buried agn , compared to @xmath230% in lower luminosity ulirgs .
it is not surprising , therefore , that in @xcite we found that the bulk of our @xmath1@xmath22 sources showed agn - like mid - ir spectra , usually accompanied by silicate absorption features .
because of the biased selection of our sample , we can not test whether or not obscured agn form a larger fraction of @xmath1@xmath22 ulirgs than observed for the local ulirgs .
we did show that , consistent with the well - known luminosity evolution between @xmath1@xmath22 and today , the number density of our silicate - absorption sources is about an order of magnitude higher than observed locally ( for ulirgs with comparable silicate absorption @xcite ) .
however , silicate absorption alone is not a strong enough diagnostic of buried agn since it can in principle arise from a range of different spatial scales ( i.e. host obscuration ) . in this study
, we find that beyond the presence of silicate absorption , the full 2 - 8@xmath0 m seds of our @xmath1@xmath22 sources , the presence and strength of 3.4@xmath0 m amorphous hydrocarbons and 3.0@xmath0 m water ice absorption features are also comparable to those observed in nearby heavily obscured ulirgs .
this suggests similar obscuration geometries .
their closest counterparts in the local universe are believed to be dominated by buried nuclei ( i.e. compact sources with steep dust temperature gradients ) most likely harboring agn as discussed in the preceeding sections ( see e.g. * ? ? ? * for a more complete discussion ) .
the lack of strong 4.67@xmath0 m co absorption however is inconsistent with this picture , although the dearth of such observations locally makes it difficult to interpret ( see 5.5 ) .
we find evidence of pah emission even among the weak - pah ( i.e. agn - dominated in the mid - ir ) sources , suggesting that star - formation is present at least in some of these sources . scaling from the 3.3@xmath0 m luminosities we expect that sfrs of order @xmath2100@xmath57/yr could be present .
this is typically insufficient to dominate the bolometric output of these galaxies .
the combination of deeply obscured , likely quasar , nuclei and ulirg - strength star - formation supports the view of co - evolution of black - hole growth and host galaxy stellar mass growth suggested by a variety of observational and theoretical considerations ( see e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
\1 ) in this paper we present deep rest - frame 28@xmath0 m spectra for 11 @xmath1@xmath22 ulirgs .
+ 2 ) our sources tend to have redder 34@xmath0 m continua than local ulirgs .
the few local sources with comparably red spectra , show continuum - dominated 2 - 10@xmath0 m spectra with weak - pah and strong silicate absorption ( consistent with their being predominantly agn - dominated ) .
they are therefore overall good analogs to our sources .
+ 3 ) we found that nearly half of the sources with sufficient coverage ( 43% ) show water ice absorption at 3@xmath0 m ( compared with 65% for local ulirgs ) .
there is no clear trend of detectability with increasing silicate feature depth .
+ 4 ) the 3.4@xmath0 m hac absorption is seen in 40% of the sample compared with 9% for local ulirgs .
the hac - to - siilcate ratio , @xmath65 , is 25 times stronger that the typical ratio observed in the milky way , but consistent with the obscured nuclei of some local ulirgs .
this is consistent with compact nuclei buried in dust with a steep temperature gradient distribution .
+ 5 ) three of our sources show tentative signs of the co 4.67@xmath0 m feature .
if genuine co absorption , these features show a strong asymmetry between the short- and long-@xmath33 branches of this feature .
moreover , the observed strength is still significantly lower than observed in local ulirgs .
stacking analysis suggests this feature is not present in the bulk of our sample or is very weak ( @[email protected] ) .
+ 6 ) one source , mips22530 , has a clear 3.3@xmath0 m pah emission feature detected .
the ratio of the 6.2 to 3.3@xmath0 m pah strengths is consistent with that observed in local ulirgs , and the overall strength of both features implies a starburst - dominated source , consistent with earlier results .
three more sources ( mips16080 mips22303 , and mips22633 ) show tentative evidence for the 3.3@xmath0 m pah .
the stacked spectra show evidence for pah emission in the bulk of the weak - pah sources implying likely significant star - formation in conjunction with the obscured black - hole accretion .
+ we thank the anonymous referee for their thoughtful comments , which have improved this paper .
we also wish to thank brian siana and kalliopi dasyra for helpful discussions .
this work is based on observations made with the _ spitzer _ space telescope , which is operated by the jet propulsion laboratory , california institute of technology under a contract with nasa .
support for this work was provided by nasa through an award issued by jpl / caltech .
gibb , e. l. , whittet , d. c. b. , boogert , a. c. a. , & tielens , a. g. g. m. 2004 , , 151 , 35 greenberg , j. m. , li , a. , mendoza - gomez , c. x. , schutte , w. a. , gerakines , p. a. , de groot , m. 1995 , apj , 455 , l177 rigby , j. r. , et al .
2008 , , 675 , 262 risaliti , g. , et al .
2006 , , 365 , 303 sajina , a. , yan , l. , armus , l. , choi , p. , fadda d. , helou , g. , spoon , h. 2007a , apj , 664 , 713 sajina , a. , yan , l. , lacy , m. , huynh , m. , 2007b , apjl , 667 , 17 sajina , a. , yan , l. , lutz , d. , et al . 2008 , apj , 683 , 659 sani , e. , et al .
2008 , , 675 , 96 siana , b. , et al .
2009 , arxiv:0904.1742 sirocky , m. m. , levenson , n. a. , elitzur , m. , spoon , h. w. w. , & armus , l. 2008 , , 678 , 729 smith , j. d. t. , et al .
2007 , , 656 , 770 solomon , p. m. , rivolo , a. r. , barrett , j. , & yahil , a. 1987 , , 319 , 730 spoon , h. w. w. , keane , j. v. , tielens , a. g. g. m. , lutz , d. , moorwood , a. f. m. , & laurent , o. 2002 , , 385 , 1022 spoon , h. w. w. , moorwood , a. f. m. , pontoppidan , k. m. , cami , j. , kregel , m. , lutz , d. , & tielens , a. g. g. m. 2003 , , 402 , 499 spoon , h. w. w. , et al . 2004 , , 154 , 184 spoon , h. w. w. , keane , j. v. , cami , j. , lahuis , f. , tielens , a. g. g. m. , armus , l. , & charmandaris , v. 2005 , astrochemistry : recent successes and current challenges , 231 , 281 spoon , h. w. w. , marshall , j. a. , houck , j. r. , elitzur , m. , hao , l. , armus , l. , brandl , b. r. , & charmandaris , v. 2007 , , 654 , l49 cccccccc 464 & 1.85 &
& 6@xmath27120s ( s ) & & & 4@xmath2740@xmath2760s ( m ) & 30@xmath27120s ( s ) + 8392 & 1.9 & 2@xmath2760s(s ) & & & 12@xmath27120s ( s ) & & 30@xmath27120s ( s ) + 15880 & 1.68 & & 6@xmath27120s ( s ) & & & 4@xmath278@xmath2760s ( m ) & + 16059 & 2.325 & & 6@xmath27120s ( s ) & & & 4@xmath2740@xmath2760s ( m ) & 30@xmath27120s ( s ) + 16080 & 2.007 & & 6@xmath27120s ( s ) & & & 4@xmath2740@xmath2760s ( m ) & 30@xmath27120s ( s ) + 16152 & 1.83 & & & & 6@xmath27120s ( s ) & 6@xmath2760s ( s ) & 30@xmath27120s ( s ) + 22303 & 2.34 & & 6@xmath27120s ( s ) & & & 4@xmath278@xmath2760s ( m ) & 30@xmath27120s ( s ) + 22432 & 1.59 & & & 1@xmath27240s(s ) & 6@xmath27120s(s ) & & + 22530 & 1.952 & & 6@xmath27120s ( s ) & & & 4@xmath2740@xmath2760s ( m ) & 30@xmath27120s ( s ) + 22548 & 2.12 & & & & 6@xmath27120s ( s ) & 6@xmath2760s ( s ) & 30@xmath27120s ( s ) + 22633 & 2.36 & & & & 6@xmath27120s ( s ) & 4@xmath278@xmath2760s ( m ) & 30@xmath27120s ( s ) ccccc 464 & 11.07 & 11.66 & 3.12 & 2.85 + 8392 & 11.63 & 12.11 & 3.63 & 0.91 + 15880 & 11.31 & 11.89 & 2.52 & 2.90 + 16059 & 11.74 & 12.12 & 3.15 & 0.97 + 16080 & 11.49 & 11.95 & 3.38 & 0.95 + 16152 & 11.68 & 11.90 & 2.12 & 1.59 + 22303 & 11.95 & 12.42 & 3.30 & 1.08 + 22432 & 11.41 & 11.90 & 3.75 & 2.76 + 22530 & 11.10 & 11.51 & 1.90 & 4.76 + 22548 & 11.88 & 12.03 & 2.26 & 1.54 + 22633 & 11.54 & 12.06 & 3.37 & -0.30 cccccc 464 & 1.85 & 1.75 & & [email protected] & [email protected] + 8392 & 1.90 & 1.90 & & @xmath160.45 & @xmath154.9 + 15880 & 1.68 & 1.62 & & @xmath160.28 & [email protected] + 16059 & 2.325(keck ) & 2.325 & @xmath160.22 & [email protected] & [email protected] + 16080 & 2.007(keck ) & 2.007 & [email protected] & [email protected] & [email protected] + 16152 & 1.83 & 1.84 & @xmath160.28 & @xmath160.16 & [email protected] + 22303 & 2.34 & 2.28 & [email protected] & @xmath160.24 & [email protected] + 22432 & 1.59 & 1.60 & & & [email protected] + 22530 & 1.952(keck ) & 1.952 & @xmath160.34 & & @xmath153.7 + 22548 & @xmath152.12 & 2.15 & @xmath160.26 & @xmath160.16 & [email protected] + 22633 & 2.36 & 2.20 & [email protected] & [email protected] & [email protected] | we present the first detections of the 3@xmath0 m water ice and 3.4@xmath0 m amorphous hydrocarbon ( hac ) absorption features in @xmath1@xmath22 ulirgs .
these are based on deep rest - frame 28@xmath0 m _ spitzer _ irs spectra of 11 sources selected for their appreciable silicate absorption .
the hac - to - silicate ratio for our @xmath1@xmath22 sources is typically higher by a factor of 25 than that observed in the milky way . this hac
` excess ' suggests compact nuclei with steep temperature gradients as opposed to predominantly host obscuration . beside the above molecular absorption features , we detect the 3.3@xmath0 m pah emission feature in one of our sources with three more individual spectra showing evidence for it .
stacking analysis suggests that water ice , hydrocarbons , and pah are likely present in the bulk of this sample even when not individually detected .
the most unexpected result of our study is the lack of clear detections of the 4.67@xmath0 m co gas absorption feature .
only three of the sources show tentative signs of this feature and at significantly lower levels than has been observed in local ulirgs .
overall , we find that the closest local analogs to our sources , in terms of 34@xmath0 m color , hac - to - silicate and ice - to - silicate ratios , as well as low pah equivalent widths are sources dominated by deeply obscured nuclei .
such sources form only a small fraction of ulirgs locally and are commonly believed to be dominated by buried agn .
our sample suggests that , in absolute number , such buried agn are at least an order of magnitude more common at @xmath1@xmath22 than today .
the presence of pah suggests that significant levels of star - formation are present even if the obscured agn typically dominate the power budget . |
a disc galaxy is primarily a set of stars which rotate around the centre at near - circular orbits at the angular velocity @xmath1 .
the observed spiral and bar - like structures are customarily treated as perturbations in such a differentially rotating disc .
thus , one might expect that the typical pattern speeds of these structures , @xmath0 , should be of the order of some average star angular velocity , @xmath2 . in reality
, however , the pattern speeds @xmath0 are of the order of the characteristic precession speed of star orbits , @xmath3 , which is only a small fraction of @xmath2 . recall that @xmath4 for near - circular orbits , where @xmath5 is the epicyclic frequency . according to lynden - bell ( 1979 ) , if stars involved in the formation of the structures satisfy the inequality @xmath6 then each star orbit as a whole , but not the individual stars , participates in the perturbations .
consequently , for studying such structures , it is more reasonable to use the model of the disc of star orbits precessing with different speeds than the commonly used concept of the differentially rotating disc of individual stars .
the main conjecture of the present paper is that the galactic spirals and bars are the normal modes in such a model of the disc galaxy . in view of its simplicity , this approach , as we will see below , allows one to clarify the underlying physical mechanisms in the formation of galactic structures .
the lynden - bell inequality ( [ in1 ] ) mentioned above are justified by results of numerous calculations ( e.g. , athanassoula & sellwood ( 1986 ) for bar - modes or lin , yuan & shu ( 1969 ) for spirals ) . and
@xmath7 for the plummer potential ( thick solid lines ) , and the minimum , mean , and maximum patterns speeds ( thin solid , dash - dotted , and dashed lines , respectively ) from the list of athanassoula & sellwood ( see their table 1 ) ; b ) the ratios @xmath8 for @xmath0 from a).,width=529 ] fig.[fig_shust]a shows @xmath1 and @xmath7 for the plummer potential @xmath9 used by athanassoula & sellwood ( 1986 ) in their @xmath10-body bar - mode analysis . the solid thin horizontal line @xmath11 , the dash - dotted line @xmath12 , and the dashed line @xmath13 correspond , respectively , to the minimum , mean , and maximum pattern speeds from the list given by athanassoula & sellwood ( 1986 ) in their table 1 .
fig.[fig_shust]b represents the ratios @xmath8 ( @xmath14 ) for these modes . for the first mode localized inside the circle @xmath15 ( thin dashed lines in fig .
[ fig_shust]b ) , typical ratios are of the order of 0.1 . for the second mode ( mean pattern speed ) , these ratios are typically of the order of 0.2 .
even for the fastest mode ( @xmath16 ) , typically , @xmath17 , taking into account that this mode is more concentrated to the centre .
therefore it is little wonder that the pattern speeds calculated below ( section 5 ) in the framework of our approach repeat very accurately those of athanassoula & sellwood ( 1986 ) in all cases , including the most rapid bar - modes .
, for the model of the galaxy from lin , yuan , & shu ( 1969).,width=529 ] on the other hand , it is evident from the curves in fig .
[ fig_shust]b that the inequality ( [ in1 ] ) and hence our approach fail at sufficiently large radii .
this is certainly of no importance for determining the pattern speeds , since the bar - modes are mainly localized near the centre . roughly speaking
, the ratio @xmath8 becomes of the order of 1 near the corotation radius ( i.e. , at @xmath18 , and @xmath19 , for the first , second , and third modes , respectively ) .
fig.[fig_shmidt]a , b show similar graphs for the classical spiral example from lin , yuan & shu ( 1969 ) where the model of our galaxy was considered .
the pattern speed used here corresponds to the mean of the interval ( @xmath20 km s@xmath21 kpc@xmath21 ) suggested in the cited paper , i.e. , the horizontal line in fig.[fig_shmidt]a is @xmath22 km s@xmath21 kpc@xmath21 .
fig.[fig_shmidt]b shows that the analysis of the eigen - modes of galactic disc for this model could be performed in the framework of our approach ( without the restrictions of the wkbj theory ) .
note that substantially higher values of @xmath0 , compared to those from lin et al .
( see , e.g. , blitz 1983 ) , are most likely due to an independent peripheral stage of the spiral structure .
such structures can be analyzed using the integral equations described in the end of section 2 .
we picture the disc as comprising a large number of slowly deforming and rotating elliptical rings , along which an individual stars moves so fast it is no more than a blur . small changes with radius in the shape or orientation of rings give rise to regions of enhanced surface density ( see , e.g. , fig 6 - 11 of binney & tremaine 1987 ) . in the absence of self gravity any initial pattern of over - densities
will wind up because the part of the pattern at radius @xmath23 will pecess at a rate @xmath24 that varies significantly with @xmath23 .
rings that have inclined minor axes exert gravitational torques on one another .
these torques cause the angular momenta and radial actions of the rings slowly to change , thus deforming the rings and altering any initial pattern of surface density .
the torques also affect the precession rates of the ring s the minor axes . the interaction between rings is strongest when the rings are physically adjacent , i.e. , have similar values of the actions . if we neglect the non - negligible range of the interaction , the disc has something in common with a goods train of deformable wagons connected by springs .
self gravity allows waves of ring deformation to propagate through the disc just as the springs permit waves of wagon displacement and deformation to propagate down the train .
we show that these waves can generate a pattern that rotates rigidly at an angular frequency that is larger than the fastest precession rate in the disc .
the picture has points in common with lynden - bell s ( 1979 ) theory of bar formation in that it focuses on slowly precessing orbital rings , but it does not conclude , with lynden - bell , that the bar rotates at a rate that compromises between the fastest and slowest precession rates of the consituent rings . instead we find that both bars and spirals , in common with many wave phenomena , move faster than the underlying medium .
several studies have shown that the rotation rates of bars in n - body models conform to our predictions rather than those of lynden - bell .
the model of precessing orbits can be introduced directly as is done in section 2 after eq .
( [ kern ] ) .
however , this model is rigorously justified by the use of the regular procedure of the perturbation theory in the small lynden - bell parameter @xmath25 from ( 1 ) , with averaging over fast radial oscillations of stars .
this programme is implemented in the beginning of section 2 .
the resulting integral equations for disc normal modes derived in two ways , namely , ( i ) directly in the model of precessing orbits , and ( ii ) with the help of averaging , are identical .
lynden - bell ( 1979 ) showed that the quantity @xmath26 ( @xmath27 and @xmath28 are the usual actions ) is an adiabatic invariant , provided that the inequality ( [ in1 ] ) is satisfied , i.e. , for sufficiently slow perturbations .
this means that the orbits evolve by the action of the potential of low - frequency modes only through change in the angular momentum @xmath29 , with @xmath30 held constant .
assume that in the initial moment @xmath31 the distribution function of orbits is @xmath32 .
let @xmath33 be a small variation of the orbit angular momentum owing to the action of a small potential perturbation for the time from @xmath31 to the current moment @xmath34 .
the flow of the phase fluid is incompressible , so the phase element with the initial density @xmath35 should come into the point @xmath36 at the moment @xmath34 . accordingly , the euler differential of the distribution function , i.e. , its perturbation , is equal to @xmath37 ( @xmath38 ) .
@xmath39 depends on the derivative @xmath40 .
so it is not suprising that just this derivative is found to be critical to our theory of disc slow modes .
henceforth , this derivative is referred to as the lynden - bell derivative of the distribution function . ] .
it turns out that the behaviour of the lynden - bell derivative determines the type of eigen - modes ( bar - like or spiral ) . to develop the above considerations , immediately resulting from the remarkable paper by lynden - bell ( 1979 ) , into the theory of low - frequency disc modes
, one should invoke the dynamical equations and the expression for the 2d disc potential ( see sections 2,3 ) . as a result
, we obtain the integral equations for these eigen - modes ( see eqs .
( [ ie1 ] ) and ( [ ie2 ] ) below ) .
if a potential satisfies these integral equations , the orbits evolve in such a way that the pattern rotates uniformly , with a certain angular speed @xmath0 .
the physics of our bar - like and spiral solutions consists just in the fact that these are ( slow ) normal modes .
all specificity ( compared to general modes ) is concerned with their slowness .
the ensuing sections of this paper present the following material . in section 2 , we derive the basic integral equation of the theory ( by two methods ) .
the basic integral equation is then reduced to the form of the classical eigen - value problem .
section 3 is devoted to the analysis of the general properties of the basic integral equation .
in particular , the crucial role of the lynden - bell derivative of the distribution function is revealed . in section 4 , this derivative is examined in more detail by the example of the schwarzschild model . in two following sections ,
the specific solutions of the basic integral equation are given . to demonstrate the capabilities of the proposed theory , test models studied earlier by the @xmath10-body method ( athanassoula & sellwood 1986 ) are adopted ( section 5 ) .
the theory gives results in close agreement with results of the @xmath10-body simulations .
a wide variety of spiral modes are studied in section 6 . in the last section 7
, a brief summary of the results is given .
the most convenient variables that should be used under studying the low - frequency modes are @xmath26 and @xmath41 , where @xmath30 is the lynden - bell adiabatic invariant ( lynden - bell 1979 ) , while @xmath42 are the usual actions .
the angle variables corresponding to @xmath43 are @xmath44 .
the variables @xmath45 are also actions , for which @xmath46 are the canonically conjugate angles .
one can show that for the unperturbed state @xmath47 where @xmath34 is the time , @xmath48 is the precession speed of the orbit with the actions @xmath49 , @xmath50 and @xmath51 are the frequencies of star radial and azimuthal oscillations : @xmath52 @xmath53 , @xmath54 is the unperturbed star energy expressed in terms of @xmath43 , @xmath55 is the star velocity , @xmath56 is the equilibrium potential . in the limit of near - circular orbits @xmath57 , @xmath58 , where @xmath59 is the epicyclic frequency .
eq.([eq2 ] ) means that the angular variable @xmath60 is slow because @xmath61 , @xmath62 .
it is easy to show that @xmath63 , where @xmath64 is the azimuth of the orbit s minor axis ( see below , fig.[fig_alpha ] ) . for later use
we give here the coordinates of stars @xmath23 , @xmath65 , expressed in terms of @xmath49 , @xmath66 .
the radius @xmath67 is determined by solving the equation @xmath68 - l^2 /{r'}^2}}\ ] ] ( it is assumed that @xmath69 ) .
the slow angular variable @xmath70 where @xmath71 the distribution function @xmath72 is governed by the collisionless boltzmann equation @xmath73 , \label{cbe}\ ] ] where @xmath74 is the hamiltonian of a star in the self - consistent gravitational field @xmath75 , @xmath76 @xmath77 $ ] denotes the poisson bracket , @xmath78
\equiv \frac{\p h}{\p { \bar w}_1 } \frac{\p f}{\p j_f } + \frac{\p h}{\p { \bar w}_2 } \frac{\p f}{\p l }
- \frac{\p f}{\p { \bar w}_1 } \frac{\p h}{\p j_f } - \frac{\p f}{\p { \bar w}_2 } \frac{\p h}{\p l}.\ ] ] to linearize the collisionless boltzmann equation , we assume that @xmath79 @xmath80 where @xmath81 and @xmath82 are the unperturbed and perturbed distribution functions , respectively , @xmath83 is the perturbation of the potential , @xmath84 is the azimuthal wavenumber , @xmath85 is the frequency ( @xmath86 ) . as a result
, we obtain the equation @xmath87 to find the desired low - frequency solutions , one can use the perturbation theory in the small lynden - bell parameter @xmath25 from ( [ in1 ] ) .
let @xmath88 be the perturbation series in powers of @xmath25 , so that @xmath89 is obtained from ( [ kint ] ) by neglecting the terms proportional to @xmath24 and @xmath90 : @xmath91 , i.e. @xmath92 is an arbitrary function of the integrals of motion , which is subsequently specified by using the periodicity condition for the solution of the next approximation , the wave frequency @xmath85 is not determined in the first approximation . ] .
the equation for @xmath93 takes the form @xmath94 given the periodicity of functions @xmath93 and @xmath95 , averaging ( [ kint2 ] ) over @xmath66 in the interval @xmath96 yields @xmath97 the azimuthal number @xmath84 can take only even values .
formally , it follows from the relation @xmath98 , which remains periodic in @xmath99 ( with a period equal to @xmath100 ) only for even @xmath84 .
physically , a closed precessing orbit is an oval symmetric relative to the centre .
so the torque from the perturbations with odd @xmath84 will break rather than rotate such orbits , which is valid only for modes with even @xmath101 . among them , the bi - symmetric mode @xmath102 is predominant .
for instance , just this bar - mode is most likely the only unstable mode .
indeed , the excees of @xmath103 over @xmath104 is due to the self - gravitation .
this effect is proportional to the function @xmath105 , as it is seen from ( [ ie2 ] ) . from ( [ pi2 ] )
, it follows that the function @xmath105 decreases with @xmath101 .
so , the growth of the modes with @xmath106 owing to cr and olr would be much less than for the @xmath102 mode .
the ilr resonance for @xmath106 , however , will prevent the growth . ] . calculating the perturbed surface density , @xmath107 and using the expression for the 2d disc potential ,
we find @xmath108 where @xmath109^{1/2}$ ] .
now one can change in the last formula from @xmath110 , @xmath111 to @xmath112 taking into account that @xmath113 ( @xmath114 , @xmath115 ) : @xmath116 } { r_{12}},\ ] ] with @xmath117 .
finally , averaging the potential @xmath95 over @xmath66,we obtain the following integral equation is correct only for unstable frequencies ( im@xmath118 ) .
thus roots with ( im@xmath119 ) should be omitted . ] : @xmath120 where @xmath121 , @xmath122 the function @xmath105 can be reduced to the following suitable form : @xmath123 where @xmath124 the function @xmath125 ( multiplied by the factor @xmath126 ) has a meaning of torque acting from the harmonical distribution of orbits , @xmath127 , with the action @xmath128 to the orbit with the action @xmath49 and orientation @xmath129 .
it is notable that the integral equation ( [ ie1 ] ) in terms of the function @xmath130 has the form of the classical eigen - value problem , where the eigen - value is directly the pattern speed @xmath0 .
indeed , determining @xmath131 through @xmath130 from ( [ kint3 ] ) and substituting it into ( [ ie1 ] ) , one obtains @xmath132 with the kernel @xmath133 .
\label{kern}\ ] ] eq .
( [ ie2 ] ) is the basic integral equation of the theory .
the integral equation ( [ ie2 ] ) is much simpler than the general integral equations for disc normal modes derived earlier by kalnajs ( 1965 ) and shu ( 1970 ) .
an ordinary desktop pc is quite capable of solving the integral equation ( [ ie2 ] ) .
the most important advantage of the present approach is the fact that it makes clear the underlying physical mechanisms of the instability processes developing in the disc . to reveal these physical mechanisms by using the integral equations of kalnajs or shu
would be very difficult , and the same is true for the @xmath10-body simulations . the formal derivation of the integral equations ( [ ie1],[ie2 ] ) for the low - frequency modes of gravitating discs given above is rigorous .
however , the physical situation will be more transparent after deriving the equivalent integral equation in another way using explicitly the fact that each orbit as a unit should take part in slow perturbations of interest . accordingly , let us consider the distribution function of closed precessing orbits , @xmath134 , such that @xmath135 is the mass of stars at orbits within a given interval @xmath136 , @xmath64 is the azimuth of minor axis , so that the orbit precession speed @xmath137 .
the collisionless kinetic equation for such a distribution function is @xmath138 where one takes into account that @xmath139 for the slow modes of interest ( lynden - bell 1979 ) , and @xmath140 , where @xmath141 is the torque acting on the orbit with given @xmath142 .
note that such a distribution function @xmath143 and kinetic equation ( [ clke ] ) were earlier suggested by v.polyachenko ( 1992 ) . by linearizing the equation ( [ clke ] ) and assuming that @xmath144
, one can obtain the equation that is analogous to eq.([kint3 ] ) : @xmath145 where @xmath146 is the perturbation of the torque .
let us check that the equations ( [ kint7s ] ) and ( [ kint3 ] ) are actually identical .
first of all , it is easy to show that @xmath147 where @xmath148 is the potential averaged over the selected orbit ; this potential is produced by all other orbits of a system : @xmath149 where @xmath150 is the linear mass density characteristic for each orbit ( @xmath151 is the total star velocity , @xmath152 is the current length of the orbit , @xmath153 is the radius - vector at the orbit ) , @xmath154 is the mass of stars on the orbit , @xmath155 ( @xmath156 and @xmath157 are the current azimuths at two orbits ) . in the formulae ( [ mom1 ] ) and ( [ phibar1 ] )
, we used the same notation @xmath158 for the averaged potential as earlier in the derivation of the integral equation ( [ ie1 ] ) , since in both cases it was actually the same quantity . to convince oneself that this is correct
, one can immediately compare the two expressions for @xmath158 , taking into account that a star is within the interval @xmath159 during @xmath160 .
thereafter it remains to make sure that the slow angular variable @xmath161 and the azimuth @xmath64 of the minor axis are identical .
indeed , ( [ w2e ] ) can be rewritten in more informative manner : @xmath162 where @xmath163 is the time it takes for the rotation of a star through the angle @xmath164 between the azimuth of the minor axis and the current azimuth @xmath156 .
then the identity @xmath165 is clear from fig.[fig_alpha ] . after invoking poisson s equation
, we obtain the integral equation that is coincident with eq .
( [ ie1 ] ) .
in addition to the derivations above , let us indicate the simplest way to use eq.([frch ] ) of the introduction for obtaining the relation ( [ kint3 ] ) .
the canonical hamilton equations in variables @xmath166 are @xmath167 where @xmath168 is the hamiltonian .
since the angular variable @xmath161 is `` slow '' , for studying the low - frequency modes , the procedure of averaging the equations of motion over a quick variable ( in this case , @xmath66 ) is appropriate . as a result ,
we obtain ( see , e.g. , arnold 1989 ) @xmath169 the first equations means the adiabatic invariance of @xmath30 , while the second equation determines the evolution of the angular momentum @xmath29 . assuming that @xmath131 corresponds to a normal mode , from ( [ kaneq3 ] ) one can find : @xmath170 integrating the last equation over @xmath34 from @xmath171 to @xmath34 , we find @xmath172 taking into account that the perturbation is switched off at @xmath173 . substituting ( [ delta_l ] ) into ( [ frch ] )
we obtain ( [ kint3 ] ) .
it is clear that the perturbed precession speed is the same periodic function of time as @xmath33 . ) , which proves the identity @xmath63 .
dotted and solid ovals show orientations of an orbit at times @xmath174 and @xmath175 , respectively ; the orbit is turned by an angle @xmath176 .
the azimuth of a star initially ( at @xmath174 ) coincides with the azimuth of the minor axis ( point @xmath177 ) .
the position of the star at @xmath178 is shown by point @xmath179 ; its azumuth is @xmath180 ( dash - dotted arc ) , and the variation of the azimuth is @xmath181 ( dashed arc).,width=264 ] recall that the integral equation ( [ ie2 ] ) is valid only for even @xmath84 .
however , one can obtain the integral equation describing the low - frequency modes for @xmath84 divisible by @xmath182 ( where @xmath182 is an arbitrary integer ) in the form ( [ ie2 ] ) , when replacing @xmath183 with @xmath184 respectively . the case @xmath185 is particularly interesting .
the corresponding basic integral equation is suitable for studying low - frequency modes of discs around black holes , in planetary rings , and so forth .
since only two types of the potentials , @xmath186 and @xmath187 are available , in which all orbits have zero precession speeds , there are only two cases ( @xmath185 and @xmath188 ) where our approximation that @xmath66 is fast is valid in an extensive region .
as it follows from ( [ pi2 ] ) , the function @xmath105 is real and symmetric : @xmath189 , @xmath190 .
let us restrict ourselves to the regular solutions for the averaged potential @xmath131 .
taking into account ( [ kint3 ] ) , one can safely divide both sides of ( [ ie2 ] ) by @xmath191 , multiply by @xmath192 , and integrate over @xmath49 . calculating the imaginary part of the resulting equation
, one finds @xmath193 by using ( [ kint3 ] ) , eq.([imom1 ] ) can be reduced to the form @xmath194 where @xmath195 denotes the angular momentum of the mode that can be obtained from the general expression of lynden - bell & kalnajs ( 1972 ) retaining the only term which dominates for the low - frequency modes : @xmath196 it is found that the type of solution of the basic integral equation depends crucially on the behaviour of the derivative @xmath191 .
below , we consider two cases . \1 .
assume that @xmath191 is strictly positive everywhere in the phase space of the system .
then @xmath195 is negative , as well as the energy of the mode @xmath197 , since @xmath198 ( lynden - bell & kalnajs , 1972 ) .
hence , from ( [ imom2 ] ) , one obtains @xmath199 .
the corresponding real eigen - functions , @xmath130 , describe non - spiral solutions .
we identify them with the inner parts of bar - modes , i.e. , bars themselves .
obviously , in the case @xmath200 under consideration , the integral equation ( [ ie2 ] ) determines merely the pattern speed @xmath201 of a mode .
the question now arises of whether there are some external reasons for growth of bar - modes .
it is found that the growth can be due to the exchange by the angular momentum with the resonance stars at the corotation and the outer lindblad resonance ( olr ) .
these resonances are located rather far from the central regions where the modes are mainly localized . in this respect
these resonances are external relative to the modes .
of course , in the end the correctness of our approach is justified by the comparison with results of @xmath10-body simulations ( see section 5 ) .
the corresponding growth rate can be calculated by the formula @xmath202 where @xmath203 , and the expressions for the rates of exchange by the angular momentum at cr ( @xmath204 ) and olr ( @xmath205 ) can be obtained from the general formulas of lynden - bell & kalnajs ( 1972 ) through their minor transformation : @xmath206 , \label{er1}\ ] ] where @xmath207 is the dirac delta - function , @xmath208 , and @xmath209 is the fourier coefficient corresponding to cr ( @xmath210 ) and olr ( @xmath211 ) in the expansion of the potential in the series in @xmath212 .
the interaction of stars in the resonance regions with the gravitational potential of the mode leads to the spiral responses . for the corotation resonance , as an example ,
the response consists of two parts .
the resonance response of the disc , without taking its self - gravity into account , produces spirals of angular length about @xmath213 ( see , e.g. , polyachenko 2002a ) .
the self - gravity also extends the resonance spirals by @xmath213 ( polyachenko 2002b ) .
therefore , the total angular length is about @xmath214 , which is typical for the majority of sb galaxies ( see , e.g. , sandage 1961 ) .
it should be emphasized that above we consider only the linear stage of the formation of structures in sb galaxies .
the observed galaxies can have strongly nonlinear bars .
the length of such bars is apparently determined by 4:1 resonance .
this is exactly the case , e.g. , for the nonlinear bar - modes computed by sellwood & athanassoula ( 1987 ) . for a symmetric bar which is free of odd fourier harmonics , in particular @xmath215 harmonic , the resonance 4:1 is the nearest to the centre , and thus is the most important one .
let us say that @xmath191 becomes negative in some regions of the phase space .
for the realistic distribution function ( see section 4 ) , these regions , if they exist , can occupy only a small fraction of the total volume of the phase space . then , in addition to the fast bars described above , the spiral - like solutions can occur .
contrary to the bars , these new modes grow due to their intrinsic instability , which is inherent to the mode itself . as follows from ( [ imom2 ] ) , for an unstable mode ( @xmath216 ) , the angular momentum @xmath217 , i.e. the contributions to @xmath195 from the regions with opposite signs of the lynden - bell derivative @xmath191 cancel each other exactly : @xmath218 .
obviously , the instability is due to the inner lindblad resonance , since the criterion of instability , @xmath219 , coincides exactly with the lynden - bell & kalnajs condition of wave exitation at ilr ( see ( [ er1 ] ) when @xmath220 ) .
this resonance becomes the source of spiral waves in the case under consideration , contrary to the commonly accepted view of the role of ilr .
accordingly , the growth rate of the unstable mode can be calculated as @xmath221 where @xmath222 , \label{lpm}\ ] ] @xmath223 and @xmath224 denote the regions of the phase space with positive and negative values of the lynden - bell derivative @xmath191 , respectively .
note that below the mode growth rates are obtained directly from the solution of the integral equation ( [ ie2 ] ) as @xmath225 .
however , it is also useful to derive a simple formula for estimate of the growth rate from ( [ gamma2 ] , [ lpm ] ) .
the simplest case is that with one narrow domain of the negative derivative @xmath191 . using the latter equality of ( [ gamma2 ] ) and approximating the integrals as the products of the average values of the integrands and small volume of the phase space domain
, we obtain : @xmath226 where @xmath33 is the width of the region with @xmath227 , and @xmath228 is calculated at the narrow region of the phase space of interest .
this estimate is consistent with the explanation of the physical mechanism of the angular momentum exchange at the ilr given by lynden - bell & kalnajs ( 1972 ) .
however , two more points need to be made : ( i ) the sign of the effect is opposite to that in the cited paper , in accordance with the fact that @xmath227 , and ( ii ) strictly speaking , their explanation is appropriate only for the case of near - circular star orbits , although , as we demonstrate below , the very elongated orbits play an essential role .
nevertheless , their considerations can readily be generalized to such a case .
let us consider the generalized schwarzschild distribution function ( shu 1970 ) @xmath229 where @xmath230 is the star energy , @xmath231 is the radius of the guiding centre : @xmath232 , @xmath233 is the star energy at the circular orbit , @xmath234 is the circular velocity , @xmath235 is the equilibrium potential .
the specific model is given by the functions @xmath236 and @xmath237 ; in the epicyclic limit , when @xmath238 , @xmath239 , @xmath240 , where @xmath241 and @xmath242 are the surface density and radial velocity dispersion respectively
. in the general case , @xmath241 and @xmath242 are expressed by @xmath236 and @xmath237 in a more complicated manner .
the lynden - bell derivative of the distribution function ( [ schw ] ) is equal to @xmath243 where @xmath244 , a prime denotes the derivative with respect to @xmath231 . as a rule , @xmath200 either in all the phase space or at least for the most part ( almost everywhere ) .
it is provided by the term @xmath245 in ( [ lbd1 ] ) .
for example , in the case of the flat rotation curve ( @xmath246 ) , this term equals @xmath247 . since @xmath238 almost everywhere in disc galaxies ,
this term is dominant .
however , there may exist some narrow regions , in which such predominance is broken .
firstly , this is possible in the regions where the function @xmath248 is sufficiently small .
for instance , with a rotation curve similar to that of our galaxy , such regions can be located at @xmath249 kpc and @xmath250 kpc .
note the obvious causes of such peculiarities in the rotation curves : ( i ) a changeover from the potential of the spherical component to the potential of the disc , and ( ii ) a sharp edge of one of the disc components .
secondly , the function @xmath237 increases rapidly when approaching the galactic centre , so that the rotation velocity @xmath251 in the central parts becomes of the order of @xmath252 .
relatively small deviations in the behaviour of the rotation curve may lead to very different solutions of the basic equation .
some other factors are also very important .
the excess of the pattern speed above the maximum precession rate is due to self - gravity .
thus the solutions have the bar - like form only in sufficiently massive discs , while for discs with relatively small masses we can obtain only spiral modes .
the number of very elongated orbits plays the essential role .
for example , the excess of such orbits over the number prescribed by the generalized schwarzschild distribution function ( [ schw ] ) strongly increases the growth rate of the spiral modes due to increase of @xmath253 for these orbits . besides
, it is seen from ( [ lbd1 ] ) that @xmath254 ( @xmath227 ) increases with @xmath255 .
the basic integral equation is studied numerically .
the unknown function @xmath256 as well as the kernel @xmath257 are considered on the @xmath258 grid in the phase space @xmath259 .
then the obtained matrix equation can be solved using standard methods of linear algebra .
it is advisable to put aside the distribution function ( [ schw ] ) until the next section and demonstrate the capabilities of the theory by applying it to test models .
for such a role , we choose two models investigated by the @xmath10-body simulations of athanassoula & sellwood ( 1986 ) .
the equilibrium potential for all the models is the plummer potential , @xmath9 .
first of all , we are dealing with a kalnajs ( 1976 ) model , which athanassoula & sellwood ( 1986 ) have denoted by ( @xmath260 , @xmath261 , @xmath262 , @xmath263 ) ( see their table 1 ) .
the computed spectrum of the matrix equation , which consists of @xmath264 eigen - values , is shown in fig.[fig1 ] .
the frequencies are real , since the lynden - bell derivative for the kalnajs models is positive everywhere in the phase space .
just few of eigen - values correspond to the discrete spectrum , while others mimic the continuous part of the spectrum ) , which defines a bounded noncompact integral operator ( _ cf .
_ reed & simon 1972 ) . ] , which is the van kampen case modes ( van kampen 1955 ; case 1960 ) .
they are located in the interval between the minimum and maximum values of the precession speed : @xmath265 .
these modes are the stable solutions of the form @xmath266 .
the peculiar van kampen modes are of no interest in studying the bar modes .
thus , in the @xmath10-body simulations of athanassoula & sellwood , the bar is always one of the modes of discrete spectra .
the latter are the eigen solutions of the basic integral equation .
both the perturbed potential and surface density of the discrete modes differ in the number of radial nodes , the nodeless mode being the mode with the maximum pattern speed . ) , for one of the kalnajs models .
the crossess are the complex pattern speeds , with the growth rates estimated by the formulas ( [ gr1 ] ) , ( [ er1 ] ) .
the diamonds are the experimental " complex pattern speeds according to athanassoula & sellwood ( 1986 ) .
the arrows point to the minimum and maximum values of the star precession speed.,width=317 ] in the spectrum given in fig.[fig1 ] one can see 5 discrete modes .
the figure shows that the pattern speeds of the first and second modes ( counting from the right ) coincide with the pattern speeds obtained in the @xmath10-body simulations to within 6% .
it is found ( see below ) that the growth rates of the other 3 modes are much lower , so they could not be observed in the numerical experiments .
rough estimates of growth rates for the most rapid modes , in which we assumed that the star orbits are circular , give @xmath267 , @xmath268 .
each of these values is composed of two parts corresponding to corotation ( cr ) and olr : @xmath269 , @xmath270 , @xmath271 , @xmath272 .
note that in this case , the olr terms are dominating .
the mode with the maximum pattern speed has the highest growth rate . obviously , the reason is that such a mode possesses the smallest corotation radius , where the disc surface density is high .
the estimate for the growth rate of the most rapid mode is nearly twise as much as one obtained by athanassoula & sellwood ( 1986 ) in the @xmath10-body simulations . for the second mode ,
the `` experimental '' and `` theoretical '' growth rates are quite close . among the reasons of the possible discrepancy
, one can mention the use of the circular orbit approximation .
recall that the estimates obtained by athanassoula & sellwood ( 1986 ) within the swing amplification approach are twice as little than their `` experimantal '' values .
another model was denoted by athanassoula & sellwood as ( @xmath260 , @xmath273 , @xmath262 , @xmath274 ) ) ( about ten models in all ) . in all cases
the results coincide very closely ( with an accuracy of 10% or better ) . besides , the eigen - frequencies of the basic integral equation for the isochrone potential coincide within a few percents with those obtained by kalnajs ( 1978 ) using his matrix equation . ] .
the computed spectrum of eigen - values of bi - symmetric modes for this model , is shown in fig.[fig5 ] .
the continuous spectrum of van kampen modes is again located in the interval between the minimum and maximum values of the precession speed : @xmath275 . ) , for the model ( @xmath260 , @xmath276 , @xmath262 , @xmath274 ) of athanassoula & sellwood ( 1986 ) .
the cross is the complex pattern speed , with the growth rate estimated by ( [ er1 ] ) .
the diamonds are the experimental " complex pattern speeds according to athanassoula & sellwood . the arrows point to the minimum and maximum values of the star precession speed.,width=317 ] since the lynden - bell derivative for the model under consideration is not positive everywhere in the phase space , the discrete spectrum contains complex modes . in fig.[fig5 ] , we can see one unstable complex and one real discrete frequencies . the complex eigen - frequency obtained from the solution of the basic integral equation
is very close to the `` experimental '' frequency , given by athanassoula & sellwood ( 1986 ) .
the spiral pattern of the unstable mode is given in fig.[fig6 ] .
, @xmath273 , @xmath262 , @xmath274 ) of athanassoula & sellwood ( 1986 ) .
the dashed circle indicates the position of the inner ilr ( the outer ilr , for this model , is at @xmath277).,width=264 ] the pattern speed of the real mode coincides with the pattern speed obtained in the @xmath10-body simulations to within 3% .
the estimate of its growth rate , in the circular orbit approximation , gives @xmath278 , the olr contribution being dominating .
this estimate is 30% larger than that obtained by athanassoula & sellwood ( 1986 ) in the @xmath10-body simulations .
the models explored above are rather artificial .
nevertheless , they were convenient for us because it was thoroughly studied in the @xmath10-body simulations by athanassoula & sellwood ( 1986 ) . besides , the analysis above points to the mechanism of bar - formation in the general case .
it is clear that bars can form in galaxies with sufficiently massive discs . in this case , the necessary inequality @xmath279 can be fulfilled , otherwise the inner lindblad resonances occur .
then , in the case of a positive lynden - bell derivative , only the continuous spectrum of the van kampen waves exists .
note that the fulfillment of the inequality @xmath279 is made difficult not only for low - mass discs but also for discs with a high concentration of mass at the centre .
the permitted frequency ( pattern speed ) of the galactic bar must be one of the eigen - values of the basic integral equation .
it means that the corresponding waves remain unchanged over many revolutions of the galaxy , despite the different rates of orbit precession .
the half - turn spirals adjacent to the bar form due to the same resonance interaction .
here we consider in greater detail the situations that arise when the lynden - bell derivative is negative in some regions of the phase space .
then a great number of available possibilities appears ; they are briefly outlined below .
an extended discussion will be published elsewhere . as a base distribution function for the analysis of the spiral solutions of the basic integral equation ,
we choose the generalized schwarzschild distribution function described in section 4 . for the standard model
, we assume the milky way - type " rotation curve ( see fig.[fig2]a ) , and the exponents @xmath280 , @xmath281 that fix the particular schwarzschild function . in our units ,
one kpc is equal to 1 , @xmath282 , @xmath283 , the gravitational constant @xmath284 , @xmath285 .
the magnitude of the rotation curve is chosen to match the disc component of the velocity curve at radii @xmath286 .
the maximum rotation velocity is @xmath287 , @xmath288 , so that for @xmath289 the ratio @xmath290 . for such parameters ,
the galactic disc is exponential almost everywhere , excluding the very central part . given the rotation curve , the equilibrium potential is @xmath291 .
in such a manner we obtain the model of a galaxy , which is similar to our galaxy ( but , of course , the model galaxy is not identical ) .
, of the circular orbits for the standard model ( solid line ) and the model with a peak " near @xmath292 ( dashed line ) .
the horizontal straight lines correspond to different pattern speeds ; ( c ) the lynden - bell derivatives , @xmath293 , for the circular orbits.,width=317 ] fig.[fig2]c shows the lynden - bell derivative , @xmath191 , for the standard model calculated by ( [ lbd1 ] ) at the line of circular orbits ( @xmath294 ) .
there is only one narrow region with @xmath295 located near the centre .
fig.[fig3 ] demonstrates how deep into the phase space the negative values of @xmath191 penetrate .
, at the phase plane @xmath259 .
the phase domain is bounded by the line of circular orbits ( from the top ) and the line corresponding to @xmath296 , where @xmath297 is the maximum value of the distribution function , @xmath298 , at each fixed value of @xmath230 ( from the right ) .
the solid curves , @xmath299 , indicate the locations of resonance orbits on the phase plane , for several values of @xmath0.,width=317 ] the basic integral equation for the standard model provides a large number of unstable modes
. they can be either trailing or leading spirals , the latter are the majority .
narrowly - localized van kampen solutions of the continuous spectrum , which occupy the wide band of real frequencies , @xmath300 , can be both the trailing and leading spirals since for each of these modes , effectively , the gravitational constant @xmath301 everywhere in the galactic disc with the assumed parameters . ] . when employing the concept of eigen - modes , just the packets of leading van kampen waves participate in the classical swing amplification process described in the original papers by goldreich & lynden - bell ( 1965 ) , julian & toomre ( 1966 ) , and toomre ( 1981 ) . if the dispersion of orbit speeds , @xmath302 , is sufficiently small ( that is always assumed ) , we obtain quasi - monochromatic packets of van kampen waves . just such wave packets moving with some group velocity from the central regions and approaching the corotation , can be amplified by the resonance effects under the transformation from the leading form to trailing one .
though each of the van kampen modes is essentially kinematic ( @xmath303 ) , the finite packets of these waves are governed by their self - gravity .
in particular , the local dispersion relation of lin shu
kalnajs should be valid for them ( away from the resonances ) .
here we touch on those improvements , which the present theory can introduce into the original scheme of the swing amplifier .
they are connected with the possibility of unstable discrete leading spiral modes . in particular
, there is then no need to reflect the short trailing waves somewhere in the central regions to close the feedback loop .
given that the discrete leading modes are monochromatic , the efficiency of the resonance amplification should grow in comparison with quasi - monochromatic packets of the van kampen waves .
these modes might be the embryo " waves for their subsequent resonance growth due to the swing amplifier .
then , apparently , from the full set of these waves , the mode with the maximum potential perturbation at the corotation is selected .
let us emphasize again the inversed " role of the ilr ( compared to the commonly accepted one ) , which acts as the generator of spiral waves .
the growth rates of spiral modes for the standard model are rather low : they provide the growth of the amplitude @xmath304 times , typically for @xmath305 years ( taking into account that @xmath306 corresponds to the rotation period @xmath307 years ) .
the main reason for such low values of growth rates follows from fig.[fig3 ] .
this figure shows that the curve @xmath308 ( for a typical unstable spiral mode ) intersects the strip of negative values of the lynden - bell derivative , @xmath191 , deep inside the phase space of the system under consideration , at @xmath259 corresponding to highly elongated orbits .
but the number of such orbits in the schwarzschild distribution function ( decreasing exponentially with the ratio @xmath309 ) is small .
the growth rates of the spiral modes can be significantly increased if one takes into account that the number of such orbits is actually very much larger .
indeed , we are dealing with the central regions of spiral galaxies , where the contribution of the spherical component , i.e. stars of population ii that have the radially elongated velocity diagram , is essential .
it implies that the spherical component can play an important active role , instead of the passive role of a rigid halo as is commonly accepted .
computation with the standard model that is accordingly modified , leads to tenfold growth rates .
the modification consists in the substitution @xmath310 provided that @xmath311 .
for such a modification , the observed surface density and radial velocity dispersion are almost unchanged .
it is also important that this modification leads to the predominance of trailing spiral modes .
the observed galactic spirals are most likely just these modes ( or similar spirals in other modifications of the standard model ) .
for instance , the modes are most often the trailing half - turn spirals as in the majority of observed galaxies .
in particular , an interesting modification consists of forming the peak " on the curve @xmath24 at sufficiently large @xmath23 , where the lynden - bell derivative can be either positive or negative ( in fig.[fig3 ] , the upper right corner of the @xmath259 phase plane ) .
we omit descriptions of the variety of cases , restricting ourselves to the only picture of the typical spiral mode ( fig.[fig4 ] ) . in conclusion
let us touch on one very important aspect of the problem .
fig.[fig2]c shows that the `` generalized '' lynden - bell derivative @xmath293 at @xmath312 can be made negative more easily compared to the case @xmath313 .
this effect is due to the inverse dependence of the large positive term @xmath314 in the expression ( [ lbd1 ] ) on @xmath182 .
hence , in the outer parts of galaxies ( for our galaxy , just beyond the solar circle ) the additional multi - arm structures are generated , which are independent from the central spiral structure ( the latter usually has two arms ) in fig.[fig2]c are shown only at sufficiently large radii to emphasize that modes with @xmath312 can not typically occupy as wide a region as the two - arm mode does .
this is explained by the different behaviour of @xmath7 and @xmath315 for @xmath312 . ] .
furthermore , one can expect that several modes will be generated there simultaneously leading to formation of the so - called flocculent structures ( elmegreen & elmegreen 1982 , 1987 ) .
all the aforementioned agrees with observations of galaxies ( see , for example , the hubble atlas of galaxies , sandage ( 1961 ) ) . as a rule , in the central part of the disc , the regular two - arm spiral
is seen , which is then superseded by a multi - arm or disordered structure .
it is appropriate to mention here that elmegreen ( 1992 ) has clearly demonstrated that three - arm spiral structures appear only on the periphery of galaxies beyond the inner lindblad resonance for the @xmath215 mode .
the new approach to the study of structures in spiral galaxies suggested above is based on a view of a galactic disc as a set of precessing orbits . following this model ,
we have obtained the simple integral equation that is a very convenient tool for studying galactic structure .
non - axisymmetric structures are considered as the normal modes of the basic integral equation ( [ ie2 ] ) .
it is found that the properties of solutions depend crucially on the behaviour of the lynden - bell derivative , @xmath191 , at the phase space of the disc .
the solutions have bar - like or spiral form depending on whether @xmath191 is positive everywhere or negative somewhere in the phase space . within the framework of the suggested approach , which does not include the swing amplifier , we succeeded in calculating the pattern speeds and growth rates of all the modes obtained earlier in the @xmath10-body simulations by athanassoula & sellwood ( 1986 ) , the coincidence of results being very high .
thus , it is shown that the bar - mode grows due to the immediate action of its gravity on the stars in the vicinity of corotation leading to the angular momentum exchange between these stars and the bar - mode .
note that athanassoula & sellwood ( 1986 ) could readily calculate exact growth rates of all the modes they computed by the formulas ( [ gr1 ] , [ er1 ] ) .
however , they preferred the language of swing amplification for rough estimates of the growth rates . to avoid confusion , we should emphasize that here we deal only with the standard fast bars .
as for the slow bars of lynden - bell ( 1979 ) , they are most likely the central parts of the unstable spiral modes ( in this sense , they are secondary ) .
note that here we leave aside the possibility of the slow bar formation in very hot centres of galactic discs ( in full analogy to the ellipsoidal deformation of spherical systems under the action of the radial orbit instability ) .
in contrast , the fast bars are primary while the adjacent spirals are secondary since the spirals form as the response of the disc near the corotation to the bar gravity .
the commonly used approach to the problem of formation of the normal spiral structures consists in using either the swing amplification mechanism ( toomre 1981 ) or its weaker counterpart ( overreflection with reference to the waser mechanism
see , e.g. , bertin & lin 1996 ) . note that the swing amplification mechanism is sometimes the only way for growth of the initial perturbations .
it seems likely that such galaxies do not have any organized spiral structure of the mode nature .
certainly , this mechanism can effectively amplify transient wave perturbations . in traditional mechanisms , the region of corotation plays the central role . in our approach , we draw attention to ilrs that can under certain conditions cause the excitation of a great variety of spiral modes . unlike the usual epicyclic idea of ilr as a certain circle , ilr in our unstable modes extends over wide regions of the galactic disc .
this is due to participation of a variety of orbits including highly elongated orbits ( see fig.[fig3 ] ) . on the one hand
, ilr can give rise to leading spiral modes , which in turn can be the embryos " for further work of the swing amplifier .
on the other hand , ilr can provide a variety of trailing spirals including those in the region between two ilrs ( see fig.[fig2]b ) .
with thanks to v.l .
polyachenko for useful discussions .
the perturbed distribution function @xmath316 is represented as a fourier series that contains resonance denominators : @xmath317 it is natural to name the terms with @xmath318 , @xmath220 , and @xmath319 as `` ilr '' , `` cr '' , and `` olr '' terms , respectively . in section 2 , we have restricted ourselves to the only ilr term . as it was shown in section 4 , this leads to the correct results ( with an accuracy of 10% ) in determination of the pattern speeds .
thus , one can conclude that the ilr - term has a dominant role .
the ratio of the denominator of the ilr term to the denominator of the cr term ( closest to ilr ) is @xmath320 where the quantity @xmath321 was already used in the introduction . strictly speaking
, this ratio is equal to the lynden - bell small parameter @xmath322 from ( [ in1 ] ) only when @xmath323 , i.e. in the central regions far enough from the corotation .
fig.[fig1s ] shows the ratio ( [ ratio ] ) for the same 3 models as in fig.[fig_shust ] .
it is seen that @xmath324 for the slowest mode , so the validity of the theory is justified for such modes .
the fact that the theory has a certain field of application justifies the main conclusions of the theory ( classification of modes , physics of instabilities ) , derived in section 3
. they can be adequate even beyond the framework of rigorous validity of our approximate theory .
for the other two modes from fig.[fig1s ] , the ratio is considerably smaller than 1 only in the central regions .
nevertheless , we can hope that the approach is still valid even for these modes , at least quantitatively , taking into account that the perturbation amplitude decrease with radius , being relatively small at the corotation .
besides , there is one more factor that weakens the role of cr term ; it is considered in the next point .
the coefficient @xmath325 in the expansion ( [ dfe1 ] ) is proportional to the linear combination of the derivatives of the unperturbed distribution function @xmath326 ( see below , item 5 ) : @xmath327 where @xmath328 thus , for the bi - symmetric mode @xmath102 one should actually calculate the ratio @xmath329 rather than ( [ ratio ] ) . in the epicyclic approximation , which is valid everywhere beyond the very central parts of galactic discs : @xmath330
fortunately , for the cr term @xmath331 , so the ratio ( [ ratio2 ] ) contains the additional small factor ( compared to ( [ ratio ] ) ) .
apparently , this may be very important for explaining why our theory is valid even for fastest modes .
intuition suggests that higher harmonics ( with @xmath332 ) can not considerably influence the large scale modes , which are the subject of our study .
the particular number of harmonics , needed for the mode computation , is determined `` experimentally '' .
the comparison with @xmath333-body simulations shows that the only ilr term provides a reasonable accuracy . in the next point
we give a more detailed analysis of the relative roles of the terms in the expansion ( [ dfe1 ] ) . for different values of the pattern speed @xmath103
: solid , dash - dotted , and dotted lines correspond to the minimum ( @xmath334 ) , mean ( @xmath335 ) , and maximum ( @xmath336 ) pattern speeds , respectively.,width=302 ] \5 .
let us derive the generalization of the basic integral equation that takes into account all a resonanses of the perturbation with fixed azimuthal number @xmath84 .
let the perturbed potential and the perturbed distribution function are expanded in fourier series @xmath337 where @xmath338 are usual radial and azimuthal actions , @xmath339 is the radial angular variable . substituting the full potential and the full distribution function @xmath340 into the collisionless boltzmann equation , one can obtain after linearization @xmath341 where the function @xmath342 is @xmath328 the second relation between @xmath343 and @xmath344 is provided by the poisson equation @xmath345 multiplying both sides of the equation by @xmath346 and introducing function @xmath347 ( see section 2 ) , one can obtain @xmath348 where @xmath349 . denoting @xmath350
the integral @xmath351 one obtains the second relation : @xmath352 ( here and below the summation over the blind index @xmath353 is assumed ) .
thus , expressing @xmath354 from ( [ cbe1 ] ) one can derive the final set of integral equations : @xmath355 for numerical calculations , it is more convinient to rewrite the equation ( [ bs1 ] ) in the form of the classical eigen - value problem .
indeed , using ( [ cbe1 ] ) one can obtain @xmath356 using the discretization of the integrals @xmath357 , the eigen - frequencies can be obtained from the condition : @xmath358 where @xmath359 is the unity matrix .
note that the basic integral equation ( [ ie2 ] ) is the special case of ( [ bs2 ] ) when @xmath360 .
in fact , the equations ( [ bs2],[det ] ) give a new formulation of the general eigen - value problem for galactic discs .
this formulation can be considered as an alternative for the well - known matrix approach of kalnajs ( apj , v.212 , p.637 ( 1977 ) ) .
let us employ the equation ( [ det ] ) to define more exactly the eigen - frequencies of the bi - symmetric @xmath102 modes in the kalnajs model @xmath361 , which was already studied in section 4 .
athanassoula & sellwood reported that there are two unstable bar - modes : @xmath362 , @xmath363 .
we found that for this model the lynden - bell derivative is positive everywhere , so only bar - modes with @xmath364 were possible . using the basic integral equation ( [ ie2 ] )
, we can determine only the pattern speed of the modes ; they found to be @xmath365 , @xmath366 .
we argued that the growth rate for these modes can be obtained , taking into account the angular momentum exchange at the resonances ( primarily , cr and olr ) .
rough estimates gave us @xmath367 and @xmath368 ; notably , the olr contribution were dominating in both cases . restricting ourselves to consideration of three terms with @xmath369 in the set of equations ( [ bs2 ] )
, we can study the impact of three main resonances : ilr , cr , and olr .
first of all , we obtained the unstable bar - modes , when all three terms are present ( run 1 ) .
then , the influence of a particular resonance can be `` measured '' separately by omitting an unnecessary term .
in such a way , we study the influence of cr term , switching off the olr term ( run 2 ) .
after that we study the impact of the olr term , omitting cr term ( run 3 ) . in all these three cases , when the principal ilr term is present , the pictures of the unstable modes in the complex @xmath370-plane resemble each other .
the situation has changed when we dropped the ilr term . in the result we obtained unrealistic picture of a sea of unstable modes with growth rates less than @xmath371 .
this fact certainly confirms the dominating role of the ilr term . *
all pattern speeds gained small positive biases , compared to the pattern speeds obtained from the basic integral equation ( [ ie2 ] ) .
* when all three terms are taken into account , both the pattern speed and the growth rate of the most unstable mode ( mode 1 ) is determined with very good accuracy .
separate consideration of cr and olr ( runs 2 and 3 ) shows that the contribution of olr term into the growth rate exceeds one of cr .
* from comparing the eigen - frequency of mode 2 obtained in runs 1 and 3 , it follows that mode s growth is entirely due to the interaction of the gravitational potential of the mode with the resonance stars at olr . in principle , equations ( [ bs2 ] ) allow to consider any number of terms , although it can be checked that the terms with @xmath332 are of less importance .
for example , taking account of terms @xmath372 will not lead to any new unstable modes .
the fastest modes become rather more unstable : @xmath373 , @xmath374 . | this paper presents a new approach to studying galactic structures .
they are considered as the low - frequency normal modes in a disc of orbits precessing at different angular speeds .
such a concept is an adequate alternative to the commonly used approach of treating the disc as a set of individual stars rotating at near - circular orbits around the centre .
the problem of determining the normal modes is reduced to a simple integral equation in the form of the classical eigen - value problem , where the eigen - value is directly equal to the pattern speed of the mode , @xmath0 .
an examination of the general properties of the basic integral equation shows that two types of solutions exist , bar - like and spiral .
the numerical solutions of both types are obtained .
the characteristic pattern speeds are of the order of the mean orbit precession speed , although for the bar - modes @xmath0 can markedly exceed the maximum precessing speed of orbits .
it is shown that the bar - mode grows due to the immediate action of its gravitational field on the stars at the resonance regions .
as for the spiral mode , its excitation is probably due to the inner lindblad resonance that can promote mode growth .
[ firstpage ] keywords : galaxies : structure . |
under large perpendicular magnetic fields @xmath0 the motion of charged carriers in two - dimensional electron systems ( 2des ) is confined to small cyclotron orbitals .
this confinement suppresses the kinetic energy of the electrons but on the other hand amplifies their coulomb interactions . in single layers
these coulomb correlations can lead to the emergence of fractional quantum hall states .
two individual but closely - spaced 2des may also exhibit a correlated state , however , the underlying physics is now also influenced by the coulomb interactions between the two systems . when the electron densities @xmath1 in both layers are identical and the individual filling factors @xmath2 are close to 1/2 ( i.e. , @xmath3 ) , the system may spontaneously develop interlayer phase coherence , provided the distance @xmath4 between the layers is sufficiently small .
this led to the prediction of josephson - type phenomena in bilayer systems @xcite nearly 20 years ago .
, l. v. keldysh and a. n. kozlov [ zh .
54 , 978 ( 1968 ) ] , yu . e lozovik and v.i .
yudson [ jetf lett .
22 , 11 ( 1975 ) ] and s. i. shevchenko [ fiz .
temp . 2 , 505 ( 1976 ) ; sov .
j. low temp .
phys . 2 , 251 ( 1977 ) ] .
] the ratio of the center - to - center layer separation @xmath4 and the magnetic length @xmath5 is commonly used to parameterize the strength of this emerging state at the bilayer s total filling factor of one .
tunneling spectroscopy experiments @xcite demonstrated the phase coherence between the two layers by showing a dramatic enhancement of the tunneling conductance at @xmath3 .
more recently , a critical behavior was observed in dc tunneling experiments @xcite , which had also been predicted @xcite but failed to appear in all prior experiments .
the purpose of this paper is to elucidate on the requirements to observe such a critical behavior .
our studies demonstrate that the coherent tunneling not only intensifies with the size of the sample but more importantly that the critical current grows linearly with the @xmath3 area . in order to deliver a thorough picture of the critical tunneling behavior
we first briefly address those experimental conditions which can easily be manipulated , such as effective layer separation @xmath6 , temperature and filling factor , before discussing the more relevant size dependence .
our data are obtained from samples of two different wafers grown in two different molecular beam epitaxy ( mbe ) machines ( hereby referred to as wafer @xmath7 and @xmath8 ) , both with a net barrier thickness of 9.6 nm , consisting of alternating alas ( 1.70 nm ) and gaas ( 0.28 nm ) layers .
wafer @xmath7 is the same as used in @xcite .
the intrinsic densities of approximately 4.0 @xmath9 @xmath10 ( wafer @xmath7 and @xmath8 ) of the two quantum wells originate from standard modulation doping .
the low - temperature mobilities exceed 450000 @xmath11v@xmath12s@xmath12 ( wafer @xmath7 ) and 500000 @xmath11v@xmath12s@xmath12 ( wafer @xmath8 ) .
the single - particle tunnel splitting @xmath13 in our double quantum wells was estimated to be approximately 150 @xmath14k .
independent electrical contact to the two layers is achieved by growing the double quantum wells onto prestructured back gates @xcite and by using additional top gates in order to exploit a selective depletion technique @xcite . as the distance between
back gate and 2des is only around 1 @xmath14 m , this overgrown back gate technique requires voltages of less than 1 v to locally deplete the contact arms and obtain independent electrical contacts to the two layers .
the samples were either patterned into hall bars of different sizes or a quasi - corbino ring @xcite .
the specific sample dimensions will be given later in the text .
characteristics are measured at several magnetic fields around @xmath3 .
the eight most inner curves are marked with symbols .
( b ) maximal ( critical ) currents as a function of the total filling factor ( data from labeled curves in subfigure ( a ) are used ) . at exactly @xmath3 the effective layer separation is 1.42 .
hall bar sample of 0.88 @xmath15 0.08 mm@xmath16 ( wafer @xmath7 ) and @xmath17mk.**,scaledwidth=49.0% ] the dc @xmath18 tunneling measurements presented in this paper are performed as sketched in figure [ fig1 ] ( b ) , by applying a tuneable dc bias voltage ( herein after referred to as 2-terminal voltage v@xmath19 ) between the two layers and detecting the current flow @xmath20 toward ground as a voltage drop over a known resistance @xcite .
as the interlayer phase coherence at @xmath3 allows to easily transfer charges between the layers , the tunnel conductance becomes enormously enhanced which is tantamount to a very small tunnel resistance .
consequently , even if there is a finite bias @xmath21 applied , the interlayer voltage probes a and b ( located close to source and drain ) may read a value for @xmath22 of close to zero ( figure [ fig1 ] ( a ) ) .
however , this coherent tunneling can be destroyed if the current ( figure [ fig1 ] ( c ) ) exceeds a critical value @xmath23 .
for the representative measurement presented in figure [ fig1 ] the critical value is roughly @xmath241.5 na . figure [ fig1 ] ( d ) is a depiction of the measured current @xmath20 over the measured voltage @xmath22 .
this representation will be used throughout this paper , and the critical currents @xmath25 now translate into the maximal positive and negative currents at v@xmath26=0 as indicated by the dashed lines .
the data here are presented in a scatter plot which show a discontinuity in the measured current and voltage characteristics when the system moves from the coherent strong tunneling regime into the weak tunneling regime .
the origin of this negative differential conductance is the sudden change in the total impedance @xmath27 , when a tunneling resistance @xmath28 of almost zero is replaced by a resistance much larger than the series resistance @xmath29 of the ( uncorrelated ) quantum hall systems , i.e. , @xmath30 .
historically , the study of coherent tunneling at @xmath3 exploits the tunneling - spectroscopy technique where the differential tunneling conductance @xmath31 is obtained in an ac measurement .
these tunneling - spectroscopy measurements ( tsm ) reveal a resonantly enhanced zero bias tunneling peak at a total filling factor of one @xcite . a critical behavior as discussed in this paper , however ,
is hidden in the dc part of the tsm which is usually not shown . if the critical current is intrinsically small , it may also be difficult to detect or conceal through the influence of the ac modulation .
it does not come as a surprise that the critical tunneling behavior ( or the signatures of the @xmath3 state in general ) , is strongest when the total filling factor of the system is exactly 1 , or when the individual filling factors are exactly 1/2 , respectively.=@xmath32 .
however , imbalancing the electron densities can be used as a tool to induce and intensify the @xmath3 state .
see spielman _
et al . _ ,
phys . rev .
b 70 , 081303 ( 2004 ) and @xcite for details . ]
deviating to larger fields ( smaller total filling factors ) or smaller fields ( larger total filling factors ) will strongly suppress the coherent tunneling . in the uncorrelated regimes , carriers
are exchanged between two individual two - dimensional electron systems , a process which now requires a finite amount of energy to overcome the coulomb repulsion between the electrons .
figure [ fig4 ] ( a ) provides a qualitative picture of this behavior .
it shows a set of several 4-terminal @xmath18 curves measured on a hall bar sample of 0.88 @xmath15 0.08 mm@xmath16 ( wafer @xmath7 ) at @xmath3 ( @xmath33 ) and for small offsets in steps of about @xmath34 thereof .
while drifting to either side of @xmath3 , the critical tunneling behavior is getting progressively suppressed around zero bias .
our data suggest that moving toward higher fields suppresses the tunneling slightly more rapidly than in the opposite low field direction as the coulomb exchange increases with @xmath35 .
figure [ fig4 ] ( b ) gives a quantitative analysis of the maximum ( critical ) current as a function of the total filling factor .
note that only the eight plots in the center of figure [ fig4 ] ( a ) allowed the determination of a maximal current .
as the total filling factor one state resides within a large parameter space , the size dependence as we are about to discuss can not be studied fully independent of other important parameters . the purpose of sec .
iv is thus to outline the effects of manipulating temperature and effective layer separation @xmath6 which largely influence the magnitude of the critical current as well .
the latter is achieved experimentally by increasing the electron densities in both layers simultaneously and adjusting the magnetic field , i.e. , @xmath36 .
figure [ fig3 ] ( a ) shows the positive and negative critical currents as a function of @xmath6 for a hall bar sample of 0.88 @xmath15 0.08 mm@xmath16 ( wafer @xmath8 ) at below 20 mk .
for @xmath37 the current appears to saturate at 2 na , even to decrease .
this however is related to the effect of the sample s very low electron density , which begins to suppress the transport current altogether ( the single layer density is roughly 1.65 @xmath15 10@xmath38 m@xmath39 ) . for intermediate @xmath6 on the other hand
the trend is clearly linear . for @xmath40
the system undergoes a phase transition and the critical tunneling behavior ( and the @xmath3 qh state ) disappears .
this value is in very good agreement with those found in magneto - transport @xcite or tunneling - spectroscopy experiments @xcite .
generally , if the onset of the @xmath3 state is observed at @xmath41 , its origin is a pure many - body effect , and in weakly tunneling samples ( i.e. , @xmath42 ) the phase - coherence would develop spontaneously @xcite .
the smooth phase transition we observed supports a `` puddle model '' as suggested by a. stern and b. i. halperin @xcite where the @xmath3 phase breaks up into domains near the phase boundary . in this model
, the type of phase transition would have to be of first order as two phases co - exist in the sample .
figure [ fig3 ] ( c ) shows three 4-terminal @xmath18 curves corresponding to a @xmath6 of 1.40 , 1.61 and 1.79 .
the expected phase transition for correlated bilayers occurring at finite temperatures is not a regular second order phase transition such as for normal superconductors at zero field or ferromagnets but a kosterlitz - thouless type of phase transition .
however , standard transport experiments are not able to judge type and form of the occurring phase transition .
nevertheless , what we see experimentally in transport ( i.e. , tunneling ) at finite temperatures is summarized in figure [ fig3 ] ( b ) . there we plot the critical current as a function of the inverse temperature for a fixed @xmath33 measured on a hall bar sample of 0.88 @xmath15 0.08 mm@xmath16 ( wafer @xmath7 ) .
figure [ fig3 ] ( d ) shows several corresponding 4-terminal @xmath18 curves . at a temperature exceeding approximately 80 mk ,
the critical current begins to decrease rapidly .
extrapolation of the data in this region indicates the suppression of the critical behavior for temperatures above 250 mk .
the @xmath3 qh state as observed in magneto - transport is very weak at 250 mk and disappears entirely at temperatures exceeding approximately 350 mk @xcite .
the exact type of this phase transition is unknown to us .
spielman @xcite found similar overall trends on the temperature and @xmath6 in tunneling spectroscopy measurements using samples which display very small critical currents .
the following study of the size dependence was motivated by magneto - transport and tunneling measurements performed on a corbino ring @xcite .
these experiments had shown a vanishing conductance across the annulus , suggesting that the bulk of the @xmath3 phase is incompressible .
as any interedge charge transfer is suppressed , tunneling would then ( generally ) only occur in the vicinity of the edges of the coherent @xmath3 system and its magnitude would have to scale with the circumference of the sample .
tunneling spectroscopy measurements on hall bar samples on the other hand indicated that the zero bias tunneling conductance may be related to the area of the sample instead @xcite . to obtain a better understanding and trying to solve this contradiction we compared the critical currents , rather the properties of the tsm tunneling peak , in terms of the circumference and area . comparing different samples of course may introduce a systematic error , yet we still found that the effect of different sample sizes had much more dramatic consequences as we will see next .
( right - hand side ) and the @xmath3 sample area @xmath43 ( left - hand side ) for @xmath441.6 for samples from wafer @xmath7 and @xmath8 .
fitting yields a linear dependence on the area .
see text for further discussions .
( b ) corresponding 4-terminal @xmath18 curves .
the sample dimensions are ( from largest to smallest ) : corbino ring : @xmath45=860 @xmath14 m , @xmath46=270 @xmath14 m ; hall bar type 1 : 0.88 @xmath15 0.08 mm@xmath16 ; hall bar type 2 : 0.15 @xmath15 0.05 mm@xmath16.**,title="fig:",scaledwidth=49.0% ] + ( right - hand side ) and the @xmath3 sample area @xmath43 ( left - hand side ) for @xmath441.6 for samples from wafer @xmath7 and @xmath8 .
fitting yields a linear dependence on the area .
see text for further discussions .
( b ) corresponding 4-terminal @xmath18 curves .
the sample dimensions are ( from largest to smallest ) : corbino ring : @xmath45=860 @xmath14 m , @xmath46=270 @xmath14 m ; hall bar type 1 : 0.88 @xmath15 0.08 mm@xmath16 ; hall bar type 2 : 0.15 @xmath15 0.05 mm@xmath16.**,title="fig:",scaledwidth=47.0% ] figure [ fig2 ] ( a ) plots the value of the critical current as a function of the sample circumference @xmath47 and as a function of the sample s @xmath3 area @xmath43 when the effective layer separation @xmath6 for all samples is @xmath48 1.6 .
the data points are labeled with the wafer index @xmath7 or @xmath8 .
our data clearly indicates that the critical current @xmath23 grows as the sample increases in size .
when we apply a linear fit in the @xmath49 diagram , we find that the trend is best described by 1.04@xmath50 for the area and 2.11@xmath51 for the circumference .
based on these data we propose that the parameter that determines coherent tunneling at @xmath3 is the sample area with a linear dependence and not the circumference .
figure [ fig2 ] ( b ) shows the corresponding 4-terminal @xmath18 curves .
please note that the sample with the largest area ( and circumference ) displayed a strong hysteresis between up- and downsweeping the applied voltage @xmath21 .
subfigure ( b ) shows the upsweep , where the negative and positive critical currents differ , i.e. , @xmath52 . in a subsequent downsweep ,
the situation and the absolute values are reversed , i.e. , @xmath53 and @xmath54(upsweep)=@xmath55(downsweep ) .
for that reason , the fits in figure [ fig2 ] ( a ) ignore the error bar of these data points . for comparison we extracted a critical current of 17 pa at @xmath44 1.6 from @xcite .
this value originates from the dc part of a tsm on a 250 @xmath15 250 @xmath14m@xmath16 sample .
we believe that the disagreement between our data and this value is related to a different single particle tunneling splitting @xmath13 , which is determined by the height and width of the tunneling barrier , and to the large enhancement of the tunneling amplitude at @xmath3 @xcite .
their double quantum well is differently designed , with 18 nm gaas quantum wells separated by a 9.9 nm barrier layer .
the center - to - center separation @xmath4 , however , is nearly identical to ours . our studies indicate that the initial notion of a length dependence on tunneling as mentioned at the beginning of this chapter is a misinterpretation due to the astonishing properties of the coherent phase .
coherent inter - edge tunneling in a corbino topology is suppressed because injected electrons do not become part of the @xmath3 state . in the excitonic condensate picture
this can be understood as it is impossible for the condensate to create an interlayer exciton by placing a hole at the site of the injected electron , when the drain lead where the hole has to originate from is far away on the other side of the bulk .
true coherent tunneling on the other hand occurs when new interlayer excitons are created or existing ones are annihilated .
these processes appear to be bulk phenomena after all .
however , as the @xmath3 phase could break up into domains near the phase boundary @xcite , it is possible that certain physical conditions may arise which yield a different dependence on the area ( and circumference ) than the one presented here .
generally , the observation of a critical behavior requires - vaguely speaking - `` sufficiently large samples '' .
the tangible sample size depends on the underlying design of the double quantum well and the sample structure , but may also be influenced by other unknown factors .
more specifically , our studies identify the area rather than the circumference as the determining parameter for coherent tunneling at @xmath3 .
a tunneling experiment in the general sense of charge transfer between two electron reservoirs through a sufficiently thin barrier is an inappropriate interpretation for the peculiar case of the total filling factor one state . instead ,
as the two layers are considered to be indistinguishable , correlated interlayer tunneling is a direct signature of interlayer phase coherence .
the model of indistinguishable layers also implies that the critical behavior and its dependence on the parameters we discussed in this paper have several implications for magneto - transport experiments performed in the regime of the total filling factor one state @xcite .
it can be shown that the leakage or tunneling current in the counterflow configuration @xcite depends on whether the driving current in the system is larger or smaller than the critical current for a given condition @xcite , i.e. , @xmath6 , temperature and sample size .
the observed gap in magneto - transport on the other hand is only slightly altered .
it nevertheless implies that temperature - activation measurements at a total filling factor of one @xcite require a much more careful interpretation of the gap energy .
what makes matters experimentally complicated is the strong dependence of the critical current @xmath23 on the size of the sample , meaning that for samples of certain dimensions the transport current may already be much larger than the critical current , even when the system is at base temperature .
this fact could be able to account for several unsettled observations such as finite dissipation in counterflow experiments for example @xcite .
future magneto - transport experiments in the regime of the coherent total filling factor one state need to take the relevance of the driving current into account .
in summary , we discussed the relevant parameters necessary for the observation of a critical behavior in the coherent @xmath3 state .
we find a linear dependence of the value of the critical tunneling current @xmath23 on the @xmath3 area of the sample .
@xmath23 will decrease when the system is brought toward the phase boundary by increasing @xmath6 or the temperature .
it is also rapidly destroyed by moving away from a total filling factor of one .
we would like to thank both a. h. macdonald and m. gilbert for discussions and comments .
our bilayer wafers were grown in collaboration with m. hauser ( max - planck institute , stuttgart ) and h .-
tranitz ( university of regensburg ) .
especially , we would like to acknowledge j. g. s. lok for his initial and intensive work on our electron double layer systems .
this project was supported by the bmbf ( german ministry of education and research ) under grant no . 01bm456 .
h. a. fertig , _ phys .
b _ * 40 * 1087 ( 1989 ) z. f. ezawa and a. iwazaki , _ int
b _ * 6 * 3205 ( 1992 ) x.
g. wen and a. zee , _ phys .
lett . _ * 69 * 1811 ( 1992 ) i. b. spielman , j. p. eisenstein , l. n. pfeiffer and k. w. west , _ phys .
lett . _ * 84 * 5808 ( 2000 ) i. b. spielman , j. p. eisenstein , l. n. pfeiffer and k. w. west , _ phys . rev .
lett . _ * 87 * 036803 ( 2001 ) z. f. ezawa and a. iwazaki , _ int
b _ * 8 * 2111 ( 1994 ) l. tiemann , w. dietsche , m. hauser and k. v. klitzing ,
_ new j. phys . _ * 10 * 045018 ( 2008 ) h. rubel , a. fischer , w. dietsche , k. v. klitzing and k. eberl , _ phys .
lett . _ * 78 * 1763 ( 1997 ) j. p.
eisenstein , l. n. pfeiffer and k. w. west _ appl .
* 57 * 2324 ( 1990 ) l. tiemann , w. dietsche , m. hauser and k. v. klitzing , _ phys .
b _ * 77 * 033306 ( 2008 ) r. d. wiersma , j. g. s. lok , l. tiemann , w. dietsche , k. von klitzing , d. schuh and w. wegscheider , _ physica e _ * 35 ( 2 ) * 320 ( 2006 ) a. d. k. finck , a. r. champagne , j. p. eisenstein , l. n. pfeiffer and k. w. west , _ phys . rev .
b _ * 78 * 075302 ( 2008 ) a. r. champagne , j. p. eisenstein , l. n. pfeiffer and k. w. west , _ phys
lett . _ * 100 * 096801 ( 2008 ) s. q. murphy , j. p. eisenstein , g. s. boebinger , l. n. pfeiffer and k. w. west , _ phys .
lett . _ * 72 * 728 ( 1994 ) i. b. spielman , ph.d .
thesis , caltech ( 2004 ) e. rossi , a. s. nunez and a. h. macdonald , _ phys .
lett . _ * 95 * 266804 ( 2005 ) a. stern and b. i. halperin , _ phys .
lett . _ * 88 * 106801 ( 2002 ) z. f. ezawa , _ quantum hall effects - field theoretical approach and related topics _ ,
2nd edition ( world scientific , 2008 ) m. kellogg , j. p. eisenstein , l. n. pfeiffer and k. w. west , _ phys .
lett . _ * 93 * 036801 ( 2004 ) e. tutuc , m. shayegan and d. a. huse , _ phys .
lett . _ * 93 * 036802 ( 2004 ) y.yoon , l. tiemann , w. dietsche and k. von klitzing , to be submitted r. d. wiersma , j. g. s lok , s. kraus , w. dietsche , k. von klitzing , d. schuh , m. bichler , h .-
tranitz and w. wegscheider _ phys . rev
. lett . _ * 93 * 266805 ( 2004 ) | we will discuss the relevant conditions to observe a critical tunneling current [ new j. phys .
10 , 045018 ( 2008 ) ] in electron - double layer systems at a total filling factor of one and find they are related to the effective layer separation and the temperature .
our studies suggest that the intensity of the critical tunneling behavior is also directly linked to the area of the sample . |
the muon lifetime is one of the key observables of today s particle physics . not only is it measured very precisely , since the current experimental error is 18 ppm @xcite , but can be described to competing accuracy within the standard model , giving rise to a strong correlation between the masses of the heavy gauge bosons . as a low energy process
, the decay is expected to be governed by an effective interaction involving only the electron , muon and their respective neutrinos .
the dynamics of the system should be corrected mostly by qed , whereas the electroweak interactions determine solely the size of the coupling constant .
the history of the calculation of the electroweak corrections , in which we are interested here , is rather long .
it started in the early eighties with the one loop contributions @xcite .
subsequently , leading terms in the top quark @xmath4 @xcite and higgs boson @xmath5 @xcite masses were derived at the two loop level . in the meantime ,
mixed electroweak and qcd corrections became available at order @xmath6 @xcite and @xmath7 @xcite .
recently , three loop leading terms in the top quark mass @xmath8 and @xmath9 have also been calculated @xcite .
as far as the pure two loop electroweak corrections are concerned , after it turned out that the subleading terms in the top quark mass expansion are comparable with the leading ones @xcite , complete fermionic and the higgs boson mass dependence of the bosonic contributions have been evaluated @xcite .
the complete bosonic part has been done in @xcite .
it is the purpose of the present paper to present the result of a new independent calculation of the fermionic contributions and , after inclusion of the bosonic corrections , also of the full electroweak corrections .
this work is organized as follows . in the next section ,
we discuss the matching procedure and fermion masses as infrared regulators , which avoid ambiguities of the definition of gamma matrices in box diagrams in noninteger dimensions .
then , we present the results for the fermionic and full contributions and compare them with previous calculations by specifying the differences in the @xmath0 boson mass prediction .
conclusions close the paper .
due to a large number of very different mass scales , it is virtually impossible to evaluate the muon decay lifetime directly within the standard model , without recourse to some approximation method .
an elegant and systematic approximation is provided by the approach based on _ effective theories_. the idea is to agree on some cutoff scale , below which all degrees of freedom are treated exactly , whereas the heavier fields are `` integrated out '' , which means that they generate effective interactions .
it should not be surprising that one first discovers experimentally the effective theories , since the dependence on the heavier scales requires higher `` resolution '' , _ i.e. _ higher energy . for precisely this reason ,
the effective theory governing muon decay , the fermi model , has been known much before the standard model . from this point of view
, one should not consider that the fermi model is used in current calculations for _ historical reasons _ , but because it is the appropriate effective theory at this energy scale .
the approximation is constructed as follows .
the lagrangian is made only from the light fields , which are the six leptons , the five quarks , the photon and the gluon . at leading order in the inverse heavy scale ,
for which we take the @xmath0 boson mass @xmath10 , a single effective operator is added , giving the lagrangian ( in the so called charge conserving form of the fermi operator ) @xmath11 where the superscript @xmath12 denotes bare quantities .
the theory is finite after mass and coupling ( @xmath13 and @xmath14 ) renormalization to all orders in @xmath13 and @xmath14 , and leading order in the fermi constant @xmath15 , which is why this parameter is not renormalized .
the _ matching procedure _ in the present case consists in requiring that the amputated renormalized green functions of the effective theory be equal to the amputated renormalized green functions of the standard model up to terms of order @xmath16 and given order in @xmath13 and @xmath14 @xmath17 which makes the muon decay amplitude the same in both models up to the specified order .
the fermi constant is then given as an expansion in @xmath13 and @xmath14 @xmath18 with @xmath19 being the born level prediction .
the quantity @xmath20 is customarily used to parametrize the higher order contributions . at the one loop level , the matching equation is schematically depicted in fig . [ matching1l ] .
we introduced there the decoupling coefficients @xcite , @xmath21 , which are different from one in the @xmath22 scheme for example , but can be neglected in the on - shell scheme .
the renormalization constant of the fermi operator @xmath23 , although trivial ( _ i.e. _ equal to one ) , has been included for generality . the matching equation , eq .
[ matching ] , can be solved in different ways .
the apparently simplest is to put all light masses and external momenta to zero , and renormalize the wave functions in the on - shell scheme . the right hand side in fig .
[ matching1l ] will then consist of only one term , proportional to @xmath24 , if we use dimensional regularization , whereas the left will only have vacuum diagrams with heavy masses .
obviously , this situation will persist to all orders .
the price to pay for this simplicity is the problem of infrared divergent box diagrams , where a product of gamma matrices occurs which does not have the form of the fermi operator . in @xcite
, this product has been defined through fierz symmetry with respect to the last line in the string , which has been implemented , for practical reasons , by means of a suitable projection operator .
although sufficient at the two loop level , this symmetry will not suffice at the three loop level , see for example fig .
[ problem ] , where due to crossings , there is no last line in this sense .
it is also not trivial that this procedure is correct even at the two loop level . for many topologies , _
e.g. _ those that contain a self energy insertion on the gauge boson line , one can convince oneself that this is indeed the case , others like the nonplanar double boxes in the purely bosonic contributions are not that easy .
a highly nontrivial test of the calculation would consist in performing the matching without generating spurious infrared divergences .
the box diagrams being ultraviolet finite can then be calculated in four dimensions , avoiding completely the problem of ambiguous gamma matrix definitions . in this work we performed the matching by keeping a common mass for all the light fermions and evaluating the box diagrams in four dimensions .
it was necessary to calculate both sides of the matching equation and reexpand them subsequently in this common mass .
the external wave function renormalization constants were not taken in the on - shell scheme , because this would introduce the usual on - shell infrared divergence . on the contrary ,
they were evaluated at zero momentum , which , in practice , is equivalent to renormalization in the @xmath22 scheme with nonvanishing decoupling coefficients .
moreover , it turned out that it is necessary to have a correct @xmath0 boson wave function renormalization constant , since the box diagrams are not gauge invariant by themselves , and in the massive case this constant cancels only in combination with vertex diagrams .
we used a photon mass regulator , but the @xmath25 renormalization constant would have been just as good . in the end , complete agreement was found with the calculation performed with massless fermions and with the projector conserving fierz symmetry with respect to the last line from @xcite .
a detailed presentation of the methods used to evaluate the bosonic contributions to @xmath26 can be found in @xcite .
the fermionic contributions introduce two additional problems .
first , some of the two loop vertex diagrams contain closed triangular fermion loops as shown in fig [ triangle ] .
the @xmath27 matrix that occurs in the trace has to be correctly defined .
we chose the naive dimensional regularization scheme @xcite , with an anticommuting @xmath27 and the four dimensional value of the trace of four gamma matrices and @xmath27 @xmath28 this choice is justified by the fact that the nonvanishing contribution of the purely four dimensional @xmath29 tensors is finite .
moreover , it has been checked in @xcite that the use of the consistent definition of t hooft and veltman @xcite gives the same result after correction of the green functions by suitable finite counterterms restoring the slavnov taylor identities .
second , the inclusion of fermions results in unstable gauge bosons , which makes a proper definition of their masses necessary if gauge invariance of @xmath30 is to be maintained @xcite .
we use the pole mass scheme , where the inverse propagator matrix @xmath31 is singular in the complex @xmath32 plane at points which can be parametrized as @xmath33 where @xmath34 is the mass and @xmath35 is the width of the boson .
this generates a fixed width breit wigner behavior of the total cross section @xmath36 as opposed to the running width parametrization actually used by the experimental collaborations for the masses and widths of the @xmath0 and @xmath37 bosons @xcite @xmath38 we translate back and forth between the two definition with the help of the following relations @xmath39 as in @xcite , we take @xmath40 as experimentally measured , whereas we assume @xmath41 to be given by the one loop qcd corrected value @xmath42 the complete result for @xmath20 at order @xmath43 and the partial contributions are given in fig .
[ dr ] .
the top quark mass and the running of the fine structure constant are taken from tab .
[ input ] , whereas the masses of the gauge bosons are translated from the experimental values given there to the pole mass scheme values with the help of eq .
[ translation ] , which in this case gives @xmath44 gev and @xmath45 gev .
.[input ] input parameters with experimental errors , where necessary for the present work .
the value of @xmath46 is the same as in @xcite for comparison purposes .
[ cols="<,<,<",options="header " , ] together with errors coming from the top quark mass and the running of the fine structure constant but without a theoretical error estimate , the @xmath10 prediction is shown against the current experimental result in fig . [ mw ] .
we have presented a new result for the complete electroweak contributions to the lifetime of the muon , which induces a shift in the @xmath0 boson mass prediction as large as @xmath2 mev for a light higgs boson , of which @xmath1 mev come from a discrepancy with the previous calculation of the fermionic contributions @xcite and the rest from the bosonic part .
the authors of @xcite corrected their evaluation and are now in full agreement with this work .
together with recent results at the three loop level @xcite , this calls for an updated fitting formula .
such a formula will be given in a subsequent publication @xcite .
the authors would like to thank a. freitas for his effort put into the comparison of the present results and the results of @xcite and for tracing an error in an earlier version of this paper , and g. weiglein for reading the manuscript .
the warm hospitality of the institute for particle physics phenomenology of the university of durham during the time when part of this work was completed is gratefully acknowledged .
this work was supported in part by the kbn grant 5p03b09320 .
a. freitas , w. hollik , w. walter and g. weiglein , phys .
b * 495 * ( 2000 ) 338 ; a. freitas , w. hollik , w. walter and g. weiglein , nucl . phys .
b * 632 * ( 2002 ) 189 .
m. awramik and m. czakon , phys .
* 89 * ( 2002 ) 241801 ; _ see also _
m. awramik and m. czakon , arxiv : hep - ph/0211041 .
a. onishchenko and o. veretin , phys .
b * 551 * ( 2003 ) 111 .
m. awramik , m. czakon , a. onishchenko and o. veretin , arxiv : hep - ph/0209084 . k. hagiwara _ et al . _
[ particle data group collaboration ] , phys .
d * 66 * ( 2002 ) 010001 .
a. sirlin , phys .
d * 22 * ( 1980 ) 971 ; w. j. marciano and a. sirlin , phys . rev .
d * 22 * ( 1980 ) 2695 [ erratum - ibid .
d * 31 * ( 1985 ) 213 ] .
j. j. van der bij and f. hoogeveen , nucl .
b * 283 * ( 1987 ) 477 ; r. barbieri , m. beccaria , p. ciafaloni , g. curci and a. vicere , phys .
b * 288 * ( 1992 ) 95 [ erratum - ibid .
b * 312 * ( 1993 ) 511 ] ; r. barbieri , m. beccaria , p. ciafaloni , g. curci and a. vicere , nucl .
b * 409 * ( 1993 ) 105 ; j. fleischer , o. v. tarasov and f. jegerlehner , phys .
b * 319 * ( 1993 ) 249 .
j. van der bij and m. j. veltman , nucl .
b * 231 * ( 1984 ) 205 ; j. j. van der bij , nucl . phys .
b * 248 * ( 1984 ) 141 ; f. halzen , b. a. kniehl and m. l. stong , z. phys .
c * 58 * ( 1993 ) 119 ; f. jegerlehner , prog . part .
* 27 * ( 1991 ) 1 .
a. djouadi and c. verzegnassi , phys .
b * 195 * ( 1987 ) 265 ; a. djouadi , nuovo cim .
a * 100 * ( 1988 ) 357 ; b. a. kniehl , nucl .
b * 347 * ( 1990 ) 86 ; f. halzen and b. a. kniehl , nucl .
b * 353 * ( 1991 ) 567 ; b. a. kniehl and a. sirlin , nucl .
b * 371 * ( 1992 ) 141 ; b. a. kniehl and a. sirlin , phys .
d * 47 * ( 1993 ) 883 ; a. djouadi and p. gambino , phys .
d * 49 * ( 1994 ) 3499 [ erratum - ibid .
d * 53 * ( 1996 ) 4111 ] .
l. avdeev , j. fleischer , s. mikhailov and o. tarasov , phys .
b * 336 * ( 1994 ) 560 [ erratum - ibid .
b * 349 * ( 1995 ) 597 ] ; k. g. chetyrkin , j. h. khn and m. steinhauser , phys . lett .
b * 351 * ( 1995 ) 331 ; k. g. chetyrkin , j. h. khn and m. steinhauser , phys . rev . lett .
* 75 * ( 1995 ) 3394 .
m. faisst , j. h. khn , t. seidensticker and o. veretin , arxiv : hep - ph/0302275 .
g. degrassi , p. gambino and a. vicini , phys .
b * 383 * ( 1996 ) 219 ; g. degrassi , p. gambino and a. sirlin , phys .
b * 394 * ( 1997 ) 188 .
s. weinberg , phys .
b * 91 * ( 1980 ) 51 ; b. a. ovrut and h. j. schnitzer , phys .
b * 100 * ( 1981 ) 403 ; w. wetzel , nucl .
b * 196 * ( 1982 ) 259 ; w. bernreuther and w. wetzel , nucl .
b * 197 * ( 1982 ) 228 [ erratum - ibid .
b * 513 * ( 1998 ) 758 ] ; w. bernreuther , annals phys .
* 151 * ( 1983 ) 127 .
w. bernreuther , z. phys .
c * 20 * ( 1983 ) 331 .
a. j. buras and p. h. weisz , nucl .
b * 333 * ( 1990 ) 66 .
m. misiak and j. urban , phys .
b * 451 * ( 1999 ) 161 .
m. s. chanowitz , m. furman and i. hinchliffe , nucl .
b * 159 * ( 1979 ) 225 .
g. t hooft and m. j. veltman , nucl .
b * 44 * ( 1972 ) 189 .
a. sirlin , phys .
* 67 * ( 1991 ) 2127 ; r. g. stuart , phys
. lett .
b * 262 * ( 1991 ) 113 ; s. willenbrock and g. valencia , phys .
b * 259 * ( 1991 ) 373 ; h. g. veltman , z. phys .
c * 62 * ( 1994 ) 235 ; m. passera and a. sirlin , phys .
d * 58 * ( 1998 ) 113010 ; p. gambino and p. a. grassi , phys .
d * 62 * ( 2000 ) 076002 ; a. r. bohm and n. l. harshman , nucl .
b * 581 * ( 2000 ) 91 .
d. y. bardin , a. leike , t. riemann and m. sachwitz , phys .
b * 206 * ( 1988 ) 539 .
t. van ritbergen and r. g. stuart , phys .
* 82 * ( 1999 ) 488 .
f. jegerlehner , arxiv : hep - ph/0105283 ; _ see also _
f. jegerlehner , j. phys .
g * 29 * ( 2003 ) 101 ; k. g. chetyrkin , j. h. khn and m. steinhauser , phys . rev .
* 75 * ( 1995 ) 3394 .
a. freitas , w. hollik , w. walter and g. weiglein , arxiv : hep - ph/0007091 , arxiv : hep - ph/0202131 . | an independent result for the two loop fermionic contributions to the muon lifetime in the standard model is obtained .
deviations are found with respect to @xcite , which result in a shift of the @xmath0 boson mass by approximately @xmath1 mev over the range of higgs boson masses from 100 gev to 1 tev .
supplied with the bosonic contributions from @xcite , this shift , due to the complete electroweak contributions , varies from @xmath2 mev to @xmath3 mev .
additionally , a new test of the matching procedure defining the fermi constant is presented , which uses fermion masses as infrared regulators . and |
@xmath4-branes interacting with higher form rr - fields ( @xmath5-form for example ) become @xmath6-branes , which was first suggested by emparan @xcite and nontrivial interactions were explicitly obtained by myers @xcite . in the presence of magnetic rr - fields ,
these interactions are caused by the motion of the branes . with the motion in compact space ( thus carrying angular momentum ) , the higher dimensional brane increases its size .
since the size is bounded in the compact space , the angular momentum is also bounded .
this fact possibly explains the stringy exclusion principle in the dual gravity setup @xcite .
recently mateos and townsend showed that this angular momentum can be given in a different guise @xcite . in some special setup of tubular @xmath7-brane
, the poynting vector of the electromagnetic field on the @xmath7 world - volume provides the angular momentum .
this can be thought of as the ` blown - up ' configuration of a bunch of @xmath2 superstrings with @xmath3-branes evenly distributed on it .
it is self - supported from collapse by the angular momentum supplied by the electric and magnetic field associated with the number of @xmath2 superstrings and @xmath3-branes respectively .
the important fact is that the tube solution preserves one quarter of supersymmetry .
this was generically possible for the intersecting @xmath8-branes with relative codimension four @xcite . in this paper
we pursue the issue further to see how this effect can be understood in the t - dual setup .
several string duality transformations will yield straightforward generalizations of supertube . in ref .
@xcite , s - dual configurations ( therefore of m theory ) to the supertube were also discussed .
these configurations will generate lots of lower dimensional descendants upon different compactifications , which are to be related with one another via u - dualities .
although these are the naive expectations , recent interests @xcite on this subject warrant to produce more explicit results .
we first show that an array of @xmath4-brane @xmath9 along some axis , say @xmath10-axis ( see eq .
( [ metric ] ) ) , when threaded vertically by superstrings over the entire volume of @xmath4-branes , is ` blown - up ' to @xmath11-brane of topology @xmath12 acquiring extra tubular two dimensions .
this is obtained by taking t - duality along various directions transverse to @xmath2 superstrings carrying @xmath3-brane charges .
the radius of the circle @xmath13 is invariant under these t - dualities .
an astonishing result is obtained when we take t - duality along the axial direction of the configuration .
the bound state of @xmath3-branes and @xmath2 superstrings becomes that of @xmath0-branes with @xmath1 superstrings moving in one direction on it .
we show that its corresponding ` blown - up ' configuration is a single helical @xmath0-brane ( @xmath8-helix ) traveling with the speed of light along its axis .
this is peculiar in that the dimensionality is not changed upon the ` blowing - up ' .
this @xmath8-helix should be related with the helical @xmath2 string discussed in ref .
@xcite via a sequence of s- and t - duality .
before transferring to the next section , we start by briefly recapitulating the results of ref .
@xcite , where the configuration of @xmath3-branes evenly arrayed along @xmath10-direction and threaded by a bunch of @xmath2 superstrings was considered .
the configuration is embedded in flat geometry parametrized as @xmath14 therefore it is free from any background gravitational effect and also there is no background field of any kind .
it was shown in ref .
@xcite that this can be considered as zero radius limit of tubular @xmath7-brane .
@xmath3-brane charge is ` dissolved ' as magnetic flux on @xmath7-brane while @xmath2 superstrings are dissolved as the electric field along @xmath10-direction . with the static gauge for the world - volume
coordinates @xmath15 on @xmath7-brane , born - infeld ( bi ) 2-form field strength is given by @xmath16 the lagrangian for the tubular @xmath7-brane is that of dirac - born - infeld ( dbi ) simplified as @xmath17 for fixed momentum @xmath18 and magnetic field @xmath19 , the hamiltonian @xmath20 is minimized at @xmath21 .
the same physics can be viewed from @xmath3-brane side . in this case
, the system is described by the matrix model .
some extended solutions including multi - supertube configurations were found in ref .
@xcite . in the next section ,
we consider the cases of planar @xmath4-brane array threaded vertically by superstrings .
these are obtained by t - dual transformation along various directions transverse to the @xmath2 superstrings carrying @xmath3-brane charge . in section iii , we study the case obtained by t - dual transformation along longitudinal direction of the original configuration . we show that the blown - up configuration is the helical @xmath0-brane moving with light velocity along its axis . in section iv , we show explicitly @xmath22 of the supersymmetry of @xmath1 minkowski vacuum is preserved in the @xmath8-helix configuration . in the last section we conclude with some discussions and remarks on further works .
the first question that arises from the supertube physics is whether similar ` blowing - up ' happens in @xmath1 setup .
one simple way to see this is to take t - duality along some directions for the original @xmath2 superstrings carrying @xmath3-brane charge and check whether the same t - duality for the supertube gives sensible ` blown - up ' configuration .
this is based on the fact that supersymmetry is preserved under t - duality @xcite . since the supertube configuration encodes those charges of @xmath3-brane array and @xmath2 superstrings , so must do its t - dual counterpart .
we first take t - duality along some directions transverse to @xmath2 superstring . the @xmath3-brane array threaded by @xmath2 superstrings
are dualized to be @xmath0 array crossed by a bunch of @xmath1 superstrings .
we expect @xmath23-brane of topology @xmath24 as the ` blown - up ' configuration .
this @xmath23-brane is nothing but t - dual version of @xmath2 supertube . to be more specific ,
we take @xmath25-direction of @xmath26 as the t - dual direction . taking t - duality directly on dbi action
is not simple .
it is very obscure in the dbi action to start from @xmath27 matrix and constrain its components ( using orbifold technique used in ref .
@xcite ) to describe @xmath8-brane array along some compact direction .
instead we take an indirect way . with the knowledge about bi fields on the @xmath23-brane which encodes the dissolved @xmath0-branes and @xmath1 superstrings
, we construct dbi action for this @xmath1 setup . on the resulting @xmath23-brane ,
the array of @xmath0-branes is dissolved as magnetic flux and the number of @xmath1 superstrings is encoded as the electric field along @xmath10-direction . in the static gauge for the additional world - volume coordinate @xmath28 , the flat geometry induced on the world - volume and the bi fields become @xmath29
one can see @xmath23 brane of topology @xmath24 is ` blown - up ' to have non - vanishing size of circle direction because dbi lagrangian constructed from the above configuration is the same as that of the supertube .
the basic reason why this gives the same results as @xmath2 case is that t - duality along @xmath25 preserves the relative codimension of @xmath7-brane and @xmath3-branes dissolved in it .
hence the same results will be obtained for further t - dualities along the directions along @xmath30 .
summing up the result for an array of @xmath4-brane @xmath9 along @xmath10-axis , we can say as follows ; when threaded vertically by superstrings over the entire volume of @xmath4-brane , it is ` blown - up ' to @xmath11-brane .
the extra two dimensions acquired is tubular extended along @xmath10-direction and embedded in the residual dimensions . in all cases the stabilized radius is the same and governed by ns field @xmath31 of the superstrings , and the magnetic field @xmath19 which is produced effectively by @xmath4-branes dissolved in the @xmath11-brane .
in this section we deal with another @xmath1 setup , by taking t - duality along @xmath10-direction .
the basic question here is about the @xmath1 counterpart of the supertube .
@xmath3-brane array threaded by @xmath2 superstrings is t - dual to @xmath0-branes along which @xmath1 superstrings are traveling in one direction .
since the former configuration is not stable in @xmath2 setup , neither should be the latter . at first sight
, one might think this latter system will be blown up to @xmath23-brane acquiring extra spherical two dimensions because there is no two dimensional object in @xmath1 setup .
we show below that this is not the case .
actually the stabilized configuration remains one - dimensional .
it turns out to be a @xmath8-helix with axial momentum flow , whose radius is the same as that of the supertube .
the basic tool is again t - duality acting on the supertube . in order to see the resulting configuration , we study the boundary conditions of @xmath1 superstring .
these can be obtained by t - dualizing the boundary conditions of @xmath2 superstrings living on the supertube ; @xmath32 t - duality along @xmath33-direction interchanges @xmath34 with @xmath35 .
with @xmath36 denoting t - dualized coordinates , the above boundary conditions are t - dualized as @xmath37 from the second condition , we note the hypersurface @xmath38 ( with the omitted conditions for other transverse directions ) defines @xmath0 world - sheet .
we take the constant @xmath39 to be zero for simplicity .
the other two conditions define the longitudinal directions of @xmath0-brane .
since both of them are neumann conditions , one can take arbitrary two independent combinations of the coordinates @xmath40 and @xmath41 to make one temporal coordinate and one spatial coordinate .
the simplest choice will be the ` orthonormal ' pair @xmath42 .
this choice is transparent if we see dual background geometry obtained by buscher s duality @xcite ; @xmath43 where one can easily see the relation between the orthonormal coordinates @xmath44 and the tilted coordinates @xmath45 ; @xmath46 in the tilted coordinates , @xmath0 brane is defined by the hypersurface @xmath47 .
the other coordinates @xmath48 are orthonormal longitudinal coordinates .
the metric induced on the hypersurface @xmath49 becomes @xmath50 from which one gets dbi lagrangian ( in static gauge as @xmath51 ) ; @xmath52 the dbi lagrangian is of the same form as that of @xmath2 case .
this is what we expected because t - duality in general leaves the @xmath8-brane action invariant .
the only difference is the change of some field strength components into the derivatives of transverse scalar @xmath53 , which now denote transverse fluctuations of @xmath0-brane with respect to the world - sheet coordinates . in the case at hand
, we can see this explicitly in the relation @xmath54 .
@xmath55 therefore @xmath31 is now the axial velocity of the @xmath8-helix .
fig.1 summarizes the configuration . as in the case of the supertube , the equations of motion just tells us that the momentum @xmath56 is conserved .
the momentum @xmath56 stabilizes the radius of @xmath8-helix as @xmath21 , at which the velocity becomes the speed of light @xmath57 and hamiltonian saturates its bound as @xmath58 .
one could consider the @xmath1 strings traveling along @xmath8-helix .
however , uniform motion of those strings are not physical because of the world - sheet reparametrization symmetry .
1 : the solid square represents @xmath0 world - sheet . in the presence of @xmath59 ,
@xmath0-brane is not static .
it is tilted with an angle of @xmath60 , thus helical .
the helix pitch is @xmath61 .
in this section we show that the above @xmath8-helix configuration preserves one quarter of supersymmetry as in the case of the supertube .
this is , in fact , easy to understand because t - duality in general preserves supersymmetry @xcite .
however , it looks not so transparent to see supersymmetry directly in the @xmath8-helix configuration . here
we give a rigorous proof of that .
we closely follow the procedure of ref .
supersymmetry is determined by the independent killing spinors @xmath62 satisfying @xmath63 where @xmath64 is the matrix defining @xmath65-transformation on the world - volume of @xmath8-branes @xcite . in @xmath1 case at hand , it is @xmath66 where @xmath67 and static gauge is chosen .
therefore the killing spinor relation ( [ killing ] ) becomes @xmath68 in @xmath1 case , chiral projection should be understood for the above spinors , so for @xmath69-component spinors @xmath70 and @xmath71 .
( see ref .
@xcite for details . ) taking into account positive chirality components , we can put @xmath72 therefore they are effectively @xmath73-component spinors . in order to delimitate
the coordinate dependent part of the spinors , we make use of the fact that killing vectors can be written as bilinears of killing spinors ; @xmath74 @xcite . in flat geometry , killing spinors can be written as @xmath75 where @xmath76 s are constant spinors and @xmath77 are lorentz transformation acting on the chiral spinors .
( note that @xmath78 are of the same chirality . ) with this in mind , one sees the killing spinor relation ( [ killing2 ] ) , i.e. , @xmath79 becomes @xmath80 in order to satisfy the above relation for arbitrary value of @xmath81 , the first two terms and the last term should vanish separately ; @xmath82 from the second relation , we see @xmath83 , from which it is clear that @xmath57 .
( this was also obtained when we insert @xmath21 into the expression of @xmath31 . )
therefore @xmath8-helix should travel with the speed of light on its axis to make the configuration supersymmetric . with this value of
@xmath31 inserted , the first relation of eq .
( [ killing3 ] ) tells us that @xmath84 hence @xmath85 can be written in terms of @xmath86 .
since @xmath86 is constrained by the projection operator @xmath87 , only @xmath88 components of killing spinors remain independent ; thereby one quarter supersymmetry of @xmath1 vacuum are preserved for the configuration .
in this paper , we studied the supertube physics in the t - dual setup . in the transversely t - dual case
, the result looks somewhat plain ; blowing - up effect happens for an array of @xmath8-branes when they are threaded by superstrings .
however , we have to mention one interesting point .
in the case of @xmath0 array crossed by superstring array , we can see why the stabilized radius is expressed by the product of @xmath56 and @xmath19 . if we take the s - duality , the role of @xmath0-branes and that of superstrings are interchanged . in order for the radius to be invariant under s - duality , its square ( by dimensional analysis ) should be proportional to @xmath89 . in the longitudinally t - dual case
, we showed that a bunch of @xmath0-branes with @xmath1 superstrings moving in one direction along the branes is ` blown - up ' to a @xmath8-helix traveling with the speed of light along its axis .
it carries @xmath22 of the supersymmetry of @xmath1 vacuum .
one could consider @xmath23-brane of topology @xmath90 as the blown - up configuration of the aforementioned @xmath1 case .
however , that configuration is not consistent with @xmath2 result because it can not be obtained by t - duality from the supertube configuration .
in fact , when one construct dbi action by assuming flat geometry in the spherical polar coordinates and ns b field over the spherical part of the topology , the stabilized radius is vanishing .
spherically blown - up configurations can be possible only with appropriate background fields .
some examples are given in ref .
@xcite and other examples include @xmath8-branes embedded in the flux branes @xcite the helical configuration of @xmath8-branes appeared in a different context . in ref .
@xcite , a double helix connected to each other by fundamental strings is obtained by t - dualizing the @xmath8-brane setup of quantum hall soliton @xcite . with some additional input of @xmath8-branes ( such as an array of @xmath91-branes along @xmath10-axis , which is transverse to their world - volume , and t - duality as before )
the super @xmath8-helix considered in this paper could be a starting point to study such a system further .
finally , we mention two related problems to be solved .
the first one concerns the well - known bound state of @xmath0/@xmath92-branes . as for an array of @xmath23-branes threaded vertically by @xmath1 superstrings
, one can apply the s - duality to change these superstrings into the same number of @xmath0-branes .
subsequent t - dualities along the direction of @xmath0-branes and another direction transverse to both branes result in the familiar @xmath0/@xmath92 bound state .
the same sequence of dualities applied to the blown - up version of the initial configuration result in the kaluza - klein monopole - like configuration @xcite .
this is very puzzling .
if blowing - up effect is preserved under u - duality , the result of supertube predicts that @xmath0/@xmath92 bound state is indeed unstable ( even though it is known as a typical one - quarter supersymmetric state ) and should be understood as a kaluza - klein monopole - like configuration .
the same argument applies to the @xmath3/@xmath93 bound state , which is to be blown - up to @xmath94 brane wrapping a trivial homology cycle .
this nontrivial observation can not be easily understood by direct consideration of @xmath0/@xmath92 without recourse to s- and t - duality .
further study on this should be pursued .
the second one is another t - duality which is not considered in this paper .
it is the t - duality along the angular direction of the supertube .
this duality is quite intriguing as it involves a fixed point . since
the resulting geometry is singular , it should be taken with care .
this will be studied elsewhere .
we thank soo - jong rey , yoonbai kim , and youngjai kiem for helpful discussions and roberto emparan for a valuable comment .
this work is supported in part by kosef through project no .
2000 - 1 - 11200 - 000 - 3 . | we study ` myers effect ' for a bunch of @xmath0-branes with @xmath1 superstrings moving in one direction along the branes .
we show that the ` blown - up ' configuration is the helical @xmath0-brane , which is self - supported from collapse by the axial momentum flow .
the tilting angle of the helix is determined by the number of @xmath0-branes .
the radius of the helix is stabilized to a certain value depending on the number of @xmath0-branes and the momentum carried by @xmath1 superstrings .
this is actually t - dual version of the supertube recently found as the ` blown - up ' configuration of a bunch of @xmath2 superstrings carrying @xmath3-brane charge .
it is found that the helical @xmath0 configuration preserves one quarter of the supersymmetry of @xmath1 vacuum .
# 1#2#3nucl . phys . *
b#1 * ( # 2 ) # 3 # 1#2#3phys .
lett .
* b#1 * ( # 2 ) # 3 # 1#2#3phys . rev . lett.*#1 * ( # 2 ) # 3 # 1#2#3phys . rev
. * d#1 * ( # 2 ) # 3 # 1#2#3ann . phys . *
# 1 * ( # 2 ) # 3 # 1#2#3phys . rep .
* # 1 * ( # 2 ) # 3 # 1#2#3rev .
mod .
phys . *
# 1 * # 1#2#3comm . math .
phys . *
# 1 * ( # 2 ) # 3 # 1#2#3mod .
. lett .
* # 1 * ( # 2 ) # 3 # 1#2#3class .
quantum grav . * # 1 * ( # 2 ) # 3 |
geometrical frustration in electronic degrees of freedom such as spin , charge , and orbit , which is often realized in the stages of highly symmetric crystals , has been one of the major topics in the field of condensed matter physics . in particular , the heavy fermion ( hf ) behavior in y@xmath16sc@xmath17mn@xmath0 [ y(sc)mn@xmath0]@xcite and liv@xmath0o@xmath1 @xcite has attracted broad interest , where such a local electronic correlation specific to the highly symmetric pyrochlore structure may be of direct relevance to the formation of the heavy quasiparticle ( qp ) state .
however , despite decades of studies , the microscopic mechanism by which the local correlation is transformed into the heavy qp mass of itinerant @xmath19-electrons in these compounds still remains controversial . in general , the development of heavy qps accompanies narrowing of the effective band width ( @xmath20 ) or an increase in the density of states ( dos ) at the fermi energy [ @xmath10 ] .
it is thus expected from a naive consideration based on heisenberg s uncertainty principle that the hf behavior should manifest itself in the spin dynamics as reduction of the spin fluctuation rate ( @xmath8 ) because these quantities are mutually linked by the following relation : @xmath22 since the effective qp mass is directly connected to the dos via the relation @xmath23^{2/3}$ ] for a three - dimensional ( 3d ) fermi gas , studies on the spin fluctuation with particular emphasis on relatively low energies should provide valuable information on the mechanism of heavy qp formation . as a probe of spin fluctuation , muon spin rotation ( @xmath24sr ) has a unique frequency window with high sensitivity for @xmath25 s@xmath26 , filling the gap between those covered by nuclear magnetic resonance ( nmr ) and neutron scattering .
this provides promising perspectives for the @xmath24sr study of @xmath19-electron hf - like compounds . here
, we establish that the longitudinal spin relaxation rate ( @xmath4 ) in the above - mentioned compounds exhibits a common feature that it asymptotically becomes independent of temperature upon cooling below a crossover temperature of @xmath27@xmath28 k. this , within the framework of fermionic qps , means that the spin fluctuation rate becomes linearly dependent on temperature , @xmath12 , which is in marked contrast to ordinary metals , where @xmath8 is independent of @xmath3 as determined by the dos at the fermi level .
considering that the transition - metal ions in these compounds comprise a pyrochlore lattice , it is naturally expected that the above - mentioned feature in the spin dynamics will be specific to the relevant lattice structure .
we argue that this is indeed the case in that the underlying spin dynamics can be understood by the spin correlation of the intersecting hubbard chains , which provides a model of the pyrochlore lattice .
this strongly indicates the crucial role of @xmath15 bands that preserve the one - dimensional ( 1d ) character specific to the undistorted pyrochlore lattice at low energies , which is consistent with a theoretical scenario that the 1d-3d crossover due to coupling between these 1d chains developing at low temperatures is the origin of the hf behavior . in the following
, we present an overview of our @xmath24sr results for y(sc)mn@xmath0@xcite , ymn@xmath0zn@xmath29in@xmath17,@xcite and liv@xmath0o@xmath1,@xcite where the temperature dependence of @xmath11 is the primary focus .
then , the temperature dependence of the spin fluctuation rate is discussed in connection with a possible scenario for the origin of heavy qps in these compounds .
we also show that the anomalous behavior of @xmath11 is accompanied by strong broadening of the linewidth under a transverse field ( @xmath30 ) , which is not explained by the hyperfine parameters extrapolated from their values at high temperatures@xcite .
as a probe of spin fluctuation in condensed matter , @xmath24sr is kin to nmr , primarily because of the fact that an implanted muon can be regarded as a light radioisotope of a proton ( @xmath31h ) . nonetheless ,
while the physics behind the mechanism of spin relaxation is common , there are two major factors , i.e. , the time windows of observation ( 10@xmath3210@xmath33 s for nmr , 10@xmath3410@xmath35 s for @xmath24sr ) and the relevant sites ( on - site for nmr , interstitial site for @xmath24sr ) , that make these two techniques distinct ( or even complementary with each other )
. it would be useful to highlight such differences by examining the longitudinal spin relaxation rate ( @xmath36 ) in simple metals . according to korringa s relation , the spin relaxation induced by @xmath37 band electrons
is given by @xmath38 where @xmath39 ( @xmath40 mhz ) and @xmath41 ( @xmath42 mhz ) are the respective gyromagnetic ratios of an electron and muon .
@xmath43 is the knight shift , which is expressed in terms of the hyperfine field ( @xmath44 ) per electron as @xmath45 @xmath46 is avogadro s number , @xmath47 is the bohr magneton , and @xmath48 is the magnetic susceptibility [ @xmath49 for uncorrelated electrons ] . equation ( [ tnt ] ) may be rewritten as @xmath50 ^ 2k_bt,\ ] ] where @xmath51 is the hyperfine field of electrons at @xmath52 orbitals ( @xmath53 ) with @xmath54 denoting the dos for the corresponding band . here
, let us consider the example of pure silver ( ag ) , in which the muon knight shift has been reported to be @xmath55 ppm @xcite . equation ( [ tnt ] ) then yields @xmath56.\ ] ] a similar estimation for silver nuclei ( e.g. , @xmath57ag , where @xmath58% ) leads to @xmath59 [ k@xmath60s ] , which is comparable to the @xmath61 value for muon . considering the melting point of silver ( @xmath62 k )
, it implies that @xmath63 s@xmath26 , which is far below the sensitive range for @xmath24sr , while it is readily observed by nmr .
thus , the @xmath37 band electron makes a negligible contribution to @xmath36 in @xmath24sr . when the electronic correlation is not negligible , it is useful to resort to a more general form of @xmath36 with the relevant spin fluctuation expressed by spin density operators , @xmath64\rangle}{2},\label{lmdsq}\ ] ] where @xmath65 ( @xmath66 being the external field ) and @xmath67 and @xmath68 are the fourier components of the hyperfine field [ @xmath69 and spin density [ @xmath70 , @xmath71 , with @xmath72 being the creation / annihilation operators for spin up(@xmath73)/down(@xmath74 ) electrons ] , respectively , @xmath75=\frac{1}{2}(ab+ba)$ ] , and @xmath76/{\rm tr}[e^{-(h'/k_bt)}]$ ] ( with @xmath77 being the hamiltonian of the electron system ) @xcite . imposing the fluctuation - dissipation theorem , which links the thermal fluctuation described by the correlation function to the imaginary part ( dissipation ) of the generalized susceptibility , @xmath78\rangle}{2},\ ] ] to eq .
( [ lmdsq ] ) with the further assumption that @xmath79 leads to @xmath80 where @xmath81 is the imaginary part of the dynamical spin susceptibility ( perpendicular to the quantization axis that is usually determined by @xmath66 ) .
the actual form of @xmath67 is determined by @xmath82 ( with @xmath83 at the muon position ) , and thus it generally depends on the muon site(s ) .
an exception is the case that the hyperfine field is predominantly determined by the fermi contact interaction , @xmath84 , so that @xmath85 irrespective of @xmath86 .
equation ( [ lmdq ] ) is then reduced to @xmath87 however , such a situation is rare for @xmath24sr because of the small charge of a muon ( same as that of a proton ) and the fact that muons are usually located at the interstitial sites .
consequently , the magnetic dipolar interaction is the primary source of hyperfine fields acting on muons , @xmath88 where @xmath89 with @xmath90 being the position of the @xmath91th electron .
more specifically , the effective hyperfine field in the paramagnetic state is given by the second moment @xmath92 ^ 2,\ ] ] where the corresponding fourier transform @xmath93 can be numerically evaluated for the specific muon site(s ) .
an example calculated for @xmath94 $ ] in y(sc)mn@xmath0 is shown in fig .
[ ymn - cr](b ) , where the muon is presumed to be at the 16@xmath95 site [ see fig . [ ymn - cr](a ) ] .
it shows a broad distribution with a peak near @xmath96 and a tail over the region @xmath97
@xmath981 @xmath26 .
a similar result is observed for the [ 1 - 10 ] direction , indicating that eq .
( [ lmdq ] ) can be approximated by the form @xmath99 following the assumption generally adopted in nmr that the density spectrum of spin fluctuation is described by the lorentzian distribution function @xmath100 with its amplitude represented by the local susceptibility @xmath101 , @xmath102 eq . ( [ rfd0 ] )
is further simplified to @xmath103 where @xmath104 is the redefined hyperfine parameter and @xmath8 is the fluctuation rate of @xmath104 at @xmath96 .
the above equation is valid when @xmath105 .
equation ( [ lrntz ] ) corresponds to the general assumption in the time domain that the correlation of a fluctuating hyperfine field can be described by the stationary gaussian
markovian process in the approximated form @xmath106 while @xmath24sr and nmr can be used to observe the spin fluctuation via the same process described by eq .
( [ rfd ] ) , their sensitive ranges are markedly different ( as already discussed in the case of @xmath37 band electrons ) . as illustrated in fig .
[ ymn - cr](c ) , the difference stems from the sensitive range of @xmath11 ( @xmath107 ) , which is determined by the accessible time window , @xmath108 , for the respective probes ( see also table [ tab ] ) .
the present study benefits greatly from this unique sensitive range of @xmath24sr to spin fluctuation .
we rely on eq .
( [ rfd ] ) to deduce @xmath8 explicitly from the experimentally determined longitudinal spin relaxation rate @xmath11 .
.sensitive ranges of spin fluctuation rate for @xmath24sr and nmr predicted from eq .
( [ rfd ] ) .
see fig .
[ ymn - cr](c ) for more details .
[ cols="^,^,^",options="header " , ]
yittrium manganite ( ymn@xmath0 ) is an intermetallic laves phase ( c15-type ) compound and was the first transition - metal system in which hf behavior was observed . as shown in fig .
[ ymn - cr ] , it comprises a 3d network of corner - shared tetrahedra with mn ions at their corners , resulting in a scheme equivalent to a cubic pyrochlore lattice .
although ymn@xmath0 exhibits magnetic order with complex helical modulation and a large volume expansion below @xmath109 k@xcite , it remains in the paramagnetic state under hydrostatic pressure ( @xmath110 gpa ) or upon the substitution of y by sc ( y@xmath16sc@xmath17mn@xmath0 , with @xmath111 ) , which is also accompanied by a large increase in the qp mass ( @xmath112 times the band mass ) as inferred from the electronic specific heat @xcite .
, where y / sc and mn atoms are indicated by green and red spheres , respectively .
mn atoms form a network of corner - shared tetrahedra known as a pyrochlore lattice .
a small yellow circle shows the 16@xmath95 site presumed to be occupied by implanted muons .
( b ) fourier transform of the hyperfine field distribution @xmath113 in y(sc)mn@xmath0 with a 16@xmath95 site at the origin .
( c ) sensitive range of spin fluctuation rate for @xmath24sr and nmr predicted from eq .
( [ rfd ] ) , where the hyperfine field is presumed to be @xmath114@xmath115 t. note that the sensitive ranges for nmr are split into two different ranges , where the high - frequency range corresponds to the region close to the limit of motional narrowing . dashed curves show @xmath11 for greater external magnetic fields ( @xmath116).,scaledwidth=47.5% ] our @xmath24sr measurements were performed using polycrystalline samples of y@xmath16sc@xmath17mn@xmath0 with various sc contents ( @xmath117 , 0.05 , 0.07 , and 0.08 , as prepared ) grown from melts in an argon arc furnace followed by annealing , where the details of sample preparation are described elsewhere @xcite .
the bulk properties of these samples including the electronic specific heat ( sommerfeld ) coefficient and uniform spin susceptibility ( @xmath101 ) were in good agreement with earlier reports ( see ref .
@xcite for more details ) .
sr spectra [ time - dependent asymmetry , @xmath118 observed for y(sc)mn@xmath0 at 2 k under two different longitudinal fields [ @xmath119 mt ( circles ) and 5 t ( triangles ) ] , where spectra for @xmath120 are shifted vertically by 0.05 with each @xmath121 for clarity .
solid curves show best fits using the function described in the text .
those for @xmath117 exhibit a slight dependence on @xmath66 , from which the hyperfine parameter is estimated , whereas the spectra are less dependent on @xmath66 for @xmath120 . ,
scaledwidth=37.5% ] figure [ tspec ] shows some examples of time - dependent @xmath24sr spectra at a low temperature ( @xmath122 k ) , where pairs of spectra obtained at two different longitudinal fields ( lf , 10 mt and 5 t ) are displayed for comparison@xcite .
the depolarization rate ( @xmath11 ) decreases with increasing sc content @xmath121 and tends to approach an asymptotic value ( as it exhibits little change between @xmath123 and 0.08 ) .
it is also noticeable that @xmath11 is mostly independent of the magnitude of the longitudinal field ( @xmath66 ) for @xmath120 , while it shows a slight variation with @xmath66 for @xmath117 .
these features can be readily understood from eq .
( [ rfd ] ) introduced in the previous section .
we also note that eq .
( [ rfd ] ) has been successfully applied to various types of magnetism including that of quasi-1d compounds @xcite .
equation ( [ rfd ] ) is modified to yield the fluctuation rate @xmath124^{1/2}\ ] ] from the experimental values of @xmath11 and @xmath101 , where the double sign corresponds to the two cases of @xmath125 ( @xmath126 ) and @xmath127 ( @xmath128 ) [ see fig .
[ ymn - cr](c ) ] .
the fact that the spectra in fig .
[ tspec ] are mostly independent of @xmath129 indicates that @xmath130 , from which eq .
( [ rfd ] ) is reduced to yield @xmath131 the magnitude of @xmath104 was determined as the gradient of the muon knight shift ( @xmath132 ) plotted against susceptibility ( i.e. , @xmath133 in the @xmath132-@xmath101 plot ) , for which additional @xmath24sr measurements under a high transverse field ( htf ) were performed on freshly synthesized samples with @xmath134 , 0.07 , and 0.09.@xcite the fast fourier transforms ( ffts ) of the htf-@xmath24sr spectra observed for these samples at various temperatures are shown in fig .
[ ymn - fft ] .
these spectra can be used to monitor the density distribution of the local internal field at the muon site [ @xmath135 via the relation @xmath136 where @xmath137 is the local field from nearby electrons .
the second moment of the local field distribution parallel to @xmath66 is directly related to the corresponding hyperfine parameter @xmath138 ( @xmath139 ) as @xmath140 in the case of isotropic hyperfine fields , where @xmath141 is the component of @xmath137 parallel to @xmath66 and the overline denotes the mean value .
the transverse relaxation rate is then given by @xmath142\cdot \frac{k_bt\chi}{n_a\mu_b^2}.\label{lmdtf}\ ] ] sr spectra observed at various temperatures for y@xmath16sc@xmath17mn@xmath0 with ( a ) @xmath134 , ( b ) 0.07 , and ( c ) 0.09.,scaledwidth=47.5% ] it is noticeable that the spectra become broad and asymmetric with decreasing @xmath3 below @xmath143 k , which seems in accordance with the onset of the curie
weiss behavior of @xmath101 ( see fig .
[ kchi ] inset ) .
the magnitude of the linewidth ( @xmath1445 mhz at 2 k ) is much greater than that expected from eq .
( [ lmdtf ] ) with @xmath138 determined by @xmath133 .
such broadening was also reported in the cases of liv@xmath0o@xmath1@xcite and ymn@xmath0zn@xmath29in@xmath17 ( see below ) .
the nearly identical linewidth for different sc contents @xmath121 shown in fig .
[ ymn - fft ] indicates that the broadening at lower temperatures is not due to extrinsic effects such as crystallographic randomness owing to the sc substitution or the formation of a spin - glass state observed for a smaller sc content @xcite . as is discussed in sect .
[ anis ] , we attribute this broadening to the strong spin fluctuation associated with geometrical frustration .
curve fitting of these spectra using the form @xmath145 yielded reasonable agreement with data for @xmath146 assuming one frequency component ( @xmath147 ) , where @xmath148 is the initial asymmetry , @xmath149 is the transverse spin depolarization rate , and @xmath150 is the initial phase .
the knight shift was then determined by @xmath151 where @xmath152 was determined by additional measurements on a reference sample of high purity silver .
meanwhile , two frequency components ( @xmath153 ) were incorporated to obtain satisfactory fits for the data below @xmath6 .
some examples of @xmath132-@xmath101 plots are shown in fig .
[ kchi]@xcite . considering the line broadening at lower temperatures ,
the hyperfine parameters are deduced from @xmath133 for @xmath146 to yield @xmath154 mhz@xmath155 ( @xmath134 ) , @xmath156 mhz@xmath155 ( @xmath123 ) , and @xmath157 mhz@xmath155 ( @xmath158 ) , where the scattering among different @xmath121 is probably due to the residual influence of the line broadening ( note that the relative variation of @xmath132 and @xmath101 is small for the relevant temperature region ) .
however , their mean value [ @xmath159 mhz/@xmath47 ] is in perfect agreement with the calculated value assuming that muons are located at the 16@xmath95 site@xcite , @xmath161 40.2 mhz/@xmath47 .
although it is inferred from the @xmath31h nmr of ymn@xmath0 that the hydrogen site is the 96@xmath162 site , which is slightly away from the 16@xmath95 site@xcite , the calculated value of the muon hyperfine parameter ( @xmath163 mhz/@xmath47 ) makes it unlikely that a muon occupies this site .
-@xmath101 plots obtained for y(sc)mn@xmath0 with ( a ) @xmath134 , ( b ) 0.07 , and ( c ) 0.09 , where the solid lines are the results of curve fitting for the data above @xmath164100 k. inset : temperature dependence of magnetic susceptibility ( @xmath101 ) used in the @xmath132-@xmath101 plot .
, scaledwidth=50.0% ] the lf-@xmath24sr time spectra in fig .
[ tspec ] were analyzed by least - squares curve fitting to deduce @xmath11 using @xmath165\\ & \simeq&[(1-a_p)\exp\{-(\lambda t)^\beta\}+a_p],\label{gzt}\end{aligned}\ ] ] where @xmath166 is the kubo - toyabe relaxation function , which is approximated by @xmath167 for high @xmath66 , @xmath168 is the power , and @xmath169 is a constant term . here , we introduce stretched exponential decay ( @xmath170 ) to reproduce the deviation of the spin dynamics from that described by the model of spin correlation with a single value of @xmath104 and/or @xmath8 at a given temperature @xcite .
the presence of the term @xmath169 is clearly inferred in the case of @xmath117 from the leveling off of the time spectra for @xmath171 @xmath24s ( see fig . [ tspec ] ; as is also needed for @xmath134 ) .
it is presumed that these deviations from single exponential decay are related to the strong line broadening observed under a transverse field .
apart from this ambiguity , excellent fits were obtained in all of the cases with @xmath169 fixed to the values deduced at 2 k. the solid curves in fig .
[ tspec ] represent the best fits obtained under these conditions@xcite . ) and power ( @xmath168 ) for y(sc)mn@xmath0 obtained by curve fitting of @xmath172 for @xmath117 ( circles ) , 0.05 ( triangles ) , 0.07 ( squares ) , and 0.08 ( inverted triangles ) .
@xmath168 for @xmath173 is fixed at 2 k. , scaledwidth=40.0% ] the temperature dependences of the parameters deduced by curve fitting are summarized in fig .
[ tparams ] .
although @xmath168 varies only slightly with @xmath121 , @xmath11 exhibits a clear tendency of becoming less dependent on temperature , i.e. , @xmath174 with increasing @xmath121 . considering the dependence of @xmath11 on @xmath8 and @xmath101 in eq .
( [ rfd ] ) , this means that @xmath8 becomes linearly dependent on @xmath3 with increasing @xmath121 .
meanwhile , the behavior of @xmath11 for @xmath175 is understood as the freezing of the mn spin fluctuation because the transition to the quasistatic spin - glass state occurs in the sample with @xmath117 below @xmath176 k ( where @xmath177 is defined as the peak muon depolarization rate under @xmath178 mt)@xcite .
the behavior of @xmath11 observed for @xmath179 shows a distinct similarity to that for liv@xmath0o@xmath1 @xcite . as shown in fig .
[ ymn - nu ] , the spin fluctuation rate in the samples with @xmath179 is in the range of @xmath180 @xmath181 s@xmath26 for @xmath182 , while it shows a steeper reduction with decreasing temperature in those with @xmath183 .
although the use of stretched exponential decay in eq .
( [ gzt ] ) prevents @xmath8 from being simply interpreted as a mean when @xmath184 , @xmath8 serves as a characteristic frequency " that describes the spin dynamics on the basis of eq .
( [ rfd ] ) @xcite .
solid lines are obtained by curve fitting using the power law @xmath185 with @xmath95 and @xmath52 being free parameters . as shown in the inset of fig .
[ ymn - nu ] , @xmath52 exceeds 2 in the case of @xmath117 , whereas it approaches unity ( @xmath186 ) for @xmath179 . ) as a function of temperature and sc content ( @xmath121 ) in y(sc)mn@xmath0 .
solid lines are the results of curve fitting using a power law ( @xmath187 ) .
inset : @xmath52 obtained by curve fitting vs @xmath121 . ,
scaledwidth=40.0% ] here , it is worth stressing that nmr and inelastic neutron scattering ( ins ) studies on the paramagnetic phase of y(sc)mn@xmath0 conducted thus far have mostly been concerned with the relatively high energy part of the spin dynamics , where they demonstrated the presence of antiferromagnetic ( af ) correlation with a characteristic frequency scale of @xmath188@xmath189 s@xmath26@xcite .
in particular , a strong hyperfine field exerted on @xmath190mn nuclei [ corresponding to @xmath191 ghz/@xmath47 in eq .
( [ rfd ] ) ] drives the sensitive range of nmr up to such high frequencies @xcite .
interestingly , the latest ins study of a single - crystalline sample ( @xmath117 ) revealed that the intensity centered at approximately @xmath192 ( in reciprocal lattice units ) exhibits anisotropic broadening along the brillouin zone boundary , which is interpreted to be due to the degeneracy of states associated with geometrical frustration @xcite .
although this might be reminiscent of short - range correlations at a low energy , as similarly reported recently for liv@xmath0o@xmath1@xcite , the details are yet to be clarified .
one of the bottlenecks in the investigation of the @xmath19-electron hf state has been the limited number of candidate compounds that exhibit bulk electronic properties attributable to heavy - qp formation .
recently , it has been reported that a ternary intermetallic compound , ymn@xmath0zn@xmath2 , exhibits a large sommerfeld coefficient ( @xmath193 mj@xmath60k@xmath194mol@xmath26 ) characteristic to the hf compounds@xcite .
it crystallizes in the cubic cecr@xmath0al@xmath2 structure with the space group of @xmath195 ( see fig . [ ymnz - cr ] ) , where the y and mn atoms respectively occupy the 8@xmath196 and 16@xmath19 sites , forming diamond and pyrochlore lattices that are common to the cubic laves phase y(sc)mn@xmath0 .
meanwhile , zn atoms at the 16@xmath95 , 48@xmath197 , and 96@xmath162 sites are located between the other two atoms , so that the pyrochlore lattice composed of the mn atoms is almost doubly expanded in comparison with that in y(sc)mn@xmath0 while keeping the tetrahedral symmetry . although the compound with this ideal composition has not yet been obtained , the partial substitution of in for zn is known to be effective for stabilizing the structure . zn@xmath198in@xmath17 , where y , mn , zn , and in atoms are indicated by blue , red , white , and green spheres , respectively .
the zn at the 96@xmath162 site is partially substituted by in to stabilize the cecr@xmath0al@xmath2 structure .
the actual compound exhibits a small amount ( @xmath199 ) of excess mn that occupies the 96@xmath162 site , where it substitutes for zn . the mn - zn tetrahedra form a network similar to the pyrochlore lattice , where the local interaction between mn atoms is relatively weaker than that in y(sc)mn@xmath0 owing to the zn atoms situated between the mn atoms .
( b ) expected muon site at the center of the mn tetrahedra ( 8@xmath200 site ) is shown.,scaledwidth=45.0% ] we have investigated the spin dynamics of mn @xmath19-electrons by @xmath24sr under a zero / longitudinal field ( zf / lf ) in a sample whose chemical composition is more precisely expressed as ymn@xmath201zn@xmath198in@xmath17 with @xmath121=2.36@xcite , where the influence of excess mn appears to be minimal [ @xmath199 = 0.11(1)]@xcite .
the sample was a mosaic of single crystals glued with varnish on a sample holder made of high - purity silver ( 12 mm@xmath202 disc ) .
@xmath24sr measurements in the range of 3004.2 k ( using a @xmath203he cryostat ) were performed under an lf ( @xmath178 mt ) to quench the depolarization due to random local fields from nuclear magnetic moments .
those in the range of 50 k0.3 mk ( with a @xmath204he cryostat ) were performed under an lf ( @xmath205 mt ) to distinguish depolarization due to the pyrochlore ( on - site ) mn from that due to the excess mn , where the yield of these signals was estimated from the field dependence of the lf-@xmath24sr spectra at 0.3 k. typical examples of zf / lf-@xmath24sr spectra obtained at 4.2 and 0.3 k are shown in fig .
[ ymnz - time ] .
the depolarization rate at 4.2 k is mostly independent of @xmath66 ( @xmath11 is unchanged between 10 and 395 mt ) , indicating that the spectra are in the limit of motional narrowing .
these spectra were analyzed by curve fitting using eq .
( [ gzt ] ) , where @xmath206 and @xmath169 was replaced with @xmath207 as the time - independent background mainly originating from muons stopped in the sample holder .
the signal - to - background ratio ( @xmath208 ) was @xmath983.8 for the @xmath203he cryostat and @xmath983 for the @xmath204he cryostat over a time range of 020 @xmath24s , allowing the reliable deduction of @xmath11 at small values ( @xmath209s@xmath26 ) .
sr time spectra of ymn@xmath201zn@xmath198in@xmath17 with @xmath121=2.36 ( @xmath199=0.11 ) measured at ( a ) 4.2 k and ( b ) 300 mk .
solid curves are best fits using eq .
( [ gzt ] ) .
the asymmetry of the spectrum under an lf of @xmath205 mt at 4.2 k is adjusted for comparison with the spectrum under an lf of @xmath178 mt .
[ ymnz - time],scaledwidth=47.5% ] while we have successfully estimated the contribution of on - site mn from the total magnetic susceptibility assuming two components of the curie
weiss term with one ( @xmath210 ) corresponding to the limit of @xmath211@xcite , it is often difficult to identify the contributions of the magnetic ions in question among those from impurities that also exhibit curie weiss - like behavior .
the true contribution of the on - site mn was confirmed by muon knight shift measurements@xcite , as was the case for y(sc)mn@xmath0 .
the htf-@xmath24sr spectra showed two frequency components with relative intensities of approximately 7 to 3 , which was consistent with the values indicated from the partial asymmetries of the corresponding components in the lf-@xmath24sr spectra .
the time spectra were then analyzed by curve fitting using eq .
( [ tfana ] ) assuming two components . the muon knight shift ( @xmath212 ) for each signal
was obtained by eq .
( [ kshift ] ) ( @xmath153 ) , where the ratio of the signal amplitudes was consistent with that observed in the zf / lf-@xmath24sr spectra , in which the contributions of the pyrochlore ( on - site ) and excess mn were readily identified .
this allowed us to unambiguously attribute the signal @xmath213 to the intrinsic pyrochlore mn .
curve fitting by the curie
weiss law yielded @xmath214 emu@xmath60k / mol and @xmath215 k , which showed nearly perfect agreement with @xmath210 in the temperature dependence .
consequently , as shown in fig .
[ ymnz - kchi ] , the @xmath213 versus @xmath210 plot exhibited a linear relationship with @xmath216 mhz/@xmath47 with a small offset of @xmath217 ppm , indicating that @xmath210 indeed originated from the on - site mn .
meanwhile , assuming that muons are located at the centers of mn tetrahedra [ 8@xmath200 site , shown in fig .
[ ymnz - cr](b ) ] , @xmath104 is calculated using the dipolar tensor for the mn spins situated at the nearest- and next - nearest - neighboring tetrahedra to yield @xmath218 mhz/@xmath47 , which is in reasonable agreement with the experimental value . ) versus magnetic susceptibility ( @xmath210 ) for intrinsic pyrochlore mn in ymn@xmath201zn@xmath198in@xmath17.,scaledwidth=40.0% ] ) and ( b ) spin fluctuation rate ( @xmath8 ) in ymn@xmath201zn@xmath198in@xmath17 with @xmath121=2.36 ( @xmath199=0.11 ) .
the solid curves in ( b ) are results of least - squares fitting using a power law ( @xmath187 ) for the range of temperatures above and below @xmath6.,scaledwidth=37.5% ] the potential influence of excess mn can be avoided by employing values of @xmath11 obtained from data under an lf with magnitude greater than 100 mt . as shown in fig .
[ ymnz - nu](a ) , @xmath11 tends to level off ( @xmath219 ) for @xmath182 , which strongly resembles the case of y(sc)mn@xmath0 with @xmath179 . combining this result with @xmath210 and @xmath138
, we obtain @xmath8 vs @xmath3 using eq .
( [ rfd ] ) as shown in fig .
[ ymnz - nu](b ) .
the linear temperature dependence of @xmath8 emerges below @xmath16410 k , confirming our previous result@xcite with improved reliability .
curve fitting using the power law [ eq .
( [ tpwr ] ) ] yields @xmath220@xcite , which is again consistent with the case of y(sc)mn@xmath0 .
it is noticeable that @xmath8 exhibits a clear kink around @xmath221 k , below which @xmath8 becomes linearly dependent on @xmath3 .
thus , @xmath6 may be regarded as a crossover temperature below which the spin fluctuation is predominantly determined by the mechanism common to y(sc)mn@xmath0 .
we also note that @xmath222 ( the depolarization rate of the signal corresponding to the on - site mn under a transverse field ) was in the range of 12 mhz below @xmath9830 k whih was much greater than that expected from the magnitude of @xmath11 [ @xmath223 mhz , see fig .
[ ymnz - nu](a ) ] , indicating the presence of additional line broadening due to the spin fluctuation associated with geometrical frustration in @xmath224 , as is also observed in y(sc)mn@xmath0 and liv@xmath0o@xmath1 .
lithium vanadate is the only compound that exhibits hf behavior among the numerous metal oxides so far studied , and therefore it has attracted much attention since its discovery in the late 90s.@xcite the formation of a heavy - qp state below a characteristic temperature ( @xmath22530 k ) is suggested by its large sommerfeld coefficient ( @xmath226 mj / mol@xmath60k@xmath227 ) and other bulk properties that are hallmarks of typical @xmath197-electron hf compounds .
moreover , it has been inferred from the result of recent photoemission spectroscopy examination that a dos peak slightly above @xmath228 develops for @xmath229.@xcite in our previous @xmath24sr study , we showed using a powder specimen of liv@xmath0o@xmath1 that the observed @xmath24sr spectra consisted of two components that could be distinguished by the response of the depolarization rate to an external magnetic field ( @xmath66).@xcite furthermore , the component that exhibited the weaker dependence on @xmath66 ( with longitudinal spin relaxation rate @xmath230 and fractional yield @xmath231 ) was mostly independent of temperature below @xmath232 k , from which we suggested that the corresponding fluctuation rate derived from the redfield theory for the _ local _ spin systems was also independent of temperature ( @xmath233 s@xmath26 ) .
in contrast , the depolarization rate associated with the other signal ( @xmath234 , with @xmath235 ) was readily suppressed by @xmath66 , which was ascribed to slowly fluctuating local magnetic moments ( @xmath236@xmath237 s@xmath26 ) .
although the occurrence of such phase separation was confirmed by a subsequent @xmath24sr study of high - quality single - crystalline samples , the increased yield @xmath197 ( @xmath238 ) implied that clarifying the origin of @xmath239 was essential to our understanding of the electronic state.@xcite figure [ lmd ] shows the muon depolarization rate ( @xmath230 ) under a zero external field , deduced by curve fitting using the sum of two components with exponential damping:@xcite @xmath240 although the data are scarce , particularly at higher temperatures , @xmath230 is only weakly dependent on temperature with a tendency to level off with decreasing temperature , which is qualitatively similar to the behavior of @xmath241 ( fig .
[ lmd ] , inset ) .
these features are remarkably similar to those observed for y(sc)mn@xmath0 ( @xmath179 ) and ymn@xmath0zn@xmath29in@xmath17 .
o@xmath1 reproduced using data in ref .
@xcite for the component showing fast fluctuation ( @xmath239 ) .
inset : temperature ( @xmath3 ) dependence of @xmath101 with @xmath3 having a logarithmic scale .
solid curves are results of curve fitting assuming two components , where the curve approaching 0.005 emu / mol is attributed to the intrinsic contribution of v ( @xmath241 , see text for more details ) . , scaledwidth=37.5% ] a more appropriate evaluation of @xmath239 from @xmath230 using eq .
( [ rfd ] ) for an itinerant system was recently carried out.@xcite since the muon knight shift data obtained from the previous @xmath24sr measurement were not sufficient to evaluate the temperature dependence of @xmath212 for the relevant component ( particularly at low temperatures),@xcite we extracted it from the bulk susceptibility ( @xmath101 ) . more specifically , considering the behavior of @xmath101 observed in single - crystalline samples , which tends to saturate at lower temperatures,@xcite we attributed the divergent behavior of @xmath101 at @xmath242 to unknown paramagnetic impurities ( obeying the curie law ) and decomposed the data into two parts : @xmath243 where @xmath244 is the weiss temperature .
curve fitting using eq .
( [ chit ] ) yielded @xmath245 emu@xmath60k / mol , @xmath246 k , and @xmath247 emu@xmath60k / mol , implying that the behavior of @xmath241 was in good agreement with that of single crystals [ thus , we use @xmath248 in eq .
( [ rfd ] ) ] . assuming that the curie term originates from free v spins ( [email protected]@xmath47 ) , the fractional yield of the impurity phase estimated as @xmath250 was 1.6% of the total volume .
this was much smaller than that of primary @xmath24sr signals ( either @xmath197 or @xmath251 with @xmath231 ) , indicating that the paramagnetism of the impurity phase was irrelevant to the interpretation of @xmath24sr data .
another important quantity in eq .
( [ rfd ] ) is @xmath104 .
the muon knight shift measurements of both powder and single - crystalline samples yielded @xmath252 ghz/@xmath47 , which corresponds to the component from which @xmath239 was deduced.@xcite apart from the large error due to the broad linewidth , this @xmath253 was in good agreement with the calculated @xmath253 of 0.143 ghz/@xmath47 for muons that occupy a site at the center of a cyclic vanadium hexamer ( as inferred from the observed @xmath24sr linewidth due to _ nuclear _ magnetic moments at high temperatures ) and are subject to the magnetic dipolar fields from vanadium ions.@xcite since the experimental value of the hyperfine parameter had a large uncertainty due to the broad linewidth , we used the calculated value for the evaluation of @xmath239@xcite .
the reevaluated @xmath239 is plotted in fig .
[ lvo - nu ] together with ins data,@xcite where one can observe that @xmath239 lies on a straight line , indicating its proportional relationship with temperature ( @xmath254 ) over a @xmath3 range of three decades below @xmath6 .
this is again strikingly similar to the behavior of the spin fluctuation rate observed in y(sc)mn@xmath0 and ymn@xmath0zn@xmath29in@xmath17 at lower temperatures ( @xmath255 ) .
we also note that the corresponding low - energy excitation ( also suggested in an earlier neutron scattering study@xcite ) has been confirmed by a recent ins experiment on high - quality samples@xcite . ) in liv@xmath0o@xmath1 as a function of temperature .
the thin solid line shows a linear @xmath3 dependence ( @xmath12 ) .
inelastic neutron scattering data are also shown for comparison , where the dashed curve shows the linewidth @xmath256 at @xmath257 @xmath26 with @xmath258 mev.@xcite , scaledwidth=37.5% ]
as mentioned earlier , the heavy qp mass is phenomenologically understood to originate from the sharp increase in the dos slightly above the fermi level and the associated flattening of the band dispersion , i.e. , @xmath259 ( with @xmath14 determined by the dispersion relation of electrons ) . in rare - earth compounds ,
such enhancement is induced by the conversion of local @xmath197-electron degrees of freedom into @xmath260 by the kondo effect , which is observed as a peak structure of @xmath260 near @xmath228 .
the observation of such a structure ( @xmath984 mev above @xmath228 ) in liv@xmath0o@xmath1 by photoemission spectroscopy@xcite appears to favor the kondo mechanism established for @xmath197-electron compounds as a common microscopic origin of the hf behavior in the relevant compounds .
however , further attempts to obtain support for this scenario have been elusive . a theoretical model to project @xmath19-electron states ( 1.5 electrons per v@xmath261 ion ) onto the kondo model by splitting them into two sub - bands by electronic correlation
had to introduce an unusually large kondo coupling ( @xmath262 k ) to overcome the competing effect of hund coupling.@xcite our @xmath24sr study of a single - crystalline sample provided evidence against the formation of a spin - singlet state as it showed the presence of the local vanadium moments at low temperatures far below @xmath263 ( interpreted as the kondo temperature ) , where the local " spins would disappear in this scenario.@xcite from this view point , the importance of a highly symmetric crystal structure and the potential influence of geometrical frustration have been stressed by various authors , leading to a wide variety of theoretical models.@xcite the coexistence of slowly fluctuating local moments with the heavy - qp state at low temperatures in y(sc)mn@xmath0 and ymn@xmath0zn@xmath29in@xmath17 supports the above - mentioned argument for liv@xmath0o@xmath1 .
the presence of local moments over the relatively long time range of @xmath264@xmath265 s indicates that the conventional kondo coupling ( which virtually eliminates local spins over a time scale longer than @xmath266@xmath267 s , where @xmath268 is the exchange energy between conduction electrons and @xmath197 electrons ) is not in effect , thereby suggesting a different origin of the heavy - qp state in these compounds . according to a theoretical investigation of intersecting hubbard chains as a model of the pyrochlore sublattice in liv@xmath0o@xmath1 ,
the low - energy part of the spin dynamics is predicted to be described by a spin - spin correlation whose relaxation rate is proportional to temperature@xcite .
the behavior @xmath269 which is commonly observed in liv@xmath0o@xmath1 , y(sc)mn@xmath0 ( @xmath179 , where the system is far from the spin - glass instability ) , and ymn@xmath0zn@xmath29in@xmath17 at lower temperatures , is perfectly consistent with the above prediction , suggesting that the @xmath15 orbitals associated with mn / v atoms retain their 1d character at low temperatures ( energies ) . in particular , the spin fluctuation rate ( @xmath239 ) in liv@xmath0o@xmath1 deduced from @xmath24sr is perfectly in line with the relaxation rate ( @xmath270 for @xmath271 k ) observed over the low - energy region of the ins spectrum @xcite .
this implies that both @xmath24sr ( sensitive over @xmath272 @xmath981 @xmath26 ) and ins probe common parts of the spin fluctuation spectrum of liv@xmath0o@xmath1 , suggesting that a similar situation is realized for y(sc)mn@xmath0 and ymn@xmath0zn@xmath29in@xmath17 .
phenomenologically , this may be interpreted to mean that the heavy qps develop in the manner @xmath13 at lower temperatures . among the many theoretical models for the microscopic origin of the hf state in liv@xmath0o@xmath1 , that proposed by fujimoto regards the quasi-1d character of the @xmath15 bands associated with the pyrochlore lattice ( consisting of intersecting chains of @xmath15 orbitals ) as an essential basis for the description of the electronic state as it is expected that the hybridization between the 1d bands will be strongly suppressed owing to the geometrical configuration ( frustration)@xcite . this model incorporates the hybridization as a perturbation to the 1d hubbard bands , which yields an energy scale ( @xmath6 ) that characterizes the crossover from 1d to 3d as the fermi liquid state develops with decreasing temperature below @xmath6 . the calculated specific heat coefficient taking account of the latter as the leading correction to the self - energy yields a large value that is consistent with the experimentally observed values .
the progression of hybridization also induces the enhancement of the 3d - like spin correlation , which should appear as the enhancement of uniform susceptibility for @xmath255 , while the spin fluctuation is dominated by the staggered component of 1d hubbard chains .
here we discuss the origin of the strong broadening of the tf-@xmath24sr linewidth observed in y(sc)mn@xmath0 at lower tempertures ( @xmath255 ) , which is also expected to be relevant to the other two compounds .
since the muon site has a @xmath273 point symmetry for the cubic @xmath274 structure , the presumed muon site suggests a powder pattern for @xmath275 that is characterized by two singularities ( edges ) at @xmath276 and @xmath277 due to the uniaxially anisotropic dipolar fields@xcite , @xmath278 ( see the inset of fig .
[ ymn - pp ] ) .
accordingly , we attempt to reproduce tf-@xmath24sr time spectra by curve fittting using the following form : @xmath279 \int_{\omega_\parallel}^{\omega_\perp } p(\omega_\mu ) \cos(\omega_\mu t + \phi_0 ) d \omega_\mu , \label{eq : uniax}\ ] ] where stretched exponential damping was incorporated to consider the additional line broadening due to spin fluctuation .
sr spectra for y@xmath16sc@xmath17mn@xmath0 with @xmath121 = 0.09 obtained at 2 k under a transverse field of 6 t. dotted and broken curves are those calculated using @xmath275 shown in the inset respectively assuming isotropic and anisotropic local spin susceptibility ( see text for more details ) , and solid curves show the result of curve fitting with the uniaxial @xmath275 plus additional line broadening .
the data points and curves are displayed on a rotating reference frame with a frequency of 811 mhz .
, scaledwidth=37.5% ] figure [ ymn - pp ] shows result of such curve fitting for a time spectrum obtained at 2 k in the time domain with the inset showing the calculated @xmath275 for ( i ) isotropic local spin susceptibility ( black dotted curve ) and ( ii ) assuming in - plane anisotropy ( blue dashed curve ) , where the mn moments are allowed to fluctuate within the easy plane perpendicular to the threefold rotation - inversion axis on the mn site ( which is parallel to the [ 111 ] direction ) .
the latter means that the local spin susceptibility probed by a muon strongly depends on the field direction .
then , the local spin susceptibility becomes considerably greater than the isotropic case when the external field is parallel to the principal axis , and thereby @xmath101 in eq .
( [ lmdtf ] ) should be replaced with the anisotropic @xmath224 ( thus enhancing @xmath149 ) .
the main panel of fig .
[ ymn - pp ] shows the corresponding @xmath280 ( black dotted and blue broken curves ) calculated for the respective @xmath275 using the parameters @xmath148 , @xmath11 , @xmath168 , and @xmath150 obtained from the curve fitting using eq .
( [ eq : uniax ] ) .
one can observe that the spectrum is not reproduced by the two types of calculated curve and that further broadening is required .
this clearly indicates the occurrence of excess line broadening , which is not explained by the anisotropy of the knight shift .
we note that a similar result was obtained for the tf-@xmath24sr time spectra in liv@xmath0o@xmath1 .
such broadening may be a manifestation of strong spin fluctuation due to the geometrical frustration . in any case , the increase in @xmath149 for @xmath255 is commonly observed in y(sc)mn@xmath0 , ymn@xmath0zn@xmath29in@xmath17 , and liv@xmath0o@xmath1 , and it is expected to be an important clue to understanding the hf - like behavior of these pyrochlore antiferromagnets .
we have shown that the longitudinal spin relaxation rate in the @xmath19-electron hf - like compounds commonly exhibits a tendency to level off , @xmath281 , below a characteristic temperature @xmath6 .
this implies that the spin fluctuation rate becomes linearly dependent on temperature , @xmath12 , in these compounds for @xmath255 .
in particular , such behavior in liv@xmath0o@xmath1 is consistent with the implications of ins experiments , providing a basis for the coherent understanding of the low - energy spin dynamics through the theoretical model of intersecting 1d hubbard chains that simulates a pyrochlore lattice .
the persistent quasi-1d spin dynamics coexists with the enhanced local susceptibility at lower temperatures ( @xmath255 ) , which is also common to two other @xmath19-electron hf systems , y(sc)mn@xmath0 and ymn@xmath0zn@xmath29in@xmath17 .
these observations strongly indicate that a geometrically constrained @xmath15 band is the primary requirement for the formation of heavy qps , with the 1d - to-3d crossover as a possible mechanism of effective mass enhancement .
the results quoted in this review were achieved under collaborations with t. yamazaki , y. tabata , and h. nakamura for y(sc)mn@xmath0 , with y. okamoto and z. hiroi for ymn@xmath0zn@xmath29in@xmath17 , and with h. ueda , c. urano , s. kondo , m. nohara , h. takagi , y. matsushita , and y. ueda for liv@xmath0o@xmath1 .
we would like to thank m. hiraishi , t. masuda , a. koda , k. m. kojima , k. ohishi , i. kawasaki , i. watanabe , and w. higemoto for their active involvement in @xmath24sr experiments .
we also thank k. tomiyasu , s. fujimoto , h. tsunetugu , and y. kuramoto for helpful discussions .
the work was partly supported by the inter - university research program of institute of materials structure science , kek ( proposal nos . 2011a0011 , 2012a0051 , 2013a0089 , 2014a0238 ) .
s. kondo , d. c. johnston , c. a. swenson , f. borsa , a. v. mahajan , l. l. miller , t. gu , a. i. goldman , m. b. maple , d. a. gajewski , e. j. freeman , n. r. dilley , r. p. dickey , j. merrin , k. kojima , g. m. luke , y. j. uemura , o. chmaissem , and j. d. jorgensen , phys .
lett . * 78 * , 3729 ( 1997 ) .
o. hartmann , r. wppling , k. aggarval , l. asch , a. kratzer , g. m. kalvius , f. j. litterst , a. yaouanc , p. dalmas de rotier , b. barbara , f. n. gygax , b. hitti , e. lippelt , and a. schenck , hyperfine interact .
* 64 * , 711 ( 1990 ) . | spin fluctuation is presumed to be one of the key properties in understanding the microscopic origin of heavy - fermion - like behavior in the class of transition - metal compounds , including liv@xmath0o@xmath1 , y(sc)mn@xmath0 , and ymn@xmath0zn@xmath2 . in this review ,
we demonstrate by our recent study of muon spin rotation / relaxation that the temperature ( @xmath3 ) dependence of the longitudinal spin relaxation rate ( @xmath4 ) in these compounds exhibits a common trend of leveling off to a constant value ( @xmath5 . ) below a characteristic temperature , @xmath6 .
this is in marked contrast to the behavior predicted for normal metals from the korringa relation , @xmath7 , where the spin fluctuation rate ( @xmath8 ) in the pauli paramagnetic state is given as a constant , @xmath9 $ ] [ with @xmath10 being the density of states at the fermi energy ] .
thus , the observed behavior of @xmath11 implies that the spin fluctuation rate becomes linearly dependent on temperature , @xmath12 , suggesting that heavy quasiparticles develop in a manner satisfying @xmath13 at lower temperatures ( @xmath14 determined by the electronic dispersion ) . considering that the theory of spin correlation for intersecting hubbard chains as a model of pyrochlore lattice predicts @xmath12 , our finding strongly indicates the crucial role of @xmath15 bands which preserve the one - dimensional character at low energies due to the geometrical frustration specific to the undistorted pyrochlore lattice . |
wireless systems require fundamental and crisp understanding of design principles and control mechanisms to efficiently manage network resources .
resource allocation policies lie at the heart of wireless communication networks , since they aim at guaranteeing the required quality of service ( qos ) at the user level , while ensuring efficient and optimized operation at the network level to maximize operators revenue .
resource allocation management in wireless communications may include a wide spectrum of network functionalities , such as scheduling , transmission rate control , power control , bandwidth reservation , call admission control , transmitter assignment , and handover @xcite . in this survey ,
a resource allocation policy is defined by the following components : _ i _ ) a multiple access technique and a scheduling component that distributes resources among users subject to individual qos requirements ; _ ii _ ) a signaling strategy that allows simultaneous transmission of independent data streams to the scheduled users ; and _ iii _ ) rate allocation and power control that guarantee qos and harness potential interference .
[ fig : mu - mimo - layered - tasks ] illustrates these components and the interconnection between them .
the figure highlights the fact that each function of the resource allocation strategy can be performed in either optimal or suboptimal way , which is elaborated upon below .
the multiple access schemes can be classified as orthogonal or non - orthogonal .
the former is a conventional scheme that assigns radio resources , e.g. , code , sub - carrier , or time slot , to one user per transmission interval .
the main characteristic of orthogonal multiple access schemes is their reliability , since there is no need to deal with co - resource interference
. the resource allocation policy can optimize with reasonable complexity several performance metrics , such as throughput , fairness , and qos @xcite .
the multiplexing gain , i.e. , the number of scheduled users , is limited by the number of available radio resources in the system . in non - orthogonal multiple access ,
a set of users concurrently superimpose their transmissions over the same radio resource , and potentially interfere with each other . in this scheme
the co - resource interference can be mitigated by signal processing and transmission techniques implemented at the transmitter and/or receiver sides .
such techniques exploit different resource domains , e.g. , power , code , or spatial domain , and a combination of them are envisaged to cope with the high data rate demands and system efficiency expected in the next generation of wireless networks @xcite .
hereinafter , we focus on multiple access schemes based on multi - antenna transceivers operating at the spatial domain , i.e. , multiple - input multiple - output ( mimo ) .
mimo communication , where a multi - antenna base station ( bs ) or access point ( ap ) transmits one or many data streams to one or multiple user equipments simultaneously , is a key technology to provide high throughput in broadband wireless communication systems .
mimo systems have evolved from a fundamental research concept to real - world deployment , and they have been integrated in state - of - the - art wireless network standards @xcite , e.g. , ieee 802.11n , 802.11ac wlan , 802.16e ( mobile wimax ) , 802.16 m ( wimax ) , 802.20 ( mbwa ) , 802.22 ( wran ) , 3gpp long - term evolution ( lte ) and lte - advanced ( e - utra ) .
resource allocation is particularly challenging in wireless communication systems mainly due to the wireless medium variability and channel randomness , which renders the overall performance location - dependent and time - varying @xcite .
nevertheless , high spectral efficiency and multiplexing gains can be attained in mimo systems since multiple data streams can be conveyed to independent users . by exploiting the spatial degrees of freedom ( dof ) offered by multiple antennas we can avoid system resource wastage @xcite .
multiuser ( mu ) mimo systems have been extensively investigated over the last years from both theoretical and practical perspective . in a recent evolution of mu - mimo technology , known as massive mimo or large - scale mimo @xcite ,
few hundreds of antennas are employed at the bs to send simultaneously different data streams to tens of users .
massive mimo has been identified as one of the promising air interface technologies to address the massive capacity requirement required demanded by 5 g networks @xcite .
+ the downlink transmission is particularly challenging in mu - mimo scenarios because the geographic location of the receivers is random and joint detection can not be performed .
the main goal is to convey independent data streams to a set of properly selected users , attaining spatial multiplexing gain offered by mu - mimo .
however , determining such a users set is a very challenging task , which depends on all elements of the resource allocation strategy , e.g. , individual qos requirements , signaling schemes , rate allocation and power control strategies implemented at the transmitter .
mimo systems allow for a plethora of mighty signal processing techniques that enhance the system performance by exploiting a multi - dimensional pool of resources .
this pool is composed of resources with different nature , e.g. , signal spaces , transmission powers , time slots , sub - carriers , codes , and users .
efficient allocation designs over such a large set of resources implies that a tradeoff between optimality and feasibility . on the one hand
, optimality can be reached by solving optimization problems over a set of integer and continuous variables , which may be a thoroughly complex task .
feasibility , on the other hand , implies that suboptimal resource allocation takes place by relaxing and reformulating optimization problems whose solutions can be found by practical and reliable algorithms .
the paper is organized as follows . in section [ section : preliminaries ]
we present the basic ideas behind mimo wireless communications , introduce mu - mimo systems , and discuss the main challenges . in section
[ section : system - set - ups ] we introduce the most commonly studied mu - mimo channel and system models , their characteristics and conventional assumptions .
section [ section : precoding ] is devoted to signal design and precoding schemes under different conditions of channel information . in section [ section : metrics - spatial - compatibility ] we introduce the most common metrics of spatial compatibility , which are used to categorize users and reduce the scheduling complexity .
section [ section : performance - uf ] presents a classification of optimization criteria and describes the usual constraints considered in mu - mimo systems . in section [ section : algorithms ]
we propose a classification of the several techniques to address the user scheduling problem .
the specific characteristics , limitations and use cases for each technique are discussed .
section [ section : algorithms : selection - partial - csit ] is focused on scheduling algorithms with partial channel information at the transmitter .
we categorize the existing approaches and present guidelines to minimize complexity and improve efficiency . in section [ section : power - allocation ]
we present the most common power allocation schemes and discuss their role in mu - mimo systems .
finally , we conclude the paper in section [ section : remarks ] . the reader can find a list of technical terms and abbreviations summarized in table [ table : abbreviations ] .
we adopt the following notation : matrices and vectors are set in upper and lower boldface , respectively .
@xmath0 , @xmath1 , @xmath2 , @xmath3 denote the transpose , the hermitian transpose , the absolute value , and the @xmath4-norm , respectively .
@xmath5 , @xmath6 denote the rank and null space of matrix @xmath7 . @xmath8 and @xmath9 denote the subspace and orthogonal subspace spanned by the columns of matrix @xmath7 .
calligraphic letters , e.g. @xmath10 , denote sets , and @xmath11 denotes cardinality .
@xmath12 is the set of nonnegative real numbers and @xmath13 is the space of @xmath14 matrices .
@xmath15 is the complex gaussian distribution with mean @xmath16 and covariance matrix @xmath7 .
@xmath17 $ ] denotes expectation .
a mimo system employs multiple antennas at the transmitter ( @xmath18 ) and receiver ( @xmath19 ) sides to improve communication performance .
the seminal works @xcite provide a mathematical motivation behind multiple antenna processing and communications .
theoretical analysis has shown that the spectral efficiency , i.e. , the amount of error - free bits per second per hertz ( bps / hz ) , follows the scaling low @xmath20 , without increasing the power or bandwidth requirements .
the signal processing techniques in multi - antenna systems can be classified as _ spatial diversity techniques _ and _ spatial multiplexing techniques _
@xcite . spatial diversity techniques ( see @xcite and references therein ) , provide transmission reliability and minimize error rates .
this is attained by transforming a fading wireless channel into an additive white gaussian noise ( awgn)-like channel , i.e. , one can mitigate signal degradation due to fading @xcite .
the probability that multiple statistically independent channels experience simultaneously deep fading gets very low as the number of independent paths increases .
the spatial diversity techniques can be applied at both transmission and reception sides of the link .
transmit diversity schemes include space diversity , polarization diversity , time diversity , frequency diversity , and angle diversity .
examples of receive diversity schemes are selection combining , maximum ratio combining ( mrc ) , and equal gain combining @xcite .
this paper focuses on spatial multiplexing techniques , which exploit the dof provided by mimo .
spatial multiplexing is tightly related to multiuser communications and smart antennas processing @xcite . in multiuser systems ,
spatial multiplexing gains can be attained by steering signals toward specific receivers , such that the power to intended users is boosted .
simultaneously , co - channel interference to unintended users can be partially or completely suppressed . in mu - mimo systems , the available resources ( power ,
bandwidth , antennas , codes , or time slots ) must be assigned among @xmath21 active users .
there are two kinds of multiuser channels : the downlink channel , also known as _ broadcast channel _ ( bc ) , where a single transmitter sends different messages to many receivers ; and the uplink channel , also called _ multiple access channel _ ( mac ) , where many transmitters communicate with a single receiver .
there are several _ explicit _ differences between bc and mac . in the former ,
the transmitted signal is a combination of the signals intended for all co - scheduled users , subject to total transmit power , @xmath22 , constraints .
in contrast , in the mac channel , the signal from the @xmath23-th user is affected by other co - scheduled users , subject to individual power constraints , i.e. @xmath24 , @xcite .
there exists an _
implicit _ connection between bc and mac , known as _ duality _ , which establishes the relationship between the capacity regions of both access channels @xcite .
the bc - mac duality has been fundamental to define optimal policies for power allocation , signaling , and qos guaranteeing in mu - mimo systems , see @xcite .
the capacity regions include operative point where transmission to multiple users do not interfere with each other . every transmission is performed over orthogonal signaling dimensions , which is a signal separation called duplexing @xcite .
this operation is performed by allocating communications across different time slots , known as time - division - duplex ( tdd ) , or across separated frequency bands , known as frequency - division - duplex ( fdd ) . in the literature of mu - mimo
, two types of diversity are studied : _ spatial multiplexing diversity _ and _ multiuser diversity _ ( mudiv ) .
the former is a consequence of the independent fading across mimo links of different users .
this means that independent data streams can be transmitted over parallel spatial channels , increasing the system capacity @xcite .
the latter arises when users that are geographically far apart have channels that fade independently at any point in time .
such independent fading processes can be exploited so that users with specific channel conditions are simultaneously scheduled @xcite .
there are two modes of transmission in mimo systems , see fig .
[ fig : su - mu - mimo ] : single user ( su ) and multiuser ( mu ) mode .
the su - mimo mode improves the performance of a single user , allocating one or many data streams in the same radio resource . in the mu - mimo mode
, different data streams are sent to different users such that a performance metric is optimized , e.g. , the average sum rate . + selecting between su- or mu - mimo transmission modes
depends on the accuracy of the channel state information at the transmitter ( csit ) , the amount of allowed interference , the target rate per user , the number of user , the signal - to - noise ratio ( snr ) regime , and the achievable capacity in each mode @xcite .
nonetheless , by assuming sufficient csit knowledge , mu - mimo processing techniques provide several performance gains @xcite : multiple antennas attain diversity gain , which improves bit error rates ( ber ) ; directivity gains realized by mudiv , since the spatial signatures of the users are uncorrelated , which mitigates inter - user interference ( iui ) ; immunity to propagation limitations in su - mimo , such as rank loss or antenna correlation ; and multiplexing gains that scale , at most , with the minimum number of deployed antennas .
+ in mu - mimo bc systems , the overall performance depends on how efficiently the resource allocation algorithms manage the hyper - dimensional pool of resources ( carriers , time slots , codes , power , antennas , users , etc . ) .
consider a system with a transmitter equipped with @xmath18 antennas , and let @xmath25 be the set of all active users , illustrated in fig .
[ fig : system - processing - blocks ] . to qualitatively determine the objectives of scheduling , we provide the following definitions : [ defn : quality - of - service ] _ quality of service_. we say that qos defines a set of prescribed network-/user - based performance targets ( e.g. , peak rates , error rates , average delays , or queue stability ) , that can be measured , improved , and guaranteed for a specific upper layer application .
[ defn : user_scheduling ] _ user scheduling_. we say that a set of radio resources ( e.g. time slot , codes , sub - channels , powers , etc . ) , has been assigned to a group of scheduled users , @xmath26 , so that a global performance metric is optimized subject to power and qos constraints . moreover
, each user @xmath27 , achieves non - zero rate with successful information reception .
consider that each user @xmath28 , is equipped with @xmath29 antennas . by having more receive than transmit antennas ( @xmath30 )
, one can solve a selection problem to achieve mudiv gains in fading fluctuating channels @xcite . a fundamental task in resource management
is to select a subset of users @xmath26 , and assign resource to it , so that a given performance metric is optimized . for the sake of illustration , consider a single - transmitter scenario , and let us formulate a general user scheduling problem for a single resource ( sub - carrier , time slot , or code ) as follows : rrcl & & & _ k = 1 ^k _ ( k ) u_(k ) [ eq : general - combinatorial - problem : cost - function ] + & & & _ k = 1^k _ k p_k p [ eq : general - combinatorial - problem : total - power ] + & & & 0 _ k p_k p_k k | [ eq : general - combinatorial - problem : individual - power ] + & & & _ k = 1^k _ k c_p [ eq : general - combinatorial - problem : set - cardinality ] + & & & _ k \{0 , 1 } k | [ eq : general - combinatorial - problem : binary - vars ] our goal is to maximize the sum of the utility functions , @xmath31 , @xmath32 , which depends on several parameters : the multiuser mimo channel , @xmath33 , the allocated power , @xmath34 , the individual data queues , @xmath35 , and the encoding order , @xmath36 .
the qos requirements can be included in the definition of @xmath31 , as individual weights [ cf .
section [ section : performance - uf ] ] .
equations ( [ eq : general - combinatorial - problem : total - power ] ) and ( [ eq : general - combinatorial - problem : individual - power ] ) define total and individual power constraints , which are set according to the scenario [ cf . section [ section : system - set - ups ] ] .
the term @xmath37 is a binary variable with value equal to 1 if the @xmath38-th user is scheduled and 0 otherwise .
the set of selected user is given by @xmath39 .
the system operates in mu - mimo mode if @xmath40 , where @xmath41 in ( [ eq : general - combinatorial - problem : set - cardinality ] ) denotes the maximum number of users or transmitted data streams , that can be sent over @xmath18 antennas .
if the solution of ( [ eq : general - combinatorial - problem : cost - function ] ) only exists for @xmath42 , the system operates in su - mimo mode .
in such a case , the optimization problem can be formulated to attain mudiv , multiplexing ( high rate ) , and diversity ( high reliability ) gains , see @xcite .
depending on the csit , the type of signaling design and coding applied to the data , theoretical analysis show that the number of users with optimal nonzero power is upper bounded and it becomes loose as the number of transmit antennas grow large .
numerical results comparing the upper bound of @xmath43 for several coding techniques can be found in @xcite .
] as @xmath44 , @xcite . in practical systems ,
multiplexing gain can be scaled up to @xmath45 , by means of linear signal processing [ cf .
section [ section : precoding : linear ] ] . the mathematical formulation in ( [ eq : general - combinatorial - problem : cost - function ] ) resembles a knapsack or subset - sum problem @xcite , which is known to be non - polynomial time complete ( np - c ) .
although the users are fixed items that must be chosen to construct @xmath46 , their associated utility functions change according to the channel conditions and the resource allocation of the co - selected users .
this implies that the optimization variables are , in general , globally coupled . finding the optimal set @xmath46 , is a combinatorial problem due to the binary variables @xmath47 , and the encoding order @xmath36 .
moreover , depending on @xmath48 , problem ( [ eq : general - combinatorial - problem : cost - function ] ) might deal with non - convex functions on the multiple parameters , e.g. @xmath21 , @xmath18 , @xmath29 , etc .
the feasibility of ( [ eq : general - combinatorial - problem : cost - function ] ) relies on the constraints and processing , e.g. the precoding schemes , the power allocation , the csit accuracy , @xcite , [ cf .
section [ section : performance - uf ] ] .
the scheduling problem can be solved optimally by exhaustively searching ( exs ) over all possible set sizes and user permutations . however , the computational complexity of exs is prohibitively high , even for small values of @xmath21 @xcite .
furthermore , problem ( [ eq : general - combinatorial - problem : cost - function ] ) can be modified to include additional dimensions , such as multiple carriers for ofdma systems ( e.g. @xcite ) or codes for cdma systems ( e.g. @xcite ) .
channel knowledge at the transmitter can be modeled taking into account instantaneous or statistical information , e.g. variance , covariance , angles of arrival / departure , and dominant path in line - of - sight @xcite .
there are two main strategies used to obtain csi , reciprocity and feedback , which provide different feedback requirements and robustness to csi errors .
the former , known as open - loop feedback , uses uplink channel information to define the downlink channel in the next transmission interval .
it is suitable for tdd since transmit directions are in identical frequencies , and the channel can be reversed .
the latter , known as closed - loop feedback , requires sending the downlink channel to the transmitter using dedicated pilots , and is commonly used in fdd .
the majority of the papers reviewed in section [ section : algorithms ] assume closed - loop with full or limited channel feedback , and we refer the reader to @xcite and references therein for further discussion on csi acquisition and its impact on system performance . to perform multiple antenna processing , interference mitigation , user scheduling , efficient power allocation , and to profit from mudiv ,
knowledge of csit is compulsory .
the complete lack of csit reduces the multiplexing gain to one , and can not use mudiv for boosting the achievable capacity @xcite . in such scenarios , the optimum resource allocation and transmission schemes
are performed over orthogonal dimensions @xcite . in the literature of mu - mimo ,
a large number of works assume full csi ( error - free ) , at both the receiver ( csir ) and transmitter ( csit ) sides . in practical systems ,
a strong downlink pilot channel , provided by the transmitter , is available to the users , hence the csir estimation error is negligible relative to that of the csit @xcite . for simplicity , it is widely assumed that csir is perfectly known at the mobile terminals . achieving full csit ( ideal noiseless and delay - free feedback ) is highly challenging in practice . feeding
back the csi requires rates that grow rapidly with the transmit power and the number of antennas @xcite .
however , by assuming full csit , it is possible to derive upper bounds on the performance of different signal processing techniques and scheduling algorithms .
the information - theoretic and numerical results using full csit provide useful insights regarding the system performance bounds ( e.g. , @xcite ) .
resource allocation strategies that optimize spectral efficiency , fairness , power consumption , and error probability can be designed to characterize optimal operating points @xcite .
analytical results for full csit reveal the role of each parameter in the system , e.g. , number of deployed transmit and receive antennas , number of active users , snr regime , etc . if channel knowledge is obtained via partial ( rate - limited ) feedback , the information available at the transmitter has finite resolution , resulting in quantization errors .
_ partial csit _ is comprised of two quantities : channel quality information ( cqi ) and channel direction information ( cdi ) @xcite .
the cqi measures the achievable sinr , the channel magnitude , or any other function of the link quality .
the cdi is the quantized version of the original channel direction , which is determined using codebooks [ cf .
section [ section : precoding : partial - csit ] ] .
the transmitter uses both indicators for scheduling [ cf .
section [ section : algorithms : selection - partial - csit ] ] , and the cqi is particularly used for power control , link adaptation , and interference management @xcite .
the csi feedback interval highly depends on the users mobility , and even for short - range communications ( e.g. , wifi ) , immediate feedback is needed to achieve and maintain good performance @xcite . by considering high mobility and limited feedback rates , one can not rely on instantaneous or full csi . in such cases ,
the transmitters perform resource allocation based on _ statistical _ csi , which vary over larger time scales than the instantaneous csi .
the statistics for the downlink and uplink are reciprocal in both fdd and tdd , which can be used to perform resource allocation , see @xcite and references therein .
table [ tab : channel - information ] summarizes the different types of csit in mu - mimo systems , and the antenna configuration at the receivers .
partial csit refers to quantized channel information , which will be elaborated upon sections [ section : precoding : partial - csit ] and [ section : algorithms : selection - partial - csit ] . 0.48l * 2>x & * miso * & * mimo * + * full csit * & @xcite & @xcite + * statistical * & @xcite & @xcite + * outdated * & @xcite & @xcite + * correlated * & @xcite & @xcite + * partial csit * & @xcite & @xcite + the signal processing and scheduling algorithms described in the following sections have been developed and studied for single - hop mu - mimo scenarios .
we have classified the scenarios in two groups , see fig . [
fig : generic - deployments ] : single transmitter and multiple transmitters scenarios . the implemented resource allocation strategies , user scheduling , and signal processing techniques depend on the number of coordinated transmitters , the number of antennas ( @xmath18 and @xmath19 ) , the number of users ( @xmath21 ) , the snr regime , and the csit accuracy .
the system optimization relies on close - loop ( e.g. @xcite ) or open - loop ( e.g. @xcite ) feedback , to achieve spatial multiplexing gains , multiuser diversity gains , and to combat interference . in cellular systems , there are two main sources of interference @xcite : other active devices in the same co - channel and same cell , i.e. , intra - cell or iui ; and from transmissions in other cells , i.e. , inter - cell interference ( ici ) . the techniques to mitigate iui and ici depend on the type of scenario and the optimization criterion .
there exist a number of scenarios where the interference can not be reduced , see @xcite , whose characteristics are described in the following definition : [ defn : interference - limited - system ] _ interference - limited system_. an mu - mimo system is said to be interference limited if the performance metric saturates ( ceiling effect ) with the transmit snr .
this might occur due to csit inaccuracy , highly correlated multiuser channels ( iui ) , and irreducible ici .
+ the objective of mu - mimo processing is to accommodate many users per resource .
therefore , resource allocation strategies are commonly analyzed at the basic resource unit , e.g. , code , single - carrier , time - slot , or frequency - time resource block .
this can be done regardless the global system model ( single - carrier , ofdm , or cdma ) , since the same resource allocation strategy is applied over all resources , e.g. , @xcite .
we adopt a signal model using the most general approach in the reviewed references .
consider a scenario where the transmitter is equipped with @xmath18 antennas , and @xmath21 active users are equipped with @xmath19 antennas .
let @xmath49 , be the discrete - time complex baseband mimo channel of the @xmath23-th user for a given carrier .
the received signal can be expressed as : @xmath50 where @xmath51 , is the joint transmitted signal for all users .
the mimo channel is usually assumed to be ergodic , i.e. , it evolves over time and frequency in an independent and identically distributed ( i.i.d .
the channel is commonly modeled as _
rayleigh fading _
, which is suitable for non line - of - sight communications .
the complete spatial statistics can be described by the second - order moments of the channel @xcite .
define the channel covariance matrix as @xmath52 $ ] , which depends on the antenna configuration , propagation environment , scattering conditions and mobility .
the channel can be decomposed as @xmath53 , where @xmath54 has i.i.d .
entries with distribution @xmath55 .
assuming spatially uncorrelated rayleigh fading channels , i.e. , @xmath56 , @xmath32 and some @xmath57 , is the most common practice in the literature @xcite .
physically , this implies rich scattering environments with sufficient antenna spacing at both ends of the radio link @xcite . under these conditions ,
the fading paths between the multi - antenna transmitter and receiver become independent .
the mu - mimo channels are modeled as narrow band , experiencing frequency - flat ( constant ) fading , where there is no inter - symbol interference @xcite .
a common simplification used in ofdm broadband systems , is to assume multiple flat - fading sub - channels @xcite .
the received signal per sub - channel can be modeled as ( [ eq : received_signal ] ) , avoiding frequency selectivity @xcite .
more realistic channel models for broadband mimo systems and their performance analysis were proposed in @xcite .
the rayleigh model can be thought as a particular case of the asymmetric _ ricean fading _ channel model . in this case , the entries of @xmath54 have non - zero mean and there exists a dominant line - of - sight ( los ) component that increases the average snr @xcite .
several authors model the mimo channel such that the correlation at transmit and receive antennas is distinguishable , and @xmath58 , @xmath32 , see references in table [ tab : channel - information ] .
there are two approaches to model and analyze performance under mimo correlation , the _ jointly correlated model _ @xcite and the simplified _ kronecker model _ @xcite .
the former assumes separability between transmit and receive eigen - directions , and characterizes their mutual dependence .
the latter assumes complete correlation separability between the transmitter and receiver arrays , see @xcite and references therein .
we refer to @xcite for a comprehensive analysis of rayleigh and rician correlated mu - mimo channels .
note that in some mimo propagation scenarios with uncorrelated antennas , the mimo capacity can be low as compared to the siso one due to the _ keyhole _ or _ pinhole _ effect .
this is related to environments where rich scattering around the transmitter and receiver leads to low correlation of the signals , while other propagation effects , like diffraction or waveguiding , lead to a rank reduction of the transfer function matrix @xcite .
the mimo channels @xmath59 , may also include large - scale fading effects due to shadowing and path loss @xcite . depending on the type of access technique ( ofdma , cdma , or tdma ) , the channel model can take into account multi - path components , correlation , doppler spread , and angular properties @xcite .
another common assumption to avoid frequency dependency ( particularly in low mobility scenarios ) , is to account for _ block - fading channels _
this means that the csi is constant ( within a coherence time duration ) for a block of consecutive channel uses before changing independently for the next block .
the noise @xmath60 , is usually modeled by i.i.d .
normalized entries according to a circular normal ( complex ) gaussian distribution with zero mean and unit variance @xcite .
the transmitted signal , @xmath61 , can be defined according to the encoding applied over the user data , the number of spatial streams per user , and the power allocation . if linear precoding is used [ cf .
section [ section : precoding ] ] , the transmitted signal is defined as @xmath62 where @xmath63 , is the precoding matrix , @xmath64 is the data signal , and @xmath65 is the number of multiplexed data streams of user @xmath23 . in single - transmitter mu - mimo scenarios with signal models defined by ( [ eq : received_signal ] ) and ( [ eq : transmitted_signal ] ) ,
the ici is negligible or assumed to be part of the additive background noise .
therefore , iui is the main performance limiting factor , which can be addressed by precoding the user data , [ cf .
section [ section : precoding ] ] .
the active users might experience similar average long - term channel gains ( large - scale fading or path loss ) and snr regimes .
for some practical cellular systems , assuming homogeneous users is valid if open - loop power control is used to compensate for cell - interior and cell - edge path losses .
therefore , the resultant effective multiuser channels have quasi - identical variances @xcite .
the user distribution affects the fading statistics of @xmath66 @xmath32 , and the general design of resource allocation algorithms @xcite .
a possible single - transmitter mu - mimo scenario arises in satellite communications .
however , due to the characteristics of the satellite channels , marginal mimo gains can be realized .
the absence of scatterers in the satellite vicinity yields a rician - type channel with a strong line - of - sight component , turning off the capabilities of mimo processing .
due to the large coverage area in satellite communications , the users have heterogeneous long - term channel gains , which directly affects the resource allocation decisions . regardless of the limited literature about mu - mimo satellite communications , recent works show promising results and discussions about how intensive frequency reuse , user scheduling , and multibeam signal processing will be implemented in next generation broadband satellite systems @xcite . in this scenarios , the channel models and assumptions aforementioned are applied . yet
, some additional considerations are made so that signaling and connections between network entities can be modeled .
deploying several transmitters across a geographic area can provide reliable communication for heterogeneous mobile terminals ( different path losses or snr regimes relative to each transmitter ) .
this kind of infra - structure based systems include cellular and wireless local area networks ( wlan ) .
the resource allocation and access control can be performed based on csi and knowledge of the interference structure @xcite .
if the transmitters are allowed to cooperate , e.g. , through a central processing unit ( cu ) , iui can be mitigated using csit for signal design [ cf .
sections [ section : introduction : needof - csit ] and [ section : precoding ] ] . global knowledge or estimation of the interference can be used to avoid poor spectral efficiency or inaccurate assignment of radio resources .
this paper focuses on scenarios where roaming ( mobility ) and reuse of resources are central management tasks .
the premise behind cellular communications , is to exploit the power falloff with distance of signal propagation to reuse the same channel at spatially - separated locations .
this means that the serving area is divided in non overlapping cells .
any cell site within a neighborhood can not use the same frequency channel , which makes the same reused frequency channels sufficiently far apart @xcite . in traditional cellular systems ,
a given user belongs to only one cell at a time and resource allocation is performed unilaterally by its serving bs ( non - cooperative approach in fig .
[ fig : generic - deployments ] ) .
each transmitter serves its own set of users , transmission parameters are adjusted in a selfish manner by measuring ici ( simple interference - awareness ) , and there is no information exchange between bss @xcite .
if frequency reuse is employed , the bs can make autonomous resource allocation decisions and be sure that no uncoordinated ici appears within the cell @xcite .
however , in many practical systems , universal frequency reuse is applied , which means that neighboring cells can access the same frequencies and time - slots simultaneously .
this might increase the ici and potentially degrade performance @xcite .
the mitigation of ici is a fundamental problem since the transmit strategy chosen by one bs will affect the reception quality of the users served by adjacent bss .
a cluster of bss can coordinate the resource allocation , scheduling decisions , and ici mitigation techniques ( cooperative approach in fig .
[ fig : generic - deployments ] ) .
dynamic clustering is an ongoing research topic ( see @xcite and references therein ) , which promises to meet the requirements established in the third generation partnership project ( 3gpp ) standards @xcite .
different forms of ici control have been proposed over the last years .
extensions of space multiple access techniques for multi - cell systems have received several names , _ coordinated multi - point _ ( comp ) @xcite , _ network mimo _
@xcite , or _ joint signal transmission / processing _ ( jt ) @xcite .
these techniques exploit the spatial dimensions , serving multiple users ( specially cell edge users @xcite ) , while mitigating ici of clustered bss .
for these approaches , a cluster can be treated as a super cell , for which mathematical models from the single - cell scenario can be applied straightforwardly , e.g. @xcite .
if user data is shared among bss , the use of _ proactive _ interference mitigation within a cluster can take place .
this implies that coordinated bss do not separately design their physical ( phy ) and media access control layer parameters . instead , the bss coordinate their coding and decoding , exploiting knowledge of global data and csi @xcite
however , to guarantee large performance gains for these systems , several conditions must be met @xcite : global csi and data sharing availability , which scales up requirements for channel estimation , backhaul capacity , and cooperation ; coherent joint transmission and accurate synchronization ; and centralized resource allocation algorithms , which may be infeasible in terms of computation load and scalability .
there is another approach of multi - cell cooperation , coined as _ coordinated scheduling _ ( cs ) with _ coordinated beamforming _ ( cbf ) , which is a form of coordinated transmission for interference mitigation @xcite .
cs / cbf refers to the partial or total sharing of csi between bss to estimate spatial signaling , power allocation , and scheduling without sharing data or performing signal - level synchronization @xcite .
cbf implies that each bs has a disjoint set of users to serve , but selects transmit strategies jointly with all other bss to reduce ici . in this approach , exchange of user data
is not necessary , but control information and csi can be exchanged to simultaneously transmit to a particular set of users @xcite .
cbf is more suitable than jt for practical implementations , since it requires less information exchange .
nevertheless , csi acquisition , control signaling , and coordinated scheduling are challenging tasks due to limited feedback bandwidth and finite capacity backhaul @xcite .
this paper covers two wireless technologies , whose specifications already support mu - mimo communications : lte - advanced for cellular networks and ieee 802.11ac for wireless local area networks ( wlan ) .
@xmath67 _ lte - advanced _ : this is the 3gpp cellular system standard for 4 g and beyond communications @xcite .
several capabilities have been added to the lte standard to increase capacity demands and integrated a large number of features in the access network . among these attributes ,
the ones related to mimo processing are the most relevant in the context of this paper : enhanced downlink mimo , multi - point and coordinated transmission schemes , and multi - antenna enhancements . due to the fact that lte is a cellular technology , most of the studied deployments in the literature lie in the category of multiple transmitters scenarios ,
see fig .
[ fig : generic - deployments ] .
mu - mimo communication has been incorporated in lte with the following maximum values : four users in mu - mimo configuration , two layers ( spatial streams ) per user , four simultaneous layers , and robust csi tracking .
practical antenna deployments at the transmitter use dual - polarized arrays , and the expected number of co - scheduled users for such configurations will be two for most cases @xcite .
lte provides a mechanism to improve performance by switching between mu or su - mimo mode on a per sub - frame basis , based on csi , traffic type , and data loads .
the goal of dynamic mode switching is to balance the spectral efficiency of cell edge and average cell users .
this can be achieved , for instance , by using transmit diversity for users at the cell edge , or implementing spatial multiplexing for cell center users @xcite .
@xmath67 _ ieee 802.11ac _ : this wlan standard supports multiuser downlink transmission , and the number of simultaneous data streams is limited by the number of antennas at the transmitter . in mu - mimo mode , it is possible to simultaneously transmit up to 8 independent data streams and up to 4 users @xcite .
spatial multiplexing is achieved using different modulation and coding schemes per stream .
mu - mimo transmission prevents the user equipment with less antennas to limit the achievable capacity of other multiple antenna users , which generates rate gains for all receivers .
a unique _ compressed explicit feedback _ protocol , based on channel sounding sequences , guarantees interoperability and is used to estimate csi and define the steering matrices ( beamforming ) .
other methods for channel estimation are described in @xcite and references therein .
compared to cellular networks , wlans usually have fewer users moving at lower speeds , the aps are less powerful that the bss and the network topology is ad hoc .
although multiple aps could be deployed , most of the works in the literature focus on mu - mimo systems with a single transmitter .
the spatial dimension provided by the multi - antenna transceivers can be used to create independent channelization schemes . in this way
, the transmitter serves different users simultaneously over the same time slot and frequency band , which is known as space - division multiple access ( sdma ) @xcite .
the spatial steering of independent signals consists of manipulating their amplitude and phases ( the concept of _ beamforming _ in classic array signal processing ) , in order to add them up constructively in desired directions and destructively in the undesired ones @xcite . by jointly encoding all ( co - resource ) signals using channel information , it is possible to increase the signal - to - interference - plus - noise ratio ( sinr ) at the intended receiver and mitigate interference for non - intended receivers . in the literature of mu - mimo systems
, the term beamforming refers to the signal steering by means of beams to achieve sdma .
the term _ precoding _ is used to denote the scaling and rotation of the set of beams , so that , their power and spatial properties are modified according to a specific goal .
hereinafter , we use the term precoding to describe the signal processing ( i.e. , beam vector / matrix computation , scaling , rotation , and projection ) , applied to the independent signals prior to transmission . in this section
we describe the most common precoding techniques used in mu - mimo scenarios , and their characteristics according to csit .
table [ tab : precoding ] summarizes the precoding schemes used in the surveyed literature , as well as their associated methods for user selection , which will be elaborated upon in the following sections .
an important performance metric determined by the precoders @xmath68 , @xmath32 , related to the delivered energy through the mu - mimo channels , is provided in the following definition .
[ defn : effective_channel_gains ] _ effective channel gain_. it is the magnitude of the channel projected onto its associated precoding weight . let @xmath69 , be the _ effective channel _ after spatial steering , thus , the effective channel gain is given by : _ i _ ) @xmath70 for miso scenarios ; _ ii _ ) for mimo scenarios , it is given by a function of the eigenvalues of @xmath71 : @xmath72 or @xmath73 .
* 6>x * method * & * utility - based * & * csi - mapping * & * metaheuristic(stochastic ) * & * classicoptimization * & * exhaustive * * search * + * dpc * & @xcite & - & @xcite & - & - + * thp * & @xcite & - & - & - & - + * vp * & @xcite & @xcite & - & - & - + * mrt * & @xcite & @xcite & @xcite & - & - + * zfbf * & @xcite & @xcite & @xcite & @xcite & @xcite + * zfdp * & @xcite & @xcite & - & - & - + * szf * & - & @xcite & @xcite & - & - + * cizf * & - & @xcite & - & - & - + * bd * & @xcite & @xcite & @xcite & @xcite & - + * mmse / slnr * & @xcite & @xcite , @xcite , @xcite & @xcite & - & - + * adaptive * & @xcite & @xcite & - & @xcite & - + * codebook - based * & @xcite & @xcite & - & - & - + from an information - theoretic perspective , the optimal transmit strategy for the mu - mimo bc is _ dirty paper coding _ ( dpc ) @xcite , and theoretical results showed that such strategy achieves the entire bc capacity region @xcite .
the principle behind this optimum coding technique is that the transmitter knows the interference for each user .
therefore , interference can be pre - subtracted ( from the information theoretical standpoint ) before transmission , which yields the capacity of an interference free channel .
dpc is a non - linear process that requires successive encoding and decoding , whose performance depends on the particular sequential order @xmath36 assigned to the co - scheduled users @xcite . although the implementation complexity of dpc for practical systems is prohibitively high , it establishes the fundamental capacity limits for mu - mimo broadcast channels @xcite .
suboptimal yet more practical non - linear precoding schemes have been proposed as an alternative to dpc @xcite .
the error rate and interference can be minimized at the symbol level by the tomlinson - harashima precoding ( thp ) , which is not limited by the number of transmit or receive antennas @xcite . by modifying or perturbing the characteristics of the transmitted signal ,
the power consumption can be minimized ( compared to traditional channel inversion filtering ) , using the non - linear vector perturbation ( vp ) scheme @xcite .
this technique requires a multidimensional integer - lattice least squares optimization , whose solution can be found by several approaches .
linear precoding is a generalization of traditional sdma @xcite , which matches the signal on both ends of the radio link .
this is attained by decoupling the input data into orthogonal spatial beams and allocating power according to csit @xcite .
the precoders weights are vectors or matrices jointly designed at the transmitter , according to several parameters : the type of csit , the coding order , the performance metric ( e.g. , mean - square error ( mse ) , error probability , power consumption , or achievable snr ) , and the system constraints ( e.g. , power and qos ) . the optimal precoder technique ( linear filtering in the spatial domain ) , would be able to balance between signal power maximization and interference power minimization @xcite .
precoding design subject to general constraints can be performed using standard optimization techniques ( see @xcite for a comprehensive survey ) or by heuristic approaches , e.g. , maximizing the signal - to - leakage - plus - noise ration ( slnr ) @xcite .
however , determining the optimal precoders is an np - hard problem for many performance metrics @xcite , whose evaluation is performed by computationally demanding algorithms @xcite .
therefore , many works in the literature focus on more suboptimal , yet practical , schemes that can achieve spatial multiplexing gains with low computational complexity . a large number of linear precoding techniques have been developed , for which having more transmit than receive antennas , i.e. , @xmath74 , is a condition required in most cases .
there are some particular precoding schemes ( e.g. slnr ) that can be implemented if @xmath75 , but the system becomes interference limited in the moderate and high signal - to - noise ( snr ) regimes , [ cf . definition [ defn : interference - limited - system ] ]
. therefore , power control and recursive adaptation of the precoders are mandatory to operate in those snr regimes @xcite . in a miso antenna configuration , the matched filter or maximum ratio transmission ( mrt ) precoding maximizes the signal power at the intended users .
this is performed by projecting the data symbol onto the beamforming vector given by the spatial direction of the intended channel @xcite .
a similar precoding scheme for mimo scenarios is the singular value decomposition ( svd ) beamforming @xcite , which uses the eigenvectors of the channel as beamforming weights .
the zero - forcing beamforming and can asymptotically achieve the dpc performance when @xmath76 @xcite . ]
( zfbf ) @xcite , also known as channel inversion precoding @xcite , completely suppresses iui in miso scenarios .
this technique is based on prefiltering the transmit signal vector by means of the moore - penrose inverse @xcite .
an extension of zfbf for mimo scenarios is the block diagonalization ( bd ) precoding , where multiple data streams can be transmitted per user @xcite . in miso scenarios ,
the regularized channel inversion ( often coined as mmse precoding ) enhances zf , taking into account the noise variance to improve performance in the low snr regime @xcite .
an extension of mmse for the general mimo scenario is the regularized bd @xcite .
zero - forcing dirty paper ( zfdp ) coding @xcite , is a technique designed for miso settings . for a given user @xmath23
, zfdp suppresses the interference coming from the next encoded users @xmath77 , combining qr decomposition @xcite and dpc .
extensions of zfdp for the mimo scenarios were defined in @xcite ( an iterative svd method ) , and @xcite ( combining zf , dpc , and eigen - beamforming ) .
successive zero - forcing ( szf ) was proposed in @xcite , for mimo settings .
szf partially suppresses iui , by encoding users similar to zfdp , but dpc is not applied in the encoding process .
the generalization of precoding schemes based on zf for multiple antenna receivers is not trivial .
this is because applying miso decompositions methods to the mimo channel is equivalent to treating each receive antenna as an independent user .
this process does not completely exploit the multiplexing and diversity gains of mimo systems @xcite .
the aforementioned precoding schemes can be classified as user - level precoders , i.e. , independent codewords intended to different users are transmitted simultaneously .
there is another class of precoding schemes , where simultaneous transmitted symbols are addressed to different users @xcite .
this class of symbol - level precoding , e.g. constructive interference zero forcing ( cizf ) precoding @xcite , has been developed for miso settings .
cizf constructively correlates the interference among the spatial streams , rather than decorrelating them completely as in the case of user - level precoding schemes . in the literature of limited feedback systems ,
csit acquisition relies on that are known a priori at the transmitter and receiver sides .
the codewords can be deterministic or randomly constructed , which defines the type of signal processing applied to achieve spatial multiplexing and interference mitigation @xcite : @xmath67 _ channel quantization and precoding _ : the codebook @xmath78 , is used by user @xmath23 to quantize its channel direction with @xmath79 bits .
this means that each user feeds back the index @xmath80 , of the most co - linear codeword to its channel , [ cf .
section [ section : metrics - spatial - compatibility : spatial - clustering ] ] .
this is illustrated in fig .
[ fig : hyperslab_cones ] , where the user @xmath23 would report the index @xmath80 , related to the cone where @xmath33 has been clustered .
the precoding weights are usually computed using zfbf over the quantized channels .
due to quantization errors , the signals can not be perfectly orthogonalized , and the sum - rate reaches a ceiling as the snr regime increases @xcite .
in other words , resource allocation is performed over non - orthogonal spatial directions .
the optimal codebook design has not been fully solved in the literature . however , if the channel is assumed to have spatially i.i.d . entries , ( @xmath81 , and homogeneous long - term channel gains , @xmath32 ) , off - line designs of @xmath82 can be realized using different approaches , e.g. , the grassmannian design @xcite , random vector quantization ( vq ) @xcite , quantization cell approximation ( qca ) @xcite , and other techniques described in @xcite .
this sort of isotropically distributed codebooks achieve acceptable performance , since they mirror the statistical properties of the eigen - directions of @xmath33 , @xmath32 . for correlated channels ( @xmath83 ) ,
non - uniform or skewed codebooks must be constructed , taking into account the statistical characteristics of the dominant eigen - directions of @xmath84 , @xmath32 , see @xcite .
@xmath67 _ random beamforming ( rbf ) _
@xcite : the available precoding vectors , @xmath85 , are constructed at the transmitter according to a known distribution , codewords that are i.i.d . , according to the stationary distribution of the unquantized beamforming vector @xcite . ] or by methods that yield random orthonormal basis in @xmath86 , @xcite .
the transmitter sends pilot symbols to different user through the beams in @xmath87 , and each user feeds back the index @xmath80 , of its best beam [ cf .
section [ section : algorithms : selection - partial - csit ] ] .
rbf is an extension of the opportunistic beamforming scheme in @xcite , and attempts to sustain multiuser diversity over fading channels with partial csit .
random changes of amplitude and phase at each transmit antenna result in channels experiencing accelerated fluctuations , which enhances the mimo processing gains .
this approach is only effective when the number of users is large , .
however , it does not achieve the full multiplexing gain at high snr @xcite , and suffers from large quantization error as @xmath18 grows @xcite . ]
@xmath88 , and the number of antennas at the transmitter @xmath18 is moderate @xcite . in some scenarios
the performance can be enhanced by power control , or employing statistical channel information to compensate poor mudiv .
we hasten to say that defining the optimum @xmath79 , either for @xmath82 or @xmath87 , is not a trivial optimization problem . in practice ,
only a tradeoff between feed back load and performance should be sought . under mild conditions
, @xmath79 can be minimized for certain codebook designs while guaranteeing diversity gains @xcite .
large codebooks provide accurate csit at a price of factors , large feedback signaling overhead and memory requirements at the receiver , which increase exponentially with @xmath79 @xcite .
recent research work have focused on transmission cooperation and coordination in multi - cellular and heterogeneous networks .
the signal processing techniques performed at the transmitters depend on several aspects @xcite : shared or non - shared user data ; global or local csit ; full or partial csi ; level of synchronization ; per - transmitter power , backhaul , and delay constraints ; full or limited coordination ; heterogeneous snr regime across the receivers ; and the number of clustered transmitters . in multi - cell scenarios , assuming full csit and shared user data across the transmitter , the precoders can be computed by techniques described in section [ section : precoding : linear ] .
however , the power should be constrained on a per - transmitter basis , instead of global power allocation as in single - transmitter scenarios @xcite .
the precoding design can be oriented to optimize different objective functions , e.g. power minimization or sum - rate maximization .
optimizing such functions deals , in general , with non - convex problems , whose solutions are non - linear precoders @xcite .
precoding weights designed for common utility functions , e.g. weighted - sum - rate maximization or mean - square - error ( mse ) minimization , can be realized via standard optimization techniques and sophisticated high - performance algorithms , see @xcite and references therein .
however , a more pragmatic approach for mu - mimo , is given by linear precoding schemes ( e.g. zfbf , slnr , or bd ) , which can be extended for multi - cell systems and providing reliable and low - complexity solutions @xcite . in _
network mimo _ scenarios , multi - cell bd precoding can remove ici for the clustered cells using centralized processing at a central controller ( full coordination ) @xcite . in miso
multi - cell scenarios , linear precoding can be performed in a distributed fashion or with partial coordination .
there are two fundamental schemes @xcite : mrt ( a competitive or egoistic scheme ) and inter - cell interference cancellation ( icic is a cooperative or altruistic scheme ) . for the mrt scheme ,
the transmitters ignore the ici that they cause to unintended receivers in their vicinity .
the goal is to maximize the received signal power of local users ( effective in non - cooperative systems ) .
the icic is a zf - based scheme , which means that the transmitters design their precoders so that no interference is cause to non - intended receivers ( effective in cooperative systems ) .
the design of icic / zf precoding is subject to constraints in the number of transmit and receive antennas , and its effectiveness highly depends on the csit across the transmitters . a close - to - optimal distributed precoding scheme ( for an arbitrary snr ) , is attained by balancing mrt and icic , whose mathematical formulation is discussed in @xcite .
if the system operates under a limited feedback constraint , the csit is acquired using the principles described in section [ section : precoding : partial - csit ] .
precoding design extensions from single to multi - cell scenarios are not straightforward .
for instance , in multi - cell cooperative systems , the codebooks sizes may significantly increase @xcite , and depending on the delay tolerance and codebook granularity , the clustered transmitters should switch between mrt and icic schemes @xcite .
several multi - cell cbf designs attempt to suppress iui and ici simultaneously with minimum coordination or control signaling between transmitters .
these schemes rely on instantaneous or statistical csi , and can be classified as @xcite : _ hierarchical _ and _ coupled _ precoders .
the hierarchical approach is implemented in systems where each transmitter suppresses ici using zf .
this is attained through the sequential construction of outer and inner precoders that reduce ici and iui , respectively .
the ici cancellation is achieved by aligning interference subspaces at the receivers , whilst the iui is suppressed at the transmitters using linear precoding and local csi , e.g. , @xcite .
the coupled or nested - structure approach , is implemented in interference limited system , [ cf . definition [ defn : interference - limited - system ] ] .
the precoding design ( beamforming weights and power allocation ) optimizes an objective function subject to a set of qos constraints , [ cf .
section [ section : performance - uf ] ] .
this kind of optimization problems has been extensively studied in the literature , covering power allocation @xcite and beamforming design @xcite . under certain conditions ,
two specific problem formulations admit global optimal solutions via standard optimization in multi - cellular scenarios @xcite : _
i ) power minimization subject to individual sinr constraints_. the cbf approach in @xcite mitigates ici by iteratively adjusting beamforming weights and transmit powers according to the experienced interference per user . _
ii ) the maximization of minimum sinr subject to power constraints_. the authors in @xcite tackled this problem using the perron - frobenius theory @xcite .
the joint beamforming and power allocation design is reformulated as an eigenvalue - eigenvector optimization problem , whose optimal solution can be found by geometrically - fast convergent algorithms . in the lte specification , csi acquisition
is called _ implicit feedback _ and relies on the following parameters @xcite : rank indicator ( ri ) , which defines the number of data streams recommended for su - mimo transmission ; precoding matrix index ( pmi ) , which is the index of the best precoding matrix in the codebook ; and channel quality indicator ( cqi ) , which contains information of the channel quality corresponding to the reported ri and pmi @xcite .
the ri and pmi indexes provide the cdi of the mimo channels , while the cqi indicates the strength of the corresponding spatial direction .
the precoding matrices are defined at the bss using different approaches @xcite : codebook - based precoding or non - codebook - based precoding ( arbitrary precoder selection based on ri ) .
the standard also supports a dual codebook structure in mu - mimo mode .
this increases codebook granularity , suppresses iui more efficiently , and enhances the overall performance .
the idea is that one codebook tracks the wideband and long - term channel characteristics , while the other tracks the frequency - selective and short - term channel variations @xcite . in this way
, the transmitter has more flexibility and accuracy when designing the precoding matrix .
another supported scheme is the _ per unitary basis stream user and rate control _ ( pu2rc ) , which allows multiple orthonormal bases per codebook , increasing quantization granularity @xcite .
although pu2rc uses deterministic codebooks , random codebooks can be used to simplify theoretical analysis , e.g. @xcite .
the optimum number of bases in the codebook depends on the number of active users @xmath21 , and should be optimized to maximize mudiv and multiplexing gains @xcite .
lte standard supports interference mitigation based on linear precoding schemes , such as slnr and zfbf with quantized channels @xcite .
the precoders can be dynamically recalculated after each csi update , i.e. , tracking channel variations .
linear precoding based on quantized channels outperforms codebook - based precoding ( rbf or pu2rc ) , when then number of active users is small , i.e. , _ sparse networks _ , @xmath89 @xcite .
the reverse holds as mudiv increases , @xmath90 , @xcite .
one popular technique to construct the steering or precoding matrix is through svd precoding @xcite .
since knowledge of csit is mandatory and feedback rates are limited , the receivers represent their estimated precoding matrix with orthogonal columns using givens rotations @xcite . then , the set of calculated parameters ( angles ) at the receivers are adjusted , quantized , and fed back to the transmitter . in general , the final precoding matrix calculated by the transmitter will be different from the weights reported by the users due to the orthogonalization process @xcite .
another practical approach is to compute the steering matrix using linear precoding schemes , e.g. zfbf or mmse @xcite . in @xcite ,
the authors used bd and regularized bd with geometric - mean decomposition for a mu - mimo scenario .
the designs goal is to balance the achievable snr across the users , meeting the requirements of the standard . if the number of transmit antennas is larger than the total number of receive antennas , the additional dof can be used to efficiently nullify inter - stream interference @xcite .
let @xmath26 and @xmath91 be two sets of user with at least one non - common user , where @xmath92 and @xmath93 are their associated channels .
a metric for spatial compatibility is a function of the csit that maps the spatial properties of the multiuser mimo channels to a positive scalar value quantifying how efficiently such channels can be separated in space @xcite , i.e. , @xmath94 .
consider the two subsets @xmath46 and @xmath95 , a metric of spatial compatibility can be used to estimate the achievable performance of their associated multiuser channels , e.g. , having @xmath96 may imply that @xmath46 is the set that achieves the maximum capacity .
the mapping function @xmath97 can be used for scheduling purposes , depending on its definition and other system parameters ( e.g. snr regime and @xmath18 ) , which will be discussed in section [ section : algorithms : aggregated - utility : csi - mapping ] .
below we provide two definitions related to the design of user scheduling policies .
[ defn : compatible_set_of_users ] _ spatially compatible users_. a feasible set of users @xmath46 , is spatially compatible if the multiuser mimo channels , @xmath98 , can be separated in the spatial domain by means of beamforming / precoding .
[ defn : user_grouping ] _ user grouping_. it is the task of forming a subset of users @xmath46 , according a compatibility criterion , e.g. spatial separability in definition [ defn : compatible_set_of_users ] , to maximize the resource allocation and scheduling efficiency .
one of the objectives of mu - mimo technology is to multiplex independent data streams to different users , which implies that only a subset of the transmitted data symbols are useful for each co - scheduled user . in such a scenario , a fundamental problem is to mitigate iui , i.e. , suppress the information intended to other receivers . there exist several signal processing techniques that can achieve such a goal , e.g. , linear precoding or interference alignment .
the effectiveness of such techniques rely upon the characteristics of the subspaces spanned by the mimo channels , i.e , iui is a function of the overlapped interference subspaces @xcite .
consider a set of user @xmath46 with @xmath21 users , let @xmath99^{t}$ ] , be the aggregated interference matrix of the user @xmath23 such that @xmath100 , which is a necessary condition to suppress iui @xcite .
define @xmath101 , as the subspace spanned by the channels of the subset of users @xmath102 , and let @xmath103 , be its orthogonal complement subspace .
in other words , @xmath104 spans the null space of @xmath105 , i.e. , @xmath106 .
the channel of the @xmath23-th user can be expressed as the sum of two vectors @xmath107 , each one representing the projection of @xmath33 onto the subspaces @xmath108 and @xmath104 respectively , as illustrated in fig .
[ fig : space - projections ] . for all user - level
zf - based precoding schemes described in section [ section : precoding : linear ] , @xmath109 , i.e. , @xmath104 contains all overlapped interference subspaces of channel @xmath33 . the component @xmath110 is related to the signal degradation of the user @xmath23 due to channel correlation , whereas @xmath111 defines the _ zero - forcing direction _ , i.e. the spatial direction that is free of iui .
the squared magnitude of @xmath111 is known as the null space projection ( nsp ) , and directly computes the effective channel gain obtained by zf precoding @xcite .
given by its projection onto the subspaces @xmath108 and @xmath104.,title="fig : " ] + a common approach in the literature is to define the mapping @xmath112 as a function of the exact or approximated _ effective channel gains _ [ cf . definition [ defn : effective_channel_gains ] ] .
in other words , the spatial compatibility metric has to consider the precoding scheme , the snr regime , and the available spatial dof .
analytical results in @xcite have shown that the set of users @xmath46 that maximizes the product of their effective channel gains .
] is also the set that maximizes the sum - rate in the high snr regime .
the nsp has been extensively used for user grouping and scheduling purposes whenever zf - based precoding is implemented [ cf .
section [ section : algorithms : aggregated - utility ] ] . note that a set of spatially compatible users would have large nsp components , which means that their associated channels are semi - orthogonal . the computation of nsp for the general mu - mimo scenario where @xmath113 @xmath32 , can be performed by several approaches , namely svd @xcite , orthogonal projection matrix @xcite , gram - schmidt orthogonalization ( gso ) @xcite , qr decomposition @xcite , products of the partial correlation coefficients @xcite , and ratio of determinants @xcite .
several works find sets of spatially compatible users by computing an approximation of the nsp ( e.g. @xcite ) , which yields efficient scheduling designs that require low computational complexity .
moreover , recent works ( @xcite and references therein ) , have proposed efficient methods to compute and track null spaces for a set of co - scheduled users , which is fundamental to mitigate iui for zf - based precoding in single and multiple transmitter scenarios .
@xmath114 where @xmath115 $ ] , and @xmath116 , implies that the channels are spatially uncorrelated or orthogonal .
the coefficient of correlation indicates how efficiently the transmitter can serve user @xmath117 without affecting user @xmath118 , and vice versa . for particular mu - miso scenarios , in which the set of scheduled users is subject to the cardinality constraint @xmath119 ,
the nsp is a function of @xmath120 , which can be evaluated from ( [ eq : coefficient_correlation ] ) simplifying the user pairing and scheduling design @xcite .
several works focused on sum - rate analysis ( e.g. , @xcite ) limit the cardinality to @xmath119 for simplicity and tractability .
other works analyzed scenarios with more practical constraints and considered two or four users , which are common values of @xmath21 in standardized systems , see @xcite .
the metric ( [ eq : coefficient_correlation ] ) has been extensively used to define policies for user grouping . in mu - mimo systems a set of user @xmath46 is called @xmath121-orthogonal if @xmath122 for every @xmath123 with @xmath124 @xcite .
the users can be grouped into disjoint sets according to a desired threshold @xmath121 , and the semi - orthogonal sets can be scheduled over independent carriers ( e.g. @xcite ) , codes ( e.g. @xcite ) , or time slots ( e.g. @xcite ) .
the optimum value of @xmath121 depends on the deployment parameters ( @xmath21 and @xmath18 ) and is usually calculated through simulations , since for @xmath125 it is very hard or impossible to find the optimal @xmath121 in closed - form
. nevertheless , there exist closed - form expressions to compute the ergodic capacity as a function of @xmath121 for mu - miso scenarios where the transmitter has two antennas @xmath126 and the users have homogeneous ( * ? ? ?
7 ) or heterogeneous @xcite large - scale fading gains . the generalization of the hyperslab clustering as a function of two parameters , @xmath127 and the minimum acceptable channel magnitude . in this way only sets of spatially compatible and strong channels can be constructed . ] for mimo settings is straightforward , i.e. @xmath128 , @xmath129 , and the coefficient of correlation between matrices can be defined as in @xcite . the parameter @xmath130 shown in fig .
[ fig : hyperslab_cones ] can be adjusted to fix the maximum co - linearity between cones @xcite .
several scheduling algorithms based on spatial clustering ( e.g. , @xcite ) , can achieve mudiv gains and improve the overall performance by adjusting the parameter @xmath127 ( or threshold @xmath121 ) according to the number of competing users @xcite , the snr regime , the large - scale fading gain , and the precoding scheme @xcite . in @xcite , it was shown that the optimal cardinality of the user set grouped based on spatial clustering is a function of @xmath21 , @xmath127 , and @xmath18 .
the statistical properties of ( [ eq : coefficient_correlation ] ) have been extensively studied in the literature cf .
@xcite , and such properties depend on the type of csit ( full or partial ) , the mimo channel distribution , and the system parameters ( e.g. @xmath18 and @xmath79 ) . can be clustered within cones given a set @xmath131 formed either by orthonormal basis or codebook words .
, title="fig : " ] + consider a set of users @xmath46 and its mu - mimo channel @xmath98 ; the spatial compatibility can be measured as a function of the corresponding eigenvalues : 1 .
the _ orthogonality defect _
@xcite is a metric derived from hadamard s inequality @xcite that measures how close a basis is to orthogonal .
it quantifies the energy degradation of the channel matrix due to the correlation between all its column vectors .
user grouping algorithms based on this metric have been developed for homogeneous mimo systems , e.g. @xcite .
2 . the determinant geometrically represents the volume of the parallelepiped defined by the column vectors of the channel matrix .
larger determinant values implies that the column vectors of a matrix are more orthogonal @xcite .
the _ matrix volume ratio _ used in @xcite measures the volume reduction of @xmath98 w.r.t .
@xmath132 , where @xmath133 is a candidate user attempting to be grouped .
different approximations of such a metric can be done using the arithmetic - geometric mean inequality over the squared singular values of @xmath98 , e.g. , @xcite .
3 . the _ geometrical angle _ ( see @xcite and references therein ) , is a metric similar to the matrix volume ratio that measures the spatial compatibility as a function of the all possible correlation coefficients [ cf .
( [ eq : coefficient_correlation ] ) ] between the basis of two subspaces .
it can be computed from the eigenvalues of two mimo matrices @xmath134 , @xmath135 , whose subspaces have different dimensions , i.e. , @xmath136 .
if the elements of a multiuser channel matrix are highly correlated , the matrix is said to be ill - conditioned , i.e. , it is close to singular and can not be inverted . in numerical analysis , the _ condition number _ @xcite quantifies whether a matrix is well- or ill - conditioned , and is computed as the ratio between the maximum and minimum eigenvalues .
this metric is used to measure how the eigenvalues of @xmath98 spread out due to spatial correlation .
the ratio between the condition numbers of two mimo channels can be used to quantify their spatial distance and compatibility @xcite .
for a given user @xmath27 , the spatial compatibility between @xmath33 and @xmath105 , can be measured by a function of @xmath137 , which is the angle between @xmath108 and @xmath138 , see fig .
[ fig : space - projections ] .
the following metrics assess the spatial compatibility based on geometrical properties of the multiuser channels : 1 . the _ principal angle between subspaces _ @xcite measures the relative orientation of the basis of @xmath139 regarding the basis of @xmath140 . given the bases of @xmath141 and @xmath142 ,
the principal angle is associated with the largest coefficient of correlation [ cf .
( [ eq : coefficient_correlation ] ) ] between the bases of both subspaces .
the _ chordal distance _ is extensively used in limited feedback systems for codebook design @xcite , user grouping and scheduling @xcite .
this metric can be computed either from the principal angles between subspaces or from the projection matrices of @xmath141 and @xmath142 in heterogeneous mu - mimo scenarios , i.e. , @xmath136 @xmath143 with @xmath123 .
the _ subspace collinearity _ @xcite quantifies how similar the subspaces spanned by two channel matrices are , following the rationale behind ( [ eq : coefficient_correlation ] ) .
given @xmath144 the subspace collinearity of the matrices compares the singular values and the spatial alignment of their associated singular vectors .
other metrics measuring the distance between subspaces are the _ weighted likelihood similarity measure _ , the _ subspace projection measure _ , and the _ fubini - study similarity metric_. these metrics have been recently propounded in @xcite and used for user grouping based on statistical csi , i.e. , @xmath145 .
l * 5>x & * sum rate * & * fairness * & * wsr * & * qos * & * round robin * + * miso * , @xmath146 & @xcite & @xcite & @xcite & @xcite & @xcite + * miso * , @xmath147 & @xcite & @xcite & @xcite & @xcite & @xcite + * mimo * , @xmath146 & @xcite & @xcite & @xcite & @xcite & @xcite + * mimo * , @xmath147 & @xcite & @xcite & @xcite & @xcite & @xcite + the optimization criteria determines the optimal resource allocation strategy @xcite , and can be classified in two groups according to the objective function and constraints @xcite .
_ i _ ) _ phy layer _ based criteria , where a objective function , @xmath148 , must be optimized and channel information is the only input to the resource allocation algorithms .
_ ii _ ) _ cross layer _ based criteria , where optimization of @xmath148 , takes into account qos requirements ( defined by upper layers ) and channel information , see fig . [
fig : system - processing - blocks ] .
this section presents an overview and classification of the objective functions ( criteria ) , and their associated constraints in the mu - mimo literature .
a summary of the content and general organization of this section are presented in fig .
[ fig : utility - functions - and - constraints ] and table [ tab : optimization - criteria ] .
two relevant concepts in multiuser system optimization are defined below .
+ the system optimization is achieved by maximizing the sum of individual utility functions , @xmath149 , @xmath150 , subject to a set of power constraints and set cardinality ( @xmath43 ) .
global optimization based on concave and differentiable utility functions is desirable , since efficient numerical methods can be applied to guarantee optimality .
a common family of functions studied in the literature is defined by the @xmath151__-fair _ _ utility function @xcite , which can be used to characterize the objective of the resource allocation strategies .
the @xmath151__-fair _ _ function is mathematically expressed as @xcite : u ( \{r_k}_k=1^k ) = _ k=1^k ( r_k ) , & [ eq : alpha - fair - utility - function-1 ] + _ k=1^k , & [ eq : alpha - fair - utility - function-2 ] where @xmath152 denotes the achievable transmission rate of the @xmath23-th user .
changing the values of @xmath151 yields different priorities to the users , and can be used to define performance tradeoffs ( e.g. throughput - fairness ) .
for instance , @xmath153 , yields maximum fairness , @xmath154 , generates the sum - rate utility , and @xmath155 defines the minimum - rate function .
these functions are , in general , non - concave , and several methods have been developed to optimize them subject to power and qos constraints , see @xcite for an in - depth review . a more common formulation of the weighted sum rate in mu - mimo scenarios is given by the following expression : @xmath156 where @xmath157 is a non - negative time - varying weight , which defines the priority of the @xmath23-th user .
this expression is related to the @xmath151-fair function , e.g. if @xmath158 , then ( [ eq : weighted - sum - rate - function ] ) converges to ( [ eq : alpha - fair - utility - function-2 ] ) with @xmath154 .
the dynamic adaptation of @xmath157 can provide different levels of fairness ( temporal , utilitarian , rate - proportional , fair - queuing , etc . ) over the resource allocation at the _ user level _ @xcite .
another approach to provide fairness is defining @xmath159 as a function of the rates statistics .
for instance , one might use the cumulative distribution function ( cdf ) of @xmath160 @xmath32 , which can be empirically estimated at the transmitter @xcite .
fairness at the _ system level _ is another optimization criterion , measured over the average rates , transmit powers , time - slots , or any other allocated resource .
different metrics to quantify fairness can be found in the literature of resource allocation @xcite . for a single resource
, one can use the gini index @xcite , jain s index @xcite , or other fairness measures in @xcite and reference therein .
generalized fairness metrics over multiple resources allocated simultaneously were defined in @xcite and references therein .
the weights in ( [ eq : weighted - sum - rate - function ] ) might be determined by upper layer requirements , such as buffer sizes , traffic / service priority , packet loads , or any other qos - related metric .
the individual weights affect the scheduling decisions , and can be used to improve precoding design @xcite . a special case of ( [ eq : weighted - sum - rate - function ] )
is given by setting @xmath158 @xmath32 , which represents the sum rate .
this criterion defines the maximum amount of error - free information successfully received by a set of co - scheduled users , regardless of fairness @xcite .
the resource allocation strategies are usually assessed in terms of sum rate , since it quantifies the effectiveness of an algorithm and simplifies scheduling rules .
however , maximizing the sum rate in scenarios where the users have heterogeneous snrs might be very inefficient , since users with poor channel conditions would experience starvation @xcite .
a classical solution to such a problem is assigning the weights @xmath161 , according to a fairness criterion .
the works listed under the _ fairness _ category in table [ tab : optimization - criteria ] , optimize the long - term proportional fairness , extending the original idea from @xcite to mu - mimo scenarios .
round robin is a channel - unaware scheduling method , that allocates the same amount of time to all competing users .
it is the simplest form of fair resource - access , but neglects mudiv , the amount of occupied resources , and qos requirements @xcite .
equal opportunity for resource sharing may not mean equal distribution of resources , which results in wastage , redundant allocation , or resource starvation @xcite .
nonetheless , round robin might be useful in ultra dense heterogeneous networks , where transitions from non - los to strong los channels alter the performance of conventional resource allocation strategies @xcite . the works listed under the round robin category in table [ tab : optimization - criteria ] ,
use the method as a benchmark to assess mudiv gains in mu - mimo systems .
it is worth mentioning that in the reviewed literature , the system performance is assessed over a large number of channel realizations .
for example , using the average sum rate per channel use , also called spectral efficiency , measured in bit / s / hz .
other common performance metrics are @xcite : the _ throughput _ or average rates achieved for packet transmission with practical modulations and realistic coding schemes ; and the _ goodput _ , which is the total average rate successfully delivered to the scheduled users , including layer 2 overheads and packet retransmission due to physical errors @xcite .
the wsr optimization problem can incorporate different types of qos requirements .
the qos defines a network- or user - based objective [ cf . definition [ defn : quality - of - service ] ] , and the weights @xmath161 , establish priorities according to the service provided to each user .
taking into account qos requirements reduces the flexibility of the resource allocation strategies .
for instance , in the case of opportunistic transmission , achieving a target qos reduces the value of csit , i.e. , the channel conditions have to be weighted by requirements coming from the upper layers @xcite .
system optimization considering qos has a broader scope and once information from upper layers is considered , a more complex cross - layer optimization problem arises .
the qos is defined according to the system model , which might include individual traffic characteristics or sinr requirements .
satisfying qos requirements can be realized at different time granularities : each user must achieve a prescribed performance metric , e.g. instantaneous sinr or data rate @xcite ; or a network performance metric must reach a stable behavior over time , e.g. attaining finite buffer sizes @xcite .
below , we provide a classification of the optimization criteria subject to qos constraints in mu - mimo scenarios .
the qos can be guaranteed as long as individual sinr requirements are met , which also satisfies target error and peak rates .
the qos can be defined by a monotonic and bijective function of the sinr : one example is the bit error rate ( ber ) , given a particular modulation and coding scheme ( mcs ) @xcite ; another example is the shannon capacity @xcite .
different tradeoffs between global and individual performance are achieved under sinr constraints : maximizing the wsr @xcite , max - min weighted sinr @xcite , sum power minimization @xcite , and other hybrid formulations , see @xcite and references therein .
it is worth noticing that in systems with sinr requirements , it is usually assumed that for each channel realization , a set of compatible users @xmath46 has been previously selected .
the system optimization is realized by power allocation and precoder design @xcite .
this means that @xmath46 must be a feasible set of users [ cf .
definition [ defn : feasible_set_of_users ] ] , and user scheduling is required only if the constraints associated with @xmath46 are infeasible @xcite . in that case
, it is necessary to relax the initial conditions taking one of the following actions : 1 . reducing the number of selected users applying _ admission control _
@xcite or _ user removal _
schemes @xcite .
these procedures are implemented in systems where the achievable sinrs take any positive real value , e.g. @xcite .
the removal process can be thought as a form of scheduling , used to generate a feasible set of users [ cf . definition [ defn : feasible_set_of_users ] ] . by identifying users with unfeasible constraints
, one can drop them temporarily ( according to priorities ) , or re - schedule them over orthogonal resources ( e.g. other carriers or time slots ) @xcite .
2 . by relaxing the individual qos requirements @xmath150 , one can achieve resource allocation feasibility [ cf . definition [ defn : resource_feasibility ] ] .
for instance , consider homogeneous target rates , then all scheduled users can be balanced to a common sinr @xcite . for heterogeneous qos , one can iterate over different sinr levels per user ( e.g. provided a set of mcss ) , so that link adaptation can take place , see @xcite .
cross - layer algorithms subject to rate constraints can be designed so that @xmath46 and its associated resources are jointly optimized .
analytical results in @xcite show that under mild conditions , there exists , with probability one , a feasible set of users @xmath46 that can be found though low - complexity algorithms [ cf .
section [ section : algorithms : aggregated - utility ] ] .
the transmission errors ( bit , frame , symbol , packet ) are usually due to noisy channels and inaccurate csit . for the former factor
, the errors occur due to the effect of non - ideal channel coding and finite blocklength channel coding .
the error rates can be reduced by implementing stronger channel codes and longer blocklengths .
the error rates also occur due to csi quantization ( limited feedback ) , estimation errors , delays ( channel outage ) , and distortion during the feedback @xcite .
therefore , iui can not be fully suppressed by precoding techniques [ cf . definition [ defn : interference - limited - system ] ] , which degrades the achievable sinrs and ber figures .
nonetheless , multi - antenna receivers can implement efficient processing techniques to increase their sinrs and combat quantization errors , e.g. , quantization - based combining ( qbc ) @xcite .
under partial csit conditions , any practical rate adaptation scheme achieves performance between the worst - case outage rate and the ideal - case , where the achievable rate equals the mutual information @xcite .
a network - based optimization requires cross - layer algorithms taking as inputs the queue state information ( qsi ) , csit , and their inter - dependence @xcite .
there exist a rich literature on optimization subject to queue or packet delay constraints , specially for systems with orthogonal resource allocation ( e.g. one user per time - slot or sub - carrier ) , see @xcite for an in - depth overview .
according to the qsi relevance , mu - mimo systems fall in two categories
@xcite : 1 . _
systems with infinite backlogs _ : in this scenario , the transmitter is assumed to always have data to send and infinite buffer capacity . in other words , the transmitter fully knows the intended data for each user .
the objective of the resource allocation strategies is , in general , to maximize the wsr [ cf .
section [ section : performance - uf : weighted - sum - rate ] ] .
the weight @xmath157 of the @xmath23-th user is used to reach a desired throughput - fairness tradeoff , instead of expressing the urgency of data flows .
thus , it can be assumed that the queues are balanced and users can not be discriminated based on them @xcite .
the service provided is delay - insensitive , and the traffic is managed so that the scheduled users attain non - zero rates @xcite .
mu - mimo systems without qsi information fall into this category .
systems with bursty traffic and limited buffer size _ : in these systems , packet data models with stochastic traffic arrivals are considered .
the performance assessment is analyzed from the delaying and queuing perspectives .
the wsr maximization attempts to balance between opportunistic channel access and urgency of data flows .
the resource allocation algorithm must guarantee finite average buffer occupancy , i.e. , _ stability of the queues lengths _ for all users @xcite .
it is worth mentioning that by little s theorem @xcite , achieving stability in the average queues is equivalent to minimize the average packet delay in the steady state . the system model might consider different sources of delay , e.g. , buffer congestion or destination unavailability ( outage delay ) , which define the type of scheduling policy to be implemented @xcite .
the choice of the system model and problem formulation depend on the desired tradeoff between tractability and accuracy in each particular scenario @xcite . in mu - mimo systems where queue stability is optimized
, the scheduling algorithm has to guarantee that the average queue lengths of all users are bounded .
simultaneously , the available csit should be opportunistically exploited for throughput maximization @xcite .
this goal can be achieved by establishing queue - based wsr as the optimization criterion .
the weights @xmath157 in ( [ eq : weighted - sum - rate - function ] ) can be defined according to the system requirements : considering the packet queue length or packet departure rates at each scheduling interval @xcite ; considering fairness with heterogeneous traffic rates @xcite ; or considering service - oriented requirements , e.g. , ber , delay tolerance , and packet dropping ratio for real- or non - real - time services @xcite .
the queue lengths not only define the priority or urgency of the traffic , but they can also define the encoding order , @xmath36 in ( [ eq : general - combinatorial - problem : cost - function ] ) , see @xcite . in general , queue lengths
are not symmetric or balanced , which means that the wsr optimization is dominated by the qsi , rather than by the csit @xcite .
the characteristics of the mu - mimo system model and the set of constraints define the resource allocation strategy that achieves stability .
several factors may affect the performance under qos constraints : csit accuracy @xcite , csit statistics @xcite , and the number of resources used for channel estimation @xcite ; the spatial compatibility between co - scheduled user and the snr regime @xcite ; the user priority based on the type of service @xcite ; or the number of simultaneous spatial streams @xcite .
the overall optimization can also consider delays due to packet losses and retransmissions .
this is supported by ack / nak ( handshake ) signaling exchange in the upper layers .
such a protocol is used for automatic repeat request ( arq ) , to convey an error - free logical channel to the application layers @xcite .
works in @xcite and references therein , discuss algorithms to solve the wsr with queue constraints taking into account arq protocols .
we refer the reader to @xcite for a comprehensive texts on cross - layer design under queue and delay constraints .
* 6>x * method * & * utility - based * & * csi - mapping * & * metaheuristic(stochastic ) * & * classicoptimization * & * exhaustivesearch * + * miso * , @xmath146 & @xcite & @xcite & @xcite & @xcite & @xcite + * miso * , @xmath147 & @xcite & @xcite & @xcite & - & @xcite + * mimo * , @xmath146 & @xcite & @xcite & @xcite & @xcite & @xcite + * mimo * , @xmath147 & @xcite & @xcite & - & @xcite & - + resource allocation can be performed over orthogonal radio resources @xcite , where the allocation decisions are made using individual estimates of ber , sinr , or rate . scheduling the best user per resource
is performed by low - complexity algorithms that do not depend on full csit .
for instance , a sorting - based algorithm at the transmitter can schedule users based on the cqi .
in contrast , the scheduling algorithms for mu - mimo systems attempt to allocate resources in a non - orthogonal fashion .
the users can not compute their achievable ber or rates since those metrics depend on the csit , powers , and precoders assigned to other co - scheduled users . to illustrate the scheduling in problem ( [ eq : general - combinatorial - problem : cost - function ] ) , consider a mu - mimo system with a single transmitter equipped with @xmath18 antennas , and @xmath21 competing users , each equipped with @xmath19 antennas . assuming that linear precoding is implemented [ cf .
section [ section : precoding : linear ] ] , the maximum number of co - scheduled users per resource is given by @xmath162 .
define @xmath163 as the _ search space _ set containing all possible user groups with limited cardinality .
contains the sets of cardinality one , i.e. , su - mimo configuration .
efficient scheduling rules must dynamically switch between su- and mu - mimo configurations @xcite . ]
the optimal set , @xmath164 , that solves ( [ eq : general - combinatorial - problem : cost - function ] ) lies in a search space of size @xmath165 .
the set @xmath164 can be found by brute - force exhaustive search over @xmath166 , which is computationally prohibitive if @xmath167 .
the search space , and in turn the complexity of the problem , varies according to the system model and constraints : the values of @xmath18 and @xmath19 , the number of streams per user , individual qos or power constraints , the set of mcss , etc . + by plugging the utility function @xmath148 , [ cf .
section [ section : performance - uf ] ] , into ( [ eq : general - combinatorial - problem : cost - function ] ) , we obtain a mathematical formulation of the mu - mimo resource allocation problem . the optimal scheduling rule that solves ( [ eq : general - combinatorial - problem : cost - function ] ) in a feasible and efficient way is still an open problem .
a number of algorithms have been proposed in the literature to circumvent the high complexity of finding @xmath164 , and its associated resource allocation .
a sub - discipline of optimization theory , known as heuristic search @xcite , provides solutions ( approximation algorithms ) to a category of discrete problems .
such problems can not be solved any other way or whose solutions take a very long time to be computed ( e.g. np - hard problems ) .
user scheduling falls in such a problem category , and the common approach in the literature is to design suboptimal heuristic algorithms that balance complexity and performance .
the optimal solution , @xmath164 , is usually computed for benchmarking in simple mu - mimo scenarios . in this section ,
we provide a classification of the most common scheduling algorithms in the literature of mu - mimo .
table [ tab : algorithms ] and fig .
[ fig : algorithm_calssification ] show the scheduling algorithms classification , the scenarios in which they are applied , and the methodology they follow .
most algorithms operate at the single resource level , i.e. , considering optimization over one sub - carrier or time - slot .
extensions to scenarios with multiple carriers or codes are straightforward , e.g. , @xcite .
let us reformulate ( [ eq : general - combinatorial - problem : cost - function ] ) as a multiuser wsr maximization problem as follows @xmath168 where @xmath169 and @xmath170 summarize the precoding weights and the powers assigned to the users in @xmath46 , and @xmath171 defines an encoding order .
the formulation in ( [ eq : inner - outer - problem ] ) illustrates the fact that the resource allocation problem can be decomposed into different subproblems , which are related to different network layers .
furthermore , under specific system settings and constraints the subproblems can be decoupled .
the weights @xmath157 , @xmath150 , can be defined according to the service or application delivered to the users , [ cf .
section [ section : performance - uf : wsr - with - qos ] ] . the inner problem in ( [ eq : inner - outer - problem ] )
attempts to maximize the wsr for a given set @xmath46 .
finding a solution requires joint admission control , precoding design , and power control , e.g. @xcite .
a solution to the inner problem exists when the multiuser channel associated with @xmath46 , has specific spatial dimension defined by the precoding scheme .
the channel dimensions must be guaranteed by the solution of the outer problem .
if linear precoding schemes are implemented , [ cf .
section [ section : precoding : linear ] ] , the original problem can be relaxed : the cardinality of @xmath46 is upper bounded by @xmath172 ; the encoding order @xmath171 can be omitted since it does not affect the precoding performance ; and power allocation can be computed via convex optimization or by heuristic algorithms [ cf . section [ section : power - allocation ] ] . finding a solution to the combinatorial outer problem implies
that the scheduler accomplishes several goals : it exploits mudiv to maximize multiplexing gains ; and it finds the multiuser channel with the best set of spatially compatible users [ cf .
definition [ defn : compatible_set_of_users ] ] .
problem ( [ eq : inner - outer - problem ] ) might include constraints over several discrete parameter sets , such as @xmath46 , the encoding orders , mcss , spatial streams per user , etc .
the complexity of the combinatorial problem grows exponentially with each discrete set size @xcite . due to the hardness of problem ( [ eq : inner - outer - problem ] ) , one common approach is to find suboptimal solutions through heuristic _
greedy opportunistic algorithms_. given a fixed precoding structure and csit , the set @xmath46 is constructed by a sequence of decisions .
every newly selected user finds a locally maximum for an objective function .
the optimized utility @xmath148 can be defined by the wsr ( direct approach ) or by a metric of spatial compatibility ( indirect approach ) .
the greedy scheduling rule requires low computational complexity and is easy to implement .
however , it does not guarantee neither performance nor convergence to the optimal solution @xcite .
the seminal works @xcite proposed a simple construction of @xmath46 , by iteratively selecting users as long as the aggregated objective function improves its value . in @xcite , a greedy user selection ( gus ) based on sum - rate maximization was proposed .
first , the user with the highest rate is selected , and the following selected user will maximize the total sum - rate . this method was improved in @xcite , using sequential water - filling to eliminate users with zero transmit power after selection and power allocation .
the algorithms in @xcite might be computational demanding because calculation of the precoders and power allocation is required each iteration .
furthermore , in the case of imperfect csi , the above algorithms fail to determine the optimal number of selected users . a generic structure of this kind of user selection
is presented in algorithm [ alg : algorithms : direct - approach ] . in this approach
, the transmitter must have channel information so that every iteration the precoders @xmath173 , @xmath150 , are recalculated to solve the inner problem in ( [ eq : inner - outer - problem ] ) . if linear precoder schemes are implemented , the maximum number of iterations is at most @xmath172 .
since the transmitter knows the precoders and the transmit powers , the achievable rates , sinr , or bers are known each iteration , allowing the scheduler to identify when it is worth adding more users to @xmath46 .
this approach is particularly effective if there are no constraints on the cardinality of @xmath46 , or if the network is sparse , i.e. , @xmath174 @xcite .
for instance , if zf - based precoding is used , the multiplexing gain is not maximized at the low snr regime , i.e. , the total number of scheduled users must be strictly less than @xmath172 . in the high snr regime
, multiplexing gains can be maximized if full csit is available @xcite , otherwise the system may become interference limited [ cf . definition [ defn : interference - limited - system ] ] .
several works have modified the structure of algorithm [ alg : algorithms : direct - approach ] to reduce complexity and improve performance , see @xcite .
the works @xcite showed that the solution space can be shrunk at each iteration , reducing the computational complexity . based on the characteristics of the powers that solve the inner problem in ( [ eq : inner - outer - problem ] ) , the scheduler defines a reduced set of candidate users , @xmath175 , for the next iteration .
nonetheless , such methods also exhibit the flaw of not being able to identify the optimal cardinality , @xmath43 . as such , there are redundant users in the selected user subset , i.e. , users that can be deleted from the selected user subset to yield a performance increase .
this is an inherent flaw of any greedy incremental algorithm , due to the non - iterative cumulative user selection procedure . in @xcite , it is proposed to add the _ delete _ and _ swap _ operations as means to tackle the redundant user issue .
however , this approach increases the complexity , even as compared to gus , as it involves matrix inversions , projections , and an iterative procedure .
moreover , its performance is sensitive to csi inaccuracies due to the increased error in the projection operation using imperfect csi , and it does not necessarily find the best user subset in the imperfect csi case .
set @xmath176 , select the user @xmath23 with the strongest channel , update @xmath177 , calculate precoders and powers , compute the achievable rate @xmath178 in ( [ eq : inner - outer - problem ] ) .
find the user @xmath179 that maximizes @xmath180 .
if @xmath181 and @xmath182 , then update @xmath183 , modify @xmath175 , and go back to step 2 .
+ otherwise go to step 4 .
compute final precoders , powers , and wsr for @xmath46 . in this approach , the inner and outer problems in ( [ eq : inner - outer - problem ] ) are decoupled . to illustrate the structure of the resource allocation strategy ,
let us reformulate ( [ eq : inner - outer - problem ] ) as two problems that must be solved sequentially : ^map = f ( ( ) ) & [ eq : csi - mapping - problem-1 ] + _ k ^map _ k _ 2 ( 1 + _ k ) & [ eq : csi - mapping - problem-2 ] equation ( [ eq : csi - mapping - problem-1 ] ) describes a combinatorial user grouping problem , where @xmath112 is a metric of spatial compatibility , [ cf .
section [ section : metrics - spatial - compatibility ] ] .
the user grouping is an np - c problem , whose solution is found via exs @xcite . observe that finding a solution to problem ( [ eq : csi - mapping - problem-1 ] ) does not require the computation of the precoders and powers .
instead , it depends on csit , algebraic operations defined by a given mapping function @xmath112 , and an iterative procedure .
it is worth noting that ( [ eq : csi - mapping - problem-1 ] ) may include the weights @xmath157 to cope with qos or fairness requirements , see @xcite .
in contrast , provided the set @xmath184 , problem ( [ eq : csi - mapping - problem-2 ] ) is tractable and can be solved efficiently via convex optimization @xcite .
early works on mu - mimo systems ( e.g. @xcite ) pointed out that under certain conditions of the weights , @xmath157 , and snr regime , the objective function of the inner problem in ( [ eq : inner - outer - problem ] ) , is dominated by the geometry of the multiuser channel @xmath98 .
the rational behind the reformulation in ( [ eq : csi - mapping - problem-1])-([eq : csi - mapping - problem-2 ] ) , is that scheduling a set of spatially compatible users [ cf .
definition [ defn : compatible_set_of_users ] ] , is crucial to suppress iui .
refining the power allocation and precoding weights in ( [ eq : csi - mapping - problem-2 ] ) requires less computational effort than solving the combinatorial problem ( [ eq : csi - mapping - problem-1 ] ) .
a generic structure of this scheduling algorithms is presented in algorithm [ alg : algorithms : indirect - approach ] .
the complexity of finding a solution to problem ( [ eq : csi - mapping - problem-1 ] ) can be simplified under certain conditions .
assuming that zf - based precoding schemes are used , results in @xcite establish that the optimal admitted set of users achieves full multiplexing gains , i.e. , @xmath185 , if the system operates in the high snr regime .
therefore , the cardinality constraint in ( [ eq : csi - mapping - problem-1 ] ) is given by @xmath186 , which shrinks the search space size .
a common approach in the literature is to assume that @xmath172 users can be co - scheduled when solving problem ( [ eq : csi - mapping - problem-1 ] ) , and refining the set @xmath46 when solving problem ( [ eq : csi - mapping - problem-2 ] ) . the majority of the works that have proposed algorithms to solve problem ( [ eq : csi - mapping - problem-1 ] ) , e.g. @xcite , use zf precoding for the second problem ( [ eq : csi - mapping - problem-2 ] ) .
consequently , the nsp is the metric of spatial compatibility that maximizes the sum - rate for the set of selected users @xmath46 , [ cf .
section [ section : metrics - spatial - compatibility : projection - spaces ] ] .
a considerable amount of research has been focused on reducing the complexity of the nsp computation , e.g. , by reusing previous nsp calculations at each iteration , implementing efficient multiuser channel decompositions , or computing approximations of the nsp .
the methods highly depend on @xmath18 , @xmath19 , the csit accuracy , and the system model .
different works have proposed a search space reduction per iteration , recalculating and reducing the number of competing users so that spatial compatibility is preserved .
a common approach is to preselect a group of candidate users , @xmath175 in algorithm [ alg : algorithms : indirect - approach ] , based on spatial clustering [ cf .
section [ section : metrics - spatial - compatibility : spatial - clustering ] ] .
the candidate users at the next iteration will exhibit channel directions fulfilling an @xmath121-orthogonality criterion .
this approach , coined as semi - orthogonal user selection ( sus ) , was originally proposed by yoo and goldsmith @xcite .
several variations of this technique can be found in the literature , see @xcite .
once that @xmath46 has been constructed , it may be necessary to modify it , so that problem ( [ eq : csi - mapping - problem-2 ] ) yields the maximum wsr . a common approach to refine @xmath46 ,
is to apply user removal techniques [ cf .
section [ section : performance - uf : wsr - with - qos ] ] .
this might be necessary if : the selected channels lacks spatial compatibility ; the system operates in extreme snr regimes ; or if the constraints related to @xmath46 turn problem ( [ eq : csi - mapping - problem-2 ] ) infeasible @xcite .
if linear precoding and water - filling are used to solve ( [ eq : csi - mapping - problem-2 ] ) , the set @xmath46 can be refined by dropping users that do not achieve a target rate or whose allocated power is zero .
authors in @xcite pointed out that regardless of the number of deployed antennas , it is more efficient to schedule users so that @xmath187 .
this means that the scheduler must seek a tradeoff between maximizing spatial multiplexing gains , and optimizing the wsr for a small set of spatially compatible users . to solve this cardinality problem , some algorithms , e.g. @xcite , maximize the wsr by comparing the solution of ( [ eq : csi - mapping - problem-2 ] ) for multiple sets that solve ( [ eq : csi - mapping - problem-1 ] )
this is performed by sequentially constructing groups of spatially compatible users with different cardinalities , and then selecting the best group based on the achievable wsr . for the general heterogeneous mu - mimo scenario ,
@xmath188 , @xmath32 , the function @xmath112 in problem ( [ eq : csi - mapping - problem-1 ] ) can be defined by a metric of subspace compatibility [ cf .
section [ section : metrics - spatial - compatibility : other - metrics ] ] .
the user grouping procedure using such a metric can operate in a greedy fashion , as in algorithm [ alg : algorithms : indirect - approach ] , but every scheduler design depends on the system model .
it is worth pointing out that in heterogeneous systems , maximizing spatial compatibility is not enough to guarantee wsr maximization .
multidimensional metrics measure the potentially achievable spatial separability but not the effective channel gain per spatial dimension , [ cf . definition [ defn : effective_channel_gains ] ] .
efficient user selection algorithms in scenarios with heterogeneous users should not only consider spatial compatibility , but also an estimation of the achievable wsr , sinrs , or effective channel gains @xcite .
moreover , dynamic allocation of the number of data streams per user must be included in the scheduler , to take into account csit accuracy and fulfill qos constraints .
users with small signal spaces experience less iui , and results in @xcite show that under certain conditions , transforming a mimo channel into an equivalent miso channel can improve spatial separability at the expense of multiplexing gain .
the reliability of scheduling based on metrics of spatial compatibility is highly sensitive to csit accuracy .
most of the works in the literature ( e.g. @xcite ) , perform scheduling by evaluating a form of spatial correlation between mimo channels of different user .
however , the correlation between antennas at each receiver is usually not taken into account .
such a correlation may affect the achievable sinr , depending on the signal processing at the receiver , e.g. , receive zf processing @xcite , or receive combining @xcite .
the inner correlation of the channel @xmath33 of the @xmath23-th user can be measured by its rank or by the magnitude ( dispersion ) of its eigenvalues @xcite .
transmitting multiple spatial streams per user can not be reliable , specially for high inner correlation . a common approach to simplify the scheduling designs
is by limiting the number of streams per user , e.g. , using only the strongest eigenvector per selected user @xcite . for each channel @xmath33
, there exists a dominant spatial direction that can be selected for transmission without creating severe iui , or performance degradation after precoding . in this way ,
problem ( [ eq : csi - mapping - problem-1 ] ) can be solved based on antenna or eigen - direction selection , e.g. , @xcite .
set @xmath176 , select the user @xmath23 with the strongest channel , update @xmath177 . find the user @xmath179 that maximizes @xmath189 .
if @xmath181 , then update @xmath183 , preselect users in @xmath175 , and go back to step 2 .
+ otherwise go to step 4 .
perform operations to refine @xmath46 .
compute precoders and powers to solve ( [ eq : csi - mapping - problem-2 ] ) for @xmath46 .
several works in the literature address problem ( [ eq : inner - outer - problem ] ) using stochastic optimization methods , which find close to optimal solutions to complex and non - convex mixed problems .
these non - traditional or metaheuristic methods are an alternative to classical programming techniques @xcite .
the mathematical proof of convergence to the optimal solution can not be demonstrated , and analytical results or performance analysis can not be derived .
however , these methods may find a close to optimal solution with high probability , without the need of computing derivatives or satisfying convexity requirements ( as in standard optimization techniques ) .
the principle behind metaheuristic methods is the following : a set of feasible solutions is iteratively combined and modified until some convergence criterion is met .
the term stochastic comes from the fact that initial conditions are generally chosen randomly . for each iteration ,
the internal parameters change according to dynamic probability functions or even randomly .
the algorithms designed under this approach are based on certain characteristics and behavior of natural phenomena , e.g. , swarm of insects , genetics , biological or molecular systems @xcite . in the context of mu - mimo systems
, two techniques have been used to solve the user scheduling problem : genetic algorithms ( ga ) and particle swarm optimization ( pso ) algorithms .
ga are bio - inspired or evolutionary algorithms , that can solve multiple objective optimization problems with many mixed continuous - discrete variables , and poorly behaved non - convex solution spaces .
unlikely standard optimization methods that rely on a single starting point ( e.g. algorithms sensitive to initial conditions @xcite ) , ga start with a set of points ( population ) .
these points increase the argument domain of the optimized function and avoid local optimal problems due to the evaluation of the solution space @xcite .
ga encode potential solutions to the optimization problem in data structures called chromosomes .
due to the binary nature of the variables @xmath47 in problem ( [ eq : general - combinatorial - problem : cost - function ] ) , the chromosomes can be represented as binary strings containing valid configurations of the variables @xmath47 , [ cf .
( [ eq : general - combinatorial - problem : set - cardinality ] ) ] .
a set of chromosomes is referred to as a population , and at each iteration , the chromosomes combine , exchange critical information , mutate , and eventually evolve toward the optimal solution @xcite .
practical application of ga to solve the wsr maximization problem ( [ eq : inner - outer - problem ] ) have been proposed for different precoding schemes : zfbf @xcite , zfdp @xcite , and dpc @xcite . recall that given a set @xmath46 , the performance of dpc is sensitive to a given user order @xmath171 selected out of @xmath190 valid orders .
the works @xcite have proposed heuristic algorithms to define @xmath171 .
since the chromosome structure contains different optimization variables , it can include not only the set of selected users but can also the order in which the users are encoded for zfdp @xcite and dpc @xcite .
pso are behavior - inspired algorithms that mimic the distributed behavior of social organisms ( particles ) .
the particles use their own intelligence ( based on an individual utility function ) and the swarm intelligence ( global utility function ) , to discover good directions that can be followed and verified by the swarm .
pso initially defines the swarm as a set of particles randomly sampled from the search space .
each particle has two associated parameters that define how fast it can move toward the optimal solution : position and velocity . in this approach , each particle wanders around in the search space to collect information about the achievable utility function at different locations .
this means that the particle looks for the configuration that steers toward the global maximum or the average direction of the swarm .
the particles exchange information regarding the best directions , and adjust their positions and velocities accordingly until the swarm converge to the same solution @xcite .
authors in @xcite designed pso algorithms to solve problem ( [ eq : inner - outer - problem ] ) using bd precoding .
such an optimization uses the utility function , @xmath148 , of the direct and indirect approaches described in section [ section : algorithms : aggregated - utility ] .
the pso approach can define the swarm direction using the function @xmath178 in the inner problem ( [ eq : inner - outer - problem ] ) , or the mapping function @xmath112 in ( [ eq : csi - mapping - problem-1 ] ) .
this illustrates the fact that the computational complexity of metaheuristic algorithms depends on the optimized objective function .
algorithm [ alg : algorithms : metaheuristic ] shows a sketch of the steps used to solve problem ( [ eq : inner - outer - problem ] ) following the ga or pso approaches .
the constants @xmath191 and @xmath192 are arbitrarily defined to limit the overall computational complexity and to speed up convergence .
the convergence criterion can be defined by the maximum number of iterations or a performance threshold .
the details of the operations required to generate new chromosomes or particles , [ step 3 in algorithm [ alg : algorithms : metaheuristic ] ] , can be found in the reference aforementioned .
the robustness of ga and pso lies in the fact that the best solutions are systematically combined , and the algorithms have reasonable immunity from getting stuck in local minimums .
metaheuristic algorithms are efficient techniques for problems where the desired solution does not need to be computed in short time scales .
although bio - inspired algorithms have been used to optimize resource allocation in wireless networks , e.g. @xcite , they seem to have limited application in practical mu - mimo systems . this is because the channel conditions , rate demands , and even the number of competing users may change very rapidly over time .
metaheuristic algorithms require centralized processing and the convergence time remains a critical issue @xcite .
they might suffer from a high computational cost because the objective function is evaluated for each member of the population ( swarm ) .
nevertheless , these methods can provide performance benchmarks to assess other suboptimal , yet practical , resource allocation algorithms . for the @xmath193 iteration , extract @xmath191 elements ( chromosomes or particles ) from the search space @xmath194 .
for the @xmath117-th iteration , evaluate the objective function for each element in @xmath195 , select the @xmath192 elements with the largest objective function , and discard the other bad elements . perform operations over the remaining elements to generate an improved population / swarm , @xmath196 , for iteration @xmath197 . if convergence criterion is met , then compute precoders and powers for the solution set .
+ otherwise go to step 2 .
classical optimization techniques @xcite , have been used for several resource allocation problems in mu - mimo systems , e.g. , wsr maximization , sum power minimization , max - min sinr , and other objective functions , see @xcite and references therein .
numerous works have included user scheduling for the optimization of the wsr @xcite , sum power minimization @xcite , or some metrics of spatial compatibility @xcite .
a large number of optimization problems in mu - mimo systems are non - convex or non - polynomial time solvable , depending on the system models and constraints @xcite . due to the high complexity of the resource allocation problem
, the conventional optimization solutions must relax the original problem , approximate non - convex constraints in an iterative fashion , or change the domain of the optimization , sacrificing optimality for the sake of tractability .
nevertheless , by employing a fixed structure for the precoders , the original problem can be simplified , and efficient resource allocation can be performed .
authors in @xcite presented different approaches to solve ( [ eq : general - combinatorial - problem : cost - function ] ) in a ofdma mu - mimo scenario .
the mathematical formulation renders the combinatorial problem into a convex problem . by analyzing the entire search space , the proposed algorithms perform channel assignment and signal space selection ,
i.e. , the algorithm defines the scheduled users per carrier and the number of spatial streams per user .
such a reformulation of ( [ eq : general - combinatorial - problem : cost - function ] ) yields algorithms that decouple the channel assignment in the ofdma system , i.e. , the global optimization is attained by solving problem ( [ eq : inner - outer - problem ] ) per sub - carrier .
furthermore , the objective function of the inner problem in ( [ eq : inner - outer - problem ] ) is defined in terms of rates and powers , which optimizes the total capacity and power consumption . for classical programming techniques ,
the set of constraints might turn the problem infeasible . the resource allocation algorithms proposed in @xcite identify the assignment sets for which time sharing is the best resource allocation strategy . as discussed in section [
section : performance - uf : wsr - with - qos ] , admission control or user removal techniques can be implemented as a form of scheduling to guarantee feasibility .
authors in @xcite have reformulated the user scheduling problem in multi - carrier scenarios as an extension of problems ( [ eq : csi - mapping - problem-1 ] ) and ( [ eq : csi - mapping - problem-2 ] ) . a close to optimal solution
can be found by a sequence of convex and linear programming optimizations . to reduce complexity of the combinatorial problem ( [ eq : csi - mapping - problem-1 ] ) , the mapping function @xmath112 is defined as the magnitude of the mimo channels , and a close to optimal scheduling sequence is designed based on that metric .
according to @xcite , the generalization of ( [ eq : csi - mapping - problem-2 ] ) for ofdma systems yields a non - convex problem , but quasi - optimal solutions can be attained by applying the classical dual decomposition method .
the works @xcite have modeled the user scheduling problem in ofdma systems as a channel assignment problem , which can be solved by efficient algorithms that run in strong polynomial time @xcite . a novel reformulation of the wsr maximization problem , ( [ eq : inner - outer - problem ] ) , as a semi - definite program in @xcite , sheds some light on non - explicit convex properties of the joint precoding design , power allocation , and users selection problem .
the reformulation requires semi - define relaxation , and the solution is found by combining convex optimization and sub - gradient projection methods .
another approach in @xcite extends the max - min fair rate allocation problem in @xcite so that joint optimization of precoding , power allocation , and user selection was attained .
the problem was formulated as the difference of convex functions and solved by the branch and bound ( bb ) technique , which is a method used to solve mixed - integer programming problems @xcite .
the approach in @xcite yields an interference limited systems [ cf . definition [ defn : interference - limited - system ] ] , i.e. , it is allowed that @xmath198 , which highly increases the search space and computational complexity . notice that even if the search space is constrained so that @xmath199 , there are @xmath200 possible user schedules in a miso configuration that will be enumerated by the bb technique . as discussed in section [ section : performance - uf ] , admission control is a form of user selection that may be required to guarantee feasibility . in @xcite , the joint sum power minimization and maximization of @xmath43
was formulated as a second - order cone program ( socp ) .
the set @xmath46 is refined iteratively by dropping the user with the largest gap to its target sinr , ( i.e. the most infeasible user ) , and optimal precoders and powers are found by the socp optimization .
although the approaches in @xcite provide suboptimal solutions due to the relaxation of non - convex problems , they show that joint optimization of the three variable sets @xmath46 , @xmath170 , and @xmath169 in ( [ eq : inner - outer - problem ] ) , is feasible through convex optimization .
other classical optimization methods have been used in the single - carrier miso scenario with zf precoding .
quadratic optimization was used in @xcite to solve problem ( [ eq : csi - mapping - problem-1 ] ) , by optimizing a heuristic function @xmath112 that linearly combines channel magnitudes and spatial correlation of the mimo channels . in @xcite , a heuristic objective function that approximates
the nsp has been proposed , and the set @xmath46 was found via integer programming .
the authors in @xcite jointly addressed scheduling and wsr maximization subject to qos constraints using geometric programming .
however , similar to the approach in @xcite , the optimization in @xcite iterates over all solutions of the search space , which limits its application for practical scenarios .
other works model the resource allocation problem ( [ eq : csi - mapping - problem-1 ] ) , as mathematical problems that have been extensively studied @xcite : the minimum set cover problem from graph theory ( e.g. , @xcite ) ; and the sum assignment problem ( e.g. , @xcite ) .
the modeling used in these approaches requires the calculation of weights or costs values associated with every possible subset in the solution space , i.e. , @xmath201 , where @xmath202 depends on the system parameters .
the solutions of such general problems can be found by heuristic algorithms . due to the exponentially increasing complexity on @xmath18 , @xmath19 , and @xmath21 ,
the algorithms based on classical optimization are suitable for offline implementation , benchmarking , and to assess other suboptimal heuristic resource allocation strategies for relatively small values of @xmath21 @xcite . in section [ section : system - set - ups :
multi - cell - scenarios ] , we have described the types of transmission schemes in multi - cellular scenarios .
different levels of coordination are required to improve and maintain the performance of users geographically spread within the coverage area . for users located at the cell edge
, the simultaneous transmission from different clustered bss can boost their throughputs ( e.g. comp @xcite ) , whereas users close to their bs do not require interaction with adjacent bss . in full signal - level coordinated systems ( e.g. , network mimo or jt ) , the global user data and csi knowledge enables a cu to perform cs using the approaches described in section [ section : algorithms : aggregated - utility ] and section [ section : algorithms : selection - partial - csit ] . assuming partial coordination ,
the cs / cbf can offer flexibility , scalability , and viability , since centralized or distributed algorithms can be implemented with limited signaling overhead and using local csit .
the minimum level of cooperation between transmitters can be attained through orthogonal resource allocation .
for instance , in lte systems a technique known as _ dynamic point blanking _ , prevents transmission at a certain time - frequency resource to reduce interference over such a resource used at a neighboring transmission point @xcite .
the optimization of cs and cbf can be done based on the methodologies described in section [ section : algorithms : aggregated - utility ] , i.e. , the two problems can be addressed either independently or jointly .
if cs and cbf are decoupled , cs can be tackled by processing shared information related to statistical csi or performance metrics extracted from local csi [ cf .
section [ section : metrics - spatial - compatibility ] ] .
the cbf optimization can be solved by optimal or heuristic techniques for precoding design and power allocation , see @xcite for a comprehensive review .
a joint optimization of cs and cbf requires iterative calculation of the precoders and selected users , since both optimization spaces are coupled @xcite .
efficient scheduling rules must determine the set of users that maximizes the performance , and generates less interference to neighboring cells .
the following classification presents some existing methods to implement cs and cbf in multi - cell networks @xcite . in full coordinated multi - cellular systems ,
a cluster of bss defines a super - cell or virtual distributed antenna system with per - bs power constraints .
full coordination allows different transmission strategies : _ i _ ) all clustered bss send data to all the users in @xmath46 ; _ ii _ ) some bss jointly serve a subset of @xmath46 ; _ iii _ ) each bs serves a set of users associated to it ; or _ iv _ ) each bs performs single transmission ( st ) and serves only one user per scheduling interval , which is known as the interference channel model ( ifc ) , see @xcite .
the fist and second strategies are exclusive of fully coordinated systems since payload data is shared among bss .
the other strategies are also implemented for distributed cs / cbf .
recent works have extended these transmission schemes to heterogeneous networks , e.g. , @xcite .
accounting for global csit provides flexibility to the cs / cbf design , and several ici cancellation techniques can be applied . using classical optimization ,
the authors in @xcite solved the joint cs / cbf optimization , formulated as an extension of problem ( [ eq : inner - outer - problem ] ) for heterogeneous networks .
the authors in @xcite addressed the cs by iteratively evaluating the precoders and powers , following the methodology described in section [ section : algorithms : aggregated - utility : objective - function ] .
the approach in @xcite tackled the cbf problem using interference alignment @xcite , whereas the cs was sequentially solved in a greedy fashion . a dynamic transmission switching between st and jt
was presented in @xcite , where a centralized cs exploited individual rate statistics ( cdf scheduling ) . in @xcite ,
problem ( [ eq : csi - mapping - problem-1 ] ) was solved using metrics of spatial compatibility ( nsp - based user selection ) , and a centralized resource allocation solved problem ( [ eq : csi - mapping - problem-2 ] ) , [ cf .
section [ section : algorithms : aggregated - utility : csi - mapping ] ] .
practical cs schemes were proposed in @xcite for heterogeneous cellular systems , where macro and small cells coordinate scheduling decisions using an the approach described in section [ section : algorithms : aggregated - utility : objective - function ] .
the cbf was optimized using linear precoding and dynamic su / mu - mimo switching @xcite . given a pool of precoding weights
, each bs picks a different subset for transmission every scheduling interval , and switches them periodically , in a temporal beam - reuse fashion @xcite .
this method requires minimum coordination and uses information regarding the allocated beams per bs and the interference environment .
this allows practical scheduling and mitigates the flashlight effect ( changes of the active beams at neighboring transmitters ) .
another approach in @xcite , performs user scheduling by solving a series of assignment problems heuristically , and the precoding weights are optimized every iteration .
the clustered bss jointly select users and assign resources in a sequential fashion .
the first bs selects its users and broadcast its decision , then the second bs selects its user based on the decision made by the first bs and so on , see @xcite . the user selection per bs
can be done using the approaches presented in section [ section : algorithms : aggregated - utility ] , and the scheduling order of the cells can be assigned according to interference levels or using the round robin approach .
the cs can be performed based on interference constraints as in @xcite .
the bss schedule their users and perform resource allocation sequentially , so that the ici generated to previously selected users is below a threshold .
if the interference can not be mitigated as desired , the interfering bss remain silent during the scheduling interval , which is a low - complexity dynamic clustering strategy . a distributed optimization policy interconnects clustered bss through a cu or master bs , limiting the message exchange to local csi and scheduling control signaling @xcite .
problem ( [ eq : inner - outer - problem ] ) has three optimization variable sets , namely , @xmath46 , @xmath169 , and @xmath170 , and their joint optimization is difficult to solve optimally and distributively .
still , authors in @xcite proposed an iterative approach that decouples and optimizes some optimization variables , while the others remain fixed .
yu _ et.al _ in @xcite
have followed such an approach and developed a strategy to tackle ( [ eq : inner - outer - problem ] ) in a semi - distributed manner , demanding limited communication between bss and the cu : fix the first two set of variables and optimize the third set , then fix the second and third sets and optimize the first one , and so on until convergence .
the bss jointly update parameters associated with all variable sets in an semi - distributed fashion . by fixing the set of users and precoder weights ,
the power allocation can be performed based on the information shared between bss , see @xcite . by fixing the precoders and powers
, the user scheduling can be performed using the approaches described in sections [ section : algorithms : aggregated - utility ] . by fixing the set of users and power allocation ,
the precoder weights can be computed using distributed algorithms , see @xcite .
a semi - distributed cs was proposed in @xcite , where the users are selected using an approximation of the nsp [ cf .
section [ section : metrics - spatial - compatibility : projection - spaces ] ] , and the cbf is computed with local csi .
a similar approach was proposed in @xcite , where the bss dedicate spatial dof to cancel ici for some selected users at neighboring cells .
this approach performs semi - distributed spatial clustering for cs [ cf . section [ section : metrics - spatial - compatibility : spatial - clustering ] ] , and distributed linear precoding for cbf .
in mu - mimo systems with partial csit , resource allocation algorithms can multiplex up to @xmath18 users accounting for the feedback of one scalar ( cqi ) and one index ( cdi ) @xcite , [ cf .
section [ section : introduction : needof - csit ] ] .
the feedback information load is proportional to the number of deployed users and antennas , but scheduling usually requires rough quantization resolution in the cqi to differentiate between high and low rate users @xcite .
the cdi plays a more relevant role to achieve spatial multiplexing , thus , it requires a higher quantization granularity .
user scheduling requires two main steps : _ i _ ) the transmitter sends pilots for csi acquisition , the users quantize their channels and feed back the cqi and cdi ; _ ii _ ) subsets of users and beams are selected for data transmission based on a particular performance metric @xcite . as discussed in section [ section : precoding : partial - csit ] , there are two approaches to acquire channel information using codebooks , i.e. , quantizing either the channel , or the precoder that better fits the channel . in the following , we classified different scheduling approaches , based on the type of quantization .
for the sake of exposition , assume a miso scenario ( @xmath203 ) , linear precoding , and equal power allocation ( @xmath204 , @xmath32 ) .
the @xmath205 of the @xmath23-th user can be given by its sinr defined as @xcite : @xmath206 where @xmath207 , are the precoding vectors extracted from the cdis ( e.g. using the zfbf ) .
@xmath208 is the actual quantized unit - norm channel or @xmath209 given by @xmath210 where @xmath211 is a predefined codebook , [ cf .
section [ section : precoding : partial - csit ] ] .
the @xmath23-th user determines its cdi using ( [ eq : pcsit_channel_quantization_best_beam_user_k ] ) , according to a minimum distance criterion , see @xcite for miso or @xcite for mimo setting .
the solution of ( [ eq : pcsit_channel_quantization_best_beam_user_k ] ) has the following geometrical interpretation : the user @xmath23 selects the most co - linear or correlated codeword to its channel , which is equivalent to find the cone containing @xmath33 , as illustrated in fig .
[ fig : hyperslab_cones ] .
notice that the @xmath212 takes into account channel magnitudes , quantization errors of the cdi , and the spatial compatibility of channel @xmath33 regarding @xmath82 .
however , to compute the precoders @xmath207 , it is necessary to know a priory @xmath46 , and the associated cdis @xmath213 , which results in a chicken - and - egg problem .
therefore , the exact value of ( [ eq : pcsit_sinr_user_k ] ) is unknown at the transmitter or receiver sides .
several works have proposed different ways to approximate ( [ eq : pcsit_sinr_user_k ] ) , mainly by its upper / lower bounds or expected value , cf .
if the transmitter has statistical csi knowledge , the formulation of ( [ eq : pcsit_sinr_user_k ] ) must take into account the covariance matrices , @xmath84 , @xmath32 , to have a more accurate estimate of the interference powers @xcite .
the cqi can also be given by the channel magnitude , i.e. , @xmath214 , but this measure ignores the valuable information contained in the codebook ( spatial compatibility ) , and neglects quantization errors .
the number of competing users , the feedback load , and scheduling complexity can be reduced by setting thresholds to the cqis .
this means that only users with strong channels will be considered for scheduling @xcite .
multiplexing and mudiv gains are realized by carefully tuning the threshold according to statistics ( e.g. @xcite ) or numerical characterization ( e.g. , @xcite ) of the cqi , for a set of fixed parameters @xmath79 , @xmath18 , @xmath21 , and @xmath121 , [ cf .
section [ section : metrics - spatial - compatibility : spatial - clustering ] ] . in some scenarios where the maximum number of spatial streams per user is at most one
, each receiver uses either a weight vector @xcite or a combining technique @xcite to transform its mimo channel matrix into an _ effective _ miso channel vector .
different approaches to reduce feedback load and compute the cqi and cdi can be implemented ( e.g. @xcite ) . in general
, the cqi requires less bits than the cdi , but an optimized bit allocation must take into account the snr regime , the number of competing users @xmath21 @xcite , and practical quantization levels ( e.g. mcss ) @xcite , [ cf .
section [ section : algorithms : selection - partial - csit : two - stage - feedback ] ] .
assume that the cqi and cdi are fed back by all competing users , so that the quantized channels can be reconstructed at the transmitter .
resource allocation can be performed using the methods described in section [ section : algorithms ] or by low - complexity approaches based on antenna selection , e.g. @xcite .
zf precoding is the most common technique used in scenarios with quantized channels , hence , scheduling based on metrics of spatial compatibility [ cf .
section [ section : metrics - spatial - compatibility ] ] , can be directly applied , e.g. , @xcite .
the user selection for the general mimo scenario can be performed by computing effective miso channels or by treating each receive antenna as a virtual user @xcite .
it is worth mentioning that zf becomes interference - limited in the high snr regime with fixed @xmath79 .
however , if @xmath21 or @xmath79 scale with the snr , the achievable rates can be improved and the quantization errors can be mitigated @xcite . a common approach to combat iui ( generated by quantization errors ) , is by reducing the multiplexing gain .
finding the optimal number of data streams is an optimization problem that depends on the system parameters , e.g. snr regime , mobility , @xmath21 , @xmath79 , and @xmath18 , see @xcite . if @xmath215 , it is likely to have a codebook with limited granularity ( fixed @xmath79 ) yielding irreducible iui .
in such a case , the subset of selected users @xmath46 , must have cardinality strictly less than @xmath18 , and their cdis must have high resolution . in that way , the precoding performance can be improved and additional spatial dof can be used to cancel interference out @xcite .
the results in @xcite show that , the number of selected users highly depends on @xmath79 , whose optimal value is a function of the snr regime . for extreme values of the snr ( very low or high ) ,
the optimal transmission schemes are tdma or su - mimo , which completely avoid iui caused by inaccurate csit .
experimental results in @xcite show that selecting about @xmath216 of the total available beams maximizes performance in different mu - mimo wlan configurations .
it is worth noting that the snr regime plays an important role in the scheduling rule design .
authors in @xcite suggested the following guidelines to improve performance : 1 . in the high snr regime with fixed codebook size @xmath217 ,
the system becomes interference limited [ cf .
definition [ defn : interference - limited - system ] ] .
the scheduling rules should prioritize users whose fed back channels reduce the quantization error , i.e. , the cdi and the available spatial dof defined the attainable performance .
results in @xcite show that the error quantization can be mitigated either in the large user regime ( @xmath76 ) , or the high resolution regime ( @xmath218 ) .
if the variance of the noise and the iui are comparable , both the cqi and cdi should be taken into account for user selection .
3 . in the low snr regime ,
the system is noise limited and the scheduling rules should prioritize the cqi , since the cdi or error quantization play a negligible role .
these guidelines for scheduling design can be applied to mu - mimo scenarios with full csit , and they have been extended to multi - cellular cooperative systems in @xcite .
channel information acquisition based on rbf uses a codebook @xmath85 , to define the precoders and provide high flexibility for scheduling , [ cf . section [ section : precoding : partial - csit ] ] .
in contrast to the methods described in section [ section : algorithms ] , where the precoders are unknown before user selection , the approach using rbf can efficiently compute a performance metric that includes the effective channel gain and interference due to quantization errors .
however , the accuracy of the quantized channel information is limited by the basis ( rbf ) or bases ( pu2rc ) that comprise the codebook @xcite . in general , user scheduling
is performed by joint user selection and precoding allocation , and the overall complexity depends on the parameters @xmath18 , @xmath79 , and @xmath21 .
the set @xmath46 that solves problem ( [ eq : inner - outer - problem ] ) , can be found as illustrated in phase 1 of fig .
[ fig : beam_selection ] , which is based on the algorithm proposed in @xcite assuming that @xmath219 and @xmath74 . @xmath67 _ training phase _ : the precoding vectors in @xmath87 can be generated according to a known distribution and chosen randomly .
the transmitter sends pilots on all the spatial beams @xmath220 , so that all users can estimate their channels .
the codebook design can be simplified by defining codewords as the orthonormal basis of @xmath221 , as is @xcite . for this codebook design , antenna selection
is implicitly performed at the receiver side .
this means that each receive antenna is treated as an individual user , which can reduce scheduling complexity and feedback load .
another approach to construct @xmath87 , is by eigen - codebook design @xcite .
each user computes eigen - decomposition of its covariance matrix @xmath84 , @xmath32 , and feeds back the eigenvector associated to its maximum eigenvalue .
this approach requires extra feedback load , but can provide flexibility to the scheduler .
the quantization granularity of @xmath87 can be enhanced if the codebook comprises multiple bases , as in pu2rc @xcite .
@xmath67 _ csi feedback _ : the users compute a cqi metric per beam , and report their largest cqi to the transmitter .
the _ effective sinr _ of the @xmath23-th user in the @xmath222-th beam can be defined as @xcite : @xmath223 and the index of the best precoder is given by @xmath224 where @xmath225 , is an index subset of active codewords defined by the transmitter so that @xmath226 .
the user designates its cdi@xmath227 and cqi@xmath228 with @xmath229 , and @xmath230 is a quantization function , see @xcite and discussion in section [ section : algorithms : selection - partial - csit : two - stage - feedback ] . for @xmath231 , evaluating ( [ eq : pcsit_rbf_sinr_user_k ] ) and ( [ eq : pcsit_rbf_best_beam_user_k ] ) can be performed in different ways @xcite : using receiver combining techniques , e.g. mrc , if one beam is assigned per user ; or equalization techniques , e.g. mmse @xcite , when multiple beams are assigned per user .
the cqi computation is not limited to the achievable sinr or peak rate , e.g. , it can be computed from sufficient statistics of the channel conditions and past cqis @xcite .
for instance , authors in @xcite defined the cqi as an estimation of the minimum power required to achieve a target sinr . in @xcite ,
the cqi is defined as a function of the current and previous channel conditions , which is an approach similar to the cdf - based scheduling @xcite .
this probabilistic method uses the channel or rate distributions to schedule the users that are more likely to achieve high performance @xcite .
one can preselect the competing users based on a performance threshold , which reduces feedback and scheduling complexity . accounting for i.i.d .
channels allows to simplify the threshold design , since the channel directions follow a uniform distributions @xcite .
the @xmath23-th user feeds back information only if its associated cqi is above a predefined threshold , @xmath232 , a scheme called _ selective multiuser diversity _ , see @xcite .
an extension of such a method is the multi - threshold selection , where the scheduling rule also takes into account the mcss supported over each link , e.g. @xcite .
if the system optimization involves queue stability constraints , other thresholds based on qsi can be imposed on top of @xmath232 , provided that the users have knowledge of their respective queue lengths @xcite .
preselection can be performed in the spatial domain [ cf .
section [ section : metrics - spatial - compatibility : spatial - clustering ] ] , where the users meet a constrained in the quantization error .
the set of competing users can be defined as @xmath233 , for a predefined threshold @xmath121 , see @xcite . in the large user regime ( @xmath76 ) with fixed @xmath18 and @xmath19 , the mudiv , for both full and partial csit , e.g. @xcite , show diminishing returns of mudiv over i.i.d .
rayleigh fading channels , i.e. , the capacity gain flattens as @xmath234 , see @xcite . ]
scales as @xmath235 @xcite , which suggests that preselection based on channel statistics , long - term throughput , or even randomly ( e.g. , @xcite ) , can reduce the feedback load and achieve mudiv gains .
@xmath67 _ user selection _
: the transmitter must find the set of users that maximizes the performance metric , e.g. wsr .
there are several scheduling approaches , e.g. , treating each receive antenna as a single user , assigning at most one beam per user , or allocating multiple beams per user .
the simplest selection is assigning the @xmath222-th beam to user @xmath236 , which is defined as @xmath237 if the quantization granularity of the function @xmath230 is low , it may happen that ( [ eq : pcsit_rbf_best_user_per_beam ] ) accepts more than one solution , in which case @xmath238 can be chosen randomly among the best user for beam @xmath222 @xcite . if the elements of @xmath33 are i.i.d .
, each selected user can not achieve maximum sinr for more than two beams on one antenna provided that @xmath239 .
this is a valuable design guideline that allows to reduce scheduling complexity @xcite .
analytical results in @xcite show that @xmath18 must be scaled proportional to @xmath240 , so that user starvation is avoided .
once that the set @xmath241 has been found , power allocation and link adaptation can be performed .
owing to the fact that the quantization granularity is fixed and bounded ( @xmath242 ) , the precoders @xmath243 can not fully separate users in the spatial domain .
hence , iui is unavoidable and particularly harmful in high snr .
moreover , in the sparse user regime , @xmath215 , rbf can not benefit from mudiv and performs poorly .
several user scheduling algorithms have been proposed in @xcite to handle sparsity and limited mudiv over mmwave channels .
the optimum number of beams that maximizes the wsr is a function of the number of competing users and the snr regime @xcite .
analytical results establish that for rbf with fixed @xmath21 , and extreme values of the snr regime , the optimal number of active beams is one , see e.g. , @xcite . ]
@xcite . depending on the channel conditions and the system constraints ,
user scheduling might yield a dynamic switching between tdma ( time - sharing ) and mu - mimo transmission .
this can take place if iui is very high , or if the individual rate or power constraints are violated @xcite .
therefore , efficient operation in an arbitrary snr requires a reduced number of scheduled users and active beams , i.e. , @xmath244 .
analytical results in @xcite show that @xmath245 can be attained if @xmath76 for a fixed @xmath79 .
in general , the optimal set of selected users is such that @xmath246 , @xcite . finding the best subset of users and beams is a combinatorial problem whose optimal solution is found by exs @xcite .
one has to enumerate all user combinations , compute their associated performance metrics @xmath148 , and select the subset with maximum @xmath148 , as illustrated in phase 2 of fig .
[ fig : beam_selection ] .
one alternative is to formulate an integer programming optimization , as in @xcite , or to perform greedy selection , see @xcite and section [ section : algorithms : aggregated - utility ] .
, the user selection is done by finding the strongest user per beam / antenna / spatial dimension associated with @xmath87 .
phase 1 matches a user with a beam based on the cqi .
phase 2 maximizes the utility function by selecting a subset of users and their associated beams according to some enumeration or search technique.,title="fig : " ] + several extension to the approach presented in fig .
[ fig : beam_selection ] have been developed : improving codebook resolution , e.g. @xcite ; dynamically updating multi - bases codebooks @xcite ; computing the cqi by different functions , e.g. @xcite ; combining rbf with deterministic precoding schemes , e.g. @xcite ; dynamically adjusting the number of active beams , e.g. @xcite , etc . to enhanced the quantization granularity ,
authors in @xcite generate several sets of @xmath217 beams , and the users report the best cqi and cdi per set .
the user selection in @xcite combines rbf and szf for the miso scenario : given an initial random precoding @xmath247 , its associated user @xmath248 solves @xmath249 , and feeds back a finer cdi that is used to generate the precoder @xmath250 for the second user , @xmath251 , and so on . in this way
, the @xmath23-th selected user will only receive interference from the previous @xmath252 selected users .
this algorithm fits particularly well in heterogeneous systems , where users can be clustered according to their locations or channel statistics , e.g. , @xcite . in systems with qos constraints [ cf . section [ section : performance - uf : wsr - with - qos ] ] , the opportunistic nature of the algorithm in fig .
[ fig : beam_selection ] changes and plays a secondary role , which has a direct impact in the mudiv gains .
this is because the competing users are preselected based on qsi , since scheduling exclusively based on cqis incurs in delay penalties @xcite . in @xcite , a subset of user
is preselected using qsi , and only those users feed back their csi .
the final selection is done as shown in fig .
[ fig : beam_selection ] , where the cqis are computed as the effective sinrs weighted by their respective qsis . notice that the parameters of the deployment may affect the latency experienced by the users .
for instance , large values of @xmath79 might result in more scheduled users since iui can be suppressed more efficiently . several works in the literature ( cf .
@xcite ) , have studied the user selection and precoding design problems in terms of the fed back information required to implement them .
the total amount of feedback bits per user , @xmath253 , is partitioned such that @xmath254 , where each part is used to solve a problem , as illustrated in fig .
[ fig : two - phase - feedback ] .
notice that the beam allocation problem in fig .
[ fig : beam_selection ] and the bit allocation problem fig .
[ fig : two - phase - feedback ] , bear some resemblance to the methodology described in section [ section : algorithms : aggregated - utility : csi - mapping ] . the user selection problem ( [ eq : csi - mapping - problem-1 ] ) ,
is solved using some properties of the csit of all competing users , and the precoding design and power allocation problem ( [ eq : csi - mapping - problem-2 ] ) is solved only for the set of users previously selected or grouped . the communication phases between the transmitter and the receivers in fig .
[ fig : two - phase - feedback ] , are part of a training or pilot mode @xcite , which is applied to mu - mimo system with partial csit . taking into account the type of implemented quantization , some remarks are in order @xcite : 1
. for high - rate feedback systems , a large value of @xmath253 for channel quantization might be more efficient .
the accurate cdi can be used for precoding design and to enhance of the multiplexing gain at high snr @xcite .
2 . for low - rate feedback systems with small values of @xmath253 , the rbf or pu2rc schemes , [ cf . section [ section : algorithms : selection - partial - csit : rbf ] ] , are more suitable since they can meet rate constraints and simplify the scheduling . in scenarios with limited feedback rates , @xmath255 ,
it is desirable to : _
i _ ) achieve significant mudiv using @xmath256 bits for the cqi ; _ ii _ ) design codebooks using @xmath257 bits of resolution for the cdi , to efficiently mitigate quantization errors , iui , and achieve multiplexing gains ; and _ iii _ ) reduce the scheduling complexity and limit the number of users that feedback @xmath257 bits .
the optimal partition of @xmath253 depends on system parameters ( @xmath19 , @xmath18 , and @xmath21 ) , snr regime , and type of precoding scheme @xcite .
moreover , practical systems not only limit the amount of bits per user , @xmath253 , but also the total feedback rate , @xmath258 , e.g. , if the users share the feedback link , see @xcite .
observe that the multiuser scheduling in fig .
[ fig : two - phase - feedback ] requires csi feedback and signaling of the scheduling decisions to the selected users .
these feedback and control loop introduce a non - negligible overhead , @xmath259 , and latency in the system that must be carefully considered in the scheduling rule design @xcite .
in the first feedback phase in fig . [
fig : two - phase - feedback ] , the competing users send a coarse quantized version of their achievable performance metric .
the @xmath260 is quantized using @xmath256 bits , where @xmath261 is the @xmath117-th quantization level .
the set of users in @xmath46 , can be chosen in a ranking - based per - beam selection , as in phase 1 of fig .
[ fig : beam_selection ] , or per - antenna selection as in @xcite .
a number of works have shown that depending on the system parameters , choosing @xmath262 can be enough to attain the mudiv gain of the unquantized cqi , see @xcite .
results in @xcite and references therein , show that efficient precoding design is attained by assigning roughly @xmath263 out of @xmath253 bits to the cqi , and allocating the remaining bits to the cdi .
if the system operates in the large user regime ( @xmath76 ) , small values of @xmath256 are enough to achieve asymptotically optimal system performance , see @xcite and references therein .
mudiv is defined by users with sinr above a threshold , and constructing @xmath46 depends on cdis , since spatially compatible users will be grouped with high probability @xcite . in the high resolution regime ( @xmath264 ) , an efficient partition meets @xmath265 , and the fraction of bits dedicated for cdi scales proportionally to @xmath266 , see @xcite . if @xmath215 , larger values of @xmath256 are required to circumvent the lack of mudiv and properly identify strong users .
however , in the high snr regime is more beneficial to have @xmath267 , since quantization errors in the cdi are the main performance - limiting factor @xcite .
mudiv clearly affects the optimal partition of @xmath79 .
numerical results in @xcite show the effects of @xmath21 over the feedback rates , and the interdependence between mudiv and @xmath257 .
accounting for link adaptation , @xmath256 must be chosen so that the cqi fully exploits the granularity of the available mcss , either for channel quantization @xcite or rbf @xcite .
the value of @xmath256 depends on the type of metric that defines the cqi .
for example , the sinr and the achievable rate are real numbers , whereas other metrics might already be defined as integer numbers , see @xcite . in the second feedback phase in fig .
[ fig : two - phase - feedback ] , each user in @xmath46 uses @xmath257 bits to report its cdi .
this mean knowledge of refined channel information at the transmitter , e.g. , more accurate spatial directions or effective channel gains .
if the system is constrained in the total feedback bandwidth , @xmath268 , heterogeneous and dynamically allocated bits can be used to quantize the cdi . by assigning different number of bits per user , i.e. , @xmath269 , @xmath150 , efficiency at the cdi feedback phase
can be achieved @xcite .
the dynamic assignment of @xmath269 is more effective in scenarios where the users have heterogeneous average long - term channel gains .
the bit allocation rules might consider the following guidelines
@xcite : 1 . for high snr users , large values of @xmath269
reduce quantization errors and mitigate iui .
this is particularly important in scenarios where the cdi is used to compute the precoder weights ( e.g. , mmse @xcite or zfbf @xcite ) , [ cf .
section [ section : algorithms : selection - partial - csit : qunatized - csi ] ] .
2 . for low snr users ,
the performance is noise limited , i.e. , the noise variance is larger than the interference , and the value of @xmath269 is relatively less important to optimize performance .
dynamic assignment of @xmath269 can be extended to sequential transmissions with error correction mechanisms ( e.g. arq @xcite ) .
such an assignment can be used to further refine @xmath46 , by temporarily dropping users retransmitting the same packet multiple times .
4 . in multiple - transmitter scenarios ,
the level of coordination and transmission scheme [ cf .
section [ section : algorithms : multi - cell - selection - classification ] ] , define the methodology to assign the value of @xmath269 at each transmitter @xcite . results in @xcite suggest that grouping users based on their locations or long - term channel gains ( e.g. , @xcite ) , yields a more efficient assignment of the feedback bandwidth and simplify the user grouping .
another parameter to be considered when defining @xmath256 and @xmath257 , is the time used for hand - shaking between feedback phases . the total time used for training and data transmission
must be kept below the coherence time of the channel , so that the impact of delayed / outdated csi is minimized @xcite .
otherwise , increasing @xmath257 would not be enough to achieve mimo gains @xcite .
analytical results in @xcite show that for multi - basis codebooks ( e.g. , pu2rc ) , large values of @xmath257 usually do not provide considerable capacity gains , and the overall performance is highly sensitive to mudiv . a design rule to bear in mind
is that the accuracy of the csit is more valuable than mudiv in practical mu - mimo scenarios @xcite .
+ the scheduling algorithms for mu - mimo in lte - advanced rely on quantized csi : _ i _ ) the ri / pmi metrics contain information regarding the su - mimo channel ( due to backward compatibility ) projected onto the subspace defined by the codebook ; _ ii _ ) the cqi metric indicates the energy of the projection , and it might include ici and noise @xcite . since the aggregated multiuser channel can be constructed by concatenating the users pmi metrics , and the cqis are known at the transmitter , the scheduling algorithms described in the subsections above can be directly used in cellular scenarios . in general , the scheduling algorithms in lte - advanced are proprietary and implementation - specific and
there is no standardized procedure to define them @xcite .
it is worth noting that the amount of available csit defines the transmission mode ( su- or mu - mimo ) , and the scheduling decisions .
practical scheduling rules with switching mode can be defined by evaluating the achievable performance of su - mimo and mu - mimo ( with multi - rank transmission ) modes , and simply choosing the one with better performance @xcite
. however , robust switching between these modes still requires research to guarantee adaptability to changes in channel and traffic conditions , as well as viable computational complexity and scalability @xcite .
authors in @xcite , highlight the fact that for known precoding matrices at the transmitter , it is possible to generate look - up - tables ( luts ) that contain information regarding spatial compatibility and potential interference .
resource allocation based on luts could be used to significantly reduce the complexity of the scheduling algorithms .
there are several factors that must be considered for cs / cbf design in cellular systems @xcite : dynamically determining whether or not coordination is required , switching between jt and cs / cbf , optimizing cluster formation and sizing , and scalability issues .
the resource allocation strategies must also include constraints on backhaul bandwidth , csi acquisition , latency , qos , mobility , and synchronization . in the ieee 802.11ac standard ,
mu - mimo transmission consists of two tasks @xcite : _ i _ ) to identify the users that belong to the transmission set , which is implementation - specific and might depend on priority weights or traffic related parameters [ cf .
section [ section : performance - uf ] ] ; _ ii _ ) the assignment of a group i d for each set of co - scheduled users , and the downlink transmission signaling . for a given time instance
, the ap constructs a group i d assignment table .
the number of rows is defined by the available number of i d groups , and the columns are given by the number of associated users .
the condition to have a proper group i d , is to have different users assigned to each one of the available stream positions . selecting the proper group i d and its associated scheduled users is equivalent to a coloring problem over a two dimensional array , whose optimal solution can be found by exs .
a heuristic algorithm was proposed in @xcite , where the positions of the users in a particular group i d are determined according to the probability of occurrence .
the group i d selection and success allocation probability can be improved by modeling the discrete searching process as a linear sum assignment problem , for which there exist reliable polynomial - time algorithms @xcite . for the general mu - mimo scenario , a selected user can receive more than one data stream , and
all of them must use the same mcs @xcite .
the authors in @xcite proposed a user selection algorithm inspired by @xcite , originally designed for cellular scenarios .
the scheme applied bd with geometric - mean decomposition to guarantee equal mcs allocation over all spatial streams assigned to a single user .
that work highlights the need of careful stream allocation per selected user , where the best number of streams is not necessarily equal to rank of the channel .
another approach to guarantee maximum sum rate with joint link adaptation ( mcs selection ) , bd , and user / stream selection was proposed in @xcite .
the authors extended the scheduling algorithm in @xcite , originally designed for cellular scenarios , to wlans with partial csit .
the link adaptation is performed by a machine learning classifier that provides robustness to csi inaccuracy .
numerical results suggest that estimation of the iui is required to allocate mcs more efficiently .
the scheduling complexity can be classified in two types @xcite : implementation complexity and computational complexity .
the former , refers to the amount of signaling overhead and information exchanged among different network entities .
the latter refers to the processing time required to execute a certain algorithm at the transmitter or cu .
the implementation complexity is assumed to be reduced whenever the channel information is quantized [ cf .
section [ section : precoding : partial - csit ] ] .
this type of complexity can be estimated as a function of the system parameters @xmath21 , @xmath79 , @xmath19 , @xmath18 , and the number of clustered transmitters .
nevertheless , to the best of our knowledge , there is no published work or reference framework that analyze the complexity in a fundamental and general way , at least for the mu - mimo systems considered here .
most research in the field of limited csi feedback has focused on the following issues related to complexity @xcite : _ i _ ) reducing the number of parameters to be quantized , e.g. , bits to quantize the cqi and cdi ; _ ii _ ) reducing the required quantized feedback resolution ; and _ iii _ ) constraining the signaling and message exchange within a cluster , and minimizing the overall coordination . in the mu - mimo literature ,
most works focus on the computational complexity required to select users , to extract the precoders , and to perform power allocation .
there are several metrics to estimate the computational complexity : _ i _ ) number of real or complex operations .
_ ii _ ) number of flops , where a flop is defined as a floating point operation @xcite , e.g. , a real addition , multiplication , or division .
_ iii _ ) using the big - oh notation , @xmath270 , which is proportional to the term that dominates the number of elementary operations needed to execute an algorithm @xcite .
_ iv _ ) other metrics , such as the number of vector / matrix operations , the number of iterations needed for convergence , or results based on simulations and execution time .
therefore , there is no unified approach to characterize the computational complexity for scheduling algorithms , but the aforementioned metrics , summarized in table [ tab : complexity ] , provide a coarse assessment of the computational load order .
we hasten to say that another sort of complexity analysis can be performed for mu - mimo systems , in which precoders and powers are jointly optimized for a fixed set of users @xmath46 , [ cf .
section [ section : performance - uf : wsr - with - qos ] ] .
the resource allocation problem only involves @xmath169 , @xmath170 , and is formulated as the optimization of a certain system - level utility function , @xmath148 , [ cf .
section [ section : performance - uf : weighted - sum - rate ] ] , subject to resource budget constraints , see @xcite and references therein . in these cases ,
the set of scheduled users is not optimized , and the complexity is normally defined as a function of @xmath18 , @xmath19 , and the characteristics of @xmath148 , e.g. , convexity or non - convexity .
0.48l x * metric & * references + flop - based & @xcite + real / complex operations & @xcite + @xmath270 & @xcite + other metrics & @xcite + * *
the optimal power allocation @xmath170 in ( [ eq : inner - outer - problem ] ) depends on the csit availability , the type of precoding scheme implemented at the transmitter , the individual user priorities , and the objective function .
accounting for full csit in single - transmitter scenarios , the optimal power allocation is usually known in analytical closed - form for several performance metrics , e.g. , ber , sinr , wsr , or fairness @xcite . assuming that @xmath46 is fixed , and linear precoding is implemented at the transmitter
, a network - centric power control algorithm finds the optimal @xmath170 for wsr maximization through convex optimization , i.e. , using the water - filling principle , see @xcite .
the linear precoding decouples the signals into orthogonal spatial directions mitigating the iui , and water - filling allocates powers according to the effective channel gains [ cf . definition [ defn : effective_channel_gains ] ] . in scenarios with
a fixed feasible set of users @xmath46 , [ cf . definition [ defn : feasible_set_of_users ] ] , and assuming that @xmath169 is coupled with @xmath170 , the power allocation may involve , for example , a wsr maximization with rate control , sum power minimization with individual sinr constraints , or a max - min sinr problem , see @xcite . for this kind of problems ,
power control algorithms are designed based on optimization theory ( see @xcite ) , and the perron - frobenius theory ] , such as wireless sensor or multi - hop networks @xcite . ] for non - negative matrices ( see @xcite ) .
in contrast to wsr maximization , where water - filling assigns more power to stronger channels , other objective functions require to allocate powers in a different way .
for instance , balancing the power over all users so that weak users are assigned more power to improve their error rates @xcite , assigning powers according to target sinrs @xcite , or considering queue stability constraints @xcite .
some works model @xmath169 so that the power is absorbed into it , and optimizing the directions and magnitudes of the precoder weights implicitly performs power control , e.g. , @xcite .
the interested reader is referred to @xcite for a comprehensive survey on theory and algorithms for joint optimization of @xmath169 and @xmath170 .
equal power allocation ( epa ) is a suboptimal strategy used to simplify the evaluation of the utility function @xmath148 , especially when the transmitter does not have full csi knowledge @xcite . in certain scenarios ,
epa allows a more tractable system performance analysis or derivation of statistics based on @xmath21 , @xmath18 , @xmath19 , or @xmath79 , which generally yields closed - form expressions . assuming full csit , if the wsr maximization problem [ cf .
section [ section : performance - uf : weighted - sum - rate ] ] is optimized using zf - based or bd - based precoding schemes , the epa ( directly proportional to the individual weights @xmath157 , @xmath150 ) , asymptotically achieves the performance of the optimal water - filling power allocation at high snr @xcite .
the authors in @xcite discussed the secondary role of power allocation when optimizing problem ( [ eq : inner - outer - problem ] ) .
the paper suggested that more efforts should to be taken to find @xmath46 and the corresponding @xmath169 , which may justify the adoption of epa under specific snr conditions . in systems with partial csit ,
power allocation requires numerical methods that depend on the optimized performance metric @xcite .
consider a practical scenario where @xmath79 is finite and fixed , and @xmath46 has been found by one of the methods described in section [ section : algorithms : selection - partial - csit ] .
in such a scenario , the system is limited by interference [ cf .
definition [ defn : interference - limited - system ] ] , at the moderate and high snr regimes , since the interference components in the denominator of ( [ eq : pcsit_sinr_user_k ] ) and ( [ eq : pcsit_rbf_sinr_user_k ] ) are non - zero .
therefore , one of the main objectives of power control is to perform interference management .
however , to compute the optimal @xmath170 for wsr maximization , the transmitter must have knowledge of all interference components for all users .
this results in non - convex optimization problems , for which efficient algorithms ( depending on the operating snr regime ) can provide suboptimal solutions @xcite .
moreover , global knowledge of the interference components at the transmitter may not be attainable in limited feedback systems ( e.g. only cqis are known ) , and adopting epa is a common practice in the reviewed mu - mimo literature .
l * 3>x * method * & * epa * & * water - filling * & * adaptive * + * utility - based * & @xcite & @xcite & @xcite + * csi - mapping * & @xcite & @xcite & @xcite + * metaheuristic ( stochastic ) * & @xcite & @xcite & @xcite + * classic optimization * & @xcite & @xcite & @xcite + * exhaustive search * & @xcite & @xcite & - + in multi - transmitter scenarios , the level of coordination , the availability of user data and csit , and the precoding scheme determine the best power allocation policy
. in full coordinated scenarios , all transmitters belong to the same infrastructure , and a cu determines the power allocation across the cluster . assuming global csit , a fixed set @xmath46 , and linear precoding schemes , the power allocation that optimizes a global performance metric subject to per - transmitter power constraints
can be realized numerically or through water - filling , see @xcite . for scenarios with partial coordination between transmitters , assuming that @xmath46 is globally known , the power allocation depends on the precoding schemes computed from local csi and individual priorities @xmath157 , @xmath150 , known by each transmitter , see @xcite .
table [ tab : power - allocation ] summarizes the reviewed scheduling methods and their associated power allocation strategies : epa , water - filling , and adaptive methods .
the references listed under the _ adaptive _ category depend on the particular objective function and system constraints , see tables [ tab : precoding ] and [ tab : optimization - criteria ] . those references
jointly optimize @xmath170 and @xmath169 via conventional optimization methods @xcite and iterative algorithms .
there are many challenges to be solved for resource management in mu - mimo systems .
identify and standardize the best practices to group and schedule users is fundamental to profit from mudiv .
this is tightly related to affordable - complexity algorithm design to efficiently manage the allocation of multiple users , carriers , time slots , antennas , and transmitters .
one of the main requirements to obtain mudiv and multiplexing gains is knowledge of the csi . in practical systems , this is challenging because the feedback load rapidly increases with @xmath21 and @xmath18 .
multiple transmit antennas require additional pilot overhead proportional to @xmath18 , if the users need to learn the complete mimo channel @xcite .
further research is required to completely understand performance bounds advance transmission schemes , e.g. , dynamic su-/mu - mimo switching and multi-/single - stream allocation .
the viability of user scheduling and simultaneous transmission in multi - cell deployments must be further investigated .
several topics must be included in future studies : synchronization issues in mu - mimo , metrics for user grouping in hetnets , scheduling signaling load , coordinated power control , latency and delay due to scheduling , mobility , the impact of realistic traffic models , backhaul constraints , and semi - distributed control algorithms .
the scheduling policies for the next generation of wireless technologies must balance between the high cost of csi acquisition and the benefits of cooperative transmission .
the current performance of mu - mimo processing is still limited in 4 g cellular communication systems .
this is because user terminals do not perform interference estimation , and in all cases only csi feedback for the su - mimo mode ( su - ri , pmi , and cqi ) is employed @xcite .
most of the reviewed works assume continuous power control , but lte only supports discrete power control in the downlink , through a user - specific data - to - pilot - power offset parameter @xcite .
these facts must be considered to properly assess current and emerging noma schemes in multiuser scenarios , as well as their limitations , see @xcite .
resource management rule have a fundamental role to play in future generations of wireless networks , where new and evolving technologies such as mmwave and massive mimo have already been considered @xcite .
such rule will require novel precoding designs and an assertive usage of the multiuser dimension to provide substantial spectral efficiency , energy efficiency and user satisfaction .
l * 4>x * parameter * & * miso * , @xmath146 & * miso * , @xmath147 & * mimo * , @xmath146 & * mimo * , @xmath147 + * capacity * & @xcite & @xcite & @xcite & @xcite + * queues * & @xcite & @xcite & @xcite & @xcite + * high snr * & @xcite & @xcite & @xcite , @xcite & @xcite + * low snr * & @xcite & @xcite & @xcite @xcite & @xcite + * large * @xmath21 & @xcite , @xcite & @xcite & @xcite , @xcite & @xcite , @xcite + * large * @xmath18 & @xcite & @xcite & @xcite , @xcite & @xcite + * bits * @xmath256 , @xmath257 & @xcite & @xcite & - & @xcite , @xcite + & & * fdd * & frequency - division - duplex & & * pmi * & precoding matrix index + * adc * & analog - to - digital converter & & * ga * & genetic algorithms & & * qos * & quality of service + * ap * & access point & & * gus * & greedy user selection & & * qca * & quantization cell approximation + * ack / nack * & acknowledgement handshake & & * hetnet * & heterogeneous network & & * qsi * & queue state information + * awgn * & additive white gaussian noise & & * ici * & inter - cell interference & & * rf * & radio frequency + * arq * & automatic repeat request & & * ifc * & interference channel & & * rbf * & random beamforming + * bs * & base station & & * iui * & inter - user interference & & * ri * & rank indicator + * ber * & bit error rate & & * jt * & joint signal transmission / processing & & * rss * & received signal strength + * bd * & block diagonalization & & * los * & line - of - sight & & * sus * & semi - orthogonal user selection + * bb * & branch and bound & & * lte * & long term evolution & & * slnr * & signal to leakage plus noise ratio + * bc * & broadcast channel & & * mrt * & maximum ratio transmission & & * snr * & signal - to - noise ratio + * cu * & central processing unit & & * mse * & mean - square - error & & * st * & single transmission + * cdi * & channel direction information & & * mmse * & minimum mse filter & & * su - mimo * & single user mimo + * cqi * & channel quality information & & * mcs * & modulation and coding scheme & & * siso * & single - input single - output + * csi * & channel state information & & * mac * & multiple access channel & & * sdv * & singular value decomposition + * cdma * & code division multiple access & & * mimo * & multiple - input multiple - output & & * sdma * & space - division multiple access + * cizf * & constructive interference zf & & * miso * & multiple - input single - output & & * szf * & successive zf + * cbf * & coordinated beamforming & & * mudiv * & multiuser diversity & & * tdma * & time division multiple access + * comp * & coordinated multi - point & & * mu - mimo * & multiuser mimo & & * tdd * & time - division - duplex + * cs * & coordinated scheduling & & * noma * & non - orthogonal multiple access & & * thp * & tomlinson - harashima precoding + * csir * & csi at the receiver & & * np * & non - deterministic polynomial time problem & & * wi - fi * & trademark of ieee 802.11 + * csit * & csi at the transmitter & & * np - c * & np complete & & * vp * & vector perturbation + * cdf * & cumulative distribution function & & * nsp * & null space projection & & * vq * & vector quantization + * dof * & degrees - of - freedom & & * ofdma * & orthogonal frequency - division multiple access & & * wsr * & weighted sum rate + * dpc * & dirty paper coding & & * ofdm * & orthogonal frequency - division multiplexing & & * wlan * & wireless local area network + * ee * & energy efficiency & & * pso * & particle swarm optimization & & * zf * & zero forcing + * epa * & equal power allocation & & * pu2rc * & per unitary basis stream user and rate control & & * zfbf * & zf beamforming + * exs * & exhaustive search & & * phy * & physical layer & & * zfdp * & zf dirty paper + several authors have focused their work on asymptotic analysis and scaling laws of performance metrics or utility functions @xmath148 .
the analytical results depend on the _ system parameters _ , @xmath21 , @xmath19 , @xmath18 , @xmath79 , @xmath22 , and @xmath271 @xmath150 .
the information - theoretic results derived in several works provide fundamental limits of achievable values of @xmath148 from the user and system perspectives .
they also shed light on the relevance of each parameter , the conditions where the parameters are interchangeable , the potential and limitations of mu - mimo systems , and judicious guidelines for the overall system design . for each utility function @xmath148 , and user priorities @xmath157 @xmath32 , there exist optimal and suboptimal operation points that can be characterized according to the system parameters and their respective regimes .
the capacity of mu - mimo systems has been assessed in various asymptotic regimes : high snr ( @xmath272 ) , low snr ( @xmath273 ) , large number of users ( @xmath234 ) , large number of transmit antennas ( @xmath274 ) , and large codebook resolution ( @xmath218 ) .
table [ tab : analytical - results ] summarizes the system parameters and the antenna configurations , i.e. , miso ( @xmath203 ) or mimo ( @xmath231 ) , and @xmath275 .
every single reference in the table has its own system model , assumptions , and constraints , studying one parameter and the corresponding effects on the performance @xmath148 , and other fixed parameters .
the table is by no means exhaustive ; a comprehensive taxonomy of the asymptotic analytical results is out of the scope of this paper .
our aim is to provide a list of organized results from the reviewed paper , so that the interested reader may use each reference for further studies .
notice that the first row in table [ tab : analytical - results ] points to references that analyze the capacity as a function of different parameters , while the second row refers to works that provide analytical results of mu - mimo systems with queue constraints .
y. l. lee , t. c. chuah , j. loo , and a. vinel , `` recent advances in radio resource management for heterogeneous lte / lte - a networks , '' _ ieee communications surveys & tutorials _ , vol .
16 , no . 4 , pp . 21422180 , fourth quarter 2014 .
l. dai , b. wang , y. yuan , s. han , c .-
i , and z. wang , `` non - orthogonal multiple access for 5 g : solutions , challenges , opportunities , and future research trends , '' _ ieee communications magazine _ , vol .
53 , no . 9 , pp . 7481 , september 2015 .
l. lu , g. y. li , a. l. swindlehurst , a. ashikhmin , and r. zhang , `` an overview of massive mimo : benefits and challenges , '' _ ieee j. of s. topics in signal processing _ , vol . 8 , no . 5 , pp . 742758 , oct .
2014 .
f. capozzi , g. piro , l. grieco , g. boggia , and p. camarda , `` downlink packet scheduling in lte cellular networks : key design issues and a survey , '' _ ieee communications surveys & tutorials _ , vol .
15 , no . 2 , pp .
678700 , second 2013 .
d. gesbert , m. kountouris , r. heath , c .- b .
chae , and t. salzer , `` from single user to multiuser communications : shifting the mimo paradigm , '' _ ieee signal processing magazine _ , vol .
24 , no . 5 ,
3646 , sept 2007 .
j. mietzner , r. schober , l. lampe , w. gerstacker , and p. hoeher , `` multiple - antenna techniques for wireless communications - a comprehensive literature survey , '' _ ieee communications surveys & tutorials _ , vol .
11 , no . 2 , pp .
87105 , second 2009 .
chen and l .- c .
wang , `` enhancing coverage and capacity for multiuser mimo systems by utilizing scheduling , '' _ ieee transactions on wireless communications _ ,
vol . 5 , no . 5 , pp . 11481157 , may 2006 .
d. love , r. heath , v. lau , d. gesbert , b. rao , and m. andrews , `` an overview of limited feedback in wireless communication systems , '' _ ieee journal on selected areas in communications _ , vol .
26 , no . 8 , pp . 13411365 , october 2008 .
s. jafar and a. goldsmith , `` isotropic fading vector broadcast channels : the scalar upper bound and loss in degrees of freedom , '' _ ieee trans . on information theory _ , vol .
51 , no . 3 , pp .
848857 , march 2005 .
i. koutsopoulos and l. tassiulas , `` the impact of space division multiplexing on resource allocation : a unified treatment of tdma , ofdma and cdma , '' _ ieee transactions on communications _ , vol .
56 , no . 2 ,
pp . 260269 , february 2008 .
o. elijah , c. y. leow , t. a. rahman , s. nunoo , and s. z. iliya , `` a comprehensive survey of pilot contamination in massive mimo - 5 g system , '' _ ieee communications surveys & tutorials _
18 , no . 2 , pp .
905923 , secondquarter 2016 .
h. weingarten , y. steinberg , and s. shamai , `` the capacity region of the gaussian multiple - input multiple - output broadcast channel , '' _ ieee transactions on information theory _ ,
52 , no . 9 , pp . 39363964 , sept 2006 .
e. bjornson and e. jorswieck , _ optimal resource allocation in coordinated multi - cell systems _ , ser .
foundations and trends(r ) in communications and information.1em plus 0.5em minus 0.4emnow publishers incorporated , 2013 .
m. kobayashi , g. caire , and d. gesbert , `` transmit diversity versus opportunistic beamforming in data packet mobile downlink transmission , '' _ ieee trans .
on commun .
_ , vol .
55 , no . 1
, pp . 151157 , 2007 . j. zhang , m. kountouris , j. andrews , and r. heath , `` multi - mode transmission for the mimo broadcast channel with imperfect channel state information , '' _ ieee transactions on communications _
59 , no . 3 , pp .
803814 , march 2011 .
g. dimic and n. sidiropoulos ,
`` on downlink beamforming with greedy user selection : performance analysis and a simple new algorithm , '' _ ieee transactions signal processing _ , vol .
53 , no .
3857 3868 , oct
. 2005 .
j. wang , d. love , and m. zoltowski , `` user selection with zero - forcing beamforming achieves the asymptotically optimal sum rate , '' _ ieee transactions on signal processing _ ,
56 , no . 8 , pp . 37133726 , aug 2008 .
t. yoo and a. goldsmith , `` on the optimality of multiantenna broadcast scheduling using zero - forcing beamforming , '' _ ieee journal on selected areas in commun .
_ , vol . 24 , no . 3 , pp .
528 541 , march 2006 .
s. huang , h. yin , h. li , and v. leung , `` decremental user selection for large - scale multi - user mimo downlink with zero - forcing beamforming , '' _ ieee wireless communications letters _ , vol . 1 , no . 5 , pp .
480483 , october 2012 .
a. razi , d. ryan , i. collings , and j. yuan , `` sum rates , rate allocation , and user scheduling for multi - user mimo vector perturbation precoding , '' _ ieee transactions on wireless communications _ , vol . 9 , no . 1 , pp .
356365 , january 2010 .
y. shi , q. yu , w. meng , and z. zhang , `` maximum product of effective channel gains : an innovative user selection algorithm for downlink multi - user multiple input and multiple output , '' _ wireless communications and mobile computing _ , 2012 .
t. maciel and a. klein , `` on the performance , complexity , and fairness of suboptimal resource allocation for multiuser mimo - ofdma systems , '' _ ieee trans .
_ , vol .
59 , no . 1 ,
pp . 406419 , 2010 .
chung , l .- c .
wang , and c .- j .
chang , `` a low - complexity beamforming - based scheduling to downlink ofdma / sdma systems with multimedia traffic , '' _ wireless networks _ , vol .
17 , no . 3 , pp .
611620 , 2011 .
d. christopoulos , s. chatzinotas , i. krikidis , and b. ottersten , `` constructive interference in linear precoding systems : power allocation and user selection , '' _ arxiv:1303.7454 _ , 2013 .
[ online ] .
available : http://arxiv.org/abs/1303.7454 e. matskani , n. sidiropoulos , z .- q .
luo , and l. tassiulas , `` convex approximation techniques for joint multiuser downlink beamforming and admission control , '' _ ieee transactions on wireless communications _ , vol . 7 , no . 7 , pp .
26822693 , july 2008 .
v. lau , `` optimal downlink space - time scheduling design with convex utility functions - multiple - antenna systems with orthogonal spatial multiplexing , '' _ ieee transactions on vehicular technology _ ,
54 , no . 4 , pp .
13221333 , july 2005 .
l. cottatellucci , m. debbah , g. gallinaro , r. mueller , m. neri , and r. rinaldo , `` interference mitigation techniques for broadband satellite systems , '' in _ aiaa international communications satellite systems conference _
, june 2006 . m. kobayashi and g. caire , `` joint beamforming and scheduling for a multi - antenna downlink with imperfect transmitter channel knowledge , '' _
ieee journal on selected areas in communications _ ,
vol . 25 , no . 7 , pp . 14681477 , september 2007 .
x. zhang , e. jorswieck , b. ottersten , and a. paulraj , `` user selection schemes in multiple antenna broadcast channels with guaranteed performance , '' in _ ieee workshop on signal processing advances in wireless communications _ , june 2007 , pp .
15 . c. swannack , e. uysal - biyikoglu , and g. wornell , `` low complexity multiuser scheduling for maximizing throughput in the mimo broadcast channel , '' in _ allerton conf .
on commun . , control , and computing _ , 2004 .
tsai , c .- j .
chang , f .-
c . ren , and c .- m .
yen , `` adaptive radio resource allocation for downlink ofdma / sdma systems with multimedia traffic , '' _ ieee transactions on wireless communications _ , vol . 7 , no . 5 , pp .
17341743 , may 2008 . h. shirani - mehr , h. papadopoulos , s. a. ramprashad , and g. caire , `` joint scheduling and arq for mu - mimo downlink in the presence of inter - cell interference , '' _ ieee transactions on communications _
59 , no . 2 ,
pp . 578589 , 2011 .
d. hammarwall , m. bengtsson , and b. ottersten , `` beamforming and user selection in sdma systems utilizing channel statistics and instantaneous snr feedback , '' in _ ieee int .
conf . on acoustics , speech and signal processing _ ,
vol . 3 , april 2007 , pp . 113116 .
n. seifi , m. matthaiou , and m. viberg , `` coordinated user scheduling in the multi - cell mimo downlink , '' in _ ieee international conference on acoustics , speech and signal processing _ , may 2011 , pp .
28402843 .
h. huh , a. tulino , and g. caire , `` network mimo with linear zero - forcing beamforming : large system analysis , impact of channel estimation , and reduced - complexity scheduling , '' _ ieee transactions on information theory _ , vol .
58 , no . 5 , pp .
29112934 , may 2012 .
ku , l .- c .
wang , and y .- l .
liu , `` joint antenna beamforming , multiuser scheduling , and power allocation for hierarchical cellular systems , '' _ ieee journal on selected areas in communications _ , vol .
33 , no . 5 , pp .
896909 , may 2015 .
z. shen , r. chen , j. andrews , r. heath , and b. evans , `` low complexity user selection algorithms for multiuser mimo systems with block diagonalization , '' _ ieee transactions on signal processing _
54 , no . 9 , pp .
3658 3663 , sept .
2006 .
r. elliott and w. krzymien , `` downlink scheduling via genetic algorithms for multiuser single - carrier and multicarrier mimo systems with dirty paper coding , '' _ ieee transactions vehicular technology _ , vol .
58 , no . 7 , pp . 3247 3262 , sept . 2009 .
tran and e .- k .
hong , `` multiuser diversity for successive zero - forcing dirty paper coding : greedy scheduling algorithms and asymptotic performance analysis , '' _ ieee transactions on signal processing _ ,
58 , no . 6 , pp .
34113416 , june 2010 .
b. h. wang , h. t. hui , and y. t. yu , `` maximum volume criterion for user selection of multiuser mimo downlink with multiantenna users and block diagonalization beamforming , '' in _ ieee antennas and propagation society international symposium _ , july 2010 , pp . 14 .
o. souihli and t. ohtsuki , `` joint feedback and scheduling scheme for service - differentiated multiuser mimo systems , '' _ ieee transactions on wireless communications _
, vol . 9 , no . 2 , pp . 528533 , february 2010 .
k. jagannathan , s. borst , p. whiting , and e. modiano , `` scheduling of multi - antenna broadcast systems with heterogeneous users , '' _
ieee j. on s. areas in commun .
_ , vol . 25 , no . 7 , pp . 14241434 , sep . 2007 .
r. chen , r. heath , and j. andrews , `` transmit selection diversity for unitary precoded multiuser spatial multiplexing systems with linear receivers , '' _ ieee transactions on signal processing _
55 , no . 3 , pp .
11591171 , march 2007 .
r. chen , z. shen , j. andrews , and r. heath , `` multimode transmission for multiuser mimo systems with block diagonalization , '' _ ieee trans . on signal processing _
56 , no . 7 , pp .
32943302 , july 2008 .
s. sigdel and w. krzymien , `` simplified fair scheduling and antenna selection algorithms for multiuser mimo orthogonal space - division multiplexing downlink , '' _ ieee transactions on vehicular technology _ , vol .
58 , no . 3 , pp .
13291344 , march 2009 .
tran , m. bengtsson , and b. ottersten , `` iterative precoder design and user scheduling for block - diagonalized systems , '' _ ieee transactions on signal processing _ ,
60 , no . 7 ,
37263739 , july 2012 .
r. c. elliott , s. sigdel , and w. a. krzymien , `` low complexity greedy , genetic and hybrid user scheduling algorithms for multiuser mimo systems with successive zero - forcing , '' _ transactions on emerging telecommunications technologies _
23 , no . 7 , pp . 604617 , 2012 .
y. hei , x. li , k. yi , and h. yang , `` novel scheduling strategy for downlink multiuser mimo system : particle swarm optimization , '' _ science in china series f : information sciences _ , vol .
52 , no . 12 , pp . 22792289 , dec .
b. lim , w. krzymien , and c. schlegel , `` efficient sum rate maximization and resource allocation in block - diagonalized space - division multiplexing , '' _ ieee transactions on vehicular technology _ , vol .
58 , no . 1 ,
478484 , jan 2009 .
y. cheng , s. li , j. zhang , f. roemer , b. song , m. haardt , y. zhou , and m. dong , `` an efficient transmission strategy for the multicarrier multiuser mimo downlink , '' _ ieee transactions on vehicular technology _ , vol .
63 , no . 2 ,
628642 , feb 2014 .
j. nam , a. adhikary , j .- y .
ahn , and g. caire , `` joint spatial division and multiplexing : opportunistic beamforming , user grouping and simplified downlink scheduling , '' _ ieee journal of selected topics in signal processing _ , vol . 8 , no . 5 , pp . 876890 , oct 2014 .
a. adhikary , j. nam , j .- y .
ahn , and g. caire , `` joint spatial division and multiplexing : the large - scale array regime , '' _ ieee transactions on information theory _ , vol .
59 , no .
10 , pp . 64416463 , oct 2013 .
r. stridh , m. bengtsson , and b. ottersten , `` system evaluation of optimal downlink beamforming with congestion control in wireless communication , '' _ ieee transactions on wireless communications _ , vol . 5 , no . 4 , pp .
743751 , april 2006 .
m. kountouris , r. de francisco , d. gesbert , d. slock , and t. salzer , `` low complexity scheduling and beamforming for multiuser mimo systems , '' in _ ieee workshop on signal processing advances in wireless communications _
, july 2006 , pp . 15 .
d. hammarwall , m. bengtsson , and b. ottersten , `` utilizing the spatial information provided by channel norm feedback in sdma systems , '' _ ieee trans . on signal processing _ ,
56 , no . 7 , pp .
32783293 , july 2008 .
g. lee and y. sung , `` asymptotically optimal simple user scheduling for massive mimo downlink with two - stage beamforming , '' in _ ieee international workshop on signal processing advances in wireless communications _ ,
june 2014 , pp .
v. raghavan , r. heath , and a. sayeed , `` systematic codebook designs for quantized beamforming in correlated mimo channels , '' _ ieee journal on selected areas in communications _ , vol . 25 , no . 7 , pp . 12981310 , september 2007 .
v. raghavan , j. choi , and d. love , `` design guidelines for limited feedback in the spatially correlated broadcast channel , '' _ ieee transactions on communications _ , vol . 63 , no . 7 , pp .
25242540 , july 2015 .
m. schellmann , l. thiele , t. haustein , and v. jungnickel , `` spatial transmission mode switching in multiuser mimo - ofdm systems with user fairness , '' _ ieee transactions on vehicular technology _
59 , no . 1 ,
235247 , jan 2010 .
e. bjornson , d. hammarwall , and b. ottersten , `` exploiting quantized channel norm feedback through conditional statistics in arbitrarily correlated mimo systems , '' _ ieee transactions on signal processing _
57 , no .
10 , pp . 40274041 , oct 2009 .
e. bjornson , e. jorswieck , and b. ottersten , `` impact of spatial correlation and precoding design in ostbc mimo systems , '' _ ieee transactions on wireless communications _
, vol . 9 , no . 11 , pp . 35783589 , november 2010 .
b. godana and t. ekman , `` parametrization based limited feedback design for correlated mimo channels using new statistical models , '' _ ieee transactions on wireless communications _ , vol . 12 , no . 10 , pp . 51725184 , october 2013
. x. rao , l. ruan , and v. lau , `` limited feedback design for interference alignment on mimo interference networks with heterogeneous path loss and spatial correlations , '' _ ieee transactions on signal processing _
61 , no .
25982607 , may 2013 .
moon , c. lee , s .-
lee , and i. lee , `` joint user scheduling and adaptive intercell interference cancelation for miso downlink cellular systems , '' _ ieee transactions on vehicular technology _ , vol .
62 , no . 1 ,
172181 , jan 2013 .
j. wagner , y .- c .
liang , and r. zhang , `` on the balance of multiuser diversity and spatial multiplexing gain in random beamforming , '' _ ieee trans . on wireless commun .
_ , vol . 7 , no . 7 , pp .
25122525 , 2008 .
e. conte , s. tomasin , and n. benvenuto , `` a comparison of scheduling strategies for mimo broadcast channel with limited feedback on ofdm systems , '' _ eurasip journal on wireless communications and networking _ , vol .
2010 , no . 1 , p. 968703
, 2010 .
j. vicario , r. bosisio , c. antn - haro , and u. spagnolini , `` beam selection strategies for orthogonal random beamforming in sparse networks , '' _ ieee transactions on wireless communications _ , vol . 7 , no . 9 , pp .
33853396 , september 2008 .
m. kountouris , r. de francisco , d. gesbert , d. slock , and t. salzer , `` efficient metrics for scheduling in mimo broadcast channels with limited feedback , '' in _ ieee international conference on acoustics , speech and signal processing _ , vol . 3 , april 2007 , pp .
iii109iii112 .
m. kountouris , d. gesbert , and t. salzer , `` enhanced multiuser random beamforming : dealing with the not so large number of users case , '' _ ieee journal on selected areas in communications _ , vol .
26 , no . 8 , pp . 15361545 , october 2008 . m. kountouris , r. de francisco , d. gesbert , d. slock , and t. salzer , `` multiuser diversity - multiplexing tradeoff in mimo broadcast channels with limited feedback , '' in _ asilomar conference on signals , systems , and computers _ , oct 2006 , pp . 364368 .
b. khoshnevis , w. yu , and y. lostanlen , `` two - stage channel quantization for scheduling and beamforming in network mimo systems : feedback design and scaling laws , '' _ ieee journal on selected areas in communications _ , vol . 31 , no . 10 , pp . 20282042 , october 2013
. r. zakhour and d. gesbert , `` a two - stage approach to feedback design in multi - user mimo channels with limited channel state information , '' in _ ieee international symposium on personal , indoor and mobile radio communications _ , sept 2007 , pp .
chae , d. mazzarese , n. jindal , and r. heath , `` coordinated beamforming with limited feedback in the mimo broadcast channel , '' _ ieee journal on selected areas in communications _ , vol .
26 , no . 8 , pp . 15051515 , october 2008 . j. chen and v. lau , `` large deviation delay analysis of queue - aware multi - user mimo systems with two - timescale mobile - driven feedback , '' _ ieee transactions on signal processing _ ,
61 , no .
16 , pp . 40674076 , aug 2013 .
m. trivellato , f. boccardi , and h. huang , `` on transceiver design and channel quantization for downlink multiuser mimo systems with limited feedback , '' _ ieee journal on selected areas in communications _ , vol .
26 , no . 8 , pp . 14941504 , october 2008 .
m. wang , f. li , j. evans , and s. dey , `` dynamic multi - user mimo scheduling with limited feedback in lte - advanced , '' in _ ieee international symposium on personal indoor and mobile radio communications _ ,
sept 2012 , pp .
16271632 .
v. lau and y. k. r. kwok , _ channel - adaptive technologies and cross - layer designs for wireless systems with multiple antennas : theory and applications_.1em plus 0.5em minus 0.4emjohn wiley and sons , 2006 .
t. e. bogale , l. b. le , a. haghighat , and l. vandendorpe , `` on the number of rf chains and phase shifters , and scheduling design with hybrid analog - digital beamforming , '' _ ieee transactions on wireless communications _ ,
15 , no . 5 , pp .
33113326 , may 2016 .
t. rappaport , r. heath , r. daniels , and j. murdock , _
millimeter wave wireless communications _ , ser . prentice hall communications engineering and emerging technologies series from ted rappaport.1em plus 0.5em minus 0.4empearson education , 2014 .
r. w. heath , n. gonzlez - prelcic , s. rangan , w. roh , and a. m. sayeed , `` an overview of signal processing techniques for millimeter wave mimo systems , '' _ ieee journal of selected topics in signal processing _ ,
10 , no . 3 , pp .
436453 , april 2016 .
h. shokri - ghadikolaei , c. fischione , g. fodor , p. popovski , and m. zorzi , `` millimeter wave cellular networks : a mac layer perspective , '' _ ieee trans . on commun .
_ , vol .
63 , no .
10 , pp . 34373458 , oct 2015 .
e. jorswieck and h. boche , _ majorization and matrix - monotone functions in wireless communications_.1em plus 0.5em minus 0.4emfoundations and trends(r ) in communications and information theory , 2006 , vol
. 6 .
v. raghavan , j. h. kotecha , and a. sayeed , `` why does the kronecker model result in misleading capacity estimates ? '' _ ieee transactions on information theory _ , vol . 56 , no . 10 , pp . 48434864 , oct 2010 .
n. czink , b. bandemer , g. vazquez - vilar , l. jalloul , c. oestges , and a. paulraj , `` spatial separation of multi - user mimo channels , '' in _ ieee international symposium on personal , indoor and mobile radio communications _ , sept 2009 , pp .
10591063 .
p. almers , f. tufvesson , and a. molisch , `` keyhole effect in mimo wireless channels : measurements and theory , '' _ ieee transactions on wireless communications , _ , vol . 5 , no . 12 , pp .
35963604 , dec 2006 . p. arapoglou , k. liolis , m. bertinelli , a. panagopoulos , p. cottis , and r. de gaudenzi , `` mimo over satellite : a review , '' _ ieee communications surveys & tutorials _ ,
vol . 13 , no . 1 ,
2751 , first 2011 .
d. gesbert , s. hanly , h. huang , s. shamai shitz , o. simeone , and w. yu , `` multi - cell mimo cooperative networks : a new look at interference , '' _ ieee journal on selected areas in communications _
28 , no . 9 , pp . 13801408 , december 2010 .
d. lee , h. seo , b. clerckx , e. hardouin , d. mazzarese , s. nagata , and k. sayana , `` coordinated multipoint transmission and reception in lte - advanced : deployment scenarios and operational challenges , '' _ ieee commun . magazine _ , vol .
50 , no . 2 ,
pp . 148155 , february 2012 .
liu , y .- h .
dai , and z .- q .
luo , `` coordinated beamforming for miso interference channel : complexity analysis and efficient algorithms , '' _ ieee transactions on signal processing _ , vol .
59 , no . 3 , pp .
11421157 , march 2011 .
b. hochwald , c. peel , and a. swindlehurst , `` a vector - perturbation technique for near - capacity multiantenna multiuser communication - part ii : perturbation , '' _ ieee transactions on communications _ , vol .
53 , no . 3 , pp .
537544 , march 2005 .
e. jorswieck , e. larsson , and d. danev , `` complete characterization of the pareto boundary for the miso interference channel , '' _ ieee trans . on signal processing _ ,
56 , no .
10 , pp . 52925296 , oct 2008 .
m. hong and z .- q .
luo , `` academic press library in signal processing : communications and radar signal processing , '' s. theodoridis and r. chellappa , eds.1em plus 0.5em minus 0.4emelsevier science , 2013 , vol . 2 , ch .
8 - signal processing and optimal resource allocation for the interference channel .
e. bjornson , m. bengtsson , and b. ottersten , `` optimal multiuser transmit beamforming : a difficult problem with a simple solution structure , '' _ ieee signal processing magazine _ , vol .
31 , no . 4 , pp .
142148 , july 2014 .
p. c. weeraddana , m. codreanu , m. latva - aho , a. ephremides , and f. carlo , _ weighted sum - rate maximization in wireless networks : a review_.1em plus 0.5em minus 0.4emfoundations and trends ( r ) in networking , 2012 , vol .
s. christensen , r. agarwal , e. carvalho , and j. cioffi , `` weighted sum - rate maximization using weighted mmse for mimo - bc beamforming design , '' _ ieee transactions on wireless communications _
, vol . 7 , no . 12 , pp . 47924799 , december 2008 .
c. peel , b. hochwald , and a.l .
swindlehurst , `` a vector - perturbation technique for near - capacity multiantenna multiuser communication - part i : channel inversion and regularization , '' _ ieee transactions on communications _ , vol .
53 , no . 1 ,
195202 , jan .
2005 .
q. spencer , a. swindlehurst , and m. haardt , `` zero - forcing methods for downlink spatial multiplexing in multiuser mimo channels , '' _ ieee trans . on signal processing _ ,
52 , no . 2 ,
461471 , feb 2004 .
m. alodeh , s. chatzinotas , and b. ottersten , `` constructive multiuser interference in symbol level precoding for the miso downlink channel , '' _ ieee transactions on signal processing _ , vol .
63 , no . 9 , pp . 22392252 , may 2015 .
d. love , r. heath , and t. strohmer , `` grassmannian beamforming for multiple - input multiple - output wireless systems , '' _ ieee transactions on information theory _ ,
49 , no .
10 , pp . 27352747 , oct 2003 .
a. silva , r. holakouei , d. castanheira , a. gameiro , and r. dinis , `` a novel distributed power allocation scheme for coordinated multicell systems , '' _ eurasip j. wireless commun . and net .
_ , vol . 30 , 2013 .
s. kaviani , o. simeone , w. krzymien , and s. shamai , `` linear precoding and equalization for network mimo with partial cooperation , '' _ ieee trans . on veh .
_ , vol .
61 , no . 5 , pp .
20832096 , jun 2012 .
e. bjornson , r. zakhour , d. gesbert , and b. ottersten , `` cooperative multicell precoding : rate region characterization and distributed strategies with instantaneous and statistical csi , '' _ ieee transactions on signal processing _
58 , no . 8 , pp . 42984310 , 2010 .
b. e. godana and d. gesbert , `` egoistic vs. altruistic beamforming in multi - cell systems with feedback and back - haul delays , '' _ eurasip journal on wireless commun . and net .
2013 , no . 1 , p. 253
, 2013 .
y. huang , c. w. tan , and b. rao , `` joint beamforming and power control in coordinated multicell : max - min duality , effective network and large system transition , '' _ ieee transactions on wireless communications _ , vol .
12 , no . 6 , pp .
27302742 , june 2013 .
e. bjornson , j. hoydis , m. kountouris , and m. debbah , `` massive mimo systems with non - ideal hardware : energy efficiency , estimation , and capacity limits , '' _ ieee transactions on information theory _
60 , no . 11 ,
pp . 71127139 , nov 2014 .
o. e. ayach , s. rajagopal , s. abu - surra , z. pi , and r. w. heath , `` spatially sparse precoding in millimeter wave mimo systems , '' _ ieee transactions on wireless communications _ ,
vol . 13 , no . 3 , pp .
14991513 , march 2014 .
z. gao , l. dai , d. mi , z. wang , m. a. imran , and m. z. shakir , `` mmwave massive - mimo - based wireless backhaul for the 5 g ultra - dense network , '' _ ieee wireless commun .
_ , vol . 22 , no . 5 , pp .
1321 , oct .
2015 .
x. gao , l. dai , and a. m. sayeed , `` low rf - complexity technologies for 5 g millimeter - wave mimo systems with large antenna arrays , '' _ corr _ , jul 2016 .
[ online ] .
available : https://arxiv.org/abs/1607.04559v1 x. yu , j. c. shen , j. zhang , and k. b. letaief , `` alternating minimization algorithms for hybrid precoding in millimeter wave mimo systems , '' _ ieee journal of selected topics in signal processing _
10 , no . 3 , pp .
485500 , april 2016 .
a. de la oliva , j. a. hernandez , d. larrabeiti , and a. azcorra , `` an overview of the cpri specification and its application to c - ran - based lte scenarios , '' _ ieee communications magazine _ , vol .
54 , no . 2 ,
pp . 152159 , february 2016 .
e. castaeda , a. silva , and a. gameiro , `` contemporary issues in wireless communications , '' m. khatib , ed.1em plus 0.5em minus 0.4emintech , 2014 , ch .
user selection and precoding techniques for rate maximization in broadcast miso systems .
a. manolakos , y. noam , and a. goldsmith , `` null space learning in cooperative mimo cellular networks using interference feedback , '' _ ieee transactions on wireless communications _ ,
14 , no . 7 , pp .
39613977 , july 2015 .
d. park , h. seo , h. kwon , and b. g. lee , `` wireless packet scheduling based on the cumulative distribution function of user transmission rates , '' _ ieee transactions on communications _
53 , no . 11 , pp . 19191929 , nov 2005 .
d. bartolome and a. perez - neira , `` spatial scheduling in multiuser wireless systems : from power allocation to admission control , '' _ ieee transactions on wireless communications _ ,
vol . 5 , no . 8 ,
20822091 , aug 2006 .
park , h. park , h. kong , and i. lee , `` new beamforming techniques based on virtual sinr maximization for coordinated multi - cell transmission , '' _ ieee transactions on wireless communications _ ,
11 , no . 3 , pp .
10341044 , march 2012 .
d. lpez - prez , m. ding , h. claussen , and a. h. jafari , `` towards 1 gbps / ue in cellular systems : understanding ultra - dense small cell deployments , '' _ ieee communications surveys & tutorials _ ,
17 , no . 4 , pp . 20782101 , fourthquarter 2015 .
d. feng , c. jiang , g. lim , l. j. cimini , g. feng , and g. y. li , `` a survey of energy - efficient wireless communications , '' _ ieee communications surveys & tutorials _ ,
15 , no . 1 ,
167178 , first 2013 . h. mahdavi - doost , m. ebrahimi , and a. khandani , `` characterization of sinr region for interfering links with constrained power , '' _ ieee trans . on inf .
56 , no . 6 , pp . 2816 2828 , june 2010 .
e. castaeda , r. samano - robles , and a. gameiro , `` sum rate maximization via joint scheduling and link adaptation for interference - coupled wireless systems , '' _ eurasip journal on wireless communications and networking _ , vol .
2013 , no . 1 , p. 268
, 2013 .
l. georgiadis , m. j. neely , and l. tassiulas , _ resource allocation and cross - layer control in wireless networks _ , ser . foundations and trends(r ) in networking.1em plus 0.5em minus 0.4emnow publishers incorporated , 2006 .
r. baldemair , t. irnich , k. balachandran , e. dahlman , g. mildh , y. seln , s. parkvall , m. meyer , and a. osseiran , `` ultra - dense networks in millimeter - wave frequencies , '' _ ieee communications magazine _ , vol .
53 , no . 1 ,
pp . 202208 , january 2015 .
j. li , e. bjornson , t. svensson , t. eriksson , and m. debbah , `` joint precoding and load balancing optimization for energy - efficient heterogeneous networks , '' _ ieee transactions on wireless communications _ , vol .
14 , no .
10 , pp . 58105822 , oct 2015 .
d. liu , l. wang , y. chen , m. elkashlan , k. k. wong , r. schober , and l. hanzo , `` user association in 5 g networks : a survey and an outlook , '' _ ieee communications surveys & tutorials _
18 , no . 2 ,
10181044 , secondquarter 2016 . h. t. cheng and w. zhuang , `` an optimization framework for balancing throughput and fairness in wireless networks with qos support , '' _ ieee trans . on wireless commun .
_ , vol . 7 , no . 2 , pp .
584593 , feb .
a. b. sediq , r. h. gohary , r. schoenen , and h. yanikomeroglu , `` optimal tradeoff between sum - rate efficiency and jain s fairness index in resource allocation , '' _ ieee transactions on wireless communications _ , vol .
12 , no . 7 , pp .
34963509 , july 2013 .
a. b. sediq , r. schoenen , h. yanikomeroglu , and g. senarath , `` optimized distributed inter - cell interference coordination ( icic ) scheme using projected subgradient and network flow optimization , '' _ ieee trans .
on commun .
_ , vol .
63 , no . 1 ,
107124 , jan 2015 .
e. bjornson , l. sanguinetti , and m. kountouris , `` deploying dense networks for maximal energy efficiency : small cells meet massive mimo , '' _ ieee journal on selected areas in communications _ , vol .
34 , no . 4 , pp . 832847 , april 2016 .
e. bjornson , e. a. jorswieck , m. debbah , and b. ottersten , `` multiobjective signal processing optimization : the way to balance conflicting metrics in 5 g systems , '' _ ieee signal processing magazine _ , vol . 31 , no . 6 , pp .
1423 , nov 2014 .
a. sharifian , r. schoenen , and h. yanikomeroglu , `` joint realtime and nonrealtime flows packet scheduling and resource block allocation in wireless ofdma networks , '' _ ieee transactions on vehicular technology _ ,
65 , no . 4 , pp . 25892607 , april 2016 .
e. bjornson , m. kountouris , m. bengtsson , and b. ottersten , `` receive combining vs. multi - stream multiplexing in downlink systems with multi - antenna users , '' _ ieee transactions on signal processing _
61 , no . 13 , pp .
34313446 , july 2013 . z. zhang , k. long , j. wang , and f. dressler , `` on swarm intelligence inspired self - organized networking : its bionic mechanisms , designing principles and optimization approaches , '' _ ieee communications surveys & tutorials _
16 , no . 1 ,
513537 , first 2014 . l. h. d. sampaio , m. h. adaniya , m. de paula marques , p. j. e. jeszensky , and t. abro , `` ant colony optimization for resource allocation and anomaly detection in communication networks , '' t. abro , ed.1em plus 0.5em minus 0.4emintech publisher , 2013 .
g. athanasiou , p. c. weeraddana , c. fischione , and l. tassiulas ,
`` optimizing client association for load balancing and fairness in millimeter - wave wireless networks , '' _ ieee / acm transactions on networking _ , vol .
23 , no . 3 , pp .
836850 , june 2015 .
d. bethanabhotla , o. y. bursalioglu , h. c. papadopoulos , and g. caire , `` optimal user - cell association for massive mimo wireless networks , '' _ ieee transactions on wireless communications _ , vol .
15 , no . 3 , pp .
18351850 , march 2016 .
e. castaeda , a. silva , r. samano - robles , and a. gameiro , `` low - complexity user selection for rate maximization in mimo broadcast channels with downlink beamforming , '' _ the scientific world journal _ , pp . 113 , 2013 .
e. pateromichelakis , m. shariat , a. u. quddus , and r. tafazolli , `` on the evolution of multi - cell scheduling in 3gpp lte / lte - a , '' _ ieee communications surveys & tutorials _ , vol .
15 , no . 2 ,
pp . 701717 , 2013 .
f. rusek , d. persson , b. k. lau , e. larsson , t. marzetta , o. edfors , and f. tufvesson , `` scaling up mimo : opportunities and challenges with very large arrays , '' _ ieee signal processing magazine _ , vol .
30 , no . 1 ,
4060 , jan 2013 .
n. abu - ali , a. e. m. taha , m. salah , and h. hassanein , `` uplink scheduling in lte and lte - advanced : tutorial , survey and evaluation framework , '' _ ieee communications surveys & tutorials _ , vol .
16 , no . 3 , pp .
12391265 , third 2014 .
j. lee and n. jindal , `` high snr analysis for mimo broadcast channels : dirty paper coding versus linear precoding , '' _ ieee transactions on information theory _ , vol .
53 , no . 12 , pp .
47874792 , dec 2007 .
y. kim , h. ji , j. lee , y .- h .
nam , b. l. ng , i. tzanidis , y. li , and j. zhang , `` full dimension mimo ( fd - mimo ) : the next evolution of mimo in lte systems , '' _ ieee wireless communications _ , vol .
21 , no . 2 ,
pp . 2633 , april 2014 .
a. puglielli , a. townley , g. lacaille , v. milovanovi , p. lu , k. trotskovsky , a. whitcombe , n. narevsky , g. wright , t. courtade , e. alon , b. nikoli , and a. m. niknejad , `` design of energy- and cost - efficient massive mimo arrays , '' _ proceedings of the ieee _
104 , no . 3 , pp .
586606 , march 2016 .
j. vieira , s. malkowsky , k. nieman , z. miers , n. kundargi , l. liu , i. wong , v. wall , o. edfors , and f. tufvesson , `` a flexible 100-antenna testbed for massive mimo , '' in _ ieee globecom workshop _
, dec 2014 , pp . 287293 .
e. bjrnson and e. g. larsson , `` three practical aspects of massive mimo : intermittent user activity , pilot synchronism , and asymmetric deployment , '' _ corr _ , 2015 .
[ online ] .
available : http://arxiv.org/abs/1509.06517 t. al - naffouri , m. sharif , and b. hassibi , `` how much does transmit correlation affect the sum - rate scaling of mimo gaussian broadcast channels ? '' _ ieee transactions on communications _
57 , no . 2 ,
pp . 562572 , february 2009 . | remarkable research activities and major advances have been occurred over the past decade in multiuser multiple - input multiple - output ( mu - mimo ) systems . several transmission technologies and precoding techniques
have been developed in order to exploit the spatial dimension so that simultaneous transmission of independent data streams reuse the same radio resources .
the achievable performance of such techniques heavily depends on the channel characteristics of the selected users , the amount of channel knowledge , and how efficiently interference is mitigated . in systems where the total number of receivers is larger than the number of total transmit antennas , user selection becomes a key approach to benefit from multiuser diversity and achieve full multiplexing gain .
the overall performance of mu - mimo systems is a complex joint multi - objective optimization problem since many variables and parameters have to be optimized , including the number of users , the number of antennas , spatial signaling , rate and power allocation , and transmission technique .
the objective of this literature survey is to provide a comprehensive overview of the various methodologies used to approach the aforementioned joint optimization task in the downlink of mu - mimo communication systems .
downlink transmission , multi - user mimo , precoding , resource allocation , spatial multiplexing , user scheduling . |
we present the results of a search for variable stars in the faint sparse globular cluster e3 , located at @xmath2 , @xmath3 .
the cluster was discovered on the eso b schmidt survey of the southern sky by lauberts ( 1976 ) .
the first @xmath4 photometry of the cluster was presented by van den bergh , demers & kunkel ( 1980 ) .
numerous candidates for blue stragglers were identified .
the photoelectric photometry of frogel & twarog ( 1983 ) confirmed this finding . a subsequent study by hesser et al .
( 1984 ) using @xmath5 photoelectric and photographic observations showed a sparsely populated subgiant branch in the color - magnitude diagram .
the first ccd photometry for e3 in the @xmath4 bands , obtained with the ctio 4 m telescope , was published by mcclure et al .
these observations suggested the presence of a second sequence of stars @xmath6 mag . above the cluster main sequence , interpreted as evidence for a significant population of binary stars in e3 .
the cluster was further studied by gratton & ortolani ( 1987 ) , who provided new @xmath4 ccd photometry from the 2.2 m telescope at eso .
the data for this project was obtained with the las campanas observatory 1.0 m swope telescope during two separate runs : from april 11 to 21 , 1996 and from may 16 to 27 , 1996 .
for the first few nights of the april run the telescope was equipped with the tek1 1024x1024 ccd camera giving a pixel scale of @xmath7 . on the night of april 14
the camera was switched to the tek3 2048x2048 ccd with a pixel scale of @xmath8 . during
the may run the ford 2048x2048 ccd camera with a pixel scale of @xmath9 was used .
the main observing target on both runs was the m4 globular cluster .
several exposures of e3 were taken at the beginning of most nights .
a total of 121 long ( @xmath10 ) exposures were taken in the @xmath11 filter ( 33 , 42 and 46 with tek1 , tek3 and ford , respectively ) , six short ( @xmath12 ) exposures in @xmath11 ( 2 with tek1 , 4 with ford ) , two @xmath13 exposures in @xmath14 ( tek1 ) and two @xmath15 exposures in @xmath16 ( tek1 ) .
the preliminary processing of the ccd frames was done with the standard routines in the iraf - ccdproc package .
the images from the tek3 camera were clipped to a size of 1024x1024 pixels@xmath17 to cover roughly the same field as the tek1 images . due to the high degree of psf variability on the images taken with the ford camera ,
only the central 800x800 pixel@xmath17 sections were used .
photometry was extracted using the _ daophot / allstar _ package ( stetson 1987 ) .
a psf varying linearly with the position on the frame was used .
the psf was modeled with a moffat function .
stars were identified using the find subroutine and aperture photometry was measured with the phot subroutine .
approximately 40 bright isolated stars were initially chosen by _
daophot _ for the construction of the psf . of those the stars with profile errors greater than twice the average were rejected and the psf was recomputed .
this procedure was repeated until no such stars were left on the list .
the psf was then further refined on frames with all but the psf stars subtracted .
this procedure was applied twice .
the psf obtained in the above method was then used by _ allstar _ in profile photometry .
the image where the most stars were identified was chosen as the template .
the template star list was then transformed to the @xmath18 coordinate system of each of the frames and used as input to _
allstar _ in the fixed - position mode .
the output profile photometry was transformed to the common instrumental system of the template image and then combined into a database .
the databases were created for the long ( @xmath19 ) exposures in the @xmath11 filter only .
we have followed the procedure for selecting variables given in kaluzny et al .
( 1998 ) , where it is described in detail . from the 1541 stars in the @xmath11 database 11 variable star candidates were selected .
after the rejection of stars with noisy and/or chaotic light curves we were left with two variables .
their periods were refined using the analysis of variance method , as described by schwarzenberg - czerny ( 1989 ) .
these two variables were confirmed with isis - the image subtraction package ( alard & lupton 1998 , alard 1999 ) .
no other variables were detected using this method . in figure
[ fig : var ] we present the phased @xmath11 light curves of the two variables . table [ tab : var ] lists the parameters of these variables : name , period , @xmath11 magnitude ( @xmath20 for the pulsating variable , @xmath21 for the eclipsing binary ) , the @xmath22 and @xmath23 colors .
the variables are indicated by open circles on the finding chart in figure [ fig : map ] .
v1 is a pulsating variable , most likely of the sx phe type , judging from its short period ( 0.0853 days ) and the shape of its light curve .
v2 is an eclipsing binary with a period of 0.4490 days .
its light curve shows an absence of the constant light phase , indicating that it is a w uma type variable .
we have used the period - luminosity calibration for sx phe stars derived by mcnamara ( 1997 ) to estimate the distance modulus to v1 : @xmath24 adopting a value of reddening e(b - v)=0.30 ( harris 1996 ) we obtain a distance modulus @xmath25 .
this value is in agreement with the distance moduli found in literature : 14.55 - van den bergh et al .
( 1980 ) , 14.4 - frogel & twarog ( 1983 ) , 14.2 - gratton & ortolani ( 1987 ) , indicating that v1 is located at the same distance as the cluster .
the following period - color - luminosity calibrations for w uma type eclipsing binaries derived by rucinski ( 2000 ) were applied to v2 : @xmath26 @xmath27 the fact that @xmath22 and @xmath23 colors were determined at random phase should not influence the outcome substantially , as in the case of contact binaries the color does not change significantly throughout the cycle . using a value of @xmath28 ( schlegel et al .
1997 ) we obtained a distance modulus of 15.42 mag . from the first calibration and 14.83 mag . from the second .
the variable appears to be located behind the cluster , in the milky way halo .
.variables in e3 + [ cols="^,^,^,^,^,<",options="header " , ] [ tab : var ]
to construct the color - magnitude diagrams we combined pairs of long exposures in the @xmath29 filters . the transformation from instrumental magnitudes to the standard system
was derived from the observations of the landolt fields ( landolt 1992 ) .
the following relations were adopted : @xmath30 we have compared our @xmath4 photometry with that of mcclure et al .
the average differences in @xmath11 magnitude and @xmath22 color were computed for 6 selected stars in the range @xmath31 and were found to be @xmath32 and @xmath33 .
the colors and magnitudes for variable stars were determined following a different procedure . for variable v1
its average magnitude @xmath34 was used to place it on the cmds .
v2 was plotted with its magnitude outside of the eclipses @xmath21 . to derive the colors we used the single @xmath16 and @xmath14 exposures and interpolated the @xmath11 magnitudes from the nearest exposures in @xmath11 to those epochs .
the final values of @xmath22 and @xmath23 were taken as the average of two color determinations , with the average scatter of 0.04 mag .
the resultant @xmath0 and @xmath1 color - magnitude diagrams are shown in figure [ fig : cmd ] , with the variable v1 denoted by an open circle and v2 by an open square . only stars within @xmath35 of the cluster center are plotted .
both variables are located among candidate blue stragglers , although v2 appears to be located behind the cluster , based on the distance modulus determination in the previous section .
the cluster main sequence is apparent in both diagrams .
it exhibits considerable scatter and there is some indication of a second sequence running above it , although not as clear as in figure 3 of mcclure et al .
this would indicate that e3 could possess a significant population of binary stars .
this is in agreement with the idea proposed by van den bergh et al .
( 1980 ) that severe tidal stripping had depleted the cluster in single stars , leading to an increased binary frequency .
a gap in the main sequence near the turnoff , at @xmath36 is visible in both cmds .
this has been previously noted by mcclure et al .
( 1985 ) and shown to be more of a visual effect , as no significant discontinuities are present in the cumulative luminosity function for stars on the main sequence ( figure 6 therein ) .
this result is confirmed by our analysis . the subgiant and lower giant branches are also discernible in the diagrams , although they show substantial scatter .
this is regarded as a real feature of the cluster , as commented in literature ( i.e. hesser et al .
a number of stars blueward of the turnoff are present , possibly blue stragglers belonging to the cluster , as first noted by van den bergh et al .
our variability search in e3 resulted in the discovery of two variable stars : an sx phe variable ( v1 ) and a w uma eclipsing binary ( v2 ) .
we have applied period - luminosity and period - color - luminosity relations to the variables to obtain their distance moduli .
v1 seems to be a blue straggler belonging to e3 , based on its distance modulus and location on the cmd .
v2 is probably located behind the cluster .
alard , c. 1999 , preprint ( astro - ph/9903111 ) alard , c. , lupton , r. h. 1998 , apj , 503 , 325 frogel , j.a .
, twarog , b.a .
1983 , apj , 274 , 270 gratton , r.g . ,
ortolani , s. 1987 , a&as , 67 , 373 harris , w. e. 1996 , aj , 112 , 1487 hesser , j .
e. , mcclure , r. d. , hawarden , t. g. , cannon , r. d. , von rudloff , r. , kruger , b. , egles , d. 1984 , pasp , 96 , 406 kaluzny , j. , stanek , k. z. , krockenberger , m. , sasselov , d. d. , tonry , j. l. , & mateo , m. 1998 , aj , 115 , 1016 landolt , a. 1992 , aj , 104 , 340 lauberts , a. 1976 , a&a , 52 , 309 mcclure , r. d. , hesser , j. e. , stetson , p. b. , stryker , l. l. 1985 , pasp , 97 , 665 mcnamara , d. h. 1997 , pasp , 109 , 1221 rucinski , s. m. 2000 , preprint ( astro - ph/0001152 ) schlegel , d. j. , finkbeiner , d. p. , davis , m. 1997 , aas , 191 , 8704 schwarzenberg - czerny , a. 1989 , mnras 253 , 198 stetson , p. b. 1987 , pasp , 99 , 191 van den bergh , s. , demers , s. , kunkel , w. e. 1980 , apj , 239 , 112 | we present the results of a search for variable stars in the faint sparse globular cluster e3 . we have found two variable stars : an sx phe variable ( v1 ) and a w uma eclipsing binary ( v2 ) .
we have applied period - luminosity and period - color - luminosity relations to the variables to obtain their distance moduli .
v1 seems to be a blue straggler belonging to e3 , based on its distance modulus and location on the cmd .
v2 is probably located behind the cluster , in the milky way halo .
we also present @xmath0 and @xmath1 color magnitude diagrams of e3 . |
following the suggestion of @xcite and @xcite several teams have carried out microlensing surveys toward the galactic bulge ( gb ) . to date , well over @xmath10 microlensing events in the gb have been detected by those groups : ogle ( @xcite ) , moa ( @xcite ) , macho ( @xcite ) and eros ( @xcite ) .
thousands of detections are expected in the upcoming years from moa and ogle - iii .
it is now well understood that these observations are useful for studying the structure , kinematics and dynamics of the galaxy , and the stellar mass function , as the event rate and time - scale distributions are related to the masses and velocities of lens objects .
the magnification of a microlensing event is described by ( @xcite ) @xmath11 where @xmath12 is the projected separation of the source and lens in units of the einstein radius @xmath13 which is given by @xmath14 where @xmath15 is the lens mass , @xmath16 is the normalized lens distance and @xmath17 and @xmath18 are the observer - lens and the observer - source distances .
the time variation of @xmath19 is @xmath20 where @xmath21 , @xmath22 , @xmath23 and @xmath24 are , respectively , the minimum impact parameter in units of @xmath13 , the time of maximum magnification , the event time - scale and the transverse velocity of the lens relative to the line of sight toward the source star . from light curve alone , one can determine the values of @xmath21 , @xmath22 and @xmath25 , but not the values of @xmath15 , @xmath26 or @xmath24 .
microlensing optical depth ( the total cross section for microlensing ) is directly related to the mass density of compact objects along the line of sight ( @xcite ) . however ,
previous results have been controversial . @xcite and @xcite first predicted the optical depth of @xmath27 , assuming that all events were associated with known disk stars .
after the first several bulge events were reported by ogle ( @xcite ) , the high event rate prompted @xcite to evaluate the contribution of bulge stars in addition to the disk stars .
they estimated @xmath28 and concluded that the value could be about twice as large , if the bulge were elongated along the line of sight .
nevertheless , the first measurements of the optical depth , @xmath29 by ogle ( @xcite ) and @xmath30 by macho ( @xcite ) , were well above the predictions .
recent studies based on difference image analysis ( dia)less sensitive to the systematics of blending in crowded fields also found large optical depths : @xmath31 at @xmath32 from 99 events by macho ( @xcite ) and @xmath33 at @xmath34 from 28 events by moa ( @xcite ) .
the latter values were adjusted for the presence of the foreground disk stars , and the uncorrected measurements are considerably lower , i.e. @xmath35 and @xmath36 , respectively . to explain high optical depths
a number of authors have suggested the presence of a bar oriented along our line of sight to the gb ( @xcite ; @xcite ) , and have adopted various values of the bar orientation and mass ( @xcite ; @xcite ; @xcite ; @xcite ) .
the resulting values are in the range @xmath37 . @xcite
have shown that high optical depth measurements available at the time could not be easily reconciled with our general understanding of the galactic dynamics , and that the standard models of the galaxy would need to be revised .
@xcite raised the possibility of a systematic bias in the optical depth measurement due to the difficulties of measuring @xmath25 associated with merging unresolved sources .
when the actual source base - line flux is unknown , @xmath25 and @xmath21 are degenerate in relatively low signal - to - noise ratio ( s / n ) events ( c.f .
@xcite ; @xcite ; @xcite ; @xcite ) .
@xcite postulated that optical depth may be estimated without a bias due to blending by using only events with bright source stars , such as red clump giants ( rcg ) , in which blending is assumed to be negligible .
although the first measurements by @xcite gave a high value @xmath30 , the recent measurements based on events with rcg sources have returned lower optical depths : @xmath38 at @xmath39 from 50 events by macho ( @xcite ) , @xmath40 at @xmath41 from 16 events by eros ( @xcite ) and @xmath42 at @xmath43 from 42 events by macho ( @xcite ) . in this paper
we present a measurement of the microlensing optical depth in the direction of the gb based on the full 4-year extent of the ogle - ii monitoring data .
for analysis we select only high s / n events with bright apparent source stars in the rcg region . in order to better understand the systematics of this complex measurement , we proceed without assuming that blending is negligible . in the first part of the paper we present the photometric data ( [ sec : data ] ) ,
selection of microlensing events ( [ sec : select ] ) , computation of the detection efficiency ( [ sec : eff ] ) and estimation of the optical depth ( [ sec : opt ] ) .
this is followed , in the second part ( [ sec : disc ] and [ sec : summary ] ) , by a description of various cross checks and general discussion .
we primarily examine source confusion and the effect of reintroducing the assumption of negligible blending in our bright event sample .
the data used in this analysis were collected between 1997 and 2000 , during the second phase of the ogle experiment .
all observations were made with the 1.3-m warsaw telescope located at the las campanas observatory , chile .
the observatory is operated by the carnegie institution of washington .
the `` first generation '' camera has a site 2048 @xmath44 2048 pixel ccd detector with pixel size of 24 @xmath45 m resulting in 0.417 arcsec / pixel scale .
images of the gb fields were taken in drift - scan mode at `` medium '' readout speed with the gain 7.1 @xmath46/adu and readout noise of 6.3 @xmath46 .
a single 2048 @xmath44 8192 pixel frame covers an area of 0.24 @xmath44 0.95 deg@xmath1 .
saturation level is about 55,000 adu .
details of the instrumentation setup can be found in @xcite . in this paper
we use @xmath47-band images for the 20 central ogle - ii fields in the gb . between 138 and 555 frames
were available in each field .
the remaining 29 fields are either harder to model ( because of a prominent disk component ) , or can not be treated as co - located with the rest and averaged .
the event number statistics in those fields is not sufficient for an independent determination of @xmath9 .
a reliable map of the microlensing optical depth must await the full analysis of the ogle - iii data .
the centers of the analyzed fields are listed in table [ tbl : fld ] .
the time baseline of the survey is almost 4 years .
there are gaps between the observing seasons when gb can not be observed from earth , each about 3 months long .
the median seeing was @xmath48 .
astrometric and photometric scales are defined by multi - color maps of @xcite , comprising positions and @xmath49-band photometry of @xmath50 stars in all 49 gb fields .
photometric zero points are accurate to about 0.04 mag .
we adopt a hybrid approach to photometry with source detection and centroiding on reference images using dophot package ( @xcite ) , combined with the dia photometry to lower the point - to - point scatter ( @xcite ; @xcite ; @xcite ) . a high s / n reference image for each field was taken to be a mean of 20 frames with the best overall seeing and background .
to obtain a difference frame we convolve the reference image with the psf matching kernel and subtract the result from a given survey frame , after interpolating to the same pixel grid .
psf photometry on difference frames is performed only for objects detected on the reference image using fixed positions found by dophot .
criteria for selecting candidate microlensing events are summarized in table [ tbl : criteria ] . for the optical depth analysis
we select only those microlensing events that occurred in bright stars .
red clump giants are sufficiently numerous for acceptable counting statistics , but not too numerous .
their combined seeing disks still cover a small fraction of the survey area .
therefore , rcgs should be less affected by source confusion with stars @xmath51 mag deeper in the luminosity function ( lf ) compared to fainter stars with higher number density .
note that by itself this does not guarantee that we can entirely neglect blending in our samples .
we discuss this issue in
[ sec : level6 ] and [ sec : blending ] . only stars with more than 70 data points are considered . in the color magnitude diagram ( cmd )
we define the extended red clump giant ( rcg ) region using a condition : @xmath52 and @xmath53 ( fig .
[ fig : cmd ] ) . extinction corrected magnitudes and colors , i.e. @xmath54 and @xmath55 , are obtained for each star using the ogle - ii reddening map of the gb ( @xcite ) .
the magnitude threshold is set at @xmath56 , where @xmath57 is the mean magnitude of the rcg population in a given field .
@xmath57 is estimated according to the prescription in @xcite and accounts for the galactic bar geometry .
this procedure is very effective in rejecting most of the disk main sequence .
in contrast with previous analyses using rcg sources ( @xcite ) , our extended rcg region includes not only rcgs , but also red giants and red supper giants .
the total number of sources selected in all 20 fields was 1,084,267 ( level 0 criteria in table [ tbl : criteria ] ) .
level 1 criteria detect generic flux brightening in otherwise quiet objects , followed by microlensing model fitting at level 2 ( table [ tbl : criteria ] ) .
we define a `` peak '' as a group of consecutive data points in the light curve that includes at least 5 data points with significance @xmath58 .
at least 3 of those points should have @xmath59 .
counting of consecutive data points allows for one failure , i.e. @xmath60 between data points .
the baseline flux and the significance of each photometric measurement with respect to that baseline are established within a centered running window spanning @xmath61 d. we adopted : @xmath62 where @xmath63 and @xmath64 are the flux and flux error of the @xmath65-th data point . @xmath66 and @xmath67 are the median and variance of all data points outside the window .
there are further conditions on the number of detected peaks ( @xmath68 ) , the maximum significance level reached by any point in any of those peaks ( @xmath69 ) , and the integrated significance of the peak where @xmath69 occurred ( @xmath70 ) .
full details are in table [ tbl : criteria ] . for lower s / n peaks ( @xmath71 ) we allow less noise in the baseline and limit the reduced @xmath72 : @xmath73 .
the non - linear model fitting is performed using the minuit package ( @xcite ) which provides the best fit parameters and corresponding 68% confidence intervals ( for 1 parameter of interest ) .
we fit a single lens microlensing model in the form : @xmath74 where @xmath75 is given in equations [ eq : amp - u][eq : u ] , @xmath76 and @xmath77 are the baseline fluxes of the source star and the blended background star , respectively .
table [ tbl : criteria ] lists all requirements on the quality of the best fit model @xmath78 and on the parameters .
long period variables and novae in our data set tend to have unusually large @xmath25 or @xmath79 .
none of the 135 candidates that remain at this point are likely to be a variable star in one of those categories . in fig .
[ fig : lcte400 ] we show light curves of two objects which failed the @xmath25 cut .
psf photometry on difference frames is performed assuming a constant background at zero counts .
the psf matching kernel does not handle the extended wings of very bright stars due to lack of signal there .
this is why light curves of some stars follow the intrinsic or seeing induced variability of their much brighter neighbors ( @xcite ) .
the effect is further exacerbated by the presence of spurious objects around bright stars in dophot output .
such artifacts around microlensed stars and other variables ( primarily novae ) have to be rejected from the list of microlensing events .
this is accomplished by cross - referencing the list of positions for our bright candidate events with an analogous list constructed for events of any brightness ( both after level 02 selection only ) . in case
there is a `` sibling '' event within 20 pixels that occurred at a similar time @xmath80 d , the one with lower @xmath69 is rejected from its respective list .
most of the 54 rejected light curves are associated with a possible nova which appeared in bul_sc4 field around hjd=2450614.79 d. stars as far as 100 pixels around it were affected .
several rejected artifacts are associated with microlensing events . at level 3 ( table [ tbl : criteria ] )
our automated criteria select 81 candidate microlensing events .
all light curves in that group were also selected as microlensing events by visual inspection .
up to this point , 41 candidates flagged as microlensing by eye were rejected because of low s / n or @xmath79 .
note that strongly degenerate events can fail if they happen to have the best fit model with @xmath79 .
out of the 81 remaining events , 69 have estimates of errors in all parameters from minuit . in practice ,
i.e. with any decent initial guess , the only reason for minuit error estimates not to converge is a very wide and shallow @xmath72 surface .
events with such degenerate models have unreliable best fit parameters and are not used in the optical depth measurement ( level 4 , table [ tbl : criteria ] ) .
furthermore , we required @xmath81 ( level 5 selection in table [ tbl : criteria ] ) .
this rejects three events with a low amplitude variability visible in the baseline and caused by the nearby bright variable stars or by differential refraction .
the requirement on the goodness of fit prevents such variability from distorting the parameters of the fitted model .
two binary lens events are rejected by the @xmath82 cut from the previous paragraph .
some binary microlensing events show only weak deviations with respect to the basic case .
the detectability of such events is sensitive to the details of the temporal sampling and coverage .
the deviations that were missed may still significantly distort the results of the single lens model fitting .
we cross - referenced our event sample with other published samples from macho ( @xcite ) , moa ( @xcite ) , eros ( @xcite , and the eros alert system ) , and ogle ( @xcite ) .
we found two more binaries in addition to the ones that failed the @xmath82 condition : sc21 - 45456 ( 97-blg-1 ; @xcite ) , in which only the declining part of the light curve is covered by the ogle data , and sc20 - 69068 ( sc20 - 1793 ; @xcite ) , where the observed deviations are small .
all four binary lens candidates and their corresponding i d cross - references are listed in table [ tbl : binary ] .
a complete list of cross - references for our candidate events is available electronically on apj - web . in case of binary lens events
it is difficult to find reliable lensing parameters like @xmath25 and @xmath83 without very good coverage .
unfortunately , that is the case for most of our candidates .
furthermore , it is a complicated task to estimate the detection efficiency for binary microlensing .
the fraction of binary lens events is around 38% ( @xcite ) . in [ sec : opt ] this fraction is used to account for the optical depth contribution from binary lens events .
this avoids estimating @xmath9 due to individual binary lenses , which is subject to large uncertainties in our sample .
one of the events , namely sc33 - 553617 ( sc33_4505 in @xcite ) , is a strong parallax event , and is included in the optical depth measurement in [ sec : opt ] .
none of the remaining candidate parallax events listed in @xcite are part of our sample , including possible parallax events with weak signatures . in this section
we look for evidence of blending using only information from light curve fitting .
model degeneracy is known to be a severe problem ( @xcite ; @xcite ; @xcite ; @xcite ) .
it was therefore of critical importance , that all 62 events in our level 5 sample have exceptionally high s / n ratio and very good coverage .
this allowed us to make use of the fitted lensed flux fraction @xmath84 and its error .
[ fig : isitotal ] shows the resulting relation between the apparent magnitude of the total light ( @xmath85 ) and the best fit source magnitude ( @xmath86 ) , both corrected for extinction . considering a positive error range
@xmath87 we can construct an upper limit for the source magnitude as @xmath88 .
this leads to an unexpected conclusion that , with high probability , in 29 out of 62 events the source may actually be below @xmath89 and does not satisfy level 0 conditions .
in other words , the actual source in such an event merged with a much brighter star to make a single psf peak detected by the dophot code .
for such events the detection efficiency can not be estimated assuming one possible lensed star per database light curve .
we remove those 29 events ( level 6 , table [ tbl : criteria ] ) and for the optical depth estimate we use only the remaining 33 high quality events with low probability of having @xmath90 .
there are 32 unique events because sc31 - 111306 and sc30 - 717699 are the same event found in the overlap region between two fields . here
we limit the discussion to light curves and selecting the final event sample .
our unexpected result prompted a much more rigorous analysis of blending and a detailed discussion is deferred to
[ sec : blending ] .
we note a relatively large number of events detected in field bul_sc30 compared to the neighboring fields sc22 and sc23 .
@xcite detected a similar excess at a location about 1 deg away .
it remains to be established whether these fluctuations imply a statistically significant clustering of microlensing events .
light curves of 33 microlensing events in the final optical depth sample are shown in fig .
[ fig : lc ] , and their best fit parameters are listed in table [ tbl : candlist6 ] . in case of sc33 - 553617
we also fitted a parallax model , in addition to a standard single lens model .
the best fit parallax model is shown as dotted line in fig .
[ fig : lc ] and its parameters are given in table [ tbl : candlist6 ] .
using the notation from @xcite we found @xmath91 rad and @xmath92 au for the heliocentric orientation of the lens trajectory and the einstein ring radius at the observer plane , respectively .
these values differ by about 2.7 @xmath93 from results in @xcite .
the discrepancy is insignificant and fully accounted for by slight differences in the selection of good science frames and the details of photometric calibrations . in table
[ tbl : candlist5 ] we list level 5 events that were rejected for blending at level 6 .
examples of light curves from the group showing evidence for blending are plotted in fig .
[ fig : lclevel5 ] . for better readability tables [ tbl : candlist6 ] and [ tbl : candlist5 ] only contain symmetric ( parabolic ) 68% confidence level intervals .
the analysis uses proper asymmetric confidence intervals for 1 parameter of interest .
complete information on model fits and their uncertainties is available in the electronic form from apj web .
supplemental files include a complete list of cross - references of our microlensing events with other published work , as well as other parameters and event samples that could not fit comfortably in this paper .
we estimate this crucial ingredient of the microlensing optical depth measurement in a two - step process . in the first part
we construct a synthetic reference image for each field to simulate blending of stellar flux and the effects of source confusion on object catalogs in ogle - ii ( see [ sec : blending ] for a detailed discussion of blending ) .
the second part deals with the ability to detect microlensing events in database light curves known as sampling efficiency .
this reflects the way our actual data set was processed .
stars were detected and their positions were measured on reference images using the dophot code . in order to reduce photometric noise the dia software was later used to measure fluxes .
the transformation of the difference flux to standard magnitudes also involves dophot photometry , which determines the baseline flux .
synthetic stars were drawn from a luminosity function ( lf ) derived separately for each field .
we made composite lfs with the bright end based on the photometric maps of @xcite .
the faint end of the lf , in all fields , was taken to be the same as the hst lf in the baade s window ( @xcite ) , except for adjustments due to variable field extinction and galactic bar geometry ( compare @xcite ) .
both pieces have good completeness near @xmath94 mag , where they were joined and normalized .
we constructed synthetic reference images starting from a flat poisson background .
those images were further populated with stars at random positions down to the extinction free magnitude @xmath95 mag , matching the observed number density of bright stars in the field .
the psf flux of each star was reduced according to the extinction map in @xcite interpolated at the appropriate position from a @xmath96 resolution .
local psf shapes were also determined on a star by star basis using spatial psf models from @xcite .
poisson noise was added in the amount that approximated the statistics of real reference frames .
those simulated frames were processed with dophot and the output source catalog was constructed using procedures identical to those in the preparation of the real database . for each simulated star in the input list we found the closest star in the output dophot catalog .
obviously , the number of detected stars is much lower than the number of simulated stars due to blending , but also because at some low flux level the gb stars are so numerous that they form a smooth background . in fig .
[ fig : ioutiin ] we show the relation between the extinction corrected @xmath47 magnitude of the input star ( @xmath97 ) and the closest output star ( @xmath98 ) .
one can see that a large majority of bright stars is recovered correctly , however great numbers of faint simulated stars `` come out '' as random bright stars .
a faint star has a very high probability of blending with a bright star to form an even brighter catalog source .
this is best studied by monte - carlo simulations because the ability to separate close stars in the photometric code depends in a complicated way at least on the flux , the flux ratio , the separation and the psf shape .
each star from the input lf has equal probability of microlensing , so fig .
[ fig : ioutiin ] can be transformed into a relation between the blending parameter @xmath99 and @xmath98 as shown in fig .
[ fig : ifs ] .
this is simply a prior probability that the lensed flux is a fraction @xmath99 of the total flux , given a random event with the observed magnitude of the composite @xmath98 .
histograms of @xmath99 were obtained by binning the map in fig .
[ fig : ifs ] into 0.2 mag intervals of @xmath98 , and were used in the next step to generate random events in the efficiency simulation on database light curves .
[ fig : histfs ] presents some examples .
it must be stressed that our blending distributions are subject to some ( small ) uncertainties intrinsic to the simulation .
ids of merging objects need to be traced between the simulated input and output catalogs , which does not take place during the construction of the photometric database ( [ sec : data ] ) .
the exact allocation of flux to detected objects is affected by the finite accuracy of positions and the photon noise interacting with the deblending code . in order to simulate real sampling and photometric accuracy
we `` inject '' additional flux and photon noise into actual light curves , randomly selected from the group searched for microlensing ( [ sec : select ] ) .
simulated events have random input parameters drawn from flat distributions : @xmath100 , @xmath101 jd and @xmath102 .
the blending parameter @xmath99 is drawn from one of the histograms made in the previous section according to the baseline magnitude ( fig .
[ fig : histfs ] ) .
the shape of this histogram varies depending on how deep we integrate in @xmath97 .
the efficiency for faint stars is low and very time consuming to compute , so we looked at possible improvements .
a preliminary simulation was conducted for bul_sc3 ( one of the densest ogle - ii fields ) by generating 20,000 events down to @xmath103 .
we strictly repeated the event selection process from [ sec : select ] .
[ fig : fs ] and [ fig : isc ] compare the most important simulation results to observed distributions . in
[ sec : blending ] we will use this information to better understand blending effects . from the data shown in the top - right panel of fig .
[ fig : isc ] we find that only 1% of events in the simulated optical depth sample have input source stars fainter than our brightness threshold @xmath104 .
we verified that this fraction remains very low for other distributions of @xmath25 .
then we can safely assume @xmath105 when constructing the histograms of @xmath99 for subsequent simulations , i.e. we use stars below the dashed line in fig .
[ fig : ifs ] .
it means that our simulated source lf is the one actually measured by the ogle - ii experiment , making our calculations very insensitive to uncertainties in the faint part of the lf .
( sufficiently deep hst data are currently available only for baade s window . ) the actual detection efficiency in each field is computed from 20,000 simulated events with blending down to @xmath104 and integrated over all parameters except @xmath25 .
the results for bul_sc3 field are plotted in fig .
[ fig : eff ] .
they are very similar for other fields , which are available electronically on apj - web .
these efficiencies are reduced by about 10% when compared to the case without blending , because the main peak in @xmath99 is near 0.9 for most stars in our rcg region .
the optical depth @xmath9 can be defined as the probability that a random star is microlensed with the impact parameter @xmath106 at any given time .
it can be estimated as @xmath107 where @xmath108 is the total number of source stars monitored for microlensing , @xmath109 is the duration of the survey in days , @xmath110 is the event time - scale for the @xmath65-th event , and @xmath111 is the detection efficiency at that time - scale . in our case @xmath112 and @xmath113 d. blending simulations in [ sec : im_eff ] show that stars below our magnitude threshold have practically no chance to contribute detectable events .
this allows us to simplify and speed up the calculation by only simulating stars with @xmath114 , but also has a side effect of giving about average efficiency to a small percentage of simulated stars which crossed the threshold and have practically no efficiency in the actual survey .
equivalently , with the efficiency computation from [ sec : cat_eff ] , the number of sources that can be lensed is slightly overestimated due to blending . in principle
, the effect can be corrected by lowering @xmath108 in equation [ eq : opt ] according to the ratio between the input and output lfs integrated above @xmath89 .
depending on the field , the resulting optical depth would increase by 712 % , or @xmath115 on average .
the size of the correction is uncertain by a factor @xmath116 due to several competing biases at a few percent level . to name a few ,
they include : limited accuracy and source confusion intrinsic to our image simulations ( [ sec : im_eff ] ) , occasional dophot artifacts in the database , and second order effects like not counting in @xmath108 a tiny fraction of `` balanced '' blends above the magnitude threshold ( those with @xmath117 and near the bright end of the lf ; see [ sec : centroid_blends ] ) . over - counting of rcg sources
is partially canceled by about 2% due to artificial events having @xmath118 , an effect caused by using 0.2 mag bins to store the @xmath99 distribution ( see fig .
[ fig : histfs ] ) .
those stars can not produce detectable events in our simulation . in the end
, we rescaled the optical depth and the error by @xmath119% . according to the level 5 criteria ( table [ tbl : criteria ] ) ,
all four binary lens events were removed from the optical depth sample .
the model parameters of those events ( including @xmath25 ) are poorly constrained , and their rcg membership is uncertain . the fraction of binary lens events among all microlensing events has been estimated at 8% ( @xcite ) , 6% ( @xcite ) , and 3% ( @xcite ) . in our rcg sample , we found evidence for a binary lens fraction @xmath120 , in good agreement with previous work .
we use this value to correct our optical depth measurement for binary lens events excluded from the sample . assuming that the lens system consists of two stars having the same typical time scale , the optical depth contribution of a binary lens event is @xmath121 times that of a single lens event .
the ratio between the number of events with binary lenses and single lenses is @xmath122 .
it follows that the optical depth values and their errors have to be rescaled by a factor 1.09 .
individual optical depth estimates for each field are listed in table [ tbl : fld ] .
the errors were estimated using the formula of @xcite .
we also estimated the average optical depth in all 20 fields combined , and found @xmath4 at @xmath5 .
the effective line of sight was computed by weighting field centers with the number of source stars in each field . here
the error was estimated using a monte - carlo method from @xcite , which is consistent with a value @xmath123 from the formula of @xcite .
the left panels of fig .
[ fig : opt ] show the microlensing optical depth from ogle - ii bulge data as a function of the galactic coordinates @xmath6 and @xmath124 ( filled circles ) .
fields were binned in ranges @xmath125 with @xmath124 and @xmath6 averaged in each bin .
the errors are estimated according to @xcite .
although there is no correlation in @xmath124 , we can see a trend in the optical depth along @xmath6 .
a simple linear fit gives @xmath7\times 10^{-6}$ ] . in fig .
[ fig : opt ] we also compare the average @xmath9 from this work ( open circle ) with values published by @xcite ( open diamond ) and by @xcite ( open square ) .
all three measurements are roughly consistent .
they fall well below the results in @xcite and @xcite based on all survey stars regardless of brightness ( @xmath126 at @xmath127 ) .
however , the latter values include a dubious upward correction by about 20% for the presence of disk stars that rarely produce events , and the uncorrected estimates ( @xmath35 and @xmath128 , respectively ) are only @xmath129 higher than our present result .
when comparing estimates made by various experiments , we adjust for the measured gradient of @xmath9 .
our new optical depth value is also consistent with predictions based on the revised cobe bar model by @xcite , which has the mass @xmath130 and the viewing angle @xmath131 , and the latest cobe elongated bar model by @xcite with @xmath131 . for comparisons , the right panels of fig .
[ fig : opt ] show the effect of removing the parallax event .
the lens that caused the parallax event is most likely in the disk .
the sample without the parallax event is more representative of the bulge structure and less influenced by the disk component , giving @xmath132\times 10^{-6}$ ] .
the gradient of @xmath9 along @xmath6 becomes more clearly visible in this case , as one would predict based on the gb models . in [ sec :
tau_blends ] we use a control sample to further investigate the systematics associated with various methods of estimating @xmath9 and discuss a possible difference between results based on samples selected using models with and without blending .
here we discuss distributions of the best fit microlensing parameters for our events . in this section event sc31 - 111306
is not included because it is identical with sc30 - 717699 .
the histograms of @xmath99 are shown in fig .
[ fig : fs ] ( top - left ) .
the solid line is for 32 events in the final optical depth sample , while the dotted line represents all 61 events at level 5 .
the remaining panels in this figure show the relevant results from monte - carlo simulations in [ sec : eff ] .
as intended , our level 6 condition removed events which appeared to be highly blended photometrically . in the remaining events @xmath99
is symmetrically distributed around the peak at @xmath133 .
there are several events with `` negative blending '' , i.e. @xmath134 .
it is not allowed theoretically , but also not uncommon in fits to real data .
such unphysical best fit parameters arise due to several factors including : near degeneracy of the model in the presence of statistical fluctuations , systematic errors in photometry ( e.g. from a centroid bias or an imperfect deblending ) , occasional `` holes '' in the background of unresolved faint stars ( @xcite ) , or assuming an incorrect model . in our data
set the first two effects explain most cases .
we allow formally @xmath134 to avoid a systematic bias in the microlensing parameters . in fig .
[ fig : isc ] ( top - left ) we show histograms of @xmath86 , the extinction corrected source magnitude .
again , the solid line represents 32 events in the optical depth sample , while the dotted line is for all 61 events in the level 5 sample .
the remaining panels in this figure are for comparisons with monte - carlo simulations .
the dashed vertical line indicates the magnitude threshold @xmath104 adopted for bul_sc3 field .
the original goal of the blending analysis was to place a reasonable upper limit on the number of blends and derive @xmath9 assuming @xmath133 in rcg events , perhaps after applying a small blending correction to the ingredients going into the estimator .
instead , for about half of apparently bright ogle - ii events we found @xmath99 and @xmath135 that likely puts the actual source well below the magnitude threshold @xmath104 . in
[ sec : photo_blends ] this surprising observation is reconciled with simulations . in fig . [
fig : umin ] we show a cumulative histogram of the best fit impact parameter for our 32 events .
it shows a slight discrepancy with respect to a uniform distribution in @xmath21 .
the ks test gives a 3.4% probability for this departure from uniformity to occur by chance , not alarming in case of this particular type of test .
@xcite find a very similar feature .
slightly low detection efficiencies at large @xmath21 and significant noise in the best fit parameters while @xmath21 is forced to be positive will tend to produce the observed pattern
. the corresponding histogram of @xmath25 is shown in fig .
[ fig : te ] ( top ) .
the mean time - scale is @xmath136 d without any correction for detection efficiency .
this is close to the value of 30 d obtained by @xcite .
a proper comparison with other results and theoretical estimates requires weighting the histogram in fig .
[ fig : te ] by the inverse efficiency ( bottom ) .
the corrected mean time - scale is @xmath137 d , while @xcite found @xmath138 d. @xcite have recently compared several measured time - scale distributions to predictions from a galactic model adopted from @xcite , after it was empirically normalized by star counts .
@xcite found that our corrected @xmath25 distribution ( preprint ) is in good agreement with their model , while the results in @xcite show an excess of short - duration events .
the discrepancy is likely to originate from a different treatment of blending ( [ sec : tau_blends ] ) . our corrected value is longer than the prediction from the model by @xcite .
it is also below 45 d predicted by a bar model rotating at @xmath139 km / s , but significantly longer than 14 d without any streaming motion estimated by @xcite .
therefore , the ogle - ii data provide some support for the existence of streaming motions in the galactic bar . further analysis of the bar kinematics is beyond the scope of this paper . as mentioned in [ sec : fs ] the data of the kind shown in fig .
[ fig : isitotal ] suggest that in many ogle - ii events the true source is much fainter than it appears .
it is somewhat unexpected that a fraction of such events should reach @xmath140% , a value suggested by the number of events rejected at level 6 . on closer look
, however , a brief inspection of model fits for hundreds of the ogle - iii events from the early warning system ( ews ; @xcite ) supports this finding .
as many as 20% of sources brighter than 16 mag are still severely blended in more densely sampled and better resolved ogle - iii images .
the results from image simulations ( [ sec : im_eff ] ) help to clarify the issue . in the top - left panel of fig .
[ fig : fs ] the observed distribution of the blending parameter @xmath99 for level 5 events ( dotted line ) peaks at a low value below @xmath141 . despite a large spread in @xmath99 introduced by model fitting ,
this is inconsistent with @xmath133 in all level 5 events according to the simulation results depicted in the bottom - left panel of fig .
[ fig : fs ] .
the opposite is true for level 6 events . the observed sample is fully consistent with `` no blending '' scenario ( solid lines ) .
we also verified how efficient are level 6 criteria in rejecting most blending without removing too many unblended events .
note that our conservative magnitude threshold employs the formal error @xmath87 in addition to the intrinsic source magnitude @xmath142 ( both from model fits ) .
from fig .
[ fig : isc ] we find that out of all simulated events which make it to the optical depth sample , only 1% has @xmath118 ( top - right ) . at the same time only about 16% of unblended events are lost that way ( bottom - left ) .
the near degeneracy of the model is broken more often with high s / n light curves delivered by dia photometry .
some loss of efficiency was still unavoidable in cases of strong degeneracy , because we insisted on keeping the contamination from faint blends at a low level .
finally , the fraction of simulated events rejected for blending is very close to that in the sample of actual observed events , i.e. @xmath140% ( figs .
[ fig : fs ] and [ fig : isc ] ) , further strenghthening the evidence for numerous blends .
the fraction of blended events relative to all events can be estimated using the statistics collected from the simulation . the number of events which passed level 5 and 6 is @xmath143 and @xmath144 , respectively ( bottom - right panel of fig .
[ fig : isc ] ) . taking into account the fact that 16% of rcg events fail the level 6 cut , the number of real rcg events at level 5 is given by @xmath145 . then the number of blended events is given by @xmath146 , which is 38% of @xmath147 . using the observed events we have @xmath148 and @xmath149 ,
so an estimated 38% of all events are blended events .
this is fully consistent with our simulation .
@xcite studied blending in the eros bulge survey data . using a pixel level simulation they found that in the eros bright star sample the fraction of events due to blends is @xmath8% .
however , the eros analysis was based on a microlensing model without blending , unless it produced unacceptable @xmath72 . in [ sec : tau_blends ] we show that this would imply even more blended events in hypothetical eros samples selected using a model that allows for blending . nevertheless , both ogle and eros image simulations indicate that blends are fairly common in rcg samples .
some strong blends betray their presence by showing a detectable centroid shift of the combined light . while in principle
this information can be used to reject blended events , applying astrometric selection criteria is complicated in practice .
another problem is that positions of some blends are aligned within the measurement errors .
sensitivity can be improved by comparing an unbiased position from stacked difference frames with the mean centroid of the composite .
even then , the maximum observable shift is @xmath150 times the separation of the blend ( for a baseline dominated reference frame ) . despite these limitations we looked for astrometric shifts to check for consistency with the photometric evidence . in fig . [
fig : astrom ] we plot the differences between source positions in the reference image , typically dominated by the baseline , and a stack of difference images taken near the peak magnification .
out of 32 level 6 events ( left ) only two show a clearly detected shift above 0.2 pixels ( @xmath151 ) .
that fraction increases dramatically to 28% for level 5 events ( left and right combined ) and is entirely consistent with the value @xmath152% predicted from a distribution of blend separations in our simulated images with the detection efficiency as a function of @xmath99 . as expected , the largest shifts in fig .
[ fig : astrom ] are observed for strong blends with low @xmath99 .
therefore , we have an independent confirmation that , for bright stars , the dia photometry alone is sufficiently effective in rejecting strong blends in ogle - ii data
. please note that the relevant selection criteria in table [ tbl : criteria ] admit blended events as long as the best fit source magnitude is @xmath153 above the threshold .
this occurs only for a tiny fraction of bright stars ( fig .
[ fig : histfs ] ) and has negligible influence on the optical depth . in previous sections we showed that samples of microlensing events with apparently bright sources still contain many strong blends .
this holds for any survey with seeing comparable to or worse than the median @xmath154 in ogle - ii , assuming a similar density of stars .
it may then seem counterintuitive , that our measured optical depth ( adjusted for the gradient ) agrees with the values recently found by @xcite and @xcite , who used microlensing parameters based on a model without blending to select their rc samples . to investigate this issue we reintroduced the latter method into our analysis of the ogle - ii extended rc sample .
we performed microlensing model fits with the blend flux fixed at @xmath155 , followed by a selection of events using the same criteria as before ( table [ tbl : criteria ] ) .
note that the requirement on the source magnitude at level 6 is now identical to the second condition at level 0 .
this procedure selected 48 good quality events , all of which were confirmed visually .
this control sample was not screened for blending and is 1.5 times larger than the one used to calculate the optical depth ( [ sec : opt ] ) .
its sole purpose is to compare our analysis with previous work ( @xcite ) , particularly to assess the level of bias in the best fit parameters and the impact on the derived optical depth . in figs .
[ fig : umin4 ] and [ fig : te4 ] we compare the results of the model fitting with and without blending . from
[ sec : select ] we show the sample of 32 level 6 events ( filled circles ) and its level 5 superset ( open circles ) .
we also plot the control sample of 48 events introduced for comparisons ( open squares ) .
there are 30 events common between all three samples .
the control sample includes 10 events rejected as blends in the optical depth analysis . in events showing no evidence for blending there is little difference between the best fit model with fixed @xmath133 , and the one with freely variable @xmath99 .
for the remaining ( blended ) events we confirm that forcing @xmath155 gives overestimated @xmath21 and underestimated @xmath25 , a well known fact .
there are 3 events in which @xmath25 significantly increases after fixing @xmath133 .
those have `` negative blends '' , i.e. @xmath134 ( [ sec : fs ] ) . before the efficiency correction
, the control sample had the mean @xmath25 of 35.8 and 25.5 d corresponding to the model fits with and without blending .
event sc37 - 485133 has the best fit @xmath156 d when @xmath99 is allowed to be optimized and was removed from the estimates of the mean time - scale .
the detection efficiency @xmath157 in the control sample was computed following the methods in [ sec : eff ] except that simulated events were assumed to have @xmath158 , as in @xcite and @xcite .
the result for bul_sc3 field is shown in fig .
[ fig : eff4 ] . assuming no blending ,
the efficiency is 30% higher on average compared to @xmath157 in a blended scenario ( fig .
[ fig : eff ] ) , largely because fixing the blending parameter amounts to less degeneracy .
the optical depth estimate with the control sample is consequently @xmath159 , where the error is from @xcite formula .
although @xmath153 lower than our main result , this value explicitly demonstrates a near cancellation of the two major competing biases discussed by @xcite and @xcite . the efficiency corrected mean time - scale in the control sample
is @xmath160 d , shorter by 20% than @xmath137 d estimated in [ sec : umin_and_te ] .
event selection using model fits with @xmath161 produced a sample of roughly 50% more events with @xmath25 underestimated by @xmath162% and @xmath157 higher by @xmath152% .
although the latter assumption is incorrect , it leads to a final value of the optical depth that is statistically indistinguishable from the one we found using a proper model .
note that an estimated 2/5 out of 50% more events stated above are real rcg events , knowing that 16% of real rcg events fail the level 6 cut ( see [ sec : photo_blends ] ) .
so , the fraction of blended events in the control sample is @xmath163% ( 30/150 ) , consistent with 19% from fig .
[ fig : umin4 ] and similar to 17% estimated by @xcite in macho rcg sample .
the difference between optical depth estimates based on two different treatments of blending in our data is at an inconclusive @xmath153 level .
generally , it is possible that one of these methods of estimating @xmath9 is more biased than the other .
such bias might have contributed to somewhat higher @xmath9 values from analyses based on blended microlensing fits ( @xcite and [ sec : opt ] in this paper ) , compared to values based on fits without blending ( @xcite and our control sample ) .
the issue warrants further investigation , however finding a reliable answer will likely require very detailed simulations that are beyond the scope of this paper .
it should be emphasized how risky is the argument that unblended fit can be used when introducing it into the model does not significantly lower the @xmath72 .
the more blending in the event , the more degenerate is the model fit . with a typical s /
n stronger blends are actually harder to find using @xmath72 improvements .
we have evidence for heavy degeneracy in 12 of our level 4 candidates , and mild degeneracy in many more cases . despite having little effect on the optical depth estimate , neglecting blends biases time - scale distributions and event rates .
the degree of cancellation likely depends on the specifics of each dataset , e.g. typical seeing , and eventually will no longer be `` perfect '' as statistical errors continue to shrink . in low s
/ n events , a 5-parameter model fit with blending is substantially more degenerate than a corresponding 4-parameter fit with @xmath164 , and can be sensitive to small changes in the data .
such low level light curve deviations may be caused by weak exotic events , most commonly by parallax and binary lens events .
the problem with this type of deviations is that they are hard to include in simulated events , so they are typically ignored in the detection efficiency calculations .
it is an open question whether the presence of those weakly exotic events in samples selected using a 5-parameter model introduces a bias in the optical depth .
the issue has not been sufficiently studied . a correction for such hypothetical bias
could in principle be obtained from extensive simulations that account for the galactic disk geometry , but such calculations are not available today . there is some evidence that weakly exotic events are not significantly biasing our measurement .
a thorough search for parallax events ( including weak signals ) has been performed by @xcite using the first 3 years of the ogle - ii data .
out of 512 candidate microlensing events , a single convincing event ( herein sc33 - 553617 ) and 5 more marginal parallax events were found ( after excluding two events that are actually binaries ) .
the fraction is about 1 % .
parallax events are most likely due to disk lenses whose fraction is rather uncertain and estimated at 1020 % from disk models .
moreover , the expected time - scale is 50 days or more , so only 1020 % of all events are relevant here .
therefore , we expect that only a few per cent of the events have parallax effects . knowing that 1 event has already been found in the sample
, we expect negligible contribution from parallax events beyond what is already included . in case of binary events ,
the binary fraction from published work is 38 % , and we correct the measurement using our own estimated 6 % fraction .
we do not expect more significant binary lens events .
any bias in the optical depth due to a combination of subtle light curve changes and fitting 5-parameter models would have to be accounted for .
however , to our best knowledge , such bias has not been reported .
we found 81 microlensing events in a sample of about 1 million bright source stars in the extended rcg region .
light curves for 62 of those events are well represented by a single lens model with relatively small uncertainty in the best fit parameters . in the latter sample we investigated blending and concluded that about 38% of microlensing events with the apparent rcg sources are actually due to fainter stars below the magnitude threshold .
this large value for the fraction of blended events is fully supported by our extensive monte - carlo simulations of source confusion in the ogle - ii reference images .
the microlensing optical depth was estimated using 32 high s / n events whose source stars are still in the extended rcg region .
we measured @xmath4 at @xmath5 .
considering our estimate of the gradient along the galactic latitude @xmath6 , @xmath7\times 10^{-6}$ ] , this value is consistent with recent measurements based on rcg source stars by @xcite and @xcite .
our goal was to ensure a high quality measurement of @xmath9 using rcg stars in the galactic bulge .
the result presented here is based on a rigorous treatment of source confusion and a relatively large number of well sampled high s / n events , with the purpose of minimizing uncertainties in @xmath25 and in the number of events .
it may therefore seem surprising that our result is not out of line with the values obtained by @xcite and @xcite , who selected rcg samples using an unblended model . on the other hand ,
these authors have also argued that the two dominant biases nearly cancel each other .
the contribution due to an underestimated number of monitored sources has the opposite sign to that from underestimated time - scales in blending free model fits .
we tested the hypothesis of near cancellation of the bias by reanalyzing our microlensing sample . indeed , under the no blending assumption we found nearly the same value of @xmath159 with roughly 50% more events , 20% lower @xmath25 and 30% higher detection efficiency .
this does not change the fact that such procedure shifts the distribution of time - scales to lower values by significant amount and makes it hard to study the kinematics of the source / lens populations .
the efficiency weighted mean time - scale in our sample of 32 events is @xmath137 d , which is significantly longer than 15 d estimated by @xcite .
this is not as long as 45 d predicted by models with the streaming motion of the bar , but significantly longer than 14 d without any streaming ( @xcite ) .
therefore , our result implies some streaming in the bar ( first measured by @xcite ) .
further analysis of the time - scale distribution is planned including the information in the ogle - ii proper motion catalog ( @xcite ) .
our value of @xmath9 agrees with predictions from @xcite model with the bulge mass @xmath130 and the viewing angle @xmath131 .
it is also consistent with microlensing optical depths derived by @xcite using galactic bulge models of @xcite and @xcite .
it has been noted that @xmath9 measurements based on rcg stars ( @xcite ) fall systematically below the estimates using all sources down to the detection limit , including recent dia results ( @xcite ) .
the difference persists , although at a lower level , when the dubious correction for disk stars is not applied .
we are not aware of a convincing detailed explanation of this gap , although source confusion and model degeneracy biasing @xmath25 and @xmath21 are prime suspects .
alternatively , the difference may be related to the treatment of blending .
the significance of any potential biases due to near degeneracy of the 5-parameter microlensing curve should be established .
the systematics of event selection using model fits with free @xmath99 ( @xcite , and this work ) and with @xmath164 ( @xcite ) may be the culprit .
another area of concern for wider use of the blended microlensing curve is a possible perturbing effect on the optical depth due to the presence of unrecognized weakly exotic events discussed in [ sec : tau_blends ] and [ sec : weakparabin ] .
the event selection criteria in this analysis strongly discriminate against a high level of degeneracy in the fit ( [ sec : level345 ] ) .
it is our goal to estimate @xmath9 using all events of any magnitude and investigate the source of the above discrepancy .
a good understanding of the involved statistics will allow tapping into a much larger set of useful events . as of 2005 , in the gb fields the ogle - iii survey detects about 600 events per year . according to @xcite one
needs roughly 700800 events to distinguish a barred galactic bulge from its axisymmetric alternative , and with @xmath165 events we can hope for a 3@xmath93 measurement of the relative contributions from the bulge and disk components .
future statistical modeling of the microlensing survey data needs to reflect source confusion effects , still evident among bright stars .
we acknowledge a. gould , o. gerhard and m. smith for helpful comments .
ts acknowledges the financial support from the jsps .
pw was supported by the oppenheimer fellowship at lanl .
this work was partly supported with the following grants to bp : nsf grant ast-0204908 , and nasa grant nag5 - 12212 .
the ogle project is partly supported by the polish kbn grant 2p03d02124 to au .
afonso , c. et al .
2003 , a&a , 404 , 145 alard c. , 2000 , a&as , 144 , 363 alard c. , lupton r. h. , 1998 , apj , 503 , 325 alcock , c. et al .
1997 , apj , 486 , 697 alcock c. et al .
, 2000a , apj , 541 , 270 alcock c. et al .
, 2000b , apj , 541 , 734 binney , j. et al .
2000 , apj , 537 , l99 binney , j. gerhard , o. & spergel , d. 1997 , mnras , 288 , 365 bissantz , n. & gerhard , o. 2002 , mnras , 330 , 591 bissantz , n. , debattista , v. p. & gerhard , o. 2004 , apj , 601 , l155 bond i. a. et al . , 2001 ,
mnras , 327 , 868 cern lib . , 1998 , http://wwwasdoc.web.cern.ch/wwwasdoc/minuit/minmain.html evans n.w . , & belokurov , 2002 , apj , 567 , 119 dwek , e. et al . 1995 , apj , 445 , 716 gould , a. & an , j. h. 2002 , apj , 565 , 1381 griest , k. , et al . 1991 , apj , 372 , l79 gyuk , g. 1999 , apj , 510 , 205 han , c. & gould , a. 1995 , apj , 449 , 521 han , c. 1999 , mnras , 309 , 373 han , c. & gould , a. 2003 , apj , 592 , 172 holtzman , j. a. 1998 , apj , 115 , 1946 jaroszyski , m. , 2002 , acta astronomica , 52 , 39 jaroszyski , m. et al . , 2004 , preprint , astro - ph/0408243 kiraga , m. , & paczyski , b. 1994 , apj , 430 , l101 paczyski , b. 1986 , apj , 304 , 1 paczyski , b. 1991 , apj , 371 , l63 paczyski , b. et al .
1994 , apj . , 435 , l113 paczyski , b. 1996 , ara&a , 34 , 419 palanque - delabrouille , n. , et al .
1998 , , 332 , 1 park , b. -g .
2004 , apj , 609 , 166 peale , s. j. 1998 , apj , 509 , 177 popowski , p. et al .
2001 , in asp conference series : microlensing 2000 : a new era of microlensing astrophysics , eds .
menzies & p.d .
sackett ( san francisco : astronomical society of the pacific ) , vol .
239 , p. 244 ,
( astro - ph/0005466 ) popowski , p. et al .
2004 , preprint ( astro - ph/0410319 ) schechter , l. , mateo , m. , & saha , a. , 1993 , pasp , 105 , 1342s smith m. c. , mao , s. & woniak , p. r. , 2002 , mnras , 332 , 962 soszyski , i. et al .
2001 , apj , 552 , 731 sumi , t. , 2004 , mnras , 349 , 193 sumi t. , eyer , l. & woniak , p. r. , 2003 , mnras , 340 , 1346 sumi , t. et al . , 2003 ,
apj , 591 , 204 sumi , t. et al . , 2004a , mnras , 348 , 1439 sumi , t. et al . ,
2004b , mnras , 356 , 331 thomas , c.l .. et al .
, 2004 , astro - ph/0410341 udalski a. , zebru k. , szymaski , m. , kubiak m. , pietrzyski g. , soszyski i. , woniak p. r. 2000 , acta astronomica , 50 , 1 udalski , a. et al . 1994 , acta astronomica , 44 , 165 udalski a. et al .
2002 , acta astronomica , 52 , 217 udalski a. , kubiak m. , szymaski m. k. , 1997 , acta astronomica , 47 , 319 udalski , a. 2003 , acta astronomica , 53 , 291 wood , a. , & mao , s. 2005 , preprint ( astro - ph/0507210 ) woniak p. r. , & paczyski , b. 1997 , apj , 487 , 55 woniak p. r. , 2000 , acta astronomica , 50 , 421 woniak , p. r. , et al .
2001 , acta astronomica , 51 , 175 zhao , h. & mao , s. 1996 , mnras , 283 , 1197 zhao , h. , spergel , d. n. & rich , r. 1995 , apj , 440 , l13 color magnitude diagram of the bul_sc1 field . our extended rcg region is defined as the upper right portion of the figure with respect to the solid line .
the filled circle represents the `` center '' of the rcg distribution from sumi ( 2004 ) . ] as a function of the galactic @xmath124 and @xmath6 ( filled circles ) .
the results based on the full level 6 sample from ogle - ii ( left ) are compared to the ones obtained after the parallax event has been removed ( right ) .
the latter case is a better representation of the bulge structure since the parallax event was likely caused by a disk lens .
also shown is the average @xmath9 in all 20 fields weighted by the stellar number density ( open circle ) .
solid line is the best fit linear gradient along @xmath6 .
there is no indication of gradient along @xmath124 in this data .
open diamond and open square represent results from popowski et al .
( 2004 ) and afonso ( 2003 ) .
[ fig : opt ] , title="fig : " ] as a function of the galactic @xmath124 and @xmath6 ( filled circles ) .
the results based on the full level 6 sample from ogle - ii ( left ) are compared to the ones obtained after the parallax event has been removed ( right ) .
the latter case is a better representation of the bulge structure since the parallax event was likely caused by a disk lens .
also shown is the average @xmath9 in all 20 fields weighted by the stellar number density ( open circle ) .
solid line is the best fit linear gradient along @xmath6 .
there is no indication of gradient along @xmath124 in this data .
open diamond and open square represent results from popowski et al .
( 2004 ) and afonso ( 2003 ) . [ fig : opt ] , title="fig : " ] clccccc 1 & 1.08 & -3.62 & 35844 & 0 & 0.000000 & 0.000000 + 2 & 2.23 & -3.46 & 38278 & 1 & 2.483758 & 2.483758 + 3 & 0.11 & -1.93 & 80976 & 4 & 4.286616 & 2.226285 + 4 & 0.43 & -2.01 & 78454 & 5 & 3.145996 & 1.494732 + 20 & 1.68 & -2.47 & 58891 & 1 & 1.239517 & 1.239517 + 21 & 1.80 & -2.66 & 52930 & 0 & 0.000000 & 0.000000 + 22 & -0.26 & -2.95 & 50732 & 1 & 0.854807 & 0.854807 + 23 & -0.50 & -3.36 & 42205 & 0 & 0.000000 & 0.000000 + 30 & 1.94 & -2.84 & 47974 & 6 & 9.543987 & 4.183375 + 31 & 2.23 & -2.94 & 45825 & 1 & 2.261020 & 2.261020 + 32 & 2.34 & -3.14 & 40476 & 1 & 0.937948 & 0.937948 + 33 & 2.35 & -3.66 & 34461 & 2 & 10.415521 & 7.933656 + 34 & 1.35 & -2.40 & 61515 & 2 & 4.087111 & 2.890040 + 35 & 3.05 & -3.00 & 41146 & 2 & 3.211394 & 2.350345 + 36 & 3.16 & -3.20 & 37922 & 0 & 0.000000 & 0.000000 + 37 & 0.00 & -1.74 & 85289 & 2 & 2.211746 & 1.772319 + 38 & 0.97 & -3.42 & 40148 & 2 & 2.876526 & 2.246087 + 39 & 0.53 & -2.21 & 72842 & 3 & 1.621968 & 0.967699 + 45 & 0.98 & -3.94 & 32823 & 0 & 0.000000 & 0.000000 + 46 & 1.09 & -4.14 & 29626 & 0 & 0.000000 & 0.000000 + llr 0 & @xmath166 & + & @xmath167 & + & @xmath168 & 1,084,267 + 1 & @xmath169 & + & @xmath170 & + & @xmath171 & + & @xmath172 for @xmath173 & 821 + 2 & @xmath174 & + & @xmath175 d & + & @xmath176 & + & @xmath177 for @xmath178 & + & @xmath179 for @xmath180 & + & @xmath181 for @xmath182 & 135 + 3 & reject spurious events & 81 + 4 & error estimates converged for all parameters & 69 + 5 & @xmath183 and no cross ref with binary events & 62 + 6 & @xmath184 & 33 + lccrrrlccc sc2 - 65831 & 18:04:06.85 & @xmath18529:01:17.2 & 57 & 1459.9@xmath186 & 44.7@xmath187 & 0.572 @xmath188 & 0.40@xmath189 & 15.0@xmath190 & 0.56 + sc3 - 91382 & 17:53:09.33 & @xmath18530:01:12.1 & 135 & 956.0@xmath191 & 38.8@xmath192 & 0.068 @xmath193 & 1.05@xmath194 & 14.9@xmath191 & 0.95 + sc3 - 356103 & 17:53:33.93 & @xmath18529:46:15.6 & 54 & 1475.8@xmath195 & 46.3@xmath196 & 0.488 @xmath197 & 0.76@xmath198 & 14.7@xmath199 & 0.90 + sc3 - 371229 & 17:53:21.13 & @xmath18529:40:37.4 & 49 & 1759.6@xmath200 & 15.6@xmath201 & 0.518 @xmath202 & 0.83@xmath203 & 15.0@xmath200 & 0.76 + sc3 - 478487 & 17:53:47.38 & @xmath18530:05:26.2 & 56 & 1705.5@xmath200 & 41.2@xmath204 & 0.531 @xmath205 & 0.56@xmath206 & 13.4@xmath200 & 1.64 + sc4 - 65601 & 17:54:09.41 & @xmath18529:53:36.9 & 111 & 1767.1@xmath191 & 25.5@xmath207 & 0.621 @xmath208 & 1.06@xmath189 & 12.8@xmath200 & 0.60 + sc4 - 267762 & 17:54:21.79 & @xmath18529:53:24.0 & 147 & 971.1@xmath191 & 7.5@xmath190 & 0.214 @xmath209 & 0.84@xmath206 & 14.6@xmath200 & 1.03 + sc4 - 522952 & 17:54:38.64 & @xmath18529:33:12.8 & 121 & 1257.1@xmath200 & 23.6@xmath207 & 0.456 @xmath210 & 1.05@xmath211 & 14.1@xmath200 & 0.55 + sc4 - 568740 & 17:54:49.35 & @xmath18529:20:25.0 & 39 & 915.2@xmath190 & 22.9@xmath212 & 0.490 @xmath213 & 0.41@xmath214 & 15.2@xmath186 & 0.53 + sc4 - 708424 & 17:55:00.07 & @xmath18529:35:03.7 & 126 & 1397.8@xmath191 & 5.9@xmath190 & 0.247 @xmath215 & 0.93@xmath206 & 14.5@xmath200 & 2.35 + sc20 - 560821 & 17:59:27.19 & @xmath18528:32:31.5 & 42 & 1052.2@xmath199 & 24.2@xmath216 & 0.542 @xmath217 & 0.33@xmath218 & 14.9@xmath192 & 1.01 + sc22 - 390560 & 17:56:56.81 & @xmath18531:11:51.0 & 48 & 1074.3@xmath200 & 14.4@xmath204 & 0.731 @xmath219 & 1.48@xmath220 & 14.9@xmath199 & 0.70 + sc30 - 57488 & 18:01:02.52 & @xmath18529:00:11.6 & 64 & 1274.8@xmath200 & 16.0@xmath221 & 0.528 @xmath222 & 0.61@xmath223 & 14.8@xmath190 & 0.98 + sc30 - 165305 & 18:01:07.74 & @xmath18528:31:41.7 & 155 & 1302.0@xmath191 & 19.5@xmath199 & 0.515 @xmath224 & 1.24@xmath206 & 13.5@xmath191 & 1.61 + sc30 - 454605 & 18:01:30.89 & @xmath18528:59:26.7 & 61 & 1781.4@xmath190 & 65.5@xmath187 & 0.738 @xmath225 & 1.03@xmath223 & 14.6@xmath190 & 1.35 + sc30 - 559419 & 18:01:33.95 & @xmath18528:28:02.3 & 81 & 709.9@xmath190 & 31.5@xmath204 & 0.615 @xmath226 & 0.90@xmath203 & 13.4@xmath200 & 1.62 + sc30 - 636963 & 18:01:44.80 & @xmath18528:58:03.5 & 113 & 1775.7@xmath191 & 18.9@xmath227 & 0.357 @xmath224 & 1.03@xmath228 & 14.9@xmath200 & 0.84 + sc30 - 717699 & 18:01:52.15 & @xmath18528:32:36.5 & 38 & 1749.3@xmath199 & 46.1@xmath229 & 0.992 @xmath230 & 0.62@xmath231 & 13.3@xmath227 & 0.76 + sc31 - 111306 & 18:01:52.14 & @xmath18528:32:36.5 & 36 & 1749.6@xmath199 & 49.2@xmath232 & 0.965 @xmath233 & 0.59@xmath234 & 13.4@xmath227 & 0.96 + sc32 - 333270 & 18:03:21.74 & @xmath18528:28:50.4 & 108 & 1015.8@xmath191 & 13.0@xmath227 & 0.330 @xmath235 & 1.00@xmath236 & 13.9@xmath200 & 1.07 + sc33 - 540825 & 18:05:45.79 & @xmath18528:30:52.4 & 45 & 1552.0@xmath227 & 53.8@xmath237 & 0.000 @xmath238 & 1.12@xmath236 & 14.2@xmath200 & 0.85 + sc33 - 553617 & 18:05:46.71 & @xmath18528:25:32.1 & 74 & 647.1@xmath199 & 153.5@xmath239 & 0.452 @xmath240 & 0.75@xmath189 & 15.3@xmath200 & 2.76 + sc33 - 553617 & & & & 668.3@xmath241 & 140.9@xmath242 & 0.405 @xmath243 & 0.84@xmath234 & 15.2@xmath186 & 1.21 + sc34 - 451887 & 17:58:14.20 & @xmath18528:48:22.3 & 45 & 1576.3@xmath244 & 57.5@xmath245 & 0.535 @xmath246 & 0.94@xmath247 & 14.7@xmath199 & 0.78 + sc34 - 840343 & 17:58:37.12 & @xmath18529:06:29.9 & 62 & 798.9@xmath199 & 66.8@xmath248 & 0.377 @xmath249 & 1.29@xmath250 & 12.3@xmath190 & 0.95 + sc35 - 144974 & 18:04:09.65 & @xmath18527:44:34.9 & 54 & 684.9@xmath200 & 33.9@xmath251 & 0.463 @xmath252 & 0.68@xmath211 & 15.2@xmath200 & 0.85 + sc35 - 451130 & 18:04:33.63 & @xmath18528:07:32.2 & 68 & 998.6@xmath200 & 17.2@xmath241 & 0.534 @xmath253 & 0.79@xmath214 & 14.8@xmath190 & 1.61 + sc37 - 401293 & 17:52:32.30 & @xmath18529:58:46.8 & 22 & 1650.3@xmath227 & 66.3@xmath254 & 0.800 @xmath255 & 0.38@xmath256 & 14.6@xmath227 & 0.82 + sc37 - 645044 & 17:52:58.69 & @xmath18529:34:22.2 & 78 & 1685.7@xmath191 & 12.0@xmath227 & 0.373 @xmath205 & 1.27@xmath214 & 14.5@xmath200 & 0.91 + sc38 - 95103 & 18:01:09.74 & @xmath18529:56:18.9 & 197 & 990.5@xmath191 & 7.3@xmath190 & 0.146 @xmath257 & 0.92@xmath258 & 15.3@xmath191 & 1.00 + sc38 - 120518 & 18:01:10.23 & @xmath18529:48:55.2 & 88 & 1316.1@xmath200 & 33.9@xmath259 & 0.557 @xmath235 & 1.05@xmath260 & 15.0@xmath200 & 0.79 + sc39 - 140577 & 17:55:17.08 & @xmath18529:37:41.0 & 150 & 1281.4@xmath200 & 6.9@xmath201 & 0.126 @xmath261 & 0.92@xmath262 & 14.7@xmath186 & 1.11 + sc39 - 323517 & 17:55:36.42 & @xmath18529:42:14.0 & 29 & 702.7@xmath190 & 14.2@xmath263 & 0.930 @xmath264 & 1.19@xmath265 & 14.8@xmath201 & 0.89 + sc39 - 361372 & 17:55:28.67 & @xmath18529:33:41.7 & 154 & 682.3@xmath191 & 17.7@xmath190 & 0.181 @xmath257 & 1.07@xmath258 & 12.1@xmath191 & 0.65 + lccrrrlccc sc3 - 147406 & 17:53:09.61 & @xmath18529:46:39.3 & 14 & 1409.2@xmath186 & 57.3@xmath266 & 0.221 @xmath267 & 0.04@xmath268 & 17.4@xmath186 & 1.20 + sc3 - 576464 & 17:53:41.56 & @xmath18529:39:06.0 & 40 & 656.5@xmath190 & 43.3@xmath269 & 0.418 @xmath270 & 0.50@xmath228 & 15.7@xmath200 & 0.76 + sc3 - 577610 & 17:53:36.20 & @xmath18529:38:08.2 & 23 & 1779.7@xmath200 & 74.2@xmath271 & 0.051 @xmath272 & 0.04@xmath268 & 19.2@xmath190 & 0.69 + sc3 - 601945 & 17:53:36.63 & @xmath18529:31:21.2 & 17 & 924.4@xmath227 & 43.2@xmath273 & 0.574 @xmath274 & 0.68@xmath275 & 16.1@xmath192 & 1.05 + sc4 - 134300 & 17:54:13.04 & @xmath18529:35:14.1 & 21 & 1041.0@xmath190 & 153.3@xmath276 & 0.047 @xmath277 & 0.02@xmath268 & 19.3@xmath186 & 0.82 + sc4 - 321697 & 17:54:25.13 & @xmath18529:37:49.4 & 36 & 1615.6@xmath190 & 28.1@xmath278 & 0.522 @xmath279 & 0.80@xmath218 & 15.8@xmath190 & 0.77 + sc4 - 489287 & 17:54:42.92 & @xmath18529:44:12.1 & 15 & 1751.5@xmath200 & 15.4@xmath187 & 0.463 @xmath280 & 0.16@xmath189 & 17.2@xmath227 & 0.80 + sc4 - 500461 & 17:54:41.53 & @xmath18529:40:10.1 & 41 & 745.8@xmath199 & 30.1@xmath281 & 0.000 @xmath282 & 0.17@xmath258 & 16.4@xmath186 & 0.53 + sc4 - 624085 & 17:55:00.19 & @xmath18529:59:53.6 & 14 & 659.3@xmath199 & 29.3@xmath283 & 0.653 @xmath284 & 0.24@xmath260 & 15.8@xmath227 & 1.73 + sc4 - 719953 & 17:54:56.68 & @xmath18529:31:47.6 & 37 & 1725.8@xmath200 & 32.3@xmath285 & 0.020 @xmath286 & 0.01@xmath287 & 19.7@xmath199 & 0.66 + sc20 - 525747 & 17:59:28.45 & @xmath18528:42:53.3 & 14 & 1746.3@xmath190 & 5.3@xmath248 & 0.063 @xmath288 & 0.02@xmath194 & 19.0@xmath241 & 0.66 + sc21 - 766993 & 18:00:39.38 & @xmath18528:55:14.6 & 28 & 1119.6@xmath244 & 93.2@xmath289 & 0.116 @xmath226 & 0.07@xmath194 & 17.6@xmath186 & 0.68 + sc22 - 380074 & 17:56:58.83 & @xmath18531:14:05.6 & 16 & 587.2@xmath186 & 40.8@xmath290 & 0.090 @xmath291 & 0.03@xmath194 & 19.3@xmath207 & 0.74 + sc22 - 414328 & 17:56:59.20 & @xmath18531:03:22.3 & 31 & 1832.8@xmath192 & 51.0@xmath292 & 0.536 @xmath293 & 0.52@xmath223 & 16.2@xmath199 & 0.62 + sc23 - 524386 & 17:57:56.24 & @xmath18530:51:58.9 & 82 & 1434.8@xmath200 & 64.0@xmath251 & 0.291 @xmath209 & 0.73@xmath258 & 15.9@xmath200 & 1.02 + sc30 - 671185 & 18:01:47.58 & @xmath18528:49:04.9 & 22 & 557.6@xmath192 & 50.0@xmath229 & 0.332 @xmath294 & 0.15@xmath295 & 17.4@xmath199 & 1.05 + sc32 - 208566 & 18:03:22.47 & @xmath18529:03:43.2 & 11 & 690.0@xmath296 & 67.6@xmath297 & 0.577 @xmath298 & 0.13@xmath236 & 17.3@xmath259 & 0.74 + sc33 - 505 & 18:05:03.99 & @xmath18529:18:08.6 & 14 & 1778.5@xmath186 & 30.4@xmath299 & 0.466 @xmath300 & 0.08@xmath295 & 17.1@xmath192 & 0.84 + sc33 - 290665 & 18:05:29.67 & @xmath18528:51:03.3 & 22 & 1845.3@xmath301 & 73.5@xmath302 & 0.815 @xmath303 & 0.57@xmath304 & 15.8@xmath204 & 0.73 + sc34 - 606996 & 17:58:34.18 & @xmath18529:06:29.4 & 8 & 974.9@xmath201 & 154.1@xmath305 & 0.091 @xmath291 & 0.01@xmath287 & 19.5@xmath192 & 0.72 + sc37 - 556534 & 17:53:02.92 & @xmath18530:03:07.8 & 30 & 1786.3@xmath191 & 8.3@xmath306 & 0.054 @xmath235 & 0.05@xmath194 & 17.8@xmath186 & 0.83 + sc39 - 1073 & 17:55:10.06 & @xmath18530:10:29.5 & 24 & 1807.5@xmath190 & 14.9@xmath263 & 0.563 @xmath307 & 0.40@xmath308 & 16.3@xmath207 & 0.87 + sc39 - 28566 & 17:55:10.31 & @xmath18530:05:11.7 & 14 & 1374.0@xmath190 & 21.2@xmath309 & 0.033 @xmath310 & 0.01@xmath268 & 20.0@xmath207 & 0.75 + sc39 - 54533 & 17:55:19.53 & @xmath18529:57:57.0 & 9 & 1223.9@xmath269 & 45.6@xmath311 & 0.372 @xmath312 & 0.05@xmath295 & 17.7@xmath221 & 0.62 + sc39 - 269524 & 17:55:23.99 & @xmath18529:55:33.1 & 11 & 906.7@xmath192 & 32.2@xmath313 & 0.096 @xmath314 & 0.03@xmath268 & 16.9@xmath227 & 1.79 + sc39 - 322789 & 17:55:26.87 & @xmath18529:43:59.7 & 12 & 968.0@xmath315 & 53.2@xmath316 & 0.471 @xmath317 & 0.10@xmath206 & 16.9@xmath192 & 0.75 + sc39 - 468687 & 17:55:45.16 & @xmath18529:56:38.6 & 10 & 1210.8@xmath318 & 86.0@xmath319 & 0.572 @xmath320 & 0.13@xmath236 & 16.8@xmath195 & 0.82 + sc39 - 576039 & 17:55:43.97 & @xmath18529:24:28.7 & 7 & 620.1@xmath241 & 106.4@xmath321 & 0.360 @xmath322 & 0.05@xmath258 & 18.1@xmath315 & 1.04 + sc39 - 753576 & 17:56:01.00 & @xmath18529:25:11.9 & 6 & 1241.2@xmath201 & 88.9@xmath323 & 0.112 @xmath225 & 0.01@xmath287 & 19.0@xmath192 & 0.79 + | we present a measurement of the microlensing optical depth toward the galactic bulge based on 4 years of the ogle - ii survey .
we consider only bright sources in the extended red clump giant ( rcg ) region of the color - magnitude diagram , in 20 bulge fields covering @xmath0 deg@xmath1 between @xmath2 and @xmath3 .
using a sample of 32 events we find @xmath4 at @xmath5 .
taking into account the measured gradient along the galactic latitude @xmath6 , @xmath7\times 10^{-6}$ ] , this value is consistent with previous measurements using rcg sources and recent theoretical predictions .
we determine the microlensing parameters and select events using a model light curve that allows for flux blending .
photometric quality delivered by difference image analysis ( dia ) combined with the 1.3 median seeing of the ogle - ii images are sufficient to constrain and reject the majority of strong blends .
we find that @xmath8% of the ogle - ii events which appear to have rcg sources are actually due to much fainter stars blended with a bright companion .
we show explicitly that model fits without blending result in similar @xmath9 estimates through partial cancellation of contributions from higher detection efficiency , underestimated time - scales and larger number of selected events . the near cancellation of the optical depth bias and the fact that microlensing event selection based on models without blending discriminates against blends have been utilized by previous analyses based on rcg sources .
the latter approach , however , leads to biased time - scale distributions and event rates . consequently
, microlensing studies should carefully consider source confusion effects even for bright stars . |
the excitations of the quasi - one - dimensional ( 1d ) heisenberg antiferromagnets systems have been studied extensively since haldane predicted an energy gap in the excitation spectra of the isotropic _ integer_-spin heisenberg chains in 1983 @xcite .
now it is well established that there is an energy gap separating the singlet ground state from the triplet lowest - energy - excitation states for the _ integer_-spin heisenberg chains , contrast to the gapless excitation states of the _ half - odd - integer_-spin heisenberg systems @xcite .
this theoretical prediction has been confirmed by buyers _
et al _ @xcite in the inelastic - neutron - scattering experiments on the quasi-1d antiferromangetic compound csnicl@xmath0 .
some subsequent experimental investigations @xcite and numerical calculations @xcite also support haldane s prediction . at very low temperature ,
most of the quasi-1d antiferromagnetic materials including csnicl@xmath0 show the three - dimensional nature with the classical magnetic order , and more interestingly , energy gaps at the magnetic wavevector have also been observed in many compounds @xcite . for the case of csnicl@xmath0 ,
the observed energy gap was initially explained by a uniaxial single - ion anisotropy but now it is widely accepted that the gapped excited state belongs to one of the two longitudinal modes corresponding to the oscillations in the magnitude of the magnetic order of the quasi-1d hexagonal systems , first proposed by affleck based on a simplified version of haldane s theory @xcite .
the gapped longitudinal modes are clearly beyond the conventional spin - wave theory which produces only the transverse excitations usually referred to as magnons . a later experimental study by enderle _
@xcite using high - resolution polarized neutron scattering also confirms affleck s proposal of the longitudinal modes , and contradicts the spin - wave theory of two - magnon by ohyama and shiba @xcite or a modified spin - wave theory by plumer and caill @xcite .
there are also investigations of the longitudinal excitation states in other quasi-1d structures with the nel - like long - ranged order at low temperature such as the tetragonal kcuf@xmath1 with @xmath2 @xcite , where good agreements between the experiment and a theory based on a sine - gordon field theory have been found for the energy gap at the magnetic wavevector @xcite .
more recently , a longitudinal mode was also observed in the dimerized antiferromagnetic compound tlcucl@xmath1 under pressure with a long - ranged nel order @xcite .
we recently proposed a general microscopic many - body theory based on the magnon - density waves for the longitudinal excitations of spin-@xmath3 quantum antiferromagnetic systems @xcite .
in analogy to feynmann s theory of the low - lying excited states in the helium-4 superfluid @xcite , we identify the longitudinal excitation states in a quantum antiferromagnet with a nel - like order as the collective modes of the magnon - density waves . in application to the quasi-1d tetragonal structure of kcuf@xmath1 with @xmath2 , with no other fitting parameters than the nearest - neighbor coupling constants in the model hamiltonian , we find that our numerical results for the energy gap values at the magnetic wavevector are in general agreement with the experiments @xcite .
we hope that more experimental results for the energy spectra at other wavevectors will be available for comparison . in this article , we extend our microscopic approach to the quasi-1d hexagonal quantum antiferromagnets such as csnicl@xmath1 and rbnicl@xmath1 @xcite both with spin-1 and csmni@xmath1 with spin-5/2 @xcite .
furthermore , we also investigate the higher - order contributions to the longitudinal excitation spectra in the large-@xmath3 expansion .
the basal planes of these materials are antieferromagnetic triangular lattice with the noncollinear magnetic order .
hence there are two possible longitudinal modes in these hexagonal systems , rather than the single longitudinal mode of the bipartite systems such as the tetragonal kcuf@xmath1 .
some preliminary results for the two dimensional triangular model have been published @xcite .
we organize this article as follows . for completeness
, we outline the main results of the spin - wave theory for the quasi-1d model in sec .
ii , using the one - boson approach after two spin rotations .
we obtain the full spin - wave spectra as a function of the uniaxial single - ion anisotropy .
to our knowledge , this anisotropy dependence of the spin - wave spectra has not been published before .
we then apply our microscopic theory for the longitudinal excitations in sec .
iii , using the approximated ground state from the spin - wave theory .
the energy gaps due to the anisotropy and the energy gaps of the longitudinal modes without anisotropy are then discussed in details .
we compare our results for the longitudinal energy gaps with the experimental results for the spin-1 compounds csnicl@xmath0 and rbnicl@xmath0 and the spin-5/2 compound csmni@xmath4 .
we find good agreements for the energy gap values for csnicl@xmath0 and rbnicl@xmath0 after including the higher - order contributions in our calculations . for csmni@xmath4 which is very close to the pure 1d system we find a big discrepancy between our approximation of the gap value and the experimental results .
we conclude this article by a summary and a discussion of the possible further corrections particularly for csmni@xmath4 .
the quasi-1d materials such as csnicl@xmath0 crystallize in the hexagonal @xmath5 structure with space group @xmath6 , where @xmath7 is an alkaline - metal cation , @xmath8 is a cation of the 3@xmath9 group , and @xmath10 is a halogen anion . the magnetic ions @xmath8 constructs the hexagonal lattice in the @xmath11 plane with adjacent spins forming angles of @xmath12 , and antiparallel adjacent spins along the chain of the @xmath13 axis as shown in figs . [ fig1](a ) and ( b ) .
the lattice constants of csnicl@xmath0 , for example , are @xmath14 @xmath15 and @xmath16 @xmath15 , and the magnetic moments are carried by ni@xmath17 . the superexchange interaction between @xmath8 ( ni@xmath17 ) ions
is modeled by an @xmath18-spin heisenberg hamiltonian with a strong intrachain interaction @xmath19 and weak interchain interaction @xmath20 such as @xmath21 where the notation @xmath22 indicates the nearest - neighbor couplings only and where we have also added an ising - like single - ion anisotropy term with constant @xmath23 .
most of the intrachain couplings in @xmath5 compounds are antiferromagnetic such as in csnicl@xmath0 or rbnicl@xmath0 with easy single - site anisotropy , or csmnbr@xmath0 and rbmnbr@xmath0 with hard anisotropy @xcite .
these intrachain couplings can also be ferromagnetic ( i.e. , @xmath24 ) as in csnif@xmath0 @xcite or cscucl@xmath0 @xcite .
we consider only the antiferromagnetic couplings here .
therefore , the classical ground state of each linear chain along the @xmath13 axis ( also denoted as @xmath25-axis ) is a nel state with alternating spin - up ( blue ) and spin - down ( red ) alignments as shown in fig . 1 ( b ) .
we consider a spin - wave theory for the hamiltonian ( 1 ) based on the one - boson approach by performing two spin rotations .
firstly , we rotate the local axes of all up - spins ( blue ) by @xmath26 so that all spins along each chain align in the same down direction .
this is equivalent to the transformation @xmath27 for the first terms in eq .
( 1 ) , leaving the last two terms unchanged .
the second rotation is on the hexagonal lattice of the @xmath11 plane ( or @xmath28-plane ) on the second and third terms of eq .
( 1 ) . following singh and huse @xcite and miyake @xcite , for every triangle of the hexagonal lattices
[ see fig 1(a ) ] , we rotate the local axes of two spins along the classical direction in the @xmath28-plane to align with that of the third spin @xcite .
this is equivalent to the rotation of the @xmath29-sites of eq .
( 1 ) by the following transformation @xmath30 where @xmath31 and @xmath32 with @xmath33 at @xmath34 defined as the magnetic - ordering wavevector of the quasi-1d hexagonal systems .
the hamiltonian after these two transformations is given as @xmath35-\frac{1}{2}j'\sum\limits_{l,\varrho'}^ { \text{plane}}\big[s_l^z s_{l+\varrho'}^z + \frac{3}{4}(s_l^+s_{l+\varrho'}^++s_l^-s_{l+\varrho'}^-)\\ & -\frac{1}{4}(s_l^+s_{l+\varrho'}^-+s_l^-s_{l+\varrho'}^+ ) -2\sin(\theta_l-\theta_{l+\varrho ' } ) ( s_l^zs_{l+\varrho'}^x - s_l^xs_{l+\varrho'}^z)\big]+\tilde{\cal h}^d , \end{split}\ ] ] where @xmath36 runs through all sites , @xmath37 and @xmath38 are the nearest neighbor index vectors with coordination numbers @xmath39 along the chain and @xmath40 on the hexagonal basal planes respectively , and @xmath41 is the rotated anisotropy term .
care should be taken for the two rotations on this anisotropy term .
the first rotation of eq .
( 2 ) leaves it unchanged due to its quadratic form as mentioned before . in order to perform the second rotation of eq .
( 3 ) involving rotations of the axes of the two spins to align with the axis of the third spin on the triangular planes , we rewrite the anisotropy term of eq .
( 1 ) in the following equivalent , suitable form @xmath42.\ ] ] the transformation of eq . to the second term in eq . ( 5 ) gives @xmath43 .
\end{split}\ ] ] we notice that this anisotropy form is different from the simple form of ref .
@xcite or that of ref .
we believe that eq .
( 6 ) is the correct form suitable for the hexagonal systems .
the energy gaps in the energy spectra due to this anisotropy term will be presented later . using the canonical holstein - primakoff transformations ,
the spin operators are expressed in terms of a single set of boson operators @xmath44 and @xmath45 as , @xmath46 where @xmath47 and @xmath3 is the spin quantum number .
the hamiltonian of eq . can then be written as , after fourier transformations of the boson operators with the fourier component operators @xmath48 and @xmath49 and to the order of @xmath50 , @xmath51 where @xmath52 is the classical energy , @xmath53 with @xmath12 and @xmath54 is given by the quadratic terms in the boson operators as @xmath55\ ] ] with constants @xmath56 and @xmath57 defined by @xmath58 and @xmath59 defined as usual by @xmath60 the quadratic hamiltonian @xmath54 of eq .
is diagonalized by the usual bogoliubov transformation and can be written in terms of the new boson operators @xmath61 and @xmath62 as , @xmath63 where @xmath64 is the quantum correction to the classical ground state energy of eq . , @xmath65 and @xmath66 is the spin - wave excitation spectra , @xmath67 the first brillouin zone of a quasi-1d antiferromagnet is ploted in fig . 2 , where the magnetic wavevector @xmath68 is located at the corner of the hexagon and where other symmetry points in conventional notations are also illustrated .
we plot the spin - wave spectra of eq .
( 15 ) in fig .
[ fig3 ] for csnicl@xmath0 using the experimental values @xmath69 , @xmath70 thz and negligible anisotropy @xmath71 @xcite .
we define the ratio of the two nearest - neighbor coupling constants as @xmath72 and , for csnicl@xmath0 , @xmath73 the spin - wave energy spectra with different polarizations are obtained by folding of the wavevectors . in fig . 3 ,
several branches along the symmetry direction of @xmath74 , and @xmath75 are shown , where @xmath76 is the reduced wave vector component in the reciprocal lattice unit ( r.l.u ) with @xmath77 , and @xmath78 $ ] .
using eq .
( [ eq12 ] ) the moving in the paramagnetic brillouin zone can be written as for @xmath79 and @xmath80 , and the corresponding symmetry directions to those in reciprocal lattice unit are @xmath81 and @xmath82 respectively .
the three transverse spin - wave branches are obtained from eq . as follows .
the @xmath25-mode has the polarization along the @xmath25-axis of the hexagonal lattice where the quantum fluctuation is at @xmath83 ; the other two modes are found in the @xmath28-plane by translating the wavevector by a magnetic wavevector as @xmath84 and are denoted as @xmath85 respectively . as can be seen from fig .
3 , at the magnetic wavevector @xmath86 , the @xmath25-mode is gapless for zero anisotropy ( @xmath87 ) .
however , as mentioned earlier , an energy gap about @xmath88 has been observed by the neutron scattering experiments for csnicl@xmath1 @xcite .
this energy gap can be reproduced in the @xmath25-mode excitation by introducing an anisotropy with @xmath89 using our approximation of eq .
( 6 ) , also plotted in fig .
3 . if we use the simple form of ref .
@xcite corresponding to setting @xmath90 in eq.(6 ) , the required anisotropy is reduced by a little more than half with the value @xmath91 .
both of these values are now considered too large for csnicl@xmath1 which has negligible anisotropy .
the conclusion is that the observed gaps are not of the transverse spin - wave spectra , but belong to the longitudinal modes , as first proposed by affleck @xcite . with @xmath69 , @xmath70 and @xmath87 thz , along the symmetry direction @xmath92 , @xmath93 and @xmath94 .
also included is the gapped @xmath25-mode ( black , denoted as @xmath95 ) with @xmath89 using the anisotropy term of eq .
the solid and dash with the blue color on the lines indicate the @xmath96-mode and @xmath97-mode respectively . ] now we turn our attention to the order parameter .
the long - range order of the quasi-1d hexagonal systems is given by the three sublattice - magnetizations with the same magnitude but different orientations as shown in fig . 1 , and it is clearly noncollinear , contrast to the collinear case of the bipartite systems . in the spin - wave theory with one boson method as described above , the magnitude of the sublattice magnetization can be expressed as @xmath98 where the quantum correction @xmath99 is the magnon density defined as the ground - state expectation value of the boson number operator @xmath100 with @xmath56 and @xmath57 defined by eqs . ( 11 ) .
the numerical result of the magnon density for csnicl@xmath0 is @xmath101 at @xmath87 , giving the sublattice magnetization @xmath102 . on the other hand , using slightly different parameter @xmath103 from ref .
@xcite , we obtain @xmath104 , giving @xmath105 .
both these results compare favorably with the experimental value of @xmath106 @xcite .
as mentioned earlier , our microscopic analysis of the longitudinal modes is based on these mangon density fluctuations and there will be two such modes as discussed in details in the following section .
as mentioned before , the longitudinal excitations in a quantum antiferromagnetic system with a nel - like long - ranged order correspond to the fluctuations in the order parameter .
using the fact that the quantum correction in the order parameter is given by the magnon density @xmath99 as discussed previously in eq .
( 17 ) , the longitudinal modes can be considered as the magnon - density waves . by analogy to feynman s theory on the low - lying excited states of the helium-4 superfluid @xcite , the longitudinal excitation states can be constructed by employing the magnon - density operators @xmath107 , in contrast the transverse spin - wave excitation states constructed by the spin - flip operators @xmath108 @xcite .
the energy spectra of these longitudinal collective modes can then be easily derived by a formula first employed by feynman for the famous phonon - roton spectrum of the helium superfluid involving the structure factor of its ground state . more specifically , following feynman
as described above , the longitudinal excitation state is approximated by applying the magnon density fluctuation operator @xmath109 to the ground state @xmath110 as , latexmath:[\[\label{eq19 } by the fourier transformation of @xmath107 operators , @xmath112 with index @xmath36 running over all lattice sites .
the condition @xmath113 in eq .
( 20 ) ensures the orthogonality to the ground state . the energy spectrum for the trial excitation state of eq .
( 19 ) can be written as @xmath114|\psi_g\rangle}{\langle\psi_e|\psi_e\rangle}\ ] ] where @xmath115 is the hermitian of @xmath109 and where in the second equation we have used the ground state equation , @xmath116 .
we notice that operator @xmath117 in @xmath109 of eq .
( 20 ) is a hermitian operator , hence @xmath118 . by considering the similar excitation state @xmath119 with the energy spectrum @xmath120 , it is straightforward to derive @xcite , @xmath121 where @xmath122 is given by the ground - state expectation value of a double commutator as @xmath123\rangle_g,\ ] ] and the state normalization integral @xmath124 is the structure factor of the lattice model @xmath125 the notation @xmath126 in eqs . and indicates the ground - state expectation .
in fact , the excitation state of eqs .
( 19 ) and ( 20 ) can also be viewed as the single - mode approximation ( sma ) @xcite and the expression for @xmath127 is actually the exact first moment of the dynamic longitudinal structure factor .
we also like to point out that the relation between the longitudinal magnon - density waves and the quasiparticle magnon modes can be examined by the first commutation of the operator ( 20 ) with the hamiltonian and that the magnon - density waves represent the coherent motion of the spin @xmath128 magnon pairs , very similar to the plasmon excitations in the electronic systems with coherent motion of quasi - electron - hole pairs as discussed in details in our earlier paper @xcite .
further support for the form of eq .
( 19 ) can also be obtained by examining the general structures of the ground and excited states within the framework of the coupled - cluster method ( ccm ) @xcite .
briefly , within the ccm , the ground state is given by applying an exponentiated correlation operator @xmath129 on a reference state @xmath130 ( i.e. , the classical nel state in our case ) as @xmath131 with the multiparticle creation operator @xmath132 and the corresponding variational coefficients @xmath133 .
in our case here , @xmath132 is given by the products of the spin - flip operators @xmath134 with respect the nel state .
the excitation state within the ccm is given by the linear form as @xcite @xmath135 with the variational coefficients @xmath136 .
in fact , the spin - wave ground state as discussed in sec .
ii can be deduced by a low - order , the so - called sub2 approximation involving the two - body correlations , in the large-@xmath3 limit of the ccm @xcite .
furthermore , using the following algebra @xmath137 + \frac1{2!}[[s^z_l,{\hat s}],{\hat s}]+\cdots\end{aligned}\ ] ] where the nested commutation series in @xmath138 terminates at the first order in our case , it is not difficult to show the similarity between the excitation state of eqs .
( 19 - 20 ) and that of eq .
the clear advantage of eqs .
( 19 ) and ( 20 ) lies on its simple form and on the fact that the double commutation in @xmath122 of eq .
( 22 ) reduces the order of calculations .
furthermore , it satisfies the sum rule as described above in the ( sma ) .
we have applied these formulas to the bipartite quasi-1d antiferromagnetic systems such as kcuf@xmath1 @xcite . for the hexagonal lattice systems as discussed here , we expect that there are two longitudinal modes due to the noncollinear nature of the order parameter on the triangular basal plane . within the one - boson approach after the two spin rotations as employed here
, the two longitudinal modes with @xmath28-polarizations of the hexagonal systems can be obtained by folding of the wavevectors in the energy spectra of eq .
( 21 ) , in similar fashion to the one - boson spin - wave theory as discussed in sec .
ii and also to that of ref .
@xcite . using the hamiltonian of eqs .
, it is straightforward to derive the following double commutator with zero anisotropy ( i.e. @xmath87 ) as @xmath139 , \end{split}\ ] ] where @xmath59 is as defined in eq . and the transverse correlation functions @xmath140 and @xmath141 are defined respectively as @xmath142 all independent of index @xmath36 due to the lattice translational symmetry .
also , the contribution from the three - boson operators with @xmath143 ( the so - called cubic term ) is zero .
we notice that this cubic term has been included in perturbation theory for the correction in spin - wave spectrum @xcite . in evaluating @xmath140 and @xmath144 of eq . , we keep up to the second order in the large-@xmath3 expansions and obtain @xmath145 where @xmath146 and where @xmath147 with @xmath56 and @xmath57 as given before by eqs . .
the structure factor within the linear spin - wave approximation is independent of @xmath3 , and is given by @xmath148 we like to point out that , the calculations of both eqs . and involve up to four - boson operators .
we first discuss the general behaviors of the longitudinal spectrum of eq .
( 21 ) as a function of the ratio of the two nearest - neighbor coupling constants , @xmath149 . in the limit @xmath150 ,
the hamiltonian of ( 1 ) becomes the pure 1d systems ; the longitudinal spectrum is gapless and identical to the doublet spin - wave spectra thus forming a triplet excitation state as discussed in details ref .
this demonstrates the limitation by the spin - wave ground - state employed , particularly when applied to the integer - spin heisenberg chain where the haldane gap is expected as discussed in sec .
i. in the other limit , @xmath151 , the hamiltonian is a pure triangular antiferromagnet with the quasi - gapped longitudinal modes as discussed in details in our previous paper @xcite where we keep only the first order term in eq . in the large @xmath3-expansion ,
similar to the case of the square lattice model . as derived from eq .
together with the spin - wave @xmath25- and @xmath85 modes as derived from eq . for csnicl@xmath0 along the symmetry direction @xmath92 , @xmath93 and @xmath94 .
the longitudinal modes @xmath152 calculated from the first order approximation and after including the second term in eq . are indicated by the dash and solid lines respectively . ] along path @xmath153 of the hexagonal brillouin zone of fig .
[ fig2](b ) together with the spin - wave @xmath25 and @xmath97 modes for csnicl@xmath0 .
the longitudinal modes @xmath152 calculated from the first order approximation and after including the second term in eq . are indicated by the dash and solid lines respectively . ] for the quasi-1d materials with intermediate values of @xmath72 , the spin - wave ground state is a reasonable approximation .
we obtain nonzero energy gaps for the longitudinal excitation spectra of eq .
( 21 ) . as discussed before , following affleck @xcite , two longitudinal modes for the quasi-1d hexagonal antiferromagnets can be obtained by folding of the wavevector .
we denote one as @xmath154 with the spectrum @xmath155 and the other as @xmath156 with the spectrum @xmath157 .
we plot these two longitudinal spectra in the first and second order approximations together with the three spin - wave spectra of eq .
( 15 ) in fig . 4 near the magnetic wavector @xmath158 for the compound csnicl@xmath0 .
our numerical result for the energy gap of the lower longitudinal mode @xmath154 at @xmath158 is @xmath159 in the first order approximation in eq . .
after including the second order terms the energy gap value is now @xmath160 , in agreement with the experimental results of @xmath88 .
we also notice that the upper mode @xmath156 is higher than the @xmath154 mode by about @xmath161 at @xmath158 .
we also plot the @xmath154 mode along the path @xmath153 of the hexagonal brillouin zone in fig .
[ fig5 ] together with the spin - wave @xmath25 and @xmath97 modes . as can be seen
, the longitudinal mode is nearly flat over the whole spectrum .
for the compound rbnicl@xmath0 also with @xmath162 , using the exchange parameters @xmath163 and @xmath164 thz with a larger ratio @xmath165 @xcite , we obtain similar longitudinal modes as those of csnicl@xmath0 .
the numerical result for the energy gap of the @xmath154 mode is 1.16 thz in the first order approximation and 0.69 thz after including the second order contributions at the magnetic wavevector .
this later result is in better agreement with the experimental result of about 0.51 thz .
we like to point out that there is some difficulty in fitting of affleck s model with the experimental results for rbnicl@xmath0 @xcite .
finally we turn to the longitudinal modes for the non - integer - spin quasi-1d hexagonal systems .
the superexchange interactions in the hexagonal compound csmni@xmath1 can be described by the hamiltonian of ( 1 ) with spin quantum number @xmath166 and the nearest - neighbor coupling constants @xmath167 and @xmath168 thz and negligible anisotropy @xcite .
this system is very close to the pure 1d system with a very small ratio @xmath169 .
the linear spin - wave theory may be a poor approximation for such a system . nevertheless , with a similar analysis as before based on the spin - wave ground state ,
we obtain the @xmath154 mode energy gap value of @xmath170 thz at the magnetic wavevector @xmath158 in the first order approximation , and of @xmath171 thz after including the second order contributions .
this later value is still much larger than the experimental value of about @xmath172 thz by harrison _
et al _ @xcite , which was used to fit a modified spin - wave theory by plumer and cail @xcite .
clearly , for such systems as csmni@xmath1 , we need a better ground state than that of the spin - wave theory in our analysis .
in this paper , we have investigated the excitation states of the quasi-1d hexagonal systems as modeled by the anisotropic heisenberg hamiltonian with only the nearest - neighbor couplings .
we have obtained the three spin - wave modes and two longitudinal modes .
the energy gaps due to the anisotropy and the energy gaps of the longitudinal modes at the magnetic wavevector are investigated and compared with the experimental results for several quasi-1d hexagonal compounds .
we have also estimated the higher - order contributions in the large-@xmath3 expansions for the longitudinal energy spectra .
we like to emphasize that our analysis applies to both integer and non - integer spin systems and there are no other fitting parameters than the nearest - neighbor coupling constants and the anisotropy parameter in the model hamiltonian provided by experiments .
therefore , the good agreement for the longitudinal energy gap values between our calculations and the experimental measurements for the compounds csnicl@xmath1 and rbnicl@xmath1 are particularly satisfactory .
the compound csmni@xmath1 is very close to the pure 1d model ( i.e. , @xmath72 very small ) for which the spin - wave ground state is not reliable .
it is therefore not surprising to find big discrepancy between our estimate based on spin - wave ground state and the experimental result even after including the higher - order contributions in our calculations .
further improvement may be found on two fronts .
firstly , the contribution of the cubic term may be calculated by a perturbation theory in a similar fashion as employed in ref .
@xcite . secondly , a better ground state is needed , particularly for the compound csmni@xmath1 where the interchain coupling is particularly weak . a more sophisticated many - body theory such as the coupled - cluster method , particularly its recent variational version @xcite , may provide such an improvement .
w. j. l. buyers _
lett . * 56 * , 371 ( 1986 ) ; r. morra , phys . rev .
b * 38 * , 543 ( 1988 ) ; z. tun , w. j. l. buyers , r. l. armstrong , k. hirakawa , and b. briat , phys . rev .
b * 42 * , 4677 ( 1990 ) .
m. steiner , k. kakurai , j. k. kjems , d. petitgrand , and r. pynn , j. appl .
phys * 61 * , 3953 ( 1987 ) ; i. a. zaliznyak , l .- p .
regnault , and d. petitgrand , phys .
b * 50 * , 15824 ( 1994 ) ; i. a. zaliznyak , s .- h .
lee , and s. v. petrov , phys .
lett . * 87 * , 17202 ( 2001 ) . | we investigate the excited states of the quasi - one - dimensional quantum antiferromagnets on hexagonal lattices , including the longitudinal modes based on the magnon - density waves .
a model hamiltonian with a uniaxial single - ion anisotropy is first studied by a spin - wave theory based on the one - boson method ; the ground state thus obtained is employed for the study of the longitudinal modes .
the full energy spectra of both the transverse modes ( i.e. , magnons ) and the longitudinal modes are obtained as functions of the nearest - neighbor coupling and the anisotropy constants .
we have found two longitudinal modes due to the non - collinear nature of the triangular antiferromagnetic order , similar to that of the phenomenological field theory approach by affleck .
the excitation energy gaps due to the anisotropy and the energy gaps of the longitudinal modes without anisotropy are then investigated .
we then compare our results for the longitudinal energy gaps at the magnetic wavevectors with the experimental results for several antiferromagnetic compounds with both integer and non - integer spin quantum numbers , and we find good agreement after the higher - order contributions are included in our calculations .
pacs numbers : : 75.10.jm , 75.30.ds , 75.50.ee . |
blue compact dwarf ( bcd ) galaxies provide evidence that isolated , small , low - metallicity galaxies may experience vigorous star formation in the present epoch , challenging our current understanding of galaxy formation .
the very nature of bcds remains puzzling in spite of thirty years of research since they were first recognized as an interesting class of galaxy @xcite .
for instance , it is still debated whether some primeval " bcds are perhaps in the process of undergoing their very first starburst , or whether bcds undergo several episodic starbursts separated by long quiescent intervals , what triggers the star formation in bcds , and whether starburst - induced supernova - driven winds can cause the transformation of bcds into one of the other known types of dwarf galaxy ( * ? ? ? * for a recent review ) .
while star - formation rates ( sfrs ) are central to discussions of the evolution of bcds , there are surprisingly few quantitative sfrs for bcds available in the literature .
the earliest and most quoted quantitative sfrs are based on direct measures of the ultraviolet luminosity of massive stars in formation .
@xcite modeled the international ultraviolet explorer spectra of seven bcds .
the implied sfrs for a salpeter initial mass function ( imf ) with index 2.35 and mass cutoffs at 0.1 and 100m@xmath3 ( which we adopt for all measurements given in this paper ) range from about @xmath6 to 25m@xmath3yr@xmath4 .
subsequent efforts preferentially used the reprocessed ultraviolet radiation of massive stars as measures for bcd sfrs , derived from either dust , or ionized gas emission .
@xcite investigated a sample of 15 bcds , and derived sfrs based on far - infrared ( fir ) and h@xmath7 luminosity .
their values range from about 0.1 to 30m@xmath3yr@xmath4 .
@xcite studied the sfrs of a sample of 48 compact star - forming dwarf galaxies in voids . using spectroscopically derived h@xmath8 luminosities , they determined values that range from 0.02 to 2.2m@xmath3yr@xmath4 , with a mean sfr of about 0.5m@xmath3yr@xmath4 .
most recently , @xcite derived the sfrs of 27 bcds from the second byurakan survey ( sbs ) .
their sfrs are based on fir ( 60@xmath0 m ) and radio continuum ( 1.4ghz ) luminosities .
the sfrs derived from fir luminosity span the range from about 0.9 to 80m@xmath3yr@xmath4 .
radio luminosities were converted to sfrs assuming that the continuum is due to thermal , free - free emission . these sfrs range from approximately 2.5 to 270m@xmath3yr@xmath4 , with a mean of about 18m@xmath3yr@xmath4 .
the values available in the literature suggest that the sfrs of bcds might vary over five orders of magnitude , from 10@xmath1 to 10@xmath9m@xmath3yr@xmath4 .
however , the mean sfrs given in the existing investigations are noted to also vary over two orders of magnitude , from @xmath10m@xmath3yr@xmath4 to 10m@xmath3yr@xmath4 .
it is therefore not clear whether and to what extent the wide range of bcd sfrs results from the use of different techniques , assumptions , or sample selections , employed in the available studies .
clearly , a systematic exploration of the parameter space of bcd sfrs is both warranted and timely .
accurate sfrs are needed before any correlations of the sfr with other physical properties can be discussed .
this paper is meant to be a first step toward establishing a baseline of systematic , quantitative sfrs for bcds .
it was motivated by our desire to measure the sfrs of bcds using a uniform approach and by employing techniques that are prone to minimal systematic errors . specifically , as star - formation may occur throughout bcds and not just in their very active centers , we excluded spectroscopic methods .
these require corrections for the ( unknown ) amount of flux missed outside of the aperture .
we also wanted to use a measurement that is free from the potentially large and uncertain corrections imposed by internal and foregound dust extinctions .
this immediately also ruled out ultraviolet continuum and optical emission line indicators of the sfr .
we therefore decided to examine the sfrs of bcds based on their radio continuum and fir luminosities .
fortunately , public databases of fir and radio flux densities that include bcds are readily at hand . for completeness , and to investigate the extent of and role played by obscuration in our sample , a subsequent study to be presented in paper ii will explore the sfrs derived from very wide - slit h@xmath7 spectroscopy .
the sample under investigation consists of 50 of the most well known and well studied bcds .
it is in this regard that our sample consists of typical " bcds , and that the resulting sfrs are called typical " . by no means
does this investigation present an unbiased or a complete sample of bcd sfrs .
the present sample comes from those bcds that have been well studied because they were used to derive the primordial helium abundance , and have been taken from a recent compilation of 54 objects ( * ? ? ?
* hereafter it ) . for all of these galaxies , accurate redshifts , needed to calculate luminosities , are available .
the bcds in this sample also have careful determinations of the oxygen abundance of the ionized gas . since oxygen is produced by massive stars , and since massive stars are used as a quantitative measure of the sfr , a particular emphasis of the present paper is to investigate any correlations between these two quantities .
observational information for the galaxies in this study is presented in table [ tab1 ] .
this contains positions ( j2000 ) , @xmath11 m flux densities from iras ( infrared astronomical satellite ) catalogues , @xmath12ghz flux densities from the nvss ( nrao vla sky survey ) and first ( faint images of the radio sky at twenty centimeters ) catalogues and images , and heliocentric velocities .
much of this information was obtained from ned .
four of the galaxies in table [ tab1 ] , mrk 71 , sbs 1319 + 579 , sbs 1331 + 493 and sbs 1533 + 574 , were split into multiple components for their metallicity measurements ( * ? ? ?
* their table 1 , which has 54 entries ) .
since the @xmath11 m ( iras ) and 1.4ghz ( nvss ) flux densities do not have fine enough resolution to distinguish between these components , these galaxies are treated as single objects in the current study .
the metallicities ( shown in table [ tab2 ] ) for those objects are the average of the values measured for the individual components ( it ) , while the errors reflect the ranges in values .
the majority of the 50 bcds in the sample had their distances estimated based on heliocentric velocities taken from ned .
the velocities given in ned were derived from emission line or hi measurements made predominantly by izotov , thuan and collaborators @xcite , or taken from the rc3 @xcite . in cases where velocities for several components were listed , we assumed the velocity that is associated with the main body or main component of the galaxy .
distances for all but three of our sample were estimated from velocities by using a virgocentric flow model @xcite .
we adopted a hubble constant of @xmath13kms@xmath4mpc@xmath4 @xcite and a virgo distance of @xmath14mpc @xcite .
the three exceptions , where distances were estimated from analysis of the i - band brightness of tip of the red giant branch ( trgb ) stars in color - magnitude diagrams ( cmds ) of stellar populations , are ugc 04483 @xcite , vii zw 403 @xcite and mrk 0209 ( i zw 36 ) @xcite .
we present histograms showing the distributions of distances and heliocentric velocities for our bcd sample in figure [ disthist ] .
the ratios of the distances from the virgocentric flow model to the trgb distances for ugc 04483 , vii zw 403 and mrk 0209 are 1.69 , 0.23 , and @xmath151.2 , respectively .
this indicates a considerable misjudgment of the distances based on velocities and the virgocentric flow model is possible .
we believe that distance errors can potentially amount to up to a factor of a few , but probably do not exceed an order of magnitude .
it is this difficulty in assessing distance errors that led us to omit errors on the sfrs derived below .
the infrared astronomical satellite has proved an immensely valuable resource in exploring the infrared properties of galaxies since its launch in 1983 .
iras products include 60@xmath0 m flux densities for galaxies from the point source catalog ( brighter than about 0.5jy ) and faint source catalog ( typically brighter than 0.2jy ) .
these fluxes were obtained for the bcd sample , where available , from ned and from the irsa issa image server and catalog overlay service .
the resolution of the iras images at 60@xmath0 m is about @xmath16 with typical positional accuracies better than @xmath17 . of the 50 bcds in our sample 18
are detected at 60@xmath0 m by iras .
we have not made direct measurements for the flux density upper limits of the non - detections , but for most of these the 0.2jy limit of the faint source catalog should be indicative . the nrao vla sky survey @xcite is a 1.4ghz radio continuum survey covering the sky north of @xmath18 declination , made using the very large array ( vla ) operated by the national radio astronomy observatory ( nrao ) .
the observations were made using the compact d and dnc configurations of the vla .
the resulting images have an rms noise level of about 0.5mjybeam@xmath4 and a fwhm resolution of @xmath19 .
positional accuracies ( from the source measurements using a gaussian fitting algorithm ) range from about @xmath20 for sources stronger than 15mjy to @xmath21 at the survey limit .
the catalog lists sources detected above a peak brightness of 2mjybeam@xmath4 . for 13 of the galaxies in our bcd sample
an nvss catalog source is close enough ( @xmath22 ) to be a probable identification , and the flux measurement from the catalog was taken .
for another 10 of the sample galaxies , where the nvss fits image showed a local peak in intensity at the position of the galaxy , but no catalog source was present , we performed our own source measurements .
we used the miriad ( multichannel image reconstruction , image analysis and display ) task sfind @xcite to make these measurements , and the sources so detected are typically at the @xmath23 level . for non - detections , we measured the local rms noise level in the vicinity of the bcd galaxy and adopted 5 times this value as the upper limit on the nvss flux density .
uncertainties in the flux density measurements of the nvss catalogue have been described in detail ( * ? ? ?
* see also their figure 31 ) . in practice
, these uncertainties rise from @xmath24 at flux densities of 10mjy to @xmath25 close to the survey limit .
these uncertainties translate to uncertainties in estimated star - formation rates ( see section [ sfr ] ) .
digitized sky survey images have been overlayed with 1.4ghz contours from nvss for each of the 23 detected bcds ( figure [ nvss1 ] ) .
first , faint images of the radio sky at twenty centimeters @xcite , is a 1.4ghz radio continuum survey currently covering about 8500 square degrees of the north galactic cap .
the observations were taken using the vla in b configuration .
the images have a typical rms noise level of 0.15mjy with a fwhm resolution of @xmath26 .
source positional accuracies vary from @xmath27 at the 3mjy level to about @xmath20 for the faintest sources .
the source detection threshold of the survey is 1mjy . because of the much higher resolution of first compared to nvss
, some unresolved nvss objects are resolved by first into more than one component ( um 311 is one example ) . in this case
we have taken the sum of the first flux densities for the individual components as the first measurement for the whole source . from the first catalog ,
6 galaxies were matched with our bcd sample , and we obtained upper limits from the fits images for another 14 .
the flux density upper limits are again taken to be 5 times the local rms value .
the remaining 30 bcds lie outside the area of the first survey .
1.4ghz contours from first are shown on the images in figure [ nvss1 ] for each of the 6 detected bcds .
an important advantage from the higher resolution of first images is being able to identify the positional origin of the radio emission with greater accuracy than allowed by nvss .
the different resolutions are similarly an important effect to account for if calculating sfr per unit area , and also allow for distinguishing between the object of interest and confusing background sources . in the case of mrk 0209 , for example
, the first emission appears to originate slightly to the se of the region of brightest optical emission .
this may initially suggest a background source for the radio emission , but the strong consistency between the radio - derived sfr and numerous other optically - derived values for this galaxy , ( * ? ? ?
* see also discussion in section [ fircor ] ) , argue that in fact this radio source is likely to be a star - forming knot in the outskirts of mrk 0209 .
similarly , cg 0798 shows a double - peak in the nvss emission , and a first source about an arcminute to the south of the galaxy accounts for some of this emission . lacking more detailed observations ,
it is certainly possible that this may be a coincidental background source and the radio flux - density and the derived sfr for cg 0798 should strictly be treated as upper limits . given the good agreement between the radio and fir sfrs for this galaxy , however , we have chosen to treat it as a detection for the current analysis .
to avoid the effects of both spectroscopic flux losses and uncertain obscuration corrections , we use 1.4ghz and 60@xmath0 m luminosities as estimators of the star - formation rate for galaxies in our bcd sample .
here we adopt the calibrations @xmath28
@xcite for the sfr derived from the @xmath11 m luminosity , and @xmath29 @xcite for the sfr from the 1.4ghz luminosity . in both cases
the factor of 5.5 converts from the mass range @xmath30m@xmath3 to the mass range @xmath31m@xmath3 , assuming a salpeter imf ( * ? ? ?
* for example ) .
also , @xcite point out that starburst galaxies may have imfs that differ from the one adopted in equation [ eq2 ] . in the worst case a maximum decrease of the sfr by a factor around 5 is possible .
any changes to the slope of the imf , however , should affect the radio and fir luminosities equally .
the sfrs calculated for our bcd sample are given in table [ tab2 ] along with the distances to the galaxies .
the metallicities ( oxygen abundance ) also shown have been taken from it .
hi masses have been derived using hi flux densities from the literature with the distances given here , and are also shown in table [ tab2 ] .
distance uncertainties ( possibly up to a factor of a few ) are the dominant source of measurement uncertainty for the sfrs .
for this reason , the figures below do not include error bars on the sfrs .
histograms showing the distribution of the derived sfrs for our sample are presented in figure [ sfrhist ] .
we also do not calculate sfr normalized to the optical area of the galaxy .
this is firstly because of the complex optical morphologies for the objects in our sample ( see figure [ nvss1 ] ) .
secondly , the galaxies in our sample are essentially point sources for the @xmath11 m and 1.4ghz detections . instead
, we investigate below the properties not only of the total sfr , but also of the specific sfr " by which we mean the sfr normalized by hi mass .
this is a distance - independent quantity , and a probe of relative star formation that should be reasonably independent of galaxy size or mass .
details of the specific sfrs and hi masses are given in section [ results ] below .
the above calibrations of sfr from luminosity are derived independently of metallicity , although it is worth noting that the 1.4ghz calibration is based on the observed galactic relation between supernova rate and nonthermal radio luminosity .
an implicit assumption in the case of estimating sfr from the 1.4ghz luminosity is that there is no contribution from an active nucleus ( seyfert or liner activity ) to the radio emission .
this assumption seems to be justified for the sample of bcds we are investigating , since all of them have hii - like spectra .
the nvss catalog yields flux densities for a significant fraction of our sample .
table [ tab2 ] and figure [ sfrhist ] show that the sfrs derived from nvss fluxes cover almost five orders of magnitude .
clearly , there are real differences in bcd sfrs , since the range seen is much larger than the potential distance uncertainties . from the 23 nvss - detected objects we calculate a median sfr of 0.27@xmath5m@xmath3yr@xmath4 .
this compares well with the median sfr , 0.36@xmath5m@xmath3yr@xmath4 , of the 18 iras - detected bcds .
the it sample of bcds being examined , however , is likely closer to being a flux - limited than a volume - limited sample , favoring more luminous galaxies with higher sfrs , and the nvss sensitivity limit strengthens this bias . given that the radio and fir detections are subject to such malmquist bias , artificially raising the median value , the median sfr quoted above should not be over - interpreted . to provide some perspective ,
an approximate uncertainty for this median can be estimated by examining the range of sfrs spanned by the @xmath32 galaxies either side of mrk 1416 ( having the median sfr@xmath33 ) .
this gives a range of an order of magnitude , from 0.13 to 1.3@xmath5m@xmath3yr@xmath4 .
figure [ sfrvssfr ] shows the comparison between sfrs derived from 1.4ghz ( first and nvss ) and 60@xmath0 m luminosities . a good agreement exists between the nvss - based 1.4ghz sfrs and the iras - based 60@xmath0 m sfrs .
it is evident from table [ tab1 ] and this figure that the first flux densities tend to be systematically lower than those from the nvss for most of the galaxies detected by both surveys .
this is also true for many of the first upper limits where an nvss detection exists .
the probable reason for this is the difference between the two surveys in their sensitivity to extended emission .
first , made using the vlas b - array , has much higher angular resolution than nvss , made with the more compact d and dnc arrays , but at the expense of sensitivity to more extended emission .
first images are typically only sensitive to structure extended over less than about @xmath34 , while the nvss images will detect emission from sources up to about @xmath35 across @xcite . as a result the nvss flux density measurements tend to be somewhat larger than those from first , since many of the galaxies in our sample are larger than the extent over which first is most sensitive . in subsequent analysis
where 1.4ghz - derived sfrs are used we thus rely on sfrs derived from nvss measurements , both for this reason and also since more of our sample galaxies have nvss detections .
figure [ sfrvsdist ] presents sfr@xmath36 as a function of distance .
the nvss detection threshold is marked . as may be expected from the tight relation in figure [ sfrvssfr ] , a diagram showing sfr@xmath37 versus distance is very similar .
an interesting point to note from this figure is the absence of any moderate to high sfr objects closer than @xmath38mpc , that is , in the region where hubble space telescope ( hst ) observations
( with wfpc2 or nicmos instruments ) are judged able to resolve individual stars within the target galaxies .
this is mainly a result of high sfr objects having lower space - densities than low sfr galaxies , combined with the small volume being probed at this distance .
the implication here is that hst - based cmd investigations of nearby galaxies are restricted to lower - sfr objects , in good agreement with the low sfrs derived from cmd modeling .
the 1.4ghz flux on which sfr@xmath36 is based is thought to measure the present - day sfr on a similar timescale as the lifetime of massive stars @xcite .
we now explore the correlation between various chemical abundances from it and absolute and specific sfr .
the chemical elements that we emphasize are associated with progenitors of different masses , and this translates into different enrichment timescales .
specifically , oxygen is produced by massive stars and returned quickly to the interstellar medium ( ism ) by way of supernova ( sn ) type ii explosions , nitrogen is thought to be produced in intermediate - mass stars and returned to the ism in the asymptotic giant branch ( agb ) wind phase ( but see it for a contrary view ) , while iron is considered to reflect the ism s pollution by low mass stars through snia events , and should lag the most behind the oxygen enrichment . in the top panel of figure [ zvssfr ] a weak positive trend of 12+log(o / h ) ( from it ) with absolute sfr@xmath36 can be seen .
the spearman rank test shows that the null hypothesis ( no correlation between 12+log(o / h ) and log(sfr ) ) , using only the detections , has a probability of only about 0.6% .
an unweighted linear regression yields 12+log(o / h ) = 0.09log(sfr ) + 7.90 ( with the uncertainty of the slope being @xmath39 = 0.04 and the uncertainty of the intercept being @xmath40 = 0.05 ) .
the extremely metal - poor galaxies i zw 18 and sbs 0335@xmath41052 lie well away from the trend shown by the majority of the sample , effectively lowering the intercept .
if we exclude i zw 18 and sbs 0335@xmath41052 from the sample , than the spearman rank correlation coefficient increases slightly , and the probability of a chance correlation drops to just below 0.5% , while the linear regression gives 12+log(o / h ) = 0.09log(sfr ) + 7.96 ( with @xmath39 = 0.03 and @xmath40 = 0.03 ) .
it is this linear fit that is shown in figure [ zvssfr ] as a solid line , with the dashed lines reflecting the error in the intercept .
( notice the fit results reported here are slightly different from the preliminary results of @xcite because here we use more up - to - date values for the hubble constant and the virgo distance . )
we conclude that bcds with high present - day total sfrs clearly also have high oxygen abundances .
if we replace absolute sfrs with specific sfrs ( i.e. sfr / m@xmath42 ) a weak positive trend seems to remain ( hi masses are discussed below ) .
the spearman rank test , however , indicates the hypothesis that 12+log(o / h ) and log(sfr / m@xmath42 ) are correlated has a 15% probability of occuring by chance . as a result it seems
there is little correlation , if any , between oxygen abundance and sfr / m@xmath42 . in the middle panel of figure [ zvssfr ] a weak positive trend of log(n / o ) ( from it ) with absolute sfr
can be seen .
the extremely metal - poor galaxies i zw 18 and sbs 0335@xmath41052 do not distinguish themselves on this plot , lying well within the distribution of values .
the spearman rank test shows that the probability for a chance correlation is about 0.7% . a linear regression yields log(n / o ) = 0.09log(sfr ) @xmath41 1.42 ( with @xmath39 = 0.02 and @xmath40 = 0.03 ) .
conversely , the spearman rank test for log(n / o ) with the logarithm of the specific sfr , log(sfr / m@xmath42 ) , shows the probability for a chance correlation is 5% .
lastly , we investigate the dependence bewteen [ o / fe ] ( from it ) and log(sfr ) or log(sfr / m@xmath42 ) ( see bottom panel of figure [ zvssfr ] ) .
the spearman rank test gives such small correlation coefficients that the data are likely to be uncorrelated .
the linear regression of [ o / fe ] on log(sfr ) results in a constant , [ o / fe ] = 0.43 ( with @xmath39 = 0.04 and @xmath40 = 0.04 ) .
figure [ zvssfr ] shows that sbs 0335@xmath41052 is quite removed from the rest of the sample bcds , but i zw 18 lies well within the distribution .
figure [ sfrvsw20 ] explores the dependence of various quantities on galaxy mass .
the width of the hi emission line profile at 20% of peak , w@xmath43 , is taken as an indicator of galaxy total mass .
hi profile widths are often used to calculate galaxy total masses , but require radii and inclinations . as can be seen from figure
[ nvss1 ] , radii and inclinations are difficult to judge for bcds because the star - forming regions dominate their appearance .
w@xmath43 , on the other hand , also may not be a solid indicator for rotational velocity if turbulent motions are significant .
with these caveats in mind , we plot in the upper panel of figure [ sfrvsw20 ] the hi mass versus the w@xmath43 velocity .
hi and w@xmath43 data were taken from @xcite ( circles ) or @xcite ( triangles ) .
the spearman rank correlation coefficient of log(m@xmath42 ) and log(w@xmath43 ) is high , indicating a less than 0.05% probability that this is a chance correlation . in other words
, we can be confident that hi mass increases as total mass increases .
a linear regression gives log(m@xmath42 ) = 3.6log(w@xmath43 ) + 1.3 ( with @xmath39 = 0.6 and @xmath40 = 1.3 ) , and is shown in figure [ sfrvsw20 ] .
the spearman rank test for total sfr and total mass rejects the null hypothesis ; correlation by chance has about a 0.5% probability ( middle panel of figure [ sfrvsw20 ] ) .
the linear regression is log(sfr ) = 4.8log(w@xmath43 ) @xmath41 10.5 ( with @xmath39 = 1.0 and @xmath40 = 2.0 ) .
the higher sfrs are clearly being hosted by higher mass galaxies , although there is a significant scatter in the trend ( for a given rotational velocity or total mass , the sfr may vary over more than an order of magnitude ) .
finally , we also investigate the correlation between metallicity and total mass .
the lower panel of figure [ sfrvsw20 ] shows that the lowest metallicity objects , such as i zw 18 or sbs 0335@xmath41052 , do not have the lowest rotational velocities .
indeed the spearman rank test indicates that the probability of a chance correlation between 12+log(o / h ) and log(w@xmath43 ) is quite high , almost 10% .
a higher total mass does not necessitate a higher oxygen abundance .
figure [ sfrvsmhi ] shows the variation of sfr@xmath36 with hi mass .
the data were taken from @xcite ( circles ) and from @xcite ( triangles ) . because of our interest in extremely low - metallicity objects , and because it is one of the bcds with a trgb distance , we added ugc 04483 to the figure , using the hi flux from @xcite .
this data point is shown as a square .
the higher sfrs are clearly being hosted by galaxies with higher hi masses although there is a significant scatter in the trend ( again , as with w@xmath43 , for a given hi mass the sfr may vary over more than an order of magnitude ) .
this scatter is reflected in the wide range of timescales for gas consumption , from 0.3gyr to more than 10gyr , indicated by the solid lines in figure [ sfrvsmhi ] .
i zw 18 and sbs 0335@xmath41052 do not stand out in any way from the rest of the sample in this picture .
the results can be summarized as follows : 1 .
the median sfr in the it sample is 0.3@xmath5m@xmath3yr@xmath4 ( with an uncertainty up to a factor of three or so , and it may be an upper limit due to malmquist bias ) .
total sfrs in the it sample range over at least four orders of magnitude .
3 . sfrs calculated from iras and from nvss data are consistent with sfr@xmath37 = sfr@xmath36 , i.e. , these galaxies obey the fir / radio correlation valid for large spirals . 4 .
while bcds in this sample are close , [email protected] , there is still a lack of high - sfr galaxies within the range of hst single - star photometry ( about 10mpc ) .
this is because such objects have space densities low enough that they are unlikely to appear in the small volume probed out to this distance .
we find a high probability of a positive correlation between 12 + log(o / h ) or log(n / o ) and total sfr , but [ o / fe ] and total sfr appear to be uncorrelated . for all three element abundances ,
the probability of a correlation with sfr / m@xmath42 ( a distance - independent quantity ) is low .
taking w@xmath43 as an indicator of total mass , we find a high probability for a positive correlation between hi mass and total mass , and between total sfr and total mass .
the probability for a correlation of metallicity and total mass , on the other hand , is small .
the typical " timescale for gas consumption in the it bcd sample is a few gyrs .
none of these galaxies is in danger of running out of gas in the very near future .
however , these bcds will die sooner than the typical dwarf irregular galaxy .
we find excellent agreement between sfr@xmath37 and sfr@xmath36 using the calibration equations of @xcite , @xcite , and @xcite .
specifically , and as shown by figure [ sfrvssfr ] , the data are consistent with sfr@xmath37 = sfr@xmath36 .
this consistency , however , does not necessarily imply that the correct sfr has been determined . as an independent check of the sfrs from fir and radio luminosities we compare them with results from modeling of color - magnitude diagrams .
cmd studies have their own but different set of limitations , in terms of sensitivity , color biasing , and fractional galaxy area observed .
we have only two cases in the it sample for which there are both distances and present - day sfrs available from single - star photometry . for vii
zw 403 , sfr@xmath44m@xmath3yr@xmath4 and sfr@xmath45m@xmath3yr@xmath4 , while cmd modeling of hst / wfpc2 v and i photometry yields [email protected]@xmath5m@xmath3yr@xmath4 @xcite . for mrk 209 ( i zw 36 )
the sfr@xmath36 is @xmath47m@xmath3yr@xmath4 , whereas hst / nicmos j and h data were modeled to yield a sfr of @xmath48m@xmath3yr@xmath4 @xcite .
we conclude that cmd modeling validates the fir and radio sfrs to the extent that it is possible with the limited amount of data .
astrophysical interpretations of the fir / radio correlation suppose that both types of emission have their origin in newly forming massive stars .
the fir emission originates from the heating of dust by uv radiation , while the radio emission is produced by synchrotron radiation from cosmic ray electrons accelerated by shocks associated with supernova remnants @xcite .
both sfr@xmath37 and sfr@xmath36 are considered to be measures of the present - day sfr .
the synchrotron component of the radio emission is thought to dominate for a star - forming galaxy observed at 1.4ghz , with a much smaller contribution from the thermal bremsstrahlung of the hii regions . with a typical synchrotron spectrum for the nonthermal component ( @xmath49 ) , less than @xmath50 of the total 1.4ghz flux is thermal @xcite .
equation [ eq2 ] above makes the assumption that supernovae generate similar amounts of radio luminosity to that observed in our own galaxy , and the detection of any nonthermal radio luminosity in external star - forming galaxies implies the presence of magnetic fields .
the calibration above for sfr@xmath36 differs from the calibration given by @xcite .
@xcite studied sfrs in a sample of bcds drawn from the second byurakan survey , but they find sfr@xmath51 sfr@xmath37 .
their calibration for sfr@xmath36 from 1.4ghz luminosity makes the assumption that the radio emission is purely thermal , although they acknowledge that a significant part of the 1.4ghz radio emission should be nonthermal emission related to supernova events .
this explains their overestimation of 1.4ghz luminosity derived sfrs .
the calibration for sfr@xmath37 given by @xcite also differs slightly from the one adopted above ( but see discussion of sfr@xmath52 calibrations given by @xcite @xcite ) .
it can easily be shown that the objects in the sample of @xcite follow the fir / radio correlation , and indeed the sfrs derived for their sample by using the calibrations of equations [ eq1 ] and [ eq2 ] result in highly consistent sfrs from both @xmath11 m and 1.4ghz luminosities .
there are suggestions that the thermal component of the radio luminosity in bcds is larger than that typical for normal spirals .
@xcite found the radio spectra of bcds are generally flatter than those of spiral galaxies .
they comment that they find it surprising that bcds do follow the same fir / radio correlation as do normal spirals , given the strongly varying sfrs in bcds ( further discussed below ) and the lack of cosmic - ray confinement .
they suggest that a lower ratio of nonthermal - to - thermal radio emission in bcds could be offset by a similar deficiency in fir emission .
@xcite re - observed a few of these bcds and also found a variety of shapes for their radio spectra , but conclude that a nonthermal component is present in most . clearly , these data suggest that the radio spectra of bcds at 1.4ghz have differing amounts of nonthermal - to - thermal emission , hence it is indeed surprising that our investigation yields identical fir and 1.4ghz sfrs when the calibrations of equations [ eq1 ] and [ eq2 ] are applied .
there are no linear radio polarization observations of bcds in the literature , and therefore we have no direct measurements of their large - scale magnetic fields .
the presence of a one - to - one fir - radio correlation in our bcd sample suggests that magnetic fields in these systems need to be of at least sufficient strength to extract synchrotron radiation from relativistic electrons accelerated by supernova remnants .
the conversion of @xmath53 luminosity to sfr@xmath37 rests on the assumption that dust absorbs well in the uv , and that it re - radiates effectively in the ir , so that the ultraviolet radiation of massive , young stars is absorbed and re - emitted as fir luminosity .
clearly the conversion factor must depend on the amount of dust present .
there has been a persistent notion in the literature that metal - poor dwarf galaxies are virtually dust free .
this is not the case .
we emphasize that although our table [ tab2 ] shows no fir detections for i zw 18 and sbs 0335@xmath41052 , it is now known that both contain appreciable amounts of dust @xcite . in any event , for the sfr@xmath37/sfr@xmath36 ratio to be unity ,
it is not as much the total amount of dust in dwarfs that matters , but rather , that there is sufficient dust to produce a fir luminosity that offsets the radio luminosity .
this idea supports the suggestion that the local properties of individual star - forming regions are more important in generating the fir / radio correlation than the global properties of the galaxies that host them .
is metallicity likely to affect the calibrations of sfr to luminosity ?
@xcite discuss dust - to - gas ratios as a function of metallicity in dwarf galaxies .
@xcite incorporate metallicity into the conversion factor , and find it can vary by about a factor of 5 through differences in dust content ( dust - to - gas ratios ) between our galaxy and others .
in order to investigate whether there is any metallicity bias in the sfrs we derived for bcds , we display in figure [ sfrratio ] sfr@xmath36/sfr@xmath37 versus 12+log(o / h ) .
this ratio spans a range of about 0.3 to 1.8 for metallicities of 12+log(o / h ) = 7.7 to 8.3 .
as can be seen from this figure we do not have measurements at both frequencies for a number of the extremely metal - poor objects . for three objects with a @xmath11 m detection but a 1.4ghz upper limit
, we see upper limits to this ratio between 0.8 and 1.5 . for six objects with 1.4ghz detections but @xmath11 m upper limits
, this ratio shows lower limits ranging from 0.6 to 2.5 .
both i zw 18 and sbs 0335@xmath41052 fall into the latter category and have lower limits of 1.5 .
the mean metallicity of the detected sample is 12+log(o / h ) = 7.98 .
it use the scale on which the sun has 12+log(o / h ) = 8.91 , so we are effectively exploring the fir / radio correlation at about 0.1 of the solar metallicity . in our sample , the mean sfr@xmath36/sfr@xmath37 is [email protected] , identical to the expected value of 1 . no trend with metallicity
is seen . in starburst galaxies with highly episodic star - formation histories ( sfh ) , a time - delay between the maximum of the fir and
the nonthermal radio emission has been hypothesized @xcite .
thermal radiation and re - radiation respond instantaneously to the presence of newly born massive stars , whereas nonthermal radiation does not occur until the end of the life of massive stars , when they explode as sne , some 10@xmath54 yrs later .
if time - delays were important we would expect a large scatter of sfr@xmath36/sfr@xmath37 about the mean , depending on exactly when we observe a galaxy in its changing sfh .
indeed , bcds might be suspected of this effect if they experience highly episodic starbursts , whereby very short bursts ( of about 3.5 myr , * ? ? ?
* ) are separated by very long quiescence periods of a gyr or more ( but see * ? ? ?
@xcite recently modeled the size of the offset in the h@xmath7 luminosity to sfr@xmath55 calibration that would arise due to episodic starbursts .
the conversion factor ranges over two orders of magnitude for the benchmark sfhs that they explore .
we observe a 1 @xmath56 variation of @xmath460.40 about the mean ratio for sfr@xmath36/sfr@xmath37 , a much smaller scatter , although we admittedly are investigating an incomplete sample .
it is possible that time - delay effects contribute to this scatter about the mean , but they clearly do not dominate the conversion factor . to summarize , our finding of a metallicity - independent fir / radio ratio of unity implies that the conversion factors that are currently being used to turn fir and radio luminosity to sfr seem to remain valid at 0.1 of the solar metallicity .
( in paper ii , we shall explore the conversion factor that translates h@xmath7 luminosity to sfr@xmath55 for these objects . )
we also infer that bcds are required to have supernova rates , radio luminosities per sn , magnetic fields , and dust contents in some scaled proportion to those of the large spirals , to which the fir and radio calibrations were originally tailored .
a major difficulty for chemical evolution models of bcds has been to explain their low oxygen abundances in spite of their active star formation . in this section
we confront the data assembled here with current theoretical models for dwarf galaxy evolution .
we see a correlation of hi gas mass with total mass , and between total sfr and total mass , in the sense that galaxies with more total mass also have a higher gas mass , and support a higher present - day sfr .
this comes as no surprise since the large range of sfrs among different galaxies is attributed in part to the large range in galaxy masses @xcite .
we see a broad trend of increasing sfr with increasing hi mass , but there is also a wide range of sfrs for a given hi mass .
this is explained as being due to different gas - consumption timescales .
here we have a hint of some variation in the physical properties of bcds .
the gas depletion timescale of bcds in the it sample is typically a few gyrs .
this is an order of magnitude shorter than those in the dwarf irregular sample studied by @xcite .
our data also show a positive correlation between oxygen abundance and sfr , in the sense that galaxies with higher global sfrs also show higher oxygen abundance .
this similarly comes as no surprise as the massive stars that are used to gauge the sfr are also the sites of oxygen nucleosynthesis .
( however , there is some as yet unsettled discussion in the literature as to whether the oxygen so measured is related to the present generation of stars , or whether it stems from prior ones , e.g.@xcite @xcite and references therein . )
while sfr correlates with total mass and oxygen abundance correlates with sfr , there is _ no correlation between oxygen abundance and total mass _ in our sample ( see figure [ sfrvsw20 ] ) .
we further find that sfr / m@xmath42 versus hi mass shows little correlation , if any , and we also find no correlation between oxygen abundance and sfr / m@xmath42 .
@xcite first presented a relationship between metallicity and total mass for a small sample of irregular galaxies and two bcds ( i zw 18 and ii zw 40 ) .
@xcite later derived the well known metallicity - luminosity relation for irregular galaxies .
in irregular galaxies , the oxygen abundance rises with absolute blue magnitude of the host galaxy ( but see * ? ? ? * ; * ? ? ? * for the opposing view ) .
note that @xcite excluded bcds from their sample .
the metallicity - luminosity relation is frequently interpreted as a metallicity - mass relation , to the extent that it is now being used to gauge galaxy assembly in the intermediate - redshift universe @xcite .
indeed , there is a general positive correlation across the hubble sequence of oxygen abundance with m@xmath57 , that holds for over 10 orders of magnitude in m@xmath57 @xcite .
it is not as yet clear that a metallicity - luminosity relation holds for bcds alone ( * ? ? ?
* figure 7 ) .
similarly , our limited data do not support the existence of a mass - metallicity relation for bcds .
since the seminal work of @xcite , the extent to which mass loss affects the evolution of low - mass galaxies has attracted much attention .
such work is particularly interesting since low - mass systems are galactic building blocks in hierarchical scenarios of galaxy formation , and could contribute to the pollution of the intergalactic medium at early epochs .
recent models for dwarf galaxies have addressed the role of dark matter halos for the evolution of low - mass systems .
@xcite find that mass loss occurs in all galaxies with gas masses below @xmath58m@xmath3 , but complete loss of gas ( blowaway ) is possible only for galaxies with very low hi masses , less than a few times 10@xmath59m@xmath3 .
the hi masses in the it sample range from several times 10@xmath59m@xmath3 , to several times 10@xmath60m@xmath3 .
they match extremely well the hi mass range to which the models by @xcite apply .
this suggests that some mass loss occurs in all of the bcds observed here .
@xcite also point out that while the fraction of cool gas lost in dwarf galaxies is generally small , most of the metals mixed with hot gas are able to leave these galaxies .
this helps to keep their metallicities low .
@xcite re - iterate that even in the presence of a metal - enriched outflow , the predominant mechanism for gas consumption is the conversion of gas into stars , and not mass loss .
figure [ zvsmhi ] shows the relation between oxygen abundance and gas ( hi ) mass for the it sample of bcds .
three dashed lines approximate the models of @xcite for dark - to - visible ( hi ) mass ratios of 0 , 10 , and 300 , respectively . our figure may be compared with the equivalent figure for gas - rich irregular galaxies in ( * ? ? ?
* their figure 4 ) .
the data for bcds are generally found to overlap with the model parameters .
the bcd sample also shows considerable overlap with the nearby dwarfs " and low - surface - brightness dwarfs " samples to which @xcite compared their models , consistent with that used by it , is assumed for all three samples . ] .
our data thus lend additional support to the models of @xcite and @xcite .
this also means that most galaxies classed as bcds are dark - matter dominated .
@xcite predict total sfrs for a limited choice of parameters .
these sfrs are quite small , and allow a variation over a few orders of magnitude , owing to differences in dark - to - visible mass ratios and time evolution .
the small global sfrs of bcds that we measure here are broadly consistent with these predictions , although our median sfrs appear to be an order of magnitude higher than their model predictions , as is the fact that we observe a wide range of sfrs .
the effect of malmquist bias on our sample may partially explain this result , biasing our median sfrs to higher values .
@xcite emphasize that there is a considerable spread in oxygen abundances for a given hi mass , and that this can be understood because higher dark - to - visible mass ratios cause a stronger gas compression , increase the sfrs , and in doing so , also increase the oxygen abundances . a large difference in dark - to - visible mass ratios among the bcds
can help us to reconcile the lack of correlation for oxygen abundance versus hi mass and oxygen abundance versus w@xmath43 while , at the same time , sfr correlates with oxygen abundance and also shows a general trend with hi mass , albeit with much scatter .
i zw 18 and sbs 0335@xmath41052 once again stand out from the rest of the bcds in figure [ zvsmhi ] .
they actually fall in a forbidden " area of the model parameter space .
the fact that the two well - studied extremely metal poor systems lie outside these models ( i.e. , in the unphysical region below a dark - to - visible mass ratio of zero ) reinforces their unusual properties .
this discrepancy may be explainable through uncertainties in the hi masses , as , for i zw 18 and sbs 0335@xmath41052 , these only need to be lowered by 12 orders of magnitude to move them close to dark - matter dominated terrain in this figure .
this may be feasible because both i zw 18 and sbs 0335@xmath41052 have complex optical and radio morphologies @xcite .
if the applicable hi mass is that of the main body " , then we can lower the hi flux by a factor of 2 or so .
a smaller distance will also lower the hi mass ( although in the case of i zw 18 it could not be lowered by more than a factor of 23 , since otherwise it should be resolved more effectively into individual stars with hst ) .
independent data on dark - to - visible mass ratios are also somewhat uncertain due to the many assumptions involved .
while in the case of i zw 18 , radio observations support a dark - to - visible mass ratio of about 10 @xcite , the dynamical mass of sbs 0335@xmath41052 is only about 25 times that in gas and stars @xcite . in any event ,
i zw 18 and sbs 0335@xmath41052 , which are unlikely to be forming their first stars now @xcite yet show oxygen abundances close to the floor " of 12+log(o / h)@xmath617.2 allowed by the models of @xcite , could be interpreted as galaxies that have a maximum time - averaged mass - loss rate ( since their masses are a few times 10@xmath62m@xmath63 ) and are thus unable to retain metals effectively .
figure 4 of @xcite contains two galaxies , ugc 04483 and leo a , with apparently similar properties to i zw 18 in terms of oxygen abundance , hi mass , and dark - to - visible mass ratio .
ugc 04483 is also present in the bcd sample of it used here .
while @xcite show it in the forbidden " region , the smaller distance derived in the meantime from single - star photometry @xcite , has resulted in a downward revision of its hi mass .
this , combined with the 0.2 dex higher oxygen abundance of it , has moved ugc 04483 well into the dark - matter dominated regime in our figure [ zvsmhi ] .
@xcite present a case study of the sfh and the oxygen enrichment of leo a. it will be interesting to see whether their model also agrees with the broad parameters of i zw 18 ( and sbs 0335@xmath41052 ) .
one clear difference between leo a and i zw 18 ( or sbs 0335@xmath41052 ) , is that the present - day sfr of the former is much smaller , by about 3 ( 4 ) orders of magnitude , than the latter .
it is worth recalling the location of i zw 18 and sbs 0335@xmath41052 on figure [ zvssfr ] .
they appear to be too metal - poor for their sfr , or too intensely star - forming for their oxygen abundance , compared to the rest of the sample . explaining this offset as an error in the measured sfr is unlikely , requiring the actual sfr of i zw 18 ( for example ) to be more than 7 orders of magnitude less than estimated here .
the metallicity of i zw 18 has been the subject of many observations and extensive measurement , and explaining the offset as an error in the measured metallicity is similarly unlikely .
it will be interesting to see whether the models of @xcite can reproduce the high sfrs , high gas - mass fractions , and low chemical abundances of i zw 18 and sbs 0335@xmath41052 .
considering the optical and hi morphologies of i zw 18 and sbs 0335@xmath41052 , it is possible that these galaxies are involved in an interaction .
the models of @xcite do not apply to merger situations .
we investigated the sfrs of a small ( 50 objects ) but well - defined sample of bcds , that of @xcite .
one advantage of studying this sample is that ionized - gas abundances are readily available from the work of this team .
in addition , hi fluxes and profiles are in hand for many of the bcds from the work of thuan and collaborators .
this allows us to investigate correlations between sfr , ionized gas chemistry , hi mass , and total mass .
the main results are summarized at the end of section [ results ] . here , we draw some conclusions about some of the results that we think have a wider implication .
we find that sfr@xmath37 agrees with sfr@xmath36 in bcds .
this agreement exists over a wide range of sfrs ( nearly five orders of magnitude in the sample that we investigate ) , at low metallicity ( on average 0.1 of solar ) , and at small galaxy masses compared to those of normal spirals .
unless several factors have conspired to yield sfr@xmath37 = sfr@xmath36 in our sample , it is logical to conclude that the calibration relations and their underlying assumptions are valid for bcds .
furthermore , the existence of this type of scaling relation for a wide range of galaxies with different masses and sizes ( from spiral galaxies to late - type dwarfs ) , and with a wide range of interactions ( from merging to isolated ) , suggests that in terms of the fir / radio correlation , the local properties of individual star - forming regions are more important than the global properties of the galaxies that host them . while bcd studies have benefitted greatly from recent advances in far - infrared astronomy , and there is now no doubt that even very metal - poor galaxies living in the present epoch contain dust , very little is known as yet about their radio properties .
radio polarization observations are non - existent but are needed to elucidate the nature of magnetic fields in bcds
. magnetic field orientations , therefore are not well known .
because the fir / radio correlation , valid for large spiral galaxies , extends to bcds , the presence of magnetic fields at some level seems highly likely .
the role of magnetic fields is still poorly understood , however , and may contribute to the observed scatter in the fir / radio correlation .
color - magnitude diagrams of nearby bcds can serve as an external check on the validity of the sfrs derived from fir or radio luminosities .
the relevant data are difficult to obtain , since they require very high spatial resolution , and while the few existing data indicate encouraging agreement , additional cmd studies are required .
single - star photometry is , and in the near future will remain , biased toward nearby objects with small sfrs due to instrumental limitations . in paper
ii , we shall therefore investigate the sfrs of bcds in the it sample using the h@xmath7 luminosities as sfr indicators .
the small scatter in sfrs about the mean fir / radio ratio supports results of cmd models that show that bcds do not have highly episodic sfhs @xcite .
our results also indicate that the median sfrs of bcds are somewhat smaller than has been deemed typical in the past .
the gas consumption timescales are at a minimum a few hundred myr , and more typically they are of the order of a few gyrs . in this respect
, bcds appear closer to death than dwarf irregulars ( compare , for example , with the recent depletion time scales in * ? ? ?
we find that there is no mass - metallicity relation among the bcds in the it sample .
the absence of a mass - metallicity relation for bcds can be understood with the help of differential winds @xcite .
the more recent models of @xcite and @xcite , which incorporate dark matter and are applicable to the hi mass range that we observe in bcds , appear to predict well the global properties of bcds , at least to first order
. an interesting prediction of these models is that all galaxies with hi masses below 10@xmath60m@xmath3 are prone to lose mass through the feedback of star formation .
interestingly , while there is not a lot of cool gas lost , the loss of metals becomes more and more important the lower the hi mass of the galaxy ( with a peak in the metal - loss rate at an hi mass of a few times 10@xmath62m@xmath3 ) .
this indicates a progressively stronger violation of the assumption of a closed - box model for chemical evolution toward low galaxy masses .
we surmise that the absence of a mass - metallicity relation in bcds reflects just this metal - enriched mass - loss .
the basis for the mass - metallicity relation breaks down below a gas mass of about 10@xmath60m@xmath3 .
i zw 18 and sbs 0335@xmath41052 continue to stand out as extreme examples of bcds whenever their chemical properties are being investigated . in terms of total sfr , hi mass , gas - consumption timescale and w@xmath43 , they fit in well with other bcds .
but in terms of oxygen abundance versus sfr and hi mass , they are extremes compared to the rest of the sample .
i zw 18 and sbs 0335@xmath41052 have gas masses of a few times 10@xmath62m@xmath3 where dwarf galaxies , according to the models of @xcite , are particularly vulnerable to the loss of heavy elements .
a detailed prediction of the sfh and the chemical enrichment of i zw 18 based on the work of @xcite would be particularly interesting as a precursor to the results of future deep single - star photometry . chemo - dynamical models with gasping , " rather than bursting ,
sfhs for i zw 18 are already in work by @xcite .
( gasping " in this sense means long episodes of moderate star - forming activity separated by short quiescent intervals , see e.g. , @xcite @xcite . ) it will be interesting to see what physical insights can be gained from their simulations .
amh gratefully acknowledges support provided by nasa through hubble fellowship grant hst - hf-01140.01-a awarded by the space telescope science institute , which is operated by the association of universities for research in astronomy , inc .
, for nasa , under contract nas 5 - 26555 . becker , r. h. , white , r. l. , helfand , d. j. 1995 , apj , 450 , 559 cannon , j. m. , skillman , e. d. , garnett , d. r. , dufour , r. j. 2001 , aas , 198 # 09.14 carollo , c. m. , lilly , s. j. 2001 , apj , 584 , l153 condon , j. j. 1992 , ara&a , 30 , 575 condon , j. j. , cotton , w. d. , greisen , e. w. , yin , q. f. , perley , r. a. , taylor , g. b. , broderick , j. j. 1998 , aj , 115 , 1693 cram , l. , hopkins , a. , mobasher , b. , rowan - robinson , m. 1998 , apj , 507 , 155 crone , m. , schulte - ladbeck , r. e. , greggio , l. , hopp , u. 2002 , apj , 567 , 258 dale , d. a. , helou , g. , neugebauer , g. , soifer , b. t. , frayer , d. t. , condon , j. j. 2001 , aj , 122 , 1736 deeg , h .- j . , brinks , e. , duric , n. , klein , u. , skillman , e. 1993 , apj , 410 , 626 dekel , a. , silk , j. 1986 , apj , 303 , 39 de vaucouleurs , g. , de vaucouleurs , a. , corwin , j. r. , buta , r. j. , paturel , g. , fouque , p. 1991
, third reference catalogue of bright galaxies , " ( new york : springer - verlag ) dolphin , a. e. , makarova , l. , karachentsev , i. d. , karachentseva , v. e. , geisler , d. , grebel , e. k. , guhathakurta , p. , hodge , p. w. , sarajedini , a. , seitzer , p. 2001 , mnras , 324 , 249 fanelli , m. n. , oconnell , r. w. , thuan , t. x. 1988 , apj , 334 , 665 ferrara , a. , tolstoy , e. 2000 , mnras,313 , 291 freedman , w. l. , madore , b. f. , gibson , b. k. , ferrarese , l. , kelson , d. d. , sakai , s. , mould , j. r. , kennicutt , r. c. , jr . , ford , h. c. , graham , j. a.,huchra , j. p. , hughes , s. m. g. , illingworth , g. d. , macri , l. m. , stetson , p. b. 2001 , apj , 553 , 47 haarsma , d. b. , partridge , r. b. , windhorst , r. a. , richards , e. a. 2000 , apj , 544 , 641 hildago - gmez , a. m. , olofsson k. 1998 , a&a , 334 , 45 hirashita , h. , inoue , a. k. , kamaya , h. , shibai , h. 2001 , a&a , 366 , 83 hopkins a. m. , miller c. j. , connolly a. j. , genovese c. , nichol r. c. , wasserman l. 2002 , aj , 123 , 1086 huchtmeier , w.k . , richter , o.g .
1986 , a&as , 63 , 323 hunter , d. a. , hoffmman , l. 1999 , aj , 117 , 2789 izotova , i. , parnovskya , s. , izotov , y. 2000 , newar , 44 , 283 izotov , y. i. , thuan , t. x. , lipovetsky , v. a. 1994 , apj , 435 , 647 izotov , y. i. , thuan , t. x. , lipovetsky , v. a. 1997 , apjs , 108 , 1 izotov , y. i. , thuan , t. x. 1998 , apj , 500 , 188 izotov , y. i. , thuan , t. x. 1999 , apj , 511 , 639 ( it ) kennicutt , r. c. 1998 , ara&a , 36 , 189 klein , u. , weiland , h. , brinks , e. 1991 , a&a 246 , 323 kobulnicky , h. a. , zaritsky , d. 1999 , apj , 511 , 118 kraan - korteweg , r. c. 1986 , a&as , 66 , 255 kunth , d. , stlin , g. 2000 , a&ar , 10 , 1 larsen , t. i. , sommer - larsen , j. , pagel , b. e. j. 2001 , mnras , 323 , 555 lequeux , j. , rayo , j. f. , serrano , a. , peimbert , m. , torres - peimbert , s. 1979 , a&a , 80 , 155 lequeux , j. , viallefond , f. 1980 , a&a , 91 , 269 lisenfeld , u. , vlk , h.j .
, xu , c. 1997 , revmexaa , 6 , 81 lisenfeld , u. , ferrara , a. 1998 , apj , 496 , 145 mac low , m .-
m . , ferrara , a. 1999 , apj , 513 , 142 marconi , g. , matteucci , f. , tosi , m. 1994 , mnras , 270 , 35 mas - hesse , m. , kunth , d. 1999 , a&a , 349 , 765 perley , r. a. 2000 , the vla observational status summary , " ( http://www.aoc.nrao.edu/vla/obstatus/vlas/vlas.html ) popescu , c. c. , hopp , u. , rosa , m. r. 1999 , a&as , 350 , 414 pustilnik , s. a. , brinks , e. , thuan , t. x. , lipovetsky , v. a. , izotov , y. i. 2001 , aj , 121 , 1413 recchi , s. , matteucci , f. , dercole , a. , tosi , m. 2002 , a&a , in press ( astro - ph/0112450 ) sage , l. j. , salzer , j. j. , loose , h .- h .
, henkel , c. 1992 , a&a , 265 , 19 sargent , w. l. w. , searle , l. 1970 , apj , 162 , l155 schulte - ladbeck , r. e. , hopp , u. , crone , m. m. , greggio , l. 1999 , apj , 525 , 709 schulte - ladbeck , r. e. , hopp , u. , crone , m. m. , greggio , l. 2000 , in the first stars " , eds .
a. weiss , t. abel , v. hill , ( springer ) p. 293
schulte - ladbeck , r. e. , hopp , u. , greggio , l. , crone , m. m. , drozdovsky , i.o .
2001a , aj , 121 , 3007 schulte - ladbeck , r.e . , hopp , u. , greggio , l. , crone , m.m .
, drozdovsky , i.o . , hopkins , a.m. 2001b , in proc . of the 40@xmath64 meeting of the bonn - bochum graduierten kolleg dwarf galaxies and their environment " , eds .
k. s. de boer , r .- j .
dettmar , u. klein , shaker verlag , p. 83
skillman , e. d. , kennicutt , r. c. , hodge , p. w. 1989 , apj , 347 , 875 thuan , t. x. , martin , g. e. 1981 , apj , 247 , 823 thuan , t. x. , lipovetsky , v. a. , martin , j .-
m . , pustilnik , s. a. 1999a , a&as , 139 , 1 thuan , t. x. , sauvage , m. , madden , s. 1999b , apj , 516 , 783 tonry , j. l. , dressler , a. , blakeslee , j. p. , ajhar , e. a. , fletcher , a. b. , luppino , g. a. , metzger , m. r. , moore , c. b. 2001 , apj , 546 , 681 tosi , m. , greggio , l. , annibali , f. , aloisi , a. 2002 , astro - ph/0112268 van zee , l. 2001 , aj , 121 , 2003 van zee , l. , westpfahl , d. , haynes , m. p. 1998 , aj , 115 , 1000 weilbacher , p.m. , fritze-v.alvensleben , u. 2001 , a&al , 373 , 9 xu , c. 1990 , apj , 365 , l47 lccclr um 311 & 01 15 33.89 & @xmath65 51 31.9 & 2.50 & @xmath66 ( @xmath67 ) & 1709 + um 420 & 02 20 54.51 & @xmath68 33 23.6 & & @xmath69 ( @xmath70 ) & 17514 + mrk 0600 & 02 51 04.57 & @xmath71 27 13.9 & & @xmath72 & 1008 + sbs 0335@xmath41052 & 03 37 44.04 & @xmath73 02 38.5 & & @xmath74 & 4043 + mrk 1089 & 05 01 37.80 & @xmath75 15 27.7 & 4.06 & @xmath76 & 4068 + mrk 0005 & 06 42 15.53 & @xmath77 37 32.6 & 0.21 & @xmath78 & 792 + mrk 0071 & 07 28 41.40 & @xmath79 11 26.0 & 3.30 & @xmath80 & 108 + mrk 1409 & 07 45 29.72 & @xmath81 25 37.6 & & @xmath82 &
5576 + sbs 0749 + 568 & 07 53 41.53 & @xmath83 41 58.9 & & @xmath84 & 5471 + sbs 0749 + 582 & 07 53 50.93 & @xmath85 08 10.4 & & @xmath69 & 9548 + ugc 04483 & 08 37 03.00 & @xmath79 46 31.0 & & @xmath70 & 178 + sbs 0907 + 543 & 09 11 08.51 & @xmath86 10 51.2 & & @xmath87 ( @xmath88 ) & 8124 + mrk 1416 & 09 20 56.19 & @xmath89 34 04.7 & & @xmath90 ( @xmath91 ) &
2326 + sbs 0926 + 606 & 09 30 06.61 & @xmath92 26 51.8 & 0.27 & @xmath74 & 4122 + i zw 018 & 09 34 01.92 & @xmath93 14 26.1 & & @xmath74 & 742 + sbs 0940 + 544 & 09 44 16.72 & @xmath86 11 32.9 & & @xmath84 & 1638 + sbs 0943 + 561 & 09 46 46.59 & @xmath93 57 06.4 & & @xmath84 & 8850 + mrk 0022 & 09 49 30.45 & @xmath93 34 48.6 & & @xmath94 &
1551 + sbs 0948 + 532 & 09 51 31.81 & @xmath89 59 36.3 & & @xmath84 ( @xmath95 ) & 13880 + mrk 1434 & 10 34 10.15 & @xmath85 03 49.1 & & @xmath96 & 2269 + mrk 1271 & 10 56 09.00 & @xmath97 10 23.0 & & @xmath98 ( @xmath99 ) &
1013 + cg 0798 & 10 57 47.45 & @xmath100 15 39.3 & 0.53 & @xmath101 ( @xmath99 ) & 634 + mrk 0036 & 11 04 58.47 & @xmath102 08 22.1 & 0.23 & @xmath103 ( @xmath95 ) & 646 + mrk 0162 & 11 05 08.13 & @xmath104 44
50.1 & 1.29 & @xmath105 ( @xmath106 ) & 6458 + sbs 1116 + 583b & 11 19 24.79 & @xmath85 03 50.8 & & @xmath107 & 9905 + vii zw 403 & 11 27 59.90 & @xmath108 59 39.0 & 0.38 & @xmath109 & @xmath41100 + sbs 1128 + 573 & 11 31 16.33 & @xmath110 03 58.7 & & @xmath84 & 1787 + mrk 1450 & 11 38 35.62 & @xmath110 52 27.2 & 0.28 & @xmath111 & 946 + um 448 & 11 42 12.30 & @xmath68 20 03.0 & 4.14 & @xmath112 ( @xmath113 ) & 5488 + mrk 0750 & 11 50 02.67 & @xmath114 01 23.8 & 0.44 & @xmath115 ( @xmath99 ) & 754 + um 461 & 11 51 33.03 & @xmath116 22 22.8 & & @xmath72 & 1039 + um 462 & 11 52 37.15 & @xmath116 28 08.5 & 0.94 & @xmath117 & 1199 + mrk 0193 & 11 55 28.35 & @xmath110 39 52.4 & & @xmath118 & 5157 + sbs 1159 + 545 & 12 02 02.37 & @xmath86 15 49.6 & & @xmath84 ( @xmath99 ) & 3537 + sbs 1205 + 557 & 12 08 28.18 & @xmath93 25 26.5 & & @xmath87 & 1740 + sbs 1211 + 540 & 12 14 02.50 & @xmath81 45 18.0 & & @xmath119 & 907 + sbs 1222 + 614 & 12 25 05.55 & @xmath120 09 08.5 & 0.29 & @xmath121 & 706 + mrk 0209 & 12 26 16.02 & @xmath122 29 36.6 & & @xmath123 ( @xmath123 ) & 281 + sbs 1249 + 493 & 12 51 52.37 & @xmath124 03 25.5 & & @xmath94 ( @xmath95 ) & 7296 + mrk 0059 & 12 59 00.35 & @xmath125 50 42.8 & 1.97 & @xmath126 ( @xmath127 ) &
1089 + sbs 1319 + 579 & 13 21 22.47 & @xmath110 41 28.1 & & @xmath119 & 2060 + sbs 1331 + 493 & 13 33 22.93 & @xmath124 06 06.3 & & @xmath69 ( @xmath95 ) & 599 + mrk 1486 & 13 59 51.01 & @xmath110 26 24.1 & & @xmath128 & 10136 + cg 0389 & 14 17 01.76 & @xmath129 30 13.4 & & @xmath115 ( @xmath99 ) & 609 + cg 0413 & 14 22 38.84 & @xmath86 14 05.9 & & @xmath84 ( @xmath99 ) &
6176 + mrk 0475 & 14 39 05.43 & @xmath100 48 20.5 & & @xmath118 ( @xmath99 ) & 583 + cg 1258 & 14 44 00.15 & @xmath102 15 55.1 & & @xmath94 ( @xmath99 ) &
13605 + sbs 1533 + 574 & 15 34 13.31 & @xmath110 17 06.8 & 0.26 & @xmath130 & 3310 + mrk 0487 & 15 37 04.19 & @xmath93 15 48.3 & 0.30 & @xmath118 & 665 + mrk 0930 & 23 31 58.29 & @xmath131 56 49.9 & 1.25 & @xmath132 & 5485 + lcclcc um 311 & 21.3 & 1.5 & 1.1 ( 0.14 ) & @xmath133 & + um 420 & 247.9 & & @xmath1521.2 ( 9.23 ) & @xmath134 & + mrk 0600 & 14.5 & & @xmath150.077 & @xmath135 & @xmath136 + sbs 0335@xmath41052 & 58.3 & & 1.3 & @xmath137 & @xmath138 + mrk 1089 & 59.8 & 18.8 & 16.2 & @xmath139 & + mrk 0005 & 15.5 & 0.065 & @xmath150.095 & @xmath140 & @xmath141 + mrk 0071 & 3.9 & 0.065 & 0.041 & @xmath142 & @xmath143 + mrk 1409 & 83.0 & & 2.5 & @xmath144 & @xmath145 + sbs 0749 + 568 & 81.6 & & @xmath152.2 & @xmath146 & @xmath147 + sbs 0749 + 582 & 139.6 & & @xmath156.8 & @xmath148 & + ugc 04483 & 3.2 & & 0.0016 & @xmath149 & @xmath150 + sbs 0907 + 543 & 119.9 & & @xmath155.1 ( @xmath151.6 ) & @xmath151 & + mrk 1416 & 38.1 & & 0.27 ( @xmath150.15 ) & @xmath152 & @xmath153 + sbs 0926 + 606 & 63.4 & 1.4 & 1.6 & @xmath154 & + i zw 018 & 15.0 & & 0.087 & @xmath155 & @xmath156 + sbs 0940 + 544 & 28.7 & & @xmath150.27 & @xmath157 & + sbs 0943 + 561 & 130.4 & & @xmath155.6 & @xmath158 & + mrk 0022 & 27.5 & & @xmath150.24 & @xmath159 & + sbs 0948 + 532 & 202.2 & & @xmath1513.4 ( @xmath155.4 ) & @xmath159 & + mrk 1434 & 38.1 & & @xmath150.84 & @xmath160 & + mrk 1271 & 11.9 & & @xmath150.064 ( @xmath150.021 ) & @xmath161 & + cg 0798 & 9.4 & 0.061 &
0.092 ( @xmath150.012 ) & @xmath162 & + mrk 0036 & 7.7 & 0.017 & 0.020 ( @xmath150.0079 ) & @xmath163 & @xmath164 + mrk 0162 & 96.6 & 15.5 & 21.4 ( 15.5 ) & @xmath165 & + sbs 1116 + 583b & 146.4 & & @xmath156.4 & @xmath166 & + vii zw 403 & 4.4 & 0.0096 & 0.0040 & @xmath167 & @xmath168 + sbs 1128 + 573 & 32.1 & & @xmath150.34 & @xmath169 & + mrk 1450 & 19.7 & 0.14 & @xmath150.11 & @xmath170 & + um 448 & 78.5 & 32.9 & 29.5 ( 22.0 ) & @xmath171 & + mrk 0750 & 6.0 & 0.020 & 0.022 ( @xmath150.0049 ) & @xmath172 & @xmath173 + um 461 & 10.2 & & @xmath150.039 & @xmath174 & + um 462 & 17.2 & 0.36 & 0.13 & @xmath175 & + mrk 0193 & 79.4 & & @xmath152.4 & @xmath176 & @xmath177 + sbs 1159 + 545 & 56.8 & & @xmath151.1 ( @xmath150.47 ) & @xmath178 & + sbs
1205 + 557 & 32.0 & & @xmath150.36 & @xmath169 & @xmath179 + sbs 1211 + 540 & 20.1 & & @xmath150.17 & @xmath180 & @xmath181 + sbs
1222 + 614 & 16.5 & 0.10 & 0.14 & @xmath175 & + mrk 0209 & 5.8 & & 0.022 ( 0.022 ) & @xmath182 & @xmath183 + sbs 1249 + 493 & 110.7 & & @xmath153.9 ( @xmath151.7 ) & @xmath184 & + mrk 0059 & 25.2 & 1.6 & 1.3 ( 0.55 ) & @xmath161 & + sbs 1319 + 579 & 36.7 & & @xmath150.56 & @xmath185 & + sbs 1331 + 493 & 15.1 & & @xmath150.078 ( @xmath150.030 ) & @xmath186 & @xmath187 + mrk 1486 & 150.6 & & 6.2 & @xmath188 & + cg 0389 & 15.3 & & 0.14 ( @xmath150.032 ) & @xmath189 & @xmath190 + cg 0413 & 94.5 & & @xmath153.0 ( @xmath151.2 ) & @xmath191 & + mrk 0475 & 13.6 & & @xmath150.072
( @xmath150.027 ) & @xmath192 & @xmath193 + cg 1258 & 199.4 & & @xmath1512.5 ( @xmath155.6 ) & @xmath194 & + sbs 1533 + 574 & 53.2 & 0.94 & 1.7 & @xmath195 & + mrk 0487 & 15.3 & 0.092 & @xmath150.092 & @xmath196 & @xmath197 + mrk 0930 & 75.3 & 9.1 & 10.6 & @xmath198 & @xmath199 + | we determine and examine the star - formation rates of 50 well known , local blue compact dwarf ( bcd ) galaxies based on their 1.4ghz and 60@xmath0 m luminosities .
we find that in cases for which both radio and far - infrared luminosities are available , the resulting star - formation rates agree extremely well with one another .
we determine that the star - formation rates of the bcd galaxies in our sample span nearly five orders of magnitude , from approximately a few times 10@xmath1 to several times 10@xmath2m@xmath3yr@xmath4 , with a median sfr of about 0.3@xmath5m@xmath3yr@xmath4 .
we discuss trends of metallicity ( primarily oxygen abundance ) with star - formation rate , and explore the connections between sfr and galaxy mass estimates . |
one of the remarkable properties of the @xmath2 standard model is the cancellation of the triangle anomalies , which is implied by the relation @xmath3 between the electromagnetic and the left and righthanded weak charges of the leptons and quarks for each fermion family .
this relation is of fundamental importance , because it renders the standard model renormalizable .
although not being enforced by every possible scenario , it might be a dynamical consequence of an as yet unknow underlying theory . in many extensions of the standard model new bosons , the leptoquarks @xmath4 , which carry both lepton and baryon number , are predicted .
this applies both to grand unified theories @xcite as well as models based on compositeness @xcite or technicolor @xcite .
whereas in grand unified theories leptoquarks emerge as gauge bosons with masses in the range of @xmath5 and couplings which do not conserve baryon ( b ) and lepton ( l ) number , other approaches contain also leptoquarks with b- and l - conserving couplings .
these particles may have masses in the range which is accessible at the present high energy colliders .
recently , the h1 experiment @xcite at hera observed 12 events in the range of large bjorken @xmath0 for @xmath6 expecting 4.7 events .
the seven excess events of h1 cluster around @xmath7 .
the zeus experiment @xcite found 4 events for @xmath8 and @xmath9 , where 0.9 events are expected .
still the statistics is low and further measurements are needed before firm conclusions can be drawn .
if the observed excess is confirmed at larger statistics , one of the possible explanations could be single leptoquark production at a mass @xmath10 , cf .
@xcite . in this note
we summarize the current experimental limits on the couplings and masses for scalar and vector leptoquarks and investigate to which extent this interpretation is compatible with experimental results obtained at other high energy colliders .
direct searches for leptoquarks were performed in @xmath11 annihilation- @xcite , @xmath12- @xcite , and @xmath13 scattering experiments @xcite .
the currently most stringent mass limits , which are independent of the fermionic couplings , were derived by the tevatron experiments ( see @xcite ) searching for leptoquark pair production @xcite .
the following mass ranges are excluded : @xmath14 in deriving the above bounds it was assumed that the branching fractions into the charged lepton - quark and neutrino - quark final states are equal as a working hypothesis .
rigorous theoretical predictions on the branching fractions do not exist . for leptoquarks with a purely righthanded fermion coupling , however , the decay @xmath15 is not possible on tree level , leading to @xmath16 .
the h1 experiment @xcite observed a high mass excess for @xmath17 and @xmath18 final states .
if both effects are interpreted in terms of a leptoquark signal the branching fraction is @xmath19 .
a preliminary result was reported by the d0 collaboration from the analysis of run ib @xcite yielding @xmath20 the above bound on 3rd generation vector leptoquarks was derived assuming the minimal vector coupling to the gluon , @xmath21 , see section 4 . for the case
that the coupling of the vector leptoquark pair to the gluon is of the yang mills type , @xmath22 , the mass range @xmath23 is excluded . note that the bound of @xmath24 becomes weaker considering besides the anomalous coupling @xmath25 a second anomalous coupling @xmath26 , as shown in ref .
mass limits from the tevatron data for vector leptoquarks associated to the first and second fermion generation were not presented yet .
bounds on the leptoquark - fermion couplings from low energy reactions such as meson decays , meson - antimeson mixing , lepton decays and others were studied in refs .
@xcite@xcite in detail . here
the classification of leptoquarks as introduced in ref .
@xcite and @xmath27 , respectively , where @xmath28 is the weak isospin . ] was widely used .-
parity violation , cf .
@xcite , to fermion fields are described by a similar lagrangian , containing only a few terms . ] for the leptoquarks associated to the first family the bounds are @xcite @xmath29 the upper bound on the coupling for the other leptoquarks lies between these values .
here @xmath30 denotes either the lefthanded ( @xmath31 ) or the righthanded ( @xmath32 ) fermion coupling , and @xmath33 is the electric charge . in some cases
even either @xmath34 or @xmath35 has to be obeyed for @xmath36 , cf .
similar results were obtained in refs .
@xcite with a lowest bound of @xmath37 although these bounds are partly affected by low energy hadronic matrix elements , the above numbers set a scale .
the cross section for single leptoquark production in @xmath13 scattering @xmath38 was calculated in refs . @xcite .
because the leptoquark widths @xmath39 are small compared to the masses @xmath40 one may work in the narrow width approximation .
the integrated leptoquark production cross section for @xmath41 scattering reads @xcite @xmath42 here @xmath43 denotes the fine structure constant , @xmath44 the quark ( antiquark ) density , and @xmath45 is the branching ratio in the production channel .
@xmath46 accounts for the leptoquark spin with @xmath47 and @xmath48 . for @xmath49
the production cross sections are estimated to be @xmath50 using the parton densities @xcite for @xmath51 and @xmath52 .
the integrated luminosities of the h1 and zeus experiment are @xmath53 and @xmath54 , respectively .
if the 7 excess events of the h1 experiment @xcite are interpreted in terms of a leptoquark signal , the respective fermionic couplings are found to be @xmath55 for leptoquark states produced in @xmath56 or @xmath57 fusion , respectively . here , eq .
( [ estim ] ) was used for an estimate , accounting for an acceptance of @xmath58 @xcite .
the values for the fermionic couplings @xmath59 for vector leptoquarks are given by @xmath60 from the above numbers .
the corresponding numbers obtained for the 3 excess events of the zeus experiment , at an acceptance of @xmath58 @xcite , are @xmath61 the leptoquark couplings are found to lie in the allowed range @xmath62 and @xmath63 , respectively , for the up and down quarks , while for the antiquarks they are larger and widely excluded by the bounds given in eqs .
( [ excl1],[excl2 ] ) .
since for scalar leptoquarks the @xmath64-distribution is flat , while for vector leptoquarks it behaves @xmath65 , cf .
@xcite , this distribution can in principle be used to get an information on the spin of the produced state .
although at the current statistics one can not get a decisive answer , it is instructive to look on the @xmath64 averages . as an example , one obtains for the 7 events of h1 @xmath66 in the range @xmath67 .
the average @xmath64-values for a scalar or vector leptoquark in this range are @xmath68 and @xmath69 , respectively , while for the standard deep inelastic @xmath70 cross section @xmath71 is obtained for comparison .
this shows , how difficult a distinction between the different hypotheses might be , since the average value for the case of vector leptoquarks is nearly the same as for the deep inelastic sample in this range of @xmath64 . at a larger statistics a @xmath72-analysis may yield a definite answer .
aside of the fusion process @xmath73 , leptoquarks should also be produced via the processes @xmath74 and @xmath75 .
since the gluon density is rapidly falling in the range of larger values of @xmath0 , the latter reaction yields only a small contribution to the cross section for large leptoquark masses .
both processes contain a collinear singularity in the limit @xmath76 , which has to be regulated for the inclusive cross section .
for the differential distributions a lower cut on @xmath77 is applied .
a part of these events would show an apparent two - jet signature , with one jet at a mass of @xmath40 .
since the leptoquark and the final state parton are balanced in @xmath77 , the electron is more difficult to isolate from the hadronic fragments of the leptoquark both kinematically , and possibly also due to @xmath78 mesons contained in the jet . for scalar leptoquark production
both processes were studied in ref .
@xcite referring to the parton parametrization of ref . @xcite . for @xmath79 and @xmath80
the integrated cross section for @xmath81 amounts to @xmath82 , i.e. @xmath83 events , for @xmath84 and @xmath85 .
the single production of leptoquarks at proton colliders proceeds either through the reaction @xmath86
@xcite or through charged lepton quark ( antiquark ) fusion @xcite . for the latter process
the electron in the initial state is provided by @xmath11 pair production of a photon which is collinearly radiated off a quark of the second proton ( antiproton ) .
the cross sections for single scalar and vector - leptoquark production in @xmath87-fusion are @xcite @xmath88 for a mass @xmath36 and @xmath89 . for the second process , the production cross section for leptoquarks at a mass @xmath36 is given by @xmath90 for values of @xmath89 the cross section amounts to @xmath91 .
these cross sections are too small to be measured at tevatron currently .
complementary to single leptoquark production and @xmath92 scattering at lep2 and at tevatron , respectively . since @xmath93 ,
the dominant contributions come from the interference terms between the respective @xmath94-channel contributions and leptoquark exchange in the @xmath95 or @xmath96 channel , cf .
@xcite . at the current luminosities
it is not possible @xcite to constrain the fermionic couplings further by these reactions in the mass range @xmath36 . ]
, which relies on the fermionic couplings of leptoquarks , one may consider the pair production annihilation , @xmath97 , @xmath98 and @xmath99 scattering were calculated in refs .
@xcite . for
a survey on the search potential , see ref .
@xcite . ] of scalar and vector leptoquarks at hadron colliders .
even if the fermionic couplings of leptoquarks are small , @xmath100 , leptoquarks may be produced through quark antiquark and gluon gluon fusion since the couplings to the gluon are well determined .
the pair production cross sections were calculated in refs .
@xcite for scalar and in refs .
@xcite for vector leptoquarks by different methods .
while in refs .
@xcite only the case of vector leptoquarks carrying a yang mills type coupling and in refs .
@xcite as well the case of the minimal vector coupling to the gluon were considered , the general case of anomalous couplings , @xmath25 and @xmath26 , of vector leptoquarks to the gluon was delt with in ref .
@xcite . in figure 1
the total pair production cross section is shown for scalar leptoquarks using the cteq3 ( lo ) parametrization @xcite for the parton densities .
for @xmath101 the scalar pair production cross section amounts to @xmath102 , taking @xmath103 both as the factorization and renormalization scale .
other choices as @xmath104 lead to a cross section which is larger by a factor of @xmath105 for @xmath36 , showing the scale dependence in lowest order .
the pair production cross section for vector leptoquarks depends sizeably on the anomalous couplings @xmath25 and @xmath26 of the vector leptoquarks to the gluon , see figure 1 .
if the coupling would be of the yang mills type , i.e. @xmath106 , the event rate could reach about 1000 events for @xmath36 at each tevatron experiment . even in the case of the minimal vector coupling , @xmath107 and @xmath108 ,
the rate amounts to 100 events .
both these cases are likely to be excluded due to the event rates measured at the tevatron experiments .
however , since a prediction on the vector couplings @xmath25 and @xmath26 does not exist , a global limit can only be derived minimizing the integrated cross section for these parameters . for a mass of @xmath36 one obtains @xmath109 for events with a @xmath110 final state assuming @xmath111 . for @xmath112 @xmath113
is obtained .
the corresponding anomalous couplings are @xmath114 and @xmath115 .
these leading order estimates show that the pair production process of leptoquarks may yield up to 10 .. 16 signal events at an integrated luminosity of @xmath116 for scalar leptoquarks and , in the presence of only the two anomalous couplings @xmath25 and @xmath26 , for vector leptoquarks at the minimal cross section , up to 20 .. 30 events , depending on the value of the factorization and renormalization scale k - factor both for scalar and vector leptoquark pair production is not yet available . ] and the yet unknown branching ratios .
the detection of the @xmath110 final state is complicated due to different background reactions , see e.g. @xcite , which can only be reliably studied by the experiments themselves . given the uncertainty in the branching fraction @xmath117 , both scalar and vector leptoquarks seem to be not yet excluded by the tevatron data .
on the basis of the present statistics bounds on the anomalous couplings @xmath25 and @xmath26 for vector leptoquarks can be derived up to masses of @xmath118 .
an interpretation of the recently reported excess of events in the range of high @xmath0 and @xmath1 at hera in terms of single leptoquark production yields leptoquark - fermion couplings , @xmath119 , which are compatible with other limits , particularly those obtained from low energy reactions , for @xmath120 and @xmath121 .
still more statistics is needed to clearly establish an effect . if the observed excess of events is due to leptoquarks , these states should as well be detected in the reaction channel @xmath122 at hera at an increased integrated luminosity .
the measurement of this reaction allows to investigate the coupling of the state @xmath123 to gluons and provides an independent possibility to determine its spin
. the fermilab experiments , at a higher statistics , will be able to either clearly confirm or exclude the interpretation of the current excess of events found at hera as being due to leptoquark production . for vector leptoquarks already at the current statistics bounds on the range of the couplings @xmath25 and @xmath26 can be derived .
it is likely that both yang - mills type and minimal couplings of the leptoquarks to the gluon are already exluded by the observed event rates .
on the other hand the tevatron data do not exclude scalar and vector leptoquarks , for some range in @xmath25 and @xmath26 , at a mass @xmath36 .
the search for leptoquark pair production at tevatron is complementary to the search for single leptoquark production at hera both due to the couplings involved in the production mechanism , as well as with respect to the alternative possibility to investigate the spin of the produced state in case of a confirmation of the effect
. * acknowledgement . *
it is a pleasure for me to thank c. berger , e. boos , w. buchmller , r. eichler , m. klein , a. kryukov , e. reya , r. rckl , g. wolf and p.m. zerwas for discussions on the present topic during the last years .
i would like to thank f. jegerlehner , m. klein , p. sding and a. wagner for reading the manuscript .
999 h. georgi and s.l .
glashow , phys rev .
lett.*32 * ( 1974 ) 275 ; + h. georgi , in : _ particles and fields _ , a.i.p .
( 1975 ) , ed .
carlson ; + h. fritzsch and p. minkowski , ann .
( ny ) * 93 * ( 1975 ) 193 ; + for reviews see : p. langacker , phys .
rep . * 72 * ( 1981 ) 185 ; g.g .
ross , _ grand unified theories _
( benjamin , reading , ma , 1984 ) .
pati and a. salam , phys . rev . * d10 * ( 1974 ) 275 ; + j.l .
hewett and t.g .
rizzo , phys .
* 183 * ( 1989 ) 193 ; + h. murayama and t. yanagida , mod .
* a7 * ( 1992 ) 147 ; + p.h .
frampton , mod .
* a7 * ( 1992 ) 559 . for
a survey see : l. lyons , progr .
* 10 * ( 1982 ) 227 ; + h. terazawa , phys . rev .
* d22 * ( 1980 ) 184 , and earlier references quoted therein ; + l.f .
abbott and e. farhi , phys . lett .
* b101 * ( 1981 ) 69 ; nucl . phys .
* b189 * ( 1981 ) 547 ; + w. bardeen and v. visnjic , nucl .
* b194 * ( 1982 ) 151 ; + w. buchmller , r. peccei and t. yanagida , phys . lett . *
b124 * ( 1983 ) 67 ; + r. barbieri , a. masiero , and g. veneziano , phys . lett . *
b128 * ( 1983 ) 179 ; + b. schrempp and f. schrempp , desy 84055 ( 1984 ) ; phys . lett . *
b153 * ( 1985 ) 101 ; + p.c .
pati , phys .
* b228 * ( 1989 ) 228 .
e. farhi and l. susskind , phys . rep . * 74 * ( 1981 ) 277 and references therein .
c. adloff et al . ,
h1 collaboration , desy 97024 , hep - ex/9702012 . j. breitweg et al . , zeus collaboration , desy 97025 , hep - ex/9702015 .
behrend et al . ,
cello collaboration , phys .
* b178 * ( 1986 ) 452 ; a : * 184 * ( 1987 ) 417 ; + w. bartel et al . , jade collaboration , z. phys .
* c36 * ( 1987 ) 15 ; + g.n .
kim et al . ,
amy collaboration , phys .
* b240 * ( 1990 ) 243 ; + d. alexander et al . ,
opal collaboration , phys .
b275 * ( 1992 ) 123 ; phys .
b263 * ( 1991 ) 123 ; + d. decamp et al .
, aleph collaboration , phys .
* 216 * ( 1992 ) 253 ; + b. adeva et al .
, l3 collaboration , phys . lett .
* b261 * ( 1991 ) 169 ; o. adriani , et al . ,
l3 collaboration , phys .
* 236 * ( 1993 ) 1 ; + p. abreu et al .
, delphi collaboration , phys .
lett . * b275 * ( 1992 ) 222 ; * b316 * ( 1993 ) 620 ; + j.k .
mizukoshi , o.j.p .
eboli , and m.c .
garcia , nucl .
b443 * ( 1995 ) 20 .
j. alitti et al . , ua2 collaboration , phys .
* b274 * ( 1992 ) 507 .
f. abe et al . , cdf collaboration , phys . rev . *
d48 * ( 1993 ) r3939 ; phys .
* 75 * ( 1995 ) 1012 ; + s. abachi et al .
, d0 collaboration , phys .
* 72 * ( 1994 ) 965 ; * 75 * ( 1995 ) 3618 .
f. abe et al . , cdf collaboration , fermilab pub96450e , subm .
m. derrick et al .
, zeus collaboration , phys .
* b306 * ( 1993 ) 173 ; desy 94204f ; + i. abt et al . ,
h1 collaboration , nucl .
b396 * ( 1993 ) 3 ; + t. ahmed et al . , h1 collaboration , z. phys . *
c64 * ( 1994 ) 545 ; + s. aid et al . ,
h1 collaboration , phys .
b353 * ( 1995 ) 578 ; * b369 * ( 1996 ) 173 ; + m. derrick et al . ,
zeus collaboration , desy 96161 .
k. w. merritt , contr .
to the dpf meeting , aug .
1996 , hep - ex/9701009 .
j.a . grifols and a. mndez , phys . rev . *
d26 * ( 1982 ) 324 ; + i. antoniadis , l. baulieu , and f. delduc , z. phys .
* c23 * ( 1984 ) 119 ; + e. eichten , i. hinchliffe , k.d .
lane , and c. quigg , rev .
phys . * 56 * ( 1984 ) 579 ; + g. altarelli and r. rckl , phys . lett .
* b144 * ( 1984 ) 126 ; + r.n .
mohapatra , g. segre , and l. wolfenstein , phys . lett .
* b145 * ( 1984 ) 433 ; + s. dawson , e. eichten , and c. quigg , phys . rev .
* d31 * ( 1985 ) 1581 .
p. arnold and c. wendt , phys . rev .
* d33 * ( 1986 ) 1873 ; + g.v .
borisov , y.f .
pirogov , and k.r .
rudakov , z. phys .
* c36 * ( 1987 ) 217 .
hewett , t.g .
rizzo , s. pakvasa , h.e .
haber , and a. pomarol , in : proc . of the workshop ` physics at current accelerators and supercolliders ' , argonne , june 1993 , eds .
hewett , a.r .
white , and d. zeppenfeld , ( anl , argonne , 1993 ) , p. 539 ; + t.g .
rizzo , slac
pub967284 , hep - ph/9609267 .
j. blmlein , e. boos , and a. kryukov , desy 96174 , hep - ph/9610408 , z. phys .
* c * in print .
d0 home page : http://d0sig0.fnal.gov/public/d0_physics.html#hot .
, march 19 , 1997 .
w. buchmller and d. wyler , phys . lett . *
b177 * ( 1986 ) 377 . s. davidson , d. bailey , and b.a .
campbell , z. phys . *
c61 * ( 1994 ) 613 .
m. leurer , phys .
* 71 * ( 1993 ) 1324 ; phys . rev .
* d49 * ( 1994 ) 333 ; phys .
* d50 * ( 1994 ) 536 .
w. buchmller , r. rckl , and d. wyler , phys . lett . *
b191 * ( 1987 ) 442 .
hewett , in : proc . of the 1990 summer study on high energy physics - research directions for the decade , snowmass , co , july 1990 , ed .
e. berger , ( world scientific , singapore , 1992 ) p. 566 ; + d. choudhury and s. raychaudhuri , cern
th/9726 , hep - ph/9702392 .
j. wudka , phys . lett . *
b167 * ( 1986 ) 337 .
a. dobado , m.j .
herrero , and c. muoz , phys . lett .
* b191 * ( 1987 ) 449 .
lai , j. botts , j. huston , j.g .
morfin , j.f .
owens , j.w .
qiu , w.k .
tung , and h. weerts , phys . rev .
* d51 * ( 1995 ) 4763 .
e. eichten , i. hinchliffe , k.d .
lane , and c. quigg , rev .
* 56 * ( 1984 ) 579 .
j.l . hewett and s. pakvasa ,
b227 * ( 1989 ) 178 ; + t.m . aliev and a.a .
bayramov , sov .
* 52 * ( 1990 ) 689 ; + t.g .
rizzo , phys .
* d44 * ( 1991 ) 186 ; + t.m . aliev and kh .
a. mustafaev , yad . fiz .
* 53 * ( 1991 ) 771 ; + j.e .
cieza montalvo and o.j.p .
eboli , phys .
* d47 * ( 1993 ) 837 ; + h. nadeau and d. london , phys . rev .
* d47 * ( 1993 ) 3742 ; + g. belanger , d. london and h. nadeau , phys . rev .
* d49 * ( 1994 ) 3140 ; + s. atag , a. celikel , and s. sultansoy , phys . lett .
* b326 * ( 1994 ) 185 ; + o. cakir and s. atag , j. phys .
* g21 * ( 1995 ) 1189 ; + v. ilin , a. pukhov , v. savrin , a. semenov , and w. von schlippe , inp - msu-95 - 22 - 386 ; hep - ph/9506334 ; phys . lett .
* b351 * ( 1995 ) 504 ; + t.m .
aliev , d.a .
demir , e. iltan , and n.k .
pak , metu - phys - hep-95 - 14 , hep - ph/9511389 ; + f. cuypers , mpi - pht-95 - 129 ( 1996 ) , hep - ph/9602355 ; nucl .
b474 * ( 1996 ) 57 ; + t.m .
aliev , e. iltan , and n.k .
pak , phys . rev . *
d54 * ( 1996 ) 4263 ; + m.a .
doncheski and s. godfrey , ocip - c-96 - 1 ; hep - ph/9608368 ; phys . rev .
* d49 * ( 1994 ) 6220 ; phys . rev .
* d51 * ( 1995 ) 1040 .
a. dobado , m.j .
herrero , and c. muoz , phys . lett . * b207 * ( 1988 ) 97 ; + j.l .
hewett and s. pakvasa , phys .
* d37 * ( 1988 ) 3165 ; + j.e .
cieza montalvo and o.j.p .
eboli , phys . rev . *
d50 * ( 1994 ) 331 .
j. ohnemus , s. rudaz , t.f .
walsh , and p.m. zerwas , phys
b334 * ( 1994 ) 203 ; + e. reya , private communication .
a. djouadi , j. ng , and t.g .
rizzo , slac
j.l . hewett and t.g .
rizzo , phys .
* d36 * ( 1987 ) 3367 ; + r. rckl and p. zerwas , in : proc . of the workshop on physics at at future accelerators , la thuile , italy , january 1987 , cern 8707 , * vol . 2 * p. 223 ; + h. dreiner , j. ellis , d.v .
nanopoulos , n.d .
tracas , and n.d .
vlachos , mod .
* 3a * ( 1988 ) 443 ; + j.l .
hewett and t.g .
rizzo , in : proc . of the 1990 summer study on high energy physics - research directions for the decade , snowmass , co , july 1990 , ed .
e. berger , ( world scientific , singapore , 1992 ) p. 562 ; + m. doncheski and j.l .
hewett , z. phys .
* c56 * ( 1992 ) 209 .
j.l . hewett and t.g .
rizzo , in ref .
@xcite ; + n.d .
tracas and s.d.p .
vlassopulos , phys .
b220 * ( 1989 ) 285 ; + d. schaile and p.m. zerwas , in : proc .
workshop on physics at future accelerators ( la thuile and geneva , 1987 ) , ed .
mulvey , cern 8707 , * vol . 2 * , p. 251 ; + a. djouadi , m. spira , and p.m. zerwas , in : proc .
workshop on @xmath11 collisions at @xmath124 : the physics potential ( munich annecy hamburg , 1991 ) , ed .
p.m. zerwas , desy 92123b , p. 561
; + j. blmlein and r. rckl , in : proc .
workshop on @xmath11 collisions at @xmath124 : the physics potential ( munich annecy hamburg , 1991 ) , ed .
p.m. zerwas , desy 92123b , p. 595 ; + j. cieza montalvo and o.j.p .
eboli , in ref .
@xcite ; + j. blmlein and r. rckl , phys . lett .
* b304 * ( 1993 ) 337 ; + j. blmlein , desy 93132 , proc . of the 2nd workshop physics and experiments with linear @xmath11 colliders , waikoloa , hi ,
april 2630 , 1993 , eds .
harris , s. olsen , s. pakvasa , and x. tata , ( world scientific , singapore , 1993 ) vol .
* ii * , p. 524 ; + desy 93153 , proc . of the workshop , @xmath11 collisions at 500 gev , munich annecy hamburg , 1993 ,
p. zerwas , ( desy , hamburg , 1993 ) , p. 419 ; + j. blmlein and p.h .
frampton , proc . of the 2nd workshop physics and experiments with linear @xmath11 colliders , waikoloa , hi ,
april 2630 , 1993 , eds .
harris , s. olsen , s. pakvasa , and x. tata , ( world scientific , singapore , 1993 ) vol . * ii * , p. 926 ; + j. blmlein , e. boos , and a. kryukov , phys .
b292 * ( 1997 ) 150 ; + j. blmlein , e. boos , and a. pukhov , mod .
* a9 * ( 1994 ) 3007 ; + j. blmlein , e. boos , and a. kryukov , ref .
j. blmlein and e. boos , nucl . phys .
* b * ( proc . suppl . ) , eds .
t. riemann and j. blmlein , * 37b * ( 1994 ) p. 181 , and references therein .
dong - pei zhu , phys . rev .
* d22 * ( 1980 ) 2266 ; + p.r .
harrison and c.h .
llewellyn smith , nucl . phys .
* b213 * ( 1983 ) 223 ; e : * b223 * ( 1983 ) 524 ; + g.v .
borisov , y.f .
pirogov , and k.r .
rudakov , z. phys . *
c36 * ( 1987 ) 217 ; + m. de montigny and l. marleau , can . j. phys . * 68 * ( 1990 ) 612 .
b. diou , l. marleau , and g. simon , laval - phy-96 - 16 , hep - ph/9610397 ; + b. diou , l. marleau , g. simon , and m. de montigny , laval - phy-96 - 17 , hep - ph/9701285 . g. altarelli , j. ellis , g.f .
giudice , s. lola , and m.l .
mangano , cern - th/9740 , hep - ph/9703276 ; + h. dreiner and p. morawietz , hep - ph/9703279 ; + m.a .
doncheski and s. godfrey , ocip
c9702 , hep - ph/9703285 ; + j. kalinowski , r. rckl , h. spiesberger , and p.m. zerwas , desy 97037 , hep - ph/9703288 ; + k.s .
babu , c. kolda , j. march russel , and f. wilczek , iassns
hep 9704 , hep - ph/9703299 ; + v. barger , k. cheung , k. hagiwara , and d. zeppenfeld , madph97991 ; hep - ph/9703311 ; + m. suzuki , lbnl401211 , hep - ph/9703316 . figure 1 : integrated cross sections for scalar vector leptoquark pair production at the tevatron , @xmath125 choosing the renormalization and factorization scale by @xmath111 .
dash - dotted line : scalar leptoquarks ; fill lines : ym : yang - mills type coupling @xmath126 ; mc : minmal vector coupling @xmath127 ; dashed lines : other choices for the anomalous couplings : mm : @xmath128 , m1 : @xmath129 .
the asterisk denotes the minimum of the pairproduction cross section for vector leptoquarks with respect to the anomalous couplings at @xmath130 . | a summary is given on the experimental bounds for the couplings and masses of scalar and vector leptoquarks associated to the first fermion generation .
we investigate to which extent an interpretation of the recently reported excess of events in the large @xmath0 and @xmath1 range at hera in terms of single leptoquark production is compatible with other experimental results .
desy 97032 + hep - ph/9703287 + march 8 , 1997 + z. phys .
* c * ( 1997 ) , in print .
* on the expectations for leptoquarks * + + johannes blmlein + + _ platanenallee 6 , d15735 zeuthen , germany
_ + |
over the years , computation modeling and hydrodynamic calculations of supernova explosions due to core - collapse of massive stars have improved our theoretical understanding of how compact objects form . at present ,
two types of supernovae are actively discussed in the literature : ( i ) after iron is synthesized within a massive star , the core completes its final nuclear burning state .
it then becomes degenerate and keeps growing in mass due to silicon shell burning outside the core .
eventually , it crosses the chandrasekhar mass limit and ensues collapse .
when the core reaches nuclear density , it bounces back and explodes the star as a supernova .
( bethe 1990 ; mezzacappa 2005 ; woosley & bloom 2006 ; kotake et al 2006 ; janka et al .
( ii ) on the other hand , if the core is not massive enough to ignite neon , nuclear burning within the core stops .
the core becomes degenerate due to cooling by neutrino emission . because of carbon shell burning outside the core , the mass of the core approaches the chandrasekhar limit .
when the core density and temperature rise above a certain threshold , electron capture by nuclei ( such as @xmath3 , @xmath4 , @xmath5 , and @xmath6 ) takes place , decreasing degenerate electron pressure support . as a result , the core contracts rapidly and eventually bounces back , driving a supernova explosion .
this scenario is referred to as electron capture supernova ( ecs ) ( miyaji et al .
1980 ; nomoto 1987 ) .
although the general idea of ecs has been established for several decades , the mechanism of how electron capture occurs in these relatively less massive cores is not yet settled .
regardless of the supernova mechanism , it is expected that some asymmetries develop during the collapse , imparting a recoil kick to the nascent neutron star ( see harrison et al .
1993 ; frail et al . 1994 ; lyne & loimer 1994 ) .
the observed proper motions of 233 galactic pulsars show that there is a class of neutron stars receiving high recoil velocities at birth ( hobbs et al . 2005 ) . on the other hand , pfahl et al .
( 2002 ) studied the orbital parameters of low eccentric be / x - ray binaries , and pointed out that these neutron stars could only have received low kick velocities ( < 50 km / s ) at the time of formation .
there is thus evidence for the existence of two different neutron star formation mechanisms ( van den heuvel 2004 , 2007 ) , potentially distinguishable by the magnitude of the natal kick of the neutron star . currently , there are 8 dns observed in our galaxy .
their orbital parameters are summarized in table 1 .
the magnitude of the natal kick and the progenitor mass of the second - born neutron star ( ns2 ) in these systems , except psr j1906 + 0746 , have been studied previously by willems et .
al . ( 2004 , 2006 ) , thorsett et . al .
( 2005 ) , stairs et al .
( 2006 ) , and wang et al .
( 2006 ) . in their study , wang et al .
( 2006 ) studied the formation of _ all _ dns except psr j1906 + 0746 , and derived constraints on the natal kick magnitude and progenitor mass of ns2 by studying the orbital dynamics of asymmetric supernova explosions , but not their kinematic history .
the other authors focused on psr b1534 + 12 , psr b1913 + 16 , and psr j0737 - 3039 and used both orbital dynamics of asymmetric supernova explosions and the measured proper motion of these systems to constrain the natal kick magnitude and progenitor mass of ns2 .
the authors traced the motion of the systems in the galaxy backwards in time and assumed each crossing of the binary through the galactic disk within the age of the dns to be a possible birth site of the dns .
however , the backward calculation of the motion in the galaxy requires the knowledge of the radial velocity which can not be measured .
willems et al .
( 2004 , 2006 ) , thorsett et al ( 2005 ) and stairs et al .
( 2006 ) therefore considered possible radial velocities drawn from either a uniform or gaussian radial velocity distribution . _
_ in this paper , we avoid the assumption of a present - day radial velocity distribution by carrying out monte carlo simulations of the motion of the observed dns in the galaxy that is forward in time instead of backward . for this purpose ,
we distribute populations of dns progenitors in the galaxy according to a double exponential distribution function .
each system in the populations is assigned a velocity equal to the vector sum of the local galactic rotational velocity and an isotropic kick velocity with a magnitude generated from the kick velocity distribution function , derived from the supernova orbital dynamics constraints for each of the observed systems .
compared to willems et al.(2004 ) , thorsett et al .
( 2005 ) , and wang et al .
( 2006 ) , we also relax the constraint that ns2 s progenitor needs to be more massive than 2.1 m@xmath7 , which is a convectional limit for a helium star that explodes in a core collapse supernova . instead , we restrict the progenitor to be more massive than ns2 only . furthermore we study _ all _ eight known dns in our galaxy , using the up - to - date observational parameters listed in table 1 . _ _ the methodology of the adopted analysis is outlined in more detail in 2 , while results for the individual systems and comparison with earlier studies are discussed in 3 . in 4
, we summarize our results and discuss their implications for the supernova forming the second - born neutron star in the observed dns .
@l@ccccc@ccc@c@c@c@ psr b1534 + 12 & 15 37 09.96 & 11 55 55.55 & 1.02 & 1.34(1 ) & -25.05(2 ) & 250 & 1.3332(10 ) & 1.3452(10 ) & 3.28 & 0.274 & 25(155 ) @xmath8 3.8 + psr b1913 + 16 & 19 15 28.00 & 16 06 27.40 & 8.3(1.4 ) & -3.27(35 ) & -1.04(42 ) & 110 & 1.4408(3 ) & 1.3873(3 ) & 2.80 & 0.617 & 18(162 ) @xmath8 6 + psr j0737 - 3039 & 07 37 51.25 & -30 39 40.71 & 1.15 & -3.82(62 ) & 2.13(23 ) & 210 & 1.337(5 ) & 1.250(5 ) & 1.26 & 0.0878 & < 15 + psr j1518 + 4904 & 15 18 16.80 & 49 04 34.25 & 0.625 & -0.67(4 ) & -8.53(4 ) & 20000 & @xmath9 & @xmath10 & 24.7 & 0.249 & + psr j1756 - 2251 & 17 56 46.63 & -22 51 59.40 & 2.5 & -0.7(2 ) & & 443 & 1.312(17 ) & @xmath11 & 2.70 & 0.181 & + psr j1811 - 1736 & 18 11 55.03 & -17 36 37.70 & 6.0 & & & 1830 & @xmath12 & @xmath13 & 40.7 & 0.828 & + psr j1829 + 2456 & 18 29 34.60 & 24 56 19.00 & 1.2 & & & 12400 & @xmath14 & @xmath15 & 6.36 & 0.139 & + psr j1906 + 0746 & 19 06 48.67 & 07 46 28.60 & 5.4 & & & 0.112 & 1.365(18 ) & 1.248(18 ) & 1.75 & 0.0853 & +
tauris @xmath16 van den heuvel ( 2004 ) reviewed the general scenario for forming dns .
it starts with a binary consisting of two massive zams stars ( @xmath17 ) .
the primary star , which is initially the more massive one , will leave the main sequence first and explode as the first supernova , in which it forms the first - born neutron star ( ns1 ) in this binary
. there may be mass transfer from the primary to the secondary star before that supernova takes place .
as the secondary star leaves the main sequence , wind mass transfer from the secondary to ns1 can occur , leading to the formation of a x - ray binary .
while ns1 is accreting mass from its companion , it is spun up or `` recycled '' , and its spin axis is expected to align with the orbital angular momentum .
later on , the x - ray binary may end up in a common envelope phase , during which ns1 is engulfed by the extended envelope of its companion .
the orbit shrinks and circularizes rapidly due to the frictional forces acting on ns1 as it moves through the envelope of the companion .
the lost orbital energy is deposited into the gas envelope , causing it to become gravitationally unbound and get ejected , leaving behind the companion s bare helium core .
the helium star eventually explodes in a supernova and becomes ns2 , orbiting the recycled , milli - second pulsar ns1 in a compact binary orbit .
mass transfer from the helium star to ns1 can also occur prior to the second supernova explosion . here
, our goal is to use the current dns binary properties listed in table 1 to constrain as best possible the magnitude of the natal kick and the mass of the progenitor of ns2 in the 8 known dns in our galaxy at the time of the second supernova .
we also predict the unknown radial velocity of each system , as well as the total transverse velocity and ns1 spin tilt angle for those systems where these properties have not been measured yet .
we achieve these goals by means of monte carlo simulations , incorporating orbital dynamics during supernova explosions and kinematic of binaries in the galactic potential , as detailed in what follows .
we start our analysis by deriving a probability distribution function ( pdf ) for the natal kick of ns2 , which depends on the semi - major axis and orbital eccentricity immediately after the second supernova ( @xmath18 and @xmath19 respectively ) . as the tidal interaction between two neutron stars in dns is extremely weak ( bildsten & cutler 1992 ) , the major mechanism that changes the orbital semi - major axis and eccentricity of dns is the emission of gravitational waves .
to account for gravitational radiation driven orbital evolution , we use the current observational parameters as initial conditions and integrate equations ( 35 ) and ( 36 ) in junker & schfer ( 1992 ) backwards in time . because of the uncertainty in the true age of the observed dns ( see kiziltan et al .
2009 for details ) , we randomly draw the integration time @xmath20 from a uniform distribution of ages between 0 and @xmath21 , where @xmath21 is the characteristic age of the dns .
if @xmath21 is greater than the age of our galaxy , which is about 10 gyr , we use 10 gyr as the upper limit for @xmath20 instead . for psr b1534 + 12 , psr b1913 + 16 , and psr j0737 - 3039
, different upper limits are adopted for @xmath20 , as discussed in more detail in the next section .
once @xmath18 and @xmath22 are known from the gravitational radiation integration , we randomly draw the natal kick velocity ( @xmath23 ) from a uniform distribution of values between 0 and 2500 km / s ( i.e. , we assume no prior kick distribution derived from pulsar samples analysis ) , and the kick direction @xmath24 from an isotropic distribution . here
@xmath25 is the polar angle between the natal kick velocity and the instantaneous orbital velocity of ns2 s progenitor at the time of the second supernova explosion , and @xmath26 is the corresponding azimuthal angle ( see figure 1 in kalogera 2000 for a graphical representation ) . using the conservation laws of orbital energy and angular momentum expressed by equations ( 19 ) and ( 20 ) in willems et al .
( 2005 ) , we obtain the orbital semi - major axis ( @xmath27 ) immediately before the second supernova and the mass of ns2 s helium star ( @xmath28 ) .
the pre - supernova orbit is assumed to be tidally circularized at the time of the second supernova explosion . on the other hand ,
the spin tilt angle of ns1 , which is the angle between ns1 axis and the post supernova orbital angular momentum axis , can be calculated by following equation ( 3 ) in kalogera ( 2000 ) .
also , using equations ( 12 ) in the same paper , we find the polar angle @xmath29 between the natal kick and the pre - supernova orbital angular momental axis . immediately before the second supernova explosion
, there may be mass transfer from ns2 s progenitor to ns1 .
if the radius @xmath30 of ns2 s progenitor is greater than its roche lobe radius @xmath31 , which is approximated by equation ( 2 ) in eggleton ( 1983 ) , roche lobe overflow ( rlo ) takes place . since
ns2 s progenitor is a bare helium star before the second supernova , we use equation ( a10 ) in kalogera & webbink ( 1998 ) to calculate @xmath30 . note that the left hand side of that equation should be @xmath32 instead of @xmath33 . in order to have dynamically stable rlo , so that ns2 s progenitor does not coalesce with ns1 to form an isolated black hole , the mass ratio @xmath28 to @xmath34 needs to be less than 3.5 ( ivanova et al .
even if ns2 s progenitor does not overflow its roche lobe , there may be mass transfer by stellar wind .
in addition , @xmath28 is restricted to be less than 8 m@xmath7 , which is a conventional limit that a helium star will become a neutron star instead of a black hole ( see , e.g. , figure 1 in belczynski et al .
2002 ; table 16.4 in tauris & van den heuvel 2004 ) .
the conservation laws of orbital energy and angular momentum allow for real solutions only for @xmath28 , @xmath23 , @xmath25 , @xmath26 , and @xmath27 values satisfying the orbital dynamics constraints given by equation ( 21)(27 ) in willems et al .
( 2005 ) . for psr b1534 + 12 , psr b1913 + 16 , and psr j0737 - 3039 ,
these constraints are supplemented with the measured ns1 spin tilt angle constraints . applying these constraints together with the constraints imposing stability of any mass transfer taking place right before the second supernova explosion yields a pdf of the natal kick velocity of ns2 .
a corresponding pdf for the kick velocity @xmath35 imparted to the binary s center of mass is obtained using equations ( 28)(32 ) in willems et al .
in all past studies , kinematic analyses have either been skipped ( wang et al .
2006 ) or simplified to one dimension ( piran & shaviv 2004 ) or have suffered by ad hoc assumptions about the un - measurable radial velocities . willems et al .
( 2006 ) used general dns population synthesis models to guide this assumption . in this sutdy
we employ a new method that still uses forward in time kinematic simulations , but they are tied to each observed system instead of being general .
we perform a monte carlo simulation of the motion of dns in the galactic potential .
for this purpose , we adopt a reference frame with origin at the galactic center , with @xmath36-axis pointing to the northern galactic pole , and with the @xmath37-axis pointing in the direction from the sun to the galactic center . with respect to this reference frame
, the sun is located at @xmath38 kpc ( ghez et al .
2008 , gillessen et al .
2009 , reid et al .
2009 , joshi 2007 ) and has a peculiar velocity of @xmath39km / s ( bienaym 1999 ) . with the given @xmath40 and peculiar velocity , the sun is moving at a velocity @xmath41 km / s . to simulate the motion of dns in the galaxy ,
we randomly distribute a population of newly formed dns in the galactic disk according to a double exponential distribution function @xmath42 here r and z are the cylindrical galactic coordinates , @xmath43 and @xmath44 are the galactic scale length and height respectively .
the distribution is normalized to unity by setting @xmath45 .
since the progenitors of dns are distributed in the same way as the massive stellar binaries in our galaxy , we choose @xmath43 = 2.8 kpc and @xmath44 = 0.07 kpc ( see joshi 2007 ) .
for all of the observed dns except psr b1534 + 12 , we use a population of @xmath210@xmath46 simulated binaries .
we use a population of @xmath210@xmath47 in the analysis of psr b1534 + 12 , due to the limiting chance of kicking it to the currently high galactic altitude and simultaneously satisfying the constraints derived from the orbital dynamics of supernova explosions .
if the scale height @xmath44 were larger than 70 pc , it would reduce the difficulty in satisfying the galactic dynamics constraints of psr b1534 + 12 . a smaller population of simulated binaries would then be needed for the analysis of this system .
the initial center - of - mass velocity of the dns is obtained by summing up the kick velocity @xmath48 imparted to the binary center of mass during the second supernova explosion and the local galactic rotational velocity @xmath49 . given the unknown orientation of the binary orbit in the galactic reference frame , the systemic kick velocity @xmath48 is assumed to be distributed isotropically in space .
the magnitude of @xmath48 is drawn from the velocity distribution derived in the previous step specific to each of the observed systems . starting from the randomly generated initial position and velocity ,
the motion of the dns is calculated by numerically integrating the equations of motion from the dns birth time to the current epoch .
the equations of motion are derived from the galactic potential of carlberg & innanen ( 1987 ) with updated model parameters of kuijken & gilmore ( 1989 ) .
the dns birth time is chosen according to the dns age @xmath20 generated in the first step of the analysis .
lastly , we use the measured current position of the observed dns to further constraint the ranges of possible @xmath23 and @xmath28 .
we set the tolerance of this constraint to be within 100 pc from the observed position .
we test the dependence of our results on this tolerance level by repeating the analysis with a 10 pc , or 50 pc tolerance , but our results do not show any significant changes . for psr j1518 + 4904 and psr j1756 - 2251 , the position constraint is supplemented with the measured current proper motion constraint . for psr b1534 + 12 , psr b1913 + 16 , and psr j0737 - 3039 , the position constraint is supplemented with both the measured proper motion and ns1 spin tilt angle constraints .
the analysis outlined in the previous section allows us to derive pdfs for the kick velocity @xmath23 , ns2 s immediate pre - supernova progenitor mass @xmath28 , the spin tilt @xmath50 of ns1 , the systemic radial velocity @xmath51 , and the current transverse velocity @xmath52 .
our primary interest is in @xmath23 and @xmath28 , because they can shed light on the type of supernova that formed ns2 , which will be discussed in more detail in the next section .
the remaining quantities give us some predictions on currently unavailable parameters of the 8 known dns in our galaxy . instead of focusing on the most likely values as studies have done in the past
, we derive confidence levels for the unknown parameters and they are summarized in table 2 .
since the @xmath23 and @xmath28 constraints are correlated , we present our results in the form of confidence level plots of a 2 dimensional joint probability distribution of @xmath23 - @xmath28 , for each dns binary , in addition to the corresponding 1 dimensional pdf s of @xmath23 and @xmath53 . as pointed out by willems et al ( 2006 ) , physical parameter constraints are affected by the dimensionality of the pdf used , because of inherent correlations between parameters and projection effects .
this variation is stronger when one solely examines the most likely values ( i.e. , values at the peak of the pdf ) and weakens as broader confidence levels are considered . in what follows ,
the confidence levels of @xmath23 and @xmath28 are derived from the 2d joint pdf . to calculate the 2d confidence levels , we first bin the data into a 2d grid and normalize the total probability within the 2d grid to unity .
then , we add the probability of each bin from the highest to the lowest , until the sum best matches the desired confidence level . on the other hand ,
the ranges of @xmath51 , @xmath52 , and @xmath50 are derived from the corresponding 1d pdf . in this case , the confidence level is found by having the shortest range of bins that has a total probability that best matches the desired confidence level .
lccccc psr b1534 + 12 & 170 - 260 ( 150 - 270 ) & 2.00 - 2.90 ( 1.34 - 3.40 ) & @xmath54 - 50 ( @xmath55 - 145 ) & 121 @xmath8 6 & 25.0 @xmath8 3.8 + psr b1913 + 16 & 200 - 410 ( 190 - 450 ) & 1.40 - 3.30 ( 1.39 - 5.00 ) & @xmath56 - 85 ( @xmath57 - 315 ) & 135 @xmath8 25 & 18 or 162 @xmath8 6 + psr j0737 - 3039 & 5 - 50 ( 5 - 120 ) & 1.25 - 1.55 ( 1.25 - 1.90 ) & 8 - 44 ( @xmath58 - 76 ) & @xmath59 & 0 - 2.0 ( 0 - 7.8 ) + psr j1518 + 4904 & 20 - 80 ( 5 - 110 ) & 1.80 - 3.30 ( 1.49 - 4.70 ) & @xmath60 - 8 ( @xmath61 - 48 ) & 25.4 @xmath8 3.6 & 2 - 18 ( 0 - 32 ) + psr j1756 - 2251 & 5 - 80 ( 5 - 185 ) & 1.25 - 1.90 ( 1.25 - 2.65 ) & @xmath62 - 32 ( @xmath55 - 98 ) & 6.5 - 28.5 ( 6.0 - 64.5 ) & 0 - 3.4 ( 0 - 16.4 ) + psr j1811 - 1736 & 0 - 170 ( 0 - 310 ) & 1.11 - 4.00 ( 1.11 - 8.00 ) & 0 - 112 ( @xmath63 - 210 ) & 30 - 114 ( 4 - 202 ) & 0 - 10.2 ( 0 - 57.6 ) + psr j1829 + 2456 & 5 - 85 ( 5 - 225 ) & 1.40 - 2.70 ( 1.36 - 6.10 ) & @xmath60 - 26 ( @xmath64 - 66 ) & 12 - 54 ( 4 - 102 ) & 0 - 8.5 ( 0 - 30.0 ) + psr j1906 + 0746 & 5 - 170 ( 5 - 510 ) & 1.25 - 2.90 ( 1.25 - 4.80 ) & 12 - 103 ( @xmath65 - 297 ) & 89 - 174 ( 21 - 376 ) & 0 - 6.7 ( 0 - 51.5 ) + in this binary , the masses of ns1 and ns2 are 1.33 @xmath66 and 1.35 @xmath66 respectively . the orbital period is @xmath67 days , with an eccentricity @xmath68 .
the angle between ns1 s spin axis and the orbital angular momentum axis is @xmath69 or @xmath70 .
the binary is currently located out of the galactic plane , at @xmath71 and @xmath72 , and 1.02 kpc away from us .
the observed proper motion @xmath73 mas yr@xmath74 and @xmath75 mas yr@xmath74 . at the measured distance
this implies @xmath76 km s@xmath74 and @xmath77 km s@xmath74 , in the reference frame of the sun .
also , we use the spin down age , which is 210 myr given by arzoumanian et al . 1999 , as the upper limit of psr b1534 + 12 s age , instead of the characteristic age .
furthermore , the current orientation of the orbital angular momentum axis is also constrained for this sytem .
the orientation axis can be described by the inclination angle @xmath78 and the angle of the orbital ascending node on the plane of the sky ( @xmath79 ) .
the sine of the inclination angle is @xmath80 = 0.975 ( stairs et al .
the angle @xmath79 is @xmath81 or @xmath82 , reckoned north through east .
the two solutions correspond to @xmath83 and @xmath84 respectively ( bogdanov et al .
2002 ) . since tidal effects between the two neutron stars
are negligible , the orbital angular momentum axis keeps a fixed orientation in space . from the kinematic analysis of the galactic motion ,
the constraints on the present - day orbital inclination and proper motion can therefore be used to determine the systemic kick component @xmath85 parallel to the post - supernova orbital angular momentum axis .
following kalogera ( 1996 ) and wex et al .
( 2000 ) , @xmath86 can also be expressed analytically as @xmath87 where @xmath88 @xmath89 @xmath90 here , @xmath91 ( @xmath92 ) is the @xmath37 , @xmath93 , or @xmath36 component of the natal kick velocity in the frame centered on ns2 s progenitor ( see figure 1 in kalogera for a graphical representation ) .
note that the sign of @xmath86 needs to be consistent with the sign of @xmath94 .
hence , through equation ( 2 ) , the @xmath86 derived from observations gives us a constraint on the y and z components of the natal kick , as well as @xmath50 .
the results are shown in figures 1 and 2 .
right before the second supernova , there must have been rlo from the progenitor of ns2 to ns1 , as noted previously by thorsett et al .
as seen in the 2d @xmath23-@xmath28 joint probability distribution in figure 1 , the kick imparted to ns2 at birth is between 150 and 270 km / s , while the progenitor mass @xmath28 is between 1.34 and 3.40 m@xmath0 , both at 95% confidence .
as seen in figure 2 , polar kicks ( i.e. , @xmath95 or @xmath96 ) are unlikely , and the possible kick directions are asymmetric about the pre - supernova orbital plane ( i.e. , @xmath29 = 90@xmath97 ) .
this asymmetry comes from the constraint on the orientation of the post - supernova orbital angular momentum axis on the sky .
moreover , we arrive at the same conclusion as thorsett et al .
( 2005 ) that @xmath98 is very unlikely .
the current radial velocity of this binary is between -110 and 145 km / s , at 95% confidence , and it is likely to be moving away from us . within the observationally constrained range , the @xmath52 pdf is roughly flat . when comparing with recent analyses by willems et al .
( 2004 ) , thorsett et al .
( 2005 ) , and wang et .
2006 ) , our limits on @xmath23 and @xmath28 are consistent wtih thorsett et al .
( 2005 ) , and are more constrained than willlems et al .
( 2004 ) and wang et al .
( 2006 ) , because willems et al .
( 2004 ) did not have the spin tilt @xmath50 measurement and wang et al .
( 2006 ) did not have the proper motion and position constraints in their analysis .
our limits on the angle @xmath29 , which is the polar angle between the natal kick direction and the pre - supernova orbital angular momentum , are not as constrained as the limits shown in thorsett et al .
( 2005 ) . through recent further investigation
, it has been found that this difference is due to a code typo related to the galactic latitude of the source , which fortunately affected significantly only the @xmath29 results ( dewey & stairs 2010 , private communication ) .
the masses of ns1 and ns2 in this binary are 1.44 @xmath66 and 1.38 @xmath66 respectively .
the orbital period is @xmath99 days , with an eccentricity @xmath100 .
the angle between the ns1 spin axis and the post supernova orbital angular momentum axis is @xmath101 or @xmath102 . at the current time , this binary is about 300 pc away the galactic plane , at @xmath103 and @xmath104 , and 8.3 kpc away from us .
the observed proper motion in right ascension and declination is @xmath105 mas yr@xmath74 and @xmath106 mas yr@xmath74 . at the measured distance
, this implies @xmath107 km s@xmath74 and @xmath108 km s@xmath74 , in the reference frame of the sun . in addition , instead of the characteristic age , we use the spin down age , which is 80 myr ( arzoumanian et al .
1999 ) , as the upper limit on psr b1913 + 16 s age .
the results are displayed in figures 3 and 4 .
as shown in the 2d @xmath23-@xmath28 joint probability distribution in figure 3 , the kick imparted to ns2 at birth is between 190 and 450 km s@xmath74 , while the progenitor mass @xmath28 is between 1.4 and 5.0 m@xmath0 , both at 95% confidence .
as seen from figure 4 , both polar and planar kick directions are allowed .
the current radial velocity is between @xmath57 and 315 km s@xmath74 at 95% confidence . within the observationally constrained range ,
the @xmath52 pdf is roughly flat , with a slight preference for transverse velocities of @xmath2120 km / s .
furthermore , although observations show a degeneracy in @xmath50 at @xmath109 and @xmath110 , our analysis shows that @xmath50 is much more likely to be @xmath109 .
in addition , although our confidence intervals of @xmath23 and @xmath28 are consistent with willems et al .
( 2004 ) and wang et .
al . ( 2006 ) , our reported intervals are tighter , because willems et al . ( 2004 ) and wang et al . ( 2006 ) reported the allowed ranges of @xmath23 and @xmath28 instead of confidence intervals .
the masses of the two neutron stars in this binary are 1.34 and 1.25 m@xmath0 , where the more massive one is ns1 .
they are orbiting each other with a period of 0.102 days , and an eccentricity of 0.0878 . at the current time
, the binary is close to the galactic plane , at @xmath111 and @xmath112 . through direct measurement of geometric parallax ,
deller et al . (
2009 ) measured the distance to psr j0737 - 3039 to be 1.15 kpc , and the proper motion in r.a . and declination to be @xmath1133.82 and 2.13 mas yr@xmath74 respectively .
at the measured distance , this implies @xmath114 km s@xmath74 and @xmath115 km s@xmath74 , in the reference frame of the sun .
in addition , lorimer et al .
( 2007 ) constrained the possible age range to be either @xmath116 or @xmath117 myr .
we adopt these age ranges in our analysis of psr j0737 - 3039 . as the radio pulse profile of psr j0737 - 3039a
( i.e. , ns1 ) has not changed significantly through six years of observation , the spin tilt angle of j0737 - 3039a with respect to the post supernova orbital angular momentum axis is believed to be small . however , there is no well constrained value for this angle .
ferdman et al .
( 2008 ) and breton ( 2008 ) argue that there are two possible models to explain the observed pulse profile : the single - cone and two - cone emission model . in this context , the number of emission cones means the number of emitting cones that intersect with our line of sight .
each model gives a different spin tilt angle estimate .
ferdman et al .
( 2008 ) estimate this angle to be < 15@xmath97 for the single - cone model and < 6.1@xmath97 for the two - cone model , both at 68.3 % confidence . due to this uncertainty , we take the spin tilt angle @xmath50to be < 15@xmath97 in our work , which corresponds to 68.3 % confidence value in the single - cone model and 95.4 % confidence value in the two - cone model in ferdman et al ( 2008 ) .
the results are displayed in figures 5 - 7 .
as shown in the 2d @xmath23-@xmath28 joint probability distribution , the kick recieved by ns2 ( i.e. , psr j0737 - 3039b ) at birth is between 5 and 120 km s@xmath74 , while the progenitor mass @xmath28 is between 1.25 and 1.90 m@xmath0 , both at 95% confidence .
the probability distribution of the kick direction is symmetric about the pre - supernova orbital plane ( i.e. , @xmath118 = 90@xmath97 ) , and planar kicks are more favorable than polar kicks .
the pdf of the semi - major axis immediately before the second supernova ( @xmath27 ) peaks at two narrow ranges : @xmath119 and @xmath120 @xmath121 .
the first peak corresponds to an age of @xmath116 myr , and the second peak to an age of @xmath117 myr .
at the current time , psr j0737 - 3039 is more likely to be moving away from us , with a radial velocity between @xmath58 and 76 km / s with 95% confidence . within the observationally constrained range , the @xmath52 pdf peaks at @xmath220 km / s .
furthermore , rlo from ns2 s progenitor to ns1 before the second supernova explosion can not be avoided . to test the dependence of the presented results on the adopted upper limit for @xmath50
, we repeated the analysis for the upper limits of @xmath122 and @xmath123 .
the results do not show any significant differences compared to those obtained with the @xmath124 upper limit .
this is not surprising , because the @xmath50 pdf as shown in figure 8 , which is obtained from the same constraints as before except raising the upper limit of @xmath50 to 60@xmath97 , peaks strongly at 0@xmath97 and decays exponentially with increasing values of @xmath50 . comparatively , our confidence intervals on @xmath23 and @xmath28 are consistent with stairs et al .
( 2006 ) and willems et al .
( 2006 ) , but are much more constrained than wang et al .
( 2006 ) , mainly because they did not have the proper motion constraint in their analyses .
the masses of the two neutron stars in this binary are 0.72 and 2.00 m@xmath0 , where the less massive one is ns1 .
however , each one has an uncertainty of @xmath20.5 m@xmath0 . in our analysis
, we choose @xmath34 to be 1.23 and @xmath125 to be 1.49 , so that @xmath34 is greater than the smallest measured neutron star mass , which is about 1.2 @xmath126 .
the characteristic age is 20 gyr , which is larger than a hubble time .
thus , we set the upper limit on psr j1518 s age to be 10 gyr instead .
the two neutron stars are orbiting each other with a period of 8.63 days , and an eccentricity of 0.249 . at the current time
, this binary is out of the galatic plane , at @xmath127 and @xmath128 , and at a distance of 0.625 kpc from the sun . the proper motion in r.a . and
dec . is @xmath129 and @xmath130 mas yr@xmath74 . at the measured distance , this implies @xmath131 and @xmath132 km s@xmath74 , in the reference frame of the sun .
there is no spin tilt measurement available for this system .
the results are shown in figures 9 and 10 . as seen in the 2d @xmath23-@xmath28 joint probability distribution ,
the kick imparted to ns2 at birth is between 5 and 110 km / s , while the progenitor mass @xmath28 is between 1.5 and 4.7 m@xmath0 , both at 95% confidence .
both polar and planar kick directions are possible , but the planar kicks have a higher probability . the spin tilt angle @xmath50 is smaller than 32@xmath97 at 95% confidence .
the radial velocity is found to be between -62 and 48 km / s at 95% confidence , and the system is currently likely moving towards us . within the observationally constrained ranges ,
the @xmath52 pdf is roughly flat .
our limits of @xmath23 and @xmath28 are much more constrained than those of wang et al .
( 2006 ) , primarily because they did not have the proper motion and galactic position constraints included in their analysis .
the masses of the two neutron stars in this binary are 1.31 and 1.26 m@xmath0 , where the more massive one is ns1 .
the characteristic age is 443 myr .
the two neutron stars are in an orbit with a period of 0.320 days , and an eccentricity of 0.181 . at the current time , this binary is close to the galactic plane , at @xmath133 and @xmath134 , and at a distance of 2.5 kpc away from us .
only proper motion @xmath135 is measured , which is -0.7 mas / yr . at the measured distance , this implies @xmath136 km / s , in the reference frame of the sun .
there is currently no spin tilt measurement available for this system .
the results are shown in figures 11 and 12 . from the 2d @xmath23-@xmath28
joint probability distribution , the kick received by ns2 at birth is between 5 and 185 km s@xmath74 , while the progenitor mass @xmath28 is between 1.25 and 2.65 m@xmath0 , both with 95% confidence .
both polar and planar kicks are allowed , but planar kicks are more favorable than polar ones . the spin tilt angle of ns1 is smaller than 16.4@xmath97 with 95% confidence . at the current time and at the measured distance ,
psr j1756 - 2251 has a total transverse velocity between 6.0 and 64.5 km s@xmath74 , and a radial velocity between @xmath55 and 98 km s@xmath74 , both with 95% confidence .
furthermore , this binary is equally likely to be moving towards us as away from us . when comparing with wang et al .
( 2006 ) , we have tighter constrained limits on @xmath23 and @xmath28 , due to the fact that we have the additional constraint on the proper motion @xmath135 .
the two neutron stars have masses of 1.62 and 1.11 m@xmath0 , where the more massive one is ns1 .
however , each mass measurement has an uncertainty of @xmath20.5 m@xmath0 .
psr j1811 has a characteristic age of 1830 myr .
the neutron stars are in an orbit with a period of 18.8 days , and an eccentricity of 0.828 . at the current time , this binary is close to the galactic plane , at @xmath137 and @xmath138 , and at a distance of 6.0 kpc away from us . like psr j1756 - 2251 , there are currently no spin tilt nor proper motion measurements available .
the results are displayed in figures 13 and 14 .
the 2d @xmath23-@xmath28 joint probability distribution shows that the kick given to ns2 at birth is less than 310 km s@xmath74 , while the progenitor mass @xmath28 is between 1.11 and 8.0 m@xmath0 , both with 95% confidence .
both polar and planar kicks are allowed , but the planar kicks are more favorable than the polar ones .
the spin tilt angle @xmath50 is smaller than 57.6@xmath97 with 95% confidence . at the measured distance
, the total transverse velocity of psr j1811 - 1736 at the current epoch is between 4 and 202 km s@xmath74 , and the radial velocity between -90 and 210 km s@xmath74 with 95% confidence , which means this binary is likely moving away from us .
although our upper boundary of the 95% confidence intervals of @xmath28 is the same as the upper limit found by wang et al .
( 2006 ) , the value of our lower boundary is less than their lower limit .
this is because wang et al .
( 2006 ) put a conservative lower limit of 2.1 m@xmath0 on the progenitor mass of a neutron star , which we did not . on the other hand ,
our limits on @xmath23 are more constrained than those of wang et al .
( 2006 ) , as we compute confidence levels instead of allowed ranges of solution as wang et al .
( 2006 ) . the two neutron stars of this binary have masses of 1.14 and 1.36 m@xmath0 , where the less massive one is ns1 .
however , there is an uncertainty of @xmath20.5 m@xmath0 in the mass measurements .
the characteristic age is 12.4 gyr .
for the same reason as psr j1518 + 4904 , we set an the upper limit of 10 gyr on the age of this binary in our analysis .
the neutron stars are orbiting each other with a period of 1.18 days , and an orbital eccentricity of 0.139 . at current time , this binary is out of the galactic plane , at @xmath139 and @xmath140 , and it is 1.2 kpc away from us .
although there is no spin tilt nor proper motion measurements available at current time , lorimer et al .
( 2005 ) derived that the transverse velocity is smaller than 118 km s@xmath74 .
the results are shown in figure 15 .
the @xmath23-@xmath28 joint probability distribution shows that the kick given to ns2 at birth is between 5 and 225 km s@xmath74 , while the progenitor mass @xmath28 is between 1.4 and 6.1 m@xmath0 with 95% confidence .
the pdfs of @xmath50 and @xmath29 are similar to those of psr j1756 - 2251 .
both polar and planar kicks are allowed , but planar kicks are more probable . the spin tilt angle @xmath50 is less than 30.0@xmath97 with 95% confidence . at the known distance , the total transverse velocity of psr j1829 + 2456 at the current time is between 4 and 102 km s@xmath74 , and the radial velocity between @xmath64 and 66 km s@xmath74 with 95% confidence .
in addition , the probability that this system is moving towards the sun is @xmath220% higher than that it is moving away from the sun . comparing our @xmath23 and @xmath28 results with wang et al .
( 2006 ) , we have more constrained limits , as they do not include any kinematic constraints .
the masses of the neutron stars are 1.37 and 1.25 m@xmath0 , where the more massive one is ns1 . unlike the other 7 binaries , the observed pulsar is ns2 .
this implies the binary is relatively young , with a charateristic age of 0.112 myr only .
the neutron stars are in an orbit with a period of 0.166 days , and an orbital eccentrirctiy of 0.0853 . at current time , this binary is close to the galactic plane , at @xmath141 and @xmath142 , and it is 5.4 kpc away from us .
again , there are no spin tilt nor proper motion measurements available yet .
the results are displayed on figure 16 .
the @xmath23-@xmath28 joint probability distribution shows that the kick given to ns2 at birth is between 5 and 510 km s@xmath74 , while the progenitor mass @xmath28 is between 1.25 and 4.80 @xmath126 with 95% confidence .
the @xmath29 and @xmath50 pdf s are similar to those of psr j1756 - 2251 .
both polar and planar kicks are possible , with the planar kicks having a higher probability .
the polar angle @xmath50 is less than 51.5@xmath97 with 95% confidence .
at the measured distance , the total transverse velocity of psr j1906 + 0746 at the current time is between 21 and 376 km s@xmath74 , and the radial velocity is between @xmath65 and 297 km s@xmath74 with 95% confidence .
it is most likely that this binary is moving away from us .
furthermore , right before its supernova , the progenitor of ns2 must have been overflowing its roche lobe , transferring mass to ns1 .
our limits on @xmath23 and @xmath28 are tighter constrained than wang et al .
( 2006 ) , because they did not have a detailed analysis on this system .
instead , they calculated the allowed @xmath23 with an assumption that @xmath28 was between 2.1 and 8.0 m@xmath0 .
the anisotropic mass and neutrino ejection in a supernova explosion gives a recoil natal kick to the newly formed neutron star .
iron core collapse supernova and ecs produce natal kicks of different strength because of the different physical conditions of the nascent neutron star immediately before the supernovae .
there are several reviews ( bethe 1990 , mezzacappa 2005 , woosley & bloom 2006 , kotake et al 2006 , and janka et al .
2007 ) that summarize our current understanding of the iron core collapse supernova . after iron
is synthesized in the core , the massive star reaches the final stage of hydrostatic nuclear burning , because the synthesis of any heavier elements requires energy instead of releasing energy .
when the iron core grows by silicon shell burning around the core to a mass above the chandrasekhar mass limit , electron degeneracy pressure can not support the core any more and it starts to collapse .
the iron core continues to collapse until it reaches nuclear density ( i.e. , @xmath143 g @xmath144 ) . since nuclear matter
has a much lower compressibility , the collapse decelerates and the core bounces back because of the increased nuclear matter pressure .
this drives a shock wave and eventually explodes the outer layers of the massive star away in a supernova explosion . on the other hand ,
the ecs scenario is described in the reviews of miyaji et al .
( 1980 ) and nomoto ( 1987 ) .
after carbon burning within the core of a post main sequence massive star , a @xmath145 core is formed .
neon can not be ignited if the mass of the @xmath145 core is less than the critical mass of 1.37 m@xmath0 for neon ignition . as the @xmath145 core is cooled by neutrino emissions ,
it becomes strongly degenerate .
when the mass of the core grows by carbon shell burning and approaches the chandrasekhar mass limit , electron captures onto @xmath6 and @xmath5 take place .
this leads to a decrease in the electron degeneracy pressure and the chandrasekhar mass limit , which induces a rapid contraction of the core .
the rapid core contraction ignites the oxygen deflagration , which incinerates materials into nuclear statistical equilibrium ( nse ) . due to the rapid electron capture onto nse elements , the collapse of the core accelerates .
similar to the iron core collapse picture , when the core reaches the nuclear matter density , it bounces back and drives a shock wave that eventually explodes the star in a supernova .
gutirrez et al .
( 2005 ) show that the abundance of @xmath6 in @xmath145 core needs to be greater than 15% , in order to have nse developed through electron captures .
however , in simulations of evolving massive agb stars by siess ( 2007 ) , the abundance of @xmath6 after carbon burning is smaller by a factor of @xmath210 than what is required to drive an explosion by electron captures . therefore , the mechanism of developing an ecs is not a settled issue yet .
in addition , podsiadlowski et al .
( 2004 ) studied the binary evolution and the dynamics of core collapse , and suggested that ecs could only occur in interacting binaries , but not in single stars . in recent simulations of supernova explosions from the collapse of @xmath145 cores by kitaura et al .
( 2006 ) , they found these supernovae are powered by neutrino heating and neutrino - driven wind of the nascent neutron star .
scheck et al .
( 2006 ) showed that the shock wave in these supernovae can propagate outwards on a relatively short timescale after the rebound of the core , which means the non - radial hydrodynamics instabilities do not have time to merge and grow to global asymmetry before the anisotropic pattern freezes out in the accelerating outward motion of the shock wave . as a result ,
_ _ the natal kick of an ecs due to anisotropic mass ejection is expected to be fairly small(see scheck et al .
furthermore , as the @xmath145 core that eventually explodes in an ecs is not massive enough to ignite neon , _ _ the mass of the neutron star progenitor should be less than the progenitors that explode in an iron core collapse supernova .
_ _ _ _ the present analysis is the first to include _ all _ currently known dns systems and account of all orbital and kinematic constraints .
we also employ a novel method for dealing with the uncertainty due to the un - measurable radial velocities and we focus on the derived pdfs and associated confidence levels , instead of just the most likely values , which can be misleading at times .
our results are summarized in table 2 ; the derived constraints are consistent with earlier studies ( when available ) but typically limit parameters to narrower ranges , and in this sense they represent the best available constraints on the formation of neutron stars in dns systems . _ _ in the context of our current understanding of massive star core collapse , we can use our results to draw a number of conclusions : + ( 1 ) psr j0737 - 3039 has a @xmath23 upper limit of 120 km / s and a @xmath28 upper limit of 1.9 @xmath126 at 95% confidence .
therefore , the formation of ns2 ( i.e. , pulsar b ) likely occurred through an ecs event .
this is consistent with the speculation of podsiadlowski et al .
( 2005 ) based on the space velocity and orbital eccentricity of psr j0737 - 3039 . for psr b1534 + 12 and psr b1913
+ 16 , @xmath23 are 150 - 270 km / s and 190 - 450 km / s ( 95% ) respectively , which means ns2 in both system must have received a significant recoil natal kick at birth .
psr b1534 + 12 has a @xmath28 upper limit of 3.4 @xmath126 ( 95% ) , while psr b1913 + 16 has a @xmath28 upper limit of 5.0 @xmath28 ( 95% ) .
because of the relatively high @xmath23 ranges and @xmath28 upper limits , ns2 in both systems are probably formed through an iron core collapse supernova event .
+ ( 2 ) psr j1518 + 4904 has a @xmath146 upper limit of 110 km / s and a @xmath28 upper limit of 4.7 @xmath126 at 95% confidence .
even though ns2 likely received a low recoil velocity at birth , we can not firmly conclude which type of supernova occurred during the formation of ns2 , due to the relatively high @xmath28 upper limit . +
( 3 ) psr j1756 - 2251 has a @xmath23 upper limit of 80 km / s and a @xmath28 upper limit of 1.90@xmath126 at 60% confidence , hence the formation of ns2 might possibly relate to an ecs event .
however , the @xmath23 and @xmath28 upper limits are 185 km / s and 2.65 @xmath126 respectively at 95% confidence , which means the likelihood of ns2 formed through an ecs event is lower than that of psr j0737 - 3039 .
+ ( 4 ) psr j1811 - 1736 , psr j1829 + 2456 , and psr j1906 + 0746 have @xmath23 upper limits of several hundreds of km / s , and @xmath28 upper limits ranging from 4.80 to 8.0 @xmath126 ( i.e. , the conventional limit on neutron star progenitor mass ) at 95% confidence .
since both low and high @xmath23 and @xmath28 are possible , we can not conclude which type of supernova event is more favorable for the formation history of ns2 in these systems .
+ ( 5 ) for all of the known dns except psr b1534 + 12 , all available constraints are consistent with imparting a polar or planar kick to ns2 at birth . for psr b1534 + 12
, the kick direction is constrained to be polar .
+ ( 6 ) for psr b1534 + 12 , psr j0737 - 3039 and psr j1906 + 0746 , the pre - supernova orbit is so tight that they can not avoid a rlo from ns2 progenitor to ns1 just before the second supernova explosion .
+ ( 7 ) furthermore , despite the low orbital eccentricity in psr j1829 + 2456 and psr j1906 + 746 , ns2 could have a high progenitor mass and have received a high recoil natal kick at birth , as shown in the 2d joint pdf of these systems ( see figures 15 , and 16 ) .
+ ( 8) we also tested the dependence of our psr j0737 - 3039 results on the current upper limit of the spin - orbital misalignment angle @xmath50 of ns1 ( i.e. , pulsar a ) . as shown in figure 8 , the @xmath50 pdf peaks strongly at @xmath147 ,
so choosing an upper limit of 15@xmath97 or 60@xmath97 does not affect our results noticably .
arzoumanian , z. , cordes , j. m. , & wasserman , i. 1999 , , 520 , 696 belczynski , k. , kalogera , v. , bulik , t. 2002 , , 572 , 407 bethe , h. a. 1990 , rev .
phys . , 62 , 801 bienaym , o. 1999 , , 341 , 86 bildsten , l. , cutler , c. 1992 , , 400 , 175 bogdanov , s. , pruszyska , m. , lewandowski , w. , & wolszczan , a. 2002 , , 581 , 495 breton , r. p. 2009
thesis at mcgill university burgay , m. , damico , n. , possenti , a. , manchester , r. n. , lyne , a. g. , et al .
2003 , , 426 , 531 champion , d. j. , lorimer , d. r. , mclaughlin , m.a . ,
cordes , j. m. , arzoumanian , z. , et al .
2004 , , 350 , l61 champion , d. j. , lorimer , d. r. , mclaughlin , m. a. , xilourix , k. m. , arzoumanian , z. , et al .
2005 , , 363 , 929 carlberg , r. g. , & innanen , k. a. 1987 , , 94 , 666 corongiu , a. , kramer , m. , stappers , b. w. , lyne , a. g. , jessner , a. , et al .
2007 , , 462 , 703 damour , t. , & taylor , j. h. 1991 , , 366 , 501 deller , a. t. , bailes , m. , & tingay , s. j. 2009 , science , 323 , 1327 eggleton , p. p. 1983
, , 268 , 368 faulkner , a. j. , kramer , m. , lyne , a. g. , manchester , r. n. , mclaughlin , m. a. , et al .
2005 , , 618 , l119 ferdman , r.d .
2008 , phd .
thesis at the university of british columbia ferdman , r. d. , stairs , i. h. , kramer , m. , manchester , r. n. , lyne , a. g. , et al . in 40 years of pulsars : millisecond pulsar , magnetars and more .
aip conference proceedings , 2008 , pp .
474 - 478 frail , d. a. , goss , w. m. , & whiteoak , j. b. z. 1994 , , 437 , 781 ghez , a. m. , salim , s. , weinberg , n. n. , lu , j. r. , do , t. , et al . , 2008 , , 689 , 1044 gillessen , s. , eisenhauer , f. , trippe , s. , alexander , t. , genzel , r. , et al .
2009 , , 692 , 1075 gutirrez , j. , canal , r. , garca - berro , e. 2005 , , 435 , 231 harrison , p. a. , lyne , a. g. , & anderson , b. 1993 , , 261 , 113 hobbs , g. , lorimer , d. r. , lyne , a. g. , kramer , m. 2005 , , 360 , 974 hulse , r. a. , & taylor , j. h. 1975 , , 195 , l51 ivanova , n. , belczynski , k. , kalogera , v. , rasio , f. a. , taam , r. e. 2003 , , 592 , 475 janka , h .- th . ,
langanke , k. , marek , a. , martnez - pindeo , g. , & mller , b. 2007 , , 442 , 38 janssen , g. h. , stappers , b. w. , kramer , m. , nice , d. j. , jessner , a. , et al .
2008 , , 490 , 753 joshi , y. c. 2007 , , 378 , 768 junker , w. , & schfer , g. 1992 , , 254 , 146 kalogera , v. 1996 , , 471 , 352 kalogera , v. 2000 , , 541 , 319 kalogera , v. , & webbink , r. f. 1998 , , 493 , 351 kasian , l. in 40 years of pulsars : millisecond pulsars , magnetars and more .
aip conference proceedings , 2008 , pp .
485 - 487 kitaura , f. s. , jankka , h .- th . , & hillebrandt , w. 2006 , , 450 , 345 kiziltan b. , & thorsett s. e. 2009 , arxiv0909.1562 konacki , m. , wolszczan , a. , & stairs , i. h. 2003 , , 589 , 495 kotake , k. , sato , k. , & takahashi , k. 2006 , rep .
, 69 , 971 kuijken , k. , & gilmore , g. 1989 , , 239 , 571 lorimer , d. r. , stairs , i. h. , freire , p. c. , cordes , j. m. , camilo , f. , et al . 2006 , ,
640 , 428 lorimer , d. r. , freire , p. c. c. , stairs , i. h. , kramer , m. , mclaughlin , m. a. , et al .
2007 , , 379 , 1217 lyne , a. g. , & loimer , d. r. 1994 , , 369 , 127 lyne , a. g. , camilo , f. , manchester , r. n. , bell , j. f. , kaspi , v. m. , et al .
2000 , , 312 , 698 lyne , a. g. , burgay , m. , kramer , m. , possenti , a. , manchester , r. n. , et al .
2004 , science , 303 , 1153 mezzacappa , a. 2005 , annu .
nucl . part .
, 55 , 467 miyaji , s. , nomoto , k. , yokoi , k. , & sugimoto , d. 1980 , , 32 , 303 nomoto , k. 1987 , , 332 , 206 pfahl , e. , rappaport , s. , podsiadlowski , p. , spruit , h. 2002 , , 574 , 364 podsiadlowski , ph . ,
langer , n. , poelarends , a. j. t. , rappaport , s. , heger , a. , et al .
2004 , , 612 , 1044 podsiadlowski , ph . ,
dewi , j. d. m. , lesaffre , p. , miller , j. c. , newton , w. g. , et al .
2005 , , 361 , 1243 piran , t. , shaviv , n. j. 2004 , , 94 , 051102 reid , m. j. , menten , k. m. , zheng , x. w. , brunthaler , a. , moscadelli , l. , et al .
2009 , , 700 , 137 scheck , l. , plewa t. , janka , h .- th . ,
kifonidis , k. , & mller , e. 2004 , , 92 , 011103 scheck , l. , kifonidis , k. , janka , h .- th . , &
mller , e. 2006 , , 457 , 963 siess , l. 2007 , , 476 , 893 stairs , i. h. 2004 , science , 304 , 547 stairs , i. h. in 40 years of pulsars : millisecond pulsars , magnetars and more .
aip conference proceedings , 2008 , pp .
424 - 432 stairs , i. h. , thorsett , s. e. , & arzoumanian , z. 2004 , , 93 , 141101 stairs , i. h. , thorsett , s. e. , dewey , r. j. , kramer , m. , & mcphee , c. a. 2006 , , 373 , l50 stairs , i. h. , thorsett , s. e. , taylor , j. h. , & wolszczan , a. 2002 , , 581 , 501 tauris , t. m. , & van den heuvel e. p. j. 2004 , in compact stellar x - ray sources , ed . w. h. g. lewin & m. van der klis ( cambridge : cambridge univ . press ) , arxiv : astro - ph/0303456 taylor , j. h. , fowler , l. a. , & mcculloch , p. m. 1979 , , 277 , 437 taylor , j. h. , hulse , r. a. , fowler , l. a. , gullahorn , g. e. , & rankin , j. m. 1976 , , 206 , l53 taylor , j. h. , & weisberg , j. m. 1982 , , 253 , 908 taylor , j. h. , & weisberg , j. m. 1989 , , 345 , 434 thorsett , s. e. , & chakrabarty , d. 1999 , , 512 , 288 thorsett , s. e. , dewey , r. j. , & stairs , i. h. 2005 , , 619 , 1036 van den heuvel , e. p. j. , in the proceedings of the 5th integral workshop on the integral universe , 2004 , p. 185
van den heuvel , e. p. j. , in the multicolored landscape of compact objects and their explosive origins .
aip conference proceedings , 2007 , pp .
589 - 606 wang , c. , lai , d. , & han , j. l. 2006 , , 639 , 1007 wex , n. , kalogera , v. , & kramer , m. 2000 , , 528 , 401 willems , b. , kalogera , v. , & henninger , m. 2004 , , 616 , 414 willems , b. , henninger , m. , levin , t. , ivanova , n. , kalogera , v. , et al .
2005 , , 625 , 324 willems , b. , kaplan , j. , fragos , t. , kalogera , v. , & belczynski , k. 2006 , , 74 , 043003 wolszczan , a. 1991 , , 350 , 688 woosley , s. e. , & bloom , j. s. 2006 , , 44 , 507 | since the discovery of the first double neutron star ( dns ) system in 1975 by hulse and taylor , there are currently 8 confirmed dns in our galaxy . for every system , the masses of both neutron stars , the orbital semi - major axis and eccentricity are measured , and proper motion is known for half of the systems . using the orbital parameters and kinematic information , if available , as constraints for _ all system _ _ , we investigate the immediate progenitor mass of the second - born neutron star and the magnitude of the supernova kick it received at birth , with the primary goal to understand the core collapse mechanism leading to neutron star formation .
compared to earlier studies , we use a novel method to address the uncertainty related to the unknown radial velocity of the observed systems . for psr b1534 + 12 and psr b1913 + 16 ,
the kick magnitudes are 150 - 270 km / s and 190 - 450 km / s ( with 95% confidence ) respectively , and the progenitor masses of the 2nd born neutron stars are 1.3 - 3.4 m@xmath0 and 1.4 - 5.0 m@xmath0 ( 95% ) , respectively .
these suggest that the 2nd born neutron star was formed by an iron core collapse supernova in both systems . for psr j0737 - 3039 , on the other hand ,
the kick magnitude is only 5 - 120 km / s ( 95% ) , and the progenitor mass of the 2nd born neutron star is 1.3 - 1.9 m@xmath0 ( 95% ) . because of the relatively low progenitor mass and kick magnitude , the formation of the 2nd born neutron star in psr j0737 - 3039 is potentially connected to an electron capture supernova of a massive @xmath1 white dwarf . for
the remaining 5 galactic dns , the kick magnitude ranges from several tens to several hundreds of km / s , and the progenitor mass of the 2nd formed neutron star can be as low as @xmath21.5 m@xmath0 , or as high as @xmath28 m@xmath0 .
therefore in these systems , it is not clear which type of supernova is more likely to form the 2nd neutron star . _ _ _ |
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.